From df1828b208137c35c43aaf93d00941774d998f9b Mon Sep 17 00:00:00 2001 From: Nicolas Cannasse Date: Sun, 21 Feb 2016 13:49:42 +0100 Subject: [PATCH] added pcre16 support, TestBaseTypes passing --- hl.vcxproj | 20 +- include/pcre/config.h | 60 + include/pcre/pcre.h | 677 +++ include/pcre/pcre_chartables.c | 198 + include/pcre/pcre_compile.c | 9732 ++++++++++++++++++++++++++++++ include/pcre/pcre_dfa_exec.c | 3674 +++++++++++ include/pcre/pcre_exec.c | 7173 ++++++++++++++++++++++ include/pcre/pcre_fullinfo.c | 245 + include/pcre/pcre_globals.c | 86 + include/pcre/pcre_internal.h | 2800 +++++++++ include/pcre/pcre_newline.c | 210 + include/pcre/pcre_string_utils.c | 211 + include/pcre/pcre_tables.c | 727 +++ include/pcre/pcre_xclass.c | 268 + include/pcre/ucp.h | 224 + src/alloc.c | 6 +- src/hl.h | 17 +- src/hlc.h | 10 +- src/objdata.c | 415 -- src/std/buffer.c | 12 +- src/std/obj.c | 439 +- src/std/regexp.c | 93 +- 22 files changed, 26863 insertions(+), 434 deletions(-) create mode 100644 include/pcre/config.h create mode 100644 include/pcre/pcre.h create mode 100644 include/pcre/pcre_chartables.c create mode 100644 include/pcre/pcre_compile.c create mode 100644 include/pcre/pcre_dfa_exec.c create mode 100644 include/pcre/pcre_exec.c create mode 100644 include/pcre/pcre_fullinfo.c create mode 100644 include/pcre/pcre_globals.c create mode 100644 include/pcre/pcre_internal.h create mode 100644 include/pcre/pcre_newline.c create mode 100644 include/pcre/pcre_string_utils.c create mode 100644 include/pcre/pcre_tables.c create mode 100644 include/pcre/pcre_xclass.c create mode 100644 include/pcre/ucp.h delete mode 100644 src/objdata.c diff --git a/hl.vcxproj b/hl.vcxproj index 1cc3790b3..487eadaf7 100644 --- a/hl.vcxproj +++ b/hl.vcxproj @@ -66,21 +66,24 @@ true + include\pcre;$(IncludePath) true + include/pcre;$(IncludePath) false + include/pcre;$(IncludePath) false + include/pcre;$(IncludePath) - Level3 Disabled WIN32;_DEBUG;_CONSOLE;%(PreprocessorDefinitions) StreamingSIMDExtensions2 @@ -143,8 +146,17 @@ + + + + + + + + + + - @@ -161,6 +173,10 @@ + + + + diff --git a/include/pcre/config.h b/include/pcre/config.h new file mode 100644 index 000000000..bd5ebd5f6 --- /dev/null +++ b/include/pcre/config.h @@ -0,0 +1,60 @@ +#define COMPILE_PCRE16 +#undef SUPPORT_JIT +#define PCRE_STATIC + +#ifdef _MSC_VER +# pragma warning(disable:4710) // inline disabled +# pragma warning(disable:4711) // inline activated +# pragma warning(disable:4242) // loss of data +# pragma warning(disable:4244) // loss of data +#endif + +/* The value of PARENS_NEST_LIMIT specifies the maximum depth of nested + parentheses (of any kind) in a pattern. This limits the amount of system + stack that is used while compiling a pattern. */ +#define PARENS_NEST_LIMIT 250 + +/* The value of LINK_SIZE determines the number of bytes used to store links + as offsets within the compiled regex. The default is 2, which allows for + compiled patterns up to 64K long. This covers the vast majority of cases. + However, PCRE can also be compiled to use 3 or 4 bytes instead. This allows + for longer patterns in extreme cases. */ +#define LINK_SIZE 2 + +/* This limit is parameterized just in case anybody ever wants to change it. + Care must be taken if it is increased, because it guards against integer + overflow caused by enormously large patterns. */ +#define MAX_NAME_COUNT 10000 + +/* This limit is parameterized just in case anybody ever wants to change it. + Care must be taken if it is increased, because it guards against integer + overflow caused by enormously large patterns. */ +#define MAX_NAME_SIZE 32 + +/* The value of NEWLINE determines the default newline character sequence. + PCRE client programs can override this by selecting other values at run + time. In ASCII environments, the value can be 10 (LF), 13 (CR), or 3338 + (CRLF); in EBCDIC environments the value can be 21 or 37 (LF), 13 (CR), or + 3349 or 3365 (CRLF) because there are two alternative codepoints (0x15 and + 0x25) that are used as the NL line terminator that is equivalent to ASCII + LF. In both ASCII and EBCDIC environments the value can also be -1 (ANY), + or -2 (ANYCRLF). */ +#define NEWLINE 10 + +/* The value of MATCH_LIMIT determines the default number of times the + internal match() function can be called during a single execution of + pcre_exec(). There is a runtime interface for setting a different limit. + The limit exists in order to catch runaway regular expressions that take + for ever to determine that they do not match. The default is set very large + so that it does not accidentally catch legitimate cases. */ +#define MATCH_LIMIT 10000000 + +/* The above limit applies to all calls of match(), whether or not they + increase the recursion depth. In some environments it is desirable to limit + the depth of recursive calls of match() more strictly, in order to restrict + the maximum amount of stack (or heap, if NO_RECURSE is defined) that is + used. The value of MATCH_LIMIT_RECURSION applies only to recursive calls of + match(). To have any useful effect, it must be less than the value of + MATCH_LIMIT. The default is to use the same value as MATCH_LIMIT. There is + a runtime method for setting a different limit. */ +#define MATCH_LIMIT_RECURSION MATCH_LIMIT diff --git a/include/pcre/pcre.h b/include/pcre/pcre.h new file mode 100644 index 000000000..bf6351f88 --- /dev/null +++ b/include/pcre/pcre.h @@ -0,0 +1,677 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + +/* This is the public header file for the PCRE library, to be #included by +applications that call the PCRE functions. + + Copyright (c) 1997-2014 University of Cambridge + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + +#ifndef _PCRE_H +#define _PCRE_H + +/* The current PCRE version information. */ + +#define PCRE_MAJOR 8 +#define PCRE_MINOR 38 +#define PCRE_PRERELEASE +#define PCRE_DATE 2015-11-23 + +/* When an application links to a PCRE DLL in Windows, the symbols that are +imported have to be identified as such. When building PCRE, the appropriate +export setting is defined in pcre_internal.h, which includes this file. So we +don't change existing definitions of PCRE_EXP_DECL and PCRECPP_EXP_DECL. */ + +#if defined(_WIN32) && !defined(PCRE_STATIC) +# ifndef PCRE_EXP_DECL +# define PCRE_EXP_DECL extern __declspec(dllimport) +# endif +# ifdef __cplusplus +# ifndef PCRECPP_EXP_DECL +# define PCRECPP_EXP_DECL extern __declspec(dllimport) +# endif +# ifndef PCRECPP_EXP_DEFN +# define PCRECPP_EXP_DEFN __declspec(dllimport) +# endif +# endif +#endif + +/* By default, we use the standard "extern" declarations. */ + +#ifndef PCRE_EXP_DECL +# ifdef __cplusplus +# define PCRE_EXP_DECL extern "C" +# else +# define PCRE_EXP_DECL extern +# endif +#endif + +#ifdef __cplusplus +# ifndef PCRECPP_EXP_DECL +# define PCRECPP_EXP_DECL extern +# endif +# ifndef PCRECPP_EXP_DEFN +# define PCRECPP_EXP_DEFN +# endif +#endif + +/* Have to include stdlib.h in order to ensure that size_t is defined; +it is needed here for malloc. */ + +#include + +/* Allow for C++ users */ + +#ifdef __cplusplus +extern "C" { +#endif + +/* Public options. Some are compile-time only, some are run-time only, and some +are both. Most of the compile-time options are saved with the compiled regex so +that they can be inspected during studying (and therefore JIT compiling). Note +that pcre_study() has its own set of options. Originally, all the options +defined here used distinct bits. However, almost all the bits in a 32-bit word +are now used, so in order to conserve them, option bits that were previously +only recognized at matching time (i.e. by pcre_exec() or pcre_dfa_exec()) may +also be used for compile-time options that affect only compiling and are not +relevant for studying or JIT compiling. + +Some options for pcre_compile() change its behaviour but do not affect the +behaviour of the execution functions. Other options are passed through to the +execution functions and affect their behaviour, with or without affecting the +behaviour of pcre_compile(). + +Options that can be passed to pcre_compile() are tagged Cx below, with these +variants: + +C1 Affects compile only +C2 Does not affect compile; affects exec, dfa_exec +C3 Affects compile, exec, dfa_exec +C4 Affects compile, exec, dfa_exec, study +C5 Affects compile, exec, study + +Options that can be set for pcre_exec() and/or pcre_dfa_exec() are flagged with +E and D, respectively. They take precedence over C3, C4, and C5 settings passed +from pcre_compile(). Those that are compatible with JIT execution are flagged +with J. */ + +#define PCRE_CASELESS 0x00000001 /* C1 */ +#define PCRE_MULTILINE 0x00000002 /* C1 */ +#define PCRE_DOTALL 0x00000004 /* C1 */ +#define PCRE_EXTENDED 0x00000008 /* C1 */ +#define PCRE_ANCHORED 0x00000010 /* C4 E D */ +#define PCRE_DOLLAR_ENDONLY 0x00000020 /* C2 */ +#define PCRE_EXTRA 0x00000040 /* C1 */ +#define PCRE_NOTBOL 0x00000080 /* E D J */ +#define PCRE_NOTEOL 0x00000100 /* E D J */ +#define PCRE_UNGREEDY 0x00000200 /* C1 */ +#define PCRE_NOTEMPTY 0x00000400 /* E D J */ +#define PCRE_UTF8 0x00000800 /* C4 ) */ +#define PCRE_UTF16 0x00000800 /* C4 ) Synonyms */ +#define PCRE_UTF32 0x00000800 /* C4 ) */ +#define PCRE_NO_AUTO_CAPTURE 0x00001000 /* C1 */ +#define PCRE_NO_UTF8_CHECK 0x00002000 /* C1 E D J ) */ +#define PCRE_NO_UTF16_CHECK 0x00002000 /* C1 E D J ) Synonyms */ +#define PCRE_NO_UTF32_CHECK 0x00002000 /* C1 E D J ) */ +#define PCRE_AUTO_CALLOUT 0x00004000 /* C1 */ +#define PCRE_PARTIAL_SOFT 0x00008000 /* E D J ) Synonyms */ +#define PCRE_PARTIAL 0x00008000 /* E D J ) */ + +/* This pair use the same bit. */ +#define PCRE_NEVER_UTF 0x00010000 /* C1 ) Overlaid */ +#define PCRE_DFA_SHORTEST 0x00010000 /* D ) Overlaid */ + +/* This pair use the same bit. */ +#define PCRE_NO_AUTO_POSSESS 0x00020000 /* C1 ) Overlaid */ +#define PCRE_DFA_RESTART 0x00020000 /* D ) Overlaid */ + +#define PCRE_FIRSTLINE 0x00040000 /* C3 */ +#define PCRE_DUPNAMES 0x00080000 /* C1 */ +#define PCRE_NEWLINE_CR 0x00100000 /* C3 E D */ +#define PCRE_NEWLINE_LF 0x00200000 /* C3 E D */ +#define PCRE_NEWLINE_CRLF 0x00300000 /* C3 E D */ +#define PCRE_NEWLINE_ANY 0x00400000 /* C3 E D */ +#define PCRE_NEWLINE_ANYCRLF 0x00500000 /* C3 E D */ +#define PCRE_BSR_ANYCRLF 0x00800000 /* C3 E D */ +#define PCRE_BSR_UNICODE 0x01000000 /* C3 E D */ +#define PCRE_JAVASCRIPT_COMPAT 0x02000000 /* C5 */ +#define PCRE_NO_START_OPTIMIZE 0x04000000 /* C2 E D ) Synonyms */ +#define PCRE_NO_START_OPTIMISE 0x04000000 /* C2 E D ) */ +#define PCRE_PARTIAL_HARD 0x08000000 /* E D J */ +#define PCRE_NOTEMPTY_ATSTART 0x10000000 /* E D J */ +#define PCRE_UCP 0x20000000 /* C3 */ + +/* Exec-time and get/set-time error codes */ + +#define PCRE_ERROR_NOMATCH (-1) +#define PCRE_ERROR_NULL (-2) +#define PCRE_ERROR_BADOPTION (-3) +#define PCRE_ERROR_BADMAGIC (-4) +#define PCRE_ERROR_UNKNOWN_OPCODE (-5) +#define PCRE_ERROR_UNKNOWN_NODE (-5) /* For backward compatibility */ +#define PCRE_ERROR_NOMEMORY (-6) +#define PCRE_ERROR_NOSUBSTRING (-7) +#define PCRE_ERROR_MATCHLIMIT (-8) +#define PCRE_ERROR_CALLOUT (-9) /* Never used by PCRE itself */ +#define PCRE_ERROR_BADUTF8 (-10) /* Same for 8/16/32 */ +#define PCRE_ERROR_BADUTF16 (-10) /* Same for 8/16/32 */ +#define PCRE_ERROR_BADUTF32 (-10) /* Same for 8/16/32 */ +#define PCRE_ERROR_BADUTF8_OFFSET (-11) /* Same for 8/16 */ +#define PCRE_ERROR_BADUTF16_OFFSET (-11) /* Same for 8/16 */ +#define PCRE_ERROR_PARTIAL (-12) +#define PCRE_ERROR_BADPARTIAL (-13) +#define PCRE_ERROR_INTERNAL (-14) +#define PCRE_ERROR_BADCOUNT (-15) +#define PCRE_ERROR_DFA_UITEM (-16) +#define PCRE_ERROR_DFA_UCOND (-17) +#define PCRE_ERROR_DFA_UMLIMIT (-18) +#define PCRE_ERROR_DFA_WSSIZE (-19) +#define PCRE_ERROR_DFA_RECURSE (-20) +#define PCRE_ERROR_RECURSIONLIMIT (-21) +#define PCRE_ERROR_NULLWSLIMIT (-22) /* No longer actually used */ +#define PCRE_ERROR_BADNEWLINE (-23) +#define PCRE_ERROR_BADOFFSET (-24) +#define PCRE_ERROR_SHORTUTF8 (-25) +#define PCRE_ERROR_SHORTUTF16 (-25) /* Same for 8/16 */ +#define PCRE_ERROR_RECURSELOOP (-26) +#define PCRE_ERROR_JIT_STACKLIMIT (-27) +#define PCRE_ERROR_BADMODE (-28) +#define PCRE_ERROR_BADENDIANNESS (-29) +#define PCRE_ERROR_DFA_BADRESTART (-30) +#define PCRE_ERROR_JIT_BADOPTION (-31) +#define PCRE_ERROR_BADLENGTH (-32) +#define PCRE_ERROR_UNSET (-33) + +/* Specific error codes for UTF-8 validity checks */ + +#define PCRE_UTF8_ERR0 0 +#define PCRE_UTF8_ERR1 1 +#define PCRE_UTF8_ERR2 2 +#define PCRE_UTF8_ERR3 3 +#define PCRE_UTF8_ERR4 4 +#define PCRE_UTF8_ERR5 5 +#define PCRE_UTF8_ERR6 6 +#define PCRE_UTF8_ERR7 7 +#define PCRE_UTF8_ERR8 8 +#define PCRE_UTF8_ERR9 9 +#define PCRE_UTF8_ERR10 10 +#define PCRE_UTF8_ERR11 11 +#define PCRE_UTF8_ERR12 12 +#define PCRE_UTF8_ERR13 13 +#define PCRE_UTF8_ERR14 14 +#define PCRE_UTF8_ERR15 15 +#define PCRE_UTF8_ERR16 16 +#define PCRE_UTF8_ERR17 17 +#define PCRE_UTF8_ERR18 18 +#define PCRE_UTF8_ERR19 19 +#define PCRE_UTF8_ERR20 20 +#define PCRE_UTF8_ERR21 21 +#define PCRE_UTF8_ERR22 22 /* Unused (was non-character) */ + +/* Specific error codes for UTF-16 validity checks */ + +#define PCRE_UTF16_ERR0 0 +#define PCRE_UTF16_ERR1 1 +#define PCRE_UTF16_ERR2 2 +#define PCRE_UTF16_ERR3 3 +#define PCRE_UTF16_ERR4 4 /* Unused (was non-character) */ + +/* Specific error codes for UTF-32 validity checks */ + +#define PCRE_UTF32_ERR0 0 +#define PCRE_UTF32_ERR1 1 +#define PCRE_UTF32_ERR2 2 /* Unused (was non-character) */ +#define PCRE_UTF32_ERR3 3 + +/* Request types for pcre_fullinfo() */ + +#define PCRE_INFO_OPTIONS 0 +#define PCRE_INFO_SIZE 1 +#define PCRE_INFO_CAPTURECOUNT 2 +#define PCRE_INFO_BACKREFMAX 3 +#define PCRE_INFO_FIRSTBYTE 4 +#define PCRE_INFO_FIRSTCHAR 4 /* For backwards compatibility */ +#define PCRE_INFO_FIRSTTABLE 5 +#define PCRE_INFO_LASTLITERAL 6 +#define PCRE_INFO_NAMEENTRYSIZE 7 +#define PCRE_INFO_NAMECOUNT 8 +#define PCRE_INFO_NAMETABLE 9 +#define PCRE_INFO_STUDYSIZE 10 +#define PCRE_INFO_DEFAULT_TABLES 11 +#define PCRE_INFO_OKPARTIAL 12 +#define PCRE_INFO_JCHANGED 13 +#define PCRE_INFO_HASCRORLF 14 +#define PCRE_INFO_MINLENGTH 15 +#define PCRE_INFO_JIT 16 +#define PCRE_INFO_JITSIZE 17 +#define PCRE_INFO_MAXLOOKBEHIND 18 +#define PCRE_INFO_FIRSTCHARACTER 19 +#define PCRE_INFO_FIRSTCHARACTERFLAGS 20 +#define PCRE_INFO_REQUIREDCHAR 21 +#define PCRE_INFO_REQUIREDCHARFLAGS 22 +#define PCRE_INFO_MATCHLIMIT 23 +#define PCRE_INFO_RECURSIONLIMIT 24 +#define PCRE_INFO_MATCH_EMPTY 25 + +/* Request types for pcre_config(). Do not re-arrange, in order to remain +compatible. */ + +#define PCRE_CONFIG_UTF8 0 +#define PCRE_CONFIG_NEWLINE 1 +#define PCRE_CONFIG_LINK_SIZE 2 +#define PCRE_CONFIG_POSIX_MALLOC_THRESHOLD 3 +#define PCRE_CONFIG_MATCH_LIMIT 4 +#define PCRE_CONFIG_STACKRECURSE 5 +#define PCRE_CONFIG_UNICODE_PROPERTIES 6 +#define PCRE_CONFIG_MATCH_LIMIT_RECURSION 7 +#define PCRE_CONFIG_BSR 8 +#define PCRE_CONFIG_JIT 9 +#define PCRE_CONFIG_UTF16 10 +#define PCRE_CONFIG_JITTARGET 11 +#define PCRE_CONFIG_UTF32 12 +#define PCRE_CONFIG_PARENS_LIMIT 13 + +/* Request types for pcre_study(). Do not re-arrange, in order to remain +compatible. */ + +#define PCRE_STUDY_JIT_COMPILE 0x0001 +#define PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE 0x0002 +#define PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE 0x0004 +#define PCRE_STUDY_EXTRA_NEEDED 0x0008 + +/* Bit flags for the pcre[16|32]_extra structure. Do not re-arrange or redefine +these bits, just add new ones on the end, in order to remain compatible. */ + +#define PCRE_EXTRA_STUDY_DATA 0x0001 +#define PCRE_EXTRA_MATCH_LIMIT 0x0002 +#define PCRE_EXTRA_CALLOUT_DATA 0x0004 +#define PCRE_EXTRA_TABLES 0x0008 +#define PCRE_EXTRA_MATCH_LIMIT_RECURSION 0x0010 +#define PCRE_EXTRA_MARK 0x0020 +#define PCRE_EXTRA_EXECUTABLE_JIT 0x0040 + +/* Types */ + +struct real_pcre; /* declaration; the definition is private */ +typedef struct real_pcre pcre; + +struct real_pcre16; /* declaration; the definition is private */ +typedef struct real_pcre16 pcre16; + +struct real_pcre32; /* declaration; the definition is private */ +typedef struct real_pcre32 pcre32; + +struct real_pcre_jit_stack; /* declaration; the definition is private */ +typedef struct real_pcre_jit_stack pcre_jit_stack; + +struct real_pcre16_jit_stack; /* declaration; the definition is private */ +typedef struct real_pcre16_jit_stack pcre16_jit_stack; + +struct real_pcre32_jit_stack; /* declaration; the definition is private */ +typedef struct real_pcre32_jit_stack pcre32_jit_stack; + +/* If PCRE is compiled with 16 bit character support, PCRE_UCHAR16 must contain +a 16 bit wide signed data type. Otherwise it can be a dummy data type since +pcre16 functions are not implemented. There is a check for this in pcre_internal.h. */ +#ifndef PCRE_UCHAR16 +#define PCRE_UCHAR16 unsigned short +#endif + +#ifndef PCRE_SPTR16 +#define PCRE_SPTR16 const PCRE_UCHAR16 * +#endif + +/* If PCRE is compiled with 32 bit character support, PCRE_UCHAR32 must contain +a 32 bit wide signed data type. Otherwise it can be a dummy data type since +pcre32 functions are not implemented. There is a check for this in pcre_internal.h. */ +#ifndef PCRE_UCHAR32 +#define PCRE_UCHAR32 unsigned int +#endif + +#ifndef PCRE_SPTR32 +#define PCRE_SPTR32 const PCRE_UCHAR32 * +#endif + +/* When PCRE is compiled as a C++ library, the subject pointer type can be +replaced with a custom type. For conventional use, the public interface is a +const char *. */ + +#ifndef PCRE_SPTR +#define PCRE_SPTR const char * +#endif + +/* The structure for passing additional data to pcre_exec(). This is defined in +such as way as to be extensible. Always add new fields at the end, in order to +remain compatible. */ + +typedef struct pcre_extra { + unsigned long int flags; /* Bits for which fields are set */ + void *study_data; /* Opaque data from pcre_study() */ + unsigned long int match_limit; /* Maximum number of calls to match() */ + void *callout_data; /* Data passed back in callouts */ + const unsigned char *tables; /* Pointer to character tables */ + unsigned long int match_limit_recursion; /* Max recursive calls to match() */ + unsigned char **mark; /* For passing back a mark pointer */ + void *executable_jit; /* Contains a pointer to a compiled jit code */ +} pcre_extra; + +/* Same structure as above, but with 16 bit char pointers. */ + +typedef struct pcre16_extra { + unsigned long int flags; /* Bits for which fields are set */ + void *study_data; /* Opaque data from pcre_study() */ + unsigned long int match_limit; /* Maximum number of calls to match() */ + void *callout_data; /* Data passed back in callouts */ + const unsigned char *tables; /* Pointer to character tables */ + unsigned long int match_limit_recursion; /* Max recursive calls to match() */ + PCRE_UCHAR16 **mark; /* For passing back a mark pointer */ + void *executable_jit; /* Contains a pointer to a compiled jit code */ +} pcre16_extra; + +/* Same structure as above, but with 32 bit char pointers. */ + +typedef struct pcre32_extra { + unsigned long int flags; /* Bits for which fields are set */ + void *study_data; /* Opaque data from pcre_study() */ + unsigned long int match_limit; /* Maximum number of calls to match() */ + void *callout_data; /* Data passed back in callouts */ + const unsigned char *tables; /* Pointer to character tables */ + unsigned long int match_limit_recursion; /* Max recursive calls to match() */ + PCRE_UCHAR32 **mark; /* For passing back a mark pointer */ + void *executable_jit; /* Contains a pointer to a compiled jit code */ +} pcre32_extra; + +/* The structure for passing out data via the pcre_callout_function. We use a +structure so that new fields can be added on the end in future versions, +without changing the API of the function, thereby allowing old clients to work +without modification. */ + +typedef struct pcre_callout_block { + int version; /* Identifies version of block */ + /* ------------------------ Version 0 ------------------------------- */ + int callout_number; /* Number compiled into pattern */ + int *offset_vector; /* The offset vector */ + PCRE_SPTR subject; /* The subject being matched */ + int subject_length; /* The length of the subject */ + int start_match; /* Offset to start of this match attempt */ + int current_position; /* Where we currently are in the subject */ + int capture_top; /* Max current capture */ + int capture_last; /* Most recently closed capture */ + void *callout_data; /* Data passed in with the call */ + /* ------------------- Added for Version 1 -------------------------- */ + int pattern_position; /* Offset to next item in the pattern */ + int next_item_length; /* Length of next item in the pattern */ + /* ------------------- Added for Version 2 -------------------------- */ + const unsigned char *mark; /* Pointer to current mark or NULL */ + /* ------------------------------------------------------------------ */ +} pcre_callout_block; + +/* Same structure as above, but with 16 bit char pointers. */ + +typedef struct pcre16_callout_block { + int version; /* Identifies version of block */ + /* ------------------------ Version 0 ------------------------------- */ + int callout_number; /* Number compiled into pattern */ + int *offset_vector; /* The offset vector */ + PCRE_SPTR16 subject; /* The subject being matched */ + int subject_length; /* The length of the subject */ + int start_match; /* Offset to start of this match attempt */ + int current_position; /* Where we currently are in the subject */ + int capture_top; /* Max current capture */ + int capture_last; /* Most recently closed capture */ + void *callout_data; /* Data passed in with the call */ + /* ------------------- Added for Version 1 -------------------------- */ + int pattern_position; /* Offset to next item in the pattern */ + int next_item_length; /* Length of next item in the pattern */ + /* ------------------- Added for Version 2 -------------------------- */ + const PCRE_UCHAR16 *mark; /* Pointer to current mark or NULL */ + /* ------------------------------------------------------------------ */ +} pcre16_callout_block; + +/* Same structure as above, but with 32 bit char pointers. */ + +typedef struct pcre32_callout_block { + int version; /* Identifies version of block */ + /* ------------------------ Version 0 ------------------------------- */ + int callout_number; /* Number compiled into pattern */ + int *offset_vector; /* The offset vector */ + PCRE_SPTR32 subject; /* The subject being matched */ + int subject_length; /* The length of the subject */ + int start_match; /* Offset to start of this match attempt */ + int current_position; /* Where we currently are in the subject */ + int capture_top; /* Max current capture */ + int capture_last; /* Most recently closed capture */ + void *callout_data; /* Data passed in with the call */ + /* ------------------- Added for Version 1 -------------------------- */ + int pattern_position; /* Offset to next item in the pattern */ + int next_item_length; /* Length of next item in the pattern */ + /* ------------------- Added for Version 2 -------------------------- */ + const PCRE_UCHAR32 *mark; /* Pointer to current mark or NULL */ + /* ------------------------------------------------------------------ */ +} pcre32_callout_block; + +/* Indirection for store get and free functions. These can be set to +alternative malloc/free functions if required. Special ones are used in the +non-recursive case for "frames". There is also an optional callout function +that is triggered by the (?) regex item. For Virtual Pascal, these definitions +have to take another form. */ + +#ifndef VPCOMPAT +PCRE_EXP_DECL void *(*pcre_malloc)(size_t); +PCRE_EXP_DECL void (*pcre_free)(void *); +PCRE_EXP_DECL void *(*pcre_stack_malloc)(size_t); +PCRE_EXP_DECL void (*pcre_stack_free)(void *); +PCRE_EXP_DECL int (*pcre_callout)(pcre_callout_block *); +PCRE_EXP_DECL int (*pcre_stack_guard)(void); + +PCRE_EXP_DECL void *(*pcre16_malloc)(size_t); +PCRE_EXP_DECL void (*pcre16_free)(void *); +PCRE_EXP_DECL void *(*pcre16_stack_malloc)(size_t); +PCRE_EXP_DECL void (*pcre16_stack_free)(void *); +PCRE_EXP_DECL int (*pcre16_callout)(pcre16_callout_block *); +PCRE_EXP_DECL int (*pcre16_stack_guard)(void); + +PCRE_EXP_DECL void *(*pcre32_malloc)(size_t); +PCRE_EXP_DECL void (*pcre32_free)(void *); +PCRE_EXP_DECL void *(*pcre32_stack_malloc)(size_t); +PCRE_EXP_DECL void (*pcre32_stack_free)(void *); +PCRE_EXP_DECL int (*pcre32_callout)(pcre32_callout_block *); +PCRE_EXP_DECL int (*pcre32_stack_guard)(void); +#else /* VPCOMPAT */ +PCRE_EXP_DECL void *pcre_malloc(size_t); +PCRE_EXP_DECL void pcre_free(void *); +PCRE_EXP_DECL void *pcre_stack_malloc(size_t); +PCRE_EXP_DECL void pcre_stack_free(void *); +PCRE_EXP_DECL int pcre_callout(pcre_callout_block *); +PCRE_EXP_DECL int pcre_stack_guard(void); + +PCRE_EXP_DECL void *pcre16_malloc(size_t); +PCRE_EXP_DECL void pcre16_free(void *); +PCRE_EXP_DECL void *pcre16_stack_malloc(size_t); +PCRE_EXP_DECL void pcre16_stack_free(void *); +PCRE_EXP_DECL int pcre16_callout(pcre16_callout_block *); +PCRE_EXP_DECL int pcre16_stack_guard(void); + +PCRE_EXP_DECL void *pcre32_malloc(size_t); +PCRE_EXP_DECL void pcre32_free(void *); +PCRE_EXP_DECL void *pcre32_stack_malloc(size_t); +PCRE_EXP_DECL void pcre32_stack_free(void *); +PCRE_EXP_DECL int pcre32_callout(pcre32_callout_block *); +PCRE_EXP_DECL int pcre32_stack_guard(void); +#endif /* VPCOMPAT */ + +/* User defined callback which provides a stack just before the match starts. */ + +typedef pcre_jit_stack *(*pcre_jit_callback)(void *); +typedef pcre16_jit_stack *(*pcre16_jit_callback)(void *); +typedef pcre32_jit_stack *(*pcre32_jit_callback)(void *); + +/* Exported PCRE functions */ + +PCRE_EXP_DECL pcre *pcre_compile(const char *, int, const char **, int *, + const unsigned char *); +PCRE_EXP_DECL pcre16 *pcre16_compile(PCRE_SPTR16, int, const char **, int *, + const unsigned char *); +PCRE_EXP_DECL pcre32 *pcre32_compile(PCRE_SPTR32, int, const char **, int *, + const unsigned char *); +PCRE_EXP_DECL pcre *pcre_compile2(const char *, int, int *, const char **, + int *, const unsigned char *); +PCRE_EXP_DECL pcre16 *pcre16_compile2(PCRE_SPTR16, int, int *, const char **, + int *, const unsigned char *); +PCRE_EXP_DECL pcre32 *pcre32_compile2(PCRE_SPTR32, int, int *, const char **, + int *, const unsigned char *); +PCRE_EXP_DECL int pcre_config(int, void *); +PCRE_EXP_DECL int pcre16_config(int, void *); +PCRE_EXP_DECL int pcre32_config(int, void *); +PCRE_EXP_DECL int pcre_copy_named_substring(const pcre *, const char *, + int *, int, const char *, char *, int); +PCRE_EXP_DECL int pcre16_copy_named_substring(const pcre16 *, PCRE_SPTR16, + int *, int, PCRE_SPTR16, PCRE_UCHAR16 *, int); +PCRE_EXP_DECL int pcre32_copy_named_substring(const pcre32 *, PCRE_SPTR32, + int *, int, PCRE_SPTR32, PCRE_UCHAR32 *, int); +PCRE_EXP_DECL int pcre_copy_substring(const char *, int *, int, int, + char *, int); +PCRE_EXP_DECL int pcre16_copy_substring(PCRE_SPTR16, int *, int, int, + PCRE_UCHAR16 *, int); +PCRE_EXP_DECL int pcre32_copy_substring(PCRE_SPTR32, int *, int, int, + PCRE_UCHAR32 *, int); +PCRE_EXP_DECL int pcre_dfa_exec(const pcre *, const pcre_extra *, + const char *, int, int, int, int *, int , int *, int); +PCRE_EXP_DECL int pcre16_dfa_exec(const pcre16 *, const pcre16_extra *, + PCRE_SPTR16, int, int, int, int *, int , int *, int); +PCRE_EXP_DECL int pcre32_dfa_exec(const pcre32 *, const pcre32_extra *, + PCRE_SPTR32, int, int, int, int *, int , int *, int); +PCRE_EXP_DECL int pcre_exec(const pcre *, const pcre_extra *, PCRE_SPTR, + int, int, int, int *, int); +PCRE_EXP_DECL int pcre16_exec(const pcre16 *, const pcre16_extra *, + PCRE_SPTR16, int, int, int, int *, int); +PCRE_EXP_DECL int pcre32_exec(const pcre32 *, const pcre32_extra *, + PCRE_SPTR32, int, int, int, int *, int); +PCRE_EXP_DECL int pcre_jit_exec(const pcre *, const pcre_extra *, + PCRE_SPTR, int, int, int, int *, int, + pcre_jit_stack *); +PCRE_EXP_DECL int pcre16_jit_exec(const pcre16 *, const pcre16_extra *, + PCRE_SPTR16, int, int, int, int *, int, + pcre16_jit_stack *); +PCRE_EXP_DECL int pcre32_jit_exec(const pcre32 *, const pcre32_extra *, + PCRE_SPTR32, int, int, int, int *, int, + pcre32_jit_stack *); +PCRE_EXP_DECL void pcre_free_substring(const char *); +PCRE_EXP_DECL void pcre16_free_substring(PCRE_SPTR16); +PCRE_EXP_DECL void pcre32_free_substring(PCRE_SPTR32); +PCRE_EXP_DECL void pcre_free_substring_list(const char **); +PCRE_EXP_DECL void pcre16_free_substring_list(PCRE_SPTR16 *); +PCRE_EXP_DECL void pcre32_free_substring_list(PCRE_SPTR32 *); +PCRE_EXP_DECL int pcre_fullinfo(const pcre *, const pcre_extra *, int, + void *); +PCRE_EXP_DECL int pcre16_fullinfo(const pcre16 *, const pcre16_extra *, int, + void *); +PCRE_EXP_DECL int pcre32_fullinfo(const pcre32 *, const pcre32_extra *, int, + void *); +PCRE_EXP_DECL int pcre_get_named_substring(const pcre *, const char *, + int *, int, const char *, const char **); +PCRE_EXP_DECL int pcre16_get_named_substring(const pcre16 *, PCRE_SPTR16, + int *, int, PCRE_SPTR16, PCRE_SPTR16 *); +PCRE_EXP_DECL int pcre32_get_named_substring(const pcre32 *, PCRE_SPTR32, + int *, int, PCRE_SPTR32, PCRE_SPTR32 *); +PCRE_EXP_DECL int pcre_get_stringnumber(const pcre *, const char *); +PCRE_EXP_DECL int pcre16_get_stringnumber(const pcre16 *, PCRE_SPTR16); +PCRE_EXP_DECL int pcre32_get_stringnumber(const pcre32 *, PCRE_SPTR32); +PCRE_EXP_DECL int pcre_get_stringtable_entries(const pcre *, const char *, + char **, char **); +PCRE_EXP_DECL int pcre16_get_stringtable_entries(const pcre16 *, PCRE_SPTR16, + PCRE_UCHAR16 **, PCRE_UCHAR16 **); +PCRE_EXP_DECL int pcre32_get_stringtable_entries(const pcre32 *, PCRE_SPTR32, + PCRE_UCHAR32 **, PCRE_UCHAR32 **); +PCRE_EXP_DECL int pcre_get_substring(const char *, int *, int, int, + const char **); +PCRE_EXP_DECL int pcre16_get_substring(PCRE_SPTR16, int *, int, int, + PCRE_SPTR16 *); +PCRE_EXP_DECL int pcre32_get_substring(PCRE_SPTR32, int *, int, int, + PCRE_SPTR32 *); +PCRE_EXP_DECL int pcre_get_substring_list(const char *, int *, int, + const char ***); +PCRE_EXP_DECL int pcre16_get_substring_list(PCRE_SPTR16, int *, int, + PCRE_SPTR16 **); +PCRE_EXP_DECL int pcre32_get_substring_list(PCRE_SPTR32, int *, int, + PCRE_SPTR32 **); +PCRE_EXP_DECL const unsigned char *pcre_maketables(void); +PCRE_EXP_DECL const unsigned char *pcre16_maketables(void); +PCRE_EXP_DECL const unsigned char *pcre32_maketables(void); +PCRE_EXP_DECL int pcre_refcount(pcre *, int); +PCRE_EXP_DECL int pcre16_refcount(pcre16 *, int); +PCRE_EXP_DECL int pcre32_refcount(pcre32 *, int); +PCRE_EXP_DECL pcre_extra *pcre_study(const pcre *, int, const char **); +PCRE_EXP_DECL pcre16_extra *pcre16_study(const pcre16 *, int, const char **); +PCRE_EXP_DECL pcre32_extra *pcre32_study(const pcre32 *, int, const char **); +PCRE_EXP_DECL void pcre_free_study(pcre_extra *); +PCRE_EXP_DECL void pcre16_free_study(pcre16_extra *); +PCRE_EXP_DECL void pcre32_free_study(pcre32_extra *); +PCRE_EXP_DECL const char *pcre_version(void); +PCRE_EXP_DECL const char *pcre16_version(void); +PCRE_EXP_DECL const char *pcre32_version(void); + +/* Utility functions for byte order swaps. */ +PCRE_EXP_DECL int pcre_pattern_to_host_byte_order(pcre *, pcre_extra *, + const unsigned char *); +PCRE_EXP_DECL int pcre16_pattern_to_host_byte_order(pcre16 *, pcre16_extra *, + const unsigned char *); +PCRE_EXP_DECL int pcre32_pattern_to_host_byte_order(pcre32 *, pcre32_extra *, + const unsigned char *); +PCRE_EXP_DECL int pcre16_utf16_to_host_byte_order(PCRE_UCHAR16 *, + PCRE_SPTR16, int, int *, int); +PCRE_EXP_DECL int pcre32_utf32_to_host_byte_order(PCRE_UCHAR32 *, + PCRE_SPTR32, int, int *, int); + +/* JIT compiler related functions. */ + +PCRE_EXP_DECL pcre_jit_stack *pcre_jit_stack_alloc(int, int); +PCRE_EXP_DECL pcre16_jit_stack *pcre16_jit_stack_alloc(int, int); +PCRE_EXP_DECL pcre32_jit_stack *pcre32_jit_stack_alloc(int, int); +PCRE_EXP_DECL void pcre_jit_stack_free(pcre_jit_stack *); +PCRE_EXP_DECL void pcre16_jit_stack_free(pcre16_jit_stack *); +PCRE_EXP_DECL void pcre32_jit_stack_free(pcre32_jit_stack *); +PCRE_EXP_DECL void pcre_assign_jit_stack(pcre_extra *, + pcre_jit_callback, void *); +PCRE_EXP_DECL void pcre16_assign_jit_stack(pcre16_extra *, + pcre16_jit_callback, void *); +PCRE_EXP_DECL void pcre32_assign_jit_stack(pcre32_extra *, + pcre32_jit_callback, void *); +PCRE_EXP_DECL void pcre_jit_free_unused_memory(void); +PCRE_EXP_DECL void pcre16_jit_free_unused_memory(void); +PCRE_EXP_DECL void pcre32_jit_free_unused_memory(void); + +#ifdef __cplusplus +} /* extern "C" */ +#endif + +#endif /* End of pcre.h */ diff --git a/include/pcre/pcre_chartables.c b/include/pcre/pcre_chartables.c new file mode 100644 index 000000000..1e20ec29d --- /dev/null +++ b/include/pcre/pcre_chartables.c @@ -0,0 +1,198 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + +/* This file contains character tables that are used when no external tables +are passed to PCRE by the application that calls it. The tables are used only +for characters whose code values are less than 256. + +This is a default version of the tables that assumes ASCII encoding. A program +called dftables (which is distributed with PCRE) can be used to build +alternative versions of this file. This is necessary if you are running in an +EBCDIC environment, or if you want to default to a different encoding, for +example ISO-8859-1. When dftables is run, it creates these tables in the +current locale. If PCRE is configured with --enable-rebuild-chartables, this +happens automatically. + +The following #includes are present because without them gcc 4.x may remove the +array definition from the final binary if PCRE is built into a static library +and dead code stripping is activated. This leads to link errors. Pulling in the +header ensures that the array gets flagged as "someone outside this compilation +unit might reference this" and so it will always be supplied to the linker. */ + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + +const pcre_uint8 PRIV(default_tables)[] = { + +/* This table is a lower casing table. */ + + 0, 1, 2, 3, 4, 5, 6, 7, + 8, 9, 10, 11, 12, 13, 14, 15, + 16, 17, 18, 19, 20, 21, 22, 23, + 24, 25, 26, 27, 28, 29, 30, 31, + 32, 33, 34, 35, 36, 37, 38, 39, + 40, 41, 42, 43, 44, 45, 46, 47, + 48, 49, 50, 51, 52, 53, 54, 55, + 56, 57, 58, 59, 60, 61, 62, 63, + 64, 97, 98, 99,100,101,102,103, + 104,105,106,107,108,109,110,111, + 112,113,114,115,116,117,118,119, + 120,121,122, 91, 92, 93, 94, 95, + 96, 97, 98, 99,100,101,102,103, + 104,105,106,107,108,109,110,111, + 112,113,114,115,116,117,118,119, + 120,121,122,123,124,125,126,127, + 128,129,130,131,132,133,134,135, + 136,137,138,139,140,141,142,143, + 144,145,146,147,148,149,150,151, + 152,153,154,155,156,157,158,159, + 160,161,162,163,164,165,166,167, + 168,169,170,171,172,173,174,175, + 176,177,178,179,180,181,182,183, + 184,185,186,187,188,189,190,191, + 192,193,194,195,196,197,198,199, + 200,201,202,203,204,205,206,207, + 208,209,210,211,212,213,214,215, + 216,217,218,219,220,221,222,223, + 224,225,226,227,228,229,230,231, + 232,233,234,235,236,237,238,239, + 240,241,242,243,244,245,246,247, + 248,249,250,251,252,253,254,255, + +/* This table is a case flipping table. */ + + 0, 1, 2, 3, 4, 5, 6, 7, + 8, 9, 10, 11, 12, 13, 14, 15, + 16, 17, 18, 19, 20, 21, 22, 23, + 24, 25, 26, 27, 28, 29, 30, 31, + 32, 33, 34, 35, 36, 37, 38, 39, + 40, 41, 42, 43, 44, 45, 46, 47, + 48, 49, 50, 51, 52, 53, 54, 55, + 56, 57, 58, 59, 60, 61, 62, 63, + 64, 97, 98, 99,100,101,102,103, + 104,105,106,107,108,109,110,111, + 112,113,114,115,116,117,118,119, + 120,121,122, 91, 92, 93, 94, 95, + 96, 65, 66, 67, 68, 69, 70, 71, + 72, 73, 74, 75, 76, 77, 78, 79, + 80, 81, 82, 83, 84, 85, 86, 87, + 88, 89, 90,123,124,125,126,127, + 128,129,130,131,132,133,134,135, + 136,137,138,139,140,141,142,143, + 144,145,146,147,148,149,150,151, + 152,153,154,155,156,157,158,159, + 160,161,162,163,164,165,166,167, + 168,169,170,171,172,173,174,175, + 176,177,178,179,180,181,182,183, + 184,185,186,187,188,189,190,191, + 192,193,194,195,196,197,198,199, + 200,201,202,203,204,205,206,207, + 208,209,210,211,212,213,214,215, + 216,217,218,219,220,221,222,223, + 224,225,226,227,228,229,230,231, + 232,233,234,235,236,237,238,239, + 240,241,242,243,244,245,246,247, + 248,249,250,251,252,253,254,255, + +/* This table contains bit maps for various character classes. Each map is 32 +bytes long and the bits run from the least significant end of each byte. The +classes that have their own maps are: space, xdigit, digit, upper, lower, word, +graph, print, punct, and cntrl. Other classes are built from combinations. */ + + 0x00,0x3e,0x00,0x00,0x01,0x00,0x00,0x00, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + + 0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03, + 0x7e,0x00,0x00,0x00,0x7e,0x00,0x00,0x00, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + + 0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + 0xfe,0xff,0xff,0x07,0x00,0x00,0x00,0x00, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + 0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0x07, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + + 0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03, + 0xfe,0xff,0xff,0x87,0xfe,0xff,0xff,0x07, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + + 0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0xff, + 0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + + 0x00,0x00,0x00,0x00,0xff,0xff,0xff,0xff, + 0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + + 0x00,0x00,0x00,0x00,0xfe,0xff,0x00,0xfc, + 0x01,0x00,0x00,0xf8,0x01,0x00,0x00,0x78, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + + 0xff,0xff,0xff,0xff,0x00,0x00,0x00,0x00, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x80, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + +/* This table identifies various classes of character by individual bits: + 0x01 white space character + 0x02 letter + 0x04 decimal digit + 0x08 hexadecimal digit + 0x10 alphanumeric or '_' + 0x80 regular expression metacharacter or binary zero +*/ + + 0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 0- 7 */ + 0x00,0x01,0x01,0x01,0x01,0x01,0x00,0x00, /* 8- 15 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 16- 23 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */ + 0x01,0x00,0x00,0x00,0x80,0x00,0x00,0x00, /* - ' */ + 0x80,0x80,0x80,0x80,0x00,0x00,0x80,0x00, /* ( - / */ + 0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /* 0 - 7 */ + 0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x80, /* 8 - ? */ + 0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /* @ - G */ + 0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* H - O */ + 0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* P - W */ + 0x12,0x12,0x12,0x80,0x80,0x00,0x80,0x10, /* X - _ */ + 0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /* ` - g */ + 0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* h - o */ + 0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* p - w */ + 0x12,0x12,0x12,0x80,0x80,0x00,0x00,0x00, /* x -127 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */ + +/* End of pcre_chartables.c */ diff --git a/include/pcre/pcre_compile.c b/include/pcre/pcre_compile.c new file mode 100644 index 000000000..4d3b3139d --- /dev/null +++ b/include/pcre/pcre_compile.c @@ -0,0 +1,9732 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + +/* PCRE is a library of functions to support regular expressions whose syntax +and semantics are as close as possible to those of the Perl 5 language. + + Written by Philip Hazel + Copyright (c) 1997-2014 University of Cambridge + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + + +/* This module contains the external function pcre_compile(), along with +supporting internal functions that are not used by other modules. */ + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#define NLBLOCK cd /* Block containing newline information */ +#define PSSTART start_pattern /* Field containing pattern start */ +#define PSEND end_pattern /* Field containing pattern end */ + +#include "pcre_internal.h" + + +/* When PCRE_DEBUG is defined, we need the pcre(16|32)_printint() function, which +is also used by pcretest. PCRE_DEBUG is not defined when building a production +library. We do not need to select pcre16_printint.c specially, because the +COMPILE_PCREx macro will already be appropriately set. */ + +#ifdef PCRE_DEBUG +/* pcre_printint.c should not include any headers */ +#define PCRE_INCLUDED +#include "pcre_printint.c" +#undef PCRE_INCLUDED +#endif + + +/* Macro for setting individual bits in class bitmaps. */ + +#define SETBIT(a,b) a[(b)/8] |= (1 << ((b)&7)) + +/* Maximum length value to check against when making sure that the integer that +holds the compiled pattern length does not overflow. We make it a bit less than +INT_MAX to allow for adding in group terminating bytes, so that we don't have +to check them every time. */ + +#define OFLOW_MAX (INT_MAX - 20) + +/* Definitions to allow mutual recursion */ + +static int + add_list_to_class(pcre_uint8 *, pcre_uchar **, int, compile_data *, + const pcre_uint32 *, unsigned int); + +static BOOL + compile_regex(int, pcre_uchar **, const pcre_uchar **, int *, BOOL, BOOL, int, int, + pcre_uint32 *, pcre_int32 *, pcre_uint32 *, pcre_int32 *, branch_chain *, + compile_data *, int *); + + + +/************************************************* +* Code parameters and static tables * +*************************************************/ + +/* This value specifies the size of stack workspace that is used during the +first pre-compile phase that determines how much memory is required. The regex +is partly compiled into this space, but the compiled parts are discarded as +soon as they can be, so that hopefully there will never be an overrun. The code +does, however, check for an overrun. The largest amount I've seen used is 218, +so this number is very generous. + +The same workspace is used during the second, actual compile phase for +remembering forward references to groups so that they can be filled in at the +end. Each entry in this list occupies LINK_SIZE bytes, so even when LINK_SIZE +is 4 there is plenty of room for most patterns. However, the memory can get +filled up by repetitions of forward references, for example patterns like +/(?1){0,1999}(b)/, and one user did hit the limit. The code has been changed so +that the workspace is expanded using malloc() in this situation. The value +below is therefore a minimum, and we put a maximum on it for safety. The +minimum is now also defined in terms of LINK_SIZE so that the use of malloc() +kicks in at the same number of forward references in all cases. */ + +#define COMPILE_WORK_SIZE (2048*LINK_SIZE) +#define COMPILE_WORK_SIZE_MAX (100*COMPILE_WORK_SIZE) + +/* This value determines the size of the initial vector that is used for +remembering named groups during the pre-compile. It is allocated on the stack, +but if it is too small, it is expanded using malloc(), in a similar way to the +workspace. The value is the number of slots in the list. */ + +#define NAMED_GROUP_LIST_SIZE 20 + +/* The overrun tests check for a slightly smaller size so that they detect the +overrun before it actually does run off the end of the data block. */ + +#define WORK_SIZE_SAFETY_MARGIN (100) + +/* Private flags added to firstchar and reqchar. */ + +#define REQ_CASELESS (1 << 0) /* Indicates caselessness */ +#define REQ_VARY (1 << 1) /* Reqchar followed non-literal item */ +/* Negative values for the firstchar and reqchar flags */ +#define REQ_UNSET (-2) +#define REQ_NONE (-1) + +/* Repeated character flags. */ + +#define UTF_LENGTH 0x10000000l /* The char contains its length. */ + +/* Table for handling escaped characters in the range '0'-'z'. Positive returns +are simple data values; negative values are for special things like \d and so +on. Zero means further processing is needed (for things like \x), or the escape +is invalid. */ + +#ifndef EBCDIC + +/* This is the "normal" table for ASCII systems or for EBCDIC systems running +in UTF-8 mode. */ + +static const short int escapes[] = { + 0, 0, + 0, 0, + 0, 0, + 0, 0, + 0, 0, + CHAR_COLON, CHAR_SEMICOLON, + CHAR_LESS_THAN_SIGN, CHAR_EQUALS_SIGN, + CHAR_GREATER_THAN_SIGN, CHAR_QUESTION_MARK, + CHAR_COMMERCIAL_AT, -ESC_A, + -ESC_B, -ESC_C, + -ESC_D, -ESC_E, + 0, -ESC_G, + -ESC_H, 0, + 0, -ESC_K, + 0, 0, + -ESC_N, 0, + -ESC_P, -ESC_Q, + -ESC_R, -ESC_S, + 0, 0, + -ESC_V, -ESC_W, + -ESC_X, 0, + -ESC_Z, CHAR_LEFT_SQUARE_BRACKET, + CHAR_BACKSLASH, CHAR_RIGHT_SQUARE_BRACKET, + CHAR_CIRCUMFLEX_ACCENT, CHAR_UNDERSCORE, + CHAR_GRAVE_ACCENT, ESC_a, + -ESC_b, 0, + -ESC_d, ESC_e, + ESC_f, 0, + -ESC_h, 0, + 0, -ESC_k, + 0, 0, + ESC_n, 0, + -ESC_p, 0, + ESC_r, -ESC_s, + ESC_tee, 0, + -ESC_v, -ESC_w, + 0, 0, + -ESC_z +}; + +#else + +/* This is the "abnormal" table for EBCDIC systems without UTF-8 support. */ + +static const short int escapes[] = { +/* 48 */ 0, 0, 0, '.', '<', '(', '+', '|', +/* 50 */ '&', 0, 0, 0, 0, 0, 0, 0, +/* 58 */ 0, 0, '!', '$', '*', ')', ';', '~', +/* 60 */ '-', '/', 0, 0, 0, 0, 0, 0, +/* 68 */ 0, 0, '|', ',', '%', '_', '>', '?', +/* 70 */ 0, 0, 0, 0, 0, 0, 0, 0, +/* 78 */ 0, '`', ':', '#', '@', '\'', '=', '"', +/* 80 */ 0, ESC_a, -ESC_b, 0, -ESC_d, ESC_e, ESC_f, 0, +/* 88 */-ESC_h, 0, 0, '{', 0, 0, 0, 0, +/* 90 */ 0, 0, -ESC_k, 0, 0, ESC_n, 0, -ESC_p, +/* 98 */ 0, ESC_r, 0, '}', 0, 0, 0, 0, +/* A0 */ 0, '~', -ESC_s, ESC_tee, 0,-ESC_v, -ESC_w, 0, +/* A8 */ 0,-ESC_z, 0, 0, 0, '[', 0, 0, +/* B0 */ 0, 0, 0, 0, 0, 0, 0, 0, +/* B8 */ 0, 0, 0, 0, 0, ']', '=', '-', +/* C0 */ '{',-ESC_A, -ESC_B, -ESC_C, -ESC_D,-ESC_E, 0, -ESC_G, +/* C8 */-ESC_H, 0, 0, 0, 0, 0, 0, 0, +/* D0 */ '}', 0, -ESC_K, 0, 0,-ESC_N, 0, -ESC_P, +/* D8 */-ESC_Q,-ESC_R, 0, 0, 0, 0, 0, 0, +/* E0 */ '\\', 0, -ESC_S, 0, 0,-ESC_V, -ESC_W, -ESC_X, +/* E8 */ 0,-ESC_Z, 0, 0, 0, 0, 0, 0, +/* F0 */ 0, 0, 0, 0, 0, 0, 0, 0, +/* F8 */ 0, 0, 0, 0, 0, 0, 0, 0 +}; + +/* We also need a table of characters that may follow \c in an EBCDIC +environment for characters 0-31. */ + +static unsigned char ebcdic_escape_c[] = "@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_"; + +#endif + + +/* Table of special "verbs" like (*PRUNE). This is a short table, so it is +searched linearly. Put all the names into a single string, in order to reduce +the number of relocations when a shared library is dynamically linked. The +string is built from string macros so that it works in UTF-8 mode on EBCDIC +platforms. */ + +typedef struct verbitem { + int len; /* Length of verb name */ + int op; /* Op when no arg, or -1 if arg mandatory */ + int op_arg; /* Op when arg present, or -1 if not allowed */ +} verbitem; + +static const char verbnames[] = + "\0" /* Empty name is a shorthand for MARK */ + STRING_MARK0 + STRING_ACCEPT0 + STRING_COMMIT0 + STRING_F0 + STRING_FAIL0 + STRING_PRUNE0 + STRING_SKIP0 + STRING_THEN; + +static const verbitem verbs[] = { + { 0, -1, OP_MARK }, + { 4, -1, OP_MARK }, + { 6, OP_ACCEPT, -1 }, + { 6, OP_COMMIT, -1 }, + { 1, OP_FAIL, -1 }, + { 4, OP_FAIL, -1 }, + { 5, OP_PRUNE, OP_PRUNE_ARG }, + { 4, OP_SKIP, OP_SKIP_ARG }, + { 4, OP_THEN, OP_THEN_ARG } +}; + +static const int verbcount = sizeof(verbs)/sizeof(verbitem); + + +/* Substitutes for [[:<:]] and [[:>:]], which mean start and end of word in +another regex library. */ + +static const pcre_uchar sub_start_of_word[] = { + CHAR_BACKSLASH, CHAR_b, CHAR_LEFT_PARENTHESIS, CHAR_QUESTION_MARK, + CHAR_EQUALS_SIGN, CHAR_BACKSLASH, CHAR_w, CHAR_RIGHT_PARENTHESIS, '\0' }; + +static const pcre_uchar sub_end_of_word[] = { + CHAR_BACKSLASH, CHAR_b, CHAR_LEFT_PARENTHESIS, CHAR_QUESTION_MARK, + CHAR_LESS_THAN_SIGN, CHAR_EQUALS_SIGN, CHAR_BACKSLASH, CHAR_w, + CHAR_RIGHT_PARENTHESIS, '\0' }; + + +/* Tables of names of POSIX character classes and their lengths. The names are +now all in a single string, to reduce the number of relocations when a shared +library is dynamically loaded. The list of lengths is terminated by a zero +length entry. The first three must be alpha, lower, upper, as this is assumed +for handling case independence. The indices for graph, print, and punct are +needed, so identify them. */ + +static const char posix_names[] = + STRING_alpha0 STRING_lower0 STRING_upper0 STRING_alnum0 + STRING_ascii0 STRING_blank0 STRING_cntrl0 STRING_digit0 + STRING_graph0 STRING_print0 STRING_punct0 STRING_space0 + STRING_word0 STRING_xdigit; + +static const pcre_uint8 posix_name_lengths[] = { + 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 6, 0 }; + +#define PC_GRAPH 8 +#define PC_PRINT 9 +#define PC_PUNCT 10 + + +/* Table of class bit maps for each POSIX class. Each class is formed from a +base map, with an optional addition or removal of another map. Then, for some +classes, there is some additional tweaking: for [:blank:] the vertical space +characters are removed, and for [:alpha:] and [:alnum:] the underscore +character is removed. The triples in the table consist of the base map offset, +second map offset or -1 if no second map, and a non-negative value for map +addition or a negative value for map subtraction (if there are two maps). The +absolute value of the third field has these meanings: 0 => no tweaking, 1 => +remove vertical space characters, 2 => remove underscore. */ + +static const int posix_class_maps[] = { + cbit_word, cbit_digit, -2, /* alpha */ + cbit_lower, -1, 0, /* lower */ + cbit_upper, -1, 0, /* upper */ + cbit_word, -1, 2, /* alnum - word without underscore */ + cbit_print, cbit_cntrl, 0, /* ascii */ + cbit_space, -1, 1, /* blank - a GNU extension */ + cbit_cntrl, -1, 0, /* cntrl */ + cbit_digit, -1, 0, /* digit */ + cbit_graph, -1, 0, /* graph */ + cbit_print, -1, 0, /* print */ + cbit_punct, -1, 0, /* punct */ + cbit_space, -1, 0, /* space */ + cbit_word, -1, 0, /* word - a Perl extension */ + cbit_xdigit,-1, 0 /* xdigit */ +}; + +/* Table of substitutes for \d etc when PCRE_UCP is set. They are replaced by +Unicode property escapes. */ + +#ifdef SUPPORT_UCP +static const pcre_uchar string_PNd[] = { + CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET, + CHAR_N, CHAR_d, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_pNd[] = { + CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET, + CHAR_N, CHAR_d, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_PXsp[] = { + CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET, + CHAR_X, CHAR_s, CHAR_p, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_pXsp[] = { + CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET, + CHAR_X, CHAR_s, CHAR_p, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_PXwd[] = { + CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET, + CHAR_X, CHAR_w, CHAR_d, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_pXwd[] = { + CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET, + CHAR_X, CHAR_w, CHAR_d, CHAR_RIGHT_CURLY_BRACKET, '\0' }; + +static const pcre_uchar *substitutes[] = { + string_PNd, /* \D */ + string_pNd, /* \d */ + string_PXsp, /* \S */ /* Xsp is Perl space, but from 8.34, Perl */ + string_pXsp, /* \s */ /* space and POSIX space are the same. */ + string_PXwd, /* \W */ + string_pXwd /* \w */ +}; + +/* The POSIX class substitutes must be in the order of the POSIX class names, +defined above, and there are both positive and negative cases. NULL means no +general substitute of a Unicode property escape (\p or \P). However, for some +POSIX classes (e.g. graph, print, punct) a special property code is compiled +directly. */ + +static const pcre_uchar string_pL[] = { + CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET, + CHAR_L, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_pLl[] = { + CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET, + CHAR_L, CHAR_l, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_pLu[] = { + CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET, + CHAR_L, CHAR_u, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_pXan[] = { + CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET, + CHAR_X, CHAR_a, CHAR_n, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_h[] = { + CHAR_BACKSLASH, CHAR_h, '\0' }; +static const pcre_uchar string_pXps[] = { + CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET, + CHAR_X, CHAR_p, CHAR_s, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_PL[] = { + CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET, + CHAR_L, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_PLl[] = { + CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET, + CHAR_L, CHAR_l, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_PLu[] = { + CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET, + CHAR_L, CHAR_u, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_PXan[] = { + CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET, + CHAR_X, CHAR_a, CHAR_n, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_H[] = { + CHAR_BACKSLASH, CHAR_H, '\0' }; +static const pcre_uchar string_PXps[] = { + CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET, + CHAR_X, CHAR_p, CHAR_s, CHAR_RIGHT_CURLY_BRACKET, '\0' }; + +static const pcre_uchar *posix_substitutes[] = { + string_pL, /* alpha */ + string_pLl, /* lower */ + string_pLu, /* upper */ + string_pXan, /* alnum */ + NULL, /* ascii */ + string_h, /* blank */ + NULL, /* cntrl */ + string_pNd, /* digit */ + NULL, /* graph */ + NULL, /* print */ + NULL, /* punct */ + string_pXps, /* space */ /* Xps is POSIX space, but from 8.34 */ + string_pXwd, /* word */ /* Perl and POSIX space are the same */ + NULL, /* xdigit */ + /* Negated cases */ + string_PL, /* ^alpha */ + string_PLl, /* ^lower */ + string_PLu, /* ^upper */ + string_PXan, /* ^alnum */ + NULL, /* ^ascii */ + string_H, /* ^blank */ + NULL, /* ^cntrl */ + string_PNd, /* ^digit */ + NULL, /* ^graph */ + NULL, /* ^print */ + NULL, /* ^punct */ + string_PXps, /* ^space */ /* Xps is POSIX space, but from 8.34 */ + string_PXwd, /* ^word */ /* Perl and POSIX space are the same */ + NULL /* ^xdigit */ +}; +#define POSIX_SUBSIZE (sizeof(posix_substitutes) / sizeof(pcre_uchar *)) +#endif + +#define STRING(a) # a +#define XSTRING(s) STRING(s) + +/* The texts of compile-time error messages. These are "char *" because they +are passed to the outside world. Do not ever re-use any error number, because +they are documented. Always add a new error instead. Messages marked DEAD below +are no longer used. This used to be a table of strings, but in order to reduce +the number of relocations needed when a shared library is loaded dynamically, +it is now one long string. We cannot use a table of offsets, because the +lengths of inserts such as XSTRING(MAX_NAME_SIZE) are not known. Instead, we +simply count through to the one we want - this isn't a performance issue +because these strings are used only when there is a compilation error. + +Each substring ends with \0 to insert a null character. This includes the final +substring, so that the whole string ends with \0\0, which can be detected when +counting through. */ + +static const char error_texts[] = + "no error\0" + "\\ at end of pattern\0" + "\\c at end of pattern\0" + "unrecognized character follows \\\0" + "numbers out of order in {} quantifier\0" + /* 5 */ + "number too big in {} quantifier\0" + "missing terminating ] for character class\0" + "invalid escape sequence in character class\0" + "range out of order in character class\0" + "nothing to repeat\0" + /* 10 */ + "internal error: invalid forward reference offset\0" + "internal error: unexpected repeat\0" + "unrecognized character after (? or (?-\0" + "POSIX named classes are supported only within a class\0" + "missing )\0" + /* 15 */ + "reference to non-existent subpattern\0" + "erroffset passed as NULL\0" + "unknown option bit(s) set\0" + "missing ) after comment\0" + "parentheses nested too deeply\0" /** DEAD **/ + /* 20 */ + "regular expression is too large\0" + "failed to get memory\0" + "unmatched parentheses\0" + "internal error: code overflow\0" + "unrecognized character after (?<\0" + /* 25 */ + "lookbehind assertion is not fixed length\0" + "malformed number or name after (?(\0" + "conditional group contains more than two branches\0" + "assertion expected after (?(\0" + "(?R or (?[+-]digits must be followed by )\0" + /* 30 */ + "unknown POSIX class name\0" + "POSIX collating elements are not supported\0" + "this version of PCRE is compiled without UTF support\0" + "spare error\0" /** DEAD **/ + "character value in \\x{} or \\o{} is too large\0" + /* 35 */ + "invalid condition (?(0)\0" + "\\C not allowed in lookbehind assertion\0" + "PCRE does not support \\L, \\l, \\N{name}, \\U, or \\u\0" + "number after (?C is > 255\0" + "closing ) for (?C expected\0" + /* 40 */ + "recursive call could loop indefinitely\0" + "unrecognized character after (?P\0" + "syntax error in subpattern name (missing terminator)\0" + "two named subpatterns have the same name\0" + "invalid UTF-8 string\0" + /* 45 */ + "support for \\P, \\p, and \\X has not been compiled\0" + "malformed \\P or \\p sequence\0" + "unknown property name after \\P or \\p\0" + "subpattern name is too long (maximum " XSTRING(MAX_NAME_SIZE) " characters)\0" + "too many named subpatterns (maximum " XSTRING(MAX_NAME_COUNT) ")\0" + /* 50 */ + "repeated subpattern is too long\0" /** DEAD **/ + "octal value is greater than \\377 in 8-bit non-UTF-8 mode\0" + "internal error: overran compiling workspace\0" + "internal error: previously-checked referenced subpattern not found\0" + "DEFINE group contains more than one branch\0" + /* 55 */ + "repeating a DEFINE group is not allowed\0" /** DEAD **/ + "inconsistent NEWLINE options\0" + "\\g is not followed by a braced, angle-bracketed, or quoted name/number or by a plain number\0" + "a numbered reference must not be zero\0" + "an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)\0" + /* 60 */ + "(*VERB) not recognized or malformed\0" + "number is too big\0" + "subpattern name expected\0" + "digit expected after (?+\0" + "] is an invalid data character in JavaScript compatibility mode\0" + /* 65 */ + "different names for subpatterns of the same number are not allowed\0" + "(*MARK) must have an argument\0" + "this version of PCRE is not compiled with Unicode property support\0" +#ifndef EBCDIC + "\\c must be followed by an ASCII character\0" +#else + "\\c must be followed by a letter or one of [\\]^_?\0" +#endif + "\\k is not followed by a braced, angle-bracketed, or quoted name\0" + /* 70 */ + "internal error: unknown opcode in find_fixedlength()\0" + "\\N is not supported in a class\0" + "too many forward references\0" + "disallowed Unicode code point (>= 0xd800 && <= 0xdfff)\0" + "invalid UTF-16 string\0" + /* 75 */ + "name is too long in (*MARK), (*PRUNE), (*SKIP), or (*THEN)\0" + "character value in \\u.... sequence is too large\0" + "invalid UTF-32 string\0" + "setting UTF is disabled by the application\0" + "non-hex character in \\x{} (closing brace missing?)\0" + /* 80 */ + "non-octal character in \\o{} (closing brace missing?)\0" + "missing opening brace after \\o\0" + "parentheses are too deeply nested\0" + "invalid range in character class\0" + "group name must start with a non-digit\0" + /* 85 */ + "parentheses are too deeply nested (stack check)\0" + "digits missing in \\x{} or \\o{}\0" + ; + +/* Table to identify digits and hex digits. This is used when compiling +patterns. Note that the tables in chartables are dependent on the locale, and +may mark arbitrary characters as digits - but the PCRE compiling code expects +to handle only 0-9, a-z, and A-Z as digits when compiling. That is why we have +a private table here. It costs 256 bytes, but it is a lot faster than doing +character value tests (at least in some simple cases I timed), and in some +applications one wants PCRE to compile efficiently as well as match +efficiently. + +For convenience, we use the same bit definitions as in chartables: + + 0x04 decimal digit + 0x08 hexadecimal digit + +Then we can use ctype_digit and ctype_xdigit in the code. */ + +/* Using a simple comparison for decimal numbers rather than a memory read +is much faster, and the resulting code is simpler (the compiler turns it +into a subtraction and unsigned comparison). */ + +#define IS_DIGIT(x) ((x) >= CHAR_0 && (x) <= CHAR_9) + +#ifndef EBCDIC + +/* This is the "normal" case, for ASCII systems, and EBCDIC systems running in +UTF-8 mode. */ + +static const pcre_uint8 digitab[] = + { + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 0- 7 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 8- 15 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 16- 23 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - ' */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* ( - / */ + 0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c, /* 0 - 7 */ + 0x0c,0x0c,0x00,0x00,0x00,0x00,0x00,0x00, /* 8 - ? */ + 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* @ - G */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* H - O */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* P - W */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* X - _ */ + 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* ` - g */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* h - o */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* p - w */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* x -127 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */ + +#else + +/* This is the "abnormal" case, for EBCDIC systems not running in UTF-8 mode. */ + +static const pcre_uint8 digitab[] = + { + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 0- 7 0 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 8- 15 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 16- 23 10 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 32- 39 20 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 40- 47 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 48- 55 30 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 56- 63 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - 71 40 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 72- | */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* & - 87 50 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 88- 95 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - -103 60 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 104- ? */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 112-119 70 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 120- " */ + 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* 128- g 80 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* h -143 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144- p 90 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* q -159 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160- x A0 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* y -175 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* ^ -183 B0 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */ + 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* { - G C0 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* H -207 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* } - P D0 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* Q -223 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* \ - X E0 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* Y -239 */ + 0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c, /* 0 - 7 F0 */ + 0x0c,0x0c,0x00,0x00,0x00,0x00,0x00,0x00};/* 8 -255 */ + +static const pcre_uint8 ebcdic_chartab[] = { /* chartable partial dup */ + 0x80,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /* 0- 7 */ + 0x00,0x00,0x00,0x00,0x01,0x01,0x00,0x00, /* 8- 15 */ + 0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /* 16- 23 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */ + 0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /* 32- 39 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 40- 47 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 48- 55 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 56- 63 */ + 0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - 71 */ + 0x00,0x00,0x00,0x80,0x00,0x80,0x80,0x80, /* 72- | */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* & - 87 */ + 0x00,0x00,0x00,0x80,0x80,0x80,0x00,0x00, /* 88- 95 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - -103 */ + 0x00,0x00,0x00,0x00,0x00,0x10,0x00,0x80, /* 104- ? */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 112-119 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 120- " */ + 0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /* 128- g */ + 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* h -143 */ + 0x00,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* 144- p */ + 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* q -159 */ + 0x00,0x00,0x12,0x12,0x12,0x12,0x12,0x12, /* 160- x */ + 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* y -175 */ + 0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* ^ -183 */ + 0x00,0x00,0x80,0x00,0x00,0x00,0x00,0x00, /* 184-191 */ + 0x80,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /* { - G */ + 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* H -207 */ + 0x00,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* } - P */ + 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* Q -223 */ + 0x00,0x00,0x12,0x12,0x12,0x12,0x12,0x12, /* \ - X */ + 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* Y -239 */ + 0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /* 0 - 7 */ + 0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x00};/* 8 -255 */ +#endif + + +/* This table is used to check whether auto-possessification is possible +between adjacent character-type opcodes. The left-hand (repeated) opcode is +used to select the row, and the right-hand opcode is use to select the column. +A value of 1 means that auto-possessification is OK. For example, the second +value in the first row means that \D+\d can be turned into \D++\d. + +The Unicode property types (\P and \p) have to be present to fill out the table +because of what their opcode values are, but the table values should always be +zero because property types are handled separately in the code. The last four +columns apply to items that cannot be repeated, so there is no need to have +rows for them. Note that OP_DIGIT etc. are generated only when PCRE_UCP is +*not* set. When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */ + +#define APTROWS (LAST_AUTOTAB_LEFT_OP - FIRST_AUTOTAB_OP + 1) +#define APTCOLS (LAST_AUTOTAB_RIGHT_OP - FIRST_AUTOTAB_OP + 1) + +static const pcre_uint8 autoposstab[APTROWS][APTCOLS] = { +/* \D \d \S \s \W \w . .+ \C \P \p \R \H \h \V \v \X \Z \z $ $M */ + { 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0 }, /* \D */ + { 1, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 1, 1, 1, 1 }, /* \d */ + { 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 1, 1, 1, 1 }, /* \S */ + { 0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0 }, /* \s */ + { 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0 }, /* \W */ + { 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 1, 1, 1, 1 }, /* \w */ + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0 }, /* . */ + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0 }, /* .+ */ + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0 }, /* \C */ + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, /* \P */ + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, /* \p */ + { 0, 1, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0 }, /* \R */ + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0 }, /* \H */ + { 0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 0, 0, 1, 0, 0, 1, 0, 0 }, /* \h */ + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0 }, /* \V */ + { 0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 0 }, /* \v */ + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0 } /* \X */ +}; + + +/* This table is used to check whether auto-possessification is possible +between adjacent Unicode property opcodes (OP_PROP and OP_NOTPROP). The +left-hand (repeated) opcode is used to select the row, and the right-hand +opcode is used to select the column. The values are as follows: + + 0 Always return FALSE (never auto-possessify) + 1 Character groups are distinct (possessify if both are OP_PROP) + 2 Check character categories in the same group (general or particular) + 3 TRUE if the two opcodes are not the same (PROP vs NOTPROP) + + 4 Check left general category vs right particular category + 5 Check right general category vs left particular category + + 6 Left alphanum vs right general category + 7 Left space vs right general category + 8 Left word vs right general category + + 9 Right alphanum vs left general category + 10 Right space vs left general category + 11 Right word vs left general category + + 12 Left alphanum vs right particular category + 13 Left space vs right particular category + 14 Left word vs right particular category + + 15 Right alphanum vs left particular category + 16 Right space vs left particular category + 17 Right word vs left particular category +*/ + +static const pcre_uint8 propposstab[PT_TABSIZE][PT_TABSIZE] = { +/* ANY LAMP GC PC SC ALNUM SPACE PXSPACE WORD CLIST UCNC */ + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, /* PT_ANY */ + { 0, 3, 0, 0, 0, 3, 1, 1, 0, 0, 0 }, /* PT_LAMP */ + { 0, 0, 2, 4, 0, 9, 10, 10, 11, 0, 0 }, /* PT_GC */ + { 0, 0, 5, 2, 0, 15, 16, 16, 17, 0, 0 }, /* PT_PC */ + { 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0 }, /* PT_SC */ + { 0, 3, 6, 12, 0, 3, 1, 1, 0, 0, 0 }, /* PT_ALNUM */ + { 0, 1, 7, 13, 0, 1, 3, 3, 1, 0, 0 }, /* PT_SPACE */ + { 0, 1, 7, 13, 0, 1, 3, 3, 1, 0, 0 }, /* PT_PXSPACE */ + { 0, 0, 8, 14, 0, 0, 1, 1, 3, 0, 0 }, /* PT_WORD */ + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, /* PT_CLIST */ + { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3 } /* PT_UCNC */ +}; + +/* This table is used to check whether auto-possessification is possible +between adjacent Unicode property opcodes (OP_PROP and OP_NOTPROP) when one +specifies a general category and the other specifies a particular category. The +row is selected by the general category and the column by the particular +category. The value is 1 if the particular category is not part of the general +category. */ + +static const pcre_uint8 catposstab[7][30] = { +/* Cc Cf Cn Co Cs Ll Lm Lo Lt Lu Mc Me Mn Nd Nl No Pc Pd Pe Pf Pi Po Ps Sc Sk Sm So Zl Zp Zs */ + { 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 }, /* C */ + { 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 }, /* L */ + { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 }, /* M */ + { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 }, /* N */ + { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1 }, /* P */ + { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 1, 1, 1 }, /* S */ + { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0 } /* Z */ +}; + +/* This table is used when checking ALNUM, (PX)SPACE, SPACE, and WORD against +a general or particular category. The properties in each row are those +that apply to the character set in question. Duplication means that a little +unnecessary work is done when checking, but this keeps things much simpler +because they can all use the same code. For more details see the comment where +this table is used. + +Note: SPACE and PXSPACE used to be different because Perl excluded VT from +"space", but from Perl 5.18 it's included, so both categories are treated the +same here. */ + +static const pcre_uint8 posspropstab[3][4] = { + { ucp_L, ucp_N, ucp_N, ucp_Nl }, /* ALNUM, 3rd and 4th values redundant */ + { ucp_Z, ucp_Z, ucp_C, ucp_Cc }, /* SPACE and PXSPACE, 2nd value redundant */ + { ucp_L, ucp_N, ucp_P, ucp_Po } /* WORD */ +}; + +/* This table is used when converting repeating opcodes into possessified +versions as a result of an explicit possessive quantifier such as ++. A zero +value means there is no possessified version - in those cases the item in +question must be wrapped in ONCE brackets. The table is truncated at OP_CALLOUT +because all relevant opcodes are less than that. */ + +static const pcre_uint8 opcode_possessify[] = { + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 0 - 15 */ + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 16 - 31 */ + + 0, /* NOTI */ + OP_POSSTAR, 0, /* STAR, MINSTAR */ + OP_POSPLUS, 0, /* PLUS, MINPLUS */ + OP_POSQUERY, 0, /* QUERY, MINQUERY */ + OP_POSUPTO, 0, /* UPTO, MINUPTO */ + 0, /* EXACT */ + 0, 0, 0, 0, /* POS{STAR,PLUS,QUERY,UPTO} */ + + OP_POSSTARI, 0, /* STARI, MINSTARI */ + OP_POSPLUSI, 0, /* PLUSI, MINPLUSI */ + OP_POSQUERYI, 0, /* QUERYI, MINQUERYI */ + OP_POSUPTOI, 0, /* UPTOI, MINUPTOI */ + 0, /* EXACTI */ + 0, 0, 0, 0, /* POS{STARI,PLUSI,QUERYI,UPTOI} */ + + OP_NOTPOSSTAR, 0, /* NOTSTAR, NOTMINSTAR */ + OP_NOTPOSPLUS, 0, /* NOTPLUS, NOTMINPLUS */ + OP_NOTPOSQUERY, 0, /* NOTQUERY, NOTMINQUERY */ + OP_NOTPOSUPTO, 0, /* NOTUPTO, NOTMINUPTO */ + 0, /* NOTEXACT */ + 0, 0, 0, 0, /* NOTPOS{STAR,PLUS,QUERY,UPTO} */ + + OP_NOTPOSSTARI, 0, /* NOTSTARI, NOTMINSTARI */ + OP_NOTPOSPLUSI, 0, /* NOTPLUSI, NOTMINPLUSI */ + OP_NOTPOSQUERYI, 0, /* NOTQUERYI, NOTMINQUERYI */ + OP_NOTPOSUPTOI, 0, /* NOTUPTOI, NOTMINUPTOI */ + 0, /* NOTEXACTI */ + 0, 0, 0, 0, /* NOTPOS{STARI,PLUSI,QUERYI,UPTOI} */ + + OP_TYPEPOSSTAR, 0, /* TYPESTAR, TYPEMINSTAR */ + OP_TYPEPOSPLUS, 0, /* TYPEPLUS, TYPEMINPLUS */ + OP_TYPEPOSQUERY, 0, /* TYPEQUERY, TYPEMINQUERY */ + OP_TYPEPOSUPTO, 0, /* TYPEUPTO, TYPEMINUPTO */ + 0, /* TYPEEXACT */ + 0, 0, 0, 0, /* TYPEPOS{STAR,PLUS,QUERY,UPTO} */ + + OP_CRPOSSTAR, 0, /* CRSTAR, CRMINSTAR */ + OP_CRPOSPLUS, 0, /* CRPLUS, CRMINPLUS */ + OP_CRPOSQUERY, 0, /* CRQUERY, CRMINQUERY */ + OP_CRPOSRANGE, 0, /* CRRANGE, CRMINRANGE */ + 0, 0, 0, 0, /* CRPOS{STAR,PLUS,QUERY,RANGE} */ + + 0, 0, 0, /* CLASS, NCLASS, XCLASS */ + 0, 0, /* REF, REFI */ + 0, 0, /* DNREF, DNREFI */ + 0, 0 /* RECURSE, CALLOUT */ +}; + + + +/************************************************* +* Find an error text * +*************************************************/ + +/* The error texts are now all in one long string, to save on relocations. As +some of the text is of unknown length, we can't use a table of offsets. +Instead, just count through the strings. This is not a performance issue +because it happens only when there has been a compilation error. + +Argument: the error number +Returns: pointer to the error string +*/ + +static const char * +find_error_text(int n) +{ +const char *s = error_texts; +for (; n > 0; n--) + { + while (*s++ != CHAR_NULL) {}; + if (*s == CHAR_NULL) return "Error text not found (please report)"; + } +return s; +} + + + +/************************************************* +* Expand the workspace * +*************************************************/ + +/* This function is called during the second compiling phase, if the number of +forward references fills the existing workspace, which is originally a block on +the stack. A larger block is obtained from malloc() unless the ultimate limit +has been reached or the increase will be rather small. + +Argument: pointer to the compile data block +Returns: 0 if all went well, else an error number +*/ + +static int +expand_workspace(compile_data *cd) +{ +pcre_uchar *newspace; +int newsize = cd->workspace_size * 2; + +if (newsize > COMPILE_WORK_SIZE_MAX) newsize = COMPILE_WORK_SIZE_MAX; +if (cd->workspace_size >= COMPILE_WORK_SIZE_MAX || + newsize - cd->workspace_size < WORK_SIZE_SAFETY_MARGIN) + return ERR72; + +newspace = (PUBL(malloc))(IN_UCHARS(newsize)); +if (newspace == NULL) return ERR21; +memcpy(newspace, cd->start_workspace, cd->workspace_size * sizeof(pcre_uchar)); +cd->hwm = (pcre_uchar *)newspace + (cd->hwm - cd->start_workspace); +if (cd->workspace_size > COMPILE_WORK_SIZE) + (PUBL(free))((void *)cd->start_workspace); +cd->start_workspace = newspace; +cd->workspace_size = newsize; +return 0; +} + + + +/************************************************* +* Check for counted repeat * +*************************************************/ + +/* This function is called when a '{' is encountered in a place where it might +start a quantifier. It looks ahead to see if it really is a quantifier or not. +It is only a quantifier if it is one of the forms {ddd} {ddd,} or {ddd,ddd} +where the ddds are digits. + +Arguments: + p pointer to the first char after '{' + +Returns: TRUE or FALSE +*/ + +static BOOL +is_counted_repeat(const pcre_uchar *p) +{ +if (!IS_DIGIT(*p)) return FALSE; +p++; +while (IS_DIGIT(*p)) p++; +if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE; + +if (*p++ != CHAR_COMMA) return FALSE; +if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE; + +if (!IS_DIGIT(*p)) return FALSE; +p++; +while (IS_DIGIT(*p)) p++; + +return (*p == CHAR_RIGHT_CURLY_BRACKET); +} + + + +/************************************************* +* Handle escapes * +*************************************************/ + +/* This function is called when a \ has been encountered. It either returns a +positive value for a simple escape such as \n, or 0 for a data character which +will be placed in chptr. A backreference to group n is returned as negative n. +When UTF-8 is enabled, a positive value greater than 255 may be returned in +chptr. On entry, ptr is pointing at the \. On exit, it is on the final +character of the escape sequence. + +Arguments: + ptrptr points to the pattern position pointer + chptr points to a returned data character + errorcodeptr points to the errorcode variable + bracount number of previous extracting brackets + options the options bits + isclass TRUE if inside a character class + +Returns: zero => a data character + positive => a special escape sequence + negative => a back reference + on error, errorcodeptr is set +*/ + +static int +check_escape(const pcre_uchar **ptrptr, pcre_uint32 *chptr, int *errorcodeptr, + int bracount, int options, BOOL isclass) +{ +/* PCRE_UTF16 has the same value as PCRE_UTF8. */ +BOOL utf = (options & PCRE_UTF8) != 0; +const pcre_uchar *ptr = *ptrptr + 1; +pcre_uint32 c; +int escape = 0; +int i; + +GETCHARINCTEST(c, ptr); /* Get character value, increment pointer */ +ptr--; /* Set pointer back to the last byte */ + +/* If backslash is at the end of the pattern, it's an error. */ + +if (c == CHAR_NULL) *errorcodeptr = ERR1; + +/* Non-alphanumerics are literals. For digits or letters, do an initial lookup +in a table. A non-zero result is something that can be returned immediately. +Otherwise further processing may be required. */ + +#ifndef EBCDIC /* ASCII/UTF-8 coding */ +/* Not alphanumeric */ +else if (c < CHAR_0 || c > CHAR_z) {} +else if ((i = escapes[c - CHAR_0]) != 0) + { if (i > 0) c = (pcre_uint32)i; else escape = -i; } + +#else /* EBCDIC coding */ +/* Not alphanumeric */ +else if (c < CHAR_a || (!MAX_255(c) || (ebcdic_chartab[c] & 0x0E) == 0)) {} +else if ((i = escapes[c - 0x48]) != 0) { if (i > 0) c = (pcre_uint32)i; else escape = -i; } +#endif + +/* Escapes that need further processing, or are illegal. */ + +else + { + const pcre_uchar *oldptr; + BOOL braced, negated, overflow; + int s; + + switch (c) + { + /* A number of Perl escapes are not handled by PCRE. We give an explicit + error. */ + + case CHAR_l: + case CHAR_L: + *errorcodeptr = ERR37; + break; + + case CHAR_u: + if ((options & PCRE_JAVASCRIPT_COMPAT) != 0) + { + /* In JavaScript, \u must be followed by four hexadecimal numbers. + Otherwise it is a lowercase u letter. */ + if (MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0 + && MAX_255(ptr[2]) && (digitab[ptr[2]] & ctype_xdigit) != 0 + && MAX_255(ptr[3]) && (digitab[ptr[3]] & ctype_xdigit) != 0 + && MAX_255(ptr[4]) && (digitab[ptr[4]] & ctype_xdigit) != 0) + { + c = 0; + for (i = 0; i < 4; ++i) + { + register pcre_uint32 cc = *(++ptr); +#ifndef EBCDIC /* ASCII/UTF-8 coding */ + if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */ + c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10)); +#else /* EBCDIC coding */ + if (cc >= CHAR_a && cc <= CHAR_z) cc += 64; /* Convert to upper case */ + c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10)); +#endif + } + +#if defined COMPILE_PCRE8 + if (c > (utf ? 0x10ffffU : 0xffU)) +#elif defined COMPILE_PCRE16 + if (c > (utf ? 0x10ffffU : 0xffffU)) +#elif defined COMPILE_PCRE32 + if (utf && c > 0x10ffffU) +#endif + { + *errorcodeptr = ERR76; + } + else if (utf && c >= 0xd800 && c <= 0xdfff) *errorcodeptr = ERR73; + } + } + else + *errorcodeptr = ERR37; + break; + + case CHAR_U: + /* In JavaScript, \U is an uppercase U letter. */ + if ((options & PCRE_JAVASCRIPT_COMPAT) == 0) *errorcodeptr = ERR37; + break; + + /* In a character class, \g is just a literal "g". Outside a character + class, \g must be followed by one of a number of specific things: + + (1) A number, either plain or braced. If positive, it is an absolute + backreference. If negative, it is a relative backreference. This is a Perl + 5.10 feature. + + (2) Perl 5.10 also supports \g{name} as a reference to a named group. This + is part of Perl's movement towards a unified syntax for back references. As + this is synonymous with \k{name}, we fudge it up by pretending it really + was \k. + + (3) For Oniguruma compatibility we also support \g followed by a name or a + number either in angle brackets or in single quotes. However, these are + (possibly recursive) subroutine calls, _not_ backreferences. Just return + the ESC_g code (cf \k). */ + + case CHAR_g: + if (isclass) break; + if (ptr[1] == CHAR_LESS_THAN_SIGN || ptr[1] == CHAR_APOSTROPHE) + { + escape = ESC_g; + break; + } + + /* Handle the Perl-compatible cases */ + + if (ptr[1] == CHAR_LEFT_CURLY_BRACKET) + { + const pcre_uchar *p; + for (p = ptr+2; *p != CHAR_NULL && *p != CHAR_RIGHT_CURLY_BRACKET; p++) + if (*p != CHAR_MINUS && !IS_DIGIT(*p)) break; + if (*p != CHAR_NULL && *p != CHAR_RIGHT_CURLY_BRACKET) + { + escape = ESC_k; + break; + } + braced = TRUE; + ptr++; + } + else braced = FALSE; + + if (ptr[1] == CHAR_MINUS) + { + negated = TRUE; + ptr++; + } + else negated = FALSE; + + /* The integer range is limited by the machine's int representation. */ + s = 0; + overflow = FALSE; + while (IS_DIGIT(ptr[1])) + { + if (s > INT_MAX / 10 - 1) /* Integer overflow */ + { + overflow = TRUE; + break; + } + s = s * 10 + (int)(*(++ptr) - CHAR_0); + } + if (overflow) /* Integer overflow */ + { + while (IS_DIGIT(ptr[1])) + ptr++; + *errorcodeptr = ERR61; + break; + } + + if (braced && *(++ptr) != CHAR_RIGHT_CURLY_BRACKET) + { + *errorcodeptr = ERR57; + break; + } + + if (s == 0) + { + *errorcodeptr = ERR58; + break; + } + + if (negated) + { + if (s > bracount) + { + *errorcodeptr = ERR15; + break; + } + s = bracount - (s - 1); + } + + escape = -s; + break; + + /* The handling of escape sequences consisting of a string of digits + starting with one that is not zero is not straightforward. Perl has changed + over the years. Nowadays \g{} for backreferences and \o{} for octal are + recommended to avoid the ambiguities in the old syntax. + + Outside a character class, the digits are read as a decimal number. If the + number is less than 8 (used to be 10), or if there are that many previous + extracting left brackets, then it is a back reference. Otherwise, up to + three octal digits are read to form an escaped byte. Thus \123 is likely to + be octal 123 (cf \0123, which is octal 012 followed by the literal 3). If + the octal value is greater than 377, the least significant 8 bits are + taken. \8 and \9 are treated as the literal characters 8 and 9. + + Inside a character class, \ followed by a digit is always either a literal + 8 or 9 or an octal number. */ + + case CHAR_1: case CHAR_2: case CHAR_3: case CHAR_4: case CHAR_5: + case CHAR_6: case CHAR_7: case CHAR_8: case CHAR_9: + + if (!isclass) + { + oldptr = ptr; + /* The integer range is limited by the machine's int representation. */ + s = (int)(c -CHAR_0); + overflow = FALSE; + while (IS_DIGIT(ptr[1])) + { + if (s > INT_MAX / 10 - 1) /* Integer overflow */ + { + overflow = TRUE; + break; + } + s = s * 10 + (int)(*(++ptr) - CHAR_0); + } + if (overflow) /* Integer overflow */ + { + while (IS_DIGIT(ptr[1])) + ptr++; + *errorcodeptr = ERR61; + break; + } + if (s < 8 || s <= bracount) /* Check for back reference */ + { + escape = -s; + break; + } + ptr = oldptr; /* Put the pointer back and fall through */ + } + + /* Handle a digit following \ when the number is not a back reference. If + the first digit is 8 or 9, Perl used to generate a binary zero byte and + then treat the digit as a following literal. At least by Perl 5.18 this + changed so as not to insert the binary zero. */ + + if ((c = *ptr) >= CHAR_8) break; + + /* Fall through with a digit less than 8 */ + + /* \0 always starts an octal number, but we may drop through to here with a + larger first octal digit. The original code used just to take the least + significant 8 bits of octal numbers (I think this is what early Perls used + to do). Nowadays we allow for larger numbers in UTF-8 mode and 16-bit mode, + but no more than 3 octal digits. */ + + case CHAR_0: + c -= CHAR_0; + while(i++ < 2 && ptr[1] >= CHAR_0 && ptr[1] <= CHAR_7) + c = c * 8 + *(++ptr) - CHAR_0; +#ifdef COMPILE_PCRE8 + if (!utf && c > 0xff) *errorcodeptr = ERR51; +#endif + break; + + /* \o is a relatively new Perl feature, supporting a more general way of + specifying character codes in octal. The only supported form is \o{ddd}. */ + + case CHAR_o: + if (ptr[1] != CHAR_LEFT_CURLY_BRACKET) *errorcodeptr = ERR81; else + if (ptr[2] == CHAR_RIGHT_CURLY_BRACKET) *errorcodeptr = ERR86; else + { + ptr += 2; + c = 0; + overflow = FALSE; + while (*ptr >= CHAR_0 && *ptr <= CHAR_7) + { + register pcre_uint32 cc = *ptr++; + if (c == 0 && cc == CHAR_0) continue; /* Leading zeroes */ +#ifdef COMPILE_PCRE32 + if (c >= 0x20000000l) { overflow = TRUE; break; } +#endif + c = (c << 3) + cc - CHAR_0 ; +#if defined COMPILE_PCRE8 + if (c > (utf ? 0x10ffffU : 0xffU)) { overflow = TRUE; break; } +#elif defined COMPILE_PCRE16 + if (c > (utf ? 0x10ffffU : 0xffffU)) { overflow = TRUE; break; } +#elif defined COMPILE_PCRE32 + if (utf && c > 0x10ffffU) { overflow = TRUE; break; } +#endif + } + if (overflow) + { + while (*ptr >= CHAR_0 && *ptr <= CHAR_7) ptr++; + *errorcodeptr = ERR34; + } + else if (*ptr == CHAR_RIGHT_CURLY_BRACKET) + { + if (utf && c >= 0xd800 && c <= 0xdfff) *errorcodeptr = ERR73; + } + else *errorcodeptr = ERR80; + } + break; + + /* \x is complicated. In JavaScript, \x must be followed by two hexadecimal + numbers. Otherwise it is a lowercase x letter. */ + + case CHAR_x: + if ((options & PCRE_JAVASCRIPT_COMPAT) != 0) + { + if (MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0 + && MAX_255(ptr[2]) && (digitab[ptr[2]] & ctype_xdigit) != 0) + { + c = 0; + for (i = 0; i < 2; ++i) + { + register pcre_uint32 cc = *(++ptr); +#ifndef EBCDIC /* ASCII/UTF-8 coding */ + if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */ + c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10)); +#else /* EBCDIC coding */ + if (cc >= CHAR_a && cc <= CHAR_z) cc += 64; /* Convert to upper case */ + c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10)); +#endif + } + } + } /* End JavaScript handling */ + + /* Handle \x in Perl's style. \x{ddd} is a character number which can be + greater than 0xff in utf or non-8bit mode, but only if the ddd are hex + digits. If not, { used to be treated as a data character. However, Perl + seems to read hex digits up to the first non-such, and ignore the rest, so + that, for example \x{zz} matches a binary zero. This seems crazy, so PCRE + now gives an error. */ + + else + { + if (ptr[1] == CHAR_LEFT_CURLY_BRACKET) + { + ptr += 2; + if (*ptr == CHAR_RIGHT_CURLY_BRACKET) + { + *errorcodeptr = ERR86; + break; + } + c = 0; + overflow = FALSE; + while (MAX_255(*ptr) && (digitab[*ptr] & ctype_xdigit) != 0) + { + register pcre_uint32 cc = *ptr++; + if (c == 0 && cc == CHAR_0) continue; /* Leading zeroes */ + +#ifdef COMPILE_PCRE32 + if (c >= 0x10000000l) { overflow = TRUE; break; } +#endif + +#ifndef EBCDIC /* ASCII/UTF-8 coding */ + if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */ + c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10)); +#else /* EBCDIC coding */ + if (cc >= CHAR_a && cc <= CHAR_z) cc += 64; /* Convert to upper case */ + c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10)); +#endif + +#if defined COMPILE_PCRE8 + if (c > (utf ? 0x10ffffU : 0xffU)) { overflow = TRUE; break; } +#elif defined COMPILE_PCRE16 + if (c > (utf ? 0x10ffffU : 0xffffU)) { overflow = TRUE; break; } +#elif defined COMPILE_PCRE32 + if (utf && c > 0x10ffffU) { overflow = TRUE; break; } +#endif + } + + if (overflow) + { + while (MAX_255(*ptr) && (digitab[*ptr] & ctype_xdigit) != 0) ptr++; + *errorcodeptr = ERR34; + } + + else if (*ptr == CHAR_RIGHT_CURLY_BRACKET) + { + if (utf && c >= 0xd800 && c <= 0xdfff) *errorcodeptr = ERR73; + } + + /* If the sequence of hex digits does not end with '}', give an error. + We used just to recognize this construct and fall through to the normal + \x handling, but nowadays Perl gives an error, which seems much more + sensible, so we do too. */ + + else *errorcodeptr = ERR79; + } /* End of \x{} processing */ + + /* Read a single-byte hex-defined char (up to two hex digits after \x) */ + + else + { + c = 0; + while (i++ < 2 && MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0) + { + pcre_uint32 cc; /* Some compilers don't like */ + cc = *(++ptr); /* ++ in initializers */ +#ifndef EBCDIC /* ASCII/UTF-8 coding */ + if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */ + c = c * 16 + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10)); +#else /* EBCDIC coding */ + if (cc <= CHAR_z) cc += 64; /* Convert to upper case */ + c = c * 16 + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10)); +#endif + } + } /* End of \xdd handling */ + } /* End of Perl-style \x handling */ + break; + + /* For \c, a following letter is upper-cased; then the 0x40 bit is flipped. + An error is given if the byte following \c is not an ASCII character. This + coding is ASCII-specific, but then the whole concept of \cx is + ASCII-specific. (However, an EBCDIC equivalent has now been added.) */ + + case CHAR_c: + c = *(++ptr); + if (c == CHAR_NULL) + { + *errorcodeptr = ERR2; + break; + } +#ifndef EBCDIC /* ASCII/UTF-8 coding */ + if (c > 127) /* Excludes all non-ASCII in either mode */ + { + *errorcodeptr = ERR68; + break; + } + if (c >= CHAR_a && c <= CHAR_z) c -= 32; + c ^= 0x40; +#else /* EBCDIC coding */ + if (c >= CHAR_a && c <= CHAR_z) c += 64; + if (c == CHAR_QUESTION_MARK) + c = ('\\' == 188 && '`' == 74)? 0x5f : 0xff; + else + { + for (i = 0; i < 32; i++) + { + if (c == ebcdic_escape_c[i]) break; + } + if (i < 32) c = i; else *errorcodeptr = ERR68; + } +#endif + break; + + /* PCRE_EXTRA enables extensions to Perl in the matter of escapes. Any + other alphanumeric following \ is an error if PCRE_EXTRA was set; + otherwise, for Perl compatibility, it is a literal. This code looks a bit + odd, but there used to be some cases other than the default, and there may + be again in future, so I haven't "optimized" it. */ + + default: + if ((options & PCRE_EXTRA) != 0) switch(c) + { + default: + *errorcodeptr = ERR3; + break; + } + break; + } + } + +/* Perl supports \N{name} for character names, as well as plain \N for "not +newline". PCRE does not support \N{name}. However, it does support +quantification such as \N{2,3}. */ + +if (escape == ESC_N && ptr[1] == CHAR_LEFT_CURLY_BRACKET && + !is_counted_repeat(ptr+2)) + *errorcodeptr = ERR37; + +/* If PCRE_UCP is set, we change the values for \d etc. */ + +if ((options & PCRE_UCP) != 0 && escape >= ESC_D && escape <= ESC_w) + escape += (ESC_DU - ESC_D); + +/* Set the pointer to the final character before returning. */ + +*ptrptr = ptr; +*chptr = c; +return escape; +} + + + +#ifdef SUPPORT_UCP +/************************************************* +* Handle \P and \p * +*************************************************/ + +/* This function is called after \P or \p has been encountered, provided that +PCRE is compiled with support for Unicode properties. On entry, ptrptr is +pointing at the P or p. On exit, it is pointing at the final character of the +escape sequence. + +Argument: + ptrptr points to the pattern position pointer + negptr points to a boolean that is set TRUE for negation else FALSE + ptypeptr points to an unsigned int that is set to the type value + pdataptr points to an unsigned int that is set to the detailed property value + errorcodeptr points to the error code variable + +Returns: TRUE if the type value was found, or FALSE for an invalid type +*/ + +static BOOL +get_ucp(const pcre_uchar **ptrptr, BOOL *negptr, unsigned int *ptypeptr, + unsigned int *pdataptr, int *errorcodeptr) +{ +pcre_uchar c; +int i, bot, top; +const pcre_uchar *ptr = *ptrptr; +pcre_uchar name[32]; + +c = *(++ptr); +if (c == CHAR_NULL) goto ERROR_RETURN; + +*negptr = FALSE; + +/* \P or \p can be followed by a name in {}, optionally preceded by ^ for +negation. */ + +if (c == CHAR_LEFT_CURLY_BRACKET) + { + if (ptr[1] == CHAR_CIRCUMFLEX_ACCENT) + { + *negptr = TRUE; + ptr++; + } + for (i = 0; i < (int)(sizeof(name) / sizeof(pcre_uchar)) - 1; i++) + { + c = *(++ptr); + if (c == CHAR_NULL) goto ERROR_RETURN; + if (c == CHAR_RIGHT_CURLY_BRACKET) break; + name[i] = c; + } + if (c != CHAR_RIGHT_CURLY_BRACKET) goto ERROR_RETURN; + name[i] = 0; + } + +/* Otherwise there is just one following character */ + +else + { + name[0] = c; + name[1] = 0; + } + +*ptrptr = ptr; + +/* Search for a recognized property name using binary chop */ + +bot = 0; +top = PRIV(utt_size); + +while (bot < top) + { + int r; + i = (bot + top) >> 1; + r = STRCMP_UC_C8(name, PRIV(utt_names) + PRIV(utt)[i].name_offset); + if (r == 0) + { + *ptypeptr = PRIV(utt)[i].type; + *pdataptr = PRIV(utt)[i].value; + return TRUE; + } + if (r > 0) bot = i + 1; else top = i; + } + +*errorcodeptr = ERR47; +*ptrptr = ptr; +return FALSE; + +ERROR_RETURN: +*errorcodeptr = ERR46; +*ptrptr = ptr; +return FALSE; +} +#endif + + + +/************************************************* +* Read repeat counts * +*************************************************/ + +/* Read an item of the form {n,m} and return the values. This is called only +after is_counted_repeat() has confirmed that a repeat-count quantifier exists, +so the syntax is guaranteed to be correct, but we need to check the values. + +Arguments: + p pointer to first char after '{' + minp pointer to int for min + maxp pointer to int for max + returned as -1 if no max + errorcodeptr points to error code variable + +Returns: pointer to '}' on success; + current ptr on error, with errorcodeptr set non-zero +*/ + +static const pcre_uchar * +read_repeat_counts(const pcre_uchar *p, int *minp, int *maxp, int *errorcodeptr) +{ +int min = 0; +int max = -1; + +while (IS_DIGIT(*p)) + { + min = min * 10 + (int)(*p++ - CHAR_0); + if (min > 65535) + { + *errorcodeptr = ERR5; + return p; + } + } + +if (*p == CHAR_RIGHT_CURLY_BRACKET) max = min; else + { + if (*(++p) != CHAR_RIGHT_CURLY_BRACKET) + { + max = 0; + while(IS_DIGIT(*p)) + { + max = max * 10 + (int)(*p++ - CHAR_0); + if (max > 65535) + { + *errorcodeptr = ERR5; + return p; + } + } + if (max < min) + { + *errorcodeptr = ERR4; + return p; + } + } + } + +*minp = min; +*maxp = max; +return p; +} + + + +/************************************************* +* Find first significant op code * +*************************************************/ + +/* This is called by several functions that scan a compiled expression looking +for a fixed first character, or an anchoring op code etc. It skips over things +that do not influence this. For some calls, it makes sense to skip negative +forward and all backward assertions, and also the \b assertion; for others it +does not. + +Arguments: + code pointer to the start of the group + skipassert TRUE if certain assertions are to be skipped + +Returns: pointer to the first significant opcode +*/ + +static const pcre_uchar* +first_significant_code(const pcre_uchar *code, BOOL skipassert) +{ +for (;;) + { + switch ((int)*code) + { + case OP_ASSERT_NOT: + case OP_ASSERTBACK: + case OP_ASSERTBACK_NOT: + if (!skipassert) return code; + do code += GET(code, 1); while (*code == OP_ALT); + code += PRIV(OP_lengths)[*code]; + break; + + case OP_WORD_BOUNDARY: + case OP_NOT_WORD_BOUNDARY: + if (!skipassert) return code; + /* Fall through */ + + case OP_CALLOUT: + case OP_CREF: + case OP_DNCREF: + case OP_RREF: + case OP_DNRREF: + case OP_DEF: + code += PRIV(OP_lengths)[*code]; + break; + + default: + return code; + } + } +/* Control never reaches here */ +} + + + +/************************************************* +* Find the fixed length of a branch * +*************************************************/ + +/* Scan a branch and compute the fixed length of subject that will match it, +if the length is fixed. This is needed for dealing with backward assertions. +In UTF8 mode, the result is in characters rather than bytes. The branch is +temporarily terminated with OP_END when this function is called. + +This function is called when a backward assertion is encountered, so that if it +fails, the error message can point to the correct place in the pattern. +However, we cannot do this when the assertion contains subroutine calls, +because they can be forward references. We solve this by remembering this case +and doing the check at the end; a flag specifies which mode we are running in. + +Arguments: + code points to the start of the pattern (the bracket) + utf TRUE in UTF-8 / UTF-16 / UTF-32 mode + atend TRUE if called when the pattern is complete + cd the "compile data" structure + recurses chain of recurse_check to catch mutual recursion + +Returns: the fixed length, + or -1 if there is no fixed length, + or -2 if \C was encountered (in UTF-8 mode only) + or -3 if an OP_RECURSE item was encountered and atend is FALSE + or -4 if an unknown opcode was encountered (internal error) +*/ + +static int +find_fixedlength(pcre_uchar *code, BOOL utf, BOOL atend, compile_data *cd, + recurse_check *recurses) +{ +int length = -1; +recurse_check this_recurse; +register int branchlength = 0; +register pcre_uchar *cc = code + 1 + LINK_SIZE; + +/* Scan along the opcodes for this branch. If we get to the end of the +branch, check the length against that of the other branches. */ + +for (;;) + { + int d; + pcre_uchar *ce, *cs; + register pcre_uchar op = *cc; + + switch (op) + { + /* We only need to continue for OP_CBRA (normal capturing bracket) and + OP_BRA (normal non-capturing bracket) because the other variants of these + opcodes are all concerned with unlimited repeated groups, which of course + are not of fixed length. */ + + case OP_CBRA: + case OP_BRA: + case OP_ONCE: + case OP_ONCE_NC: + case OP_COND: + d = find_fixedlength(cc + ((op == OP_CBRA)? IMM2_SIZE : 0), utf, atend, cd, + recurses); + if (d < 0) return d; + branchlength += d; + do cc += GET(cc, 1); while (*cc == OP_ALT); + cc += 1 + LINK_SIZE; + break; + + /* Reached end of a branch; if it's a ket it is the end of a nested call. + If it's ALT it is an alternation in a nested call. An ACCEPT is effectively + an ALT. If it is END it's the end of the outer call. All can be handled by + the same code. Note that we must not include the OP_KETRxxx opcodes here, + because they all imply an unlimited repeat. */ + + case OP_ALT: + case OP_KET: + case OP_END: + case OP_ACCEPT: + case OP_ASSERT_ACCEPT: + if (length < 0) length = branchlength; + else if (length != branchlength) return -1; + if (*cc != OP_ALT) return length; + cc += 1 + LINK_SIZE; + branchlength = 0; + break; + + /* A true recursion implies not fixed length, but a subroutine call may + be OK. If the subroutine is a forward reference, we can't deal with + it until the end of the pattern, so return -3. */ + + case OP_RECURSE: + if (!atend) return -3; + cs = ce = (pcre_uchar *)cd->start_code + GET(cc, 1); /* Start subpattern */ + do ce += GET(ce, 1); while (*ce == OP_ALT); /* End subpattern */ + if (cc > cs && cc < ce) return -1; /* Recursion */ + else /* Check for mutual recursion */ + { + recurse_check *r = recurses; + for (r = recurses; r != NULL; r = r->prev) if (r->group == cs) break; + if (r != NULL) return -1; /* Mutual recursion */ + } + this_recurse.prev = recurses; + this_recurse.group = cs; + d = find_fixedlength(cs + IMM2_SIZE, utf, atend, cd, &this_recurse); + if (d < 0) return d; + branchlength += d; + cc += 1 + LINK_SIZE; + break; + + /* Skip over assertive subpatterns */ + + case OP_ASSERT: + case OP_ASSERT_NOT: + case OP_ASSERTBACK: + case OP_ASSERTBACK_NOT: + do cc += GET(cc, 1); while (*cc == OP_ALT); + cc += 1 + LINK_SIZE; + break; + + /* Skip over things that don't match chars */ + + case OP_MARK: + case OP_PRUNE_ARG: + case OP_SKIP_ARG: + case OP_THEN_ARG: + cc += cc[1] + PRIV(OP_lengths)[*cc]; + break; + + case OP_CALLOUT: + case OP_CIRC: + case OP_CIRCM: + case OP_CLOSE: + case OP_COMMIT: + case OP_CREF: + case OP_DEF: + case OP_DNCREF: + case OP_DNRREF: + case OP_DOLL: + case OP_DOLLM: + case OP_EOD: + case OP_EODN: + case OP_FAIL: + case OP_NOT_WORD_BOUNDARY: + case OP_PRUNE: + case OP_REVERSE: + case OP_RREF: + case OP_SET_SOM: + case OP_SKIP: + case OP_SOD: + case OP_SOM: + case OP_THEN: + case OP_WORD_BOUNDARY: + cc += PRIV(OP_lengths)[*cc]; + break; + + /* Handle literal characters */ + + case OP_CHAR: + case OP_CHARI: + case OP_NOT: + case OP_NOTI: + branchlength++; + cc += 2; +#ifdef SUPPORT_UTF + if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]); +#endif + break; + + /* Handle exact repetitions. The count is already in characters, but we + need to skip over a multibyte character in UTF8 mode. */ + + case OP_EXACT: + case OP_EXACTI: + case OP_NOTEXACT: + case OP_NOTEXACTI: + branchlength += (int)GET2(cc,1); + cc += 2 + IMM2_SIZE; +#ifdef SUPPORT_UTF + if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]); +#endif + break; + + case OP_TYPEEXACT: + branchlength += GET2(cc,1); + if (cc[1 + IMM2_SIZE] == OP_PROP || cc[1 + IMM2_SIZE] == OP_NOTPROP) + cc += 2; + cc += 1 + IMM2_SIZE + 1; + break; + + /* Handle single-char matchers */ + + case OP_PROP: + case OP_NOTPROP: + cc += 2; + /* Fall through */ + + case OP_HSPACE: + case OP_VSPACE: + case OP_NOT_HSPACE: + case OP_NOT_VSPACE: + case OP_NOT_DIGIT: + case OP_DIGIT: + case OP_NOT_WHITESPACE: + case OP_WHITESPACE: + case OP_NOT_WORDCHAR: + case OP_WORDCHAR: + case OP_ANY: + case OP_ALLANY: + branchlength++; + cc++; + break; + + /* The single-byte matcher isn't allowed. This only happens in UTF-8 mode; + otherwise \C is coded as OP_ALLANY. */ + + case OP_ANYBYTE: + return -2; + + /* Check a class for variable quantification */ + + case OP_CLASS: + case OP_NCLASS: +#if defined SUPPORT_UTF || defined COMPILE_PCRE16 || defined COMPILE_PCRE32 + case OP_XCLASS: + /* The original code caused an unsigned overflow in 64 bit systems, + so now we use a conditional statement. */ + if (op == OP_XCLASS) + cc += GET(cc, 1); + else + cc += PRIV(OP_lengths)[OP_CLASS]; +#else + cc += PRIV(OP_lengths)[OP_CLASS]; +#endif + + switch (*cc) + { + case OP_CRSTAR: + case OP_CRMINSTAR: + case OP_CRPLUS: + case OP_CRMINPLUS: + case OP_CRQUERY: + case OP_CRMINQUERY: + case OP_CRPOSSTAR: + case OP_CRPOSPLUS: + case OP_CRPOSQUERY: + return -1; + + case OP_CRRANGE: + case OP_CRMINRANGE: + case OP_CRPOSRANGE: + if (GET2(cc,1) != GET2(cc,1+IMM2_SIZE)) return -1; + branchlength += (int)GET2(cc,1); + cc += 1 + 2 * IMM2_SIZE; + break; + + default: + branchlength++; + } + break; + + /* Anything else is variable length */ + + case OP_ANYNL: + case OP_BRAMINZERO: + case OP_BRAPOS: + case OP_BRAPOSZERO: + case OP_BRAZERO: + case OP_CBRAPOS: + case OP_EXTUNI: + case OP_KETRMAX: + case OP_KETRMIN: + case OP_KETRPOS: + case OP_MINPLUS: + case OP_MINPLUSI: + case OP_MINQUERY: + case OP_MINQUERYI: + case OP_MINSTAR: + case OP_MINSTARI: + case OP_MINUPTO: + case OP_MINUPTOI: + case OP_NOTMINPLUS: + case OP_NOTMINPLUSI: + case OP_NOTMINQUERY: + case OP_NOTMINQUERYI: + case OP_NOTMINSTAR: + case OP_NOTMINSTARI: + case OP_NOTMINUPTO: + case OP_NOTMINUPTOI: + case OP_NOTPLUS: + case OP_NOTPLUSI: + case OP_NOTPOSPLUS: + case OP_NOTPOSPLUSI: + case OP_NOTPOSQUERY: + case OP_NOTPOSQUERYI: + case OP_NOTPOSSTAR: + case OP_NOTPOSSTARI: + case OP_NOTPOSUPTO: + case OP_NOTPOSUPTOI: + case OP_NOTQUERY: + case OP_NOTQUERYI: + case OP_NOTSTAR: + case OP_NOTSTARI: + case OP_NOTUPTO: + case OP_NOTUPTOI: + case OP_PLUS: + case OP_PLUSI: + case OP_POSPLUS: + case OP_POSPLUSI: + case OP_POSQUERY: + case OP_POSQUERYI: + case OP_POSSTAR: + case OP_POSSTARI: + case OP_POSUPTO: + case OP_POSUPTOI: + case OP_QUERY: + case OP_QUERYI: + case OP_REF: + case OP_REFI: + case OP_DNREF: + case OP_DNREFI: + case OP_SBRA: + case OP_SBRAPOS: + case OP_SCBRA: + case OP_SCBRAPOS: + case OP_SCOND: + case OP_SKIPZERO: + case OP_STAR: + case OP_STARI: + case OP_TYPEMINPLUS: + case OP_TYPEMINQUERY: + case OP_TYPEMINSTAR: + case OP_TYPEMINUPTO: + case OP_TYPEPLUS: + case OP_TYPEPOSPLUS: + case OP_TYPEPOSQUERY: + case OP_TYPEPOSSTAR: + case OP_TYPEPOSUPTO: + case OP_TYPEQUERY: + case OP_TYPESTAR: + case OP_TYPEUPTO: + case OP_UPTO: + case OP_UPTOI: + return -1; + + /* Catch unrecognized opcodes so that when new ones are added they + are not forgotten, as has happened in the past. */ + + default: + return -4; + } + } +/* Control never gets here */ +} + + + +/************************************************* +* Scan compiled regex for specific bracket * +*************************************************/ + +/* This little function scans through a compiled pattern until it finds a +capturing bracket with the given number, or, if the number is negative, an +instance of OP_REVERSE for a lookbehind. The function is global in the C sense +so that it can be called from pcre_study() when finding the minimum matching +length. + +Arguments: + code points to start of expression + utf TRUE in UTF-8 / UTF-16 / UTF-32 mode + number the required bracket number or negative to find a lookbehind + +Returns: pointer to the opcode for the bracket, or NULL if not found +*/ + +const pcre_uchar * +PRIV(find_bracket)(const pcre_uchar *code, BOOL utf, int number) +{ +for (;;) + { + register pcre_uchar c = *code; + + if (c == OP_END) return NULL; + + /* XCLASS is used for classes that cannot be represented just by a bit + map. This includes negated single high-valued characters. The length in + the table is zero; the actual length is stored in the compiled code. */ + + if (c == OP_XCLASS) code += GET(code, 1); + + /* Handle recursion */ + + else if (c == OP_REVERSE) + { + if (number < 0) return (pcre_uchar *)code; + code += PRIV(OP_lengths)[c]; + } + + /* Handle capturing bracket */ + + else if (c == OP_CBRA || c == OP_SCBRA || + c == OP_CBRAPOS || c == OP_SCBRAPOS) + { + int n = (int)GET2(code, 1+LINK_SIZE); + if (n == number) return (pcre_uchar *)code; + code += PRIV(OP_lengths)[c]; + } + + /* Otherwise, we can get the item's length from the table, except that for + repeated character types, we have to test for \p and \P, which have an extra + two bytes of parameters, and for MARK/PRUNE/SKIP/THEN with an argument, we + must add in its length. */ + + else + { + switch(c) + { + case OP_TYPESTAR: + case OP_TYPEMINSTAR: + case OP_TYPEPLUS: + case OP_TYPEMINPLUS: + case OP_TYPEQUERY: + case OP_TYPEMINQUERY: + case OP_TYPEPOSSTAR: + case OP_TYPEPOSPLUS: + case OP_TYPEPOSQUERY: + if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2; + break; + + case OP_TYPEUPTO: + case OP_TYPEMINUPTO: + case OP_TYPEEXACT: + case OP_TYPEPOSUPTO: + if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP) + code += 2; + break; + + case OP_MARK: + case OP_PRUNE_ARG: + case OP_SKIP_ARG: + case OP_THEN_ARG: + code += code[1]; + break; + } + + /* Add in the fixed length from the table */ + + code += PRIV(OP_lengths)[c]; + + /* In UTF-8 mode, opcodes that are followed by a character may be followed by + a multi-byte character. The length in the table is a minimum, so we have to + arrange to skip the extra bytes. */ + +#if defined SUPPORT_UTF && !defined COMPILE_PCRE32 + if (utf) switch(c) + { + case OP_CHAR: + case OP_CHARI: + case OP_NOT: + case OP_NOTI: + case OP_EXACT: + case OP_EXACTI: + case OP_NOTEXACT: + case OP_NOTEXACTI: + case OP_UPTO: + case OP_UPTOI: + case OP_NOTUPTO: + case OP_NOTUPTOI: + case OP_MINUPTO: + case OP_MINUPTOI: + case OP_NOTMINUPTO: + case OP_NOTMINUPTOI: + case OP_POSUPTO: + case OP_POSUPTOI: + case OP_NOTPOSUPTO: + case OP_NOTPOSUPTOI: + case OP_STAR: + case OP_STARI: + case OP_NOTSTAR: + case OP_NOTSTARI: + case OP_MINSTAR: + case OP_MINSTARI: + case OP_NOTMINSTAR: + case OP_NOTMINSTARI: + case OP_POSSTAR: + case OP_POSSTARI: + case OP_NOTPOSSTAR: + case OP_NOTPOSSTARI: + case OP_PLUS: + case OP_PLUSI: + case OP_NOTPLUS: + case OP_NOTPLUSI: + case OP_MINPLUS: + case OP_MINPLUSI: + case OP_NOTMINPLUS: + case OP_NOTMINPLUSI: + case OP_POSPLUS: + case OP_POSPLUSI: + case OP_NOTPOSPLUS: + case OP_NOTPOSPLUSI: + case OP_QUERY: + case OP_QUERYI: + case OP_NOTQUERY: + case OP_NOTQUERYI: + case OP_MINQUERY: + case OP_MINQUERYI: + case OP_NOTMINQUERY: + case OP_NOTMINQUERYI: + case OP_POSQUERY: + case OP_POSQUERYI: + case OP_NOTPOSQUERY: + case OP_NOTPOSQUERYI: + if (HAS_EXTRALEN(code[-1])) code += GET_EXTRALEN(code[-1]); + break; + } +#else + (void)(utf); /* Keep compiler happy by referencing function argument */ +#endif + } + } +} + + + +/************************************************* +* Scan compiled regex for recursion reference * +*************************************************/ + +/* This little function scans through a compiled pattern until it finds an +instance of OP_RECURSE. + +Arguments: + code points to start of expression + utf TRUE in UTF-8 / UTF-16 / UTF-32 mode + +Returns: pointer to the opcode for OP_RECURSE, or NULL if not found +*/ + +static const pcre_uchar * +find_recurse(const pcre_uchar *code, BOOL utf) +{ +for (;;) + { + register pcre_uchar c = *code; + if (c == OP_END) return NULL; + if (c == OP_RECURSE) return code; + + /* XCLASS is used for classes that cannot be represented just by a bit + map. This includes negated single high-valued characters. The length in + the table is zero; the actual length is stored in the compiled code. */ + + if (c == OP_XCLASS) code += GET(code, 1); + + /* Otherwise, we can get the item's length from the table, except that for + repeated character types, we have to test for \p and \P, which have an extra + two bytes of parameters, and for MARK/PRUNE/SKIP/THEN with an argument, we + must add in its length. */ + + else + { + switch(c) + { + case OP_TYPESTAR: + case OP_TYPEMINSTAR: + case OP_TYPEPLUS: + case OP_TYPEMINPLUS: + case OP_TYPEQUERY: + case OP_TYPEMINQUERY: + case OP_TYPEPOSSTAR: + case OP_TYPEPOSPLUS: + case OP_TYPEPOSQUERY: + if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2; + break; + + case OP_TYPEPOSUPTO: + case OP_TYPEUPTO: + case OP_TYPEMINUPTO: + case OP_TYPEEXACT: + if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP) + code += 2; + break; + + case OP_MARK: + case OP_PRUNE_ARG: + case OP_SKIP_ARG: + case OP_THEN_ARG: + code += code[1]; + break; + } + + /* Add in the fixed length from the table */ + + code += PRIV(OP_lengths)[c]; + + /* In UTF-8 mode, opcodes that are followed by a character may be followed + by a multi-byte character. The length in the table is a minimum, so we have + to arrange to skip the extra bytes. */ + +#if defined SUPPORT_UTF && !defined COMPILE_PCRE32 + if (utf) switch(c) + { + case OP_CHAR: + case OP_CHARI: + case OP_NOT: + case OP_NOTI: + case OP_EXACT: + case OP_EXACTI: + case OP_NOTEXACT: + case OP_NOTEXACTI: + case OP_UPTO: + case OP_UPTOI: + case OP_NOTUPTO: + case OP_NOTUPTOI: + case OP_MINUPTO: + case OP_MINUPTOI: + case OP_NOTMINUPTO: + case OP_NOTMINUPTOI: + case OP_POSUPTO: + case OP_POSUPTOI: + case OP_NOTPOSUPTO: + case OP_NOTPOSUPTOI: + case OP_STAR: + case OP_STARI: + case OP_NOTSTAR: + case OP_NOTSTARI: + case OP_MINSTAR: + case OP_MINSTARI: + case OP_NOTMINSTAR: + case OP_NOTMINSTARI: + case OP_POSSTAR: + case OP_POSSTARI: + case OP_NOTPOSSTAR: + case OP_NOTPOSSTARI: + case OP_PLUS: + case OP_PLUSI: + case OP_NOTPLUS: + case OP_NOTPLUSI: + case OP_MINPLUS: + case OP_MINPLUSI: + case OP_NOTMINPLUS: + case OP_NOTMINPLUSI: + case OP_POSPLUS: + case OP_POSPLUSI: + case OP_NOTPOSPLUS: + case OP_NOTPOSPLUSI: + case OP_QUERY: + case OP_QUERYI: + case OP_NOTQUERY: + case OP_NOTQUERYI: + case OP_MINQUERY: + case OP_MINQUERYI: + case OP_NOTMINQUERY: + case OP_NOTMINQUERYI: + case OP_POSQUERY: + case OP_POSQUERYI: + case OP_NOTPOSQUERY: + case OP_NOTPOSQUERYI: + if (HAS_EXTRALEN(code[-1])) code += GET_EXTRALEN(code[-1]); + break; + } +#else + (void)(utf); /* Keep compiler happy by referencing function argument */ +#endif + } + } +} + + + +/************************************************* +* Scan compiled branch for non-emptiness * +*************************************************/ + +/* This function scans through a branch of a compiled pattern to see whether it +can match the empty string or not. It is called from could_be_empty() +below and from compile_branch() when checking for an unlimited repeat of a +group that can match nothing. Note that first_significant_code() skips over +backward and negative forward assertions when its final argument is TRUE. If we +hit an unclosed bracket, we return "empty" - this means we've struck an inner +bracket whose current branch will already have been scanned. + +Arguments: + code points to start of search + endcode points to where to stop + utf TRUE if in UTF-8 / UTF-16 / UTF-32 mode + cd contains pointers to tables etc. + recurses chain of recurse_check to catch mutual recursion + +Returns: TRUE if what is matched could be empty +*/ + +static BOOL +could_be_empty_branch(const pcre_uchar *code, const pcre_uchar *endcode, + BOOL utf, compile_data *cd, recurse_check *recurses) +{ +register pcre_uchar c; +recurse_check this_recurse; + +for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE); + code < endcode; + code = first_significant_code(code + PRIV(OP_lengths)[c], TRUE)) + { + const pcre_uchar *ccode; + + c = *code; + + /* Skip over forward assertions; the other assertions are skipped by + first_significant_code() with a TRUE final argument. */ + + if (c == OP_ASSERT) + { + do code += GET(code, 1); while (*code == OP_ALT); + c = *code; + continue; + } + + /* For a recursion/subroutine call, if its end has been reached, which + implies a backward reference subroutine call, we can scan it. If it's a + forward reference subroutine call, we can't. To detect forward reference + we have to scan up the list that is kept in the workspace. This function is + called only when doing the real compile, not during the pre-compile that + measures the size of the compiled pattern. */ + + if (c == OP_RECURSE) + { + const pcre_uchar *scode = cd->start_code + GET(code, 1); + const pcre_uchar *endgroup = scode; + BOOL empty_branch; + + /* Test for forward reference or uncompleted reference. This is disabled + when called to scan a completed pattern by setting cd->start_workspace to + NULL. */ + + if (cd->start_workspace != NULL) + { + const pcre_uchar *tcode; + for (tcode = cd->start_workspace; tcode < cd->hwm; tcode += LINK_SIZE) + if ((int)GET(tcode, 0) == (int)(code + 1 - cd->start_code)) return TRUE; + if (GET(scode, 1) == 0) return TRUE; /* Unclosed */ + } + + /* If the reference is to a completed group, we need to detect whether this + is a recursive call, as otherwise there will be an infinite loop. If it is + a recursion, just skip over it. Simple recursions are easily detected. For + mutual recursions we keep a chain on the stack. */ + + do endgroup += GET(endgroup, 1); while (*endgroup == OP_ALT); + if (code >= scode && code <= endgroup) continue; /* Simple recursion */ + else + { + recurse_check *r = recurses; + for (r = recurses; r != NULL; r = r->prev) + if (r->group == scode) break; + if (r != NULL) continue; /* Mutual recursion */ + } + + /* Completed reference; scan the referenced group, remembering it on the + stack chain to detect mutual recursions. */ + + empty_branch = FALSE; + this_recurse.prev = recurses; + this_recurse.group = scode; + + do + { + if (could_be_empty_branch(scode, endcode, utf, cd, &this_recurse)) + { + empty_branch = TRUE; + break; + } + scode += GET(scode, 1); + } + while (*scode == OP_ALT); + + if (!empty_branch) return FALSE; /* All branches are non-empty */ + continue; + } + + /* Groups with zero repeats can of course be empty; skip them. */ + + if (c == OP_BRAZERO || c == OP_BRAMINZERO || c == OP_SKIPZERO || + c == OP_BRAPOSZERO) + { + code += PRIV(OP_lengths)[c]; + do code += GET(code, 1); while (*code == OP_ALT); + c = *code; + continue; + } + + /* A nested group that is already marked as "could be empty" can just be + skipped. */ + + if (c == OP_SBRA || c == OP_SBRAPOS || + c == OP_SCBRA || c == OP_SCBRAPOS) + { + do code += GET(code, 1); while (*code == OP_ALT); + c = *code; + continue; + } + + /* For other groups, scan the branches. */ + + if (c == OP_BRA || c == OP_BRAPOS || + c == OP_CBRA || c == OP_CBRAPOS || + c == OP_ONCE || c == OP_ONCE_NC || + c == OP_COND || c == OP_SCOND) + { + BOOL empty_branch; + if (GET(code, 1) == 0) return TRUE; /* Hit unclosed bracket */ + + /* If a conditional group has only one branch, there is a second, implied, + empty branch, so just skip over the conditional, because it could be empty. + Otherwise, scan the individual branches of the group. */ + + if (c == OP_COND && code[GET(code, 1)] != OP_ALT) + code += GET(code, 1); + else + { + empty_branch = FALSE; + do + { + if (!empty_branch && could_be_empty_branch(code, endcode, utf, cd, + recurses)) empty_branch = TRUE; + code += GET(code, 1); + } + while (*code == OP_ALT); + if (!empty_branch) return FALSE; /* All branches are non-empty */ + } + + c = *code; + continue; + } + + /* Handle the other opcodes */ + + switch (c) + { + /* Check for quantifiers after a class. XCLASS is used for classes that + cannot be represented just by a bit map. This includes negated single + high-valued characters. The length in PRIV(OP_lengths)[] is zero; the + actual length is stored in the compiled code, so we must update "code" + here. */ + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + case OP_XCLASS: + ccode = code += GET(code, 1); + goto CHECK_CLASS_REPEAT; +#endif + + case OP_CLASS: + case OP_NCLASS: + ccode = code + PRIV(OP_lengths)[OP_CLASS]; + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + CHECK_CLASS_REPEAT: +#endif + + switch (*ccode) + { + case OP_CRSTAR: /* These could be empty; continue */ + case OP_CRMINSTAR: + case OP_CRQUERY: + case OP_CRMINQUERY: + case OP_CRPOSSTAR: + case OP_CRPOSQUERY: + break; + + default: /* Non-repeat => class must match */ + case OP_CRPLUS: /* These repeats aren't empty */ + case OP_CRMINPLUS: + case OP_CRPOSPLUS: + return FALSE; + + case OP_CRRANGE: + case OP_CRMINRANGE: + case OP_CRPOSRANGE: + if (GET2(ccode, 1) > 0) return FALSE; /* Minimum > 0 */ + break; + } + break; + + /* Opcodes that must match a character */ + + case OP_ANY: + case OP_ALLANY: + case OP_ANYBYTE: + + case OP_PROP: + case OP_NOTPROP: + case OP_ANYNL: + + case OP_NOT_HSPACE: + case OP_HSPACE: + case OP_NOT_VSPACE: + case OP_VSPACE: + case OP_EXTUNI: + + case OP_NOT_DIGIT: + case OP_DIGIT: + case OP_NOT_WHITESPACE: + case OP_WHITESPACE: + case OP_NOT_WORDCHAR: + case OP_WORDCHAR: + + case OP_CHAR: + case OP_CHARI: + case OP_NOT: + case OP_NOTI: + + case OP_PLUS: + case OP_PLUSI: + case OP_MINPLUS: + case OP_MINPLUSI: + + case OP_NOTPLUS: + case OP_NOTPLUSI: + case OP_NOTMINPLUS: + case OP_NOTMINPLUSI: + + case OP_POSPLUS: + case OP_POSPLUSI: + case OP_NOTPOSPLUS: + case OP_NOTPOSPLUSI: + + case OP_EXACT: + case OP_EXACTI: + case OP_NOTEXACT: + case OP_NOTEXACTI: + + case OP_TYPEPLUS: + case OP_TYPEMINPLUS: + case OP_TYPEPOSPLUS: + case OP_TYPEEXACT: + + return FALSE; + + /* These are going to continue, as they may be empty, but we have to + fudge the length for the \p and \P cases. */ + + case OP_TYPESTAR: + case OP_TYPEMINSTAR: + case OP_TYPEPOSSTAR: + case OP_TYPEQUERY: + case OP_TYPEMINQUERY: + case OP_TYPEPOSQUERY: + if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2; + break; + + /* Same for these */ + + case OP_TYPEUPTO: + case OP_TYPEMINUPTO: + case OP_TYPEPOSUPTO: + if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP) + code += 2; + break; + + /* End of branch */ + + case OP_KET: + case OP_KETRMAX: + case OP_KETRMIN: + case OP_KETRPOS: + case OP_ALT: + return TRUE; + + /* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO, + MINUPTO, and POSUPTO and their caseless and negative versions may be + followed by a multibyte character. */ + +#if defined SUPPORT_UTF && !defined COMPILE_PCRE32 + case OP_STAR: + case OP_STARI: + case OP_NOTSTAR: + case OP_NOTSTARI: + + case OP_MINSTAR: + case OP_MINSTARI: + case OP_NOTMINSTAR: + case OP_NOTMINSTARI: + + case OP_POSSTAR: + case OP_POSSTARI: + case OP_NOTPOSSTAR: + case OP_NOTPOSSTARI: + + case OP_QUERY: + case OP_QUERYI: + case OP_NOTQUERY: + case OP_NOTQUERYI: + + case OP_MINQUERY: + case OP_MINQUERYI: + case OP_NOTMINQUERY: + case OP_NOTMINQUERYI: + + case OP_POSQUERY: + case OP_POSQUERYI: + case OP_NOTPOSQUERY: + case OP_NOTPOSQUERYI: + + if (utf && HAS_EXTRALEN(code[1])) code += GET_EXTRALEN(code[1]); + break; + + case OP_UPTO: + case OP_UPTOI: + case OP_NOTUPTO: + case OP_NOTUPTOI: + + case OP_MINUPTO: + case OP_MINUPTOI: + case OP_NOTMINUPTO: + case OP_NOTMINUPTOI: + + case OP_POSUPTO: + case OP_POSUPTOI: + case OP_NOTPOSUPTO: + case OP_NOTPOSUPTOI: + + if (utf && HAS_EXTRALEN(code[1 + IMM2_SIZE])) code += GET_EXTRALEN(code[1 + IMM2_SIZE]); + break; +#endif + + /* MARK, and PRUNE/SKIP/THEN with an argument must skip over the argument + string. */ + + case OP_MARK: + case OP_PRUNE_ARG: + case OP_SKIP_ARG: + case OP_THEN_ARG: + code += code[1]; + break; + + /* None of the remaining opcodes are required to match a character. */ + + default: + break; + } + } + +return TRUE; +} + + + +/************************************************* +* Scan compiled regex for non-emptiness * +*************************************************/ + +/* This function is called to check for left recursive calls. We want to check +the current branch of the current pattern to see if it could match the empty +string. If it could, we must look outwards for branches at other levels, +stopping when we pass beyond the bracket which is the subject of the recursion. +This function is called only during the real compile, not during the +pre-compile. + +Arguments: + code points to start of the recursion + endcode points to where to stop (current RECURSE item) + bcptr points to the chain of current (unclosed) branch starts + utf TRUE if in UTF-8 / UTF-16 / UTF-32 mode + cd pointers to tables etc + +Returns: TRUE if what is matched could be empty +*/ + +static BOOL +could_be_empty(const pcre_uchar *code, const pcre_uchar *endcode, + branch_chain *bcptr, BOOL utf, compile_data *cd) +{ +while (bcptr != NULL && bcptr->current_branch >= code) + { + if (!could_be_empty_branch(bcptr->current_branch, endcode, utf, cd, NULL)) + return FALSE; + bcptr = bcptr->outer; + } +return TRUE; +} + + + +/************************************************* +* Base opcode of repeated opcodes * +*************************************************/ + +/* Returns the base opcode for repeated single character type opcodes. If the +opcode is not a repeated character type, it returns with the original value. + +Arguments: c opcode +Returns: base opcode for the type +*/ + +static pcre_uchar +get_repeat_base(pcre_uchar c) +{ +return (c > OP_TYPEPOSUPTO)? c : + (c >= OP_TYPESTAR)? OP_TYPESTAR : + (c >= OP_NOTSTARI)? OP_NOTSTARI : + (c >= OP_NOTSTAR)? OP_NOTSTAR : + (c >= OP_STARI)? OP_STARI : + OP_STAR; +} + + + +#ifdef SUPPORT_UCP +/************************************************* +* Check a character and a property * +*************************************************/ + +/* This function is called by check_auto_possessive() when a property item +is adjacent to a fixed character. + +Arguments: + c the character + ptype the property type + pdata the data for the type + negated TRUE if it's a negated property (\P or \p{^) + +Returns: TRUE if auto-possessifying is OK +*/ + +static BOOL +check_char_prop(pcre_uint32 c, unsigned int ptype, unsigned int pdata, + BOOL negated) +{ +const pcre_uint32 *p; +const ucd_record *prop = GET_UCD(c); + +switch(ptype) + { + case PT_LAMP: + return (prop->chartype == ucp_Lu || + prop->chartype == ucp_Ll || + prop->chartype == ucp_Lt) == negated; + + case PT_GC: + return (pdata == PRIV(ucp_gentype)[prop->chartype]) == negated; + + case PT_PC: + return (pdata == prop->chartype) == negated; + + case PT_SC: + return (pdata == prop->script) == negated; + + /* These are specials */ + + case PT_ALNUM: + return (PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N) == negated; + + /* Perl space used to exclude VT, but from Perl 5.18 it is included, which + means that Perl space and POSIX space are now identical. PCRE was changed + at release 8.34. */ + + case PT_SPACE: /* Perl space */ + case PT_PXSPACE: /* POSIX space */ + switch(c) + { + HSPACE_CASES: + VSPACE_CASES: + return negated; + + default: + return (PRIV(ucp_gentype)[prop->chartype] == ucp_Z) == negated; + } + break; /* Control never reaches here */ + + case PT_WORD: + return (PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N || + c == CHAR_UNDERSCORE) == negated; + + case PT_CLIST: + p = PRIV(ucd_caseless_sets) + prop->caseset; + for (;;) + { + if (c < *p) return !negated; + if (c == *p++) return negated; + } + break; /* Control never reaches here */ + } + +return FALSE; +} +#endif /* SUPPORT_UCP */ + + + +/************************************************* +* Fill the character property list * +*************************************************/ + +/* Checks whether the code points to an opcode that can take part in auto- +possessification, and if so, fills a list with its properties. + +Arguments: + code points to start of expression + utf TRUE if in UTF-8 / UTF-16 / UTF-32 mode + fcc points to case-flipping table + list points to output list + list[0] will be filled with the opcode + list[1] will be non-zero if this opcode + can match an empty character string + list[2..7] depends on the opcode + +Returns: points to the start of the next opcode if *code is accepted + NULL if *code is not accepted +*/ + +static const pcre_uchar * +get_chr_property_list(const pcre_uchar *code, BOOL utf, + const pcre_uint8 *fcc, pcre_uint32 *list) +{ +pcre_uchar c = *code; +pcre_uchar base; +const pcre_uchar *end; +pcre_uint32 chr; + +#ifdef SUPPORT_UCP +pcre_uint32 *clist_dest; +const pcre_uint32 *clist_src; +#else +utf = utf; /* Suppress "unused parameter" compiler warning */ +#endif + +list[0] = c; +list[1] = FALSE; +code++; + +if (c >= OP_STAR && c <= OP_TYPEPOSUPTO) + { + base = get_repeat_base(c); + c -= (base - OP_STAR); + + if (c == OP_UPTO || c == OP_MINUPTO || c == OP_EXACT || c == OP_POSUPTO) + code += IMM2_SIZE; + + list[1] = (c != OP_PLUS && c != OP_MINPLUS && c != OP_EXACT && c != OP_POSPLUS); + + switch(base) + { + case OP_STAR: + list[0] = OP_CHAR; + break; + + case OP_STARI: + list[0] = OP_CHARI; + break; + + case OP_NOTSTAR: + list[0] = OP_NOT; + break; + + case OP_NOTSTARI: + list[0] = OP_NOTI; + break; + + case OP_TYPESTAR: + list[0] = *code; + code++; + break; + } + c = list[0]; + } + +switch(c) + { + case OP_NOT_DIGIT: + case OP_DIGIT: + case OP_NOT_WHITESPACE: + case OP_WHITESPACE: + case OP_NOT_WORDCHAR: + case OP_WORDCHAR: + case OP_ANY: + case OP_ALLANY: + case OP_ANYNL: + case OP_NOT_HSPACE: + case OP_HSPACE: + case OP_NOT_VSPACE: + case OP_VSPACE: + case OP_EXTUNI: + case OP_EODN: + case OP_EOD: + case OP_DOLL: + case OP_DOLLM: + return code; + + case OP_CHAR: + case OP_NOT: + GETCHARINCTEST(chr, code); + list[2] = chr; + list[3] = NOTACHAR; + return code; + + case OP_CHARI: + case OP_NOTI: + list[0] = (c == OP_CHARI) ? OP_CHAR : OP_NOT; + GETCHARINCTEST(chr, code); + list[2] = chr; + +#ifdef SUPPORT_UCP + if (chr < 128 || (chr < 256 && !utf)) + list[3] = fcc[chr]; + else + list[3] = UCD_OTHERCASE(chr); +#elif defined SUPPORT_UTF || !defined COMPILE_PCRE8 + list[3] = (chr < 256) ? fcc[chr] : chr; +#else + list[3] = fcc[chr]; +#endif + + /* The othercase might be the same value. */ + + if (chr == list[3]) + list[3] = NOTACHAR; + else + list[4] = NOTACHAR; + return code; + +#ifdef SUPPORT_UCP + case OP_PROP: + case OP_NOTPROP: + if (code[0] != PT_CLIST) + { + list[2] = code[0]; + list[3] = code[1]; + return code + 2; + } + + /* Convert only if we have enough space. */ + + clist_src = PRIV(ucd_caseless_sets) + code[1]; + clist_dest = list + 2; + code += 2; + + do { + if (clist_dest >= list + 8) + { + /* Early return if there is not enough space. This should never + happen, since all clists are shorter than 5 character now. */ + list[2] = code[0]; + list[3] = code[1]; + return code; + } + *clist_dest++ = *clist_src; + } + while(*clist_src++ != NOTACHAR); + + /* All characters are stored. The terminating NOTACHAR + is copied form the clist itself. */ + + list[0] = (c == OP_PROP) ? OP_CHAR : OP_NOT; + return code; +#endif + + case OP_NCLASS: + case OP_CLASS: +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + case OP_XCLASS: + if (c == OP_XCLASS) + end = code + GET(code, 0) - 1; + else +#endif + end = code + 32 / sizeof(pcre_uchar); + + switch(*end) + { + case OP_CRSTAR: + case OP_CRMINSTAR: + case OP_CRQUERY: + case OP_CRMINQUERY: + case OP_CRPOSSTAR: + case OP_CRPOSQUERY: + list[1] = TRUE; + end++; + break; + + case OP_CRPLUS: + case OP_CRMINPLUS: + case OP_CRPOSPLUS: + end++; + break; + + case OP_CRRANGE: + case OP_CRMINRANGE: + case OP_CRPOSRANGE: + list[1] = (GET2(end, 1) == 0); + end += 1 + 2 * IMM2_SIZE; + break; + } + list[2] = (pcre_uint32)(end - code); + return end; + } +return NULL; /* Opcode not accepted */ +} + + + +/************************************************* +* Scan further character sets for match * +*************************************************/ + +/* Checks whether the base and the current opcode have a common character, in +which case the base cannot be possessified. + +Arguments: + code points to the byte code + utf TRUE in UTF-8 / UTF-16 / UTF-32 mode + cd static compile data + base_list the data list of the base opcode + +Returns: TRUE if the auto-possessification is possible +*/ + +static BOOL +compare_opcodes(const pcre_uchar *code, BOOL utf, const compile_data *cd, + const pcre_uint32 *base_list, const pcre_uchar *base_end, int *rec_limit) +{ +pcre_uchar c; +pcre_uint32 list[8]; +const pcre_uint32 *chr_ptr; +const pcre_uint32 *ochr_ptr; +const pcre_uint32 *list_ptr; +const pcre_uchar *next_code; +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 +const pcre_uchar *xclass_flags; +#endif +const pcre_uint8 *class_bitset; +const pcre_uint8 *set1, *set2, *set_end; +pcre_uint32 chr; +BOOL accepted, invert_bits; +BOOL entered_a_group = FALSE; + +if (*rec_limit == 0) return FALSE; +--(*rec_limit); + +/* Note: the base_list[1] contains whether the current opcode has greedy +(represented by a non-zero value) quantifier. This is a different from +other character type lists, which stores here that the character iterator +matches to an empty string (also represented by a non-zero value). */ + +for(;;) + { + /* All operations move the code pointer forward. + Therefore infinite recursions are not possible. */ + + c = *code; + + /* Skip over callouts */ + + if (c == OP_CALLOUT) + { + code += PRIV(OP_lengths)[c]; + continue; + } + + if (c == OP_ALT) + { + do code += GET(code, 1); while (*code == OP_ALT); + c = *code; + } + + switch(c) + { + case OP_END: + case OP_KETRPOS: + /* TRUE only in greedy case. The non-greedy case could be replaced by + an OP_EXACT, but it is probably not worth it. (And note that OP_EXACT + uses more memory, which we cannot get at this stage.) */ + + return base_list[1] != 0; + + case OP_KET: + /* If the bracket is capturing, and referenced by an OP_RECURSE, or + it is an atomic sub-pattern (assert, once, etc.) the non-greedy case + cannot be converted to a possessive form. */ + + if (base_list[1] == 0) return FALSE; + + switch(*(code - GET(code, 1))) + { + case OP_ASSERT: + case OP_ASSERT_NOT: + case OP_ASSERTBACK: + case OP_ASSERTBACK_NOT: + case OP_ONCE: + case OP_ONCE_NC: + /* Atomic sub-patterns and assertions can always auto-possessify their + last iterator. However, if the group was entered as a result of checking + a previous iterator, this is not possible. */ + + return !entered_a_group; + } + + code += PRIV(OP_lengths)[c]; + continue; + + case OP_ONCE: + case OP_ONCE_NC: + case OP_BRA: + case OP_CBRA: + next_code = code + GET(code, 1); + code += PRIV(OP_lengths)[c]; + + while (*next_code == OP_ALT) + { + if (!compare_opcodes(code, utf, cd, base_list, base_end, rec_limit)) + return FALSE; + code = next_code + 1 + LINK_SIZE; + next_code += GET(next_code, 1); + } + + entered_a_group = TRUE; + continue; + + case OP_BRAZERO: + case OP_BRAMINZERO: + + next_code = code + 1; + if (*next_code != OP_BRA && *next_code != OP_CBRA + && *next_code != OP_ONCE && *next_code != OP_ONCE_NC) return FALSE; + + do next_code += GET(next_code, 1); while (*next_code == OP_ALT); + + /* The bracket content will be checked by the + OP_BRA/OP_CBRA case above. */ + next_code += 1 + LINK_SIZE; + if (!compare_opcodes(next_code, utf, cd, base_list, base_end, rec_limit)) + return FALSE; + + code += PRIV(OP_lengths)[c]; + continue; + + default: + break; + } + + /* Check for a supported opcode, and load its properties. */ + + code = get_chr_property_list(code, utf, cd->fcc, list); + if (code == NULL) return FALSE; /* Unsupported */ + + /* If either opcode is a small character list, set pointers for comparing + characters from that list with another list, or with a property. */ + + if (base_list[0] == OP_CHAR) + { + chr_ptr = base_list + 2; + list_ptr = list; + } + else if (list[0] == OP_CHAR) + { + chr_ptr = list + 2; + list_ptr = base_list; + } + + /* Character bitsets can also be compared to certain opcodes. */ + + else if (base_list[0] == OP_CLASS || list[0] == OP_CLASS +#ifdef COMPILE_PCRE8 + /* In 8 bit, non-UTF mode, OP_CLASS and OP_NCLASS are the same. */ + || (!utf && (base_list[0] == OP_NCLASS || list[0] == OP_NCLASS)) +#endif + ) + { +#ifdef COMPILE_PCRE8 + if (base_list[0] == OP_CLASS || (!utf && base_list[0] == OP_NCLASS)) +#else + if (base_list[0] == OP_CLASS) +#endif + { + set1 = (pcre_uint8 *)(base_end - base_list[2]); + list_ptr = list; + } + else + { + set1 = (pcre_uint8 *)(code - list[2]); + list_ptr = base_list; + } + + invert_bits = FALSE; + switch(list_ptr[0]) + { + case OP_CLASS: + case OP_NCLASS: + set2 = (pcre_uint8 *) + ((list_ptr == list ? code : base_end) - list_ptr[2]); + break; + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + case OP_XCLASS: + xclass_flags = (list_ptr == list ? code : base_end) - list_ptr[2] + LINK_SIZE; + if ((*xclass_flags & XCL_HASPROP) != 0) return FALSE; + if ((*xclass_flags & XCL_MAP) == 0) + { + /* No bits are set for characters < 256. */ + if (list[1] == 0) return TRUE; + /* Might be an empty repeat. */ + continue; + } + set2 = (pcre_uint8 *)(xclass_flags + 1); + break; +#endif + + case OP_NOT_DIGIT: + invert_bits = TRUE; + /* Fall through */ + case OP_DIGIT: + set2 = (pcre_uint8 *)(cd->cbits + cbit_digit); + break; + + case OP_NOT_WHITESPACE: + invert_bits = TRUE; + /* Fall through */ + case OP_WHITESPACE: + set2 = (pcre_uint8 *)(cd->cbits + cbit_space); + break; + + case OP_NOT_WORDCHAR: + invert_bits = TRUE; + /* Fall through */ + case OP_WORDCHAR: + set2 = (pcre_uint8 *)(cd->cbits + cbit_word); + break; + + default: + return FALSE; + } + + /* Because the sets are unaligned, we need + to perform byte comparison here. */ + set_end = set1 + 32; + if (invert_bits) + { + do + { + if ((*set1++ & ~(*set2++)) != 0) return FALSE; + } + while (set1 < set_end); + } + else + { + do + { + if ((*set1++ & *set2++) != 0) return FALSE; + } + while (set1 < set_end); + } + + if (list[1] == 0) return TRUE; + /* Might be an empty repeat. */ + continue; + } + + /* Some property combinations also acceptable. Unicode property opcodes are + processed specially; the rest can be handled with a lookup table. */ + + else + { + pcre_uint32 leftop, rightop; + + leftop = base_list[0]; + rightop = list[0]; + +#ifdef SUPPORT_UCP + accepted = FALSE; /* Always set in non-unicode case. */ + if (leftop == OP_PROP || leftop == OP_NOTPROP) + { + if (rightop == OP_EOD) + accepted = TRUE; + else if (rightop == OP_PROP || rightop == OP_NOTPROP) + { + int n; + const pcre_uint8 *p; + BOOL same = leftop == rightop; + BOOL lisprop = leftop == OP_PROP; + BOOL risprop = rightop == OP_PROP; + BOOL bothprop = lisprop && risprop; + + /* There's a table that specifies how each combination is to be + processed: + 0 Always return FALSE (never auto-possessify) + 1 Character groups are distinct (possessify if both are OP_PROP) + 2 Check character categories in the same group (general or particular) + 3 Return TRUE if the two opcodes are not the same + ... see comments below + */ + + n = propposstab[base_list[2]][list[2]]; + switch(n) + { + case 0: break; + case 1: accepted = bothprop; break; + case 2: accepted = (base_list[3] == list[3]) != same; break; + case 3: accepted = !same; break; + + case 4: /* Left general category, right particular category */ + accepted = risprop && catposstab[base_list[3]][list[3]] == same; + break; + + case 5: /* Right general category, left particular category */ + accepted = lisprop && catposstab[list[3]][base_list[3]] == same; + break; + + /* This code is logically tricky. Think hard before fiddling with it. + The posspropstab table has four entries per row. Each row relates to + one of PCRE's special properties such as ALNUM or SPACE or WORD. + Only WORD actually needs all four entries, but using repeats for the + others means they can all use the same code below. + + The first two entries in each row are Unicode general categories, and + apply always, because all the characters they include are part of the + PCRE character set. The third and fourth entries are a general and a + particular category, respectively, that include one or more relevant + characters. One or the other is used, depending on whether the check + is for a general or a particular category. However, in both cases the + category contains more characters than the specials that are defined + for the property being tested against. Therefore, it cannot be used + in a NOTPROP case. + + Example: the row for WORD contains ucp_L, ucp_N, ucp_P, ucp_Po. + Underscore is covered by ucp_P or ucp_Po. */ + + case 6: /* Left alphanum vs right general category */ + case 7: /* Left space vs right general category */ + case 8: /* Left word vs right general category */ + p = posspropstab[n-6]; + accepted = risprop && lisprop == + (list[3] != p[0] && + list[3] != p[1] && + (list[3] != p[2] || !lisprop)); + break; + + case 9: /* Right alphanum vs left general category */ + case 10: /* Right space vs left general category */ + case 11: /* Right word vs left general category */ + p = posspropstab[n-9]; + accepted = lisprop && risprop == + (base_list[3] != p[0] && + base_list[3] != p[1] && + (base_list[3] != p[2] || !risprop)); + break; + + case 12: /* Left alphanum vs right particular category */ + case 13: /* Left space vs right particular category */ + case 14: /* Left word vs right particular category */ + p = posspropstab[n-12]; + accepted = risprop && lisprop == + (catposstab[p[0]][list[3]] && + catposstab[p[1]][list[3]] && + (list[3] != p[3] || !lisprop)); + break; + + case 15: /* Right alphanum vs left particular category */ + case 16: /* Right space vs left particular category */ + case 17: /* Right word vs left particular category */ + p = posspropstab[n-15]; + accepted = lisprop && risprop == + (catposstab[p[0]][base_list[3]] && + catposstab[p[1]][base_list[3]] && + (base_list[3] != p[3] || !risprop)); + break; + } + } + } + + else +#endif /* SUPPORT_UCP */ + + accepted = leftop >= FIRST_AUTOTAB_OP && leftop <= LAST_AUTOTAB_LEFT_OP && + rightop >= FIRST_AUTOTAB_OP && rightop <= LAST_AUTOTAB_RIGHT_OP && + autoposstab[leftop - FIRST_AUTOTAB_OP][rightop - FIRST_AUTOTAB_OP]; + + if (!accepted) return FALSE; + + if (list[1] == 0) return TRUE; + /* Might be an empty repeat. */ + continue; + } + + /* Control reaches here only if one of the items is a small character list. + All characters are checked against the other side. */ + + do + { + chr = *chr_ptr; + + switch(list_ptr[0]) + { + case OP_CHAR: + ochr_ptr = list_ptr + 2; + do + { + if (chr == *ochr_ptr) return FALSE; + ochr_ptr++; + } + while(*ochr_ptr != NOTACHAR); + break; + + case OP_NOT: + ochr_ptr = list_ptr + 2; + do + { + if (chr == *ochr_ptr) + break; + ochr_ptr++; + } + while(*ochr_ptr != NOTACHAR); + if (*ochr_ptr == NOTACHAR) return FALSE; /* Not found */ + break; + + /* Note that OP_DIGIT etc. are generated only when PCRE_UCP is *not* + set. When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */ + + case OP_DIGIT: + if (chr < 256 && (cd->ctypes[chr] & ctype_digit) != 0) return FALSE; + break; + + case OP_NOT_DIGIT: + if (chr > 255 || (cd->ctypes[chr] & ctype_digit) == 0) return FALSE; + break; + + case OP_WHITESPACE: + if (chr < 256 && (cd->ctypes[chr] & ctype_space) != 0) return FALSE; + break; + + case OP_NOT_WHITESPACE: + if (chr > 255 || (cd->ctypes[chr] & ctype_space) == 0) return FALSE; + break; + + case OP_WORDCHAR: + if (chr < 255 && (cd->ctypes[chr] & ctype_word) != 0) return FALSE; + break; + + case OP_NOT_WORDCHAR: + if (chr > 255 || (cd->ctypes[chr] & ctype_word) == 0) return FALSE; + break; + + case OP_HSPACE: + switch(chr) + { + HSPACE_CASES: return FALSE; + default: break; + } + break; + + case OP_NOT_HSPACE: + switch(chr) + { + HSPACE_CASES: break; + default: return FALSE; + } + break; + + case OP_ANYNL: + case OP_VSPACE: + switch(chr) + { + VSPACE_CASES: return FALSE; + default: break; + } + break; + + case OP_NOT_VSPACE: + switch(chr) + { + VSPACE_CASES: break; + default: return FALSE; + } + break; + + case OP_DOLL: + case OP_EODN: + switch (chr) + { + case CHAR_CR: + case CHAR_LF: + case CHAR_VT: + case CHAR_FF: + case CHAR_NEL: +#ifndef EBCDIC + case 0x2028: + case 0x2029: +#endif /* Not EBCDIC */ + return FALSE; + } + break; + + case OP_EOD: /* Can always possessify before \z */ + break; + +#ifdef SUPPORT_UCP + case OP_PROP: + case OP_NOTPROP: + if (!check_char_prop(chr, list_ptr[2], list_ptr[3], + list_ptr[0] == OP_NOTPROP)) + return FALSE; + break; +#endif + + case OP_NCLASS: + if (chr > 255) return FALSE; + /* Fall through */ + + case OP_CLASS: + if (chr > 255) break; + class_bitset = (pcre_uint8 *) + ((list_ptr == list ? code : base_end) - list_ptr[2]); + if ((class_bitset[chr >> 3] & (1 << (chr & 7))) != 0) return FALSE; + break; + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + case OP_XCLASS: + if (PRIV(xclass)(chr, (list_ptr == list ? code : base_end) - + list_ptr[2] + LINK_SIZE, utf)) return FALSE; + break; +#endif + + default: + return FALSE; + } + + chr_ptr++; + } + while(*chr_ptr != NOTACHAR); + + /* At least one character must be matched from this opcode. */ + + if (list[1] == 0) return TRUE; + } + +/* Control never reaches here. There used to be a fail-save return FALSE; here, +but some compilers complain about an unreachable statement. */ + +} + + + +/************************************************* +* Scan compiled regex for auto-possession * +*************************************************/ + +/* Replaces single character iterations with their possessive alternatives +if appropriate. This function modifies the compiled opcode! + +Arguments: + code points to start of the byte code + utf TRUE in UTF-8 / UTF-16 / UTF-32 mode + cd static compile data + +Returns: nothing +*/ + +static void +auto_possessify(pcre_uchar *code, BOOL utf, const compile_data *cd) +{ +register pcre_uchar c; +const pcre_uchar *end; +pcre_uchar *repeat_opcode; +pcre_uint32 list[8]; +int rec_limit; + +for (;;) + { + c = *code; + + /* When a pattern with bad UTF-8 encoding is compiled with NO_UTF_CHECK, + it may compile without complaining, but may get into a loop here if the code + pointer points to a bad value. This is, of course a documentated possibility, + when NO_UTF_CHECK is set, so it isn't a bug, but we can detect this case and + just give up on this optimization. */ + + if (c >= OP_TABLE_LENGTH) return; + + if (c >= OP_STAR && c <= OP_TYPEPOSUPTO) + { + c -= get_repeat_base(c) - OP_STAR; + end = (c <= OP_MINUPTO) ? + get_chr_property_list(code, utf, cd->fcc, list) : NULL; + list[1] = c == OP_STAR || c == OP_PLUS || c == OP_QUERY || c == OP_UPTO; + + rec_limit = 1000; + if (end != NULL && compare_opcodes(end, utf, cd, list, end, &rec_limit)) + { + switch(c) + { + case OP_STAR: + *code += OP_POSSTAR - OP_STAR; + break; + + case OP_MINSTAR: + *code += OP_POSSTAR - OP_MINSTAR; + break; + + case OP_PLUS: + *code += OP_POSPLUS - OP_PLUS; + break; + + case OP_MINPLUS: + *code += OP_POSPLUS - OP_MINPLUS; + break; + + case OP_QUERY: + *code += OP_POSQUERY - OP_QUERY; + break; + + case OP_MINQUERY: + *code += OP_POSQUERY - OP_MINQUERY; + break; + + case OP_UPTO: + *code += OP_POSUPTO - OP_UPTO; + break; + + case OP_MINUPTO: + *code += OP_POSUPTO - OP_MINUPTO; + break; + } + } + c = *code; + } + else if (c == OP_CLASS || c == OP_NCLASS || c == OP_XCLASS) + { +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + if (c == OP_XCLASS) + repeat_opcode = code + GET(code, 1); + else +#endif + repeat_opcode = code + 1 + (32 / sizeof(pcre_uchar)); + + c = *repeat_opcode; + if (c >= OP_CRSTAR && c <= OP_CRMINRANGE) + { + /* end must not be NULL. */ + end = get_chr_property_list(code, utf, cd->fcc, list); + + list[1] = (c & 1) == 0; + + rec_limit = 1000; + if (compare_opcodes(end, utf, cd, list, end, &rec_limit)) + { + switch (c) + { + case OP_CRSTAR: + case OP_CRMINSTAR: + *repeat_opcode = OP_CRPOSSTAR; + break; + + case OP_CRPLUS: + case OP_CRMINPLUS: + *repeat_opcode = OP_CRPOSPLUS; + break; + + case OP_CRQUERY: + case OP_CRMINQUERY: + *repeat_opcode = OP_CRPOSQUERY; + break; + + case OP_CRRANGE: + case OP_CRMINRANGE: + *repeat_opcode = OP_CRPOSRANGE; + break; + } + } + } + c = *code; + } + + switch(c) + { + case OP_END: + return; + + case OP_TYPESTAR: + case OP_TYPEMINSTAR: + case OP_TYPEPLUS: + case OP_TYPEMINPLUS: + case OP_TYPEQUERY: + case OP_TYPEMINQUERY: + case OP_TYPEPOSSTAR: + case OP_TYPEPOSPLUS: + case OP_TYPEPOSQUERY: + if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2; + break; + + case OP_TYPEUPTO: + case OP_TYPEMINUPTO: + case OP_TYPEEXACT: + case OP_TYPEPOSUPTO: + if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP) + code += 2; + break; + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + case OP_XCLASS: + code += GET(code, 1); + break; +#endif + + case OP_MARK: + case OP_PRUNE_ARG: + case OP_SKIP_ARG: + case OP_THEN_ARG: + code += code[1]; + break; + } + + /* Add in the fixed length from the table */ + + code += PRIV(OP_lengths)[c]; + + /* In UTF-8 mode, opcodes that are followed by a character may be followed by + a multi-byte character. The length in the table is a minimum, so we have to + arrange to skip the extra bytes. */ + +#if defined SUPPORT_UTF && !defined COMPILE_PCRE32 + if (utf) switch(c) + { + case OP_CHAR: + case OP_CHARI: + case OP_NOT: + case OP_NOTI: + case OP_STAR: + case OP_MINSTAR: + case OP_PLUS: + case OP_MINPLUS: + case OP_QUERY: + case OP_MINQUERY: + case OP_UPTO: + case OP_MINUPTO: + case OP_EXACT: + case OP_POSSTAR: + case OP_POSPLUS: + case OP_POSQUERY: + case OP_POSUPTO: + case OP_STARI: + case OP_MINSTARI: + case OP_PLUSI: + case OP_MINPLUSI: + case OP_QUERYI: + case OP_MINQUERYI: + case OP_UPTOI: + case OP_MINUPTOI: + case OP_EXACTI: + case OP_POSSTARI: + case OP_POSPLUSI: + case OP_POSQUERYI: + case OP_POSUPTOI: + case OP_NOTSTAR: + case OP_NOTMINSTAR: + case OP_NOTPLUS: + case OP_NOTMINPLUS: + case OP_NOTQUERY: + case OP_NOTMINQUERY: + case OP_NOTUPTO: + case OP_NOTMINUPTO: + case OP_NOTEXACT: + case OP_NOTPOSSTAR: + case OP_NOTPOSPLUS: + case OP_NOTPOSQUERY: + case OP_NOTPOSUPTO: + case OP_NOTSTARI: + case OP_NOTMINSTARI: + case OP_NOTPLUSI: + case OP_NOTMINPLUSI: + case OP_NOTQUERYI: + case OP_NOTMINQUERYI: + case OP_NOTUPTOI: + case OP_NOTMINUPTOI: + case OP_NOTEXACTI: + case OP_NOTPOSSTARI: + case OP_NOTPOSPLUSI: + case OP_NOTPOSQUERYI: + case OP_NOTPOSUPTOI: + if (HAS_EXTRALEN(code[-1])) code += GET_EXTRALEN(code[-1]); + break; + } +#else + (void)(utf); /* Keep compiler happy by referencing function argument */ +#endif + } +} + + + +/************************************************* +* Check for POSIX class syntax * +*************************************************/ + +/* This function is called when the sequence "[:" or "[." or "[=" is +encountered in a character class. It checks whether this is followed by a +sequence of characters terminated by a matching ":]" or ".]" or "=]". If we +reach an unescaped ']' without the special preceding character, return FALSE. + +Originally, this function only recognized a sequence of letters between the +terminators, but it seems that Perl recognizes any sequence of characters, +though of course unknown POSIX names are subsequently rejected. Perl gives an +"Unknown POSIX class" error for [:f\oo:] for example, where previously PCRE +didn't consider this to be a POSIX class. Likewise for [:1234:]. + +The problem in trying to be exactly like Perl is in the handling of escapes. We +have to be sure that [abc[:x\]pqr] is *not* treated as containing a POSIX +class, but [abc[:x\]pqr:]] is (so that an error can be generated). The code +below handles the special cases \\ and \], but does not try to do any other +escape processing. This makes it different from Perl for cases such as +[:l\ower:] where Perl recognizes it as the POSIX class "lower" but PCRE does +not recognize "l\ower". This is a lesser evil than not diagnosing bad classes +when Perl does, I think. + +A user pointed out that PCRE was rejecting [:a[:digit:]] whereas Perl was not. +It seems that the appearance of a nested POSIX class supersedes an apparent +external class. For example, [:a[:digit:]b:] matches "a", "b", ":", or +a digit. + +In Perl, unescaped square brackets may also appear as part of class names. For +example, [:a[:abc]b:] gives unknown POSIX class "[:abc]b:]". However, for +[:a[:abc]b][b:] it gives unknown POSIX class "[:abc]b][b:]", which does not +seem right at all. PCRE does not allow closing square brackets in POSIX class +names. + +Arguments: + ptr pointer to the initial [ + endptr where to return the end pointer + +Returns: TRUE or FALSE +*/ + +static BOOL +check_posix_syntax(const pcre_uchar *ptr, const pcre_uchar **endptr) +{ +pcre_uchar terminator; /* Don't combine these lines; the Solaris cc */ +terminator = *(++ptr); /* compiler warns about "non-constant" initializer. */ +for (++ptr; *ptr != CHAR_NULL; ptr++) + { + if (*ptr == CHAR_BACKSLASH && + (ptr[1] == CHAR_RIGHT_SQUARE_BRACKET || + ptr[1] == CHAR_BACKSLASH)) + ptr++; + else if ((*ptr == CHAR_LEFT_SQUARE_BRACKET && ptr[1] == terminator) || + *ptr == CHAR_RIGHT_SQUARE_BRACKET) return FALSE; + else if (*ptr == terminator && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET) + { + *endptr = ptr; + return TRUE; + } + } +return FALSE; +} + + + + +/************************************************* +* Check POSIX class name * +*************************************************/ + +/* This function is called to check the name given in a POSIX-style class entry +such as [:alnum:]. + +Arguments: + ptr points to the first letter + len the length of the name + +Returns: a value representing the name, or -1 if unknown +*/ + +static int +check_posix_name(const pcre_uchar *ptr, int len) +{ +const char *pn = posix_names; +register int yield = 0; +while (posix_name_lengths[yield] != 0) + { + if (len == posix_name_lengths[yield] && + STRNCMP_UC_C8(ptr, pn, (unsigned int)len) == 0) return yield; + pn += posix_name_lengths[yield] + 1; + yield++; + } +return -1; +} + + +/************************************************* +* Adjust OP_RECURSE items in repeated group * +*************************************************/ + +/* OP_RECURSE items contain an offset from the start of the regex to the group +that is referenced. This means that groups can be replicated for fixed +repetition simply by copying (because the recursion is allowed to refer to +earlier groups that are outside the current group). However, when a group is +optional (i.e. the minimum quantifier is zero), OP_BRAZERO or OP_SKIPZERO is +inserted before it, after it has been compiled. This means that any OP_RECURSE +items within it that refer to the group itself or any contained groups have to +have their offsets adjusted. That one of the jobs of this function. Before it +is called, the partially compiled regex must be temporarily terminated with +OP_END. + +This function has been extended to cope with forward references for recursions +and subroutine calls. It must check the list of such references for the +group we are dealing with. If it finds that one of the recursions in the +current group is on this list, it does not adjust the value in the reference +(which is a group number). After the group has been scanned, all the offsets in +the forward reference list for the group are adjusted. + +Arguments: + group points to the start of the group + adjust the amount by which the group is to be moved + utf TRUE in UTF-8 / UTF-16 / UTF-32 mode + cd contains pointers to tables etc. + save_hwm_offset the hwm forward reference offset at the start of the group + +Returns: nothing +*/ + +static void +adjust_recurse(pcre_uchar *group, int adjust, BOOL utf, compile_data *cd, + size_t save_hwm_offset) +{ +int offset; +pcre_uchar *hc; +pcre_uchar *ptr = group; + +while ((ptr = (pcre_uchar *)find_recurse(ptr, utf)) != NULL) + { + for (hc = (pcre_uchar *)cd->start_workspace + save_hwm_offset; hc < cd->hwm; + hc += LINK_SIZE) + { + offset = (int)GET(hc, 0); + if (cd->start_code + offset == ptr + 1) break; + } + + /* If we have not found this recursion on the forward reference list, adjust + the recursion's offset if it's after the start of this group. */ + + if (hc >= cd->hwm) + { + offset = (int)GET(ptr, 1); + if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust); + } + + ptr += 1 + LINK_SIZE; + } + +/* Now adjust all forward reference offsets for the group. */ + +for (hc = (pcre_uchar *)cd->start_workspace + save_hwm_offset; hc < cd->hwm; + hc += LINK_SIZE) + { + offset = (int)GET(hc, 0); + PUT(hc, 0, offset + adjust); + } +} + + + +/************************************************* +* Insert an automatic callout point * +*************************************************/ + +/* This function is called when the PCRE_AUTO_CALLOUT option is set, to insert +callout points before each pattern item. + +Arguments: + code current code pointer + ptr current pattern pointer + cd pointers to tables etc + +Returns: new code pointer +*/ + +static pcre_uchar * +auto_callout(pcre_uchar *code, const pcre_uchar *ptr, compile_data *cd) +{ +*code++ = OP_CALLOUT; +*code++ = 255; +PUT(code, 0, (int)(ptr - cd->start_pattern)); /* Pattern offset */ +PUT(code, LINK_SIZE, 0); /* Default length */ +return code + 2 * LINK_SIZE; +} + + + +/************************************************* +* Complete a callout item * +*************************************************/ + +/* A callout item contains the length of the next item in the pattern, which +we can't fill in till after we have reached the relevant point. This is used +for both automatic and manual callouts. + +Arguments: + previous_callout points to previous callout item + ptr current pattern pointer + cd pointers to tables etc + +Returns: nothing +*/ + +static void +complete_callout(pcre_uchar *previous_callout, const pcre_uchar *ptr, compile_data *cd) +{ +int length = (int)(ptr - cd->start_pattern - GET(previous_callout, 2)); +PUT(previous_callout, 2 + LINK_SIZE, length); +} + + + +#ifdef SUPPORT_UCP +/************************************************* +* Get othercase range * +*************************************************/ + +/* This function is passed the start and end of a class range, in UTF-8 mode +with UCP support. It searches up the characters, looking for ranges of +characters in the "other" case. Each call returns the next one, updating the +start address. A character with multiple other cases is returned on its own +with a special return value. + +Arguments: + cptr points to starting character value; updated + d end value + ocptr where to put start of othercase range + odptr where to put end of othercase range + +Yield: -1 when no more + 0 when a range is returned + >0 the CASESET offset for char with multiple other cases + in this case, ocptr contains the original +*/ + +static int +get_othercase_range(pcre_uint32 *cptr, pcre_uint32 d, pcre_uint32 *ocptr, + pcre_uint32 *odptr) +{ +pcre_uint32 c, othercase, next; +unsigned int co; + +/* Find the first character that has an other case. If it has multiple other +cases, return its case offset value. */ + +for (c = *cptr; c <= d; c++) + { + if ((co = UCD_CASESET(c)) != 0) + { + *ocptr = c++; /* Character that has the set */ + *cptr = c; /* Rest of input range */ + return (int)co; + } + if ((othercase = UCD_OTHERCASE(c)) != c) break; + } + +if (c > d) return -1; /* Reached end of range */ + +/* Found a character that has a single other case. Search for the end of the +range, which is either the end of the input range, or a character that has zero +or more than one other cases. */ + +*ocptr = othercase; +next = othercase + 1; + +for (++c; c <= d; c++) + { + if ((co = UCD_CASESET(c)) != 0 || UCD_OTHERCASE(c) != next) break; + next++; + } + +*odptr = next - 1; /* End of othercase range */ +*cptr = c; /* Rest of input range */ +return 0; +} +#endif /* SUPPORT_UCP */ + + + +/************************************************* +* Add a character or range to a class * +*************************************************/ + +/* This function packages up the logic of adding a character or range of +characters to a class. The character values in the arguments will be within the +valid values for the current mode (8-bit, 16-bit, UTF, etc). This function is +mutually recursive with the function immediately below. + +Arguments: + classbits the bit map for characters < 256 + uchardptr points to the pointer for extra data + options the options word + cd contains pointers to tables etc. + start start of range character + end end of range character + +Returns: the number of < 256 characters added + the pointer to extra data is updated +*/ + +static int +add_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr, int options, + compile_data *cd, pcre_uint32 start, pcre_uint32 end) +{ +pcre_uint32 c; +pcre_uint32 classbits_end = (end <= 0xff ? end : 0xff); +int n8 = 0; + +/* If caseless matching is required, scan the range and process alternate +cases. In Unicode, there are 8-bit characters that have alternate cases that +are greater than 255 and vice-versa. Sometimes we can just extend the original +range. */ + +if ((options & PCRE_CASELESS) != 0) + { +#ifdef SUPPORT_UCP + if ((options & PCRE_UTF8) != 0) + { + int rc; + pcre_uint32 oc, od; + + options &= ~PCRE_CASELESS; /* Remove for recursive calls */ + c = start; + + while ((rc = get_othercase_range(&c, end, &oc, &od)) >= 0) + { + /* Handle a single character that has more than one other case. */ + + if (rc > 0) n8 += add_list_to_class(classbits, uchardptr, options, cd, + PRIV(ucd_caseless_sets) + rc, oc); + + /* Do nothing if the other case range is within the original range. */ + + else if (oc >= start && od <= end) continue; + + /* Extend the original range if there is overlap, noting that if oc < c, we + can't have od > end because a subrange is always shorter than the basic + range. Otherwise, use a recursive call to add the additional range. */ + + else if (oc < start && od >= start - 1) start = oc; /* Extend downwards */ + else if (od > end && oc <= end + 1) + { + end = od; /* Extend upwards */ + if (end > classbits_end) classbits_end = (end <= 0xff ? end : 0xff); + } + else n8 += add_to_class(classbits, uchardptr, options, cd, oc, od); + } + } + else +#endif /* SUPPORT_UCP */ + + /* Not UTF-mode, or no UCP */ + + for (c = start; c <= classbits_end; c++) + { + SETBIT(classbits, cd->fcc[c]); + n8++; + } + } + +/* Now handle the original range. Adjust the final value according to the bit +length - this means that the same lists of (e.g.) horizontal spaces can be used +in all cases. */ + +#if defined COMPILE_PCRE8 +#ifdef SUPPORT_UTF + if ((options & PCRE_UTF8) == 0) +#endif + if (end > 0xff) end = 0xff; + +#elif defined COMPILE_PCRE16 +#ifdef SUPPORT_UTF + if ((options & PCRE_UTF16) == 0) +#endif + if (end > 0xffff) end = 0xffff; + +#endif /* COMPILE_PCRE[8|16] */ + +/* Use the bitmap for characters < 256. Otherwise use extra data.*/ + +for (c = start; c <= classbits_end; c++) + { + /* Regardless of start, c will always be <= 255. */ + SETBIT(classbits, c); + n8++; + } + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 +if (start <= 0xff) start = 0xff + 1; + +if (end >= start) + { + pcre_uchar *uchardata = *uchardptr; +#ifdef SUPPORT_UTF + if ((options & PCRE_UTF8) != 0) /* All UTFs use the same flag bit */ + { + if (start < end) + { + *uchardata++ = XCL_RANGE; + uchardata += PRIV(ord2utf)(start, uchardata); + uchardata += PRIV(ord2utf)(end, uchardata); + } + else if (start == end) + { + *uchardata++ = XCL_SINGLE; + uchardata += PRIV(ord2utf)(start, uchardata); + } + } + else +#endif /* SUPPORT_UTF */ + + /* Without UTF support, character values are constrained by the bit length, + and can only be > 256 for 16-bit and 32-bit libraries. */ + +#ifdef COMPILE_PCRE8 + {} +#else + if (start < end) + { + *uchardata++ = XCL_RANGE; + *uchardata++ = start; + *uchardata++ = end; + } + else if (start == end) + { + *uchardata++ = XCL_SINGLE; + *uchardata++ = start; + } +#endif + + *uchardptr = uchardata; /* Updata extra data pointer */ + } +#endif /* SUPPORT_UTF || !COMPILE_PCRE8 */ + +return n8; /* Number of 8-bit characters */ +} + + + + +/************************************************* +* Add a list of characters to a class * +*************************************************/ + +/* This function is used for adding a list of case-equivalent characters to a +class, and also for adding a list of horizontal or vertical whitespace. If the +list is in order (which it should be), ranges of characters are detected and +handled appropriately. This function is mutually recursive with the function +above. + +Arguments: + classbits the bit map for characters < 256 + uchardptr points to the pointer for extra data + options the options word + cd contains pointers to tables etc. + p points to row of 32-bit values, terminated by NOTACHAR + except character to omit; this is used when adding lists of + case-equivalent characters to avoid including the one we + already know about + +Returns: the number of < 256 characters added + the pointer to extra data is updated +*/ + +static int +add_list_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr, int options, + compile_data *cd, const pcre_uint32 *p, unsigned int except) +{ +int n8 = 0; +while (p[0] < NOTACHAR) + { + int n = 0; + if (p[0] != except) + { + while(p[n+1] == p[0] + n + 1) n++; + n8 += add_to_class(classbits, uchardptr, options, cd, p[0], p[n]); + } + p += n + 1; + } +return n8; +} + + + +/************************************************* +* Add characters not in a list to a class * +*************************************************/ + +/* This function is used for adding the complement of a list of horizontal or +vertical whitespace to a class. The list must be in order. + +Arguments: + classbits the bit map for characters < 256 + uchardptr points to the pointer for extra data + options the options word + cd contains pointers to tables etc. + p points to row of 32-bit values, terminated by NOTACHAR + +Returns: the number of < 256 characters added + the pointer to extra data is updated +*/ + +static int +add_not_list_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr, + int options, compile_data *cd, const pcre_uint32 *p) +{ +BOOL utf = (options & PCRE_UTF8) != 0; +int n8 = 0; +if (p[0] > 0) + n8 += add_to_class(classbits, uchardptr, options, cd, 0, p[0] - 1); +while (p[0] < NOTACHAR) + { + while (p[1] == p[0] + 1) p++; + n8 += add_to_class(classbits, uchardptr, options, cd, p[0] + 1, + (p[1] == NOTACHAR) ? (utf ? 0x10ffffu : 0xffffffffu) : p[1] - 1); + p++; + } +return n8; +} + + + +/************************************************* +* Compile one branch * +*************************************************/ + +/* Scan the pattern, compiling it into the a vector. If the options are +changed during the branch, the pointer is used to change the external options +bits. This function is used during the pre-compile phase when we are trying +to find out the amount of memory needed, as well as during the real compile +phase. The value of lengthptr distinguishes the two phases. + +Arguments: + optionsptr pointer to the option bits + codeptr points to the pointer to the current code point + ptrptr points to the current pattern pointer + errorcodeptr points to error code variable + firstcharptr place to put the first required character + firstcharflagsptr place to put the first character flags, or a negative number + reqcharptr place to put the last required character + reqcharflagsptr place to put the last required character flags, or a negative number + bcptr points to current branch chain + cond_depth conditional nesting depth + cd contains pointers to tables etc. + lengthptr NULL during the real compile phase + points to length accumulator during pre-compile phase + +Returns: TRUE on success + FALSE, with *errorcodeptr set non-zero on error +*/ + +static BOOL +compile_branch(int *optionsptr, pcre_uchar **codeptr, + const pcre_uchar **ptrptr, int *errorcodeptr, + pcre_uint32 *firstcharptr, pcre_int32 *firstcharflagsptr, + pcre_uint32 *reqcharptr, pcre_int32 *reqcharflagsptr, + branch_chain *bcptr, int cond_depth, + compile_data *cd, int *lengthptr) +{ +int repeat_type, op_type; +int repeat_min = 0, repeat_max = 0; /* To please picky compilers */ +int bravalue = 0; +int greedy_default, greedy_non_default; +pcre_uint32 firstchar, reqchar; +pcre_int32 firstcharflags, reqcharflags; +pcre_uint32 zeroreqchar, zerofirstchar; +pcre_int32 zeroreqcharflags, zerofirstcharflags; +pcre_int32 req_caseopt, reqvary, tempreqvary; +int options = *optionsptr; /* May change dynamically */ +int after_manual_callout = 0; +int length_prevgroup = 0; +register pcre_uint32 c; +int escape; +register pcre_uchar *code = *codeptr; +pcre_uchar *last_code = code; +pcre_uchar *orig_code = code; +pcre_uchar *tempcode; +BOOL inescq = FALSE; +BOOL groupsetfirstchar = FALSE; +const pcre_uchar *ptr = *ptrptr; +const pcre_uchar *tempptr; +const pcre_uchar *nestptr = NULL; +pcre_uchar *previous = NULL; +pcre_uchar *previous_callout = NULL; +size_t item_hwm_offset = 0; +pcre_uint8 classbits[32]; + +/* We can fish out the UTF-8 setting once and for all into a BOOL, but we +must not do this for other options (e.g. PCRE_EXTENDED) because they may change +dynamically as we process the pattern. */ + +#ifdef SUPPORT_UTF +/* PCRE_UTF[16|32] have the same value as PCRE_UTF8. */ +BOOL utf = (options & PCRE_UTF8) != 0; +#ifndef COMPILE_PCRE32 +pcre_uchar utf_chars[6]; +#endif +#else +BOOL utf = FALSE; +#endif + +/* Helper variables for OP_XCLASS opcode (for characters > 255). We define +class_uchardata always so that it can be passed to add_to_class() always, +though it will not be used in non-UTF 8-bit cases. This avoids having to supply +alternative calls for the different cases. */ + +pcre_uchar *class_uchardata; +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 +BOOL xclass; +pcre_uchar *class_uchardata_base; +#endif + +#ifdef PCRE_DEBUG +if (lengthptr != NULL) DPRINTF((">> start branch\n")); +#endif + +/* Set up the default and non-default settings for greediness */ + +greedy_default = ((options & PCRE_UNGREEDY) != 0); +greedy_non_default = greedy_default ^ 1; + +/* Initialize no first byte, no required byte. REQ_UNSET means "no char +matching encountered yet". It gets changed to REQ_NONE if we hit something that +matches a non-fixed char first char; reqchar just remains unset if we never +find one. + +When we hit a repeat whose minimum is zero, we may have to adjust these values +to take the zero repeat into account. This is implemented by setting them to +zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual +item types that can be repeated set these backoff variables appropriately. */ + +firstchar = reqchar = zerofirstchar = zeroreqchar = 0; +firstcharflags = reqcharflags = zerofirstcharflags = zeroreqcharflags = REQ_UNSET; + +/* The variable req_caseopt contains either the REQ_CASELESS value +or zero, according to the current setting of the caseless flag. The +REQ_CASELESS leaves the lower 28 bit empty. It is added into the +firstchar or reqchar variables to record the case status of the +value. This is used only for ASCII characters. */ + +req_caseopt = ((options & PCRE_CASELESS) != 0)? REQ_CASELESS:0; + +/* Switch on next character until the end of the branch */ + +for (;; ptr++) + { + BOOL negate_class; + BOOL should_flip_negation; + BOOL possessive_quantifier; + BOOL is_quantifier; + BOOL is_recurse; + BOOL reset_bracount; + int class_has_8bitchar; + int class_one_char; +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + BOOL xclass_has_prop; +#endif + int newoptions; + int recno; + int refsign; + int skipbytes; + pcre_uint32 subreqchar, subfirstchar; + pcre_int32 subreqcharflags, subfirstcharflags; + int terminator; + unsigned int mclength; + unsigned int tempbracount; + pcre_uint32 ec; + pcre_uchar mcbuffer[8]; + + /* Get next character in the pattern */ + + c = *ptr; + + /* If we are at the end of a nested substitution, revert to the outer level + string. Nesting only happens one level deep. */ + + if (c == CHAR_NULL && nestptr != NULL) + { + ptr = nestptr; + nestptr = NULL; + c = *ptr; + } + + /* If we are in the pre-compile phase, accumulate the length used for the + previous cycle of this loop. */ + + if (lengthptr != NULL) + { +#ifdef PCRE_DEBUG + if (code > cd->hwm) cd->hwm = code; /* High water info */ +#endif + if (code > cd->start_workspace + cd->workspace_size - + WORK_SIZE_SAFETY_MARGIN) /* Check for overrun */ + { + *errorcodeptr = ERR52; + goto FAILED; + } + + /* There is at least one situation where code goes backwards: this is the + case of a zero quantifier after a class (e.g. [ab]{0}). At compile time, + the class is simply eliminated. However, it is created first, so we have to + allow memory for it. Therefore, don't ever reduce the length at this point. + */ + + if (code < last_code) code = last_code; + + /* Paranoid check for integer overflow */ + + if (OFLOW_MAX - *lengthptr < code - last_code) + { + *errorcodeptr = ERR20; + goto FAILED; + } + + *lengthptr += (int)(code - last_code); + DPRINTF(("length=%d added %d c=%c (0x%x)\n", *lengthptr, + (int)(code - last_code), c, c)); + + /* If "previous" is set and it is not at the start of the work space, move + it back to there, in order to avoid filling up the work space. Otherwise, + if "previous" is NULL, reset the current code pointer to the start. */ + + if (previous != NULL) + { + if (previous > orig_code) + { + memmove(orig_code, previous, IN_UCHARS(code - previous)); + code -= previous - orig_code; + previous = orig_code; + } + } + else code = orig_code; + + /* Remember where this code item starts so we can pick up the length + next time round. */ + + last_code = code; + } + + /* In the real compile phase, just check the workspace used by the forward + reference list. */ + + else if (cd->hwm > cd->start_workspace + cd->workspace_size) + { + *errorcodeptr = ERR52; + goto FAILED; + } + + /* If in \Q...\E, check for the end; if not, we have a literal */ + + if (inescq && c != CHAR_NULL) + { + if (c == CHAR_BACKSLASH && ptr[1] == CHAR_E) + { + inescq = FALSE; + ptr++; + continue; + } + else + { + if (previous_callout != NULL) + { + if (lengthptr == NULL) /* Don't attempt in pre-compile phase */ + complete_callout(previous_callout, ptr, cd); + previous_callout = NULL; + } + if ((options & PCRE_AUTO_CALLOUT) != 0) + { + previous_callout = code; + code = auto_callout(code, ptr, cd); + } + goto NORMAL_CHAR; + } + /* Control does not reach here. */ + } + + /* In extended mode, skip white space and comments. We need a loop in order + to check for more white space and more comments after a comment. */ + + if ((options & PCRE_EXTENDED) != 0) + { + for (;;) + { + while (MAX_255(c) && (cd->ctypes[c] & ctype_space) != 0) c = *(++ptr); + if (c != CHAR_NUMBER_SIGN) break; + ptr++; + while (*ptr != CHAR_NULL) + { + if (IS_NEWLINE(ptr)) /* For non-fixed-length newline cases, */ + { /* IS_NEWLINE sets cd->nllen. */ + ptr += cd->nllen; + break; + } + ptr++; +#ifdef SUPPORT_UTF + if (utf) FORWARDCHAR(ptr); +#endif + } + c = *ptr; /* Either NULL or the char after a newline */ + } + } + + /* See if the next thing is a quantifier. */ + + is_quantifier = + c == CHAR_ASTERISK || c == CHAR_PLUS || c == CHAR_QUESTION_MARK || + (c == CHAR_LEFT_CURLY_BRACKET && is_counted_repeat(ptr+1)); + + /* Fill in length of a previous callout, except when the next thing is a + quantifier or when processing a property substitution string in UCP mode. */ + + if (!is_quantifier && previous_callout != NULL && nestptr == NULL && + after_manual_callout-- <= 0) + { + if (lengthptr == NULL) /* Don't attempt in pre-compile phase */ + complete_callout(previous_callout, ptr, cd); + previous_callout = NULL; + } + + /* Create auto callout, except for quantifiers, or while processing property + strings that are substituted for \w etc in UCP mode. */ + + if ((options & PCRE_AUTO_CALLOUT) != 0 && !is_quantifier && nestptr == NULL) + { + previous_callout = code; + code = auto_callout(code, ptr, cd); + } + + /* Process the next pattern item. */ + + switch(c) + { + /* ===================================================================*/ + case CHAR_NULL: /* The branch terminates at string end */ + case CHAR_VERTICAL_LINE: /* or | or ) */ + case CHAR_RIGHT_PARENTHESIS: + *firstcharptr = firstchar; + *firstcharflagsptr = firstcharflags; + *reqcharptr = reqchar; + *reqcharflagsptr = reqcharflags; + *codeptr = code; + *ptrptr = ptr; + if (lengthptr != NULL) + { + if (OFLOW_MAX - *lengthptr < code - last_code) + { + *errorcodeptr = ERR20; + goto FAILED; + } + *lengthptr += (int)(code - last_code); /* To include callout length */ + DPRINTF((">> end branch\n")); + } + return TRUE; + + + /* ===================================================================*/ + /* Handle single-character metacharacters. In multiline mode, ^ disables + the setting of any following char as a first character. */ + + case CHAR_CIRCUMFLEX_ACCENT: + previous = NULL; + if ((options & PCRE_MULTILINE) != 0) + { + if (firstcharflags == REQ_UNSET) + zerofirstcharflags = firstcharflags = REQ_NONE; + *code++ = OP_CIRCM; + } + else *code++ = OP_CIRC; + break; + + case CHAR_DOLLAR_SIGN: + previous = NULL; + *code++ = ((options & PCRE_MULTILINE) != 0)? OP_DOLLM : OP_DOLL; + break; + + /* There can never be a first char if '.' is first, whatever happens about + repeats. The value of reqchar doesn't change either. */ + + case CHAR_DOT: + if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE; + zerofirstchar = firstchar; + zerofirstcharflags = firstcharflags; + zeroreqchar = reqchar; + zeroreqcharflags = reqcharflags; + previous = code; + item_hwm_offset = cd->hwm - cd->start_workspace; + *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY; + break; + + + /* ===================================================================*/ + /* Character classes. If the included characters are all < 256, we build a + 32-byte bitmap of the permitted characters, except in the special case + where there is only one such character. For negated classes, we build the + map as usual, then invert it at the end. However, we use a different opcode + so that data characters > 255 can be handled correctly. + + If the class contains characters outside the 0-255 range, a different + opcode is compiled. It may optionally have a bit map for characters < 256, + but those above are are explicitly listed afterwards. A flag byte tells + whether the bitmap is present, and whether this is a negated class or not. + + In JavaScript compatibility mode, an isolated ']' causes an error. In + default (Perl) mode, it is treated as a data character. */ + + case CHAR_RIGHT_SQUARE_BRACKET: + if ((cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0) + { + *errorcodeptr = ERR64; + goto FAILED; + } + goto NORMAL_CHAR; + + /* In another (POSIX) regex library, the ugly syntax [[:<:]] and [[:>:]] is + used for "start of word" and "end of word". As these are otherwise illegal + sequences, we don't break anything by recognizing them. They are replaced + by \b(?=\w) and \b(?<=\w) respectively. Sequences like [a[:<:]] are + erroneous and are handled by the normal code below. */ + + case CHAR_LEFT_SQUARE_BRACKET: + if (STRNCMP_UC_C8(ptr+1, STRING_WEIRD_STARTWORD, 6) == 0) + { + nestptr = ptr + 7; + ptr = sub_start_of_word - 1; + continue; + } + + if (STRNCMP_UC_C8(ptr+1, STRING_WEIRD_ENDWORD, 6) == 0) + { + nestptr = ptr + 7; + ptr = sub_end_of_word - 1; + continue; + } + + /* Handle a real character class. */ + + previous = code; + item_hwm_offset = cd->hwm - cd->start_workspace; + + /* PCRE supports POSIX class stuff inside a class. Perl gives an error if + they are encountered at the top level, so we'll do that too. */ + + if ((ptr[1] == CHAR_COLON || ptr[1] == CHAR_DOT || + ptr[1] == CHAR_EQUALS_SIGN) && + check_posix_syntax(ptr, &tempptr)) + { + *errorcodeptr = (ptr[1] == CHAR_COLON)? ERR13 : ERR31; + goto FAILED; + } + + /* If the first character is '^', set the negation flag and skip it. Also, + if the first few characters (either before or after ^) are \Q\E or \E we + skip them too. This makes for compatibility with Perl. */ + + negate_class = FALSE; + for (;;) + { + c = *(++ptr); + if (c == CHAR_BACKSLASH) + { + if (ptr[1] == CHAR_E) + ptr++; + else if (STRNCMP_UC_C8(ptr + 1, STR_Q STR_BACKSLASH STR_E, 3) == 0) + ptr += 3; + else + break; + } + else if (!negate_class && c == CHAR_CIRCUMFLEX_ACCENT) + negate_class = TRUE; + else break; + } + + /* Empty classes are allowed in JavaScript compatibility mode. Otherwise, + an initial ']' is taken as a data character -- the code below handles + that. In JS mode, [] must always fail, so generate OP_FAIL, whereas + [^] must match any character, so generate OP_ALLANY. */ + + if (c == CHAR_RIGHT_SQUARE_BRACKET && + (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0) + { + *code++ = negate_class? OP_ALLANY : OP_FAIL; + if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE; + zerofirstchar = firstchar; + zerofirstcharflags = firstcharflags; + break; + } + + /* If a class contains a negative special such as \S, we need to flip the + negation flag at the end, so that support for characters > 255 works + correctly (they are all included in the class). */ + + should_flip_negation = FALSE; + + /* Extended class (xclass) will be used when characters > 255 + might match. */ + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + xclass = FALSE; + class_uchardata = code + LINK_SIZE + 2; /* For XCLASS items */ + class_uchardata_base = class_uchardata; /* Save the start */ +#endif + + /* For optimization purposes, we track some properties of the class: + class_has_8bitchar will be non-zero if the class contains at least one < + 256 character; class_one_char will be 1 if the class contains just one + character; xclass_has_prop will be TRUE if unicode property checks + are present in the class. */ + + class_has_8bitchar = 0; + class_one_char = 0; +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + xclass_has_prop = FALSE; +#endif + + /* Initialize the 32-char bit map to all zeros. We build the map in a + temporary bit of memory, in case the class contains fewer than two + 8-bit characters because in that case the compiled code doesn't use the bit + map. */ + + memset(classbits, 0, 32 * sizeof(pcre_uint8)); + + /* Process characters until ] is reached. By writing this as a "do" it + means that an initial ] is taken as a data character. At the start of the + loop, c contains the first byte of the character. */ + + if (c != CHAR_NULL) do + { + const pcre_uchar *oldptr; + +#ifdef SUPPORT_UTF + if (utf && HAS_EXTRALEN(c)) + { /* Braces are required because the */ + GETCHARLEN(c, ptr, ptr); /* macro generates multiple statements */ + } +#endif + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + /* In the pre-compile phase, accumulate the length of any extra + data and reset the pointer. This is so that very large classes that + contain a zillion > 255 characters no longer overwrite the work space + (which is on the stack). We have to remember that there was XCLASS data, + however. */ + + if (class_uchardata > class_uchardata_base) xclass = TRUE; + + if (lengthptr != NULL && class_uchardata > class_uchardata_base) + { + *lengthptr += (int)(class_uchardata - class_uchardata_base); + class_uchardata = class_uchardata_base; + } +#endif + + /* Inside \Q...\E everything is literal except \E */ + + if (inescq) + { + if (c == CHAR_BACKSLASH && ptr[1] == CHAR_E) /* If we are at \E */ + { + inescq = FALSE; /* Reset literal state */ + ptr++; /* Skip the 'E' */ + continue; /* Carry on with next */ + } + goto CHECK_RANGE; /* Could be range if \E follows */ + } + + /* Handle POSIX class names. Perl allows a negation extension of the + form [:^name:]. A square bracket that doesn't match the syntax is + treated as a literal. We also recognize the POSIX constructions + [.ch.] and [=ch=] ("collating elements") and fault them, as Perl + 5.6 and 5.8 do. */ + + if (c == CHAR_LEFT_SQUARE_BRACKET && + (ptr[1] == CHAR_COLON || ptr[1] == CHAR_DOT || + ptr[1] == CHAR_EQUALS_SIGN) && check_posix_syntax(ptr, &tempptr)) + { + BOOL local_negate = FALSE; + int posix_class, taboffset, tabopt; + register const pcre_uint8 *cbits = cd->cbits; + pcre_uint8 pbits[32]; + + if (ptr[1] != CHAR_COLON) + { + *errorcodeptr = ERR31; + goto FAILED; + } + + ptr += 2; + if (*ptr == CHAR_CIRCUMFLEX_ACCENT) + { + local_negate = TRUE; + should_flip_negation = TRUE; /* Note negative special */ + ptr++; + } + + posix_class = check_posix_name(ptr, (int)(tempptr - ptr)); + if (posix_class < 0) + { + *errorcodeptr = ERR30; + goto FAILED; + } + + /* If matching is caseless, upper and lower are converted to + alpha. This relies on the fact that the class table starts with + alpha, lower, upper as the first 3 entries. */ + + if ((options & PCRE_CASELESS) != 0 && posix_class <= 2) + posix_class = 0; + + /* When PCRE_UCP is set, some of the POSIX classes are converted to + different escape sequences that use Unicode properties \p or \P. Others + that are not available via \p or \P generate XCL_PROP/XCL_NOTPROP + directly. */ + +#ifdef SUPPORT_UCP + if ((options & PCRE_UCP) != 0) + { + unsigned int ptype = 0; + int pc = posix_class + ((local_negate)? POSIX_SUBSIZE/2 : 0); + + /* The posix_substitutes table specifies which POSIX classes can be + converted to \p or \P items. */ + + if (posix_substitutes[pc] != NULL) + { + nestptr = tempptr + 1; + ptr = posix_substitutes[pc] - 1; + continue; + } + + /* There are three other classes that generate special property calls + that are recognized only in an XCLASS. */ + + else switch(posix_class) + { + case PC_GRAPH: + ptype = PT_PXGRAPH; + /* Fall through */ + case PC_PRINT: + if (ptype == 0) ptype = PT_PXPRINT; + /* Fall through */ + case PC_PUNCT: + if (ptype == 0) ptype = PT_PXPUNCT; + *class_uchardata++ = local_negate? XCL_NOTPROP : XCL_PROP; + *class_uchardata++ = ptype; + *class_uchardata++ = 0; + xclass_has_prop = TRUE; + ptr = tempptr + 1; + continue; + + /* For the other POSIX classes (ascii, xdigit) we are going to fall + through to the non-UCP case and build a bit map for characters with + code points less than 256. If we are in a negated POSIX class + within a non-negated overall class, characters with code points + greater than 255 must all match. In the special case where we have + not yet generated any xclass data, and this is the final item in + the overall class, we need do nothing: later on, the opcode + OP_NCLASS will be used to indicate that characters greater than 255 + are acceptable. If we have already seen an xclass item or one may + follow (we have to assume that it might if this is not the end of + the class), explicitly match all wide codepoints. */ + + default: + if (!negate_class && local_negate && + (xclass || tempptr[2] != CHAR_RIGHT_SQUARE_BRACKET)) + { + *class_uchardata++ = XCL_RANGE; + class_uchardata += PRIV(ord2utf)(0x100, class_uchardata); + class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata); + } + break; + } + } +#endif + /* In the non-UCP case, or when UCP makes no difference, we build the + bit map for the POSIX class in a chunk of local store because we may be + adding and subtracting from it, and we don't want to subtract bits that + may be in the main map already. At the end we or the result into the + bit map that is being built. */ + + posix_class *= 3; + + /* Copy in the first table (always present) */ + + memcpy(pbits, cbits + posix_class_maps[posix_class], + 32 * sizeof(pcre_uint8)); + + /* If there is a second table, add or remove it as required. */ + + taboffset = posix_class_maps[posix_class + 1]; + tabopt = posix_class_maps[posix_class + 2]; + + if (taboffset >= 0) + { + if (tabopt >= 0) + for (c = 0; c < 32; c++) pbits[c] |= cbits[c + taboffset]; + else + for (c = 0; c < 32; c++) pbits[c] &= ~cbits[c + taboffset]; + } + + /* Now see if we need to remove any special characters. An option + value of 1 removes vertical space and 2 removes underscore. */ + + if (tabopt < 0) tabopt = -tabopt; + if (tabopt == 1) pbits[1] &= ~0x3c; + else if (tabopt == 2) pbits[11] &= 0x7f; + + /* Add the POSIX table or its complement into the main table that is + being built and we are done. */ + + if (local_negate) + for (c = 0; c < 32; c++) classbits[c] |= ~pbits[c]; + else + for (c = 0; c < 32; c++) classbits[c] |= pbits[c]; + + ptr = tempptr + 1; + /* Every class contains at least one < 256 character. */ + class_has_8bitchar = 1; + /* Every class contains at least two characters. */ + class_one_char = 2; + continue; /* End of POSIX syntax handling */ + } + + /* Backslash may introduce a single character, or it may introduce one + of the specials, which just set a flag. The sequence \b is a special + case. Inside a class (and only there) it is treated as backspace. We + assume that other escapes have more than one character in them, so + speculatively set both class_has_8bitchar and class_one_char bigger + than one. Unrecognized escapes fall through and are either treated + as literal characters (by default), or are faulted if + PCRE_EXTRA is set. */ + + if (c == CHAR_BACKSLASH) + { + escape = check_escape(&ptr, &ec, errorcodeptr, cd->bracount, options, + TRUE); + if (*errorcodeptr != 0) goto FAILED; + if (escape == 0) c = ec; + else if (escape == ESC_b) c = CHAR_BS; /* \b is backspace in a class */ + else if (escape == ESC_N) /* \N is not supported in a class */ + { + *errorcodeptr = ERR71; + goto FAILED; + } + else if (escape == ESC_Q) /* Handle start of quoted string */ + { + if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E) + { + ptr += 2; /* avoid empty string */ + } + else inescq = TRUE; + continue; + } + else if (escape == ESC_E) continue; /* Ignore orphan \E */ + + else + { + register const pcre_uint8 *cbits = cd->cbits; + /* Every class contains at least two < 256 characters. */ + class_has_8bitchar++; + /* Every class contains at least two characters. */ + class_one_char += 2; + + switch (escape) + { +#ifdef SUPPORT_UCP + case ESC_du: /* These are the values given for \d etc */ + case ESC_DU: /* when PCRE_UCP is set. We replace the */ + case ESC_wu: /* escape sequence with an appropriate \p */ + case ESC_WU: /* or \P to test Unicode properties instead */ + case ESC_su: /* of the default ASCII testing. */ + case ESC_SU: + nestptr = ptr; + ptr = substitutes[escape - ESC_DU] - 1; /* Just before substitute */ + class_has_8bitchar--; /* Undo! */ + continue; +#endif + case ESC_d: + for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_digit]; + continue; + + case ESC_D: + should_flip_negation = TRUE; + for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_digit]; + continue; + + case ESC_w: + for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_word]; + continue; + + case ESC_W: + should_flip_negation = TRUE; + for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word]; + continue; + + /* Perl 5.004 onwards omitted VT from \s, but restored it at Perl + 5.18. Before PCRE 8.34, we had to preserve the VT bit if it was + previously set by something earlier in the character class. + Luckily, the value of CHAR_VT is 0x0b in both ASCII and EBCDIC, so + we could just adjust the appropriate bit. From PCRE 8.34 we no + longer treat \s and \S specially. */ + + case ESC_s: + for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_space]; + continue; + + case ESC_S: + should_flip_negation = TRUE; + for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space]; + continue; + + /* The rest apply in both UCP and non-UCP cases. */ + + case ESC_h: + (void)add_list_to_class(classbits, &class_uchardata, options, cd, + PRIV(hspace_list), NOTACHAR); + continue; + + case ESC_H: + (void)add_not_list_to_class(classbits, &class_uchardata, options, + cd, PRIV(hspace_list)); + continue; + + case ESC_v: + (void)add_list_to_class(classbits, &class_uchardata, options, cd, + PRIV(vspace_list), NOTACHAR); + continue; + + case ESC_V: + (void)add_not_list_to_class(classbits, &class_uchardata, options, + cd, PRIV(vspace_list)); + continue; + + case ESC_p: + case ESC_P: +#ifdef SUPPORT_UCP + { + BOOL negated; + unsigned int ptype = 0, pdata = 0; + if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr)) + goto FAILED; + *class_uchardata++ = ((escape == ESC_p) != negated)? + XCL_PROP : XCL_NOTPROP; + *class_uchardata++ = ptype; + *class_uchardata++ = pdata; + xclass_has_prop = TRUE; + class_has_8bitchar--; /* Undo! */ + continue; + } +#else + *errorcodeptr = ERR45; + goto FAILED; +#endif + /* Unrecognized escapes are faulted if PCRE is running in its + strict mode. By default, for compatibility with Perl, they are + treated as literals. */ + + default: + if ((options & PCRE_EXTRA) != 0) + { + *errorcodeptr = ERR7; + goto FAILED; + } + class_has_8bitchar--; /* Undo the speculative increase. */ + class_one_char -= 2; /* Undo the speculative increase. */ + c = *ptr; /* Get the final character and fall through */ + break; + } + } + + /* Fall through if the escape just defined a single character (c >= 0). + This may be greater than 256. */ + + escape = 0; + + } /* End of backslash handling */ + + /* A character may be followed by '-' to form a range. However, Perl does + not permit ']' to be the end of the range. A '-' character at the end is + treated as a literal. Perl ignores orphaned \E sequences entirely. The + code for handling \Q and \E is messy. */ + + CHECK_RANGE: + while (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E) + { + inescq = FALSE; + ptr += 2; + } + oldptr = ptr; + + /* Remember if \r or \n were explicitly used */ + + if (c == CHAR_CR || c == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF; + + /* Check for range */ + + if (!inescq && ptr[1] == CHAR_MINUS) + { + pcre_uint32 d; + ptr += 2; + while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2; + + /* If we hit \Q (not followed by \E) at this point, go into escaped + mode. */ + + while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_Q) + { + ptr += 2; + if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) + { ptr += 2; continue; } + inescq = TRUE; + break; + } + + /* Minus (hyphen) at the end of a class is treated as a literal, so put + back the pointer and jump to handle the character that preceded it. */ + + if (*ptr == CHAR_NULL || (!inescq && *ptr == CHAR_RIGHT_SQUARE_BRACKET)) + { + ptr = oldptr; + goto CLASS_SINGLE_CHARACTER; + } + + /* Otherwise, we have a potential range; pick up the next character */ + +#ifdef SUPPORT_UTF + if (utf) + { /* Braces are required because the */ + GETCHARLEN(d, ptr, ptr); /* macro generates multiple statements */ + } + else +#endif + d = *ptr; /* Not UTF-8 mode */ + + /* The second part of a range can be a single-character escape + sequence, but not any of the other escapes. Perl treats a hyphen as a + literal in such circumstances. However, in Perl's warning mode, a + warning is given, so PCRE now faults it as it is almost certainly a + mistake on the user's part. */ + + if (!inescq) + { + if (d == CHAR_BACKSLASH) + { + int descape; + descape = check_escape(&ptr, &d, errorcodeptr, cd->bracount, options, TRUE); + if (*errorcodeptr != 0) goto FAILED; + + /* 0 means a character was put into d; \b is backspace; any other + special causes an error. */ + + if (descape != 0) + { + if (descape == ESC_b) d = CHAR_BS; else + { + *errorcodeptr = ERR83; + goto FAILED; + } + } + } + + /* A hyphen followed by a POSIX class is treated in the same way. */ + + else if (d == CHAR_LEFT_SQUARE_BRACKET && + (ptr[1] == CHAR_COLON || ptr[1] == CHAR_DOT || + ptr[1] == CHAR_EQUALS_SIGN) && + check_posix_syntax(ptr, &tempptr)) + { + *errorcodeptr = ERR83; + goto FAILED; + } + } + + /* Check that the two values are in the correct order. Optimize + one-character ranges. */ + + if (d < c) + { + *errorcodeptr = ERR8; + goto FAILED; + } + if (d == c) goto CLASS_SINGLE_CHARACTER; /* A few lines below */ + + /* We have found a character range, so single character optimizations + cannot be done anymore. Any value greater than 1 indicates that there + is more than one character. */ + + class_one_char = 2; + + /* Remember an explicit \r or \n, and add the range to the class. */ + + if (d == CHAR_CR || d == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF; + + class_has_8bitchar += + add_to_class(classbits, &class_uchardata, options, cd, c, d); + + continue; /* Go get the next char in the class */ + } + + /* Handle a single character - we can get here for a normal non-escape + char, or after \ that introduces a single character or for an apparent + range that isn't. Only the value 1 matters for class_one_char, so don't + increase it if it is already 2 or more ... just in case there's a class + with a zillion characters in it. */ + + CLASS_SINGLE_CHARACTER: + if (class_one_char < 2) class_one_char++; + + /* If xclass_has_prop is false and class_one_char is 1, we have the first + single character in the class, and there have been no prior ranges, or + XCLASS items generated by escapes. If this is the final character in the + class, we can optimize by turning the item into a 1-character OP_CHAR[I] + if it's positive, or OP_NOT[I] if it's negative. In the positive case, it + can cause firstchar to be set. Otherwise, there can be no first char if + this item is first, whatever repeat count may follow. In the case of + reqchar, save the previous value for reinstating. */ + + if (!inescq && +#ifdef SUPPORT_UCP + !xclass_has_prop && +#endif + class_one_char == 1 && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET) + { + ptr++; + zeroreqchar = reqchar; + zeroreqcharflags = reqcharflags; + + if (negate_class) + { +#ifdef SUPPORT_UCP + int d; +#endif + if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE; + zerofirstchar = firstchar; + zerofirstcharflags = firstcharflags; + + /* For caseless UTF-8 mode when UCP support is available, check + whether this character has more than one other case. If so, generate + a special OP_NOTPROP item instead of OP_NOTI. */ + +#ifdef SUPPORT_UCP + if (utf && (options & PCRE_CASELESS) != 0 && + (d = UCD_CASESET(c)) != 0) + { + *code++ = OP_NOTPROP; + *code++ = PT_CLIST; + *code++ = d; + } + else +#endif + /* Char has only one other case, or UCP not available */ + + { + *code++ = ((options & PCRE_CASELESS) != 0)? OP_NOTI: OP_NOT; +#if defined SUPPORT_UTF && !defined COMPILE_PCRE32 + if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR) + code += PRIV(ord2utf)(c, code); + else +#endif + *code++ = c; + } + + /* We are finished with this character class */ + + goto END_CLASS; + } + + /* For a single, positive character, get the value into mcbuffer, and + then we can handle this with the normal one-character code. */ + +#if defined SUPPORT_UTF && !defined COMPILE_PCRE32 + if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR) + mclength = PRIV(ord2utf)(c, mcbuffer); + else +#endif + { + mcbuffer[0] = c; + mclength = 1; + } + goto ONE_CHAR; + } /* End of 1-char optimization */ + + /* There is more than one character in the class, or an XCLASS item + has been generated. Add this character to the class. */ + + class_has_8bitchar += + add_to_class(classbits, &class_uchardata, options, cd, c, c); + } + + /* Loop until ']' reached. This "while" is the end of the "do" far above. + If we are at the end of an internal nested string, revert to the outer + string. */ + + while (((c = *(++ptr)) != CHAR_NULL || + (nestptr != NULL && + (ptr = nestptr, nestptr = NULL, c = *(++ptr)) != CHAR_NULL)) && + (c != CHAR_RIGHT_SQUARE_BRACKET || inescq)); + + /* Check for missing terminating ']' */ + + if (c == CHAR_NULL) + { + *errorcodeptr = ERR6; + goto FAILED; + } + + /* We will need an XCLASS if data has been placed in class_uchardata. In + the second phase this is a sufficient test. However, in the pre-compile + phase, class_uchardata gets emptied to prevent workspace overflow, so it + only if the very last character in the class needs XCLASS will it contain + anything at this point. For this reason, xclass gets set TRUE above when + uchar_classdata is emptied, and that's why this code is the way it is here + instead of just doing a test on class_uchardata below. */ + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + if (class_uchardata > class_uchardata_base) xclass = TRUE; +#endif + + /* If this is the first thing in the branch, there can be no first char + setting, whatever the repeat count. Any reqchar setting must remain + unchanged after any kind of repeat. */ + + if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE; + zerofirstchar = firstchar; + zerofirstcharflags = firstcharflags; + zeroreqchar = reqchar; + zeroreqcharflags = reqcharflags; + + /* If there are characters with values > 255, we have to compile an + extended class, with its own opcode, unless there was a negated special + such as \S in the class, and PCRE_UCP is not set, because in that case all + characters > 255 are in the class, so any that were explicitly given as + well can be ignored. If (when there are explicit characters > 255 that must + be listed) there are no characters < 256, we can omit the bitmap in the + actual compiled code. */ + +#ifdef SUPPORT_UTF + if (xclass && (xclass_has_prop || !should_flip_negation || + (options & PCRE_UCP) != 0)) +#elif !defined COMPILE_PCRE8 + if (xclass && (xclass_has_prop || !should_flip_negation)) +#endif +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + { + *class_uchardata++ = XCL_END; /* Marks the end of extra data */ + *code++ = OP_XCLASS; + code += LINK_SIZE; + *code = negate_class? XCL_NOT:0; + if (xclass_has_prop) *code |= XCL_HASPROP; + + /* If the map is required, move up the extra data to make room for it; + otherwise just move the code pointer to the end of the extra data. */ + + if (class_has_8bitchar > 0) + { + *code++ |= XCL_MAP; + memmove(code + (32 / sizeof(pcre_uchar)), code, + IN_UCHARS(class_uchardata - code)); + if (negate_class && !xclass_has_prop) + for (c = 0; c < 32; c++) classbits[c] = ~classbits[c]; + memcpy(code, classbits, 32); + code = class_uchardata + (32 / sizeof(pcre_uchar)); + } + else code = class_uchardata; + + /* Now fill in the complete length of the item */ + + PUT(previous, 1, (int)(code - previous)); + break; /* End of class handling */ + } + + /* Even though any XCLASS list is now discarded, we must allow for + its memory. */ + + if (lengthptr != NULL) + *lengthptr += (int)(class_uchardata - class_uchardata_base); +#endif + + /* If there are no characters > 255, or they are all to be included or + excluded, set the opcode to OP_CLASS or OP_NCLASS, depending on whether the + whole class was negated and whether there were negative specials such as \S + (non-UCP) in the class. Then copy the 32-byte map into the code vector, + negating it if necessary. */ + + *code++ = (negate_class == should_flip_negation) ? OP_CLASS : OP_NCLASS; + if (lengthptr == NULL) /* Save time in the pre-compile phase */ + { + if (negate_class) + for (c = 0; c < 32; c++) classbits[c] = ~classbits[c]; + memcpy(code, classbits, 32); + } + code += 32 / sizeof(pcre_uchar); + + END_CLASS: + break; + + + /* ===================================================================*/ + /* Various kinds of repeat; '{' is not necessarily a quantifier, but this + has been tested above. */ + + case CHAR_LEFT_CURLY_BRACKET: + if (!is_quantifier) goto NORMAL_CHAR; + ptr = read_repeat_counts(ptr+1, &repeat_min, &repeat_max, errorcodeptr); + if (*errorcodeptr != 0) goto FAILED; + goto REPEAT; + + case CHAR_ASTERISK: + repeat_min = 0; + repeat_max = -1; + goto REPEAT; + + case CHAR_PLUS: + repeat_min = 1; + repeat_max = -1; + goto REPEAT; + + case CHAR_QUESTION_MARK: + repeat_min = 0; + repeat_max = 1; + + REPEAT: + if (previous == NULL) + { + *errorcodeptr = ERR9; + goto FAILED; + } + + if (repeat_min == 0) + { + firstchar = zerofirstchar; /* Adjust for zero repeat */ + firstcharflags = zerofirstcharflags; + reqchar = zeroreqchar; /* Ditto */ + reqcharflags = zeroreqcharflags; + } + + /* Remember whether this is a variable length repeat */ + + reqvary = (repeat_min == repeat_max)? 0 : REQ_VARY; + + op_type = 0; /* Default single-char op codes */ + possessive_quantifier = FALSE; /* Default not possessive quantifier */ + + /* Save start of previous item, in case we have to move it up in order to + insert something before it. */ + + tempcode = previous; + + /* Before checking for a possessive quantifier, we must skip over + whitespace and comments in extended mode because Perl allows white space at + this point. */ + + if ((options & PCRE_EXTENDED) != 0) + { + const pcre_uchar *p = ptr + 1; + for (;;) + { + while (MAX_255(*p) && (cd->ctypes[*p] & ctype_space) != 0) p++; + if (*p != CHAR_NUMBER_SIGN) break; + p++; + while (*p != CHAR_NULL) + { + if (IS_NEWLINE(p)) /* For non-fixed-length newline cases, */ + { /* IS_NEWLINE sets cd->nllen. */ + p += cd->nllen; + break; + } + p++; +#ifdef SUPPORT_UTF + if (utf) FORWARDCHAR(p); +#endif + } /* Loop for comment characters */ + } /* Loop for multiple comments */ + ptr = p - 1; /* Character before the next significant one. */ + } + + /* If the next character is '+', we have a possessive quantifier. This + implies greediness, whatever the setting of the PCRE_UNGREEDY option. + If the next character is '?' this is a minimizing repeat, by default, + but if PCRE_UNGREEDY is set, it works the other way round. We change the + repeat type to the non-default. */ + + if (ptr[1] == CHAR_PLUS) + { + repeat_type = 0; /* Force greedy */ + possessive_quantifier = TRUE; + ptr++; + } + else if (ptr[1] == CHAR_QUESTION_MARK) + { + repeat_type = greedy_non_default; + ptr++; + } + else repeat_type = greedy_default; + + /* If previous was a recursion call, wrap it in atomic brackets so that + previous becomes the atomic group. All recursions were so wrapped in the + past, but it no longer happens for non-repeated recursions. In fact, the + repeated ones could be re-implemented independently so as not to need this, + but for the moment we rely on the code for repeating groups. */ + + if (*previous == OP_RECURSE) + { + memmove(previous + 1 + LINK_SIZE, previous, IN_UCHARS(1 + LINK_SIZE)); + *previous = OP_ONCE; + PUT(previous, 1, 2 + 2*LINK_SIZE); + previous[2 + 2*LINK_SIZE] = OP_KET; + PUT(previous, 3 + 2*LINK_SIZE, 2 + 2*LINK_SIZE); + code += 2 + 2 * LINK_SIZE; + length_prevgroup = 3 + 3*LINK_SIZE; + + /* When actually compiling, we need to check whether this was a forward + reference, and if so, adjust the offset. */ + + if (lengthptr == NULL && cd->hwm >= cd->start_workspace + LINK_SIZE) + { + int offset = GET(cd->hwm, -LINK_SIZE); + if (offset == previous + 1 - cd->start_code) + PUT(cd->hwm, -LINK_SIZE, offset + 1 + LINK_SIZE); + } + } + + /* Now handle repetition for the different types of item. */ + + /* If previous was a character or negated character match, abolish the item + and generate a repeat item instead. If a char item has a minimum of more + than one, ensure that it is set in reqchar - it might not be if a sequence + such as x{3} is the first thing in a branch because the x will have gone + into firstchar instead. */ + + if (*previous == OP_CHAR || *previous == OP_CHARI + || *previous == OP_NOT || *previous == OP_NOTI) + { + switch (*previous) + { + default: /* Make compiler happy. */ + case OP_CHAR: op_type = OP_STAR - OP_STAR; break; + case OP_CHARI: op_type = OP_STARI - OP_STAR; break; + case OP_NOT: op_type = OP_NOTSTAR - OP_STAR; break; + case OP_NOTI: op_type = OP_NOTSTARI - OP_STAR; break; + } + + /* Deal with UTF characters that take up more than one character. It's + easier to write this out separately than try to macrify it. Use c to + hold the length of the character in bytes, plus UTF_LENGTH to flag that + it's a length rather than a small character. */ + +#if defined SUPPORT_UTF && !defined COMPILE_PCRE32 + if (utf && NOT_FIRSTCHAR(code[-1])) + { + pcre_uchar *lastchar = code - 1; + BACKCHAR(lastchar); + c = (int)(code - lastchar); /* Length of UTF-8 character */ + memcpy(utf_chars, lastchar, IN_UCHARS(c)); /* Save the char */ + c |= UTF_LENGTH; /* Flag c as a length */ + } + else +#endif /* SUPPORT_UTF */ + + /* Handle the case of a single charater - either with no UTF support, or + with UTF disabled, or for a single character UTF character. */ + { + c = code[-1]; + if (*previous <= OP_CHARI && repeat_min > 1) + { + reqchar = c; + reqcharflags = req_caseopt | cd->req_varyopt; + } + } + + goto OUTPUT_SINGLE_REPEAT; /* Code shared with single character types */ + } + + /* If previous was a character type match (\d or similar), abolish it and + create a suitable repeat item. The code is shared with single-character + repeats by setting op_type to add a suitable offset into repeat_type. Note + the the Unicode property types will be present only when SUPPORT_UCP is + defined, but we don't wrap the little bits of code here because it just + makes it horribly messy. */ + + else if (*previous < OP_EODN) + { + pcre_uchar *oldcode; + int prop_type, prop_value; + op_type = OP_TYPESTAR - OP_STAR; /* Use type opcodes */ + c = *previous; + + OUTPUT_SINGLE_REPEAT: + if (*previous == OP_PROP || *previous == OP_NOTPROP) + { + prop_type = previous[1]; + prop_value = previous[2]; + } + else prop_type = prop_value = -1; + + oldcode = code; + code = previous; /* Usually overwrite previous item */ + + /* If the maximum is zero then the minimum must also be zero; Perl allows + this case, so we do too - by simply omitting the item altogether. */ + + if (repeat_max == 0) goto END_REPEAT; + + /* Combine the op_type with the repeat_type */ + + repeat_type += op_type; + + /* A minimum of zero is handled either as the special case * or ?, or as + an UPTO, with the maximum given. */ + + if (repeat_min == 0) + { + if (repeat_max == -1) *code++ = OP_STAR + repeat_type; + else if (repeat_max == 1) *code++ = OP_QUERY + repeat_type; + else + { + *code++ = OP_UPTO + repeat_type; + PUT2INC(code, 0, repeat_max); + } + } + + /* A repeat minimum of 1 is optimized into some special cases. If the + maximum is unlimited, we use OP_PLUS. Otherwise, the original item is + left in place and, if the maximum is greater than 1, we use OP_UPTO with + one less than the maximum. */ + + else if (repeat_min == 1) + { + if (repeat_max == -1) + *code++ = OP_PLUS + repeat_type; + else + { + code = oldcode; /* leave previous item in place */ + if (repeat_max == 1) goto END_REPEAT; + *code++ = OP_UPTO + repeat_type; + PUT2INC(code, 0, repeat_max - 1); + } + } + + /* The case {n,n} is just an EXACT, while the general case {n,m} is + handled as an EXACT followed by an UPTO. */ + + else + { + *code++ = OP_EXACT + op_type; /* NB EXACT doesn't have repeat_type */ + PUT2INC(code, 0, repeat_min); + + /* If the maximum is unlimited, insert an OP_STAR. Before doing so, + we have to insert the character for the previous code. For a repeated + Unicode property match, there are two extra bytes that define the + required property. In UTF-8 mode, long characters have their length in + c, with the UTF_LENGTH bit as a flag. */ + + if (repeat_max < 0) + { +#if defined SUPPORT_UTF && !defined COMPILE_PCRE32 + if (utf && (c & UTF_LENGTH) != 0) + { + memcpy(code, utf_chars, IN_UCHARS(c & 7)); + code += c & 7; + } + else +#endif + { + *code++ = c; + if (prop_type >= 0) + { + *code++ = prop_type; + *code++ = prop_value; + } + } + *code++ = OP_STAR + repeat_type; + } + + /* Else insert an UPTO if the max is greater than the min, again + preceded by the character, for the previously inserted code. If the + UPTO is just for 1 instance, we can use QUERY instead. */ + + else if (repeat_max != repeat_min) + { +#if defined SUPPORT_UTF && !defined COMPILE_PCRE32 + if (utf && (c & UTF_LENGTH) != 0) + { + memcpy(code, utf_chars, IN_UCHARS(c & 7)); + code += c & 7; + } + else +#endif + *code++ = c; + if (prop_type >= 0) + { + *code++ = prop_type; + *code++ = prop_value; + } + repeat_max -= repeat_min; + + if (repeat_max == 1) + { + *code++ = OP_QUERY + repeat_type; + } + else + { + *code++ = OP_UPTO + repeat_type; + PUT2INC(code, 0, repeat_max); + } + } + } + + /* The character or character type itself comes last in all cases. */ + +#if defined SUPPORT_UTF && !defined COMPILE_PCRE32 + if (utf && (c & UTF_LENGTH) != 0) + { + memcpy(code, utf_chars, IN_UCHARS(c & 7)); + code += c & 7; + } + else +#endif + *code++ = c; + + /* For a repeated Unicode property match, there are two extra bytes that + define the required property. */ + +#ifdef SUPPORT_UCP + if (prop_type >= 0) + { + *code++ = prop_type; + *code++ = prop_value; + } +#endif + } + + /* If previous was a character class or a back reference, we put the repeat + stuff after it, but just skip the item if the repeat was {0,0}. */ + + else if (*previous == OP_CLASS || *previous == OP_NCLASS || +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + *previous == OP_XCLASS || +#endif + *previous == OP_REF || *previous == OP_REFI || + *previous == OP_DNREF || *previous == OP_DNREFI) + { + if (repeat_max == 0) + { + code = previous; + goto END_REPEAT; + } + + if (repeat_min == 0 && repeat_max == -1) + *code++ = OP_CRSTAR + repeat_type; + else if (repeat_min == 1 && repeat_max == -1) + *code++ = OP_CRPLUS + repeat_type; + else if (repeat_min == 0 && repeat_max == 1) + *code++ = OP_CRQUERY + repeat_type; + else + { + *code++ = OP_CRRANGE + repeat_type; + PUT2INC(code, 0, repeat_min); + if (repeat_max == -1) repeat_max = 0; /* 2-byte encoding for max */ + PUT2INC(code, 0, repeat_max); + } + } + + /* If previous was a bracket group, we may have to replicate it in certain + cases. Note that at this point we can encounter only the "basic" bracket + opcodes such as BRA and CBRA, as this is the place where they get converted + into the more special varieties such as BRAPOS and SBRA. A test for >= + OP_ASSERT and <= OP_COND includes ASSERT, ASSERT_NOT, ASSERTBACK, + ASSERTBACK_NOT, ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND. + Originally, PCRE did not allow repetition of assertions, but now it does, + for Perl compatibility. */ + + else if (*previous >= OP_ASSERT && *previous <= OP_COND) + { + register int i; + int len = (int)(code - previous); + size_t base_hwm_offset = item_hwm_offset; + pcre_uchar *bralink = NULL; + pcre_uchar *brazeroptr = NULL; + + /* Repeating a DEFINE group is pointless, but Perl allows the syntax, so + we just ignore the repeat. */ + + if (*previous == OP_COND && previous[LINK_SIZE+1] == OP_DEF) + goto END_REPEAT; + + /* There is no sense in actually repeating assertions. The only potential + use of repetition is in cases when the assertion is optional. Therefore, + if the minimum is greater than zero, just ignore the repeat. If the + maximum is not zero or one, set it to 1. */ + + if (*previous < OP_ONCE) /* Assertion */ + { + if (repeat_min > 0) goto END_REPEAT; + if (repeat_max < 0 || repeat_max > 1) repeat_max = 1; + } + + /* The case of a zero minimum is special because of the need to stick + OP_BRAZERO in front of it, and because the group appears once in the + data, whereas in other cases it appears the minimum number of times. For + this reason, it is simplest to treat this case separately, as otherwise + the code gets far too messy. There are several special subcases when the + minimum is zero. */ + + if (repeat_min == 0) + { + /* If the maximum is also zero, we used to just omit the group from the + output altogether, like this: + + ** if (repeat_max == 0) + ** { + ** code = previous; + ** goto END_REPEAT; + ** } + + However, that fails when a group or a subgroup within it is referenced + as a subroutine from elsewhere in the pattern, so now we stick in + OP_SKIPZERO in front of it so that it is skipped on execution. As we + don't have a list of which groups are referenced, we cannot do this + selectively. + + If the maximum is 1 or unlimited, we just have to stick in the BRAZERO + and do no more at this point. However, we do need to adjust any + OP_RECURSE calls inside the group that refer to the group itself or any + internal or forward referenced group, because the offset is from the + start of the whole regex. Temporarily terminate the pattern while doing + this. */ + + if (repeat_max <= 1) /* Covers 0, 1, and unlimited */ + { + *code = OP_END; + adjust_recurse(previous, 1, utf, cd, item_hwm_offset); + memmove(previous + 1, previous, IN_UCHARS(len)); + code++; + if (repeat_max == 0) + { + *previous++ = OP_SKIPZERO; + goto END_REPEAT; + } + brazeroptr = previous; /* Save for possessive optimizing */ + *previous++ = OP_BRAZERO + repeat_type; + } + + /* If the maximum is greater than 1 and limited, we have to replicate + in a nested fashion, sticking OP_BRAZERO before each set of brackets. + The first one has to be handled carefully because it's the original + copy, which has to be moved up. The remainder can be handled by code + that is common with the non-zero minimum case below. We have to + adjust the value or repeat_max, since one less copy is required. Once + again, we may have to adjust any OP_RECURSE calls inside the group. */ + + else + { + int offset; + *code = OP_END; + adjust_recurse(previous, 2 + LINK_SIZE, utf, cd, item_hwm_offset); + memmove(previous + 2 + LINK_SIZE, previous, IN_UCHARS(len)); + code += 2 + LINK_SIZE; + *previous++ = OP_BRAZERO + repeat_type; + *previous++ = OP_BRA; + + /* We chain together the bracket offset fields that have to be + filled in later when the ends of the brackets are reached. */ + + offset = (bralink == NULL)? 0 : (int)(previous - bralink); + bralink = previous; + PUTINC(previous, 0, offset); + } + + repeat_max--; + } + + /* If the minimum is greater than zero, replicate the group as many + times as necessary, and adjust the maximum to the number of subsequent + copies that we need. If we set a first char from the group, and didn't + set a required char, copy the latter from the former. If there are any + forward reference subroutine calls in the group, there will be entries on + the workspace list; replicate these with an appropriate increment. */ + + else + { + if (repeat_min > 1) + { + /* In the pre-compile phase, we don't actually do the replication. We + just adjust the length as if we had. Do some paranoid checks for + potential integer overflow. The INT64_OR_DOUBLE type is a 64-bit + integer type when available, otherwise double. */ + + if (lengthptr != NULL) + { + int delta = (repeat_min - 1)*length_prevgroup; + if ((INT64_OR_DOUBLE)(repeat_min - 1)* + (INT64_OR_DOUBLE)length_prevgroup > + (INT64_OR_DOUBLE)INT_MAX || + OFLOW_MAX - *lengthptr < delta) + { + *errorcodeptr = ERR20; + goto FAILED; + } + *lengthptr += delta; + } + + /* This is compiling for real. If there is a set first byte for + the group, and we have not yet set a "required byte", set it. Make + sure there is enough workspace for copying forward references before + doing the copy. */ + + else + { + if (groupsetfirstchar && reqcharflags < 0) + { + reqchar = firstchar; + reqcharflags = firstcharflags; + } + + for (i = 1; i < repeat_min; i++) + { + pcre_uchar *hc; + size_t this_hwm_offset = cd->hwm - cd->start_workspace; + memcpy(code, previous, IN_UCHARS(len)); + + while (cd->hwm > cd->start_workspace + cd->workspace_size - + WORK_SIZE_SAFETY_MARGIN - + (this_hwm_offset - base_hwm_offset)) + { + *errorcodeptr = expand_workspace(cd); + if (*errorcodeptr != 0) goto FAILED; + } + + for (hc = (pcre_uchar *)cd->start_workspace + base_hwm_offset; + hc < (pcre_uchar *)cd->start_workspace + this_hwm_offset; + hc += LINK_SIZE) + { + PUT(cd->hwm, 0, GET(hc, 0) + len); + cd->hwm += LINK_SIZE; + } + base_hwm_offset = this_hwm_offset; + code += len; + } + } + } + + if (repeat_max > 0) repeat_max -= repeat_min; + } + + /* This code is common to both the zero and non-zero minimum cases. If + the maximum is limited, it replicates the group in a nested fashion, + remembering the bracket starts on a stack. In the case of a zero minimum, + the first one was set up above. In all cases the repeat_max now specifies + the number of additional copies needed. Again, we must remember to + replicate entries on the forward reference list. */ + + if (repeat_max >= 0) + { + /* In the pre-compile phase, we don't actually do the replication. We + just adjust the length as if we had. For each repetition we must add 1 + to the length for BRAZERO and for all but the last repetition we must + add 2 + 2*LINKSIZE to allow for the nesting that occurs. Do some + paranoid checks to avoid integer overflow. The INT64_OR_DOUBLE type is + a 64-bit integer type when available, otherwise double. */ + + if (lengthptr != NULL && repeat_max > 0) + { + int delta = repeat_max * (length_prevgroup + 1 + 2 + 2*LINK_SIZE) - + 2 - 2*LINK_SIZE; /* Last one doesn't nest */ + if ((INT64_OR_DOUBLE)repeat_max * + (INT64_OR_DOUBLE)(length_prevgroup + 1 + 2 + 2*LINK_SIZE) + > (INT64_OR_DOUBLE)INT_MAX || + OFLOW_MAX - *lengthptr < delta) + { + *errorcodeptr = ERR20; + goto FAILED; + } + *lengthptr += delta; + } + + /* This is compiling for real */ + + else for (i = repeat_max - 1; i >= 0; i--) + { + pcre_uchar *hc; + size_t this_hwm_offset = cd->hwm - cd->start_workspace; + + *code++ = OP_BRAZERO + repeat_type; + + /* All but the final copy start a new nesting, maintaining the + chain of brackets outstanding. */ + + if (i != 0) + { + int offset; + *code++ = OP_BRA; + offset = (bralink == NULL)? 0 : (int)(code - bralink); + bralink = code; + PUTINC(code, 0, offset); + } + + memcpy(code, previous, IN_UCHARS(len)); + + /* Ensure there is enough workspace for forward references before + copying them. */ + + while (cd->hwm > cd->start_workspace + cd->workspace_size - + WORK_SIZE_SAFETY_MARGIN - + (this_hwm_offset - base_hwm_offset)) + { + *errorcodeptr = expand_workspace(cd); + if (*errorcodeptr != 0) goto FAILED; + } + + for (hc = (pcre_uchar *)cd->start_workspace + base_hwm_offset; + hc < (pcre_uchar *)cd->start_workspace + this_hwm_offset; + hc += LINK_SIZE) + { + PUT(cd->hwm, 0, GET(hc, 0) + len + ((i != 0)? 2+LINK_SIZE : 1)); + cd->hwm += LINK_SIZE; + } + base_hwm_offset = this_hwm_offset; + code += len; + } + + /* Now chain through the pending brackets, and fill in their length + fields (which are holding the chain links pro tem). */ + + while (bralink != NULL) + { + int oldlinkoffset; + int offset = (int)(code - bralink + 1); + pcre_uchar *bra = code - offset; + oldlinkoffset = GET(bra, 1); + bralink = (oldlinkoffset == 0)? NULL : bralink - oldlinkoffset; + *code++ = OP_KET; + PUTINC(code, 0, offset); + PUT(bra, 1, offset); + } + } + + /* If the maximum is unlimited, set a repeater in the final copy. For + ONCE brackets, that's all we need to do. However, possessively repeated + ONCE brackets can be converted into non-capturing brackets, as the + behaviour of (?:xx)++ is the same as (?>xx)++ and this saves having to + deal with possessive ONCEs specially. + + Otherwise, when we are doing the actual compile phase, check to see + whether this group is one that could match an empty string. If so, + convert the initial operator to the S form (e.g. OP_BRA -> OP_SBRA) so + that runtime checking can be done. [This check is also applied to ONCE + groups at runtime, but in a different way.] + + Then, if the quantifier was possessive and the bracket is not a + conditional, we convert the BRA code to the POS form, and the KET code to + KETRPOS. (It turns out to be convenient at runtime to detect this kind of + subpattern at both the start and at the end.) The use of special opcodes + makes it possible to reduce greatly the stack usage in pcre_exec(). If + the group is preceded by OP_BRAZERO, convert this to OP_BRAPOSZERO. + + Then, if the minimum number of matches is 1 or 0, cancel the possessive + flag so that the default action below, of wrapping everything inside + atomic brackets, does not happen. When the minimum is greater than 1, + there will be earlier copies of the group, and so we still have to wrap + the whole thing. */ + + else + { + pcre_uchar *ketcode = code - 1 - LINK_SIZE; + pcre_uchar *bracode = ketcode - GET(ketcode, 1); + + /* Convert possessive ONCE brackets to non-capturing */ + + if ((*bracode == OP_ONCE || *bracode == OP_ONCE_NC) && + possessive_quantifier) *bracode = OP_BRA; + + /* For non-possessive ONCE brackets, all we need to do is to + set the KET. */ + + if (*bracode == OP_ONCE || *bracode == OP_ONCE_NC) + *ketcode = OP_KETRMAX + repeat_type; + + /* Handle non-ONCE brackets and possessive ONCEs (which have been + converted to non-capturing above). */ + + else + { + /* In the compile phase, check for empty string matching. */ + + if (lengthptr == NULL) + { + pcre_uchar *scode = bracode; + do + { + if (could_be_empty_branch(scode, ketcode, utf, cd, NULL)) + { + *bracode += OP_SBRA - OP_BRA; + break; + } + scode += GET(scode, 1); + } + while (*scode == OP_ALT); + } + + /* A conditional group with only one branch has an implicit empty + alternative branch. */ + + if (*bracode == OP_COND && bracode[GET(bracode,1)] != OP_ALT) + *bracode = OP_SCOND; + + /* Handle possessive quantifiers. */ + + if (possessive_quantifier) + { + /* For COND brackets, we wrap the whole thing in a possessively + repeated non-capturing bracket, because we have not invented POS + versions of the COND opcodes. Because we are moving code along, we + must ensure that any pending recursive references are updated. */ + + if (*bracode == OP_COND || *bracode == OP_SCOND) + { + int nlen = (int)(code - bracode); + *code = OP_END; + adjust_recurse(bracode, 1 + LINK_SIZE, utf, cd, item_hwm_offset); + memmove(bracode + 1 + LINK_SIZE, bracode, IN_UCHARS(nlen)); + code += 1 + LINK_SIZE; + nlen += 1 + LINK_SIZE; + *bracode = (*bracode == OP_COND)? OP_BRAPOS : OP_SBRAPOS; + *code++ = OP_KETRPOS; + PUTINC(code, 0, nlen); + PUT(bracode, 1, nlen); + } + + /* For non-COND brackets, we modify the BRA code and use KETRPOS. */ + + else + { + *bracode += 1; /* Switch to xxxPOS opcodes */ + *ketcode = OP_KETRPOS; + } + + /* If the minimum is zero, mark it as possessive, then unset the + possessive flag when the minimum is 0 or 1. */ + + if (brazeroptr != NULL) *brazeroptr = OP_BRAPOSZERO; + if (repeat_min < 2) possessive_quantifier = FALSE; + } + + /* Non-possessive quantifier */ + + else *ketcode = OP_KETRMAX + repeat_type; + } + } + } + + /* If previous is OP_FAIL, it was generated by an empty class [] in + JavaScript mode. The other ways in which OP_FAIL can be generated, that is + by (*FAIL) or (?!) set previous to NULL, which gives a "nothing to repeat" + error above. We can just ignore the repeat in JS case. */ + + else if (*previous == OP_FAIL) goto END_REPEAT; + + /* Else there's some kind of shambles */ + + else + { + *errorcodeptr = ERR11; + goto FAILED; + } + + /* If the character following a repeat is '+', possessive_quantifier is + TRUE. For some opcodes, there are special alternative opcodes for this + case. For anything else, we wrap the entire repeated item inside OP_ONCE + brackets. Logically, the '+' notation is just syntactic sugar, taken from + Sun's Java package, but the special opcodes can optimize it. + + Some (but not all) possessively repeated subpatterns have already been + completely handled in the code just above. For them, possessive_quantifier + is always FALSE at this stage. Note that the repeated item starts at + tempcode, not at previous, which might be the first part of a string whose + (former) last char we repeated. */ + + if (possessive_quantifier) + { + int len; + + /* Possessifying an EXACT quantifier has no effect, so we can ignore it. + However, QUERY, STAR, or UPTO may follow (for quantifiers such as {5,6}, + {5,}, or {5,10}). We skip over an EXACT item; if the length of what + remains is greater than zero, there's a further opcode that can be + handled. If not, do nothing, leaving the EXACT alone. */ + + switch(*tempcode) + { + case OP_TYPEEXACT: + tempcode += PRIV(OP_lengths)[*tempcode] + + ((tempcode[1 + IMM2_SIZE] == OP_PROP + || tempcode[1 + IMM2_SIZE] == OP_NOTPROP)? 2 : 0); + break; + + /* CHAR opcodes are used for exacts whose count is 1. */ + + case OP_CHAR: + case OP_CHARI: + case OP_NOT: + case OP_NOTI: + case OP_EXACT: + case OP_EXACTI: + case OP_NOTEXACT: + case OP_NOTEXACTI: + tempcode += PRIV(OP_lengths)[*tempcode]; +#ifdef SUPPORT_UTF + if (utf && HAS_EXTRALEN(tempcode[-1])) + tempcode += GET_EXTRALEN(tempcode[-1]); +#endif + break; + + /* For the class opcodes, the repeat operator appears at the end; + adjust tempcode to point to it. */ + + case OP_CLASS: + case OP_NCLASS: + tempcode += 1 + 32/sizeof(pcre_uchar); + break; + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + case OP_XCLASS: + tempcode += GET(tempcode, 1); + break; +#endif + } + + /* If tempcode is equal to code (which points to the end of the repeated + item), it means we have skipped an EXACT item but there is no following + QUERY, STAR, or UPTO; the value of len will be 0, and we do nothing. In + all other cases, tempcode will be pointing to the repeat opcode, and will + be less than code, so the value of len will be greater than 0. */ + + len = (int)(code - tempcode); + if (len > 0) + { + unsigned int repcode = *tempcode; + + /* There is a table for possessifying opcodes, all of which are less + than OP_CALLOUT. A zero entry means there is no possessified version. + */ + + if (repcode < OP_CALLOUT && opcode_possessify[repcode] > 0) + *tempcode = opcode_possessify[repcode]; + + /* For opcode without a special possessified version, wrap the item in + ONCE brackets. Because we are moving code along, we must ensure that any + pending recursive references are updated. */ + + else + { + *code = OP_END; + adjust_recurse(tempcode, 1 + LINK_SIZE, utf, cd, item_hwm_offset); + memmove(tempcode + 1 + LINK_SIZE, tempcode, IN_UCHARS(len)); + code += 1 + LINK_SIZE; + len += 1 + LINK_SIZE; + tempcode[0] = OP_ONCE; + *code++ = OP_KET; + PUTINC(code, 0, len); + PUT(tempcode, 1, len); + } + } + +#ifdef NEVER + if (len > 0) switch (*tempcode) + { + case OP_STAR: *tempcode = OP_POSSTAR; break; + case OP_PLUS: *tempcode = OP_POSPLUS; break; + case OP_QUERY: *tempcode = OP_POSQUERY; break; + case OP_UPTO: *tempcode = OP_POSUPTO; break; + + case OP_STARI: *tempcode = OP_POSSTARI; break; + case OP_PLUSI: *tempcode = OP_POSPLUSI; break; + case OP_QUERYI: *tempcode = OP_POSQUERYI; break; + case OP_UPTOI: *tempcode = OP_POSUPTOI; break; + + case OP_NOTSTAR: *tempcode = OP_NOTPOSSTAR; break; + case OP_NOTPLUS: *tempcode = OP_NOTPOSPLUS; break; + case OP_NOTQUERY: *tempcode = OP_NOTPOSQUERY; break; + case OP_NOTUPTO: *tempcode = OP_NOTPOSUPTO; break; + + case OP_NOTSTARI: *tempcode = OP_NOTPOSSTARI; break; + case OP_NOTPLUSI: *tempcode = OP_NOTPOSPLUSI; break; + case OP_NOTQUERYI: *tempcode = OP_NOTPOSQUERYI; break; + case OP_NOTUPTOI: *tempcode = OP_NOTPOSUPTOI; break; + + case OP_TYPESTAR: *tempcode = OP_TYPEPOSSTAR; break; + case OP_TYPEPLUS: *tempcode = OP_TYPEPOSPLUS; break; + case OP_TYPEQUERY: *tempcode = OP_TYPEPOSQUERY; break; + case OP_TYPEUPTO: *tempcode = OP_TYPEPOSUPTO; break; + + case OP_CRSTAR: *tempcode = OP_CRPOSSTAR; break; + case OP_CRPLUS: *tempcode = OP_CRPOSPLUS; break; + case OP_CRQUERY: *tempcode = OP_CRPOSQUERY; break; + case OP_CRRANGE: *tempcode = OP_CRPOSRANGE; break; + + /* Because we are moving code along, we must ensure that any + pending recursive references are updated. */ + + default: + *code = OP_END; + adjust_recurse(tempcode, 1 + LINK_SIZE, utf, cd, item_hwm_offset); + memmove(tempcode + 1 + LINK_SIZE, tempcode, IN_UCHARS(len)); + code += 1 + LINK_SIZE; + len += 1 + LINK_SIZE; + tempcode[0] = OP_ONCE; + *code++ = OP_KET; + PUTINC(code, 0, len); + PUT(tempcode, 1, len); + break; + } +#endif + } + + /* In all case we no longer have a previous item. We also set the + "follows varying string" flag for subsequently encountered reqchars if + it isn't already set and we have just passed a varying length item. */ + + END_REPEAT: + previous = NULL; + cd->req_varyopt |= reqvary; + break; + + + /* ===================================================================*/ + /* Start of nested parenthesized sub-expression, or comment or lookahead or + lookbehind or option setting or condition or all the other extended + parenthesis forms. */ + + case CHAR_LEFT_PARENTHESIS: + ptr++; + + /* First deal with comments. Putting this code right at the start ensures + that comments have no bad side effects. */ + + if (ptr[0] == CHAR_QUESTION_MARK && ptr[1] == CHAR_NUMBER_SIGN) + { + ptr += 2; + while (*ptr != CHAR_NULL && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++; + if (*ptr == CHAR_NULL) + { + *errorcodeptr = ERR18; + goto FAILED; + } + continue; + } + + /* Now deal with various "verbs" that can be introduced by '*'. */ + + if (ptr[0] == CHAR_ASTERISK && (ptr[1] == ':' + || (MAX_255(ptr[1]) && ((cd->ctypes[ptr[1]] & ctype_letter) != 0)))) + { + int i, namelen; + int arglen = 0; + const char *vn = verbnames; + const pcre_uchar *name = ptr + 1; + const pcre_uchar *arg = NULL; + previous = NULL; + ptr++; + while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_letter) != 0) ptr++; + namelen = (int)(ptr - name); + + /* It appears that Perl allows any characters whatsoever, other than + a closing parenthesis, to appear in arguments, so we no longer insist on + letters, digits, and underscores. */ + + if (*ptr == CHAR_COLON) + { + arg = ++ptr; + while (*ptr != CHAR_NULL && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++; + arglen = (int)(ptr - arg); + if ((unsigned int)arglen > MAX_MARK) + { + *errorcodeptr = ERR75; + goto FAILED; + } + } + + if (*ptr != CHAR_RIGHT_PARENTHESIS) + { + *errorcodeptr = ERR60; + goto FAILED; + } + + /* Scan the table of verb names */ + + for (i = 0; i < verbcount; i++) + { + if (namelen == verbs[i].len && + STRNCMP_UC_C8(name, vn, namelen) == 0) + { + int setverb; + + /* Check for open captures before ACCEPT and convert it to + ASSERT_ACCEPT if in an assertion. */ + + if (verbs[i].op == OP_ACCEPT) + { + open_capitem *oc; + if (arglen != 0) + { + *errorcodeptr = ERR59; + goto FAILED; + } + cd->had_accept = TRUE; + for (oc = cd->open_caps; oc != NULL; oc = oc->next) + { + *code++ = OP_CLOSE; + PUT2INC(code, 0, oc->number); + } + setverb = *code++ = + (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT; + + /* Do not set firstchar after *ACCEPT */ + if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE; + } + + /* Handle other cases with/without an argument */ + + else if (arglen == 0) + { + if (verbs[i].op < 0) /* Argument is mandatory */ + { + *errorcodeptr = ERR66; + goto FAILED; + } + setverb = *code++ = verbs[i].op; + } + + else + { + if (verbs[i].op_arg < 0) /* Argument is forbidden */ + { + *errorcodeptr = ERR59; + goto FAILED; + } + setverb = *code++ = verbs[i].op_arg; + if (lengthptr != NULL) /* In pass 1 just add in the length */ + { /* to avoid potential workspace */ + *lengthptr += arglen; /* overflow. */ + *code++ = 0; + } + else + { + *code++ = arglen; + memcpy(code, arg, IN_UCHARS(arglen)); + code += arglen; + } + *code++ = 0; + } + + switch (setverb) + { + case OP_THEN: + case OP_THEN_ARG: + cd->external_flags |= PCRE_HASTHEN; + break; + + case OP_PRUNE: + case OP_PRUNE_ARG: + case OP_SKIP: + case OP_SKIP_ARG: + cd->had_pruneorskip = TRUE; + break; + } + + break; /* Found verb, exit loop */ + } + + vn += verbs[i].len + 1; + } + + if (i < verbcount) continue; /* Successfully handled a verb */ + *errorcodeptr = ERR60; /* Verb not recognized */ + goto FAILED; + } + + /* Initialize for "real" parentheses */ + + newoptions = options; + skipbytes = 0; + bravalue = OP_CBRA; + item_hwm_offset = cd->hwm - cd->start_workspace; + reset_bracount = FALSE; + + /* Deal with the extended parentheses; all are introduced by '?', and the + appearance of any of them means that this is not a capturing group. */ + + if (*ptr == CHAR_QUESTION_MARK) + { + int i, set, unset, namelen; + int *optset; + const pcre_uchar *name; + pcre_uchar *slot; + + switch (*(++ptr)) + { + /* ------------------------------------------------------------ */ + case CHAR_VERTICAL_LINE: /* Reset capture count for each branch */ + reset_bracount = TRUE; + cd->dupgroups = TRUE; /* Record (?| encountered */ + /* Fall through */ + + /* ------------------------------------------------------------ */ + case CHAR_COLON: /* Non-capturing bracket */ + bravalue = OP_BRA; + ptr++; + break; + + + /* ------------------------------------------------------------ */ + case CHAR_LEFT_PARENTHESIS: + bravalue = OP_COND; /* Conditional group */ + tempptr = ptr; + + /* A condition can be an assertion, a number (referring to a numbered + group's having been set), a name (referring to a named group), or 'R', + referring to recursion. R and R&name are also permitted for + recursion tests. + + There are ways of testing a named group: (?(name)) is used by Python; + Perl 5.10 onwards uses (?() or (?('name')). + + There is one unfortunate ambiguity, caused by history. 'R' can be the + recursive thing or the name 'R' (and similarly for 'R' followed by + digits). We look for a name first; if not found, we try the other case. + + For compatibility with auto-callouts, we allow a callout to be + specified before a condition that is an assertion. First, check for the + syntax of a callout; if found, adjust the temporary pointer that is + used to check for an assertion condition. That's all that is needed! */ + + if (ptr[1] == CHAR_QUESTION_MARK && ptr[2] == CHAR_C) + { + for (i = 3;; i++) if (!IS_DIGIT(ptr[i])) break; + if (ptr[i] == CHAR_RIGHT_PARENTHESIS) + tempptr += i + 1; + } + + /* For conditions that are assertions, check the syntax, and then exit + the switch. This will take control down to where bracketed groups, + including assertions, are processed. */ + + if (tempptr[1] == CHAR_QUESTION_MARK && + (tempptr[2] == CHAR_EQUALS_SIGN || + tempptr[2] == CHAR_EXCLAMATION_MARK || + (tempptr[2] == CHAR_LESS_THAN_SIGN && + (tempptr[3] == CHAR_EQUALS_SIGN || + tempptr[3] == CHAR_EXCLAMATION_MARK)))) + { + cd->iscondassert = TRUE; + break; + } + + /* Other conditions use OP_CREF/OP_DNCREF/OP_RREF/OP_DNRREF, and all + need to skip at least 1+IMM2_SIZE bytes at the start of the group. */ + + code[1+LINK_SIZE] = OP_CREF; + skipbytes = 1+IMM2_SIZE; + refsign = -1; /* => not a number */ + namelen = -1; /* => not a name; must set to avoid warning */ + name = NULL; /* Always set to avoid warning */ + recno = 0; /* Always set to avoid warning */ + + /* Check for a test for recursion in a named group. */ + + ptr++; + if (*ptr == CHAR_R && ptr[1] == CHAR_AMPERSAND) + { + terminator = -1; + ptr += 2; + code[1+LINK_SIZE] = OP_RREF; /* Change the type of test */ + } + + /* Check for a test for a named group's having been set, using the Perl + syntax (?() or (?('name'), and also allow for the original PCRE + syntax of (?(name) or for (?(+n), (?(-n), and just (?(n). */ + + else if (*ptr == CHAR_LESS_THAN_SIGN) + { + terminator = CHAR_GREATER_THAN_SIGN; + ptr++; + } + else if (*ptr == CHAR_APOSTROPHE) + { + terminator = CHAR_APOSTROPHE; + ptr++; + } + else + { + terminator = CHAR_NULL; + if (*ptr == CHAR_MINUS || *ptr == CHAR_PLUS) refsign = *ptr++; + else if (IS_DIGIT(*ptr)) refsign = 0; + } + + /* Handle a number */ + + if (refsign >= 0) + { + while (IS_DIGIT(*ptr)) + { + if (recno > INT_MAX / 10 - 1) /* Integer overflow */ + { + while (IS_DIGIT(*ptr)) ptr++; + *errorcodeptr = ERR61; + goto FAILED; + } + recno = recno * 10 + (int)(*ptr - CHAR_0); + ptr++; + } + } + + /* Otherwise we expect to read a name; anything else is an error. When + a name is one of a number of duplicates, a different opcode is used and + it needs more memory. Unfortunately we cannot tell whether a name is a + duplicate in the first pass, so we have to allow for more memory. */ + + else + { + if (IS_DIGIT(*ptr)) + { + *errorcodeptr = ERR84; + goto FAILED; + } + if (!MAX_255(*ptr) || (cd->ctypes[*ptr] & ctype_word) == 0) + { + *errorcodeptr = ERR28; /* Assertion expected */ + goto FAILED; + } + name = ptr++; + while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0) + { + ptr++; + } + namelen = (int)(ptr - name); + if (lengthptr != NULL) skipbytes += IMM2_SIZE; + } + + /* Check the terminator */ + + if ((terminator > 0 && *ptr++ != (pcre_uchar)terminator) || + *ptr++ != CHAR_RIGHT_PARENTHESIS) + { + ptr--; /* Error offset */ + *errorcodeptr = ERR26; /* Malformed number or name */ + goto FAILED; + } + + /* Do no further checking in the pre-compile phase. */ + + if (lengthptr != NULL) break; + + /* In the real compile we do the work of looking for the actual + reference. If refsign is not negative, it means we have a number in + recno. */ + + if (refsign >= 0) + { + if (recno <= 0) + { + *errorcodeptr = ERR35; + goto FAILED; + } + if (refsign != 0) recno = (refsign == CHAR_MINUS)? + cd->bracount - recno + 1 : recno + cd->bracount; + if (recno <= 0 || recno > cd->final_bracount) + { + *errorcodeptr = ERR15; + goto FAILED; + } + PUT2(code, 2+LINK_SIZE, recno); + if (recno > cd->top_backref) cd->top_backref = recno; + break; + } + + /* Otherwise look for the name. */ + + slot = cd->name_table; + for (i = 0; i < cd->names_found; i++) + { + if (STRNCMP_UC_UC(name, slot+IMM2_SIZE, namelen) == 0) break; + slot += cd->name_entry_size; + } + + /* Found the named subpattern. If the name is duplicated, add one to + the opcode to change CREF/RREF into DNCREF/DNRREF and insert + appropriate data values. Otherwise, just insert the unique subpattern + number. */ + + if (i < cd->names_found) + { + int offset = i++; + int count = 1; + recno = GET2(slot, 0); /* Number from first found */ + if (recno > cd->top_backref) cd->top_backref = recno; + for (; i < cd->names_found; i++) + { + slot += cd->name_entry_size; + if (STRNCMP_UC_UC(name, slot+IMM2_SIZE, namelen) != 0 || + (slot+IMM2_SIZE)[namelen] != 0) break; + count++; + } + + if (count > 1) + { + PUT2(code, 2+LINK_SIZE, offset); + PUT2(code, 2+LINK_SIZE+IMM2_SIZE, count); + skipbytes += IMM2_SIZE; + code[1+LINK_SIZE]++; + } + else /* Not a duplicated name */ + { + PUT2(code, 2+LINK_SIZE, recno); + } + } + + /* If terminator == CHAR_NULL it means that the name followed directly + after the opening parenthesis [e.g. (?(abc)...] and in this case there + are some further alternatives to try. For the cases where terminator != + CHAR_NULL [things like (?(... or (?('name')... or (?(R&name)... ] + we have now checked all the possibilities, so give an error. */ + + else if (terminator != CHAR_NULL) + { + *errorcodeptr = ERR15; + goto FAILED; + } + + /* Check for (?(R) for recursion. Allow digits after R to specify a + specific group number. */ + + else if (*name == CHAR_R) + { + recno = 0; + for (i = 1; i < namelen; i++) + { + if (!IS_DIGIT(name[i])) + { + *errorcodeptr = ERR15; + goto FAILED; + } + if (recno > INT_MAX / 10 - 1) /* Integer overflow */ + { + *errorcodeptr = ERR61; + goto FAILED; + } + recno = recno * 10 + name[i] - CHAR_0; + } + if (recno == 0) recno = RREF_ANY; + code[1+LINK_SIZE] = OP_RREF; /* Change test type */ + PUT2(code, 2+LINK_SIZE, recno); + } + + /* Similarly, check for the (?(DEFINE) "condition", which is always + false. */ + + else if (namelen == 6 && STRNCMP_UC_C8(name, STRING_DEFINE, 6) == 0) + { + code[1+LINK_SIZE] = OP_DEF; + skipbytes = 1; + } + + /* Reference to an unidentified subpattern. */ + + else + { + *errorcodeptr = ERR15; + goto FAILED; + } + break; + + + /* ------------------------------------------------------------ */ + case CHAR_EQUALS_SIGN: /* Positive lookahead */ + bravalue = OP_ASSERT; + cd->assert_depth += 1; + ptr++; + break; + + /* Optimize (?!) to (*FAIL) unless it is quantified - which is a weird + thing to do, but Perl allows all assertions to be quantified, and when + they contain capturing parentheses there may be a potential use for + this feature. Not that that applies to a quantified (?!) but we allow + it for uniformity. */ + + /* ------------------------------------------------------------ */ + case CHAR_EXCLAMATION_MARK: /* Negative lookahead */ + ptr++; + if (*ptr == CHAR_RIGHT_PARENTHESIS && ptr[1] != CHAR_ASTERISK && + ptr[1] != CHAR_PLUS && ptr[1] != CHAR_QUESTION_MARK && + (ptr[1] != CHAR_LEFT_CURLY_BRACKET || !is_counted_repeat(ptr+2))) + { + *code++ = OP_FAIL; + previous = NULL; + continue; + } + bravalue = OP_ASSERT_NOT; + cd->assert_depth += 1; + break; + + + /* ------------------------------------------------------------ */ + case CHAR_LESS_THAN_SIGN: /* Lookbehind or named define */ + switch (ptr[1]) + { + case CHAR_EQUALS_SIGN: /* Positive lookbehind */ + bravalue = OP_ASSERTBACK; + cd->assert_depth += 1; + ptr += 2; + break; + + case CHAR_EXCLAMATION_MARK: /* Negative lookbehind */ + bravalue = OP_ASSERTBACK_NOT; + cd->assert_depth += 1; + ptr += 2; + break; + + default: /* Could be name define, else bad */ + if (MAX_255(ptr[1]) && (cd->ctypes[ptr[1]] & ctype_word) != 0) + goto DEFINE_NAME; + ptr++; /* Correct offset for error */ + *errorcodeptr = ERR24; + goto FAILED; + } + break; + + + /* ------------------------------------------------------------ */ + case CHAR_GREATER_THAN_SIGN: /* One-time brackets */ + bravalue = OP_ONCE; + ptr++; + break; + + + /* ------------------------------------------------------------ */ + case CHAR_C: /* Callout - may be followed by digits; */ + previous_callout = code; /* Save for later completion */ + after_manual_callout = 1; /* Skip one item before completing */ + *code++ = OP_CALLOUT; + { + int n = 0; + ptr++; + while(IS_DIGIT(*ptr)) + n = n * 10 + *ptr++ - CHAR_0; + if (*ptr != CHAR_RIGHT_PARENTHESIS) + { + *errorcodeptr = ERR39; + goto FAILED; + } + if (n > 255) + { + *errorcodeptr = ERR38; + goto FAILED; + } + *code++ = n; + PUT(code, 0, (int)(ptr - cd->start_pattern + 1)); /* Pattern offset */ + PUT(code, LINK_SIZE, 0); /* Default length */ + code += 2 * LINK_SIZE; + } + previous = NULL; + continue; + + + /* ------------------------------------------------------------ */ + case CHAR_P: /* Python-style named subpattern handling */ + if (*(++ptr) == CHAR_EQUALS_SIGN || + *ptr == CHAR_GREATER_THAN_SIGN) /* Reference or recursion */ + { + is_recurse = *ptr == CHAR_GREATER_THAN_SIGN; + terminator = CHAR_RIGHT_PARENTHESIS; + goto NAMED_REF_OR_RECURSE; + } + else if (*ptr != CHAR_LESS_THAN_SIGN) /* Test for Python-style defn */ + { + *errorcodeptr = ERR41; + goto FAILED; + } + /* Fall through to handle (?P< as (?< is handled */ + + + /* ------------------------------------------------------------ */ + DEFINE_NAME: /* Come here from (?< handling */ + case CHAR_APOSTROPHE: + terminator = (*ptr == CHAR_LESS_THAN_SIGN)? + CHAR_GREATER_THAN_SIGN : CHAR_APOSTROPHE; + name = ++ptr; + if (IS_DIGIT(*ptr)) + { + *errorcodeptr = ERR84; /* Group name must start with non-digit */ + goto FAILED; + } + while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0) ptr++; + namelen = (int)(ptr - name); + + /* In the pre-compile phase, do a syntax check, remember the longest + name, and then remember the group in a vector, expanding it if + necessary. Duplicates for the same number are skipped; other duplicates + are checked for validity. In the actual compile, there is nothing to + do. */ + + if (lengthptr != NULL) + { + named_group *ng; + pcre_uint32 number = cd->bracount + 1; + + if (*ptr != (pcre_uchar)terminator) + { + *errorcodeptr = ERR42; + goto FAILED; + } + + if (cd->names_found >= MAX_NAME_COUNT) + { + *errorcodeptr = ERR49; + goto FAILED; + } + + if (namelen + IMM2_SIZE + 1 > cd->name_entry_size) + { + cd->name_entry_size = namelen + IMM2_SIZE + 1; + if (namelen > MAX_NAME_SIZE) + { + *errorcodeptr = ERR48; + goto FAILED; + } + } + + /* Scan the list to check for duplicates. For duplicate names, if the + number is the same, break the loop, which causes the name to be + discarded; otherwise, if DUPNAMES is not set, give an error. + If it is set, allow the name with a different number, but continue + scanning in case this is a duplicate with the same number. For + non-duplicate names, give an error if the number is duplicated. */ + + ng = cd->named_groups; + for (i = 0; i < cd->names_found; i++, ng++) + { + if (namelen == ng->length && + STRNCMP_UC_UC(name, ng->name, namelen) == 0) + { + if (ng->number == number) break; + if ((options & PCRE_DUPNAMES) == 0) + { + *errorcodeptr = ERR43; + goto FAILED; + } + cd->dupnames = TRUE; /* Duplicate names exist */ + } + else if (ng->number == number) + { + *errorcodeptr = ERR65; + goto FAILED; + } + } + + if (i >= cd->names_found) /* Not a duplicate with same number */ + { + /* Increase the list size if necessary */ + + if (cd->names_found >= cd->named_group_list_size) + { + int newsize = cd->named_group_list_size * 2; + named_group *newspace = (PUBL(malloc)) + (newsize * sizeof(named_group)); + + if (newspace == NULL) + { + *errorcodeptr = ERR21; + goto FAILED; + } + + memcpy(newspace, cd->named_groups, + cd->named_group_list_size * sizeof(named_group)); + if (cd->named_group_list_size > NAMED_GROUP_LIST_SIZE) + (PUBL(free))((void *)cd->named_groups); + cd->named_groups = newspace; + cd->named_group_list_size = newsize; + } + + cd->named_groups[cd->names_found].name = name; + cd->named_groups[cd->names_found].length = namelen; + cd->named_groups[cd->names_found].number = number; + cd->names_found++; + } + } + + ptr++; /* Move past > or ' in both passes. */ + goto NUMBERED_GROUP; + + + /* ------------------------------------------------------------ */ + case CHAR_AMPERSAND: /* Perl recursion/subroutine syntax */ + terminator = CHAR_RIGHT_PARENTHESIS; + is_recurse = TRUE; + /* Fall through */ + + /* We come here from the Python syntax above that handles both + references (?P=name) and recursion (?P>name), as well as falling + through from the Perl recursion syntax (?&name). We also come here from + the Perl \k or \k'name' back reference syntax and the \k{name} + .NET syntax, and the Oniguruma \g<...> and \g'...' subroutine syntax. */ + + NAMED_REF_OR_RECURSE: + name = ++ptr; + if (IS_DIGIT(*ptr)) + { + *errorcodeptr = ERR84; /* Group name must start with non-digit */ + goto FAILED; + } + while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0) ptr++; + namelen = (int)(ptr - name); + + /* In the pre-compile phase, do a syntax check. We used to just set + a dummy reference number, because it was not used in the first pass. + However, with the change of recursive back references to be atomic, + we have to look for the number so that this state can be identified, as + otherwise the incorrect length is computed. If it's not a backwards + reference, the dummy number will do. */ + + if (lengthptr != NULL) + { + named_group *ng; + recno = 0; + + if (namelen == 0) + { + *errorcodeptr = ERR62; + goto FAILED; + } + if (*ptr != (pcre_uchar)terminator) + { + *errorcodeptr = ERR42; + goto FAILED; + } + if (namelen > MAX_NAME_SIZE) + { + *errorcodeptr = ERR48; + goto FAILED; + } + + /* Count named back references. */ + + if (!is_recurse) cd->namedrefcount++; + + /* We have to allow for a named reference to a duplicated name (this + cannot be determined until the second pass). This needs an extra + 16-bit data item. */ + + *lengthptr += IMM2_SIZE; + + /* If this is a forward reference and we are within a (?|...) group, + the reference may end up as the number of a group which we are + currently inside, that is, it could be a recursive reference. In the + real compile this will be picked up and the reference wrapped with + OP_ONCE to make it atomic, so we must space in case this occurs. */ + + /* In fact, this can happen for a non-forward reference because + another group with the same number might be created later. This + issue is fixed "properly" in PCRE2. As PCRE1 is now in maintenance + only mode, we finesse the bug by allowing more memory always. */ + + *lengthptr += 2 + 2*LINK_SIZE; + + /* It is even worse than that. The current reference may be to an + existing named group with a different number (so apparently not + recursive) but which later on is also attached to a group with the + current number. This can only happen if $(| has been previous + encountered. In that case, we allow yet more memory, just in case. + (Again, this is fixed "properly" in PCRE2. */ + + if (cd->dupgroups) *lengthptr += 4 + 4*LINK_SIZE; + + /* Otherwise, check for recursion here. The name table does not exist + in the first pass; instead we must scan the list of names encountered + so far in order to get the number. If the name is not found, leave + the value of recno as 0 for a forward reference. */ + + else + { + ng = cd->named_groups; + for (i = 0; i < cd->names_found; i++, ng++) + { + if (namelen == ng->length && + STRNCMP_UC_UC(name, ng->name, namelen) == 0) + { + open_capitem *oc; + recno = ng->number; + if (is_recurse) break; + for (oc = cd->open_caps; oc != NULL; oc = oc->next) + { + if (oc->number == recno) + { + oc->flag = TRUE; + break; + } + } + } + } + } + } + + /* In the real compile, search the name table. We check the name + first, and then check that we have reached the end of the name in the + table. That way, if the name is longer than any in the table, the + comparison will fail without reading beyond the table entry. */ + + else + { + slot = cd->name_table; + for (i = 0; i < cd->names_found; i++) + { + if (STRNCMP_UC_UC(name, slot+IMM2_SIZE, namelen) == 0 && + slot[IMM2_SIZE+namelen] == 0) + break; + slot += cd->name_entry_size; + } + + if (i < cd->names_found) + { + recno = GET2(slot, 0); + } + else + { + *errorcodeptr = ERR15; + goto FAILED; + } + } + + /* In both phases, for recursions, we can now go to the code than + handles numerical recursion. */ + + if (is_recurse) goto HANDLE_RECURSION; + + /* In the second pass we must see if the name is duplicated. If so, we + generate a different opcode. */ + + if (lengthptr == NULL && cd->dupnames) + { + int count = 1; + unsigned int index = i; + pcre_uchar *cslot = slot + cd->name_entry_size; + + for (i++; i < cd->names_found; i++) + { + if (STRCMP_UC_UC(slot + IMM2_SIZE, cslot + IMM2_SIZE) != 0) break; + count++; + cslot += cd->name_entry_size; + } + + if (count > 1) + { + if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE; + previous = code; + item_hwm_offset = cd->hwm - cd->start_workspace; + *code++ = ((options & PCRE_CASELESS) != 0)? OP_DNREFI : OP_DNREF; + PUT2INC(code, 0, index); + PUT2INC(code, 0, count); + + /* Process each potentially referenced group. */ + + for (; slot < cslot; slot += cd->name_entry_size) + { + open_capitem *oc; + recno = GET2(slot, 0); + cd->backref_map |= (recno < 32)? (1 << recno) : 1; + if (recno > cd->top_backref) cd->top_backref = recno; + + /* Check to see if this back reference is recursive, that it, it + is inside the group that it references. A flag is set so that the + group can be made atomic. */ + + for (oc = cd->open_caps; oc != NULL; oc = oc->next) + { + if (oc->number == recno) + { + oc->flag = TRUE; + break; + } + } + } + + continue; /* End of back ref handling */ + } + } + + /* First pass, or a non-duplicated name. */ + + goto HANDLE_REFERENCE; + + + /* ------------------------------------------------------------ */ + case CHAR_R: /* Recursion, same as (?0) */ + recno = 0; + if (*(++ptr) != CHAR_RIGHT_PARENTHESIS) + { + *errorcodeptr = ERR29; + goto FAILED; + } + goto HANDLE_RECURSION; + + + /* ------------------------------------------------------------ */ + case CHAR_MINUS: case CHAR_PLUS: /* Recursion or subroutine */ + case CHAR_0: case CHAR_1: case CHAR_2: case CHAR_3: case CHAR_4: + case CHAR_5: case CHAR_6: case CHAR_7: case CHAR_8: case CHAR_9: + { + const pcre_uchar *called; + terminator = CHAR_RIGHT_PARENTHESIS; + + /* Come here from the \g<...> and \g'...' code (Oniguruma + compatibility). However, the syntax has been checked to ensure that + the ... are a (signed) number, so that neither ERR63 nor ERR29 will + be called on this path, nor with the jump to OTHER_CHAR_AFTER_QUERY + ever be taken. */ + + HANDLE_NUMERICAL_RECURSION: + + if ((refsign = *ptr) == CHAR_PLUS) + { + ptr++; + if (!IS_DIGIT(*ptr)) + { + *errorcodeptr = ERR63; + goto FAILED; + } + } + else if (refsign == CHAR_MINUS) + { + if (!IS_DIGIT(ptr[1])) + goto OTHER_CHAR_AFTER_QUERY; + ptr++; + } + + recno = 0; + while(IS_DIGIT(*ptr)) + { + if (recno > INT_MAX / 10 - 1) /* Integer overflow */ + { + while (IS_DIGIT(*ptr)) ptr++; + *errorcodeptr = ERR61; + goto FAILED; + } + recno = recno * 10 + *ptr++ - CHAR_0; + } + + if (*ptr != (pcre_uchar)terminator) + { + *errorcodeptr = ERR29; + goto FAILED; + } + + if (refsign == CHAR_MINUS) + { + if (recno == 0) + { + *errorcodeptr = ERR58; + goto FAILED; + } + recno = cd->bracount - recno + 1; + if (recno <= 0) + { + *errorcodeptr = ERR15; + goto FAILED; + } + } + else if (refsign == CHAR_PLUS) + { + if (recno == 0) + { + *errorcodeptr = ERR58; + goto FAILED; + } + recno += cd->bracount; + } + + /* Come here from code above that handles a named recursion */ + + HANDLE_RECURSION: + + previous = code; + item_hwm_offset = cd->hwm - cd->start_workspace; + called = cd->start_code; + + /* When we are actually compiling, find the bracket that is being + referenced. Temporarily end the regex in case it doesn't exist before + this point. If we end up with a forward reference, first check that + the bracket does occur later so we can give the error (and position) + now. Then remember this forward reference in the workspace so it can + be filled in at the end. */ + + if (lengthptr == NULL) + { + *code = OP_END; + if (recno != 0) + called = PRIV(find_bracket)(cd->start_code, utf, recno); + + /* Forward reference */ + + if (called == NULL) + { + if (recno > cd->final_bracount) + { + *errorcodeptr = ERR15; + goto FAILED; + } + + /* Fudge the value of "called" so that when it is inserted as an + offset below, what it actually inserted is the reference number + of the group. Then remember the forward reference. */ + + called = cd->start_code + recno; + if (cd->hwm >= cd->start_workspace + cd->workspace_size - + WORK_SIZE_SAFETY_MARGIN) + { + *errorcodeptr = expand_workspace(cd); + if (*errorcodeptr != 0) goto FAILED; + } + PUTINC(cd->hwm, 0, (int)(code + 1 - cd->start_code)); + } + + /* If not a forward reference, and the subpattern is still open, + this is a recursive call. We check to see if this is a left + recursion that could loop for ever, and diagnose that case. We + must not, however, do this check if we are in a conditional + subpattern because the condition might be testing for recursion in + a pattern such as /(?(R)a+|(?R)b)/, which is perfectly valid. + Forever loops are also detected at runtime, so those that occur in + conditional subpatterns will be picked up then. */ + + else if (GET(called, 1) == 0 && cond_depth <= 0 && + could_be_empty(called, code, bcptr, utf, cd)) + { + *errorcodeptr = ERR40; + goto FAILED; + } + } + + /* Insert the recursion/subroutine item. It does not have a set first + character (relevant if it is repeated, because it will then be + wrapped with ONCE brackets). */ + + *code = OP_RECURSE; + PUT(code, 1, (int)(called - cd->start_code)); + code += 1 + LINK_SIZE; + groupsetfirstchar = FALSE; + } + + /* Can't determine a first byte now */ + + if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE; + continue; + + + /* ------------------------------------------------------------ */ + default: /* Other characters: check option setting */ + OTHER_CHAR_AFTER_QUERY: + set = unset = 0; + optset = &set; + + while (*ptr != CHAR_RIGHT_PARENTHESIS && *ptr != CHAR_COLON) + { + switch (*ptr++) + { + case CHAR_MINUS: optset = &unset; break; + + case CHAR_J: /* Record that it changed in the external options */ + *optset |= PCRE_DUPNAMES; + cd->external_flags |= PCRE_JCHANGED; + break; + + case CHAR_i: *optset |= PCRE_CASELESS; break; + case CHAR_m: *optset |= PCRE_MULTILINE; break; + case CHAR_s: *optset |= PCRE_DOTALL; break; + case CHAR_x: *optset |= PCRE_EXTENDED; break; + case CHAR_U: *optset |= PCRE_UNGREEDY; break; + case CHAR_X: *optset |= PCRE_EXTRA; break; + + default: *errorcodeptr = ERR12; + ptr--; /* Correct the offset */ + goto FAILED; + } + } + + /* Set up the changed option bits, but don't change anything yet. */ + + newoptions = (options | set) & (~unset); + + /* If the options ended with ')' this is not the start of a nested + group with option changes, so the options change at this level. If this + item is right at the start of the pattern, the options can be + abstracted and made external in the pre-compile phase, and ignored in + the compile phase. This can be helpful when matching -- for instance in + caseless checking of required bytes. + + If the code pointer is not (cd->start_code + 1 + LINK_SIZE), we are + definitely *not* at the start of the pattern because something has been + compiled. In the pre-compile phase, however, the code pointer can have + that value after the start, because it gets reset as code is discarded + during the pre-compile. However, this can happen only at top level - if + we are within parentheses, the starting BRA will still be present. At + any parenthesis level, the length value can be used to test if anything + has been compiled at that level. Thus, a test for both these conditions + is necessary to ensure we correctly detect the start of the pattern in + both phases. + + If we are not at the pattern start, reset the greedy defaults and the + case value for firstchar and reqchar. */ + + if (*ptr == CHAR_RIGHT_PARENTHESIS) + { + if (code == cd->start_code + 1 + LINK_SIZE && + (lengthptr == NULL || *lengthptr == 2 + 2*LINK_SIZE)) + { + cd->external_options = newoptions; + } + else + { + greedy_default = ((newoptions & PCRE_UNGREEDY) != 0); + greedy_non_default = greedy_default ^ 1; + req_caseopt = ((newoptions & PCRE_CASELESS) != 0)? REQ_CASELESS:0; + } + + /* Change options at this level, and pass them back for use + in subsequent branches. */ + + *optionsptr = options = newoptions; + previous = NULL; /* This item can't be repeated */ + continue; /* It is complete */ + } + + /* If the options ended with ':' we are heading into a nested group + with possible change of options. Such groups are non-capturing and are + not assertions of any kind. All we need to do is skip over the ':'; + the newoptions value is handled below. */ + + bravalue = OP_BRA; + ptr++; + } /* End of switch for character following (? */ + } /* End of (? handling */ + + /* Opening parenthesis not followed by '*' or '?'. If PCRE_NO_AUTO_CAPTURE + is set, all unadorned brackets become non-capturing and behave like (?:...) + brackets. */ + + else if ((options & PCRE_NO_AUTO_CAPTURE) != 0) + { + bravalue = OP_BRA; + } + + /* Else we have a capturing group. */ + + else + { + NUMBERED_GROUP: + cd->bracount += 1; + PUT2(code, 1+LINK_SIZE, cd->bracount); + skipbytes = IMM2_SIZE; + } + + /* Process nested bracketed regex. First check for parentheses nested too + deeply. */ + + if ((cd->parens_depth += 1) > PARENS_NEST_LIMIT) + { + *errorcodeptr = ERR82; + goto FAILED; + } + + /* All assertions used not to be repeatable, but this was changed for Perl + compatibility. All kinds can now be repeated except for assertions that are + conditions (Perl also forbids these to be repeated). We copy code into a + non-register variable (tempcode) in order to be able to pass its address + because some compilers complain otherwise. At the start of a conditional + group whose condition is an assertion, cd->iscondassert is set. We unset it + here so as to allow assertions later in the group to be quantified. */ + + if (bravalue >= OP_ASSERT && bravalue <= OP_ASSERTBACK_NOT && + cd->iscondassert) + { + previous = NULL; + cd->iscondassert = FALSE; + } + else + { + previous = code; + item_hwm_offset = cd->hwm - cd->start_workspace; + } + + *code = bravalue; + tempcode = code; + tempreqvary = cd->req_varyopt; /* Save value before bracket */ + tempbracount = cd->bracount; /* Save value before bracket */ + length_prevgroup = 0; /* Initialize for pre-compile phase */ + + if (!compile_regex( + newoptions, /* The complete new option state */ + &tempcode, /* Where to put code (updated) */ + &ptr, /* Input pointer (updated) */ + errorcodeptr, /* Where to put an error message */ + (bravalue == OP_ASSERTBACK || + bravalue == OP_ASSERTBACK_NOT), /* TRUE if back assert */ + reset_bracount, /* True if (?| group */ + skipbytes, /* Skip over bracket number */ + cond_depth + + ((bravalue == OP_COND)?1:0), /* Depth of condition subpatterns */ + &subfirstchar, /* For possible first char */ + &subfirstcharflags, + &subreqchar, /* For possible last char */ + &subreqcharflags, + bcptr, /* Current branch chain */ + cd, /* Tables block */ + (lengthptr == NULL)? NULL : /* Actual compile phase */ + &length_prevgroup /* Pre-compile phase */ + )) + goto FAILED; + + cd->parens_depth -= 1; + + /* If this was an atomic group and there are no capturing groups within it, + generate OP_ONCE_NC instead of OP_ONCE. */ + + if (bravalue == OP_ONCE && cd->bracount <= tempbracount) + *code = OP_ONCE_NC; + + if (bravalue >= OP_ASSERT && bravalue <= OP_ASSERTBACK_NOT) + cd->assert_depth -= 1; + + /* At the end of compiling, code is still pointing to the start of the + group, while tempcode has been updated to point past the end of the group. + The pattern pointer (ptr) is on the bracket. + + If this is a conditional bracket, check that there are no more than + two branches in the group, or just one if it's a DEFINE group. We do this + in the real compile phase, not in the pre-pass, where the whole group may + not be available. */ + + if (bravalue == OP_COND && lengthptr == NULL) + { + pcre_uchar *tc = code; + int condcount = 0; + + do { + condcount++; + tc += GET(tc,1); + } + while (*tc != OP_KET); + + /* A DEFINE group is never obeyed inline (the "condition" is always + false). It must have only one branch. */ + + if (code[LINK_SIZE+1] == OP_DEF) + { + if (condcount > 1) + { + *errorcodeptr = ERR54; + goto FAILED; + } + bravalue = OP_DEF; /* Just a flag to suppress char handling below */ + } + + /* A "normal" conditional group. If there is just one branch, we must not + make use of its firstchar or reqchar, because this is equivalent to an + empty second branch. */ + + else + { + if (condcount > 2) + { + *errorcodeptr = ERR27; + goto FAILED; + } + if (condcount == 1) subfirstcharflags = subreqcharflags = REQ_NONE; + } + } + + /* Error if hit end of pattern */ + + if (*ptr != CHAR_RIGHT_PARENTHESIS) + { + *errorcodeptr = ERR14; + goto FAILED; + } + + /* In the pre-compile phase, update the length by the length of the group, + less the brackets at either end. Then reduce the compiled code to just a + set of non-capturing brackets so that it doesn't use much memory if it is + duplicated by a quantifier.*/ + + if (lengthptr != NULL) + { + if (OFLOW_MAX - *lengthptr < length_prevgroup - 2 - 2*LINK_SIZE) + { + *errorcodeptr = ERR20; + goto FAILED; + } + *lengthptr += length_prevgroup - 2 - 2*LINK_SIZE; + code++; /* This already contains bravalue */ + PUTINC(code, 0, 1 + LINK_SIZE); + *code++ = OP_KET; + PUTINC(code, 0, 1 + LINK_SIZE); + break; /* No need to waste time with special character handling */ + } + + /* Otherwise update the main code pointer to the end of the group. */ + + code = tempcode; + + /* For a DEFINE group, required and first character settings are not + relevant. */ + + if (bravalue == OP_DEF) break; + + /* Handle updating of the required and first characters for other types of + group. Update for normal brackets of all kinds, and conditions with two + branches (see code above). If the bracket is followed by a quantifier with + zero repeat, we have to back off. Hence the definition of zeroreqchar and + zerofirstchar outside the main loop so that they can be accessed for the + back off. */ + + zeroreqchar = reqchar; + zeroreqcharflags = reqcharflags; + zerofirstchar = firstchar; + zerofirstcharflags = firstcharflags; + groupsetfirstchar = FALSE; + + if (bravalue >= OP_ONCE) + { + /* If we have not yet set a firstchar in this branch, take it from the + subpattern, remembering that it was set here so that a repeat of more + than one can replicate it as reqchar if necessary. If the subpattern has + no firstchar, set "none" for the whole branch. In both cases, a zero + repeat forces firstchar to "none". */ + + if (firstcharflags == REQ_UNSET) + { + if (subfirstcharflags >= 0) + { + firstchar = subfirstchar; + firstcharflags = subfirstcharflags; + groupsetfirstchar = TRUE; + } + else firstcharflags = REQ_NONE; + zerofirstcharflags = REQ_NONE; + } + + /* If firstchar was previously set, convert the subpattern's firstchar + into reqchar if there wasn't one, using the vary flag that was in + existence beforehand. */ + + else if (subfirstcharflags >= 0 && subreqcharflags < 0) + { + subreqchar = subfirstchar; + subreqcharflags = subfirstcharflags | tempreqvary; + } + + /* If the subpattern set a required byte (or set a first byte that isn't + really the first byte - see above), set it. */ + + if (subreqcharflags >= 0) + { + reqchar = subreqchar; + reqcharflags = subreqcharflags; + } + } + + /* For a forward assertion, we take the reqchar, if set. This can be + helpful if the pattern that follows the assertion doesn't set a different + char. For example, it's useful for /(?=abcde).+/. We can't set firstchar + for an assertion, however because it leads to incorrect effect for patterns + such as /(?=a)a.+/ when the "real" "a" would then become a reqchar instead + of a firstchar. This is overcome by a scan at the end if there's no + firstchar, looking for an asserted first char. */ + + else if (bravalue == OP_ASSERT && subreqcharflags >= 0) + { + reqchar = subreqchar; + reqcharflags = subreqcharflags; + } + break; /* End of processing '(' */ + + + /* ===================================================================*/ + /* Handle metasequences introduced by \. For ones like \d, the ESC_ values + are arranged to be the negation of the corresponding OP_values in the + default case when PCRE_UCP is not set. For the back references, the values + are negative the reference number. Only back references and those types + that consume a character may be repeated. We can test for values between + ESC_b and ESC_Z for the latter; this may have to change if any new ones are + ever created. */ + + case CHAR_BACKSLASH: + tempptr = ptr; + escape = check_escape(&ptr, &ec, errorcodeptr, cd->bracount, options, FALSE); + if (*errorcodeptr != 0) goto FAILED; + + if (escape == 0) /* The escape coded a single character */ + c = ec; + else + { + if (escape == ESC_Q) /* Handle start of quoted string */ + { + if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E) + ptr += 2; /* avoid empty string */ + else inescq = TRUE; + continue; + } + + if (escape == ESC_E) continue; /* Perl ignores an orphan \E */ + + /* For metasequences that actually match a character, we disable the + setting of a first character if it hasn't already been set. */ + + if (firstcharflags == REQ_UNSET && escape > ESC_b && escape < ESC_Z) + firstcharflags = REQ_NONE; + + /* Set values to reset to if this is followed by a zero repeat. */ + + zerofirstchar = firstchar; + zerofirstcharflags = firstcharflags; + zeroreqchar = reqchar; + zeroreqcharflags = reqcharflags; + + /* \g or \g'name' is a subroutine call by name and \g or \g'n' + is a subroutine call by number (Oniguruma syntax). In fact, the value + ESC_g is returned only for these cases. So we don't need to check for < + or ' if the value is ESC_g. For the Perl syntax \g{n} the value is + -n, and for the Perl syntax \g{name} the result is ESC_k (as + that is a synonym for a named back reference). */ + + if (escape == ESC_g) + { + const pcre_uchar *p; + pcre_uint32 cf; + + item_hwm_offset = cd->hwm - cd->start_workspace; /* Normally this is set when '(' is read */ + terminator = (*(++ptr) == CHAR_LESS_THAN_SIGN)? + CHAR_GREATER_THAN_SIGN : CHAR_APOSTROPHE; + + /* These two statements stop the compiler for warning about possibly + unset variables caused by the jump to HANDLE_NUMERICAL_RECURSION. In + fact, because we do the check for a number below, the paths that + would actually be in error are never taken. */ + + skipbytes = 0; + reset_bracount = FALSE; + + /* If it's not a signed or unsigned number, treat it as a name. */ + + cf = ptr[1]; + if (cf != CHAR_PLUS && cf != CHAR_MINUS && !IS_DIGIT(cf)) + { + is_recurse = TRUE; + goto NAMED_REF_OR_RECURSE; + } + + /* Signed or unsigned number (cf = ptr[1]) is known to be plus or minus + or a digit. */ + + p = ptr + 2; + while (IS_DIGIT(*p)) p++; + if (*p != (pcre_uchar)terminator) + { + *errorcodeptr = ERR57; + goto FAILED; + } + ptr++; + goto HANDLE_NUMERICAL_RECURSION; + } + + /* \k or \k'name' is a back reference by name (Perl syntax). + We also support \k{name} (.NET syntax). */ + + if (escape == ESC_k) + { + if ((ptr[1] != CHAR_LESS_THAN_SIGN && + ptr[1] != CHAR_APOSTROPHE && ptr[1] != CHAR_LEFT_CURLY_BRACKET)) + { + *errorcodeptr = ERR69; + goto FAILED; + } + is_recurse = FALSE; + terminator = (*(++ptr) == CHAR_LESS_THAN_SIGN)? + CHAR_GREATER_THAN_SIGN : (*ptr == CHAR_APOSTROPHE)? + CHAR_APOSTROPHE : CHAR_RIGHT_CURLY_BRACKET; + goto NAMED_REF_OR_RECURSE; + } + + /* Back references are handled specially; must disable firstchar if + not set to cope with cases like (?=(\w+))\1: which would otherwise set + ':' later. */ + + if (escape < 0) + { + open_capitem *oc; + recno = -escape; + + /* Come here from named backref handling when the reference is to a + single group (i.e. not to a duplicated name. */ + + HANDLE_REFERENCE: + if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE; + previous = code; + item_hwm_offset = cd->hwm - cd->start_workspace; + *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF; + PUT2INC(code, 0, recno); + cd->backref_map |= (recno < 32)? (1 << recno) : 1; + if (recno > cd->top_backref) cd->top_backref = recno; + + /* Check to see if this back reference is recursive, that it, it + is inside the group that it references. A flag is set so that the + group can be made atomic. */ + + for (oc = cd->open_caps; oc != NULL; oc = oc->next) + { + if (oc->number == recno) + { + oc->flag = TRUE; + break; + } + } + } + + /* So are Unicode property matches, if supported. */ + +#ifdef SUPPORT_UCP + else if (escape == ESC_P || escape == ESC_p) + { + BOOL negated; + unsigned int ptype = 0, pdata = 0; + if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr)) + goto FAILED; + previous = code; + item_hwm_offset = cd->hwm - cd->start_workspace; + *code++ = ((escape == ESC_p) != negated)? OP_PROP : OP_NOTPROP; + *code++ = ptype; + *code++ = pdata; + } +#else + + /* If Unicode properties are not supported, \X, \P, and \p are not + allowed. */ + + else if (escape == ESC_X || escape == ESC_P || escape == ESC_p) + { + *errorcodeptr = ERR45; + goto FAILED; + } +#endif + + /* For the rest (including \X when Unicode properties are supported), we + can obtain the OP value by negating the escape value in the default + situation when PCRE_UCP is not set. When it *is* set, we substitute + Unicode property tests. Note that \b and \B do a one-character + lookbehind, and \A also behaves as if it does. */ + + else + { + if ((escape == ESC_b || escape == ESC_B || escape == ESC_A) && + cd->max_lookbehind == 0) + cd->max_lookbehind = 1; +#ifdef SUPPORT_UCP + if (escape >= ESC_DU && escape <= ESC_wu) + { + nestptr = ptr + 1; /* Where to resume */ + ptr = substitutes[escape - ESC_DU] - 1; /* Just before substitute */ + } + else +#endif + /* In non-UTF-8 mode, we turn \C into OP_ALLANY instead of OP_ANYBYTE + so that it works in DFA mode and in lookbehinds. */ + + { + previous = (escape > ESC_b && escape < ESC_Z)? code : NULL; + item_hwm_offset = cd->hwm - cd->start_workspace; + *code++ = (!utf && escape == ESC_C)? OP_ALLANY : escape; + } + } + continue; + } + + /* We have a data character whose value is in c. In UTF-8 mode it may have + a value > 127. We set its representation in the length/buffer, and then + handle it as a data character. */ + +#if defined SUPPORT_UTF && !defined COMPILE_PCRE32 + if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR) + mclength = PRIV(ord2utf)(c, mcbuffer); + else +#endif + + { + mcbuffer[0] = c; + mclength = 1; + } + goto ONE_CHAR; + + + /* ===================================================================*/ + /* Handle a literal character. It is guaranteed not to be whitespace or # + when the extended flag is set. If we are in a UTF mode, it may be a + multi-unit literal character. */ + + default: + NORMAL_CHAR: + mclength = 1; + mcbuffer[0] = c; + +#ifdef SUPPORT_UTF + if (utf && HAS_EXTRALEN(c)) + ACROSSCHAR(TRUE, ptr[1], mcbuffer[mclength++] = *(++ptr)); +#endif + + /* At this point we have the character's bytes in mcbuffer, and the length + in mclength. When not in UTF-8 mode, the length is always 1. */ + + ONE_CHAR: + previous = code; + item_hwm_offset = cd->hwm - cd->start_workspace; + + /* For caseless UTF-8 mode when UCP support is available, check whether + this character has more than one other case. If so, generate a special + OP_PROP item instead of OP_CHARI. */ + +#ifdef SUPPORT_UCP + if (utf && (options & PCRE_CASELESS) != 0) + { + GETCHAR(c, mcbuffer); + if ((c = UCD_CASESET(c)) != 0) + { + *code++ = OP_PROP; + *code++ = PT_CLIST; + *code++ = c; + if (firstcharflags == REQ_UNSET) + firstcharflags = zerofirstcharflags = REQ_NONE; + break; + } + } +#endif + + /* Caseful matches, or not one of the multicase characters. */ + + *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARI : OP_CHAR; + for (c = 0; c < mclength; c++) *code++ = mcbuffer[c]; + + /* Remember if \r or \n were seen */ + + if (mcbuffer[0] == CHAR_CR || mcbuffer[0] == CHAR_NL) + cd->external_flags |= PCRE_HASCRORLF; + + /* Set the first and required bytes appropriately. If no previous first + byte, set it from this character, but revert to none on a zero repeat. + Otherwise, leave the firstchar value alone, and don't change it on a zero + repeat. */ + + if (firstcharflags == REQ_UNSET) + { + zerofirstcharflags = REQ_NONE; + zeroreqchar = reqchar; + zeroreqcharflags = reqcharflags; + + /* If the character is more than one byte long, we can set firstchar + only if it is not to be matched caselessly. */ + + if (mclength == 1 || req_caseopt == 0) + { + firstchar = mcbuffer[0] | req_caseopt; + firstchar = mcbuffer[0]; + firstcharflags = req_caseopt; + + if (mclength != 1) + { + reqchar = code[-1]; + reqcharflags = cd->req_varyopt; + } + } + else firstcharflags = reqcharflags = REQ_NONE; + } + + /* firstchar was previously set; we can set reqchar only if the length is + 1 or the matching is caseful. */ + + else + { + zerofirstchar = firstchar; + zerofirstcharflags = firstcharflags; + zeroreqchar = reqchar; + zeroreqcharflags = reqcharflags; + if (mclength == 1 || req_caseopt == 0) + { + reqchar = code[-1]; + reqcharflags = req_caseopt | cd->req_varyopt; + } + } + + break; /* End of literal character handling */ + } + } /* end of big loop */ + + +/* Control never reaches here by falling through, only by a goto for all the +error states. Pass back the position in the pattern so that it can be displayed +to the user for diagnosing the error. */ + +FAILED: +*ptrptr = ptr; +return FALSE; +} + + + +/************************************************* +* Compile sequence of alternatives * +*************************************************/ + +/* On entry, ptr is pointing past the bracket character, but on return it +points to the closing bracket, or vertical bar, or end of string. The code +variable is pointing at the byte into which the BRA operator has been stored. +This function is used during the pre-compile phase when we are trying to find +out the amount of memory needed, as well as during the real compile phase. The +value of lengthptr distinguishes the two phases. + +Arguments: + options option bits, including any changes for this subpattern + codeptr -> the address of the current code pointer + ptrptr -> the address of the current pattern pointer + errorcodeptr -> pointer to error code variable + lookbehind TRUE if this is a lookbehind assertion + reset_bracount TRUE to reset the count for each branch + skipbytes skip this many bytes at start (for brackets and OP_COND) + cond_depth depth of nesting for conditional subpatterns + firstcharptr place to put the first required character + firstcharflagsptr place to put the first character flags, or a negative number + reqcharptr place to put the last required character + reqcharflagsptr place to put the last required character flags, or a negative number + bcptr pointer to the chain of currently open branches + cd points to the data block with tables pointers etc. + lengthptr NULL during the real compile phase + points to length accumulator during pre-compile phase + +Returns: TRUE on success +*/ + +static BOOL +compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr, + int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes, + int cond_depth, + pcre_uint32 *firstcharptr, pcre_int32 *firstcharflagsptr, + pcre_uint32 *reqcharptr, pcre_int32 *reqcharflagsptr, + branch_chain *bcptr, compile_data *cd, int *lengthptr) +{ +const pcre_uchar *ptr = *ptrptr; +pcre_uchar *code = *codeptr; +pcre_uchar *last_branch = code; +pcre_uchar *start_bracket = code; +pcre_uchar *reverse_count = NULL; +open_capitem capitem; +int capnumber = 0; +pcre_uint32 firstchar, reqchar; +pcre_int32 firstcharflags, reqcharflags; +pcre_uint32 branchfirstchar, branchreqchar; +pcre_int32 branchfirstcharflags, branchreqcharflags; +int length; +unsigned int orig_bracount; +unsigned int max_bracount; +branch_chain bc; +size_t save_hwm_offset; + +/* If set, call the external function that checks for stack availability. */ + +if (PUBL(stack_guard) != NULL && PUBL(stack_guard)()) + { + *errorcodeptr= ERR85; + return FALSE; + } + +/* Miscellaneous initialization */ + +bc.outer = bcptr; +bc.current_branch = code; + +firstchar = reqchar = 0; +firstcharflags = reqcharflags = REQ_UNSET; + +save_hwm_offset = cd->hwm - cd->start_workspace; + +/* Accumulate the length for use in the pre-compile phase. Start with the +length of the BRA and KET and any extra bytes that are required at the +beginning. We accumulate in a local variable to save frequent testing of +lenthptr for NULL. We cannot do this by looking at the value of code at the +start and end of each alternative, because compiled items are discarded during +the pre-compile phase so that the work space is not exceeded. */ + +length = 2 + 2*LINK_SIZE + skipbytes; + +/* WARNING: If the above line is changed for any reason, you must also change +the code that abstracts option settings at the start of the pattern and makes +them global. It tests the value of length for (2 + 2*LINK_SIZE) in the +pre-compile phase to find out whether anything has yet been compiled or not. */ + +/* If this is a capturing subpattern, add to the chain of open capturing items +so that we can detect them if (*ACCEPT) is encountered. This is also used to +detect groups that contain recursive back references to themselves. Note that +only OP_CBRA need be tested here; changing this opcode to one of its variants, +e.g. OP_SCBRAPOS, happens later, after the group has been compiled. */ + +if (*code == OP_CBRA) + { + capnumber = GET2(code, 1 + LINK_SIZE); + capitem.number = capnumber; + capitem.next = cd->open_caps; + capitem.flag = FALSE; + cd->open_caps = &capitem; + } + +/* Offset is set zero to mark that this bracket is still open */ + +PUT(code, 1, 0); +code += 1 + LINK_SIZE + skipbytes; + +/* Loop for each alternative branch */ + +orig_bracount = max_bracount = cd->bracount; +for (;;) + { + /* For a (?| group, reset the capturing bracket count so that each branch + uses the same numbers. */ + + if (reset_bracount) cd->bracount = orig_bracount; + + /* Set up dummy OP_REVERSE if lookbehind assertion */ + + if (lookbehind) + { + *code++ = OP_REVERSE; + reverse_count = code; + PUTINC(code, 0, 0); + length += 1 + LINK_SIZE; + } + + /* Now compile the branch; in the pre-compile phase its length gets added + into the length. */ + + if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar, + &branchfirstcharflags, &branchreqchar, &branchreqcharflags, &bc, + cond_depth, cd, (lengthptr == NULL)? NULL : &length)) + { + *ptrptr = ptr; + return FALSE; + } + + /* Keep the highest bracket count in case (?| was used and some branch + has fewer than the rest. */ + + if (cd->bracount > max_bracount) max_bracount = cd->bracount; + + /* In the real compile phase, there is some post-processing to be done. */ + + if (lengthptr == NULL) + { + /* If this is the first branch, the firstchar and reqchar values for the + branch become the values for the regex. */ + + if (*last_branch != OP_ALT) + { + firstchar = branchfirstchar; + firstcharflags = branchfirstcharflags; + reqchar = branchreqchar; + reqcharflags = branchreqcharflags; + } + + /* If this is not the first branch, the first char and reqchar have to + match the values from all the previous branches, except that if the + previous value for reqchar didn't have REQ_VARY set, it can still match, + and we set REQ_VARY for the regex. */ + + else + { + /* If we previously had a firstchar, but it doesn't match the new branch, + we have to abandon the firstchar for the regex, but if there was + previously no reqchar, it takes on the value of the old firstchar. */ + + if (firstcharflags >= 0 && + (firstcharflags != branchfirstcharflags || firstchar != branchfirstchar)) + { + if (reqcharflags < 0) + { + reqchar = firstchar; + reqcharflags = firstcharflags; + } + firstcharflags = REQ_NONE; + } + + /* If we (now or from before) have no firstchar, a firstchar from the + branch becomes a reqchar if there isn't a branch reqchar. */ + + if (firstcharflags < 0 && branchfirstcharflags >= 0 && branchreqcharflags < 0) + { + branchreqchar = branchfirstchar; + branchreqcharflags = branchfirstcharflags; + } + + /* Now ensure that the reqchars match */ + + if (((reqcharflags & ~REQ_VARY) != (branchreqcharflags & ~REQ_VARY)) || + reqchar != branchreqchar) + reqcharflags = REQ_NONE; + else + { + reqchar = branchreqchar; + reqcharflags |= branchreqcharflags; /* To "or" REQ_VARY */ + } + } + + /* If lookbehind, check that this branch matches a fixed-length string, and + put the length into the OP_REVERSE item. Temporarily mark the end of the + branch with OP_END. If the branch contains OP_RECURSE, the result is -3 + because there may be forward references that we can't check here. Set a + flag to cause another lookbehind check at the end. Why not do it all at the + end? Because common, erroneous checks are picked up here and the offset of + the problem can be shown. */ + + if (lookbehind) + { + int fixed_length; + *code = OP_END; + fixed_length = find_fixedlength(last_branch, (options & PCRE_UTF8) != 0, + FALSE, cd, NULL); + DPRINTF(("fixed length = %d\n", fixed_length)); + if (fixed_length == -3) + { + cd->check_lookbehind = TRUE; + } + else if (fixed_length < 0) + { + *errorcodeptr = (fixed_length == -2)? ERR36 : + (fixed_length == -4)? ERR70: ERR25; + *ptrptr = ptr; + return FALSE; + } + else + { + if (fixed_length > cd->max_lookbehind) + cd->max_lookbehind = fixed_length; + PUT(reverse_count, 0, fixed_length); + } + } + } + + /* Reached end of expression, either ')' or end of pattern. In the real + compile phase, go back through the alternative branches and reverse the chain + of offsets, with the field in the BRA item now becoming an offset to the + first alternative. If there are no alternatives, it points to the end of the + group. The length in the terminating ket is always the length of the whole + bracketed item. Return leaving the pointer at the terminating char. */ + + if (*ptr != CHAR_VERTICAL_LINE) + { + if (lengthptr == NULL) + { + int branch_length = (int)(code - last_branch); + do + { + int prev_length = GET(last_branch, 1); + PUT(last_branch, 1, branch_length); + branch_length = prev_length; + last_branch -= branch_length; + } + while (branch_length > 0); + } + + /* Fill in the ket */ + + *code = OP_KET; + PUT(code, 1, (int)(code - start_bracket)); + code += 1 + LINK_SIZE; + + /* If it was a capturing subpattern, check to see if it contained any + recursive back references. If so, we must wrap it in atomic brackets. + Because we are moving code along, we must ensure that any pending recursive + references are updated. In any event, remove the block from the chain. */ + + if (capnumber > 0) + { + if (cd->open_caps->flag) + { + *code = OP_END; + adjust_recurse(start_bracket, 1 + LINK_SIZE, + (options & PCRE_UTF8) != 0, cd, save_hwm_offset); + memmove(start_bracket + 1 + LINK_SIZE, start_bracket, + IN_UCHARS(code - start_bracket)); + *start_bracket = OP_ONCE; + code += 1 + LINK_SIZE; + PUT(start_bracket, 1, (int)(code - start_bracket)); + *code = OP_KET; + PUT(code, 1, (int)(code - start_bracket)); + code += 1 + LINK_SIZE; + length += 2 + 2*LINK_SIZE; + } + cd->open_caps = cd->open_caps->next; + } + + /* Retain the highest bracket number, in case resetting was used. */ + + cd->bracount = max_bracount; + + /* Set values to pass back */ + + *codeptr = code; + *ptrptr = ptr; + *firstcharptr = firstchar; + *firstcharflagsptr = firstcharflags; + *reqcharptr = reqchar; + *reqcharflagsptr = reqcharflags; + if (lengthptr != NULL) + { + if (OFLOW_MAX - *lengthptr < length) + { + *errorcodeptr = ERR20; + return FALSE; + } + *lengthptr += length; + } + return TRUE; + } + + /* Another branch follows. In the pre-compile phase, we can move the code + pointer back to where it was for the start of the first branch. (That is, + pretend that each branch is the only one.) + + In the real compile phase, insert an ALT node. Its length field points back + to the previous branch while the bracket remains open. At the end the chain + is reversed. It's done like this so that the start of the bracket has a + zero offset until it is closed, making it possible to detect recursion. */ + + if (lengthptr != NULL) + { + code = *codeptr + 1 + LINK_SIZE + skipbytes; + length += 1 + LINK_SIZE; + } + else + { + *code = OP_ALT; + PUT(code, 1, (int)(code - last_branch)); + bc.current_branch = last_branch = code; + code += 1 + LINK_SIZE; + } + + ptr++; + } +/* Control never reaches here */ +} + + + + +/************************************************* +* Check for anchored expression * +*************************************************/ + +/* Try to find out if this is an anchored regular expression. Consider each +alternative branch. If they all start with OP_SOD or OP_CIRC, or with a bracket +all of whose alternatives start with OP_SOD or OP_CIRC (recurse ad lib), then +it's anchored. However, if this is a multiline pattern, then only OP_SOD will +be found, because ^ generates OP_CIRCM in that mode. + +We can also consider a regex to be anchored if OP_SOM starts all its branches. +This is the code for \G, which means "match at start of match position, taking +into account the match offset". + +A branch is also implicitly anchored if it starts with .* and DOTALL is set, +because that will try the rest of the pattern at all possible matching points, +so there is no point trying again.... er .... + +.... except when the .* appears inside capturing parentheses, and there is a +subsequent back reference to those parentheses. We haven't enough information +to catch that case precisely. + +At first, the best we could do was to detect when .* was in capturing brackets +and the highest back reference was greater than or equal to that level. +However, by keeping a bitmap of the first 31 back references, we can catch some +of the more common cases more precisely. + +... A second exception is when the .* appears inside an atomic group, because +this prevents the number of characters it matches from being adjusted. + +Arguments: + code points to start of expression (the bracket) + bracket_map a bitmap of which brackets we are inside while testing; this + handles up to substring 31; after that we just have to take + the less precise approach + cd points to the compile data block + atomcount atomic group level + +Returns: TRUE or FALSE +*/ + +static BOOL +is_anchored(register const pcre_uchar *code, unsigned int bracket_map, + compile_data *cd, int atomcount) +{ +do { + const pcre_uchar *scode = first_significant_code( + code + PRIV(OP_lengths)[*code], FALSE); + register int op = *scode; + + /* Non-capturing brackets */ + + if (op == OP_BRA || op == OP_BRAPOS || + op == OP_SBRA || op == OP_SBRAPOS) + { + if (!is_anchored(scode, bracket_map, cd, atomcount)) return FALSE; + } + + /* Capturing brackets */ + + else if (op == OP_CBRA || op == OP_CBRAPOS || + op == OP_SCBRA || op == OP_SCBRAPOS) + { + int n = GET2(scode, 1+LINK_SIZE); + int new_map = bracket_map | ((n < 32)? (1 << n) : 1); + if (!is_anchored(scode, new_map, cd, atomcount)) return FALSE; + } + + /* Positive forward assertions and conditions */ + + else if (op == OP_ASSERT || op == OP_COND) + { + if (!is_anchored(scode, bracket_map, cd, atomcount)) return FALSE; + } + + /* Atomic groups */ + + else if (op == OP_ONCE || op == OP_ONCE_NC) + { + if (!is_anchored(scode, bracket_map, cd, atomcount + 1)) + return FALSE; + } + + /* .* is not anchored unless DOTALL is set (which generates OP_ALLANY) and + it isn't in brackets that are or may be referenced or inside an atomic + group. */ + + else if ((op == OP_TYPESTAR || op == OP_TYPEMINSTAR || + op == OP_TYPEPOSSTAR)) + { + if (scode[1] != OP_ALLANY || (bracket_map & cd->backref_map) != 0 || + atomcount > 0 || cd->had_pruneorskip) + return FALSE; + } + + /* Check for explicit anchoring */ + + else if (op != OP_SOD && op != OP_SOM && op != OP_CIRC) return FALSE; + + code += GET(code, 1); + } +while (*code == OP_ALT); /* Loop for each alternative */ +return TRUE; +} + + + +/************************************************* +* Check for starting with ^ or .* * +*************************************************/ + +/* This is called to find out if every branch starts with ^ or .* so that +"first char" processing can be done to speed things up in multiline +matching and for non-DOTALL patterns that start with .* (which must start at +the beginning or after \n). As in the case of is_anchored() (see above), we +have to take account of back references to capturing brackets that contain .* +because in that case we can't make the assumption. Also, the appearance of .* +inside atomic brackets or in a pattern that contains *PRUNE or *SKIP does not +count, because once again the assumption no longer holds. + +Arguments: + code points to start of expression (the bracket) + bracket_map a bitmap of which brackets we are inside while testing; this + handles up to substring 31; after that we just have to take + the less precise approach + cd points to the compile data + atomcount atomic group level + +Returns: TRUE or FALSE +*/ + +static BOOL +is_startline(const pcre_uchar *code, unsigned int bracket_map, + compile_data *cd, int atomcount) +{ +do { + const pcre_uchar *scode = first_significant_code( + code + PRIV(OP_lengths)[*code], FALSE); + register int op = *scode; + + /* If we are at the start of a conditional assertion group, *both* the + conditional assertion *and* what follows the condition must satisfy the test + for start of line. Other kinds of condition fail. Note that there may be an + auto-callout at the start of a condition. */ + + if (op == OP_COND) + { + scode += 1 + LINK_SIZE; + if (*scode == OP_CALLOUT) scode += PRIV(OP_lengths)[OP_CALLOUT]; + switch (*scode) + { + case OP_CREF: + case OP_DNCREF: + case OP_RREF: + case OP_DNRREF: + case OP_DEF: + case OP_FAIL: + return FALSE; + + default: /* Assertion */ + if (!is_startline(scode, bracket_map, cd, atomcount)) return FALSE; + do scode += GET(scode, 1); while (*scode == OP_ALT); + scode += 1 + LINK_SIZE; + break; + } + scode = first_significant_code(scode, FALSE); + op = *scode; + } + + /* Non-capturing brackets */ + + if (op == OP_BRA || op == OP_BRAPOS || + op == OP_SBRA || op == OP_SBRAPOS) + { + if (!is_startline(scode, bracket_map, cd, atomcount)) return FALSE; + } + + /* Capturing brackets */ + + else if (op == OP_CBRA || op == OP_CBRAPOS || + op == OP_SCBRA || op == OP_SCBRAPOS) + { + int n = GET2(scode, 1+LINK_SIZE); + int new_map = bracket_map | ((n < 32)? (1 << n) : 1); + if (!is_startline(scode, new_map, cd, atomcount)) return FALSE; + } + + /* Positive forward assertions */ + + else if (op == OP_ASSERT) + { + if (!is_startline(scode, bracket_map, cd, atomcount)) return FALSE; + } + + /* Atomic brackets */ + + else if (op == OP_ONCE || op == OP_ONCE_NC) + { + if (!is_startline(scode, bracket_map, cd, atomcount + 1)) return FALSE; + } + + /* .* means "start at start or after \n" if it isn't in atomic brackets or + brackets that may be referenced, as long as the pattern does not contain + *PRUNE or *SKIP, because these break the feature. Consider, for example, + /.*?a(*PRUNE)b/ with the subject "aab", which matches "ab", i.e. not at the + start of a line. */ + + else if (op == OP_TYPESTAR || op == OP_TYPEMINSTAR || op == OP_TYPEPOSSTAR) + { + if (scode[1] != OP_ANY || (bracket_map & cd->backref_map) != 0 || + atomcount > 0 || cd->had_pruneorskip) + return FALSE; + } + + /* Check for explicit circumflex; anything else gives a FALSE result. Note + in particular that this includes atomic brackets OP_ONCE and OP_ONCE_NC + because the number of characters matched by .* cannot be adjusted inside + them. */ + + else if (op != OP_CIRC && op != OP_CIRCM) return FALSE; + + /* Move on to the next alternative */ + + code += GET(code, 1); + } +while (*code == OP_ALT); /* Loop for each alternative */ +return TRUE; +} + + + +/************************************************* +* Check for asserted fixed first char * +*************************************************/ + +/* During compilation, the "first char" settings from forward assertions are +discarded, because they can cause conflicts with actual literals that follow. +However, if we end up without a first char setting for an unanchored pattern, +it is worth scanning the regex to see if there is an initial asserted first +char. If all branches start with the same asserted char, or with a +non-conditional bracket all of whose alternatives start with the same asserted +char (recurse ad lib), then we return that char, with the flags set to zero or +REQ_CASELESS; otherwise return zero with REQ_NONE in the flags. + +Arguments: + code points to start of expression (the bracket) + flags points to the first char flags, or to REQ_NONE + inassert TRUE if in an assertion + +Returns: the fixed first char, or 0 with REQ_NONE in flags +*/ + +static pcre_uint32 +find_firstassertedchar(const pcre_uchar *code, pcre_int32 *flags, + BOOL inassert) +{ +register pcre_uint32 c = 0; +int cflags = REQ_NONE; + +*flags = REQ_NONE; +do { + pcre_uint32 d; + int dflags; + int xl = (*code == OP_CBRA || *code == OP_SCBRA || + *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0; + const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl, + TRUE); + register pcre_uchar op = *scode; + + switch(op) + { + default: + return 0; + + case OP_BRA: + case OP_BRAPOS: + case OP_CBRA: + case OP_SCBRA: + case OP_CBRAPOS: + case OP_SCBRAPOS: + case OP_ASSERT: + case OP_ONCE: + case OP_ONCE_NC: + d = find_firstassertedchar(scode, &dflags, op == OP_ASSERT); + if (dflags < 0) + return 0; + if (cflags < 0) { c = d; cflags = dflags; } else if (c != d || cflags != dflags) return 0; + break; + + case OP_EXACT: + scode += IMM2_SIZE; + /* Fall through */ + + case OP_CHAR: + case OP_PLUS: + case OP_MINPLUS: + case OP_POSPLUS: + if (!inassert) return 0; + if (cflags < 0) { c = scode[1]; cflags = 0; } + else if (c != scode[1]) return 0; + break; + + case OP_EXACTI: + scode += IMM2_SIZE; + /* Fall through */ + + case OP_CHARI: + case OP_PLUSI: + case OP_MINPLUSI: + case OP_POSPLUSI: + if (!inassert) return 0; + if (cflags < 0) { c = scode[1]; cflags = REQ_CASELESS; } + else if (c != scode[1]) return 0; + break; + } + + code += GET(code, 1); + } +while (*code == OP_ALT); + +*flags = cflags; +return c; +} + + + +/************************************************* +* Add an entry to the name/number table * +*************************************************/ + +/* This function is called between compiling passes to add an entry to the +name/number table, maintaining alphabetical order. Checking for permitted +and forbidden duplicates has already been done. + +Arguments: + cd the compile data block + name the name to add + length the length of the name + groupno the group number + +Returns: nothing +*/ + +static void +add_name(compile_data *cd, const pcre_uchar *name, int length, + unsigned int groupno) +{ +int i; +pcre_uchar *slot = cd->name_table; + +for (i = 0; i < cd->names_found; i++) + { + int crc = memcmp(name, slot+IMM2_SIZE, IN_UCHARS(length)); + if (crc == 0 && slot[IMM2_SIZE+length] != 0) + crc = -1; /* Current name is a substring */ + + /* Make space in the table and break the loop for an earlier name. For a + duplicate or later name, carry on. We do this for duplicates so that in the + simple case (when ?(| is not used) they are in order of their numbers. In all + cases they are in the order in which they appear in the pattern. */ + + if (crc < 0) + { + memmove(slot + cd->name_entry_size, slot, + IN_UCHARS((cd->names_found - i) * cd->name_entry_size)); + break; + } + + /* Continue the loop for a later or duplicate name */ + + slot += cd->name_entry_size; + } + +PUT2(slot, 0, groupno); +memcpy(slot + IMM2_SIZE, name, IN_UCHARS(length)); +slot[IMM2_SIZE + length] = 0; +cd->names_found++; +} + + + +/************************************************* +* Compile a Regular Expression * +*************************************************/ + +/* This function takes a string and returns a pointer to a block of store +holding a compiled version of the expression. The original API for this +function had no error code return variable; it is retained for backwards +compatibility. The new function is given a new name. + +Arguments: + pattern the regular expression + options various option bits + errorcodeptr pointer to error code variable (pcre_compile2() only) + can be NULL if you don't want a code value + errorptr pointer to pointer to error text + erroroffset ptr offset in pattern where error was detected + tables pointer to character tables or NULL + +Returns: pointer to compiled data block, or NULL on error, + with errorptr and erroroffset set +*/ + +#if defined COMPILE_PCRE8 +PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION +pcre_compile(const char *pattern, int options, const char **errorptr, + int *erroroffset, const unsigned char *tables) +#elif defined COMPILE_PCRE16 +PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION +pcre16_compile(PCRE_SPTR16 pattern, int options, const char **errorptr, + int *erroroffset, const unsigned char *tables) +#elif defined COMPILE_PCRE32 +PCRE_EXP_DEFN pcre32 * PCRE_CALL_CONVENTION +pcre32_compile(PCRE_SPTR32 pattern, int options, const char **errorptr, + int *erroroffset, const unsigned char *tables) +#endif +{ +#if defined COMPILE_PCRE8 +return pcre_compile2(pattern, options, NULL, errorptr, erroroffset, tables); +#elif defined COMPILE_PCRE16 +return pcre16_compile2(pattern, options, NULL, errorptr, erroroffset, tables); +#elif defined COMPILE_PCRE32 +return pcre32_compile2(pattern, options, NULL, errorptr, erroroffset, tables); +#endif +} + + +#if defined COMPILE_PCRE8 +PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION +pcre_compile2(const char *pattern, int options, int *errorcodeptr, + const char **errorptr, int *erroroffset, const unsigned char *tables) +#elif defined COMPILE_PCRE16 +PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION +pcre16_compile2(PCRE_SPTR16 pattern, int options, int *errorcodeptr, + const char **errorptr, int *erroroffset, const unsigned char *tables) +#elif defined COMPILE_PCRE32 +PCRE_EXP_DEFN pcre32 * PCRE_CALL_CONVENTION +pcre32_compile2(PCRE_SPTR32 pattern, int options, int *errorcodeptr, + const char **errorptr, int *erroroffset, const unsigned char *tables) +#endif +{ +REAL_PCRE *re; +int length = 1; /* For final END opcode */ +pcre_int32 firstcharflags, reqcharflags; +pcre_uint32 firstchar, reqchar; +pcre_uint32 limit_match = PCRE_UINT32_MAX; +pcre_uint32 limit_recursion = PCRE_UINT32_MAX; +int newline; +int errorcode = 0; +int skipatstart = 0; +BOOL utf; +BOOL never_utf = FALSE; +size_t size; +pcre_uchar *code; +const pcre_uchar *codestart; +const pcre_uchar *ptr; +compile_data compile_block; +compile_data *cd = &compile_block; + +/* This space is used for "compiling" into during the first phase, when we are +computing the amount of memory that is needed. Compiled items are thrown away +as soon as possible, so that a fairly large buffer should be sufficient for +this purpose. The same space is used in the second phase for remembering where +to fill in forward references to subpatterns. That may overflow, in which case +new memory is obtained from malloc(). */ + +pcre_uchar cworkspace[COMPILE_WORK_SIZE]; + +/* This vector is used for remembering name groups during the pre-compile. In a +similar way to cworkspace, it can be expanded using malloc() if necessary. */ + +named_group named_groups[NAMED_GROUP_LIST_SIZE]; + +/* Set this early so that early errors get offset 0. */ + +ptr = (const pcre_uchar *)pattern; + +/* We can't pass back an error message if errorptr is NULL; I guess the best we +can do is just return NULL, but we can set a code value if there is a code +pointer. */ + +if (errorptr == NULL) + { + if (errorcodeptr != NULL) *errorcodeptr = 99; + return NULL; + } + +*errorptr = NULL; +if (errorcodeptr != NULL) *errorcodeptr = ERR0; + +/* However, we can give a message for this error */ + +if (erroroffset == NULL) + { + errorcode = ERR16; + goto PCRE_EARLY_ERROR_RETURN2; + } + +*erroroffset = 0; + +/* Set up pointers to the individual character tables */ + +if (tables == NULL) tables = PRIV(default_tables); +cd->lcc = tables + lcc_offset; +cd->fcc = tables + fcc_offset; +cd->cbits = tables + cbits_offset; +cd->ctypes = tables + ctypes_offset; + +/* Check that all undefined public option bits are zero */ + +if ((options & ~PUBLIC_COMPILE_OPTIONS) != 0) + { + errorcode = ERR17; + goto PCRE_EARLY_ERROR_RETURN; + } + +/* If PCRE_NEVER_UTF is set, remember it. */ + +if ((options & PCRE_NEVER_UTF) != 0) never_utf = TRUE; + +/* Check for global one-time settings at the start of the pattern, and remember +the offset for later. */ + +cd->external_flags = 0; /* Initialize here for LIMIT_MATCH/RECURSION */ + +while (ptr[skipatstart] == CHAR_LEFT_PARENTHESIS && + ptr[skipatstart+1] == CHAR_ASTERISK) + { + int newnl = 0; + int newbsr = 0; + +/* For completeness and backward compatibility, (*UTFn) is supported in the +relevant libraries, but (*UTF) is generic and always supported. Note that +PCRE_UTF8 == PCRE_UTF16 == PCRE_UTF32. */ + +#ifdef COMPILE_PCRE8 + if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF8_RIGHTPAR, 5) == 0) + { skipatstart += 7; options |= PCRE_UTF8; continue; } +#endif +#ifdef COMPILE_PCRE16 + if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF16_RIGHTPAR, 6) == 0) + { skipatstart += 8; options |= PCRE_UTF16; continue; } +#endif +#ifdef COMPILE_PCRE32 + if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF32_RIGHTPAR, 6) == 0) + { skipatstart += 8; options |= PCRE_UTF32; continue; } +#endif + + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 4) == 0) + { skipatstart += 6; options |= PCRE_UTF8; continue; } + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UCP_RIGHTPAR, 4) == 0) + { skipatstart += 6; options |= PCRE_UCP; continue; } + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_NO_AUTO_POSSESS_RIGHTPAR, 16) == 0) + { skipatstart += 18; options |= PCRE_NO_AUTO_POSSESS; continue; } + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_NO_START_OPT_RIGHTPAR, 13) == 0) + { skipatstart += 15; options |= PCRE_NO_START_OPTIMIZE; continue; } + + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_LIMIT_MATCH_EQ, 12) == 0) + { + pcre_uint32 c = 0; + int p = skipatstart + 14; + while (isdigit(ptr[p])) + { + if (c > PCRE_UINT32_MAX / 10 - 1) break; /* Integer overflow */ + c = c*10 + ptr[p++] - CHAR_0; + } + if (ptr[p++] != CHAR_RIGHT_PARENTHESIS) break; + if (c < limit_match) + { + limit_match = c; + cd->external_flags |= PCRE_MLSET; + } + skipatstart = p; + continue; + } + + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_LIMIT_RECURSION_EQ, 16) == 0) + { + pcre_uint32 c = 0; + int p = skipatstart + 18; + while (isdigit(ptr[p])) + { + if (c > PCRE_UINT32_MAX / 10 - 1) break; /* Integer overflow check */ + c = c*10 + ptr[p++] - CHAR_0; + } + if (ptr[p++] != CHAR_RIGHT_PARENTHESIS) break; + if (c < limit_recursion) + { + limit_recursion = c; + cd->external_flags |= PCRE_RLSET; + } + skipatstart = p; + continue; + } + + if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_CR_RIGHTPAR, 3) == 0) + { skipatstart += 5; newnl = PCRE_NEWLINE_CR; } + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_LF_RIGHTPAR, 3) == 0) + { skipatstart += 5; newnl = PCRE_NEWLINE_LF; } + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_CRLF_RIGHTPAR, 5) == 0) + { skipatstart += 7; newnl = PCRE_NEWLINE_CR + PCRE_NEWLINE_LF; } + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_ANY_RIGHTPAR, 4) == 0) + { skipatstart += 6; newnl = PCRE_NEWLINE_ANY; } + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_ANYCRLF_RIGHTPAR, 8) == 0) + { skipatstart += 10; newnl = PCRE_NEWLINE_ANYCRLF; } + + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_BSR_ANYCRLF_RIGHTPAR, 12) == 0) + { skipatstart += 14; newbsr = PCRE_BSR_ANYCRLF; } + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_BSR_UNICODE_RIGHTPAR, 12) == 0) + { skipatstart += 14; newbsr = PCRE_BSR_UNICODE; } + + if (newnl != 0) + options = (options & ~PCRE_NEWLINE_BITS) | newnl; + else if (newbsr != 0) + options = (options & ~(PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) | newbsr; + else break; + } + +/* PCRE_UTF(16|32) have the same value as PCRE_UTF8. */ +utf = (options & PCRE_UTF8) != 0; +if (utf && never_utf) + { + errorcode = ERR78; + goto PCRE_EARLY_ERROR_RETURN2; + } + +/* Can't support UTF unless PCRE has been compiled to include the code. The +return of an error code from PRIV(valid_utf)() is a new feature, introduced in +release 8.13. It is passed back from pcre_[dfa_]exec(), but at the moment is +not used here. */ + +#ifdef SUPPORT_UTF +if (utf && (options & PCRE_NO_UTF8_CHECK) == 0 && + (errorcode = PRIV(valid_utf)((PCRE_PUCHAR)pattern, -1, erroroffset)) != 0) + { +#if defined COMPILE_PCRE8 + errorcode = ERR44; +#elif defined COMPILE_PCRE16 + errorcode = ERR74; +#elif defined COMPILE_PCRE32 + errorcode = ERR77; +#endif + goto PCRE_EARLY_ERROR_RETURN2; + } +#else +if (utf) + { + errorcode = ERR32; + goto PCRE_EARLY_ERROR_RETURN; + } +#endif + +/* Can't support UCP unless PCRE has been compiled to include the code. */ + +#ifndef SUPPORT_UCP +if ((options & PCRE_UCP) != 0) + { + errorcode = ERR67; + goto PCRE_EARLY_ERROR_RETURN; + } +#endif + +/* Check validity of \R options. */ + +if ((options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) == + (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) + { + errorcode = ERR56; + goto PCRE_EARLY_ERROR_RETURN; + } + +/* Handle different types of newline. The three bits give seven cases. The +current code allows for fixed one- or two-byte sequences, plus "any" and +"anycrlf". */ + +switch (options & PCRE_NEWLINE_BITS) + { + case 0: newline = NEWLINE; break; /* Build-time default */ + case PCRE_NEWLINE_CR: newline = CHAR_CR; break; + case PCRE_NEWLINE_LF: newline = CHAR_NL; break; + case PCRE_NEWLINE_CR+ + PCRE_NEWLINE_LF: newline = (CHAR_CR << 8) | CHAR_NL; break; + case PCRE_NEWLINE_ANY: newline = -1; break; + case PCRE_NEWLINE_ANYCRLF: newline = -2; break; + default: errorcode = ERR56; goto PCRE_EARLY_ERROR_RETURN; + } + +if (newline == -2) + { + cd->nltype = NLTYPE_ANYCRLF; + } +else if (newline < 0) + { + cd->nltype = NLTYPE_ANY; + } +else + { + cd->nltype = NLTYPE_FIXED; + if (newline > 255) + { + cd->nllen = 2; + cd->nl[0] = (newline >> 8) & 255; + cd->nl[1] = newline & 255; + } + else + { + cd->nllen = 1; + cd->nl[0] = newline; + } + } + +/* Maximum back reference and backref bitmap. The bitmap records up to 31 back +references to help in deciding whether (.*) can be treated as anchored or not. +*/ + +cd->top_backref = 0; +cd->backref_map = 0; + +/* Reflect pattern for debugging output */ + +DPRINTF(("------------------------------------------------------------------\n")); +#ifdef PCRE_DEBUG +print_puchar(stdout, (PCRE_PUCHAR)pattern); +#endif +DPRINTF(("\n")); + +/* Pretend to compile the pattern while actually just accumulating the length +of memory required. This behaviour is triggered by passing a non-NULL final +argument to compile_regex(). We pass a block of workspace (cworkspace) for it +to compile parts of the pattern into; the compiled code is discarded when it is +no longer needed, so hopefully this workspace will never overflow, though there +is a test for its doing so. */ + +cd->bracount = cd->final_bracount = 0; +cd->names_found = 0; +cd->name_entry_size = 0; +cd->name_table = NULL; +cd->dupnames = FALSE; +cd->dupgroups = FALSE; +cd->namedrefcount = 0; +cd->start_code = cworkspace; +cd->hwm = cworkspace; +cd->iscondassert = FALSE; +cd->start_workspace = cworkspace; +cd->workspace_size = COMPILE_WORK_SIZE; +cd->named_groups = named_groups; +cd->named_group_list_size = NAMED_GROUP_LIST_SIZE; +cd->start_pattern = (const pcre_uchar *)pattern; +cd->end_pattern = (const pcre_uchar *)(pattern + STRLEN_UC((const pcre_uchar *)pattern)); +cd->req_varyopt = 0; +cd->parens_depth = 0; +cd->assert_depth = 0; +cd->max_lookbehind = 0; +cd->external_options = options; +cd->open_caps = NULL; + +/* Now do the pre-compile. On error, errorcode will be set non-zero, so we +don't need to look at the result of the function here. The initial options have +been put into the cd block so that they can be changed if an option setting is +found within the regex right at the beginning. Bringing initial option settings +outside can help speed up starting point checks. */ + +ptr += skipatstart; +code = cworkspace; +*code = OP_BRA; + +(void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE, + FALSE, 0, 0, &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL, + cd, &length); +if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN; + +DPRINTF(("end pre-compile: length=%d workspace=%d\n", length, + (int)(cd->hwm - cworkspace))); + +if (length > MAX_PATTERN_SIZE) + { + errorcode = ERR20; + goto PCRE_EARLY_ERROR_RETURN; + } + +/* Compute the size of the data block for storing the compiled pattern. Integer +overflow should no longer be possible because nowadays we limit the maximum +value of cd->names_found and cd->name_entry_size. */ + +size = sizeof(REAL_PCRE) + + (length + cd->names_found * cd->name_entry_size) * sizeof(pcre_uchar); + +/* Get the memory. */ + +re = (REAL_PCRE *)(PUBL(malloc))(size); +if (re == NULL) + { + errorcode = ERR21; + goto PCRE_EARLY_ERROR_RETURN; + } + +/* Put in the magic number, and save the sizes, initial options, internal +flags, and character table pointer. NULL is used for the default character +tables. The nullpad field is at the end; it's there to help in the case when a +regex compiled on a system with 4-byte pointers is run on another with 8-byte +pointers. */ + +re->magic_number = MAGIC_NUMBER; +re->size = (int)size; +re->options = cd->external_options; +re->flags = cd->external_flags; +re->limit_match = limit_match; +re->limit_recursion = limit_recursion; +re->first_char = 0; +re->req_char = 0; +re->name_table_offset = sizeof(REAL_PCRE) / sizeof(pcre_uchar); +re->name_entry_size = cd->name_entry_size; +re->name_count = cd->names_found; +re->ref_count = 0; +re->tables = (tables == PRIV(default_tables))? NULL : tables; +re->nullpad = NULL; +#ifdef COMPILE_PCRE32 +re->dummy = 0; +#else +re->dummy1 = re->dummy2 = re->dummy3 = 0; +#endif + +/* The starting points of the name/number translation table and of the code are +passed around in the compile data block. The start/end pattern and initial +options are already set from the pre-compile phase, as is the name_entry_size +field. Reset the bracket count and the names_found field. Also reset the hwm +field; this time it's used for remembering forward references to subpatterns. +*/ + +cd->final_bracount = cd->bracount; /* Save for checking forward references */ +cd->parens_depth = 0; +cd->assert_depth = 0; +cd->bracount = 0; +cd->max_lookbehind = 0; +cd->name_table = (pcre_uchar *)re + re->name_table_offset; +codestart = cd->name_table + re->name_entry_size * re->name_count; +cd->start_code = codestart; +cd->hwm = (pcre_uchar *)(cd->start_workspace); +cd->iscondassert = FALSE; +cd->req_varyopt = 0; +cd->had_accept = FALSE; +cd->had_pruneorskip = FALSE; +cd->check_lookbehind = FALSE; +cd->open_caps = NULL; + +/* If any named groups were found, create the name/number table from the list +created in the first pass. */ + +if (cd->names_found > 0) + { + int i = cd->names_found; + named_group *ng = cd->named_groups; + cd->names_found = 0; + for (; i > 0; i--, ng++) + add_name(cd, ng->name, ng->length, ng->number); + if (cd->named_group_list_size > NAMED_GROUP_LIST_SIZE) + (PUBL(free))((void *)cd->named_groups); + } + +/* Set up a starting, non-extracting bracket, then compile the expression. On +error, errorcode will be set non-zero, so we don't need to look at the result +of the function here. */ + +ptr = (const pcre_uchar *)pattern + skipatstart; +code = (pcre_uchar *)codestart; +*code = OP_BRA; +(void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0, + &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL, cd, NULL); +re->top_bracket = cd->bracount; +re->top_backref = cd->top_backref; +re->max_lookbehind = cd->max_lookbehind; +re->flags = cd->external_flags | PCRE_MODE; + +if (cd->had_accept) + { + reqchar = 0; /* Must disable after (*ACCEPT) */ + reqcharflags = REQ_NONE; + } + +/* If not reached end of pattern on success, there's an excess bracket. */ + +if (errorcode == 0 && *ptr != CHAR_NULL) errorcode = ERR22; + +/* Fill in the terminating state and check for disastrous overflow, but +if debugging, leave the test till after things are printed out. */ + +*code++ = OP_END; + +#ifndef PCRE_DEBUG +if (code - codestart > length) errorcode = ERR23; +#endif + +#ifdef SUPPORT_VALGRIND +/* If the estimated length exceeds the really used length, mark the extra +allocated memory as unaddressable, so that any out-of-bound reads can be +detected. */ +VALGRIND_MAKE_MEM_NOACCESS(code, (length - (code - codestart)) * sizeof(pcre_uchar)); +#endif + +/* Fill in any forward references that are required. There may be repeated +references; optimize for them, as searching a large regex takes time. */ + +if (cd->hwm > cd->start_workspace) + { + int prev_recno = -1; + const pcre_uchar *groupptr = NULL; + while (errorcode == 0 && cd->hwm > cd->start_workspace) + { + int offset, recno; + cd->hwm -= LINK_SIZE; + offset = GET(cd->hwm, 0); + + /* Check that the hwm handling hasn't gone wrong. This whole area is + rewritten in PCRE2 because there are some obscure cases. */ + + if (offset == 0 || codestart[offset-1] != OP_RECURSE) + { + errorcode = ERR10; + break; + } + + recno = GET(codestart, offset); + if (recno != prev_recno) + { + groupptr = PRIV(find_bracket)(codestart, utf, recno); + prev_recno = recno; + } + if (groupptr == NULL) errorcode = ERR53; + else PUT(((pcre_uchar *)codestart), offset, (int)(groupptr - codestart)); + } + } + +/* If the workspace had to be expanded, free the new memory. Set the pointer to +NULL to indicate that forward references have been filled in. */ + +if (cd->workspace_size > COMPILE_WORK_SIZE) + (PUBL(free))((void *)cd->start_workspace); +cd->start_workspace = NULL; + +/* Give an error if there's back reference to a non-existent capturing +subpattern. */ + +if (errorcode == 0 && re->top_backref > re->top_bracket) errorcode = ERR15; + +/* Unless disabled, check whether any single character iterators can be +auto-possessified. The function overwrites the appropriate opcode values, so +the type of the pointer must be cast. NOTE: the intermediate variable "temp" is +used in this code because at least one compiler gives a warning about loss of +"const" attribute if the cast (pcre_uchar *)codestart is used directly in the +function call. */ + +if (errorcode == 0 && (options & PCRE_NO_AUTO_POSSESS) == 0) + { + pcre_uchar *temp = (pcre_uchar *)codestart; + auto_possessify(temp, utf, cd); + } + +/* If there were any lookbehind assertions that contained OP_RECURSE +(recursions or subroutine calls), a flag is set for them to be checked here, +because they may contain forward references. Actual recursions cannot be fixed +length, but subroutine calls can. It is done like this so that those without +OP_RECURSE that are not fixed length get a diagnosic with a useful offset. The +exceptional ones forgo this. We scan the pattern to check that they are fixed +length, and set their lengths. */ + +if (errorcode == 0 && cd->check_lookbehind) + { + pcre_uchar *cc = (pcre_uchar *)codestart; + + /* Loop, searching for OP_REVERSE items, and process those that do not have + their length set. (Actually, it will also re-process any that have a length + of zero, but that is a pathological case, and it does no harm.) When we find + one, we temporarily terminate the branch it is in while we scan it. */ + + for (cc = (pcre_uchar *)PRIV(find_bracket)(codestart, utf, -1); + cc != NULL; + cc = (pcre_uchar *)PRIV(find_bracket)(cc, utf, -1)) + { + if (GET(cc, 1) == 0) + { + int fixed_length; + pcre_uchar *be = cc - 1 - LINK_SIZE + GET(cc, -LINK_SIZE); + int end_op = *be; + *be = OP_END; + fixed_length = find_fixedlength(cc, (re->options & PCRE_UTF8) != 0, TRUE, + cd, NULL); + *be = end_op; + DPRINTF(("fixed length = %d\n", fixed_length)); + if (fixed_length < 0) + { + errorcode = (fixed_length == -2)? ERR36 : + (fixed_length == -4)? ERR70 : ERR25; + break; + } + if (fixed_length > cd->max_lookbehind) cd->max_lookbehind = fixed_length; + PUT(cc, 1, fixed_length); + } + cc += 1 + LINK_SIZE; + } + } + +/* Failed to compile, or error while post-processing */ + +if (errorcode != 0) + { + (PUBL(free))(re); + PCRE_EARLY_ERROR_RETURN: + *erroroffset = (int)(ptr - (const pcre_uchar *)pattern); + PCRE_EARLY_ERROR_RETURN2: + *errorptr = find_error_text(errorcode); + if (errorcodeptr != NULL) *errorcodeptr = errorcode; + return NULL; + } + +/* If the anchored option was not passed, set the flag if we can determine that +the pattern is anchored by virtue of ^ characters or \A or anything else, such +as starting with non-atomic .* when DOTALL is set and there are no occurrences +of *PRUNE or *SKIP. + +Otherwise, if we know what the first byte has to be, save it, because that +speeds up unanchored matches no end. If not, see if we can set the +PCRE_STARTLINE flag. This is helpful for multiline matches when all branches +start with ^. and also when all branches start with non-atomic .* for +non-DOTALL matches when *PRUNE and SKIP are not present. */ + +if ((re->options & PCRE_ANCHORED) == 0) + { + if (is_anchored(codestart, 0, cd, 0)) re->options |= PCRE_ANCHORED; + else + { + if (firstcharflags < 0) + firstchar = find_firstassertedchar(codestart, &firstcharflags, FALSE); + if (firstcharflags >= 0) /* Remove caseless flag for non-caseable chars */ + { +#if defined COMPILE_PCRE8 + re->first_char = firstchar & 0xff; +#elif defined COMPILE_PCRE16 + re->first_char = firstchar & 0xffff; +#elif defined COMPILE_PCRE32 + re->first_char = firstchar; +#endif + if ((firstcharflags & REQ_CASELESS) != 0) + { +#if defined SUPPORT_UCP && !(defined COMPILE_PCRE8) + /* We ignore non-ASCII first chars in 8 bit mode. */ + if (utf) + { + if (re->first_char < 128) + { + if (cd->fcc[re->first_char] != re->first_char) + re->flags |= PCRE_FCH_CASELESS; + } + else if (UCD_OTHERCASE(re->first_char) != re->first_char) + re->flags |= PCRE_FCH_CASELESS; + } + else +#endif + if (MAX_255(re->first_char) + && cd->fcc[re->first_char] != re->first_char) + re->flags |= PCRE_FCH_CASELESS; + } + + re->flags |= PCRE_FIRSTSET; + } + + else if (is_startline(codestart, 0, cd, 0)) re->flags |= PCRE_STARTLINE; + } + } + +/* For an anchored pattern, we use the "required byte" only if it follows a +variable length item in the regex. Remove the caseless flag for non-caseable +bytes. */ + +if (reqcharflags >= 0 && + ((re->options & PCRE_ANCHORED) == 0 || (reqcharflags & REQ_VARY) != 0)) + { +#if defined COMPILE_PCRE8 + re->req_char = reqchar & 0xff; +#elif defined COMPILE_PCRE16 + re->req_char = reqchar & 0xffff; +#elif defined COMPILE_PCRE32 + re->req_char = reqchar; +#endif + if ((reqcharflags & REQ_CASELESS) != 0) + { +#if defined SUPPORT_UCP && !(defined COMPILE_PCRE8) + /* We ignore non-ASCII first chars in 8 bit mode. */ + if (utf) + { + if (re->req_char < 128) + { + if (cd->fcc[re->req_char] != re->req_char) + re->flags |= PCRE_RCH_CASELESS; + } + else if (UCD_OTHERCASE(re->req_char) != re->req_char) + re->flags |= PCRE_RCH_CASELESS; + } + else +#endif + if (MAX_255(re->req_char) && cd->fcc[re->req_char] != re->req_char) + re->flags |= PCRE_RCH_CASELESS; + } + + re->flags |= PCRE_REQCHSET; + } + +/* Print out the compiled data if debugging is enabled. This is never the +case when building a production library. */ + +#ifdef PCRE_DEBUG +printf("Length = %d top_bracket = %d top_backref = %d\n", + length, re->top_bracket, re->top_backref); + +printf("Options=%08x\n", re->options); + +if ((re->flags & PCRE_FIRSTSET) != 0) + { + pcre_uchar ch = re->first_char; + const char *caseless = + ((re->flags & PCRE_FCH_CASELESS) == 0)? "" : " (caseless)"; + if (PRINTABLE(ch)) printf("First char = %c%s\n", ch, caseless); + else printf("First char = \\x%02x%s\n", ch, caseless); + } + +if ((re->flags & PCRE_REQCHSET) != 0) + { + pcre_uchar ch = re->req_char; + const char *caseless = + ((re->flags & PCRE_RCH_CASELESS) == 0)? "" : " (caseless)"; + if (PRINTABLE(ch)) printf("Req char = %c%s\n", ch, caseless); + else printf("Req char = \\x%02x%s\n", ch, caseless); + } + +#if defined COMPILE_PCRE8 +pcre_printint((pcre *)re, stdout, TRUE); +#elif defined COMPILE_PCRE16 +pcre16_printint((pcre *)re, stdout, TRUE); +#elif defined COMPILE_PCRE32 +pcre32_printint((pcre *)re, stdout, TRUE); +#endif + +/* This check is done here in the debugging case so that the code that +was compiled can be seen. */ + +if (code - codestart > length) + { + (PUBL(free))(re); + *errorptr = find_error_text(ERR23); + *erroroffset = ptr - (pcre_uchar *)pattern; + if (errorcodeptr != NULL) *errorcodeptr = ERR23; + return NULL; + } +#endif /* PCRE_DEBUG */ + +/* Check for a pattern than can match an empty string, so that this information +can be provided to applications. */ + +do + { + if (could_be_empty_branch(codestart, code, utf, cd, NULL)) + { + re->flags |= PCRE_MATCH_EMPTY; + break; + } + codestart += GET(codestart, 1); + } +while (*codestart == OP_ALT); + +#if defined COMPILE_PCRE8 +return (pcre *)re; +#elif defined COMPILE_PCRE16 +return (pcre16 *)re; +#elif defined COMPILE_PCRE32 +return (pcre32 *)re; +#endif +} + +/* End of pcre_compile.c */ diff --git a/include/pcre/pcre_dfa_exec.c b/include/pcre/pcre_dfa_exec.c new file mode 100644 index 000000000..170ce6a00 --- /dev/null +++ b/include/pcre/pcre_dfa_exec.c @@ -0,0 +1,3674 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + +/* PCRE is a library of functions to support regular expressions whose syntax +and semantics are as close as possible to those of the Perl 5 language (but see +below for why this module is different). + + Written by Philip Hazel + Copyright (c) 1997-2014 University of Cambridge + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + +/* This module contains the external function pcre_dfa_exec(), which is an +alternative matching function that uses a sort of DFA algorithm (not a true +FSM). This is NOT Perl-compatible, but it has advantages in certain +applications. */ + + +/* NOTE ABOUT PERFORMANCE: A user of this function sent some code that improved +the performance of his patterns greatly. I could not use it as it stood, as it +was not thread safe, and made assumptions about pattern sizes. Also, it caused +test 7 to loop, and test 9 to crash with a segfault. + +The issue is the check for duplicate states, which is done by a simple linear +search up the state list. (Grep for "duplicate" below to find the code.) For +many patterns, there will never be many states active at one time, so a simple +linear search is fine. In patterns that have many active states, it might be a +bottleneck. The suggested code used an indexing scheme to remember which states +had previously been used for each character, and avoided the linear search when +it knew there was no chance of a duplicate. This was implemented when adding +states to the state lists. + +I wrote some thread-safe, not-limited code to try something similar at the time +of checking for duplicates (instead of when adding states), using index vectors +on the stack. It did give a 13% improvement with one specially constructed +pattern for certain subject strings, but on other strings and on many of the +simpler patterns in the test suite it did worse. The major problem, I think, +was the extra time to initialize the index. This had to be done for each call +of internal_dfa_exec(). (The supplied patch used a static vector, initialized +only once - I suspect this was the cause of the problems with the tests.) + +Overall, I concluded that the gains in some cases did not outweigh the losses +in others, so I abandoned this code. */ + + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#define NLBLOCK md /* Block containing newline information */ +#define PSSTART start_subject /* Field containing processed string start */ +#define PSEND end_subject /* Field containing processed string end */ + +#include "pcre_internal.h" + + +/* For use to indent debugging output */ + +#define SP " " + + +/************************************************* +* Code parameters and static tables * +*************************************************/ + +/* These are offsets that are used to turn the OP_TYPESTAR and friends opcodes +into others, under special conditions. A gap of 20 between the blocks should be +enough. The resulting opcodes don't have to be less than 256 because they are +never stored, so we push them well clear of the normal opcodes. */ + +#define OP_PROP_EXTRA 300 +#define OP_EXTUNI_EXTRA 320 +#define OP_ANYNL_EXTRA 340 +#define OP_HSPACE_EXTRA 360 +#define OP_VSPACE_EXTRA 380 + + +/* This table identifies those opcodes that are followed immediately by a +character that is to be tested in some way. This makes it possible to +centralize the loading of these characters. In the case of Type * etc, the +"character" is the opcode for \D, \d, \S, \s, \W, or \w, which will always be a +small value. Non-zero values in the table are the offsets from the opcode where +the character is to be found. ***NOTE*** If the start of this table is +modified, the three tables that follow must also be modified. */ + +static const pcre_uint8 coptable[] = { + 0, /* End */ + 0, 0, 0, 0, 0, /* \A, \G, \K, \B, \b */ + 0, 0, 0, 0, 0, 0, /* \D, \d, \S, \s, \W, \w */ + 0, 0, 0, /* Any, AllAny, Anybyte */ + 0, 0, /* \P, \p */ + 0, 0, 0, 0, 0, /* \R, \H, \h, \V, \v */ + 0, /* \X */ + 0, 0, 0, 0, 0, 0, /* \Z, \z, $, $M, ^, ^M */ + 1, /* Char */ + 1, /* Chari */ + 1, /* not */ + 1, /* noti */ + /* Positive single-char repeats */ + 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ + 1+IMM2_SIZE, 1+IMM2_SIZE, /* upto, minupto */ + 1+IMM2_SIZE, /* exact */ + 1, 1, 1, 1+IMM2_SIZE, /* *+, ++, ?+, upto+ */ + 1, 1, 1, 1, 1, 1, /* *I, *?I, +I, +?I, ?I, ??I */ + 1+IMM2_SIZE, 1+IMM2_SIZE, /* upto I, minupto I */ + 1+IMM2_SIZE, /* exact I */ + 1, 1, 1, 1+IMM2_SIZE, /* *+I, ++I, ?+I, upto+I */ + /* Negative single-char repeats - only for chars < 256 */ + 1, 1, 1, 1, 1, 1, /* NOT *, *?, +, +?, ?, ?? */ + 1+IMM2_SIZE, 1+IMM2_SIZE, /* NOT upto, minupto */ + 1+IMM2_SIZE, /* NOT exact */ + 1, 1, 1, 1+IMM2_SIZE, /* NOT *+, ++, ?+, upto+ */ + 1, 1, 1, 1, 1, 1, /* NOT *I, *?I, +I, +?I, ?I, ??I */ + 1+IMM2_SIZE, 1+IMM2_SIZE, /* NOT upto I, minupto I */ + 1+IMM2_SIZE, /* NOT exact I */ + 1, 1, 1, 1+IMM2_SIZE, /* NOT *+I, ++I, ?+I, upto+I */ + /* Positive type repeats */ + 1, 1, 1, 1, 1, 1, /* Type *, *?, +, +?, ?, ?? */ + 1+IMM2_SIZE, 1+IMM2_SIZE, /* Type upto, minupto */ + 1+IMM2_SIZE, /* Type exact */ + 1, 1, 1, 1+IMM2_SIZE, /* Type *+, ++, ?+, upto+ */ + /* Character class & ref repeats */ + 0, 0, 0, 0, 0, 0, /* *, *?, +, +?, ?, ?? */ + 0, 0, /* CRRANGE, CRMINRANGE */ + 0, 0, 0, 0, /* Possessive *+, ++, ?+, CRPOSRANGE */ + 0, /* CLASS */ + 0, /* NCLASS */ + 0, /* XCLASS - variable length */ + 0, /* REF */ + 0, /* REFI */ + 0, /* DNREF */ + 0, /* DNREFI */ + 0, /* RECURSE */ + 0, /* CALLOUT */ + 0, /* Alt */ + 0, /* Ket */ + 0, /* KetRmax */ + 0, /* KetRmin */ + 0, /* KetRpos */ + 0, /* Reverse */ + 0, /* Assert */ + 0, /* Assert not */ + 0, /* Assert behind */ + 0, /* Assert behind not */ + 0, 0, /* ONCE, ONCE_NC */ + 0, 0, 0, 0, 0, /* BRA, BRAPOS, CBRA, CBRAPOS, COND */ + 0, 0, 0, 0, 0, /* SBRA, SBRAPOS, SCBRA, SCBRAPOS, SCOND */ + 0, 0, /* CREF, DNCREF */ + 0, 0, /* RREF, DNRREF */ + 0, /* DEF */ + 0, 0, 0, /* BRAZERO, BRAMINZERO, BRAPOSZERO */ + 0, 0, 0, /* MARK, PRUNE, PRUNE_ARG */ + 0, 0, 0, 0, /* SKIP, SKIP_ARG, THEN, THEN_ARG */ + 0, 0, 0, 0, /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT */ + 0, 0 /* CLOSE, SKIPZERO */ +}; + +/* This table identifies those opcodes that inspect a character. It is used to +remember the fact that a character could have been inspected when the end of +the subject is reached. ***NOTE*** If the start of this table is modified, the +two tables that follow must also be modified. */ + +static const pcre_uint8 poptable[] = { + 0, /* End */ + 0, 0, 0, 1, 1, /* \A, \G, \K, \B, \b */ + 1, 1, 1, 1, 1, 1, /* \D, \d, \S, \s, \W, \w */ + 1, 1, 1, /* Any, AllAny, Anybyte */ + 1, 1, /* \P, \p */ + 1, 1, 1, 1, 1, /* \R, \H, \h, \V, \v */ + 1, /* \X */ + 0, 0, 0, 0, 0, 0, /* \Z, \z, $, $M, ^, ^M */ + 1, /* Char */ + 1, /* Chari */ + 1, /* not */ + 1, /* noti */ + /* Positive single-char repeats */ + 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ + 1, 1, 1, /* upto, minupto, exact */ + 1, 1, 1, 1, /* *+, ++, ?+, upto+ */ + 1, 1, 1, 1, 1, 1, /* *I, *?I, +I, +?I, ?I, ??I */ + 1, 1, 1, /* upto I, minupto I, exact I */ + 1, 1, 1, 1, /* *+I, ++I, ?+I, upto+I */ + /* Negative single-char repeats - only for chars < 256 */ + 1, 1, 1, 1, 1, 1, /* NOT *, *?, +, +?, ?, ?? */ + 1, 1, 1, /* NOT upto, minupto, exact */ + 1, 1, 1, 1, /* NOT *+, ++, ?+, upto+ */ + 1, 1, 1, 1, 1, 1, /* NOT *I, *?I, +I, +?I, ?I, ??I */ + 1, 1, 1, /* NOT upto I, minupto I, exact I */ + 1, 1, 1, 1, /* NOT *+I, ++I, ?+I, upto+I */ + /* Positive type repeats */ + 1, 1, 1, 1, 1, 1, /* Type *, *?, +, +?, ?, ?? */ + 1, 1, 1, /* Type upto, minupto, exact */ + 1, 1, 1, 1, /* Type *+, ++, ?+, upto+ */ + /* Character class & ref repeats */ + 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ + 1, 1, /* CRRANGE, CRMINRANGE */ + 1, 1, 1, 1, /* Possessive *+, ++, ?+, CRPOSRANGE */ + 1, /* CLASS */ + 1, /* NCLASS */ + 1, /* XCLASS - variable length */ + 0, /* REF */ + 0, /* REFI */ + 0, /* DNREF */ + 0, /* DNREFI */ + 0, /* RECURSE */ + 0, /* CALLOUT */ + 0, /* Alt */ + 0, /* Ket */ + 0, /* KetRmax */ + 0, /* KetRmin */ + 0, /* KetRpos */ + 0, /* Reverse */ + 0, /* Assert */ + 0, /* Assert not */ + 0, /* Assert behind */ + 0, /* Assert behind not */ + 0, 0, /* ONCE, ONCE_NC */ + 0, 0, 0, 0, 0, /* BRA, BRAPOS, CBRA, CBRAPOS, COND */ + 0, 0, 0, 0, 0, /* SBRA, SBRAPOS, SCBRA, SCBRAPOS, SCOND */ + 0, 0, /* CREF, DNCREF */ + 0, 0, /* RREF, DNRREF */ + 0, /* DEF */ + 0, 0, 0, /* BRAZERO, BRAMINZERO, BRAPOSZERO */ + 0, 0, 0, /* MARK, PRUNE, PRUNE_ARG */ + 0, 0, 0, 0, /* SKIP, SKIP_ARG, THEN, THEN_ARG */ + 0, 0, 0, 0, /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT */ + 0, 0 /* CLOSE, SKIPZERO */ +}; + +/* These 2 tables allow for compact code for testing for \D, \d, \S, \s, \W, +and \w */ + +static const pcre_uint8 toptable1[] = { + 0, 0, 0, 0, 0, 0, + ctype_digit, ctype_digit, + ctype_space, ctype_space, + ctype_word, ctype_word, + 0, 0 /* OP_ANY, OP_ALLANY */ +}; + +static const pcre_uint8 toptable2[] = { + 0, 0, 0, 0, 0, 0, + ctype_digit, 0, + ctype_space, 0, + ctype_word, 0, + 1, 1 /* OP_ANY, OP_ALLANY */ +}; + + +/* Structure for holding data about a particular state, which is in effect the +current data for an active path through the match tree. It must consist +entirely of ints because the working vector we are passed, and which we put +these structures in, is a vector of ints. */ + +typedef struct stateblock { + int offset; /* Offset to opcode */ + int count; /* Count for repeats */ + int data; /* Some use extra data */ +} stateblock; + +#define INTS_PER_STATEBLOCK (int)(sizeof(stateblock)/sizeof(int)) + + +#ifdef PCRE_DEBUG +/************************************************* +* Print character string * +*************************************************/ + +/* Character string printing function for debugging. + +Arguments: + p points to string + length number of bytes + f where to print + +Returns: nothing +*/ + +static void +pchars(const pcre_uchar *p, int length, FILE *f) +{ +pcre_uint32 c; +while (length-- > 0) + { + if (isprint(c = *(p++))) + fprintf(f, "%c", c); + else + fprintf(f, "\\x{%02x}", c); + } +} +#endif + + + +/************************************************* +* Execute a Regular Expression - DFA engine * +*************************************************/ + +/* This internal function applies a compiled pattern to a subject string, +starting at a given point, using a DFA engine. This function is called from the +external one, possibly multiple times if the pattern is not anchored. The +function calls itself recursively for some kinds of subpattern. + +Arguments: + md the match_data block with fixed information + this_start_code the opening bracket of this subexpression's code + current_subject where we currently are in the subject string + start_offset start offset in the subject string + offsets vector to contain the matching string offsets + offsetcount size of same + workspace vector of workspace + wscount size of same + rlevel function call recursion level + +Returns: > 0 => number of match offset pairs placed in offsets + = 0 => offsets overflowed; longest matches are present + -1 => failed to match + < -1 => some kind of unexpected problem + +The following macros are used for adding states to the two state vectors (one +for the current character, one for the following character). */ + +#define ADD_ACTIVE(x,y) \ + if (active_count++ < wscount) \ + { \ + next_active_state->offset = (x); \ + next_active_state->count = (y); \ + next_active_state++; \ + DPRINTF(("%.*sADD_ACTIVE(%d,%d)\n", rlevel*2-2, SP, (x), (y))); \ + } \ + else return PCRE_ERROR_DFA_WSSIZE + +#define ADD_ACTIVE_DATA(x,y,z) \ + if (active_count++ < wscount) \ + { \ + next_active_state->offset = (x); \ + next_active_state->count = (y); \ + next_active_state->data = (z); \ + next_active_state++; \ + DPRINTF(("%.*sADD_ACTIVE_DATA(%d,%d,%d)\n", rlevel*2-2, SP, (x), (y), (z))); \ + } \ + else return PCRE_ERROR_DFA_WSSIZE + +#define ADD_NEW(x,y) \ + if (new_count++ < wscount) \ + { \ + next_new_state->offset = (x); \ + next_new_state->count = (y); \ + next_new_state++; \ + DPRINTF(("%.*sADD_NEW(%d,%d)\n", rlevel*2-2, SP, (x), (y))); \ + } \ + else return PCRE_ERROR_DFA_WSSIZE + +#define ADD_NEW_DATA(x,y,z) \ + if (new_count++ < wscount) \ + { \ + next_new_state->offset = (x); \ + next_new_state->count = (y); \ + next_new_state->data = (z); \ + next_new_state++; \ + DPRINTF(("%.*sADD_NEW_DATA(%d,%d,%d) line %d\n", rlevel*2-2, SP, \ + (x), (y), (z), __LINE__)); \ + } \ + else return PCRE_ERROR_DFA_WSSIZE + +/* And now, here is the code */ + +static int +internal_dfa_exec( + dfa_match_data *md, + const pcre_uchar *this_start_code, + const pcre_uchar *current_subject, + int start_offset, + int *offsets, + int offsetcount, + int *workspace, + int wscount, + int rlevel) +{ +stateblock *active_states, *new_states, *temp_states; +stateblock *next_active_state, *next_new_state; + +const pcre_uint8 *ctypes, *lcc, *fcc; +const pcre_uchar *ptr; +const pcre_uchar *end_code, *first_op; + +dfa_recursion_info new_recursive; + +int active_count, new_count, match_count; + +/* Some fields in the md block are frequently referenced, so we load them into +independent variables in the hope that this will perform better. */ + +const pcre_uchar *start_subject = md->start_subject; +const pcre_uchar *end_subject = md->end_subject; +const pcre_uchar *start_code = md->start_code; + +#ifdef SUPPORT_UTF +BOOL utf = (md->poptions & PCRE_UTF8) != 0; +#else +BOOL utf = FALSE; +#endif + +BOOL reset_could_continue = FALSE; + +rlevel++; +offsetcount &= (-2); + +wscount -= 2; +wscount = (wscount - (wscount % (INTS_PER_STATEBLOCK * 2))) / + (2 * INTS_PER_STATEBLOCK); + +DPRINTF(("\n%.*s---------------------\n" + "%.*sCall to internal_dfa_exec f=%d\n", + rlevel*2-2, SP, rlevel*2-2, SP, rlevel)); + +ctypes = md->tables + ctypes_offset; +lcc = md->tables + lcc_offset; +fcc = md->tables + fcc_offset; + +match_count = PCRE_ERROR_NOMATCH; /* A negative number */ + +active_states = (stateblock *)(workspace + 2); +next_new_state = new_states = active_states + wscount; +new_count = 0; + +first_op = this_start_code + 1 + LINK_SIZE + + ((*this_start_code == OP_CBRA || *this_start_code == OP_SCBRA || + *this_start_code == OP_CBRAPOS || *this_start_code == OP_SCBRAPOS) + ? IMM2_SIZE:0); + +/* The first thing in any (sub) pattern is a bracket of some sort. Push all +the alternative states onto the list, and find out where the end is. This +makes is possible to use this function recursively, when we want to stop at a +matching internal ket rather than at the end. + +If the first opcode in the first alternative is OP_REVERSE, we are dealing with +a backward assertion. In that case, we have to find out the maximum amount to +move back, and set up each alternative appropriately. */ + +if (*first_op == OP_REVERSE) + { + int max_back = 0; + int gone_back; + + end_code = this_start_code; + do + { + int back = GET(end_code, 2+LINK_SIZE); + if (back > max_back) max_back = back; + end_code += GET(end_code, 1); + } + while (*end_code == OP_ALT); + + /* If we can't go back the amount required for the longest lookbehind + pattern, go back as far as we can; some alternatives may still be viable. */ + +#ifdef SUPPORT_UTF + /* In character mode we have to step back character by character */ + + if (utf) + { + for (gone_back = 0; gone_back < max_back; gone_back++) + { + if (current_subject <= start_subject) break; + current_subject--; + ACROSSCHAR(current_subject > start_subject, *current_subject, current_subject--); + } + } + else +#endif + + /* In byte-mode we can do this quickly. */ + + { + gone_back = (current_subject - max_back < start_subject)? + (int)(current_subject - start_subject) : max_back; + current_subject -= gone_back; + } + + /* Save the earliest consulted character */ + + if (current_subject < md->start_used_ptr) + md->start_used_ptr = current_subject; + + /* Now we can process the individual branches. */ + + end_code = this_start_code; + do + { + int back = GET(end_code, 2+LINK_SIZE); + if (back <= gone_back) + { + int bstate = (int)(end_code - start_code + 2 + 2*LINK_SIZE); + ADD_NEW_DATA(-bstate, 0, gone_back - back); + } + end_code += GET(end_code, 1); + } + while (*end_code == OP_ALT); + } + +/* This is the code for a "normal" subpattern (not a backward assertion). The +start of a whole pattern is always one of these. If we are at the top level, +we may be asked to restart matching from the same point that we reached for a +previous partial match. We still have to scan through the top-level branches to +find the end state. */ + +else + { + end_code = this_start_code; + + /* Restarting */ + + if (rlevel == 1 && (md->moptions & PCRE_DFA_RESTART) != 0) + { + do { end_code += GET(end_code, 1); } while (*end_code == OP_ALT); + new_count = workspace[1]; + if (!workspace[0]) + memcpy(new_states, active_states, new_count * sizeof(stateblock)); + } + + /* Not restarting */ + + else + { + int length = 1 + LINK_SIZE + + ((*this_start_code == OP_CBRA || *this_start_code == OP_SCBRA || + *this_start_code == OP_CBRAPOS || *this_start_code == OP_SCBRAPOS) + ? IMM2_SIZE:0); + do + { + ADD_NEW((int)(end_code - start_code + length), 0); + end_code += GET(end_code, 1); + length = 1 + LINK_SIZE; + } + while (*end_code == OP_ALT); + } + } + +workspace[0] = 0; /* Bit indicating which vector is current */ + +DPRINTF(("%.*sEnd state = %d\n", rlevel*2-2, SP, (int)(end_code - start_code))); + +/* Loop for scanning the subject */ + +ptr = current_subject; +for (;;) + { + int i, j; + int clen, dlen; + pcre_uint32 c, d; + int forced_fail = 0; + BOOL partial_newline = FALSE; + BOOL could_continue = reset_could_continue; + reset_could_continue = FALSE; + + /* Make the new state list into the active state list and empty the + new state list. */ + + temp_states = active_states; + active_states = new_states; + new_states = temp_states; + active_count = new_count; + new_count = 0; + + workspace[0] ^= 1; /* Remember for the restarting feature */ + workspace[1] = active_count; + +#ifdef PCRE_DEBUG + printf("%.*sNext character: rest of subject = \"", rlevel*2-2, SP); + pchars(ptr, STRLEN_UC(ptr), stdout); + printf("\"\n"); + + printf("%.*sActive states: ", rlevel*2-2, SP); + for (i = 0; i < active_count; i++) + printf("%d/%d ", active_states[i].offset, active_states[i].count); + printf("\n"); +#endif + + /* Set the pointers for adding new states */ + + next_active_state = active_states + active_count; + next_new_state = new_states; + + /* Load the current character from the subject outside the loop, as many + different states may want to look at it, and we assume that at least one + will. */ + + if (ptr < end_subject) + { + clen = 1; /* Number of data items in the character */ +#ifdef SUPPORT_UTF + GETCHARLENTEST(c, ptr, clen); +#else + c = *ptr; +#endif /* SUPPORT_UTF */ + } + else + { + clen = 0; /* This indicates the end of the subject */ + c = NOTACHAR; /* This value should never actually be used */ + } + + /* Scan up the active states and act on each one. The result of an action + may be to add more states to the currently active list (e.g. on hitting a + parenthesis) or it may be to put states on the new list, for considering + when we move the character pointer on. */ + + for (i = 0; i < active_count; i++) + { + stateblock *current_state = active_states + i; + BOOL caseless = FALSE; + const pcre_uchar *code; + int state_offset = current_state->offset; + int codevalue, rrc; + int count; + +#ifdef PCRE_DEBUG + printf ("%.*sProcessing state %d c=", rlevel*2-2, SP, state_offset); + if (clen == 0) printf("EOL\n"); + else if (c > 32 && c < 127) printf("'%c'\n", c); + else printf("0x%02x\n", c); +#endif + + /* A negative offset is a special case meaning "hold off going to this + (negated) state until the number of characters in the data field have + been skipped". If the could_continue flag was passed over from a previous + state, arrange for it to passed on. */ + + if (state_offset < 0) + { + if (current_state->data > 0) + { + DPRINTF(("%.*sSkipping this character\n", rlevel*2-2, SP)); + ADD_NEW_DATA(state_offset, current_state->count, + current_state->data - 1); + if (could_continue) reset_could_continue = TRUE; + continue; + } + else + { + current_state->offset = state_offset = -state_offset; + } + } + + /* Check for a duplicate state with the same count, and skip if found. + See the note at the head of this module about the possibility of improving + performance here. */ + + for (j = 0; j < i; j++) + { + if (active_states[j].offset == state_offset && + active_states[j].count == current_state->count) + { + DPRINTF(("%.*sDuplicate state: skipped\n", rlevel*2-2, SP)); + goto NEXT_ACTIVE_STATE; + } + } + + /* The state offset is the offset to the opcode */ + + code = start_code + state_offset; + codevalue = *code; + + /* If this opcode inspects a character, but we are at the end of the + subject, remember the fact for use when testing for a partial match. */ + + if (clen == 0 && poptable[codevalue] != 0) + could_continue = TRUE; + + /* If this opcode is followed by an inline character, load it. It is + tempting to test for the presence of a subject character here, but that + is wrong, because sometimes zero repetitions of the subject are + permitted. + + We also use this mechanism for opcodes such as OP_TYPEPLUS that take an + argument that is not a data character - but is always one byte long because + the values are small. We have to take special action to deal with \P, \p, + \H, \h, \V, \v and \X in this case. To keep the other cases fast, convert + these ones to new opcodes. */ + + if (coptable[codevalue] > 0) + { + dlen = 1; +#ifdef SUPPORT_UTF + if (utf) { GETCHARLEN(d, (code + coptable[codevalue]), dlen); } else +#endif /* SUPPORT_UTF */ + d = code[coptable[codevalue]]; + if (codevalue >= OP_TYPESTAR) + { + switch(d) + { + case OP_ANYBYTE: return PCRE_ERROR_DFA_UITEM; + case OP_NOTPROP: + case OP_PROP: codevalue += OP_PROP_EXTRA; break; + case OP_ANYNL: codevalue += OP_ANYNL_EXTRA; break; + case OP_EXTUNI: codevalue += OP_EXTUNI_EXTRA; break; + case OP_NOT_HSPACE: + case OP_HSPACE: codevalue += OP_HSPACE_EXTRA; break; + case OP_NOT_VSPACE: + case OP_VSPACE: codevalue += OP_VSPACE_EXTRA; break; + default: break; + } + } + } + else + { + dlen = 0; /* Not strictly necessary, but compilers moan */ + d = NOTACHAR; /* if these variables are not set. */ + } + + + /* Now process the individual opcodes */ + + switch (codevalue) + { +/* ========================================================================== */ + /* These cases are never obeyed. This is a fudge that causes a compile- + time error if the vectors coptable or poptable, which are indexed by + opcode, are not the correct length. It seems to be the only way to do + such a check at compile time, as the sizeof() operator does not work + in the C preprocessor. */ + + case OP_TABLE_LENGTH: + case OP_TABLE_LENGTH + + ((sizeof(coptable) == OP_TABLE_LENGTH) && + (sizeof(poptable) == OP_TABLE_LENGTH)): + break; + +/* ========================================================================== */ + /* Reached a closing bracket. If not at the end of the pattern, carry + on with the next opcode. For repeating opcodes, also add the repeat + state. Note that KETRPOS will always be encountered at the end of the + subpattern, because the possessive subpattern repeats are always handled + using recursive calls. Thus, it never adds any new states. + + At the end of the (sub)pattern, unless we have an empty string and + PCRE_NOTEMPTY is set, or PCRE_NOTEMPTY_ATSTART is set and we are at the + start of the subject, save the match data, shifting up all previous + matches so we always have the longest first. */ + + case OP_KET: + case OP_KETRMIN: + case OP_KETRMAX: + case OP_KETRPOS: + if (code != end_code) + { + ADD_ACTIVE(state_offset + 1 + LINK_SIZE, 0); + if (codevalue != OP_KET) + { + ADD_ACTIVE(state_offset - GET(code, 1), 0); + } + } + else + { + if (ptr > current_subject || + ((md->moptions & PCRE_NOTEMPTY) == 0 && + ((md->moptions & PCRE_NOTEMPTY_ATSTART) == 0 || + current_subject > start_subject + md->start_offset))) + { + if (match_count < 0) match_count = (offsetcount >= 2)? 1 : 0; + else if (match_count > 0 && ++match_count * 2 > offsetcount) + match_count = 0; + count = ((match_count == 0)? offsetcount : match_count * 2) - 2; + if (count > 0) memmove(offsets + 2, offsets, count * sizeof(int)); + if (offsetcount >= 2) + { + offsets[0] = (int)(current_subject - start_subject); + offsets[1] = (int)(ptr - start_subject); + DPRINTF(("%.*sSet matched string = \"%.*s\"\n", rlevel*2-2, SP, + offsets[1] - offsets[0], (char *)current_subject)); + } + if ((md->moptions & PCRE_DFA_SHORTEST) != 0) + { + DPRINTF(("%.*sEnd of internal_dfa_exec %d: returning %d\n" + "%.*s---------------------\n\n", rlevel*2-2, SP, rlevel, + match_count, rlevel*2-2, SP)); + return match_count; + } + } + } + break; + +/* ========================================================================== */ + /* These opcodes add to the current list of states without looking + at the current character. */ + + /*-----------------------------------------------------------------*/ + case OP_ALT: + do { code += GET(code, 1); } while (*code == OP_ALT); + ADD_ACTIVE((int)(code - start_code), 0); + break; + + /*-----------------------------------------------------------------*/ + case OP_BRA: + case OP_SBRA: + do + { + ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE), 0); + code += GET(code, 1); + } + while (*code == OP_ALT); + break; + + /*-----------------------------------------------------------------*/ + case OP_CBRA: + case OP_SCBRA: + ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE + IMM2_SIZE), 0); + code += GET(code, 1); + while (*code == OP_ALT) + { + ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE), 0); + code += GET(code, 1); + } + break; + + /*-----------------------------------------------------------------*/ + case OP_BRAZERO: + case OP_BRAMINZERO: + ADD_ACTIVE(state_offset + 1, 0); + code += 1 + GET(code, 2); + while (*code == OP_ALT) code += GET(code, 1); + ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE), 0); + break; + + /*-----------------------------------------------------------------*/ + case OP_SKIPZERO: + code += 1 + GET(code, 2); + while (*code == OP_ALT) code += GET(code, 1); + ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE), 0); + break; + + /*-----------------------------------------------------------------*/ + case OP_CIRC: + if (ptr == start_subject && (md->moptions & PCRE_NOTBOL) == 0) + { ADD_ACTIVE(state_offset + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + case OP_CIRCM: + if ((ptr == start_subject && (md->moptions & PCRE_NOTBOL) == 0) || + (ptr != end_subject && WAS_NEWLINE(ptr))) + { ADD_ACTIVE(state_offset + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + case OP_EOD: + if (ptr >= end_subject) + { + if ((md->moptions & PCRE_PARTIAL_HARD) != 0) + could_continue = TRUE; + else { ADD_ACTIVE(state_offset + 1, 0); } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_SOD: + if (ptr == start_subject) { ADD_ACTIVE(state_offset + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + case OP_SOM: + if (ptr == start_subject + start_offset) { ADD_ACTIVE(state_offset + 1, 0); } + break; + + +/* ========================================================================== */ + /* These opcodes inspect the next subject character, and sometimes + the previous one as well, but do not have an argument. The variable + clen contains the length of the current character and is zero if we are + at the end of the subject. */ + + /*-----------------------------------------------------------------*/ + case OP_ANY: + if (clen > 0 && !IS_NEWLINE(ptr)) + { + if (ptr + 1 >= md->end_subject && + (md->moptions & (PCRE_PARTIAL_HARD)) != 0 && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + could_continue = partial_newline = TRUE; + } + else + { + ADD_NEW(state_offset + 1, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_ALLANY: + if (clen > 0) + { ADD_NEW(state_offset + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + case OP_EODN: + if (clen == 0 && (md->moptions & PCRE_PARTIAL_HARD) != 0) + could_continue = TRUE; + else if (clen == 0 || (IS_NEWLINE(ptr) && ptr == end_subject - md->nllen)) + { ADD_ACTIVE(state_offset + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + case OP_DOLL: + if ((md->moptions & PCRE_NOTEOL) == 0) + { + if (clen == 0 && (md->moptions & PCRE_PARTIAL_HARD) != 0) + could_continue = TRUE; + else if (clen == 0 || + ((md->poptions & PCRE_DOLLAR_ENDONLY) == 0 && IS_NEWLINE(ptr) && + (ptr == end_subject - md->nllen) + )) + { ADD_ACTIVE(state_offset + 1, 0); } + else if (ptr + 1 >= md->end_subject && + (md->moptions & (PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT)) != 0 && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + if ((md->moptions & PCRE_PARTIAL_HARD) != 0) + { + reset_could_continue = TRUE; + ADD_NEW_DATA(-(state_offset + 1), 0, 1); + } + else could_continue = partial_newline = TRUE; + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_DOLLM: + if ((md->moptions & PCRE_NOTEOL) == 0) + { + if (clen == 0 && (md->moptions & PCRE_PARTIAL_HARD) != 0) + could_continue = TRUE; + else if (clen == 0 || + ((md->poptions & PCRE_DOLLAR_ENDONLY) == 0 && IS_NEWLINE(ptr))) + { ADD_ACTIVE(state_offset + 1, 0); } + else if (ptr + 1 >= md->end_subject && + (md->moptions & (PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT)) != 0 && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + if ((md->moptions & PCRE_PARTIAL_HARD) != 0) + { + reset_could_continue = TRUE; + ADD_NEW_DATA(-(state_offset + 1), 0, 1); + } + else could_continue = partial_newline = TRUE; + } + } + else if (IS_NEWLINE(ptr)) + { ADD_ACTIVE(state_offset + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + + case OP_DIGIT: + case OP_WHITESPACE: + case OP_WORDCHAR: + if (clen > 0 && c < 256 && + ((ctypes[c] & toptable1[codevalue]) ^ toptable2[codevalue]) != 0) + { ADD_NEW(state_offset + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + case OP_NOT_DIGIT: + case OP_NOT_WHITESPACE: + case OP_NOT_WORDCHAR: + if (clen > 0 && (c >= 256 || + ((ctypes[c] & toptable1[codevalue]) ^ toptable2[codevalue]) != 0)) + { ADD_NEW(state_offset + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + case OP_WORD_BOUNDARY: + case OP_NOT_WORD_BOUNDARY: + { + int left_word, right_word; + + if (ptr > start_subject) + { + const pcre_uchar *temp = ptr - 1; + if (temp < md->start_used_ptr) md->start_used_ptr = temp; +#if defined SUPPORT_UTF && !defined COMPILE_PCRE32 + if (utf) { BACKCHAR(temp); } +#endif + GETCHARTEST(d, temp); +#ifdef SUPPORT_UCP + if ((md->poptions & PCRE_UCP) != 0) + { + if (d == '_') left_word = TRUE; else + { + int cat = UCD_CATEGORY(d); + left_word = (cat == ucp_L || cat == ucp_N); + } + } + else +#endif + left_word = d < 256 && (ctypes[d] & ctype_word) != 0; + } + else left_word = FALSE; + + if (clen > 0) + { +#ifdef SUPPORT_UCP + if ((md->poptions & PCRE_UCP) != 0) + { + if (c == '_') right_word = TRUE; else + { + int cat = UCD_CATEGORY(c); + right_word = (cat == ucp_L || cat == ucp_N); + } + } + else +#endif + right_word = c < 256 && (ctypes[c] & ctype_word) != 0; + } + else right_word = FALSE; + + if ((left_word == right_word) == (codevalue == OP_NOT_WORD_BOUNDARY)) + { ADD_ACTIVE(state_offset + 1, 0); } + } + break; + + + /*-----------------------------------------------------------------*/ + /* Check the next character by Unicode property. We will get here only + if the support is in the binary; otherwise a compile-time error occurs. + */ + +#ifdef SUPPORT_UCP + case OP_PROP: + case OP_NOTPROP: + if (clen > 0) + { + BOOL OK; + const pcre_uint32 *cp; + const ucd_record * prop = GET_UCD(c); + switch(code[1]) + { + case PT_ANY: + OK = TRUE; + break; + + case PT_LAMP: + OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || + prop->chartype == ucp_Lt; + break; + + case PT_GC: + OK = PRIV(ucp_gentype)[prop->chartype] == code[2]; + break; + + case PT_PC: + OK = prop->chartype == code[2]; + break; + + case PT_SC: + OK = prop->script == code[2]; + break; + + /* These are specials for combination cases. */ + + case PT_ALNUM: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N; + break; + + /* Perl space used to exclude VT, but from Perl 5.18 it is included, + which means that Perl space and POSIX space are now identical. PCRE + was changed at release 8.34. */ + + case PT_SPACE: /* Perl space */ + case PT_PXSPACE: /* POSIX space */ + switch(c) + { + HSPACE_CASES: + VSPACE_CASES: + OK = TRUE; + break; + + default: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z; + break; + } + break; + + case PT_WORD: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N || + c == CHAR_UNDERSCORE; + break; + + case PT_CLIST: + cp = PRIV(ucd_caseless_sets) + code[2]; + for (;;) + { + if (c < *cp) { OK = FALSE; break; } + if (c == *cp++) { OK = TRUE; break; } + } + break; + + case PT_UCNC: + OK = c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT || + c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) || + c >= 0xe000; + break; + + /* Should never occur, but keep compilers from grumbling. */ + + default: + OK = codevalue != OP_PROP; + break; + } + + if (OK == (codevalue == OP_PROP)) { ADD_NEW(state_offset + 3, 0); } + } + break; +#endif + + + +/* ========================================================================== */ + /* These opcodes likewise inspect the subject character, but have an + argument that is not a data character. It is one of these opcodes: + OP_ANY, OP_ALLANY, OP_DIGIT, OP_NOT_DIGIT, OP_WHITESPACE, OP_NOT_SPACE, + OP_WORDCHAR, OP_NOT_WORDCHAR. The value is loaded into d. */ + + case OP_TYPEPLUS: + case OP_TYPEMINPLUS: + case OP_TYPEPOSPLUS: + count = current_state->count; /* Already matched */ + if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } + if (clen > 0) + { + if (d == OP_ANY && ptr + 1 >= md->end_subject && + (md->moptions & (PCRE_PARTIAL_HARD)) != 0 && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + could_continue = partial_newline = TRUE; + } + else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || + (c < 256 && + (d != OP_ANY || !IS_NEWLINE(ptr)) && + ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) + { + if (count > 0 && codevalue == OP_TYPEPOSPLUS) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + count++; + ADD_NEW(state_offset, count); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_TYPEQUERY: + case OP_TYPEMINQUERY: + case OP_TYPEPOSQUERY: + ADD_ACTIVE(state_offset + 2, 0); + if (clen > 0) + { + if (d == OP_ANY && ptr + 1 >= md->end_subject && + (md->moptions & (PCRE_PARTIAL_HARD)) != 0 && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + could_continue = partial_newline = TRUE; + } + else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || + (c < 256 && + (d != OP_ANY || !IS_NEWLINE(ptr)) && + ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) + { + if (codevalue == OP_TYPEPOSQUERY) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + ADD_NEW(state_offset + 2, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_TYPESTAR: + case OP_TYPEMINSTAR: + case OP_TYPEPOSSTAR: + ADD_ACTIVE(state_offset + 2, 0); + if (clen > 0) + { + if (d == OP_ANY && ptr + 1 >= md->end_subject && + (md->moptions & (PCRE_PARTIAL_HARD)) != 0 && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + could_continue = partial_newline = TRUE; + } + else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || + (c < 256 && + (d != OP_ANY || !IS_NEWLINE(ptr)) && + ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) + { + if (codevalue == OP_TYPEPOSSTAR) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + ADD_NEW(state_offset, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_TYPEEXACT: + count = current_state->count; /* Number already matched */ + if (clen > 0) + { + if (d == OP_ANY && ptr + 1 >= md->end_subject && + (md->moptions & (PCRE_PARTIAL_HARD)) != 0 && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + could_continue = partial_newline = TRUE; + } + else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || + (c < 256 && + (d != OP_ANY || !IS_NEWLINE(ptr)) && + ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) + { + if (++count >= (int)GET2(code, 1)) + { ADD_NEW(state_offset + 1 + IMM2_SIZE + 1, 0); } + else + { ADD_NEW(state_offset, count); } + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_TYPEUPTO: + case OP_TYPEMINUPTO: + case OP_TYPEPOSUPTO: + ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0); + count = current_state->count; /* Number already matched */ + if (clen > 0) + { + if (d == OP_ANY && ptr + 1 >= md->end_subject && + (md->moptions & (PCRE_PARTIAL_HARD)) != 0 && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + could_continue = partial_newline = TRUE; + } + else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || + (c < 256 && + (d != OP_ANY || !IS_NEWLINE(ptr)) && + ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) + { + if (codevalue == OP_TYPEPOSUPTO) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + if (++count >= (int)GET2(code, 1)) + { ADD_NEW(state_offset + 2 + IMM2_SIZE, 0); } + else + { ADD_NEW(state_offset, count); } + } + } + break; + +/* ========================================================================== */ + /* These are virtual opcodes that are used when something like + OP_TYPEPLUS has OP_PROP, OP_NOTPROP, OP_ANYNL, or OP_EXTUNI as its + argument. It keeps the code above fast for the other cases. The argument + is in the d variable. */ + +#ifdef SUPPORT_UCP + case OP_PROP_EXTRA + OP_TYPEPLUS: + case OP_PROP_EXTRA + OP_TYPEMINPLUS: + case OP_PROP_EXTRA + OP_TYPEPOSPLUS: + count = current_state->count; /* Already matched */ + if (count > 0) { ADD_ACTIVE(state_offset + 4, 0); } + if (clen > 0) + { + BOOL OK; + const pcre_uint32 *cp; + const ucd_record * prop = GET_UCD(c); + switch(code[2]) + { + case PT_ANY: + OK = TRUE; + break; + + case PT_LAMP: + OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || + prop->chartype == ucp_Lt; + break; + + case PT_GC: + OK = PRIV(ucp_gentype)[prop->chartype] == code[3]; + break; + + case PT_PC: + OK = prop->chartype == code[3]; + break; + + case PT_SC: + OK = prop->script == code[3]; + break; + + /* These are specials for combination cases. */ + + case PT_ALNUM: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N; + break; + + /* Perl space used to exclude VT, but from Perl 5.18 it is included, + which means that Perl space and POSIX space are now identical. PCRE + was changed at release 8.34. */ + + case PT_SPACE: /* Perl space */ + case PT_PXSPACE: /* POSIX space */ + switch(c) + { + HSPACE_CASES: + VSPACE_CASES: + OK = TRUE; + break; + + default: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z; + break; + } + break; + + case PT_WORD: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N || + c == CHAR_UNDERSCORE; + break; + + case PT_CLIST: + cp = PRIV(ucd_caseless_sets) + code[3]; + for (;;) + { + if (c < *cp) { OK = FALSE; break; } + if (c == *cp++) { OK = TRUE; break; } + } + break; + + case PT_UCNC: + OK = c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT || + c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) || + c >= 0xe000; + break; + + /* Should never occur, but keep compilers from grumbling. */ + + default: + OK = codevalue != OP_PROP; + break; + } + + if (OK == (d == OP_PROP)) + { + if (count > 0 && codevalue == OP_PROP_EXTRA + OP_TYPEPOSPLUS) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + count++; + ADD_NEW(state_offset, count); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_EXTUNI_EXTRA + OP_TYPEPLUS: + case OP_EXTUNI_EXTRA + OP_TYPEMINPLUS: + case OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS: + count = current_state->count; /* Already matched */ + if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } + if (clen > 0) + { + int lgb, rgb; + const pcre_uchar *nptr = ptr + clen; + int ncount = 0; + if (count > 0 && codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + lgb = UCD_GRAPHBREAK(c); + while (nptr < end_subject) + { + dlen = 1; + if (!utf) d = *nptr; else { GETCHARLEN(d, nptr, dlen); } + rgb = UCD_GRAPHBREAK(d); + if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break; + ncount++; + lgb = rgb; + nptr += dlen; + } + count++; + ADD_NEW_DATA(-state_offset, count, ncount); + } + break; +#endif + + /*-----------------------------------------------------------------*/ + case OP_ANYNL_EXTRA + OP_TYPEPLUS: + case OP_ANYNL_EXTRA + OP_TYPEMINPLUS: + case OP_ANYNL_EXTRA + OP_TYPEPOSPLUS: + count = current_state->count; /* Already matched */ + if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } + if (clen > 0) + { + int ncount = 0; + switch (c) + { + case CHAR_VT: + case CHAR_FF: + case CHAR_NEL: +#ifndef EBCDIC + case 0x2028: + case 0x2029: +#endif /* Not EBCDIC */ + if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break; + goto ANYNL01; + + case CHAR_CR: + if (ptr + 1 < end_subject && UCHAR21TEST(ptr + 1) == CHAR_LF) ncount = 1; + /* Fall through */ + + ANYNL01: + case CHAR_LF: + if (count > 0 && codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSPLUS) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + count++; + ADD_NEW_DATA(-state_offset, count, ncount); + break; + + default: + break; + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_VSPACE_EXTRA + OP_TYPEPLUS: + case OP_VSPACE_EXTRA + OP_TYPEMINPLUS: + case OP_VSPACE_EXTRA + OP_TYPEPOSPLUS: + count = current_state->count; /* Already matched */ + if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } + if (clen > 0) + { + BOOL OK; + switch (c) + { + VSPACE_CASES: + OK = TRUE; + break; + + default: + OK = FALSE; + break; + } + + if (OK == (d == OP_VSPACE)) + { + if (count > 0 && codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSPLUS) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + count++; + ADD_NEW_DATA(-state_offset, count, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_HSPACE_EXTRA + OP_TYPEPLUS: + case OP_HSPACE_EXTRA + OP_TYPEMINPLUS: + case OP_HSPACE_EXTRA + OP_TYPEPOSPLUS: + count = current_state->count; /* Already matched */ + if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } + if (clen > 0) + { + BOOL OK; + switch (c) + { + HSPACE_CASES: + OK = TRUE; + break; + + default: + OK = FALSE; + break; + } + + if (OK == (d == OP_HSPACE)) + { + if (count > 0 && codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSPLUS) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + count++; + ADD_NEW_DATA(-state_offset, count, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ +#ifdef SUPPORT_UCP + case OP_PROP_EXTRA + OP_TYPEQUERY: + case OP_PROP_EXTRA + OP_TYPEMINQUERY: + case OP_PROP_EXTRA + OP_TYPEPOSQUERY: + count = 4; + goto QS1; + + case OP_PROP_EXTRA + OP_TYPESTAR: + case OP_PROP_EXTRA + OP_TYPEMINSTAR: + case OP_PROP_EXTRA + OP_TYPEPOSSTAR: + count = 0; + + QS1: + + ADD_ACTIVE(state_offset + 4, 0); + if (clen > 0) + { + BOOL OK; + const pcre_uint32 *cp; + const ucd_record * prop = GET_UCD(c); + switch(code[2]) + { + case PT_ANY: + OK = TRUE; + break; + + case PT_LAMP: + OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || + prop->chartype == ucp_Lt; + break; + + case PT_GC: + OK = PRIV(ucp_gentype)[prop->chartype] == code[3]; + break; + + case PT_PC: + OK = prop->chartype == code[3]; + break; + + case PT_SC: + OK = prop->script == code[3]; + break; + + /* These are specials for combination cases. */ + + case PT_ALNUM: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N; + break; + + /* Perl space used to exclude VT, but from Perl 5.18 it is included, + which means that Perl space and POSIX space are now identical. PCRE + was changed at release 8.34. */ + + case PT_SPACE: /* Perl space */ + case PT_PXSPACE: /* POSIX space */ + switch(c) + { + HSPACE_CASES: + VSPACE_CASES: + OK = TRUE; + break; + + default: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z; + break; + } + break; + + case PT_WORD: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N || + c == CHAR_UNDERSCORE; + break; + + case PT_CLIST: + cp = PRIV(ucd_caseless_sets) + code[3]; + for (;;) + { + if (c < *cp) { OK = FALSE; break; } + if (c == *cp++) { OK = TRUE; break; } + } + break; + + case PT_UCNC: + OK = c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT || + c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) || + c >= 0xe000; + break; + + /* Should never occur, but keep compilers from grumbling. */ + + default: + OK = codevalue != OP_PROP; + break; + } + + if (OK == (d == OP_PROP)) + { + if (codevalue == OP_PROP_EXTRA + OP_TYPEPOSSTAR || + codevalue == OP_PROP_EXTRA + OP_TYPEPOSQUERY) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + ADD_NEW(state_offset + count, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_EXTUNI_EXTRA + OP_TYPEQUERY: + case OP_EXTUNI_EXTRA + OP_TYPEMINQUERY: + case OP_EXTUNI_EXTRA + OP_TYPEPOSQUERY: + count = 2; + goto QS2; + + case OP_EXTUNI_EXTRA + OP_TYPESTAR: + case OP_EXTUNI_EXTRA + OP_TYPEMINSTAR: + case OP_EXTUNI_EXTRA + OP_TYPEPOSSTAR: + count = 0; + + QS2: + + ADD_ACTIVE(state_offset + 2, 0); + if (clen > 0) + { + int lgb, rgb; + const pcre_uchar *nptr = ptr + clen; + int ncount = 0; + if (codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSSTAR || + codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSQUERY) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + lgb = UCD_GRAPHBREAK(c); + while (nptr < end_subject) + { + dlen = 1; + if (!utf) d = *nptr; else { GETCHARLEN(d, nptr, dlen); } + rgb = UCD_GRAPHBREAK(d); + if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break; + ncount++; + lgb = rgb; + nptr += dlen; + } + ADD_NEW_DATA(-(state_offset + count), 0, ncount); + } + break; +#endif + + /*-----------------------------------------------------------------*/ + case OP_ANYNL_EXTRA + OP_TYPEQUERY: + case OP_ANYNL_EXTRA + OP_TYPEMINQUERY: + case OP_ANYNL_EXTRA + OP_TYPEPOSQUERY: + count = 2; + goto QS3; + + case OP_ANYNL_EXTRA + OP_TYPESTAR: + case OP_ANYNL_EXTRA + OP_TYPEMINSTAR: + case OP_ANYNL_EXTRA + OP_TYPEPOSSTAR: + count = 0; + + QS3: + ADD_ACTIVE(state_offset + 2, 0); + if (clen > 0) + { + int ncount = 0; + switch (c) + { + case CHAR_VT: + case CHAR_FF: + case CHAR_NEL: +#ifndef EBCDIC + case 0x2028: + case 0x2029: +#endif /* Not EBCDIC */ + if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break; + goto ANYNL02; + + case CHAR_CR: + if (ptr + 1 < end_subject && UCHAR21TEST(ptr + 1) == CHAR_LF) ncount = 1; + /* Fall through */ + + ANYNL02: + case CHAR_LF: + if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSSTAR || + codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSQUERY) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + ADD_NEW_DATA(-(state_offset + (int)count), 0, ncount); + break; + + default: + break; + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_VSPACE_EXTRA + OP_TYPEQUERY: + case OP_VSPACE_EXTRA + OP_TYPEMINQUERY: + case OP_VSPACE_EXTRA + OP_TYPEPOSQUERY: + count = 2; + goto QS4; + + case OP_VSPACE_EXTRA + OP_TYPESTAR: + case OP_VSPACE_EXTRA + OP_TYPEMINSTAR: + case OP_VSPACE_EXTRA + OP_TYPEPOSSTAR: + count = 0; + + QS4: + ADD_ACTIVE(state_offset + 2, 0); + if (clen > 0) + { + BOOL OK; + switch (c) + { + VSPACE_CASES: + OK = TRUE; + break; + + default: + OK = FALSE; + break; + } + if (OK == (d == OP_VSPACE)) + { + if (codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSSTAR || + codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSQUERY) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + ADD_NEW_DATA(-(state_offset + (int)count), 0, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_HSPACE_EXTRA + OP_TYPEQUERY: + case OP_HSPACE_EXTRA + OP_TYPEMINQUERY: + case OP_HSPACE_EXTRA + OP_TYPEPOSQUERY: + count = 2; + goto QS5; + + case OP_HSPACE_EXTRA + OP_TYPESTAR: + case OP_HSPACE_EXTRA + OP_TYPEMINSTAR: + case OP_HSPACE_EXTRA + OP_TYPEPOSSTAR: + count = 0; + + QS5: + ADD_ACTIVE(state_offset + 2, 0); + if (clen > 0) + { + BOOL OK; + switch (c) + { + HSPACE_CASES: + OK = TRUE; + break; + + default: + OK = FALSE; + break; + } + + if (OK == (d == OP_HSPACE)) + { + if (codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSSTAR || + codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSQUERY) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + ADD_NEW_DATA(-(state_offset + (int)count), 0, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ +#ifdef SUPPORT_UCP + case OP_PROP_EXTRA + OP_TYPEEXACT: + case OP_PROP_EXTRA + OP_TYPEUPTO: + case OP_PROP_EXTRA + OP_TYPEMINUPTO: + case OP_PROP_EXTRA + OP_TYPEPOSUPTO: + if (codevalue != OP_PROP_EXTRA + OP_TYPEEXACT) + { ADD_ACTIVE(state_offset + 1 + IMM2_SIZE + 3, 0); } + count = current_state->count; /* Number already matched */ + if (clen > 0) + { + BOOL OK; + const pcre_uint32 *cp; + const ucd_record * prop = GET_UCD(c); + switch(code[1 + IMM2_SIZE + 1]) + { + case PT_ANY: + OK = TRUE; + break; + + case PT_LAMP: + OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || + prop->chartype == ucp_Lt; + break; + + case PT_GC: + OK = PRIV(ucp_gentype)[prop->chartype] == code[1 + IMM2_SIZE + 2]; + break; + + case PT_PC: + OK = prop->chartype == code[1 + IMM2_SIZE + 2]; + break; + + case PT_SC: + OK = prop->script == code[1 + IMM2_SIZE + 2]; + break; + + /* These are specials for combination cases. */ + + case PT_ALNUM: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N; + break; + + /* Perl space used to exclude VT, but from Perl 5.18 it is included, + which means that Perl space and POSIX space are now identical. PCRE + was changed at release 8.34. */ + + case PT_SPACE: /* Perl space */ + case PT_PXSPACE: /* POSIX space */ + switch(c) + { + HSPACE_CASES: + VSPACE_CASES: + OK = TRUE; + break; + + default: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z; + break; + } + break; + + case PT_WORD: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N || + c == CHAR_UNDERSCORE; + break; + + case PT_CLIST: + cp = PRIV(ucd_caseless_sets) + code[1 + IMM2_SIZE + 2]; + for (;;) + { + if (c < *cp) { OK = FALSE; break; } + if (c == *cp++) { OK = TRUE; break; } + } + break; + + case PT_UCNC: + OK = c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT || + c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) || + c >= 0xe000; + break; + + /* Should never occur, but keep compilers from grumbling. */ + + default: + OK = codevalue != OP_PROP; + break; + } + + if (OK == (d == OP_PROP)) + { + if (codevalue == OP_PROP_EXTRA + OP_TYPEPOSUPTO) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + if (++count >= (int)GET2(code, 1)) + { ADD_NEW(state_offset + 1 + IMM2_SIZE + 3, 0); } + else + { ADD_NEW(state_offset, count); } + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_EXTUNI_EXTRA + OP_TYPEEXACT: + case OP_EXTUNI_EXTRA + OP_TYPEUPTO: + case OP_EXTUNI_EXTRA + OP_TYPEMINUPTO: + case OP_EXTUNI_EXTRA + OP_TYPEPOSUPTO: + if (codevalue != OP_EXTUNI_EXTRA + OP_TYPEEXACT) + { ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0); } + count = current_state->count; /* Number already matched */ + if (clen > 0) + { + int lgb, rgb; + const pcre_uchar *nptr = ptr + clen; + int ncount = 0; + if (codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSUPTO) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + lgb = UCD_GRAPHBREAK(c); + while (nptr < end_subject) + { + dlen = 1; + if (!utf) d = *nptr; else { GETCHARLEN(d, nptr, dlen); } + rgb = UCD_GRAPHBREAK(d); + if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break; + ncount++; + lgb = rgb; + nptr += dlen; + } + if (nptr >= end_subject && (md->moptions & PCRE_PARTIAL_HARD) != 0) + reset_could_continue = TRUE; + if (++count >= (int)GET2(code, 1)) + { ADD_NEW_DATA(-(state_offset + 2 + IMM2_SIZE), 0, ncount); } + else + { ADD_NEW_DATA(-state_offset, count, ncount); } + } + break; +#endif + + /*-----------------------------------------------------------------*/ + case OP_ANYNL_EXTRA + OP_TYPEEXACT: + case OP_ANYNL_EXTRA + OP_TYPEUPTO: + case OP_ANYNL_EXTRA + OP_TYPEMINUPTO: + case OP_ANYNL_EXTRA + OP_TYPEPOSUPTO: + if (codevalue != OP_ANYNL_EXTRA + OP_TYPEEXACT) + { ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0); } + count = current_state->count; /* Number already matched */ + if (clen > 0) + { + int ncount = 0; + switch (c) + { + case CHAR_VT: + case CHAR_FF: + case CHAR_NEL: +#ifndef EBCDIC + case 0x2028: + case 0x2029: +#endif /* Not EBCDIC */ + if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break; + goto ANYNL03; + + case CHAR_CR: + if (ptr + 1 < end_subject && UCHAR21TEST(ptr + 1) == CHAR_LF) ncount = 1; + /* Fall through */ + + ANYNL03: + case CHAR_LF: + if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSUPTO) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + if (++count >= (int)GET2(code, 1)) + { ADD_NEW_DATA(-(state_offset + 2 + IMM2_SIZE), 0, ncount); } + else + { ADD_NEW_DATA(-state_offset, count, ncount); } + break; + + default: + break; + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_VSPACE_EXTRA + OP_TYPEEXACT: + case OP_VSPACE_EXTRA + OP_TYPEUPTO: + case OP_VSPACE_EXTRA + OP_TYPEMINUPTO: + case OP_VSPACE_EXTRA + OP_TYPEPOSUPTO: + if (codevalue != OP_VSPACE_EXTRA + OP_TYPEEXACT) + { ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0); } + count = current_state->count; /* Number already matched */ + if (clen > 0) + { + BOOL OK; + switch (c) + { + VSPACE_CASES: + OK = TRUE; + break; + + default: + OK = FALSE; + } + + if (OK == (d == OP_VSPACE)) + { + if (codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSUPTO) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + if (++count >= (int)GET2(code, 1)) + { ADD_NEW_DATA(-(state_offset + 2 + IMM2_SIZE), 0, 0); } + else + { ADD_NEW_DATA(-state_offset, count, 0); } + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_HSPACE_EXTRA + OP_TYPEEXACT: + case OP_HSPACE_EXTRA + OP_TYPEUPTO: + case OP_HSPACE_EXTRA + OP_TYPEMINUPTO: + case OP_HSPACE_EXTRA + OP_TYPEPOSUPTO: + if (codevalue != OP_HSPACE_EXTRA + OP_TYPEEXACT) + { ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0); } + count = current_state->count; /* Number already matched */ + if (clen > 0) + { + BOOL OK; + switch (c) + { + HSPACE_CASES: + OK = TRUE; + break; + + default: + OK = FALSE; + break; + } + + if (OK == (d == OP_HSPACE)) + { + if (codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSUPTO) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + if (++count >= (int)GET2(code, 1)) + { ADD_NEW_DATA(-(state_offset + 2 + IMM2_SIZE), 0, 0); } + else + { ADD_NEW_DATA(-state_offset, count, 0); } + } + } + break; + +/* ========================================================================== */ + /* These opcodes are followed by a character that is usually compared + to the current subject character; it is loaded into d. We still get + here even if there is no subject character, because in some cases zero + repetitions are permitted. */ + + /*-----------------------------------------------------------------*/ + case OP_CHAR: + if (clen > 0 && c == d) { ADD_NEW(state_offset + dlen + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + case OP_CHARI: + if (clen == 0) break; + +#ifdef SUPPORT_UTF + if (utf) + { + if (c == d) { ADD_NEW(state_offset + dlen + 1, 0); } else + { + unsigned int othercase; + if (c < 128) + othercase = fcc[c]; + else + /* If we have Unicode property support, we can use it to test the + other case of the character. */ +#ifdef SUPPORT_UCP + othercase = UCD_OTHERCASE(c); +#else + othercase = NOTACHAR; +#endif + + if (d == othercase) { ADD_NEW(state_offset + dlen + 1, 0); } + } + } + else +#endif /* SUPPORT_UTF */ + /* Not UTF mode */ + { + if (TABLE_GET(c, lcc, c) == TABLE_GET(d, lcc, d)) + { ADD_NEW(state_offset + 2, 0); } + } + break; + + +#ifdef SUPPORT_UCP + /*-----------------------------------------------------------------*/ + /* This is a tricky one because it can match more than one character. + Find out how many characters to skip, and then set up a negative state + to wait for them to pass before continuing. */ + + case OP_EXTUNI: + if (clen > 0) + { + int lgb, rgb; + const pcre_uchar *nptr = ptr + clen; + int ncount = 0; + lgb = UCD_GRAPHBREAK(c); + while (nptr < end_subject) + { + dlen = 1; + if (!utf) d = *nptr; else { GETCHARLEN(d, nptr, dlen); } + rgb = UCD_GRAPHBREAK(d); + if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break; + ncount++; + lgb = rgb; + nptr += dlen; + } + if (nptr >= end_subject && (md->moptions & PCRE_PARTIAL_HARD) != 0) + reset_could_continue = TRUE; + ADD_NEW_DATA(-(state_offset + 1), 0, ncount); + } + break; +#endif + + /*-----------------------------------------------------------------*/ + /* This is a tricky like EXTUNI because it too can match more than one + character (when CR is followed by LF). In this case, set up a negative + state to wait for one character to pass before continuing. */ + + case OP_ANYNL: + if (clen > 0) switch(c) + { + case CHAR_VT: + case CHAR_FF: + case CHAR_NEL: +#ifndef EBCDIC + case 0x2028: + case 0x2029: +#endif /* Not EBCDIC */ + if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break; + + case CHAR_LF: + ADD_NEW(state_offset + 1, 0); + break; + + case CHAR_CR: + if (ptr + 1 >= end_subject) + { + ADD_NEW(state_offset + 1, 0); + if ((md->moptions & PCRE_PARTIAL_HARD) != 0) + reset_could_continue = TRUE; + } + else if (UCHAR21TEST(ptr + 1) == CHAR_LF) + { + ADD_NEW_DATA(-(state_offset + 1), 0, 1); + } + else + { + ADD_NEW(state_offset + 1, 0); + } + break; + } + break; + + /*-----------------------------------------------------------------*/ + case OP_NOT_VSPACE: + if (clen > 0) switch(c) + { + VSPACE_CASES: + break; + + default: + ADD_NEW(state_offset + 1, 0); + break; + } + break; + + /*-----------------------------------------------------------------*/ + case OP_VSPACE: + if (clen > 0) switch(c) + { + VSPACE_CASES: + ADD_NEW(state_offset + 1, 0); + break; + + default: + break; + } + break; + + /*-----------------------------------------------------------------*/ + case OP_NOT_HSPACE: + if (clen > 0) switch(c) + { + HSPACE_CASES: + break; + + default: + ADD_NEW(state_offset + 1, 0); + break; + } + break; + + /*-----------------------------------------------------------------*/ + case OP_HSPACE: + if (clen > 0) switch(c) + { + HSPACE_CASES: + ADD_NEW(state_offset + 1, 0); + break; + + default: + break; + } + break; + + /*-----------------------------------------------------------------*/ + /* Match a negated single character casefully. */ + + case OP_NOT: + if (clen > 0 && c != d) { ADD_NEW(state_offset + dlen + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + /* Match a negated single character caselessly. */ + + case OP_NOTI: + if (clen > 0) + { + unsigned int otherd; +#ifdef SUPPORT_UTF + if (utf && d >= 128) + { +#ifdef SUPPORT_UCP + otherd = UCD_OTHERCASE(d); +#endif /* SUPPORT_UCP */ + } + else +#endif /* SUPPORT_UTF */ + otherd = TABLE_GET(d, fcc, d); + if (c != d && c != otherd) + { ADD_NEW(state_offset + dlen + 1, 0); } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_PLUSI: + case OP_MINPLUSI: + case OP_POSPLUSI: + case OP_NOTPLUSI: + case OP_NOTMINPLUSI: + case OP_NOTPOSPLUSI: + caseless = TRUE; + codevalue -= OP_STARI - OP_STAR; + + /* Fall through */ + case OP_PLUS: + case OP_MINPLUS: + case OP_POSPLUS: + case OP_NOTPLUS: + case OP_NOTMINPLUS: + case OP_NOTPOSPLUS: + count = current_state->count; /* Already matched */ + if (count > 0) { ADD_ACTIVE(state_offset + dlen + 1, 0); } + if (clen > 0) + { + pcre_uint32 otherd = NOTACHAR; + if (caseless) + { +#ifdef SUPPORT_UTF + if (utf && d >= 128) + { +#ifdef SUPPORT_UCP + otherd = UCD_OTHERCASE(d); +#endif /* SUPPORT_UCP */ + } + else +#endif /* SUPPORT_UTF */ + otherd = TABLE_GET(d, fcc, d); + } + if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) + { + if (count > 0 && + (codevalue == OP_POSPLUS || codevalue == OP_NOTPOSPLUS)) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + count++; + ADD_NEW(state_offset, count); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_QUERYI: + case OP_MINQUERYI: + case OP_POSQUERYI: + case OP_NOTQUERYI: + case OP_NOTMINQUERYI: + case OP_NOTPOSQUERYI: + caseless = TRUE; + codevalue -= OP_STARI - OP_STAR; + /* Fall through */ + case OP_QUERY: + case OP_MINQUERY: + case OP_POSQUERY: + case OP_NOTQUERY: + case OP_NOTMINQUERY: + case OP_NOTPOSQUERY: + ADD_ACTIVE(state_offset + dlen + 1, 0); + if (clen > 0) + { + pcre_uint32 otherd = NOTACHAR; + if (caseless) + { +#ifdef SUPPORT_UTF + if (utf && d >= 128) + { +#ifdef SUPPORT_UCP + otherd = UCD_OTHERCASE(d); +#endif /* SUPPORT_UCP */ + } + else +#endif /* SUPPORT_UTF */ + otherd = TABLE_GET(d, fcc, d); + } + if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) + { + if (codevalue == OP_POSQUERY || codevalue == OP_NOTPOSQUERY) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + ADD_NEW(state_offset + dlen + 1, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_STARI: + case OP_MINSTARI: + case OP_POSSTARI: + case OP_NOTSTARI: + case OP_NOTMINSTARI: + case OP_NOTPOSSTARI: + caseless = TRUE; + codevalue -= OP_STARI - OP_STAR; + /* Fall through */ + case OP_STAR: + case OP_MINSTAR: + case OP_POSSTAR: + case OP_NOTSTAR: + case OP_NOTMINSTAR: + case OP_NOTPOSSTAR: + ADD_ACTIVE(state_offset + dlen + 1, 0); + if (clen > 0) + { + pcre_uint32 otherd = NOTACHAR; + if (caseless) + { +#ifdef SUPPORT_UTF + if (utf && d >= 128) + { +#ifdef SUPPORT_UCP + otherd = UCD_OTHERCASE(d); +#endif /* SUPPORT_UCP */ + } + else +#endif /* SUPPORT_UTF */ + otherd = TABLE_GET(d, fcc, d); + } + if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) + { + if (codevalue == OP_POSSTAR || codevalue == OP_NOTPOSSTAR) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + ADD_NEW(state_offset, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_EXACTI: + case OP_NOTEXACTI: + caseless = TRUE; + codevalue -= OP_STARI - OP_STAR; + /* Fall through */ + case OP_EXACT: + case OP_NOTEXACT: + count = current_state->count; /* Number already matched */ + if (clen > 0) + { + pcre_uint32 otherd = NOTACHAR; + if (caseless) + { +#ifdef SUPPORT_UTF + if (utf && d >= 128) + { +#ifdef SUPPORT_UCP + otherd = UCD_OTHERCASE(d); +#endif /* SUPPORT_UCP */ + } + else +#endif /* SUPPORT_UTF */ + otherd = TABLE_GET(d, fcc, d); + } + if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) + { + if (++count >= (int)GET2(code, 1)) + { ADD_NEW(state_offset + dlen + 1 + IMM2_SIZE, 0); } + else + { ADD_NEW(state_offset, count); } + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_UPTOI: + case OP_MINUPTOI: + case OP_POSUPTOI: + case OP_NOTUPTOI: + case OP_NOTMINUPTOI: + case OP_NOTPOSUPTOI: + caseless = TRUE; + codevalue -= OP_STARI - OP_STAR; + /* Fall through */ + case OP_UPTO: + case OP_MINUPTO: + case OP_POSUPTO: + case OP_NOTUPTO: + case OP_NOTMINUPTO: + case OP_NOTPOSUPTO: + ADD_ACTIVE(state_offset + dlen + 1 + IMM2_SIZE, 0); + count = current_state->count; /* Number already matched */ + if (clen > 0) + { + pcre_uint32 otherd = NOTACHAR; + if (caseless) + { +#ifdef SUPPORT_UTF + if (utf && d >= 128) + { +#ifdef SUPPORT_UCP + otherd = UCD_OTHERCASE(d); +#endif /* SUPPORT_UCP */ + } + else +#endif /* SUPPORT_UTF */ + otherd = TABLE_GET(d, fcc, d); + } + if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) + { + if (codevalue == OP_POSUPTO || codevalue == OP_NOTPOSUPTO) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + if (++count >= (int)GET2(code, 1)) + { ADD_NEW(state_offset + dlen + 1 + IMM2_SIZE, 0); } + else + { ADD_NEW(state_offset, count); } + } + } + break; + + +/* ========================================================================== */ + /* These are the class-handling opcodes */ + + case OP_CLASS: + case OP_NCLASS: + case OP_XCLASS: + { + BOOL isinclass = FALSE; + int next_state_offset; + const pcre_uchar *ecode; + + /* For a simple class, there is always just a 32-byte table, and we + can set isinclass from it. */ + + if (codevalue != OP_XCLASS) + { + ecode = code + 1 + (32 / sizeof(pcre_uchar)); + if (clen > 0) + { + isinclass = (c > 255)? (codevalue == OP_NCLASS) : + ((((pcre_uint8 *)(code + 1))[c/8] & (1 << (c&7))) != 0); + } + } + + /* An extended class may have a table or a list of single characters, + ranges, or both, and it may be positive or negative. There's a + function that sorts all this out. */ + + else + { + ecode = code + GET(code, 1); + if (clen > 0) isinclass = PRIV(xclass)(c, code + 1 + LINK_SIZE, utf); + } + + /* At this point, isinclass is set for all kinds of class, and ecode + points to the byte after the end of the class. If there is a + quantifier, this is where it will be. */ + + next_state_offset = (int)(ecode - start_code); + + switch (*ecode) + { + case OP_CRSTAR: + case OP_CRMINSTAR: + case OP_CRPOSSTAR: + ADD_ACTIVE(next_state_offset + 1, 0); + if (isinclass) + { + if (*ecode == OP_CRPOSSTAR) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + ADD_NEW(state_offset, 0); + } + break; + + case OP_CRPLUS: + case OP_CRMINPLUS: + case OP_CRPOSPLUS: + count = current_state->count; /* Already matched */ + if (count > 0) { ADD_ACTIVE(next_state_offset + 1, 0); } + if (isinclass) + { + if (count > 0 && *ecode == OP_CRPOSPLUS) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + count++; + ADD_NEW(state_offset, count); + } + break; + + case OP_CRQUERY: + case OP_CRMINQUERY: + case OP_CRPOSQUERY: + ADD_ACTIVE(next_state_offset + 1, 0); + if (isinclass) + { + if (*ecode == OP_CRPOSQUERY) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + ADD_NEW(next_state_offset + 1, 0); + } + break; + + case OP_CRRANGE: + case OP_CRMINRANGE: + case OP_CRPOSRANGE: + count = current_state->count; /* Already matched */ + if (count >= (int)GET2(ecode, 1)) + { ADD_ACTIVE(next_state_offset + 1 + 2 * IMM2_SIZE, 0); } + if (isinclass) + { + int max = (int)GET2(ecode, 1 + IMM2_SIZE); + if (*ecode == OP_CRPOSRANGE) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + if (++count >= max && max != 0) /* Max 0 => no limit */ + { ADD_NEW(next_state_offset + 1 + 2 * IMM2_SIZE, 0); } + else + { ADD_NEW(state_offset, count); } + } + break; + + default: + if (isinclass) { ADD_NEW(next_state_offset, 0); } + break; + } + } + break; + +/* ========================================================================== */ + /* These are the opcodes for fancy brackets of various kinds. We have + to use recursion in order to handle them. The "always failing" assertion + (?!) is optimised to OP_FAIL when compiling, so we have to support that, + though the other "backtracking verbs" are not supported. */ + + case OP_FAIL: + forced_fail++; /* Count FAILs for multiple states */ + break; + + case OP_ASSERT: + case OP_ASSERT_NOT: + case OP_ASSERTBACK: + case OP_ASSERTBACK_NOT: + { + int rc; + int local_offsets[2]; + int local_workspace[1000]; + const pcre_uchar *endasscode = code + GET(code, 1); + + while (*endasscode == OP_ALT) endasscode += GET(endasscode, 1); + + rc = internal_dfa_exec( + md, /* static match data */ + code, /* this subexpression's code */ + ptr, /* where we currently are */ + (int)(ptr - start_subject), /* start offset */ + local_offsets, /* offset vector */ + sizeof(local_offsets)/sizeof(int), /* size of same */ + local_workspace, /* workspace vector */ + sizeof(local_workspace)/sizeof(int), /* size of same */ + rlevel); /* function recursion level */ + + if (rc == PCRE_ERROR_DFA_UITEM) return rc; + if ((rc >= 0) == (codevalue == OP_ASSERT || codevalue == OP_ASSERTBACK)) + { ADD_ACTIVE((int)(endasscode + LINK_SIZE + 1 - start_code), 0); } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_COND: + case OP_SCOND: + { + int local_offsets[1000]; + int local_workspace[1000]; + int codelink = GET(code, 1); + int condcode; + + /* Because of the way auto-callout works during compile, a callout item + is inserted between OP_COND and an assertion condition. This does not + happen for the other conditions. */ + + if (code[LINK_SIZE+1] == OP_CALLOUT) + { + rrc = 0; + if (PUBL(callout) != NULL) + { + PUBL(callout_block) cb; + cb.version = 1; /* Version 1 of the callout block */ + cb.callout_number = code[LINK_SIZE+2]; + cb.offset_vector = offsets; +#if defined COMPILE_PCRE8 + cb.subject = (PCRE_SPTR)start_subject; +#elif defined COMPILE_PCRE16 + cb.subject = (PCRE_SPTR16)start_subject; +#elif defined COMPILE_PCRE32 + cb.subject = (PCRE_SPTR32)start_subject; +#endif + cb.subject_length = (int)(end_subject - start_subject); + cb.start_match = (int)(current_subject - start_subject); + cb.current_position = (int)(ptr - start_subject); + cb.pattern_position = GET(code, LINK_SIZE + 3); + cb.next_item_length = GET(code, 3 + 2*LINK_SIZE); + cb.capture_top = 1; + cb.capture_last = -1; + cb.callout_data = md->callout_data; + cb.mark = NULL; /* No (*MARK) support */ + if ((rrc = (*PUBL(callout))(&cb)) < 0) return rrc; /* Abandon */ + } + if (rrc > 0) break; /* Fail this thread */ + code += PRIV(OP_lengths)[OP_CALLOUT]; /* Skip callout data */ + } + + condcode = code[LINK_SIZE+1]; + + /* Back reference conditions and duplicate named recursion conditions + are not supported */ + + if (condcode == OP_CREF || condcode == OP_DNCREF || + condcode == OP_DNRREF) + return PCRE_ERROR_DFA_UCOND; + + /* The DEFINE condition is always false, and the assertion (?!) is + converted to OP_FAIL. */ + + if (condcode == OP_DEF || condcode == OP_FAIL) + { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); } + + /* The only supported version of OP_RREF is for the value RREF_ANY, + which means "test if in any recursion". We can't test for specifically + recursed groups. */ + + else if (condcode == OP_RREF) + { + int value = GET2(code, LINK_SIZE + 2); + if (value != RREF_ANY) return PCRE_ERROR_DFA_UCOND; + if (md->recursive != NULL) + { ADD_ACTIVE(state_offset + LINK_SIZE + 2 + IMM2_SIZE, 0); } + else { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); } + } + + /* Otherwise, the condition is an assertion */ + + else + { + int rc; + const pcre_uchar *asscode = code + LINK_SIZE + 1; + const pcre_uchar *endasscode = asscode + GET(asscode, 1); + + while (*endasscode == OP_ALT) endasscode += GET(endasscode, 1); + + rc = internal_dfa_exec( + md, /* fixed match data */ + asscode, /* this subexpression's code */ + ptr, /* where we currently are */ + (int)(ptr - start_subject), /* start offset */ + local_offsets, /* offset vector */ + sizeof(local_offsets)/sizeof(int), /* size of same */ + local_workspace, /* workspace vector */ + sizeof(local_workspace)/sizeof(int), /* size of same */ + rlevel); /* function recursion level */ + + if (rc == PCRE_ERROR_DFA_UITEM) return rc; + if ((rc >= 0) == + (condcode == OP_ASSERT || condcode == OP_ASSERTBACK)) + { ADD_ACTIVE((int)(endasscode + LINK_SIZE + 1 - start_code), 0); } + else + { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); } + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_RECURSE: + { + dfa_recursion_info *ri; + int local_offsets[1000]; + int local_workspace[1000]; + const pcre_uchar *callpat = start_code + GET(code, 1); + int recno = (callpat == md->start_code)? 0 : + GET2(callpat, 1 + LINK_SIZE); + int rc; + + DPRINTF(("%.*sStarting regex recursion\n", rlevel*2-2, SP)); + + /* Check for repeating a recursion without advancing the subject + pointer. This should catch convoluted mutual recursions. (Some simple + cases are caught at compile time.) */ + + for (ri = md->recursive; ri != NULL; ri = ri->prevrec) + if (recno == ri->group_num && ptr == ri->subject_position) + return PCRE_ERROR_RECURSELOOP; + + /* Remember this recursion and where we started it so as to + catch infinite loops. */ + + new_recursive.group_num = recno; + new_recursive.subject_position = ptr; + new_recursive.prevrec = md->recursive; + md->recursive = &new_recursive; + + rc = internal_dfa_exec( + md, /* fixed match data */ + callpat, /* this subexpression's code */ + ptr, /* where we currently are */ + (int)(ptr - start_subject), /* start offset */ + local_offsets, /* offset vector */ + sizeof(local_offsets)/sizeof(int), /* size of same */ + local_workspace, /* workspace vector */ + sizeof(local_workspace)/sizeof(int), /* size of same */ + rlevel); /* function recursion level */ + + md->recursive = new_recursive.prevrec; /* Done this recursion */ + + DPRINTF(("%.*sReturn from regex recursion: rc=%d\n", rlevel*2-2, SP, + rc)); + + /* Ran out of internal offsets */ + + if (rc == 0) return PCRE_ERROR_DFA_RECURSE; + + /* For each successful matched substring, set up the next state with a + count of characters to skip before trying it. Note that the count is in + characters, not bytes. */ + + if (rc > 0) + { + for (rc = rc*2 - 2; rc >= 0; rc -= 2) + { + int charcount = local_offsets[rc+1] - local_offsets[rc]; +#if defined SUPPORT_UTF && !defined COMPILE_PCRE32 + if (utf) + { + const pcre_uchar *p = start_subject + local_offsets[rc]; + const pcre_uchar *pp = start_subject + local_offsets[rc+1]; + while (p < pp) if (NOT_FIRSTCHAR(*p++)) charcount--; + } +#endif + if (charcount > 0) + { + ADD_NEW_DATA(-(state_offset + LINK_SIZE + 1), 0, (charcount - 1)); + } + else + { + ADD_ACTIVE(state_offset + LINK_SIZE + 1, 0); + } + } + } + else if (rc != PCRE_ERROR_NOMATCH) return rc; + } + break; + + /*-----------------------------------------------------------------*/ + case OP_BRAPOS: + case OP_SBRAPOS: + case OP_CBRAPOS: + case OP_SCBRAPOS: + case OP_BRAPOSZERO: + { + int charcount, matched_count; + const pcre_uchar *local_ptr = ptr; + BOOL allow_zero; + + if (codevalue == OP_BRAPOSZERO) + { + allow_zero = TRUE; + codevalue = *(++code); /* Codevalue will be one of above BRAs */ + } + else allow_zero = FALSE; + + /* Loop to match the subpattern as many times as possible as if it were + a complete pattern. */ + + for (matched_count = 0;; matched_count++) + { + int local_offsets[2]; + int local_workspace[1000]; + + int rc = internal_dfa_exec( + md, /* fixed match data */ + code, /* this subexpression's code */ + local_ptr, /* where we currently are */ + (int)(ptr - start_subject), /* start offset */ + local_offsets, /* offset vector */ + sizeof(local_offsets)/sizeof(int), /* size of same */ + local_workspace, /* workspace vector */ + sizeof(local_workspace)/sizeof(int), /* size of same */ + rlevel); /* function recursion level */ + + /* Failed to match */ + + if (rc < 0) + { + if (rc != PCRE_ERROR_NOMATCH) return rc; + break; + } + + /* Matched: break the loop if zero characters matched. */ + + charcount = local_offsets[1] - local_offsets[0]; + if (charcount == 0) break; + local_ptr += charcount; /* Advance temporary position ptr */ + } + + /* At this point we have matched the subpattern matched_count + times, and local_ptr is pointing to the character after the end of the + last match. */ + + if (matched_count > 0 || allow_zero) + { + const pcre_uchar *end_subpattern = code; + int next_state_offset; + + do { end_subpattern += GET(end_subpattern, 1); } + while (*end_subpattern == OP_ALT); + next_state_offset = + (int)(end_subpattern - start_code + LINK_SIZE + 1); + + /* Optimization: if there are no more active states, and there + are no new states yet set up, then skip over the subject string + right here, to save looping. Otherwise, set up the new state to swing + into action when the end of the matched substring is reached. */ + + if (i + 1 >= active_count && new_count == 0) + { + ptr = local_ptr; + clen = 0; + ADD_NEW(next_state_offset, 0); + } + else + { + const pcre_uchar *p = ptr; + const pcre_uchar *pp = local_ptr; + charcount = (int)(pp - p); +#if defined SUPPORT_UTF && !defined COMPILE_PCRE32 + if (utf) while (p < pp) if (NOT_FIRSTCHAR(*p++)) charcount--; +#endif + ADD_NEW_DATA(-next_state_offset, 0, (charcount - 1)); + } + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_ONCE: + case OP_ONCE_NC: + { + int local_offsets[2]; + int local_workspace[1000]; + + int rc = internal_dfa_exec( + md, /* fixed match data */ + code, /* this subexpression's code */ + ptr, /* where we currently are */ + (int)(ptr - start_subject), /* start offset */ + local_offsets, /* offset vector */ + sizeof(local_offsets)/sizeof(int), /* size of same */ + local_workspace, /* workspace vector */ + sizeof(local_workspace)/sizeof(int), /* size of same */ + rlevel); /* function recursion level */ + + if (rc >= 0) + { + const pcre_uchar *end_subpattern = code; + int charcount = local_offsets[1] - local_offsets[0]; + int next_state_offset, repeat_state_offset; + + do { end_subpattern += GET(end_subpattern, 1); } + while (*end_subpattern == OP_ALT); + next_state_offset = + (int)(end_subpattern - start_code + LINK_SIZE + 1); + + /* If the end of this subpattern is KETRMAX or KETRMIN, we must + arrange for the repeat state also to be added to the relevant list. + Calculate the offset, or set -1 for no repeat. */ + + repeat_state_offset = (*end_subpattern == OP_KETRMAX || + *end_subpattern == OP_KETRMIN)? + (int)(end_subpattern - start_code - GET(end_subpattern, 1)) : -1; + + /* If we have matched an empty string, add the next state at the + current character pointer. This is important so that the duplicate + checking kicks in, which is what breaks infinite loops that match an + empty string. */ + + if (charcount == 0) + { + ADD_ACTIVE(next_state_offset, 0); + } + + /* Optimization: if there are no more active states, and there + are no new states yet set up, then skip over the subject string + right here, to save looping. Otherwise, set up the new state to swing + into action when the end of the matched substring is reached. */ + + else if (i + 1 >= active_count && new_count == 0) + { + ptr += charcount; + clen = 0; + ADD_NEW(next_state_offset, 0); + + /* If we are adding a repeat state at the new character position, + we must fudge things so that it is the only current state. + Otherwise, it might be a duplicate of one we processed before, and + that would cause it to be skipped. */ + + if (repeat_state_offset >= 0) + { + next_active_state = active_states; + active_count = 0; + i = -1; + ADD_ACTIVE(repeat_state_offset, 0); + } + } + else + { +#if defined SUPPORT_UTF && !defined COMPILE_PCRE32 + if (utf) + { + const pcre_uchar *p = start_subject + local_offsets[0]; + const pcre_uchar *pp = start_subject + local_offsets[1]; + while (p < pp) if (NOT_FIRSTCHAR(*p++)) charcount--; + } +#endif + ADD_NEW_DATA(-next_state_offset, 0, (charcount - 1)); + if (repeat_state_offset >= 0) + { ADD_NEW_DATA(-repeat_state_offset, 0, (charcount - 1)); } + } + } + else if (rc != PCRE_ERROR_NOMATCH) return rc; + } + break; + + +/* ========================================================================== */ + /* Handle callouts */ + + case OP_CALLOUT: + rrc = 0; + if (PUBL(callout) != NULL) + { + PUBL(callout_block) cb; + cb.version = 1; /* Version 1 of the callout block */ + cb.callout_number = code[1]; + cb.offset_vector = offsets; +#if defined COMPILE_PCRE8 + cb.subject = (PCRE_SPTR)start_subject; +#elif defined COMPILE_PCRE16 + cb.subject = (PCRE_SPTR16)start_subject; +#elif defined COMPILE_PCRE32 + cb.subject = (PCRE_SPTR32)start_subject; +#endif + cb.subject_length = (int)(end_subject - start_subject); + cb.start_match = (int)(current_subject - start_subject); + cb.current_position = (int)(ptr - start_subject); + cb.pattern_position = GET(code, 2); + cb.next_item_length = GET(code, 2 + LINK_SIZE); + cb.capture_top = 1; + cb.capture_last = -1; + cb.callout_data = md->callout_data; + cb.mark = NULL; /* No (*MARK) support */ + if ((rrc = (*PUBL(callout))(&cb)) < 0) return rrc; /* Abandon */ + } + if (rrc == 0) + { ADD_ACTIVE(state_offset + PRIV(OP_lengths)[OP_CALLOUT], 0); } + break; + + +/* ========================================================================== */ + default: /* Unsupported opcode */ + return PCRE_ERROR_DFA_UITEM; + } + + NEXT_ACTIVE_STATE: continue; + + } /* End of loop scanning active states */ + + /* We have finished the processing at the current subject character. If no + new states have been set for the next character, we have found all the + matches that we are going to find. If we are at the top level and partial + matching has been requested, check for appropriate conditions. + + The "forced_ fail" variable counts the number of (*F) encountered for the + character. If it is equal to the original active_count (saved in + workspace[1]) it means that (*F) was found on every active state. In this + case we don't want to give a partial match. + + The "could_continue" variable is true if a state could have continued but + for the fact that the end of the subject was reached. */ + + if (new_count <= 0) + { + if (rlevel == 1 && /* Top level, and */ + could_continue && /* Some could go on, and */ + forced_fail != workspace[1] && /* Not all forced fail & */ + ( /* either... */ + (md->moptions & PCRE_PARTIAL_HARD) != 0 /* Hard partial */ + || /* or... */ + ((md->moptions & PCRE_PARTIAL_SOFT) != 0 && /* Soft partial and */ + match_count < 0) /* no matches */ + ) && /* And... */ + ( + partial_newline || /* Either partial NL */ + ( /* or ... */ + ptr >= end_subject && /* End of subject and */ + ptr > md->start_used_ptr) /* Inspected non-empty string */ + ) + ) + match_count = PCRE_ERROR_PARTIAL; + DPRINTF(("%.*sEnd of internal_dfa_exec %d: returning %d\n" + "%.*s---------------------\n\n", rlevel*2-2, SP, rlevel, match_count, + rlevel*2-2, SP)); + break; /* In effect, "return", but see the comment below */ + } + + /* One or more states are active for the next character. */ + + ptr += clen; /* Advance to next subject character */ + } /* Loop to move along the subject string */ + +/* Control gets here from "break" a few lines above. We do it this way because +if we use "return" above, we have compiler trouble. Some compilers warn if +there's nothing here because they think the function doesn't return a value. On +the other hand, if we put a dummy statement here, some more clever compilers +complain that it can't be reached. Sigh. */ + +return match_count; +} + + + + +/************************************************* +* Execute a Regular Expression - DFA engine * +*************************************************/ + +/* This external function applies a compiled re to a subject string using a DFA +engine. This function calls the internal function multiple times if the pattern +is not anchored. + +Arguments: + argument_re points to the compiled expression + extra_data points to extra data or is NULL + subject points to the subject string + length length of subject string (may contain binary zeros) + start_offset where to start in the subject string + options option bits + offsets vector of match offsets + offsetcount size of same + workspace workspace vector + wscount size of same + +Returns: > 0 => number of match offset pairs placed in offsets + = 0 => offsets overflowed; longest matches are present + -1 => failed to match + < -1 => some kind of unexpected problem +*/ + +#if defined COMPILE_PCRE8 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre_dfa_exec(const pcre *argument_re, const pcre_extra *extra_data, + const char *subject, int length, int start_offset, int options, int *offsets, + int offsetcount, int *workspace, int wscount) +#elif defined COMPILE_PCRE16 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre16_dfa_exec(const pcre16 *argument_re, const pcre16_extra *extra_data, + PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets, + int offsetcount, int *workspace, int wscount) +#elif defined COMPILE_PCRE32 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre32_dfa_exec(const pcre32 *argument_re, const pcre32_extra *extra_data, + PCRE_SPTR32 subject, int length, int start_offset, int options, int *offsets, + int offsetcount, int *workspace, int wscount) +#endif +{ +REAL_PCRE *re = (REAL_PCRE *)argument_re; +dfa_match_data match_block; +dfa_match_data *md = &match_block; +BOOL utf, anchored, startline, firstline; +const pcre_uchar *current_subject, *end_subject; +const pcre_study_data *study = NULL; + +const pcre_uchar *req_char_ptr; +const pcre_uint8 *start_bits = NULL; +BOOL has_first_char = FALSE; +BOOL has_req_char = FALSE; +pcre_uchar first_char = 0; +pcre_uchar first_char2 = 0; +pcre_uchar req_char = 0; +pcre_uchar req_char2 = 0; +int newline; + +/* Plausibility checks */ + +if ((options & ~PUBLIC_DFA_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION; +if (re == NULL || subject == NULL || workspace == NULL || + (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL; +if (offsetcount < 0) return PCRE_ERROR_BADCOUNT; +if (wscount < 20) return PCRE_ERROR_DFA_WSSIZE; +if (length < 0) return PCRE_ERROR_BADLENGTH; +if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET; + +/* Check that the first field in the block is the magic number. If it is not, +return with PCRE_ERROR_BADMAGIC. However, if the magic number is equal to +REVERSED_MAGIC_NUMBER we return with PCRE_ERROR_BADENDIANNESS, which +means that the pattern is likely compiled with different endianness. */ + +if (re->magic_number != MAGIC_NUMBER) + return re->magic_number == REVERSED_MAGIC_NUMBER? + PCRE_ERROR_BADENDIANNESS:PCRE_ERROR_BADMAGIC; +if ((re->flags & PCRE_MODE) == 0) return PCRE_ERROR_BADMODE; + +/* If restarting after a partial match, do some sanity checks on the contents +of the workspace. */ + +if ((options & PCRE_DFA_RESTART) != 0) + { + if ((workspace[0] & (-2)) != 0 || workspace[1] < 1 || + workspace[1] > (wscount - 2)/INTS_PER_STATEBLOCK) + return PCRE_ERROR_DFA_BADRESTART; + } + +/* Set up study, callout, and table data */ + +md->tables = re->tables; +md->callout_data = NULL; + +if (extra_data != NULL) + { + unsigned long int flags = extra_data->flags; + if ((flags & PCRE_EXTRA_STUDY_DATA) != 0) + study = (const pcre_study_data *)extra_data->study_data; + if ((flags & PCRE_EXTRA_MATCH_LIMIT) != 0) return PCRE_ERROR_DFA_UMLIMIT; + if ((flags & PCRE_EXTRA_MATCH_LIMIT_RECURSION) != 0) + return PCRE_ERROR_DFA_UMLIMIT; + if ((flags & PCRE_EXTRA_CALLOUT_DATA) != 0) + md->callout_data = extra_data->callout_data; + if ((flags & PCRE_EXTRA_TABLES) != 0) + md->tables = extra_data->tables; + } + +/* Set some local values */ + +current_subject = (const pcre_uchar *)subject + start_offset; +end_subject = (const pcre_uchar *)subject + length; +req_char_ptr = current_subject - 1; + +#ifdef SUPPORT_UTF +/* PCRE_UTF(16|32) have the same value as PCRE_UTF8. */ +utf = (re->options & PCRE_UTF8) != 0; +#else +utf = FALSE; +#endif + +anchored = (options & (PCRE_ANCHORED|PCRE_DFA_RESTART)) != 0 || + (re->options & PCRE_ANCHORED) != 0; + +/* The remaining fixed data for passing around. */ + +md->start_code = (const pcre_uchar *)argument_re + + re->name_table_offset + re->name_count * re->name_entry_size; +md->start_subject = (const pcre_uchar *)subject; +md->end_subject = end_subject; +md->start_offset = start_offset; +md->moptions = options; +md->poptions = re->options; + +/* If the BSR option is not set at match time, copy what was set +at compile time. */ + +if ((md->moptions & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) == 0) + { + if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0) + md->moptions |= re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE); +#ifdef BSR_ANYCRLF + else md->moptions |= PCRE_BSR_ANYCRLF; +#endif + } + +/* Handle different types of newline. The three bits give eight cases. If +nothing is set at run time, whatever was used at compile time applies. */ + +switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options : (pcre_uint32)options) & + PCRE_NEWLINE_BITS) + { + case 0: newline = NEWLINE; break; /* Compile-time default */ + case PCRE_NEWLINE_CR: newline = CHAR_CR; break; + case PCRE_NEWLINE_LF: newline = CHAR_NL; break; + case PCRE_NEWLINE_CR+ + PCRE_NEWLINE_LF: newline = (CHAR_CR << 8) | CHAR_NL; break; + case PCRE_NEWLINE_ANY: newline = -1; break; + case PCRE_NEWLINE_ANYCRLF: newline = -2; break; + default: return PCRE_ERROR_BADNEWLINE; + } + +if (newline == -2) + { + md->nltype = NLTYPE_ANYCRLF; + } +else if (newline < 0) + { + md->nltype = NLTYPE_ANY; + } +else + { + md->nltype = NLTYPE_FIXED; + if (newline > 255) + { + md->nllen = 2; + md->nl[0] = (newline >> 8) & 255; + md->nl[1] = newline & 255; + } + else + { + md->nllen = 1; + md->nl[0] = newline; + } + } + +/* Check a UTF-8 string if required. Unfortunately there's no way of passing +back the character offset. */ + +#ifdef SUPPORT_UTF +if (utf && (options & PCRE_NO_UTF8_CHECK) == 0) + { + int erroroffset; + int errorcode = PRIV(valid_utf)((pcre_uchar *)subject, length, &erroroffset); + if (errorcode != 0) + { + if (offsetcount >= 2) + { + offsets[0] = erroroffset; + offsets[1] = errorcode; + } +#if defined COMPILE_PCRE8 + return (errorcode <= PCRE_UTF8_ERR5 && (options & PCRE_PARTIAL_HARD) != 0) ? + PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8; +#elif defined COMPILE_PCRE16 + return (errorcode <= PCRE_UTF16_ERR1 && (options & PCRE_PARTIAL_HARD) != 0) ? + PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16; +#elif defined COMPILE_PCRE32 + return PCRE_ERROR_BADUTF32; +#endif + } +#if defined COMPILE_PCRE8 || defined COMPILE_PCRE16 + if (start_offset > 0 && start_offset < length && + NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset])) + return PCRE_ERROR_BADUTF8_OFFSET; +#endif + } +#endif + +/* If the exec call supplied NULL for tables, use the inbuilt ones. This +is a feature that makes it possible to save compiled regex and re-use them +in other programs later. */ + +if (md->tables == NULL) md->tables = PRIV(default_tables); + +/* The "must be at the start of a line" flags are used in a loop when finding +where to start. */ + +startline = (re->flags & PCRE_STARTLINE) != 0; +firstline = (re->options & PCRE_FIRSTLINE) != 0; + +/* Set up the first character to match, if available. The first_byte value is +never set for an anchored regular expression, but the anchoring may be forced +at run time, so we have to test for anchoring. The first char may be unset for +an unanchored pattern, of course. If there's no first char and the pattern was +studied, there may be a bitmap of possible first characters. */ + +if (!anchored) + { + if ((re->flags & PCRE_FIRSTSET) != 0) + { + has_first_char = TRUE; + first_char = first_char2 = (pcre_uchar)(re->first_char); + if ((re->flags & PCRE_FCH_CASELESS) != 0) + { + first_char2 = TABLE_GET(first_char, md->tables + fcc_offset, first_char); +#if defined SUPPORT_UCP && !(defined COMPILE_PCRE8) + if (utf && first_char > 127) + first_char2 = UCD_OTHERCASE(first_char); +#endif + } + } + else + { + if (!startline && study != NULL && + (study->flags & PCRE_STUDY_MAPPED) != 0) + start_bits = study->start_bits; + } + } + +/* For anchored or unanchored matches, there may be a "last known required +character" set. */ + +if ((re->flags & PCRE_REQCHSET) != 0) + { + has_req_char = TRUE; + req_char = req_char2 = (pcre_uchar)(re->req_char); + if ((re->flags & PCRE_RCH_CASELESS) != 0) + { + req_char2 = TABLE_GET(req_char, md->tables + fcc_offset, req_char); +#if defined SUPPORT_UCP && !(defined COMPILE_PCRE8) + if (utf && req_char > 127) + req_char2 = UCD_OTHERCASE(req_char); +#endif + } + } + +/* Call the main matching function, looping for a non-anchored regex after a +failed match. If not restarting, perform certain optimizations at the start of +a match. */ + +for (;;) + { + int rc; + + if ((options & PCRE_DFA_RESTART) == 0) + { + const pcre_uchar *save_end_subject = end_subject; + + /* If firstline is TRUE, the start of the match is constrained to the first + line of a multiline string. Implement this by temporarily adjusting + end_subject so that we stop scanning at a newline. If the match fails at + the newline, later code breaks this loop. */ + + if (firstline) + { + PCRE_PUCHAR t = current_subject; +#ifdef SUPPORT_UTF + if (utf) + { + while (t < md->end_subject && !IS_NEWLINE(t)) + { + t++; + ACROSSCHAR(t < end_subject, *t, t++); + } + } + else +#endif + while (t < md->end_subject && !IS_NEWLINE(t)) t++; + end_subject = t; + } + + /* There are some optimizations that avoid running the match if a known + starting point is not found. However, there is an option that disables + these, for testing and for ensuring that all callouts do actually occur. + The option can be set in the regex by (*NO_START_OPT) or passed in + match-time options. */ + + if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0) + { + /* Advance to a known first pcre_uchar (i.e. data item) */ + + if (has_first_char) + { + if (first_char != first_char2) + { + pcre_uchar csc; + while (current_subject < end_subject && + (csc = UCHAR21TEST(current_subject)) != first_char && csc != first_char2) + current_subject++; + } + else + while (current_subject < end_subject && + UCHAR21TEST(current_subject) != first_char) + current_subject++; + } + + /* Or to just after a linebreak for a multiline match if possible */ + + else if (startline) + { + if (current_subject > md->start_subject + start_offset) + { +#ifdef SUPPORT_UTF + if (utf) + { + while (current_subject < end_subject && + !WAS_NEWLINE(current_subject)) + { + current_subject++; + ACROSSCHAR(current_subject < end_subject, *current_subject, + current_subject++); + } + } + else +#endif + while (current_subject < end_subject && !WAS_NEWLINE(current_subject)) + current_subject++; + + /* If we have just passed a CR and the newline option is ANY or + ANYCRLF, and we are now at a LF, advance the match position by one + more character. */ + + if (UCHAR21TEST(current_subject - 1) == CHAR_CR && + (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) && + current_subject < end_subject && + UCHAR21TEST(current_subject) == CHAR_NL) + current_subject++; + } + } + + /* Advance to a non-unique first pcre_uchar after study */ + + else if (start_bits != NULL) + { + while (current_subject < end_subject) + { + register pcre_uint32 c = UCHAR21TEST(current_subject); +#ifndef COMPILE_PCRE8 + if (c > 255) c = 255; +#endif + if ((start_bits[c/8] & (1 << (c&7))) != 0) break; + current_subject++; + } + } + } + + /* Restore fudged end_subject */ + + end_subject = save_end_subject; + + /* The following two optimizations are disabled for partial matching or if + disabling is explicitly requested (and of course, by the test above, this + code is not obeyed when restarting after a partial match). */ + + if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0 && + (options & (PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT)) == 0) + { + /* If the pattern was studied, a minimum subject length may be set. This + is a lower bound; no actual string of that length may actually match the + pattern. Although the value is, strictly, in characters, we treat it as + in pcre_uchar units to avoid spending too much time in this optimization. + */ + + if (study != NULL && (study->flags & PCRE_STUDY_MINLEN) != 0 && + (pcre_uint32)(end_subject - current_subject) < study->minlength) + return PCRE_ERROR_NOMATCH; + + /* If req_char is set, we know that that pcre_uchar must appear in the + subject for the match to succeed. If the first pcre_uchar is set, + req_char must be later in the subject; otherwise the test starts at the + match point. This optimization can save a huge amount of work in patterns + with nested unlimited repeats that aren't going to match. Writing + separate code for cased/caseless versions makes it go faster, as does + using an autoincrement and backing off on a match. + + HOWEVER: when the subject string is very, very long, searching to its end + can take a long time, and give bad performance on quite ordinary + patterns. This showed up when somebody was matching /^C/ on a 32-megabyte + string... so we don't do this when the string is sufficiently long. */ + + if (has_req_char && end_subject - current_subject < REQ_BYTE_MAX) + { + register PCRE_PUCHAR p = current_subject + (has_first_char? 1:0); + + /* We don't need to repeat the search if we haven't yet reached the + place we found it at last time. */ + + if (p > req_char_ptr) + { + if (req_char != req_char2) + { + while (p < end_subject) + { + register pcre_uint32 pp = UCHAR21INCTEST(p); + if (pp == req_char || pp == req_char2) { p--; break; } + } + } + else + { + while (p < end_subject) + { + if (UCHAR21INCTEST(p) == req_char) { p--; break; } + } + } + + /* If we can't find the required pcre_uchar, break the matching loop, + which will cause a return or PCRE_ERROR_NOMATCH. */ + + if (p >= end_subject) break; + + /* If we have found the required pcre_uchar, save the point where we + found it, so that we don't search again next time round the loop if + the start hasn't passed this point yet. */ + + req_char_ptr = p; + } + } + } + } /* End of optimizations that are done when not restarting */ + + /* OK, now we can do the business */ + + md->start_used_ptr = current_subject; + md->recursive = NULL; + + rc = internal_dfa_exec( + md, /* fixed match data */ + md->start_code, /* this subexpression's code */ + current_subject, /* where we currently are */ + start_offset, /* start offset in subject */ + offsets, /* offset vector */ + offsetcount, /* size of same */ + workspace, /* workspace vector */ + wscount, /* size of same */ + 0); /* function recurse level */ + + /* Anything other than "no match" means we are done, always; otherwise, carry + on only if not anchored. */ + + if (rc != PCRE_ERROR_NOMATCH || anchored) + { + if (rc == PCRE_ERROR_PARTIAL && offsetcount >= 2) + { + offsets[0] = (int)(md->start_used_ptr - (PCRE_PUCHAR)subject); + offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject); + if (offsetcount > 2) + offsets[2] = (int)(current_subject - (PCRE_PUCHAR)subject); + } + return rc; + } + + /* Advance to the next subject character unless we are at the end of a line + and firstline is set. */ + + if (firstline && IS_NEWLINE(current_subject)) break; + current_subject++; +#ifdef SUPPORT_UTF + if (utf) + { + ACROSSCHAR(current_subject < end_subject, *current_subject, + current_subject++); + } +#endif + if (current_subject > end_subject) break; + + /* If we have just passed a CR and we are now at a LF, and the pattern does + not contain any explicit matches for \r or \n, and the newline option is CRLF + or ANY or ANYCRLF, advance the match position by one more character. */ + + if (UCHAR21TEST(current_subject - 1) == CHAR_CR && + current_subject < end_subject && + UCHAR21TEST(current_subject) == CHAR_NL && + (re->flags & PCRE_HASCRORLF) == 0 && + (md->nltype == NLTYPE_ANY || + md->nltype == NLTYPE_ANYCRLF || + md->nllen == 2)) + current_subject++; + + } /* "Bumpalong" loop */ + +return PCRE_ERROR_NOMATCH; +} + +/* End of pcre_dfa_exec.c */ diff --git a/include/pcre/pcre_exec.c b/include/pcre/pcre_exec.c new file mode 100644 index 000000000..24b23ca28 --- /dev/null +++ b/include/pcre/pcre_exec.c @@ -0,0 +1,7173 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + +/* PCRE is a library of functions to support regular expressions whose syntax +and semantics are as close as possible to those of the Perl 5 language. + + Written by Philip Hazel + Copyright (c) 1997-2014 University of Cambridge + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + +/* This module contains pcre_exec(), the externally visible function that does +pattern matching using an NFA algorithm, trying to mimic Perl as closely as +possible. There are also some static supporting functions. */ + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#define NLBLOCK md /* Block containing newline information */ +#define PSSTART start_subject /* Field containing processed string start */ +#define PSEND end_subject /* Field containing processed string end */ + +#include "pcre_internal.h" + +/* Undefine some potentially clashing cpp symbols */ + +#undef min +#undef max + +/* The md->capture_last field uses the lower 16 bits for the last captured +substring (which can never be greater than 65535) and a bit in the top half +to mean "capture vector overflowed". This odd way of doing things was +implemented when it was realized that preserving and restoring the overflow bit +whenever the last capture number was saved/restored made for a neater +interface, and doing it this way saved on (a) another variable, which would +have increased the stack frame size (a big NO-NO in PCRE) and (b) another +separate set of save/restore instructions. The following defines are used in +implementing this. */ + +#define CAPLMASK 0x0000ffff /* The bits used for last_capture */ +#define OVFLMASK 0xffff0000 /* The bits used for the overflow flag */ +#define OVFLBIT 0x00010000 /* The bit that is set for overflow */ + +/* Values for setting in md->match_function_type to indicate two special types +of call to match(). We do it this way to save on using another stack variable, +as stack usage is to be discouraged. */ + +#define MATCH_CONDASSERT 1 /* Called to check a condition assertion */ +#define MATCH_CBEGROUP 2 /* Could-be-empty unlimited repeat group */ + +/* Non-error returns from the match() function. Error returns are externally +defined PCRE_ERROR_xxx codes, which are all negative. */ + +#define MATCH_MATCH 1 +#define MATCH_NOMATCH 0 + +/* Special internal returns from the match() function. Make them sufficiently +negative to avoid the external error codes. */ + +#define MATCH_ACCEPT (-999) +#define MATCH_KETRPOS (-998) +#define MATCH_ONCE (-997) +/* The next 5 must be kept together and in sequence so that a test that checks +for any one of them can use a range. */ +#define MATCH_COMMIT (-996) +#define MATCH_PRUNE (-995) +#define MATCH_SKIP (-994) +#define MATCH_SKIP_ARG (-993) +#define MATCH_THEN (-992) +#define MATCH_BACKTRACK_MAX MATCH_THEN +#define MATCH_BACKTRACK_MIN MATCH_COMMIT + +/* Maximum number of ints of offset to save on the stack for recursive calls. +If the offset vector is bigger, malloc is used. This should be a multiple of 3, +because the offset vector is always a multiple of 3 long. */ + +#define REC_STACK_SAVE_MAX 30 + +/* Min and max values for the common repeats; for the maxima, 0 => infinity */ + +static const char rep_min[] = { 0, 0, 1, 1, 0, 0, 0, 0, 0, 1, 0, }; +static const char rep_max[] = { 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 1, }; + +#ifdef PCRE_DEBUG +/************************************************* +* Debugging function to print chars * +*************************************************/ + +/* Print a sequence of chars in printable format, stopping at the end of the +subject if the requested. + +Arguments: + p points to characters + length number to print + is_subject TRUE if printing from within md->start_subject + md pointer to matching data block, if is_subject is TRUE + +Returns: nothing +*/ + +static void +pchars(const pcre_uchar *p, int length, BOOL is_subject, match_data *md) +{ +pcre_uint32 c; +BOOL utf = md->utf; +if (is_subject && length > md->end_subject - p) length = md->end_subject - p; +while (length-- > 0) + if (isprint(c = UCHAR21INCTEST(p))) printf("%c", (char)c); else printf("\\x{%02x}", c); +} +#endif + + + +/************************************************* +* Match a back-reference * +*************************************************/ + +/* Normally, if a back reference hasn't been set, the length that is passed is +negative, so the match always fails. However, in JavaScript compatibility mode, +the length passed is zero. Note that in caseless UTF-8 mode, the number of +subject bytes matched may be different to the number of reference bytes. + +Arguments: + offset index into the offset vector + eptr pointer into the subject + length length of reference to be matched (number of bytes) + md points to match data block + caseless TRUE if caseless + +Returns: >= 0 the number of subject bytes matched + -1 no match + -2 partial match; always given if at end subject +*/ + +static int +match_ref(int offset, register PCRE_PUCHAR eptr, int length, match_data *md, + BOOL caseless) +{ +PCRE_PUCHAR eptr_start = eptr; +register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset]; +#if defined SUPPORT_UTF && defined SUPPORT_UCP +BOOL utf = md->utf; +#endif + +#ifdef PCRE_DEBUG +if (eptr >= md->end_subject) + printf("matching subject "); +else + { + printf("matching subject "); + pchars(eptr, length, TRUE, md); + } +printf(" against backref "); +pchars(p, length, FALSE, md); +printf("\n"); +#endif + +/* Always fail if reference not set (and not JavaScript compatible - in that +case the length is passed as zero). */ + +if (length < 0) return -1; + +/* Separate the caseless case for speed. In UTF-8 mode we can only do this +properly if Unicode properties are supported. Otherwise, we can check only +ASCII characters. */ + +if (caseless) + { +#if defined SUPPORT_UTF && defined SUPPORT_UCP + if (utf) + { + /* Match characters up to the end of the reference. NOTE: the number of + data units matched may differ, because in UTF-8 there are some characters + whose upper and lower case versions code have different numbers of bytes. + For example, U+023A (2 bytes in UTF-8) is the upper case version of U+2C65 + (3 bytes in UTF-8); a sequence of 3 of the former uses 6 bytes, as does a + sequence of two of the latter. It is important, therefore, to check the + length along the reference, not along the subject (earlier code did this + wrong). */ + + PCRE_PUCHAR endptr = p + length; + while (p < endptr) + { + pcre_uint32 c, d; + const ucd_record *ur; + if (eptr >= md->end_subject) return -2; /* Partial match */ + GETCHARINC(c, eptr); + GETCHARINC(d, p); + ur = GET_UCD(d); + if (c != d && c != d + ur->other_case) + { + const pcre_uint32 *pp = PRIV(ucd_caseless_sets) + ur->caseset; + for (;;) + { + if (c < *pp) return -1; + if (c == *pp++) break; + } + } + } + } + else +#endif + + /* The same code works when not in UTF-8 mode and in UTF-8 mode when there + is no UCP support. */ + { + while (length-- > 0) + { + pcre_uint32 cc, cp; + if (eptr >= md->end_subject) return -2; /* Partial match */ + cc = UCHAR21TEST(eptr); + cp = UCHAR21TEST(p); + if (TABLE_GET(cp, md->lcc, cp) != TABLE_GET(cc, md->lcc, cc)) return -1; + p++; + eptr++; + } + } + } + +/* In the caseful case, we can just compare the bytes, whether or not we +are in UTF-8 mode. */ + +else + { + while (length-- > 0) + { + if (eptr >= md->end_subject) return -2; /* Partial match */ + if (UCHAR21INCTEST(p) != UCHAR21INCTEST(eptr)) return -1; + } + } + +return (int)(eptr - eptr_start); +} + + + +/*************************************************************************** +**************************************************************************** + RECURSION IN THE match() FUNCTION + +The match() function is highly recursive, though not every recursive call +increases the recursive depth. Nevertheless, some regular expressions can cause +it to recurse to a great depth. I was writing for Unix, so I just let it call +itself recursively. This uses the stack for saving everything that has to be +saved for a recursive call. On Unix, the stack can be large, and this works +fine. + +It turns out that on some non-Unix-like systems there are problems with +programs that use a lot of stack. (This despite the fact that every last chip +has oodles of memory these days, and techniques for extending the stack have +been known for decades.) So.... + +There is a fudge, triggered by defining NO_RECURSE, which avoids recursive +calls by keeping local variables that need to be preserved in blocks of memory +obtained from malloc() instead instead of on the stack. Macros are used to +achieve this so that the actual code doesn't look very different to what it +always used to. + +The original heap-recursive code used longjmp(). However, it seems that this +can be very slow on some operating systems. Following a suggestion from Stan +Switzer, the use of longjmp() has been abolished, at the cost of having to +provide a unique number for each call to RMATCH. There is no way of generating +a sequence of numbers at compile time in C. I have given them names, to make +them stand out more clearly. + +Crude tests on x86 Linux show a small speedup of around 5-8%. However, on +FreeBSD, avoiding longjmp() more than halves the time taken to run the standard +tests. Furthermore, not using longjmp() means that local dynamic variables +don't have indeterminate values; this has meant that the frame size can be +reduced because the result can be "passed back" by straight setting of the +variable instead of being passed in the frame. +**************************************************************************** +***************************************************************************/ + +/* Numbers for RMATCH calls. When this list is changed, the code at HEAP_RETURN +below must be updated in sync. */ + +enum { RM1=1, RM2, RM3, RM4, RM5, RM6, RM7, RM8, RM9, RM10, + RM11, RM12, RM13, RM14, RM15, RM16, RM17, RM18, RM19, RM20, + RM21, RM22, RM23, RM24, RM25, RM26, RM27, RM28, RM29, RM30, + RM31, RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40, + RM41, RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50, + RM51, RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60, + RM61, RM62, RM63, RM64, RM65, RM66, RM67 }; + +/* These versions of the macros use the stack, as normal. There are debugging +versions and production versions. Note that the "rw" argument of RMATCH isn't +actually used in this definition. */ + +#ifndef NO_RECURSE +#define REGISTER register + +#ifdef PCRE_DEBUG +#define RMATCH(ra,rb,rc,rd,re,rw) \ + { \ + printf("match() called in line %d\n", __LINE__); \ + rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1); \ + printf("to line %d\n", __LINE__); \ + } +#define RRETURN(ra) \ + { \ + printf("match() returned %d from line %d\n", ra, __LINE__); \ + return ra; \ + } +#else +#define RMATCH(ra,rb,rc,rd,re,rw) \ + rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1) +#define RRETURN(ra) return ra +#endif + +#else + + +/* These versions of the macros manage a private stack on the heap. Note that +the "rd" argument of RMATCH isn't actually used in this definition. It's the md +argument of match(), which never changes. */ + +#define REGISTER + +#define RMATCH(ra,rb,rc,rd,re,rw)\ + {\ + heapframe *newframe = frame->Xnextframe;\ + if (newframe == NULL)\ + {\ + newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\ + if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\ + newframe->Xnextframe = NULL;\ + frame->Xnextframe = newframe;\ + }\ + frame->Xwhere = rw;\ + newframe->Xeptr = ra;\ + newframe->Xecode = rb;\ + newframe->Xmstart = mstart;\ + newframe->Xoffset_top = rc;\ + newframe->Xeptrb = re;\ + newframe->Xrdepth = frame->Xrdepth + 1;\ + newframe->Xprevframe = frame;\ + frame = newframe;\ + DPRINTF(("restarting from line %d\n", __LINE__));\ + goto HEAP_RECURSE;\ + L_##rw:\ + DPRINTF(("jumped back to line %d\n", __LINE__));\ + } + +#define RRETURN(ra)\ + {\ + heapframe *oldframe = frame;\ + frame = oldframe->Xprevframe;\ + if (frame != NULL)\ + {\ + rrc = ra;\ + goto HEAP_RETURN;\ + }\ + return ra;\ + } + + +/* Structure for remembering the local variables in a private frame */ + +typedef struct heapframe { + struct heapframe *Xprevframe; + struct heapframe *Xnextframe; + + /* Function arguments that may change */ + + PCRE_PUCHAR Xeptr; + const pcre_uchar *Xecode; + PCRE_PUCHAR Xmstart; + int Xoffset_top; + eptrblock *Xeptrb; + unsigned int Xrdepth; + + /* Function local variables */ + + PCRE_PUCHAR Xcallpat; +#ifdef SUPPORT_UTF + PCRE_PUCHAR Xcharptr; +#endif + PCRE_PUCHAR Xdata; + PCRE_PUCHAR Xnext; + PCRE_PUCHAR Xpp; + PCRE_PUCHAR Xprev; + PCRE_PUCHAR Xsaved_eptr; + + recursion_info Xnew_recursive; + + BOOL Xcur_is_word; + BOOL Xcondition; + BOOL Xprev_is_word; + +#ifdef SUPPORT_UCP + int Xprop_type; + unsigned int Xprop_value; + int Xprop_fail_result; + int Xoclength; + pcre_uchar Xocchars[6]; +#endif + + int Xcodelink; + int Xctype; + unsigned int Xfc; + int Xfi; + int Xlength; + int Xmax; + int Xmin; + unsigned int Xnumber; + int Xoffset; + unsigned int Xop; + pcre_int32 Xsave_capture_last; + int Xsave_offset1, Xsave_offset2, Xsave_offset3; + int Xstacksave[REC_STACK_SAVE_MAX]; + + eptrblock Xnewptrb; + + /* Where to jump back to */ + + int Xwhere; + +} heapframe; + +#endif + + +/*************************************************************************** +***************************************************************************/ + + + +/************************************************* +* Match from current position * +*************************************************/ + +/* This function is called recursively in many circumstances. Whenever it +returns a negative (error) response, the outer incarnation must also return the +same response. */ + +/* These macros pack up tests that are used for partial matching, and which +appear several times in the code. We set the "hit end" flag if the pointer is +at the end of the subject and also past the start of the subject (i.e. +something has been matched). For hard partial matching, we then return +immediately. The second one is used when we already know we are past the end of +the subject. */ + +#define CHECK_PARTIAL()\ + if (md->partial != 0 && eptr >= md->end_subject && \ + eptr > md->start_used_ptr) \ + { \ + md->hitend = TRUE; \ + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); \ + } + +#define SCHECK_PARTIAL()\ + if (md->partial != 0 && eptr > md->start_used_ptr) \ + { \ + md->hitend = TRUE; \ + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); \ + } + + +/* Performance note: It might be tempting to extract commonly used fields from +the md structure (e.g. utf, end_subject) into individual variables to improve +performance. Tests using gcc on a SPARC disproved this; in the first case, it +made performance worse. + +Arguments: + eptr pointer to current character in subject + ecode pointer to current position in compiled code + mstart pointer to the current match start position (can be modified + by encountering \K) + offset_top current top pointer + md pointer to "static" info for the match + eptrb pointer to chain of blocks containing eptr at start of + brackets - for testing for empty matches + rdepth the recursion depth + +Returns: MATCH_MATCH if matched ) these values are >= 0 + MATCH_NOMATCH if failed to match ) + a negative MATCH_xxx value for PRUNE, SKIP, etc + a negative PCRE_ERROR_xxx value if aborted by an error condition + (e.g. stopped by repeated call or recursion limit) +*/ + +static int +match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode, + PCRE_PUCHAR mstart, int offset_top, match_data *md, eptrblock *eptrb, + unsigned int rdepth) +{ +/* These variables do not need to be preserved over recursion in this function, +so they can be ordinary variables in all cases. Mark some of them with +"register" because they are used a lot in loops. */ + +register int rrc; /* Returns from recursive calls */ +register int i; /* Used for loops not involving calls to RMATCH() */ +register pcre_uint32 c; /* Character values not kept over RMATCH() calls */ +register BOOL utf; /* Local copy of UTF flag for speed */ + +BOOL minimize, possessive; /* Quantifier options */ +BOOL caseless; +int condcode; + +/* When recursion is not being used, all "local" variables that have to be +preserved over calls to RMATCH() are part of a "frame". We set up the top-level +frame on the stack here; subsequent instantiations are obtained from the heap +whenever RMATCH() does a "recursion". See the macro definitions above. Putting +the top-level on the stack rather than malloc-ing them all gives a performance +boost in many cases where there is not much "recursion". */ + +#ifdef NO_RECURSE +heapframe *frame = (heapframe *)md->match_frames_base; + +/* Copy in the original argument variables */ + +frame->Xeptr = eptr; +frame->Xecode = ecode; +frame->Xmstart = mstart; +frame->Xoffset_top = offset_top; +frame->Xeptrb = eptrb; +frame->Xrdepth = rdepth; + +/* This is where control jumps back to to effect "recursion" */ + +HEAP_RECURSE: + +/* Macros make the argument variables come from the current frame */ + +#define eptr frame->Xeptr +#define ecode frame->Xecode +#define mstart frame->Xmstart +#define offset_top frame->Xoffset_top +#define eptrb frame->Xeptrb +#define rdepth frame->Xrdepth + +/* Ditto for the local variables */ + +#ifdef SUPPORT_UTF +#define charptr frame->Xcharptr +#endif +#define callpat frame->Xcallpat +#define codelink frame->Xcodelink +#define data frame->Xdata +#define next frame->Xnext +#define pp frame->Xpp +#define prev frame->Xprev +#define saved_eptr frame->Xsaved_eptr + +#define new_recursive frame->Xnew_recursive + +#define cur_is_word frame->Xcur_is_word +#define condition frame->Xcondition +#define prev_is_word frame->Xprev_is_word + +#ifdef SUPPORT_UCP +#define prop_type frame->Xprop_type +#define prop_value frame->Xprop_value +#define prop_fail_result frame->Xprop_fail_result +#define oclength frame->Xoclength +#define occhars frame->Xocchars +#endif + +#define ctype frame->Xctype +#define fc frame->Xfc +#define fi frame->Xfi +#define length frame->Xlength +#define max frame->Xmax +#define min frame->Xmin +#define number frame->Xnumber +#define offset frame->Xoffset +#define op frame->Xop +#define save_capture_last frame->Xsave_capture_last +#define save_offset1 frame->Xsave_offset1 +#define save_offset2 frame->Xsave_offset2 +#define save_offset3 frame->Xsave_offset3 +#define stacksave frame->Xstacksave + +#define newptrb frame->Xnewptrb + +/* When recursion is being used, local variables are allocated on the stack and +get preserved during recursion in the normal way. In this environment, fi and +i, and fc and c, can be the same variables. */ + +#else /* NO_RECURSE not defined */ +#define fi i +#define fc c + +/* Many of the following variables are used only in small blocks of the code. +My normal style of coding would have declared them within each of those blocks. +However, in order to accommodate the version of this code that uses an external +"stack" implemented on the heap, it is easier to declare them all here, so the +declarations can be cut out in a block. The only declarations within blocks +below are for variables that do not have to be preserved over a recursive call +to RMATCH(). */ + +#ifdef SUPPORT_UTF +const pcre_uchar *charptr; +#endif +const pcre_uchar *callpat; +const pcre_uchar *data; +const pcre_uchar *next; +PCRE_PUCHAR pp; +const pcre_uchar *prev; +PCRE_PUCHAR saved_eptr; + +recursion_info new_recursive; + +BOOL cur_is_word; +BOOL condition; +BOOL prev_is_word; + +#ifdef SUPPORT_UCP +int prop_type; +unsigned int prop_value; +int prop_fail_result; +int oclength; +pcre_uchar occhars[6]; +#endif + +int codelink; +int ctype; +int length; +int max; +int min; +unsigned int number; +int offset; +unsigned int op; +pcre_int32 save_capture_last; +int save_offset1, save_offset2, save_offset3; +int stacksave[REC_STACK_SAVE_MAX]; + +eptrblock newptrb; + +/* There is a special fudge for calling match() in a way that causes it to +measure the size of its basic stack frame when the stack is being used for +recursion. The second argument (ecode) being NULL triggers this behaviour. It +cannot normally ever be NULL. The return is the negated value of the frame +size. */ + +if (ecode == NULL) + { + if (rdepth == 0) + return match((PCRE_PUCHAR)&rdepth, NULL, NULL, 0, NULL, NULL, 1); + else + { + int len = (char *)&rdepth - (char *)eptr; + return (len > 0)? -len : len; + } + } +#endif /* NO_RECURSE */ + +/* To save space on the stack and in the heap frame, I have doubled up on some +of the local variables that are used only in localised parts of the code, but +still need to be preserved over recursive calls of match(). These macros define +the alternative names that are used. */ + +#define allow_zero cur_is_word +#define cbegroup condition +#define code_offset codelink +#define condassert condition +#define matched_once prev_is_word +#define foc number +#define save_mark data + +/* These statements are here to stop the compiler complaining about unitialized +variables. */ + +#ifdef SUPPORT_UCP +prop_value = 0; +prop_fail_result = 0; +#endif + + +/* This label is used for tail recursion, which is used in a few cases even +when NO_RECURSE is not defined, in order to reduce the amount of stack that is +used. Thanks to Ian Taylor for noticing this possibility and sending the +original patch. */ + +TAIL_RECURSE: + +/* OK, now we can get on with the real code of the function. Recursive calls +are specified by the macro RMATCH and RRETURN is used to return. When +NO_RECURSE is *not* defined, these just turn into a recursive call to match() +and a "return", respectively (possibly with some debugging if PCRE_DEBUG is +defined). However, RMATCH isn't like a function call because it's quite a +complicated macro. It has to be used in one particular way. This shouldn't, +however, impact performance when true recursion is being used. */ + +#ifdef SUPPORT_UTF +utf = md->utf; /* Local copy of the flag */ +#else +utf = FALSE; +#endif + +/* First check that we haven't called match() too many times, or that we +haven't exceeded the recursive call limit. */ + +if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT); +if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT); + +/* At the start of a group with an unlimited repeat that may match an empty +string, the variable md->match_function_type is set to MATCH_CBEGROUP. It is +done this way to save having to use another function argument, which would take +up space on the stack. See also MATCH_CONDASSERT below. + +When MATCH_CBEGROUP is set, add the current subject pointer to the chain of +such remembered pointers, to be checked when we hit the closing ket, in order +to break infinite loops that match no characters. When match() is called in +other circumstances, don't add to the chain. The MATCH_CBEGROUP feature must +NOT be used with tail recursion, because the memory block that is used is on +the stack, so a new one may be required for each match(). */ + +if (md->match_function_type == MATCH_CBEGROUP) + { + newptrb.epb_saved_eptr = eptr; + newptrb.epb_prev = eptrb; + eptrb = &newptrb; + md->match_function_type = 0; + } + +/* Now start processing the opcodes. */ + +for (;;) + { + minimize = possessive = FALSE; + op = *ecode; + + switch(op) + { + case OP_MARK: + md->nomatch_mark = ecode + 2; + md->mark = NULL; /* In case previously set by assertion */ + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md, + eptrb, RM55); + if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) && + md->mark == NULL) md->mark = ecode + 2; + + /* A return of MATCH_SKIP_ARG means that matching failed at SKIP with an + argument, and we must check whether that argument matches this MARK's + argument. It is passed back in md->start_match_ptr (an overloading of that + variable). If it does match, we reset that variable to the current subject + position and return MATCH_SKIP. Otherwise, pass back the return code + unaltered. */ + + else if (rrc == MATCH_SKIP_ARG && + STRCMP_UC_UC_TEST(ecode + 2, md->start_match_ptr) == 0) + { + md->start_match_ptr = eptr; + RRETURN(MATCH_SKIP); + } + RRETURN(rrc); + + case OP_FAIL: + RRETURN(MATCH_NOMATCH); + + case OP_COMMIT: + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, + eptrb, RM52); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + RRETURN(MATCH_COMMIT); + + case OP_PRUNE: + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, + eptrb, RM51); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + RRETURN(MATCH_PRUNE); + + case OP_PRUNE_ARG: + md->nomatch_mark = ecode + 2; + md->mark = NULL; /* In case previously set by assertion */ + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md, + eptrb, RM56); + if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) && + md->mark == NULL) md->mark = ecode + 2; + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + RRETURN(MATCH_PRUNE); + + case OP_SKIP: + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, + eptrb, RM53); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + md->start_match_ptr = eptr; /* Pass back current position */ + RRETURN(MATCH_SKIP); + + /* Note that, for Perl compatibility, SKIP with an argument does NOT set + nomatch_mark. When a pattern match ends with a SKIP_ARG for which there was + not a matching mark, we have to re-run the match, ignoring the SKIP_ARG + that failed and any that precede it (either they also failed, or were not + triggered). To do this, we maintain a count of executed SKIP_ARGs. If a + SKIP_ARG gets to top level, the match is re-run with md->ignore_skip_arg + set to the count of the one that failed. */ + + case OP_SKIP_ARG: + md->skip_arg_count++; + if (md->skip_arg_count <= md->ignore_skip_arg) + { + ecode += PRIV(OP_lengths)[*ecode] + ecode[1]; + break; + } + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md, + eptrb, RM57); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + + /* Pass back the current skip name by overloading md->start_match_ptr and + returning the special MATCH_SKIP_ARG return code. This will either be + caught by a matching MARK, or get to the top, where it causes a rematch + with md->ignore_skip_arg set to the value of md->skip_arg_count. */ + + md->start_match_ptr = ecode + 2; + RRETURN(MATCH_SKIP_ARG); + + /* For THEN (and THEN_ARG) we pass back the address of the opcode, so that + the branch in which it occurs can be determined. Overload the start of + match pointer to do this. */ + + case OP_THEN: + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, + eptrb, RM54); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + md->start_match_ptr = ecode; + RRETURN(MATCH_THEN); + + case OP_THEN_ARG: + md->nomatch_mark = ecode + 2; + md->mark = NULL; /* In case previously set by assertion */ + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, + md, eptrb, RM58); + if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) && + md->mark == NULL) md->mark = ecode + 2; + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + md->start_match_ptr = ecode; + RRETURN(MATCH_THEN); + + /* Handle an atomic group that does not contain any capturing parentheses. + This can be handled like an assertion. Prior to 8.13, all atomic groups + were handled this way. In 8.13, the code was changed as below for ONCE, so + that backups pass through the group and thereby reset captured values. + However, this uses a lot more stack, so in 8.20, atomic groups that do not + contain any captures generate OP_ONCE_NC, which can be handled in the old, + less stack intensive way. + + Check the alternative branches in turn - the matching won't pass the KET + for this kind of subpattern. If any one branch matches, we carry on as at + the end of a normal bracket, leaving the subject pointer, but resetting + the start-of-match value in case it was changed by \K. */ + + case OP_ONCE_NC: + prev = ecode; + saved_eptr = eptr; + save_mark = md->mark; + do + { + RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64); + if (rrc == MATCH_MATCH) /* Note: _not_ MATCH_ACCEPT */ + { + mstart = md->start_match_ptr; + break; + } + if (rrc == MATCH_THEN) + { + next = ecode + GET(ecode,1); + if (md->start_match_ptr < next && + (*ecode == OP_ALT || *next == OP_ALT)) + rrc = MATCH_NOMATCH; + } + + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + ecode += GET(ecode,1); + md->mark = save_mark; + } + while (*ecode == OP_ALT); + + /* If hit the end of the group (which could be repeated), fail */ + + if (*ecode != OP_ONCE_NC && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH); + + /* Continue as from after the group, updating the offsets high water + mark, since extracts may have been taken. */ + + do ecode += GET(ecode, 1); while (*ecode == OP_ALT); + + offset_top = md->end_offset_top; + eptr = md->end_match_ptr; + + /* For a non-repeating ket, just continue at this level. This also + happens for a repeating ket if no characters were matched in the group. + This is the forcible breaking of infinite loops as implemented in Perl + 5.005. */ + + if (*ecode == OP_KET || eptr == saved_eptr) + { + ecode += 1+LINK_SIZE; + break; + } + + /* The repeating kets try the rest of the pattern or restart from the + preceding bracket, in the appropriate order. The second "call" of match() + uses tail recursion, to avoid using another stack frame. */ + + if (*ecode == OP_KETRMIN) + { + RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM65); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + ecode = prev; + goto TAIL_RECURSE; + } + else /* OP_KETRMAX */ + { + RMATCH(eptr, prev, offset_top, md, eptrb, RM66); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + ecode += 1 + LINK_SIZE; + goto TAIL_RECURSE; + } + /* Control never gets here */ + + /* Handle a capturing bracket, other than those that are possessive with an + unlimited repeat. If there is space in the offset vector, save the current + subject position in the working slot at the top of the vector. We mustn't + change the current values of the data slot, because they may be set from a + previous iteration of this group, and be referred to by a reference inside + the group. A failure to match might occur after the group has succeeded, + if something later on doesn't match. For this reason, we need to restore + the working value and also the values of the final offsets, in case they + were set by a previous iteration of the same bracket. + + If there isn't enough space in the offset vector, treat this as if it were + a non-capturing bracket. Don't worry about setting the flag for the error + case here; that is handled in the code for KET. */ + + case OP_CBRA: + case OP_SCBRA: + number = GET2(ecode, 1+LINK_SIZE); + offset = number << 1; + +#ifdef PCRE_DEBUG + printf("start bracket %d\n", number); + printf("subject="); + pchars(eptr, 16, TRUE, md); + printf("\n"); +#endif + + if (offset < md->offset_max) + { + save_offset1 = md->offset_vector[offset]; + save_offset2 = md->offset_vector[offset+1]; + save_offset3 = md->offset_vector[md->offset_end - number]; + save_capture_last = md->capture_last; + save_mark = md->mark; + + DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3)); + md->offset_vector[md->offset_end - number] = + (int)(eptr - md->start_subject); + + for (;;) + { + if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP; + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, + eptrb, RM1); + if (rrc == MATCH_ONCE) break; /* Backing up through an atomic group */ + + /* If we backed up to a THEN, check whether it is within the current + branch by comparing the address of the THEN that is passed back with + the end of the branch. If it is within the current branch, and the + branch is one of two or more alternatives (it either starts or ends + with OP_ALT), we have reached the limit of THEN's action, so convert + the return code to NOMATCH, which will cause normal backtracking to + happen from now on. Otherwise, THEN is passed back to an outer + alternative. This implements Perl's treatment of parenthesized groups, + where a group not containing | does not affect the current alternative, + that is, (X) is NOT the same as (X|(*F)). */ + + if (rrc == MATCH_THEN) + { + next = ecode + GET(ecode,1); + if (md->start_match_ptr < next && + (*ecode == OP_ALT || *next == OP_ALT)) + rrc = MATCH_NOMATCH; + } + + /* Anything other than NOMATCH is passed back. */ + + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + md->capture_last = save_capture_last; + ecode += GET(ecode, 1); + md->mark = save_mark; + if (*ecode != OP_ALT) break; + } + + DPRINTF(("bracket %d failed\n", number)); + md->offset_vector[offset] = save_offset1; + md->offset_vector[offset+1] = save_offset2; + md->offset_vector[md->offset_end - number] = save_offset3; + + /* At this point, rrc will be one of MATCH_ONCE or MATCH_NOMATCH. */ + + RRETURN(rrc); + } + + /* FALL THROUGH ... Insufficient room for saving captured contents. Treat + as a non-capturing bracket. */ + + /* VVVVVVVVVVVVVVVVVVVVVVVVV */ + /* VVVVVVVVVVVVVVVVVVVVVVVVV */ + + DPRINTF(("insufficient capture room: treat as non-capturing\n")); + + /* VVVVVVVVVVVVVVVVVVVVVVVVV */ + /* VVVVVVVVVVVVVVVVVVVVVVVVV */ + + /* Non-capturing or atomic group, except for possessive with unlimited + repeat and ONCE group with no captures. Loop for all the alternatives. + + When we get to the final alternative within the brackets, we used to return + the result of a recursive call to match() whatever happened so it was + possible to reduce stack usage by turning this into a tail recursion, + except in the case of a possibly empty group. However, now that there is + the possiblity of (*THEN) occurring in the final alternative, this + optimization is no longer always possible. + + We can optimize if we know there are no (*THEN)s in the pattern; at present + this is the best that can be done. + + MATCH_ONCE is returned when the end of an atomic group is successfully + reached, but subsequent matching fails. It passes back up the tree (causing + captured values to be reset) until the original atomic group level is + reached. This is tested by comparing md->once_target with the start of the + group. At this point, the return is converted into MATCH_NOMATCH so that + previous backup points can be taken. */ + + case OP_ONCE: + case OP_BRA: + case OP_SBRA: + DPRINTF(("start non-capturing bracket\n")); + + for (;;) + { + if (op >= OP_SBRA || op == OP_ONCE) + md->match_function_type = MATCH_CBEGROUP; + + /* If this is not a possibly empty group, and there are no (*THEN)s in + the pattern, and this is the final alternative, optimize as described + above. */ + + else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT) + { + ecode += PRIV(OP_lengths)[*ecode]; + goto TAIL_RECURSE; + } + + /* In all other cases, we have to make another call to match(). */ + + save_mark = md->mark; + save_capture_last = md->capture_last; + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb, + RM2); + + /* See comment in the code for capturing groups above about handling + THEN. */ + + if (rrc == MATCH_THEN) + { + next = ecode + GET(ecode,1); + if (md->start_match_ptr < next && + (*ecode == OP_ALT || *next == OP_ALT)) + rrc = MATCH_NOMATCH; + } + + if (rrc != MATCH_NOMATCH) + { + if (rrc == MATCH_ONCE) + { + const pcre_uchar *scode = ecode; + if (*scode != OP_ONCE) /* If not at start, find it */ + { + while (*scode == OP_ALT) scode += GET(scode, 1); + scode -= GET(scode, 1); + } + if (md->once_target == scode) rrc = MATCH_NOMATCH; + } + RRETURN(rrc); + } + ecode += GET(ecode, 1); + md->mark = save_mark; + if (*ecode != OP_ALT) break; + md->capture_last = save_capture_last; + } + + RRETURN(MATCH_NOMATCH); + + /* Handle possessive capturing brackets with an unlimited repeat. We come + here from BRAZERO with allow_zero set TRUE. The offset_vector values are + handled similarly to the normal case above. However, the matching is + different. The end of these brackets will always be OP_KETRPOS, which + returns MATCH_KETRPOS without going further in the pattern. By this means + we can handle the group by iteration rather than recursion, thereby + reducing the amount of stack needed. */ + + case OP_CBRAPOS: + case OP_SCBRAPOS: + allow_zero = FALSE; + + POSSESSIVE_CAPTURE: + number = GET2(ecode, 1+LINK_SIZE); + offset = number << 1; + +#ifdef PCRE_DEBUG + printf("start possessive bracket %d\n", number); + printf("subject="); + pchars(eptr, 16, TRUE, md); + printf("\n"); +#endif + + if (offset >= md->offset_max) goto POSSESSIVE_NON_CAPTURE; + + matched_once = FALSE; + code_offset = (int)(ecode - md->start_code); + + save_offset1 = md->offset_vector[offset]; + save_offset2 = md->offset_vector[offset+1]; + save_offset3 = md->offset_vector[md->offset_end - number]; + save_capture_last = md->capture_last; + + DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3)); + + /* Each time round the loop, save the current subject position for use + when the group matches. For MATCH_MATCH, the group has matched, so we + restart it with a new subject starting position, remembering that we had + at least one match. For MATCH_NOMATCH, carry on with the alternatives, as + usual. If we haven't matched any alternatives in any iteration, check to + see if a previous iteration matched. If so, the group has matched; + continue from afterwards. Otherwise it has failed; restore the previous + capture values before returning NOMATCH. */ + + for (;;) + { + md->offset_vector[md->offset_end - number] = + (int)(eptr - md->start_subject); + if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP; + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, + eptrb, RM63); + if (rrc == MATCH_KETRPOS) + { + offset_top = md->end_offset_top; + ecode = md->start_code + code_offset; + save_capture_last = md->capture_last; + matched_once = TRUE; + mstart = md->start_match_ptr; /* In case \K changed it */ + if (eptr == md->end_match_ptr) /* Matched an empty string */ + { + do ecode += GET(ecode, 1); while (*ecode == OP_ALT); + break; + } + eptr = md->end_match_ptr; + continue; + } + + /* See comment in the code for capturing groups above about handling + THEN. */ + + if (rrc == MATCH_THEN) + { + next = ecode + GET(ecode,1); + if (md->start_match_ptr < next && + (*ecode == OP_ALT || *next == OP_ALT)) + rrc = MATCH_NOMATCH; + } + + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + md->capture_last = save_capture_last; + ecode += GET(ecode, 1); + if (*ecode != OP_ALT) break; + } + + if (!matched_once) + { + md->offset_vector[offset] = save_offset1; + md->offset_vector[offset+1] = save_offset2; + md->offset_vector[md->offset_end - number] = save_offset3; + } + + if (allow_zero || matched_once) + { + ecode += 1 + LINK_SIZE; + break; + } + + RRETURN(MATCH_NOMATCH); + + /* Non-capturing possessive bracket with unlimited repeat. We come here + from BRAZERO with allow_zero = TRUE. The code is similar to the above, + without the capturing complication. It is written out separately for speed + and cleanliness. */ + + case OP_BRAPOS: + case OP_SBRAPOS: + allow_zero = FALSE; + + POSSESSIVE_NON_CAPTURE: + matched_once = FALSE; + code_offset = (int)(ecode - md->start_code); + save_capture_last = md->capture_last; + + for (;;) + { + if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP; + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, + eptrb, RM48); + if (rrc == MATCH_KETRPOS) + { + offset_top = md->end_offset_top; + ecode = md->start_code + code_offset; + matched_once = TRUE; + mstart = md->start_match_ptr; /* In case \K reset it */ + if (eptr == md->end_match_ptr) /* Matched an empty string */ + { + do ecode += GET(ecode, 1); while (*ecode == OP_ALT); + break; + } + eptr = md->end_match_ptr; + continue; + } + + /* See comment in the code for capturing groups above about handling + THEN. */ + + if (rrc == MATCH_THEN) + { + next = ecode + GET(ecode,1); + if (md->start_match_ptr < next && + (*ecode == OP_ALT || *next == OP_ALT)) + rrc = MATCH_NOMATCH; + } + + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + ecode += GET(ecode, 1); + if (*ecode != OP_ALT) break; + md->capture_last = save_capture_last; + } + + if (matched_once || allow_zero) + { + ecode += 1 + LINK_SIZE; + break; + } + RRETURN(MATCH_NOMATCH); + + /* Control never reaches here. */ + + /* Conditional group: compilation checked that there are no more than two + branches. If the condition is false, skipping the first branch takes us + past the end of the item if there is only one branch, but that's exactly + what we want. */ + + case OP_COND: + case OP_SCOND: + + /* The variable codelink will be added to ecode when the condition is + false, to get to the second branch. Setting it to the offset to the ALT + or KET, then incrementing ecode achieves this effect. We now have ecode + pointing to the condition or callout. */ + + codelink = GET(ecode, 1); /* Offset to the second branch */ + ecode += 1 + LINK_SIZE; /* From this opcode */ + + /* Because of the way auto-callout works during compile, a callout item is + inserted between OP_COND and an assertion condition. */ + + if (*ecode == OP_CALLOUT) + { + if (PUBL(callout) != NULL) + { + PUBL(callout_block) cb; + cb.version = 2; /* Version 1 of the callout block */ + cb.callout_number = ecode[1]; + cb.offset_vector = md->offset_vector; +#if defined COMPILE_PCRE8 + cb.subject = (PCRE_SPTR)md->start_subject; +#elif defined COMPILE_PCRE16 + cb.subject = (PCRE_SPTR16)md->start_subject; +#elif defined COMPILE_PCRE32 + cb.subject = (PCRE_SPTR32)md->start_subject; +#endif + cb.subject_length = (int)(md->end_subject - md->start_subject); + cb.start_match = (int)(mstart - md->start_subject); + cb.current_position = (int)(eptr - md->start_subject); + cb.pattern_position = GET(ecode, 2); + cb.next_item_length = GET(ecode, 2 + LINK_SIZE); + cb.capture_top = offset_top/2; + cb.capture_last = md->capture_last & CAPLMASK; + /* Internal change requires this for API compatibility. */ + if (cb.capture_last == 0) cb.capture_last = -1; + cb.callout_data = md->callout_data; + cb.mark = md->nomatch_mark; + if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH); + if (rrc < 0) RRETURN(rrc); + } + + /* Advance ecode past the callout, so it now points to the condition. We + must adjust codelink so that the value of ecode+codelink is unchanged. */ + + ecode += PRIV(OP_lengths)[OP_CALLOUT]; + codelink -= PRIV(OP_lengths)[OP_CALLOUT]; + } + + /* Test the various possible conditions */ + + condition = FALSE; + switch(condcode = *ecode) + { + case OP_RREF: /* Numbered group recursion test */ + if (md->recursive != NULL) /* Not recursing => FALSE */ + { + unsigned int recno = GET2(ecode, 1); /* Recursion group number*/ + condition = (recno == RREF_ANY || recno == md->recursive->group_num); + } + break; + + case OP_DNRREF: /* Duplicate named group recursion test */ + if (md->recursive != NULL) + { + int count = GET2(ecode, 1 + IMM2_SIZE); + pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size; + while (count-- > 0) + { + unsigned int recno = GET2(slot, 0); + condition = recno == md->recursive->group_num; + if (condition) break; + slot += md->name_entry_size; + } + } + break; + + case OP_CREF: /* Numbered group used test */ + offset = GET2(ecode, 1) << 1; /* Doubled ref number */ + condition = offset < offset_top && md->offset_vector[offset] >= 0; + break; + + case OP_DNCREF: /* Duplicate named group used test */ + { + int count = GET2(ecode, 1 + IMM2_SIZE); + pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size; + while (count-- > 0) + { + offset = GET2(slot, 0) << 1; + condition = offset < offset_top && md->offset_vector[offset] >= 0; + if (condition) break; + slot += md->name_entry_size; + } + } + break; + + case OP_DEF: /* DEFINE - always false */ + case OP_FAIL: /* From optimized (?!) condition */ + break; + + /* The condition is an assertion. Call match() to evaluate it - setting + md->match_function_type to MATCH_CONDASSERT causes it to stop at the end + of an assertion. */ + + default: + md->match_function_type = MATCH_CONDASSERT; + RMATCH(eptr, ecode, offset_top, md, NULL, RM3); + if (rrc == MATCH_MATCH) + { + if (md->end_offset_top > offset_top) + offset_top = md->end_offset_top; /* Captures may have happened */ + condition = TRUE; + + /* Advance ecode past the assertion to the start of the first branch, + but adjust it so that the general choosing code below works. If the + assertion has a quantifier that allows zero repeats we must skip over + the BRAZERO. This is a lunatic thing to do, but somebody did! */ + + if (*ecode == OP_BRAZERO) ecode++; + ecode += GET(ecode, 1); + while (*ecode == OP_ALT) ecode += GET(ecode, 1); + ecode += 1 + LINK_SIZE - PRIV(OP_lengths)[condcode]; + } + + /* PCRE doesn't allow the effect of (*THEN) to escape beyond an + assertion; it is therefore treated as NOMATCH. Any other return is an + error. */ + + else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) + { + RRETURN(rrc); /* Need braces because of following else */ + } + break; + } + + /* Choose branch according to the condition */ + + ecode += condition? PRIV(OP_lengths)[condcode] : codelink; + + /* We are now at the branch that is to be obeyed. As there is only one, we + can use tail recursion to avoid using another stack frame, except when + there is unlimited repeat of a possibly empty group. In the latter case, a + recursive call to match() is always required, unless the second alternative + doesn't exist, in which case we can just plough on. Note that, for + compatibility with Perl, the | in a conditional group is NOT treated as + creating two alternatives. If a THEN is encountered in the branch, it + propagates out to the enclosing alternative (unless nested in a deeper set + of alternatives, of course). */ + + if (condition || ecode[-(1+LINK_SIZE)] == OP_ALT) + { + if (op != OP_SCOND) + { + goto TAIL_RECURSE; + } + + md->match_function_type = MATCH_CBEGROUP; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM49); + RRETURN(rrc); + } + + /* Condition false & no alternative; continue after the group. */ + + else + { + } + break; + + + /* Before OP_ACCEPT there may be any number of OP_CLOSE opcodes, + to close any currently open capturing brackets. */ + + case OP_CLOSE: + number = GET2(ecode, 1); /* Must be less than 65536 */ + offset = number << 1; + +#ifdef PCRE_DEBUG + printf("end bracket %d at *ACCEPT", number); + printf("\n"); +#endif + + md->capture_last = (md->capture_last & OVFLMASK) | number; + if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else + { + md->offset_vector[offset] = + md->offset_vector[md->offset_end - number]; + md->offset_vector[offset+1] = (int)(eptr - md->start_subject); + + /* If this group is at or above the current highwater mark, ensure that + any groups between the current high water mark and this group are marked + unset and then update the high water mark. */ + + if (offset >= offset_top) + { + register int *iptr = md->offset_vector + offset_top; + register int *iend = md->offset_vector + offset; + while (iptr < iend) *iptr++ = -1; + offset_top = offset + 2; + } + } + ecode += 1 + IMM2_SIZE; + break; + + + /* End of the pattern, either real or forced. */ + + case OP_END: + case OP_ACCEPT: + case OP_ASSERT_ACCEPT: + + /* If we have matched an empty string, fail if not in an assertion and not + in a recursion if either PCRE_NOTEMPTY is set, or if PCRE_NOTEMPTY_ATSTART + is set and we have matched at the start of the subject. In both cases, + backtracking will then try other alternatives, if any. */ + + if (eptr == mstart && op != OP_ASSERT_ACCEPT && + md->recursive == NULL && + (md->notempty || + (md->notempty_atstart && + mstart == md->start_subject + md->start_offset))) + RRETURN(MATCH_NOMATCH); + + /* Otherwise, we have a match. */ + + md->end_match_ptr = eptr; /* Record where we ended */ + md->end_offset_top = offset_top; /* and how many extracts were taken */ + md->start_match_ptr = mstart; /* and the start (\K can modify) */ + + /* For some reason, the macros don't work properly if an expression is + given as the argument to RRETURN when the heap is in use. */ + + rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT; + RRETURN(rrc); + + /* Assertion brackets. Check the alternative branches in turn - the + matching won't pass the KET for an assertion. If any one branch matches, + the assertion is true. Lookbehind assertions have an OP_REVERSE item at the + start of each branch to move the current point backwards, so the code at + this level is identical to the lookahead case. When the assertion is part + of a condition, we want to return immediately afterwards. The caller of + this incarnation of the match() function will have set MATCH_CONDASSERT in + md->match_function type, and one of these opcodes will be the first opcode + that is processed. We use a local variable that is preserved over calls to + match() to remember this case. */ + + case OP_ASSERT: + case OP_ASSERTBACK: + save_mark = md->mark; + if (md->match_function_type == MATCH_CONDASSERT) + { + condassert = TRUE; + md->match_function_type = 0; + } + else condassert = FALSE; + + /* Loop for each branch */ + + do + { + RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4); + + /* A match means that the assertion is true; break out of the loop + that matches its alternatives. */ + + if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) + { + mstart = md->start_match_ptr; /* In case \K reset it */ + break; + } + + /* If not matched, restore the previous mark setting. */ + + md->mark = save_mark; + + /* See comment in the code for capturing groups above about handling + THEN. */ + + if (rrc == MATCH_THEN) + { + next = ecode + GET(ecode,1); + if (md->start_match_ptr < next && + (*ecode == OP_ALT || *next == OP_ALT)) + rrc = MATCH_NOMATCH; + } + + /* Anything other than NOMATCH causes the entire assertion to fail, + passing back the return code. This includes COMMIT, SKIP, PRUNE and an + uncaptured THEN, which means they take their normal effect. This + consistent approach does not always have exactly the same effect as in + Perl. */ + + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + ecode += GET(ecode, 1); + } + while (*ecode == OP_ALT); /* Continue for next alternative */ + + /* If we have tried all the alternative branches, the assertion has + failed. If not, we broke out after a match. */ + + if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH); + + /* If checking an assertion for a condition, return MATCH_MATCH. */ + + if (condassert) RRETURN(MATCH_MATCH); + + /* Continue from after a successful assertion, updating the offsets high + water mark, since extracts may have been taken during the assertion. */ + + do ecode += GET(ecode,1); while (*ecode == OP_ALT); + ecode += 1 + LINK_SIZE; + offset_top = md->end_offset_top; + continue; + + /* Negative assertion: all branches must fail to match for the assertion to + succeed. */ + + case OP_ASSERT_NOT: + case OP_ASSERTBACK_NOT: + save_mark = md->mark; + if (md->match_function_type == MATCH_CONDASSERT) + { + condassert = TRUE; + md->match_function_type = 0; + } + else condassert = FALSE; + + /* Loop for each alternative branch. */ + + do + { + RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5); + md->mark = save_mark; /* Always restore the mark setting */ + + switch(rrc) + { + case MATCH_MATCH: /* A successful match means */ + case MATCH_ACCEPT: /* the assertion has failed. */ + RRETURN(MATCH_NOMATCH); + + case MATCH_NOMATCH: /* Carry on with next branch */ + break; + + /* See comment in the code for capturing groups above about handling + THEN. */ + + case MATCH_THEN: + next = ecode + GET(ecode,1); + if (md->start_match_ptr < next && + (*ecode == OP_ALT || *next == OP_ALT)) + { + rrc = MATCH_NOMATCH; + break; + } + /* Otherwise fall through. */ + + /* COMMIT, SKIP, PRUNE, and an uncaptured THEN cause the whole + assertion to fail to match, without considering any more alternatives. + Failing to match means the assertion is true. This is a consistent + approach, but does not always have the same effect as in Perl. */ + + case MATCH_COMMIT: + case MATCH_SKIP: + case MATCH_SKIP_ARG: + case MATCH_PRUNE: + do ecode += GET(ecode,1); while (*ecode == OP_ALT); + goto NEG_ASSERT_TRUE; /* Break out of alternation loop */ + + /* Anything else is an error */ + + default: + RRETURN(rrc); + } + + /* Continue with next branch */ + + ecode += GET(ecode,1); + } + while (*ecode == OP_ALT); + + /* All branches in the assertion failed to match. */ + + NEG_ASSERT_TRUE: + if (condassert) RRETURN(MATCH_MATCH); /* Condition assertion */ + ecode += 1 + LINK_SIZE; /* Continue with current branch */ + continue; + + /* Move the subject pointer back. This occurs only at the start of + each branch of a lookbehind assertion. If we are too close to the start to + move back, this match function fails. When working with UTF-8 we move + back a number of characters, not bytes. */ + + case OP_REVERSE: +#ifdef SUPPORT_UTF + if (utf) + { + i = GET(ecode, 1); + while (i-- > 0) + { + eptr--; + if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH); + BACKCHAR(eptr); + } + } + else +#endif + + /* No UTF-8 support, or not in UTF-8 mode: count is byte count */ + + { + eptr -= GET(ecode, 1); + if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH); + } + + /* Save the earliest consulted character, then skip to next op code */ + + if (eptr < md->start_used_ptr) md->start_used_ptr = eptr; + ecode += 1 + LINK_SIZE; + break; + + /* The callout item calls an external function, if one is provided, passing + details of the match so far. This is mainly for debugging, though the + function is able to force a failure. */ + + case OP_CALLOUT: + if (PUBL(callout) != NULL) + { + PUBL(callout_block) cb; + cb.version = 2; /* Version 1 of the callout block */ + cb.callout_number = ecode[1]; + cb.offset_vector = md->offset_vector; +#if defined COMPILE_PCRE8 + cb.subject = (PCRE_SPTR)md->start_subject; +#elif defined COMPILE_PCRE16 + cb.subject = (PCRE_SPTR16)md->start_subject; +#elif defined COMPILE_PCRE32 + cb.subject = (PCRE_SPTR32)md->start_subject; +#endif + cb.subject_length = (int)(md->end_subject - md->start_subject); + cb.start_match = (int)(mstart - md->start_subject); + cb.current_position = (int)(eptr - md->start_subject); + cb.pattern_position = GET(ecode, 2); + cb.next_item_length = GET(ecode, 2 + LINK_SIZE); + cb.capture_top = offset_top/2; + cb.capture_last = md->capture_last & CAPLMASK; + /* Internal change requires this for API compatibility. */ + if (cb.capture_last == 0) cb.capture_last = -1; + cb.callout_data = md->callout_data; + cb.mark = md->nomatch_mark; + if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH); + if (rrc < 0) RRETURN(rrc); + } + ecode += 2 + 2*LINK_SIZE; + break; + + /* Recursion either matches the current regex, or some subexpression. The + offset data is the offset to the starting bracket from the start of the + whole pattern. (This is so that it works from duplicated subpatterns.) + + The state of the capturing groups is preserved over recursion, and + re-instated afterwards. We don't know how many are started and not yet + finished (offset_top records the completed total) so we just have to save + all the potential data. There may be up to 65535 such values, which is too + large to put on the stack, but using malloc for small numbers seems + expensive. As a compromise, the stack is used when there are no more than + REC_STACK_SAVE_MAX values to store; otherwise malloc is used. + + There are also other values that have to be saved. We use a chained + sequence of blocks that actually live on the stack. Thanks to Robin Houston + for the original version of this logic. It has, however, been hacked around + a lot, so he is not to blame for the current way it works. */ + + case OP_RECURSE: + { + recursion_info *ri; + unsigned int recno; + + callpat = md->start_code + GET(ecode, 1); + recno = (callpat == md->start_code)? 0 : + GET2(callpat, 1 + LINK_SIZE); + + /* Check for repeating a recursion without advancing the subject pointer. + This should catch convoluted mutual recursions. (Some simple cases are + caught at compile time.) */ + + for (ri = md->recursive; ri != NULL; ri = ri->prevrec) + if (recno == ri->group_num && eptr == ri->subject_position) + RRETURN(PCRE_ERROR_RECURSELOOP); + + /* Add to "recursing stack" */ + + new_recursive.group_num = recno; + new_recursive.saved_capture_last = md->capture_last; + new_recursive.subject_position = eptr; + new_recursive.prevrec = md->recursive; + md->recursive = &new_recursive; + + /* Where to continue from afterwards */ + + ecode += 1 + LINK_SIZE; + + /* Now save the offset data */ + + new_recursive.saved_max = md->offset_end; + if (new_recursive.saved_max <= REC_STACK_SAVE_MAX) + new_recursive.offset_save = stacksave; + else + { + new_recursive.offset_save = + (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int)); + if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY); + } + memcpy(new_recursive.offset_save, md->offset_vector, + new_recursive.saved_max * sizeof(int)); + + /* OK, now we can do the recursion. After processing each alternative, + restore the offset data and the last captured value. If there were nested + recursions, md->recursive might be changed, so reset it before looping. + */ + + DPRINTF(("Recursing into group %d\n", new_recursive.group_num)); + cbegroup = (*callpat >= OP_SBRA); + do + { + if (cbegroup) md->match_function_type = MATCH_CBEGROUP; + RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top, + md, eptrb, RM6); + memcpy(md->offset_vector, new_recursive.offset_save, + new_recursive.saved_max * sizeof(int)); + md->capture_last = new_recursive.saved_capture_last; + md->recursive = new_recursive.prevrec; + if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) + { + DPRINTF(("Recursion matched\n")); + if (new_recursive.offset_save != stacksave) + (PUBL(free))(new_recursive.offset_save); + + /* Set where we got to in the subject, and reset the start in case + it was changed by \K. This *is* propagated back out of a recursion, + for Perl compatibility. */ + + eptr = md->end_match_ptr; + mstart = md->start_match_ptr; + goto RECURSION_MATCHED; /* Exit loop; end processing */ + } + + /* PCRE does not allow THEN, SKIP, PRUNE or COMMIT to escape beyond a + recursion; they cause a NOMATCH for the entire recursion. These codes + are defined in a range that can be tested for. */ + + if (rrc >= MATCH_BACKTRACK_MIN && rrc <= MATCH_BACKTRACK_MAX) + { + if (new_recursive.offset_save != stacksave) + (PUBL(free))(new_recursive.offset_save); + RRETURN(MATCH_NOMATCH); + } + + /* Any return code other than NOMATCH is an error. */ + + if (rrc != MATCH_NOMATCH) + { + DPRINTF(("Recursion gave error %d\n", rrc)); + if (new_recursive.offset_save != stacksave) + (PUBL(free))(new_recursive.offset_save); + RRETURN(rrc); + } + + md->recursive = &new_recursive; + callpat += GET(callpat, 1); + } + while (*callpat == OP_ALT); + + DPRINTF(("Recursion didn't match\n")); + md->recursive = new_recursive.prevrec; + if (new_recursive.offset_save != stacksave) + (PUBL(free))(new_recursive.offset_save); + RRETURN(MATCH_NOMATCH); + } + + RECURSION_MATCHED: + break; + + /* An alternation is the end of a branch; scan along to find the end of the + bracketed group and go to there. */ + + case OP_ALT: + do ecode += GET(ecode,1); while (*ecode == OP_ALT); + break; + + /* BRAZERO, BRAMINZERO and SKIPZERO occur just before a bracket group, + indicating that it may occur zero times. It may repeat infinitely, or not + at all - i.e. it could be ()* or ()? or even (){0} in the pattern. Brackets + with fixed upper repeat limits are compiled as a number of copies, with the + optional ones preceded by BRAZERO or BRAMINZERO. */ + + case OP_BRAZERO: + next = ecode + 1; + RMATCH(eptr, next, offset_top, md, eptrb, RM10); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + do next += GET(next, 1); while (*next == OP_ALT); + ecode = next + 1 + LINK_SIZE; + break; + + case OP_BRAMINZERO: + next = ecode + 1; + do next += GET(next, 1); while (*next == OP_ALT); + RMATCH(eptr, next + 1+LINK_SIZE, offset_top, md, eptrb, RM11); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + ecode++; + break; + + case OP_SKIPZERO: + next = ecode+1; + do next += GET(next,1); while (*next == OP_ALT); + ecode = next + 1 + LINK_SIZE; + break; + + /* BRAPOSZERO occurs before a possessive bracket group. Don't do anything + here; just jump to the group, with allow_zero set TRUE. */ + + case OP_BRAPOSZERO: + op = *(++ecode); + allow_zero = TRUE; + if (op == OP_CBRAPOS || op == OP_SCBRAPOS) goto POSSESSIVE_CAPTURE; + goto POSSESSIVE_NON_CAPTURE; + + /* End of a group, repeated or non-repeating. */ + + case OP_KET: + case OP_KETRMIN: + case OP_KETRMAX: + case OP_KETRPOS: + prev = ecode - GET(ecode, 1); + + /* If this was a group that remembered the subject start, in order to break + infinite repeats of empty string matches, retrieve the subject start from + the chain. Otherwise, set it NULL. */ + + if (*prev >= OP_SBRA || *prev == OP_ONCE) + { + saved_eptr = eptrb->epb_saved_eptr; /* Value at start of group */ + eptrb = eptrb->epb_prev; /* Backup to previous group */ + } + else saved_eptr = NULL; + + /* If we are at the end of an assertion group or a non-capturing atomic + group, stop matching and return MATCH_MATCH, but record the current high + water mark for use by positive assertions. We also need to record the match + start in case it was changed by \K. */ + + if ((*prev >= OP_ASSERT && *prev <= OP_ASSERTBACK_NOT) || + *prev == OP_ONCE_NC) + { + md->end_match_ptr = eptr; /* For ONCE_NC */ + md->end_offset_top = offset_top; + md->start_match_ptr = mstart; + RRETURN(MATCH_MATCH); /* Sets md->mark */ + } + + /* For capturing groups we have to check the group number back at the start + and if necessary complete handling an extraction by setting the offsets and + bumping the high water mark. Whole-pattern recursion is coded as a recurse + into group 0, so it won't be picked up here. Instead, we catch it when the + OP_END is reached. Other recursion is handled here. We just have to record + the current subject position and start match pointer and give a MATCH + return. */ + + if (*prev == OP_CBRA || *prev == OP_SCBRA || + *prev == OP_CBRAPOS || *prev == OP_SCBRAPOS) + { + number = GET2(prev, 1+LINK_SIZE); + offset = number << 1; + +#ifdef PCRE_DEBUG + printf("end bracket %d", number); + printf("\n"); +#endif + + /* Handle a recursively called group. */ + + if (md->recursive != NULL && md->recursive->group_num == number) + { + md->end_match_ptr = eptr; + md->start_match_ptr = mstart; + RRETURN(MATCH_MATCH); + } + + /* Deal with capturing */ + + md->capture_last = (md->capture_last & OVFLMASK) | number; + if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else + { + /* If offset is greater than offset_top, it means that we are + "skipping" a capturing group, and that group's offsets must be marked + unset. In earlier versions of PCRE, all the offsets were unset at the + start of matching, but this doesn't work because atomic groups and + assertions can cause a value to be set that should later be unset. + Example: matching /(?>(a))b|(a)c/ against "ac". This sets group 1 as + part of the atomic group, but this is not on the final matching path, + so must be unset when 2 is set. (If there is no group 2, there is no + problem, because offset_top will then be 2, indicating no capture.) */ + + if (offset > offset_top) + { + register int *iptr = md->offset_vector + offset_top; + register int *iend = md->offset_vector + offset; + while (iptr < iend) *iptr++ = -1; + } + + /* Now make the extraction */ + + md->offset_vector[offset] = + md->offset_vector[md->offset_end - number]; + md->offset_vector[offset+1] = (int)(eptr - md->start_subject); + if (offset_top <= offset) offset_top = offset + 2; + } + } + + /* OP_KETRPOS is a possessive repeating ket. Remember the current position, + and return the MATCH_KETRPOS. This makes it possible to do the repeats one + at a time from the outer level, thus saving stack. This must precede the + empty string test - in this case that test is done at the outer level. */ + + if (*ecode == OP_KETRPOS) + { + md->start_match_ptr = mstart; /* In case \K reset it */ + md->end_match_ptr = eptr; + md->end_offset_top = offset_top; + RRETURN(MATCH_KETRPOS); + } + + /* For an ordinary non-repeating ket, just continue at this level. This + also happens for a repeating ket if no characters were matched in the + group. This is the forcible breaking of infinite loops as implemented in + Perl 5.005. For a non-repeating atomic group that includes captures, + establish a backup point by processing the rest of the pattern at a lower + level. If this results in a NOMATCH return, pass MATCH_ONCE back to the + original OP_ONCE level, thereby bypassing intermediate backup points, but + resetting any captures that happened along the way. */ + + if (*ecode == OP_KET || eptr == saved_eptr) + { + if (*prev == OP_ONCE) + { + RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM12); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + md->once_target = prev; /* Level at which to change to MATCH_NOMATCH */ + RRETURN(MATCH_ONCE); + } + ecode += 1 + LINK_SIZE; /* Carry on at this level */ + break; + } + + /* The normal repeating kets try the rest of the pattern or restart from + the preceding bracket, in the appropriate order. In the second case, we can + use tail recursion to avoid using another stack frame, unless we have an + an atomic group or an unlimited repeat of a group that can match an empty + string. */ + + if (*ecode == OP_KETRMIN) + { + RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM7); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (*prev == OP_ONCE) + { + RMATCH(eptr, prev, offset_top, md, eptrb, RM8); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + md->once_target = prev; /* Level at which to change to MATCH_NOMATCH */ + RRETURN(MATCH_ONCE); + } + if (*prev >= OP_SBRA) /* Could match an empty string */ + { + RMATCH(eptr, prev, offset_top, md, eptrb, RM50); + RRETURN(rrc); + } + ecode = prev; + goto TAIL_RECURSE; + } + else /* OP_KETRMAX */ + { + RMATCH(eptr, prev, offset_top, md, eptrb, RM13); + if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH; + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (*prev == OP_ONCE) + { + RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM9); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + md->once_target = prev; + RRETURN(MATCH_ONCE); + } + ecode += 1 + LINK_SIZE; + goto TAIL_RECURSE; + } + /* Control never gets here */ + + /* Not multiline mode: start of subject assertion, unless notbol. */ + + case OP_CIRC: + if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH); + + /* Start of subject assertion */ + + case OP_SOD: + if (eptr != md->start_subject) RRETURN(MATCH_NOMATCH); + ecode++; + break; + + /* Multiline mode: start of subject unless notbol, or after any newline. */ + + case OP_CIRCM: + if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH); + if (eptr != md->start_subject && + (eptr == md->end_subject || !WAS_NEWLINE(eptr))) + RRETURN(MATCH_NOMATCH); + ecode++; + break; + + /* Start of match assertion */ + + case OP_SOM: + if (eptr != md->start_subject + md->start_offset) RRETURN(MATCH_NOMATCH); + ecode++; + break; + + /* Reset the start of match point */ + + case OP_SET_SOM: + mstart = eptr; + ecode++; + break; + + /* Multiline mode: assert before any newline, or before end of subject + unless noteol is set. */ + + case OP_DOLLM: + if (eptr < md->end_subject) + { + if (!IS_NEWLINE(eptr)) + { + if (md->partial != 0 && + eptr + 1 >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + UCHAR21TEST(eptr) == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + RRETURN(MATCH_NOMATCH); + } + } + else + { + if (md->noteol) RRETURN(MATCH_NOMATCH); + SCHECK_PARTIAL(); + } + ecode++; + break; + + /* Not multiline mode: assert before a terminating newline or before end of + subject unless noteol is set. */ + + case OP_DOLL: + if (md->noteol) RRETURN(MATCH_NOMATCH); + if (!md->endonly) goto ASSERT_NL_OR_EOS; + + /* ... else fall through for endonly */ + + /* End of subject assertion (\z) */ + + case OP_EOD: + if (eptr < md->end_subject) RRETURN(MATCH_NOMATCH); + SCHECK_PARTIAL(); + ecode++; + break; + + /* End of subject or ending \n assertion (\Z) */ + + case OP_EODN: + ASSERT_NL_OR_EOS: + if (eptr < md->end_subject && + (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen)) + { + if (md->partial != 0 && + eptr + 1 >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + UCHAR21TEST(eptr) == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + RRETURN(MATCH_NOMATCH); + } + + /* Either at end of string or \n before end. */ + + SCHECK_PARTIAL(); + ecode++; + break; + + /* Word boundary assertions */ + + case OP_NOT_WORD_BOUNDARY: + case OP_WORD_BOUNDARY: + { + + /* Find out if the previous and current characters are "word" characters. + It takes a bit more work in UTF-8 mode. Characters > 255 are assumed to + be "non-word" characters. Remember the earliest consulted character for + partial matching. */ + +#ifdef SUPPORT_UTF + if (utf) + { + /* Get status of previous character */ + + if (eptr == md->start_subject) prev_is_word = FALSE; else + { + PCRE_PUCHAR lastptr = eptr - 1; + BACKCHAR(lastptr); + if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr; + GETCHAR(c, lastptr); +#ifdef SUPPORT_UCP + if (md->use_ucp) + { + if (c == '_') prev_is_word = TRUE; else + { + int cat = UCD_CATEGORY(c); + prev_is_word = (cat == ucp_L || cat == ucp_N); + } + } + else +#endif + prev_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0; + } + + /* Get status of next character */ + + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + cur_is_word = FALSE; + } + else + { + GETCHAR(c, eptr); +#ifdef SUPPORT_UCP + if (md->use_ucp) + { + if (c == '_') cur_is_word = TRUE; else + { + int cat = UCD_CATEGORY(c); + cur_is_word = (cat == ucp_L || cat == ucp_N); + } + } + else +#endif + cur_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0; + } + } + else +#endif + + /* Not in UTF-8 mode, but we may still have PCRE_UCP set, and for + consistency with the behaviour of \w we do use it in this case. */ + + { + /* Get status of previous character */ + + if (eptr == md->start_subject) prev_is_word = FALSE; else + { + if (eptr <= md->start_used_ptr) md->start_used_ptr = eptr - 1; +#ifdef SUPPORT_UCP + if (md->use_ucp) + { + c = eptr[-1]; + if (c == '_') prev_is_word = TRUE; else + { + int cat = UCD_CATEGORY(c); + prev_is_word = (cat == ucp_L || cat == ucp_N); + } + } + else +#endif + prev_is_word = MAX_255(eptr[-1]) + && ((md->ctypes[eptr[-1]] & ctype_word) != 0); + } + + /* Get status of next character */ + + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + cur_is_word = FALSE; + } + else +#ifdef SUPPORT_UCP + if (md->use_ucp) + { + c = *eptr; + if (c == '_') cur_is_word = TRUE; else + { + int cat = UCD_CATEGORY(c); + cur_is_word = (cat == ucp_L || cat == ucp_N); + } + } + else +#endif + cur_is_word = MAX_255(*eptr) + && ((md->ctypes[*eptr] & ctype_word) != 0); + } + + /* Now see if the situation is what we want */ + + if ((*ecode++ == OP_WORD_BOUNDARY)? + cur_is_word == prev_is_word : cur_is_word != prev_is_word) + RRETURN(MATCH_NOMATCH); + } + break; + + /* Match any single character type except newline; have to take care with + CRLF newlines and partial matching. */ + + case OP_ANY: + if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); + if (md->partial != 0 && + eptr + 1 >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + UCHAR21TEST(eptr) == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + + /* Fall through */ + + /* Match any single character whatsoever. */ + + case OP_ALLANY: + if (eptr >= md->end_subject) /* DO NOT merge the eptr++ here; it must */ + { /* not be updated before SCHECK_PARTIAL. */ + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + eptr++; +#ifdef SUPPORT_UTF + if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++); +#endif + ecode++; + break; + + /* Match a single byte, even in UTF-8 mode. This opcode really does match + any byte, even newline, independent of the setting of PCRE_DOTALL. */ + + case OP_ANYBYTE: + if (eptr >= md->end_subject) /* DO NOT merge the eptr++ here; it must */ + { /* not be updated before SCHECK_PARTIAL. */ + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + eptr++; + ecode++; + break; + + case OP_NOT_DIGIT: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ( +#if defined SUPPORT_UTF || !(defined COMPILE_PCRE8) + c < 256 && +#endif + (md->ctypes[c] & ctype_digit) != 0 + ) + RRETURN(MATCH_NOMATCH); + ecode++; + break; + + case OP_DIGIT: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ( +#if defined SUPPORT_UTF || !(defined COMPILE_PCRE8) + c > 255 || +#endif + (md->ctypes[c] & ctype_digit) == 0 + ) + RRETURN(MATCH_NOMATCH); + ecode++; + break; + + case OP_NOT_WHITESPACE: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ( +#if defined SUPPORT_UTF || !(defined COMPILE_PCRE8) + c < 256 && +#endif + (md->ctypes[c] & ctype_space) != 0 + ) + RRETURN(MATCH_NOMATCH); + ecode++; + break; + + case OP_WHITESPACE: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ( +#if defined SUPPORT_UTF || !(defined COMPILE_PCRE8) + c > 255 || +#endif + (md->ctypes[c] & ctype_space) == 0 + ) + RRETURN(MATCH_NOMATCH); + ecode++; + break; + + case OP_NOT_WORDCHAR: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ( +#if defined SUPPORT_UTF || !(defined COMPILE_PCRE8) + c < 256 && +#endif + (md->ctypes[c] & ctype_word) != 0 + ) + RRETURN(MATCH_NOMATCH); + ecode++; + break; + + case OP_WORDCHAR: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ( +#if defined SUPPORT_UTF || !(defined COMPILE_PCRE8) + c > 255 || +#endif + (md->ctypes[c] & ctype_word) == 0 + ) + RRETURN(MATCH_NOMATCH); + ecode++; + break; + + case OP_ANYNL: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + switch(c) + { + default: RRETURN(MATCH_NOMATCH); + + case CHAR_CR: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + } + else if (UCHAR21TEST(eptr) == CHAR_LF) eptr++; + break; + + case CHAR_LF: + break; + + case CHAR_VT: + case CHAR_FF: + case CHAR_NEL: +#ifndef EBCDIC + case 0x2028: + case 0x2029: +#endif /* Not EBCDIC */ + if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH); + break; + } + ecode++; + break; + + case OP_NOT_HSPACE: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + switch(c) + { + HSPACE_CASES: RRETURN(MATCH_NOMATCH); /* Byte and multibyte cases */ + default: break; + } + ecode++; + break; + + case OP_HSPACE: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + switch(c) + { + HSPACE_CASES: break; /* Byte and multibyte cases */ + default: RRETURN(MATCH_NOMATCH); + } + ecode++; + break; + + case OP_NOT_VSPACE: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + switch(c) + { + VSPACE_CASES: RRETURN(MATCH_NOMATCH); + default: break; + } + ecode++; + break; + + case OP_VSPACE: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + switch(c) + { + VSPACE_CASES: break; + default: RRETURN(MATCH_NOMATCH); + } + ecode++; + break; + +#ifdef SUPPORT_UCP + /* Check the next character by Unicode property. We will get here only + if the support is in the binary; otherwise a compile-time error occurs. */ + + case OP_PROP: + case OP_NOTPROP: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + { + const pcre_uint32 *cp; + const ucd_record *prop = GET_UCD(c); + + switch(ecode[1]) + { + case PT_ANY: + if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH); + break; + + case PT_LAMP: + if ((prop->chartype == ucp_Lu || + prop->chartype == ucp_Ll || + prop->chartype == ucp_Lt) == (op == OP_NOTPROP)) + RRETURN(MATCH_NOMATCH); + break; + + case PT_GC: + if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP)) + RRETURN(MATCH_NOMATCH); + break; + + case PT_PC: + if ((ecode[2] != prop->chartype) == (op == OP_PROP)) + RRETURN(MATCH_NOMATCH); + break; + + case PT_SC: + if ((ecode[2] != prop->script) == (op == OP_PROP)) + RRETURN(MATCH_NOMATCH); + break; + + /* These are specials */ + + case PT_ALNUM: + if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP)) + RRETURN(MATCH_NOMATCH); + break; + + /* Perl space used to exclude VT, but from Perl 5.18 it is included, + which means that Perl space and POSIX space are now identical. PCRE + was changed at release 8.34. */ + + case PT_SPACE: /* Perl space */ + case PT_PXSPACE: /* POSIX space */ + switch(c) + { + HSPACE_CASES: + VSPACE_CASES: + if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH); + break; + + default: + if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z) == + (op == OP_NOTPROP)) RRETURN(MATCH_NOMATCH); + break; + } + break; + + case PT_WORD: + if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N || + c == CHAR_UNDERSCORE) == (op == OP_NOTPROP)) + RRETURN(MATCH_NOMATCH); + break; + + case PT_CLIST: + cp = PRIV(ucd_caseless_sets) + ecode[2]; + for (;;) + { + if (c < *cp) + { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; } + if (c == *cp++) + { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } } + } + break; + + case PT_UCNC: + if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT || + c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) || + c >= 0xe000) == (op == OP_NOTPROP)) + RRETURN(MATCH_NOMATCH); + break; + + /* This should never occur */ + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } + + ecode += 3; + } + break; + + /* Match an extended Unicode sequence. We will get here only if the support + is in the binary; otherwise a compile-time error occurs. */ + + case OP_EXTUNI: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + else + { + int lgb, rgb; + GETCHARINCTEST(c, eptr); + lgb = UCD_GRAPHBREAK(c); + while (eptr < md->end_subject) + { + int len = 1; + if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); } + rgb = UCD_GRAPHBREAK(c); + if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break; + lgb = rgb; + eptr += len; + } + } + CHECK_PARTIAL(); + ecode++; + break; +#endif /* SUPPORT_UCP */ + + + /* Match a back reference, possibly repeatedly. Look past the end of the + item to see if there is repeat information following. The code is similar + to that for character classes, but repeated for efficiency. Then obey + similar code to character type repeats - written out again for speed. + However, if the referenced string is the empty string, always treat + it as matched, any number of times (otherwise there could be infinite + loops). If the reference is unset, there are two possibilities: + + (a) In the default, Perl-compatible state, set the length negative; + this ensures that every attempt at a match fails. We can't just fail + here, because of the possibility of quantifiers with zero minima. + + (b) If the JavaScript compatibility flag is set, set the length to zero + so that the back reference matches an empty string. + + Otherwise, set the length to the length of what was matched by the + referenced subpattern. + + The OP_REF and OP_REFI opcodes are used for a reference to a numbered group + or to a non-duplicated named group. For a duplicated named group, OP_DNREF + and OP_DNREFI are used. In this case we must scan the list of groups to + which the name refers, and use the first one that is set. */ + + case OP_DNREF: + case OP_DNREFI: + caseless = op == OP_DNREFI; + { + int count = GET2(ecode, 1+IMM2_SIZE); + pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size; + ecode += 1 + 2*IMM2_SIZE; + + /* Setting the default length first and initializing 'offset' avoids + compiler warnings in the REF_REPEAT code. */ + + length = (md->jscript_compat)? 0 : -1; + offset = 0; + + while (count-- > 0) + { + offset = GET2(slot, 0) << 1; + if (offset < offset_top && md->offset_vector[offset] >= 0) + { + length = md->offset_vector[offset+1] - md->offset_vector[offset]; + break; + } + slot += md->name_entry_size; + } + } + goto REF_REPEAT; + + case OP_REF: + case OP_REFI: + caseless = op == OP_REFI; + offset = GET2(ecode, 1) << 1; /* Doubled ref number */ + ecode += 1 + IMM2_SIZE; + if (offset >= offset_top || md->offset_vector[offset] < 0) + length = (md->jscript_compat)? 0 : -1; + else + length = md->offset_vector[offset+1] - md->offset_vector[offset]; + + /* Set up for repetition, or handle the non-repeated case */ + + REF_REPEAT: + switch (*ecode) + { + case OP_CRSTAR: + case OP_CRMINSTAR: + case OP_CRPLUS: + case OP_CRMINPLUS: + case OP_CRQUERY: + case OP_CRMINQUERY: + c = *ecode++ - OP_CRSTAR; + minimize = (c & 1) != 0; + min = rep_min[c]; /* Pick up values from tables; */ + max = rep_max[c]; /* zero for max => infinity */ + if (max == 0) max = INT_MAX; + break; + + case OP_CRRANGE: + case OP_CRMINRANGE: + minimize = (*ecode == OP_CRMINRANGE); + min = GET2(ecode, 1); + max = GET2(ecode, 1 + IMM2_SIZE); + if (max == 0) max = INT_MAX; + ecode += 1 + 2 * IMM2_SIZE; + break; + + default: /* No repeat follows */ + if ((length = match_ref(offset, eptr, length, md, caseless)) < 0) + { + if (length == -2) eptr = md->end_subject; /* Partial match */ + CHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + eptr += length; + continue; /* With the main loop */ + } + + /* Handle repeated back references. If the length of the reference is + zero, just continue with the main loop. If the length is negative, it + means the reference is unset in non-Java-compatible mode. If the minimum is + zero, we can continue at the same level without recursion. For any other + minimum, carrying on will result in NOMATCH. */ + + if (length == 0) continue; + if (length < 0 && min == 0) continue; + + /* First, ensure the minimum number of matches are present. We get back + the length of the reference string explicitly rather than passing the + address of eptr, so that eptr can be a register variable. */ + + for (i = 1; i <= min; i++) + { + int slength; + if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0) + { + if (slength == -2) eptr = md->end_subject; /* Partial match */ + CHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + eptr += slength; + } + + /* If min = max, continue at the same level without recursion. + They are not both allowed to be zero. */ + + if (min == max) continue; + + /* If minimizing, keep trying and advancing the pointer */ + + if (minimize) + { + for (fi = min;; fi++) + { + int slength; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM14); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0) + { + if (slength == -2) eptr = md->end_subject; /* Partial match */ + CHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + eptr += slength; + } + /* Control never gets here */ + } + + /* If maximizing, find the longest string and work backwards */ + + else + { + pp = eptr; + for (i = min; i < max; i++) + { + int slength; + if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0) + { + /* Can't use CHECK_PARTIAL because we don't want to update eptr in + the soft partial matching case. */ + + if (slength == -2 && md->partial != 0 && + md->end_subject > md->start_used_ptr) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + break; + } + eptr += slength; + } + + while (eptr >= pp) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM15); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + eptr -= length; + } + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + /* Match a bit-mapped character class, possibly repeatedly. This op code is + used when all the characters in the class have values in the range 0-255, + and either the matching is caseful, or the characters are in the range + 0-127 when UTF-8 processing is enabled. The only difference between + OP_CLASS and OP_NCLASS occurs when a data character outside the range is + encountered. + + First, look past the end of the item to see if there is repeat information + following. Then obey similar code to character type repeats - written out + again for speed. */ + + case OP_NCLASS: + case OP_CLASS: + { + /* The data variable is saved across frames, so the byte map needs to + be stored there. */ +#define BYTE_MAP ((pcre_uint8 *)data) + data = ecode + 1; /* Save for matching */ + ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */ + + switch (*ecode) + { + case OP_CRSTAR: + case OP_CRMINSTAR: + case OP_CRPLUS: + case OP_CRMINPLUS: + case OP_CRQUERY: + case OP_CRMINQUERY: + case OP_CRPOSSTAR: + case OP_CRPOSPLUS: + case OP_CRPOSQUERY: + c = *ecode++ - OP_CRSTAR; + if (c < OP_CRPOSSTAR - OP_CRSTAR) minimize = (c & 1) != 0; + else possessive = TRUE; + min = rep_min[c]; /* Pick up values from tables; */ + max = rep_max[c]; /* zero for max => infinity */ + if (max == 0) max = INT_MAX; + break; + + case OP_CRRANGE: + case OP_CRMINRANGE: + case OP_CRPOSRANGE: + minimize = (*ecode == OP_CRMINRANGE); + possessive = (*ecode == OP_CRPOSRANGE); + min = GET2(ecode, 1); + max = GET2(ecode, 1 + IMM2_SIZE); + if (max == 0) max = INT_MAX; + ecode += 1 + 2 * IMM2_SIZE; + break; + + default: /* No repeat follows */ + min = max = 1; + break; + } + + /* First, ensure the minimum number of matches are present. */ + +#ifdef SUPPORT_UTF + if (utf) + { + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(c, eptr); + if (c > 255) + { + if (op == OP_CLASS) RRETURN(MATCH_NOMATCH); + } + else + if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH); + } + } + else +#endif + /* Not UTF mode */ + { + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + c = *eptr++; +#ifndef COMPILE_PCRE8 + if (c > 255) + { + if (op == OP_CLASS) RRETURN(MATCH_NOMATCH); + } + else +#endif + if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH); + } + } + + /* If max == min we can continue with the main loop without the + need to recurse. */ + + if (min == max) continue; + + /* If minimizing, keep testing the rest of the expression and advancing + the pointer while it matches the class. */ + + if (minimize) + { +#ifdef SUPPORT_UTF + if (utf) + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM16); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(c, eptr); + if (c > 255) + { + if (op == OP_CLASS) RRETURN(MATCH_NOMATCH); + } + else + if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH); + } + } + else +#endif + /* Not UTF mode */ + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM17); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + c = *eptr++; +#ifndef COMPILE_PCRE8 + if (c > 255) + { + if (op == OP_CLASS) RRETURN(MATCH_NOMATCH); + } + else +#endif + if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH); + } + } + /* Control never gets here */ + } + + /* If maximizing, find the longest possible run, then work backwards. */ + + else + { + pp = eptr; + +#ifdef SUPPORT_UTF + if (utf) + { + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + if (c > 255) + { + if (op == OP_CLASS) break; + } + else + if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break; + eptr += len; + } + + if (possessive) continue; /* No backtracking */ + + for (;;) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM18); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (eptr-- == pp) break; /* Stop if tried at original pos */ + BACKCHAR(eptr); + } + } + else +#endif + /* Not UTF mode */ + { + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + c = *eptr; +#ifndef COMPILE_PCRE8 + if (c > 255) + { + if (op == OP_CLASS) break; + } + else +#endif + if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break; + eptr++; + } + + if (possessive) continue; /* No backtracking */ + + while (eptr >= pp) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM19); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + eptr--; + } + } + + RRETURN(MATCH_NOMATCH); + } +#undef BYTE_MAP + } + /* Control never gets here */ + + + /* Match an extended character class. In the 8-bit library, this opcode is + encountered only when UTF-8 mode mode is supported. In the 16-bit and + 32-bit libraries, codepoints greater than 255 may be encountered even when + UTF is not supported. */ + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + case OP_XCLASS: + { + data = ecode + 1 + LINK_SIZE; /* Save for matching */ + ecode += GET(ecode, 1); /* Advance past the item */ + + switch (*ecode) + { + case OP_CRSTAR: + case OP_CRMINSTAR: + case OP_CRPLUS: + case OP_CRMINPLUS: + case OP_CRQUERY: + case OP_CRMINQUERY: + case OP_CRPOSSTAR: + case OP_CRPOSPLUS: + case OP_CRPOSQUERY: + c = *ecode++ - OP_CRSTAR; + if (c < OP_CRPOSSTAR - OP_CRSTAR) minimize = (c & 1) != 0; + else possessive = TRUE; + min = rep_min[c]; /* Pick up values from tables; */ + max = rep_max[c]; /* zero for max => infinity */ + if (max == 0) max = INT_MAX; + break; + + case OP_CRRANGE: + case OP_CRMINRANGE: + case OP_CRPOSRANGE: + minimize = (*ecode == OP_CRMINRANGE); + possessive = (*ecode == OP_CRPOSRANGE); + min = GET2(ecode, 1); + max = GET2(ecode, 1 + IMM2_SIZE); + if (max == 0) max = INT_MAX; + ecode += 1 + 2 * IMM2_SIZE; + break; + + default: /* No repeat follows */ + min = max = 1; + break; + } + + /* First, ensure the minimum number of matches are present. */ + + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH); + } + + /* If max == min we can continue with the main loop without the + need to recurse. */ + + if (min == max) continue; + + /* If minimizing, keep testing the rest of the expression and advancing + the pointer while it matches the class. */ + + if (minimize) + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM20); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + } + + /* If maximizing, find the longest possible run, then work backwards. */ + + else + { + pp = eptr; + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } +#ifdef SUPPORT_UTF + GETCHARLENTEST(c, eptr, len); +#else + c = *eptr; +#endif + if (!PRIV(xclass)(c, data, utf)) break; + eptr += len; + } + + if (possessive) continue; /* No backtracking */ + + for(;;) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM21); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (eptr-- == pp) break; /* Stop if tried at original pos */ +#ifdef SUPPORT_UTF + if (utf) BACKCHAR(eptr); +#endif + } + RRETURN(MATCH_NOMATCH); + } + + /* Control never gets here */ + } +#endif /* End of XCLASS */ + + /* Match a single character, casefully */ + + case OP_CHAR: +#ifdef SUPPORT_UTF + if (utf) + { + length = 1; + ecode++; + GETCHARLEN(fc, ecode, length); + if (length > md->end_subject - eptr) + { + CHECK_PARTIAL(); /* Not SCHECK_PARTIAL() */ + RRETURN(MATCH_NOMATCH); + } + while (length-- > 0) if (*ecode++ != UCHAR21INC(eptr)) RRETURN(MATCH_NOMATCH); + } + else +#endif + /* Not UTF mode */ + { + if (md->end_subject - eptr < 1) + { + SCHECK_PARTIAL(); /* This one can use SCHECK_PARTIAL() */ + RRETURN(MATCH_NOMATCH); + } + if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH); + ecode += 2; + } + break; + + /* Match a single character, caselessly. If we are at the end of the + subject, give up immediately. */ + + case OP_CHARI: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + +#ifdef SUPPORT_UTF + if (utf) + { + length = 1; + ecode++; + GETCHARLEN(fc, ecode, length); + + /* If the pattern character's value is < 128, we have only one byte, and + we know that its other case must also be one byte long, so we can use the + fast lookup table. We know that there is at least one byte left in the + subject. */ + + if (fc < 128) + { + pcre_uint32 cc = UCHAR21(eptr); + if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH); + ecode++; + eptr++; + } + + /* Otherwise we must pick up the subject character. Note that we cannot + use the value of "length" to check for sufficient bytes left, because the + other case of the character may have more or fewer bytes. */ + + else + { + pcre_uint32 dc; + GETCHARINC(dc, eptr); + ecode += length; + + /* If we have Unicode property support, we can use it to test the other + case of the character, if there is one. */ + + if (fc != dc) + { +#ifdef SUPPORT_UCP + if (dc != UCD_OTHERCASE(fc)) +#endif + RRETURN(MATCH_NOMATCH); + } + } + } + else +#endif /* SUPPORT_UTF */ + + /* Not UTF mode */ + { + if (TABLE_GET(ecode[1], md->lcc, ecode[1]) + != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH); + eptr++; + ecode += 2; + } + break; + + /* Match a single character repeatedly. */ + + case OP_EXACT: + case OP_EXACTI: + min = max = GET2(ecode, 1); + ecode += 1 + IMM2_SIZE; + goto REPEATCHAR; + + case OP_POSUPTO: + case OP_POSUPTOI: + possessive = TRUE; + /* Fall through */ + + case OP_UPTO: + case OP_UPTOI: + case OP_MINUPTO: + case OP_MINUPTOI: + min = 0; + max = GET2(ecode, 1); + minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI; + ecode += 1 + IMM2_SIZE; + goto REPEATCHAR; + + case OP_POSSTAR: + case OP_POSSTARI: + possessive = TRUE; + min = 0; + max = INT_MAX; + ecode++; + goto REPEATCHAR; + + case OP_POSPLUS: + case OP_POSPLUSI: + possessive = TRUE; + min = 1; + max = INT_MAX; + ecode++; + goto REPEATCHAR; + + case OP_POSQUERY: + case OP_POSQUERYI: + possessive = TRUE; + min = 0; + max = 1; + ecode++; + goto REPEATCHAR; + + case OP_STAR: + case OP_STARI: + case OP_MINSTAR: + case OP_MINSTARI: + case OP_PLUS: + case OP_PLUSI: + case OP_MINPLUS: + case OP_MINPLUSI: + case OP_QUERY: + case OP_QUERYI: + case OP_MINQUERY: + case OP_MINQUERYI: + c = *ecode++ - ((op < OP_STARI)? OP_STAR : OP_STARI); + minimize = (c & 1) != 0; + min = rep_min[c]; /* Pick up values from tables; */ + max = rep_max[c]; /* zero for max => infinity */ + if (max == 0) max = INT_MAX; + + /* Common code for all repeated single-character matches. We first check + for the minimum number of characters. If the minimum equals the maximum, we + are done. Otherwise, if minimizing, check the rest of the pattern for a + match; if there isn't one, advance up to the maximum, one character at a + time. + + If maximizing, advance up to the maximum number of matching characters, + until eptr is past the end of the maximum run. If possessive, we are + then done (no backing up). Otherwise, match at this position; anything + other than no match is immediately returned. For nomatch, back up one + character, unless we are matching \R and the last thing matched was + \r\n, in which case, back up two bytes. When we reach the first optional + character position, we can save stack by doing a tail recurse. + + The various UTF/non-UTF and caseful/caseless cases are handled separately, + for speed. */ + + REPEATCHAR: +#ifdef SUPPORT_UTF + if (utf) + { + length = 1; + charptr = ecode; + GETCHARLEN(fc, ecode, length); + ecode += length; + + /* Handle multibyte character matching specially here. There is + support for caseless matching if UCP support is present. */ + + if (length > 1) + { +#ifdef SUPPORT_UCP + pcre_uint32 othercase; + if (op >= OP_STARI && /* Caseless */ + (othercase = UCD_OTHERCASE(fc)) != fc) + oclength = PRIV(ord2utf)(othercase, occhars); + else oclength = 0; +#endif /* SUPPORT_UCP */ + + for (i = 1; i <= min; i++) + { + if (eptr <= md->end_subject - length && + memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length; +#ifdef SUPPORT_UCP + else if (oclength > 0 && + eptr <= md->end_subject - oclength && + memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength; +#endif /* SUPPORT_UCP */ + else + { + CHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + } + + if (min == max) continue; + + if (minimize) + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM22); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr <= md->end_subject - length && + memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length; +#ifdef SUPPORT_UCP + else if (oclength > 0 && + eptr <= md->end_subject - oclength && + memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength; +#endif /* SUPPORT_UCP */ + else + { + CHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + } + /* Control never gets here */ + } + + else /* Maximize */ + { + pp = eptr; + for (i = min; i < max; i++) + { + if (eptr <= md->end_subject - length && + memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length; +#ifdef SUPPORT_UCP + else if (oclength > 0 && + eptr <= md->end_subject - oclength && + memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength; +#endif /* SUPPORT_UCP */ + else + { + CHECK_PARTIAL(); + break; + } + } + + if (possessive) continue; /* No backtracking */ + for(;;) + { + if (eptr <= pp) goto TAIL_RECURSE; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM23); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); +#ifdef SUPPORT_UCP + eptr--; + BACKCHAR(eptr); +#else /* without SUPPORT_UCP */ + eptr -= length; +#endif /* SUPPORT_UCP */ + } + } + /* Control never gets here */ + } + + /* If the length of a UTF-8 character is 1, we fall through here, and + obey the code as for non-UTF-8 characters below, though in this case the + value of fc will always be < 128. */ + } + else +#endif /* SUPPORT_UTF */ + /* When not in UTF-8 mode, load a single-byte character. */ + fc = *ecode++; + + /* The value of fc at this point is always one character, though we may + or may not be in UTF mode. The code is duplicated for the caseless and + caseful cases, for speed, since matching characters is likely to be quite + common. First, ensure the minimum number of matches are present. If min = + max, continue at the same level without recursing. Otherwise, if + minimizing, keep trying the rest of the expression and advancing one + matching character if failing, up to the maximum. Alternatively, if + maximizing, find the maximum number of characters and work backwards. */ + + DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max, + max, (char *)eptr)); + + if (op >= OP_STARI) /* Caseless */ + { +#ifdef COMPILE_PCRE8 + /* fc must be < 128 if UTF is enabled. */ + foc = md->fcc[fc]; +#else +#ifdef SUPPORT_UTF +#ifdef SUPPORT_UCP + if (utf && fc > 127) + foc = UCD_OTHERCASE(fc); +#else + if (utf && fc > 127) + foc = fc; +#endif /* SUPPORT_UCP */ + else +#endif /* SUPPORT_UTF */ + foc = TABLE_GET(fc, md->fcc, fc); +#endif /* COMPILE_PCRE8 */ + + for (i = 1; i <= min; i++) + { + pcre_uint32 cc; /* Faster than pcre_uchar */ + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + cc = UCHAR21TEST(eptr); + if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH); + eptr++; + } + if (min == max) continue; + if (minimize) + { + for (fi = min;; fi++) + { + pcre_uint32 cc; /* Faster than pcre_uchar */ + RMATCH(eptr, ecode, offset_top, md, eptrb, RM24); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + cc = UCHAR21TEST(eptr); + if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH); + eptr++; + } + /* Control never gets here */ + } + else /* Maximize */ + { + pp = eptr; + for (i = min; i < max; i++) + { + pcre_uint32 cc; /* Faster than pcre_uchar */ + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + cc = UCHAR21TEST(eptr); + if (fc != cc && foc != cc) break; + eptr++; + } + if (possessive) continue; /* No backtracking */ + for (;;) + { + if (eptr == pp) goto TAIL_RECURSE; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM25); + eptr--; + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + } + /* Control never gets here */ + } + } + + /* Caseful comparisons (includes all multi-byte characters) */ + + else + { + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (fc != UCHAR21INCTEST(eptr)) RRETURN(MATCH_NOMATCH); + } + + if (min == max) continue; + + if (minimize) + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM26); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (fc != UCHAR21INCTEST(eptr)) RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + } + else /* Maximize */ + { + pp = eptr; + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (fc != UCHAR21TEST(eptr)) break; + eptr++; + } + if (possessive) continue; /* No backtracking */ + for (;;) + { + if (eptr == pp) goto TAIL_RECURSE; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM27); + eptr--; + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + } + /* Control never gets here */ + } + } + /* Control never gets here */ + + /* Match a negated single one-byte character. The character we are + checking can be multibyte. */ + + case OP_NOT: + case OP_NOTI: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } +#ifdef SUPPORT_UTF + if (utf) + { + register pcre_uint32 ch, och; + + ecode++; + GETCHARINC(ch, ecode); + GETCHARINC(c, eptr); + + if (op == OP_NOT) + { + if (ch == c) RRETURN(MATCH_NOMATCH); + } + else + { +#ifdef SUPPORT_UCP + if (ch > 127) + och = UCD_OTHERCASE(ch); +#else + if (ch > 127) + och = ch; +#endif /* SUPPORT_UCP */ + else + och = TABLE_GET(ch, md->fcc, ch); + if (ch == c || och == c) RRETURN(MATCH_NOMATCH); + } + } + else +#endif + { + register pcre_uint32 ch = ecode[1]; + c = *eptr++; + if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c)) + RRETURN(MATCH_NOMATCH); + ecode += 2; + } + break; + + /* Match a negated single one-byte character repeatedly. This is almost a + repeat of the code for a repeated single character, but I haven't found a + nice way of commoning these up that doesn't require a test of the + positive/negative option for each character match. Maybe that wouldn't add + very much to the time taken, but character matching *is* what this is all + about... */ + + case OP_NOTEXACT: + case OP_NOTEXACTI: + min = max = GET2(ecode, 1); + ecode += 1 + IMM2_SIZE; + goto REPEATNOTCHAR; + + case OP_NOTUPTO: + case OP_NOTUPTOI: + case OP_NOTMINUPTO: + case OP_NOTMINUPTOI: + min = 0; + max = GET2(ecode, 1); + minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI; + ecode += 1 + IMM2_SIZE; + goto REPEATNOTCHAR; + + case OP_NOTPOSSTAR: + case OP_NOTPOSSTARI: + possessive = TRUE; + min = 0; + max = INT_MAX; + ecode++; + goto REPEATNOTCHAR; + + case OP_NOTPOSPLUS: + case OP_NOTPOSPLUSI: + possessive = TRUE; + min = 1; + max = INT_MAX; + ecode++; + goto REPEATNOTCHAR; + + case OP_NOTPOSQUERY: + case OP_NOTPOSQUERYI: + possessive = TRUE; + min = 0; + max = 1; + ecode++; + goto REPEATNOTCHAR; + + case OP_NOTPOSUPTO: + case OP_NOTPOSUPTOI: + possessive = TRUE; + min = 0; + max = GET2(ecode, 1); + ecode += 1 + IMM2_SIZE; + goto REPEATNOTCHAR; + + case OP_NOTSTAR: + case OP_NOTSTARI: + case OP_NOTMINSTAR: + case OP_NOTMINSTARI: + case OP_NOTPLUS: + case OP_NOTPLUSI: + case OP_NOTMINPLUS: + case OP_NOTMINPLUSI: + case OP_NOTQUERY: + case OP_NOTQUERYI: + case OP_NOTMINQUERY: + case OP_NOTMINQUERYI: + c = *ecode++ - ((op >= OP_NOTSTARI)? OP_NOTSTARI: OP_NOTSTAR); + minimize = (c & 1) != 0; + min = rep_min[c]; /* Pick up values from tables; */ + max = rep_max[c]; /* zero for max => infinity */ + if (max == 0) max = INT_MAX; + + /* Common code for all repeated single-byte matches. */ + + REPEATNOTCHAR: + GETCHARINCTEST(fc, ecode); + + /* The code is duplicated for the caseless and caseful cases, for speed, + since matching characters is likely to be quite common. First, ensure the + minimum number of matches are present. If min = max, continue at the same + level without recursing. Otherwise, if minimizing, keep trying the rest of + the expression and advancing one matching character if failing, up to the + maximum. Alternatively, if maximizing, find the maximum number of + characters and work backwards. */ + + DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max, + max, (char *)eptr)); + + if (op >= OP_NOTSTARI) /* Caseless */ + { +#ifdef SUPPORT_UTF +#ifdef SUPPORT_UCP + if (utf && fc > 127) + foc = UCD_OTHERCASE(fc); +#else + if (utf && fc > 127) + foc = fc; +#endif /* SUPPORT_UCP */ + else +#endif /* SUPPORT_UTF */ + foc = TABLE_GET(fc, md->fcc, fc); + +#ifdef SUPPORT_UTF + if (utf) + { + register pcre_uint32 d; + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(d, eptr); + if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH); + } + } + else +#endif /* SUPPORT_UTF */ + /* Not UTF mode */ + { + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH); + eptr++; + } + } + + if (min == max) continue; + + if (minimize) + { +#ifdef SUPPORT_UTF + if (utf) + { + register pcre_uint32 d; + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM28); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(d, eptr); + if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH); + } + } + else +#endif /*SUPPORT_UTF */ + /* Not UTF mode */ + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM29); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH); + eptr++; + } + } + /* Control never gets here */ + } + + /* Maximize case */ + + else + { + pp = eptr; + +#ifdef SUPPORT_UTF + if (utf) + { + register pcre_uint32 d; + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(d, eptr, len); + if (fc == d || (unsigned int)foc == d) break; + eptr += len; + } + if (possessive) continue; /* No backtracking */ + for(;;) + { + if (eptr <= pp) goto TAIL_RECURSE; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM30); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + eptr--; + BACKCHAR(eptr); + } + } + else +#endif /* SUPPORT_UTF */ + /* Not UTF mode */ + { + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (fc == *eptr || foc == *eptr) break; + eptr++; + } + if (possessive) continue; /* No backtracking */ + for (;;) + { + if (eptr == pp) goto TAIL_RECURSE; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM31); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + eptr--; + } + } + /* Control never gets here */ + } + } + + /* Caseful comparisons */ + + else + { +#ifdef SUPPORT_UTF + if (utf) + { + register pcre_uint32 d; + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(d, eptr); + if (fc == d) RRETURN(MATCH_NOMATCH); + } + } + else +#endif + /* Not UTF mode */ + { + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (fc == *eptr++) RRETURN(MATCH_NOMATCH); + } + } + + if (min == max) continue; + + if (minimize) + { +#ifdef SUPPORT_UTF + if (utf) + { + register pcre_uint32 d; + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM32); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(d, eptr); + if (fc == d) RRETURN(MATCH_NOMATCH); + } + } + else +#endif + /* Not UTF mode */ + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM33); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (fc == *eptr++) RRETURN(MATCH_NOMATCH); + } + } + /* Control never gets here */ + } + + /* Maximize case */ + + else + { + pp = eptr; + +#ifdef SUPPORT_UTF + if (utf) + { + register pcre_uint32 d; + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(d, eptr, len); + if (fc == d) break; + eptr += len; + } + if (possessive) continue; /* No backtracking */ + for(;;) + { + if (eptr <= pp) goto TAIL_RECURSE; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM34); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + eptr--; + BACKCHAR(eptr); + } + } + else +#endif + /* Not UTF mode */ + { + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (fc == *eptr) break; + eptr++; + } + if (possessive) continue; /* No backtracking */ + for (;;) + { + if (eptr == pp) goto TAIL_RECURSE; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM35); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + eptr--; + } + } + /* Control never gets here */ + } + } + /* Control never gets here */ + + /* Match a single character type repeatedly; several different opcodes + share code. This is very similar to the code for single characters, but we + repeat it in the interests of efficiency. */ + + case OP_TYPEEXACT: + min = max = GET2(ecode, 1); + minimize = TRUE; + ecode += 1 + IMM2_SIZE; + goto REPEATTYPE; + + case OP_TYPEUPTO: + case OP_TYPEMINUPTO: + min = 0; + max = GET2(ecode, 1); + minimize = *ecode == OP_TYPEMINUPTO; + ecode += 1 + IMM2_SIZE; + goto REPEATTYPE; + + case OP_TYPEPOSSTAR: + possessive = TRUE; + min = 0; + max = INT_MAX; + ecode++; + goto REPEATTYPE; + + case OP_TYPEPOSPLUS: + possessive = TRUE; + min = 1; + max = INT_MAX; + ecode++; + goto REPEATTYPE; + + case OP_TYPEPOSQUERY: + possessive = TRUE; + min = 0; + max = 1; + ecode++; + goto REPEATTYPE; + + case OP_TYPEPOSUPTO: + possessive = TRUE; + min = 0; + max = GET2(ecode, 1); + ecode += 1 + IMM2_SIZE; + goto REPEATTYPE; + + case OP_TYPESTAR: + case OP_TYPEMINSTAR: + case OP_TYPEPLUS: + case OP_TYPEMINPLUS: + case OP_TYPEQUERY: + case OP_TYPEMINQUERY: + c = *ecode++ - OP_TYPESTAR; + minimize = (c & 1) != 0; + min = rep_min[c]; /* Pick up values from tables; */ + max = rep_max[c]; /* zero for max => infinity */ + if (max == 0) max = INT_MAX; + + /* Common code for all repeated single character type matches. Note that + in UTF-8 mode, '.' matches a character of any length, but for the other + character types, the valid characters are all one-byte long. */ + + REPEATTYPE: + ctype = *ecode++; /* Code for the character type */ + +#ifdef SUPPORT_UCP + if (ctype == OP_PROP || ctype == OP_NOTPROP) + { + prop_fail_result = ctype == OP_NOTPROP; + prop_type = *ecode++; + prop_value = *ecode++; + } + else prop_type = -1; +#endif + + /* First, ensure the minimum number of matches are present. Use inline + code for maximizing the speed, and do the type test once at the start + (i.e. keep it out of the loop). Separate the UTF-8 code completely as that + is tidier. Also separate the UCP code, which can be the same for both UTF-8 + and single-bytes. */ + + if (min > 0) + { +#ifdef SUPPORT_UCP + if (prop_type >= 0) + { + switch(prop_type) + { + case PT_ANY: + if (prop_fail_result) RRETURN(MATCH_NOMATCH); + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + } + break; + + case PT_LAMP: + for (i = 1; i <= min; i++) + { + int chartype; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + chartype = UCD_CHARTYPE(c); + if ((chartype == ucp_Lu || + chartype == ucp_Ll || + chartype == ucp_Lt) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + case PT_GC: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + case PT_PC: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + case PT_SC: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + case PT_ALNUM: + for (i = 1; i <= min; i++) + { + int category; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + category = UCD_CATEGORY(c); + if ((category == ucp_L || category == ucp_N) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + /* Perl space used to exclude VT, but from Perl 5.18 it is included, + which means that Perl space and POSIX space are now identical. PCRE + was changed at release 8.34. */ + + case PT_SPACE: /* Perl space */ + case PT_PXSPACE: /* POSIX space */ + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + switch(c) + { + HSPACE_CASES: + VSPACE_CASES: + if (prop_fail_result) RRETURN(MATCH_NOMATCH); + break; + + default: + if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + break; + } + } + break; + + case PT_WORD: + for (i = 1; i <= min; i++) + { + int category; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + category = UCD_CATEGORY(c); + if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE) + == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + case PT_CLIST: + for (i = 1; i <= min; i++) + { + const pcre_uint32 *cp; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + cp = PRIV(ucd_caseless_sets) + prop_value; + for (;;) + { + if (c < *cp) + { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } } + if (c == *cp++) + { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; } + } + } + break; + + case PT_UCNC: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT || + c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) || + c >= 0xe000) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + /* This should not occur */ + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } + } + + /* Match extended Unicode sequences. We will get here only if the + support is in the binary; otherwise a compile-time error occurs. */ + + else if (ctype == OP_EXTUNI) + { + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + else + { + int lgb, rgb; + GETCHARINCTEST(c, eptr); + lgb = UCD_GRAPHBREAK(c); + while (eptr < md->end_subject) + { + int len = 1; + if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); } + rgb = UCD_GRAPHBREAK(c); + if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break; + lgb = rgb; + eptr += len; + } + } + CHECK_PARTIAL(); + } + } + + else +#endif /* SUPPORT_UCP */ + +/* Handle all other cases when the coding is UTF-8 */ + +#ifdef SUPPORT_UTF + if (utf) switch(ctype) + { + case OP_ANY: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); + if (md->partial != 0 && + eptr + 1 >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + UCHAR21(eptr) == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + eptr++; + ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++); + } + break; + + case OP_ALLANY: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + eptr++; + ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++); + } + break; + + case OP_ANYBYTE: + if (eptr > md->end_subject - min) RRETURN(MATCH_NOMATCH); + eptr += min; + break; + + case OP_ANYNL: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(c, eptr); + switch(c) + { + default: RRETURN(MATCH_NOMATCH); + + case CHAR_CR: + if (eptr < md->end_subject && UCHAR21(eptr) == CHAR_LF) eptr++; + break; + + case CHAR_LF: + break; + + case CHAR_VT: + case CHAR_FF: + case CHAR_NEL: +#ifndef EBCDIC + case 0x2028: + case 0x2029: +#endif /* Not EBCDIC */ + if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH); + break; + } + } + break; + + case OP_NOT_HSPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(c, eptr); + switch(c) + { + HSPACE_CASES: RRETURN(MATCH_NOMATCH); /* Byte and multibyte cases */ + default: break; + } + } + break; + + case OP_HSPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(c, eptr); + switch(c) + { + HSPACE_CASES: break; /* Byte and multibyte cases */ + default: RRETURN(MATCH_NOMATCH); + } + } + break; + + case OP_NOT_VSPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(c, eptr); + switch(c) + { + VSPACE_CASES: RRETURN(MATCH_NOMATCH); + default: break; + } + } + break; + + case OP_VSPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(c, eptr); + switch(c) + { + VSPACE_CASES: break; + default: RRETURN(MATCH_NOMATCH); + } + } + break; + + case OP_NOT_DIGIT: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(c, eptr); + if (c < 128 && (md->ctypes[c] & ctype_digit) != 0) + RRETURN(MATCH_NOMATCH); + } + break; + + case OP_DIGIT: + for (i = 1; i <= min; i++) + { + pcre_uint32 cc; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + cc = UCHAR21(eptr); + if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0) + RRETURN(MATCH_NOMATCH); + eptr++; + /* No need to skip more bytes - we know it's a 1-byte character */ + } + break; + + case OP_NOT_WHITESPACE: + for (i = 1; i <= min; i++) + { + pcre_uint32 cc; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + cc = UCHAR21(eptr); + if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0) + RRETURN(MATCH_NOMATCH); + eptr++; + ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++); + } + break; + + case OP_WHITESPACE: + for (i = 1; i <= min; i++) + { + pcre_uint32 cc; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + cc = UCHAR21(eptr); + if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0) + RRETURN(MATCH_NOMATCH); + eptr++; + /* No need to skip more bytes - we know it's a 1-byte character */ + } + break; + + case OP_NOT_WORDCHAR: + for (i = 1; i <= min; i++) + { + pcre_uint32 cc; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + cc = UCHAR21(eptr); + if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0) + RRETURN(MATCH_NOMATCH); + eptr++; + ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++); + } + break; + + case OP_WORDCHAR: + for (i = 1; i <= min; i++) + { + pcre_uint32 cc; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + cc = UCHAR21(eptr); + if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0) + RRETURN(MATCH_NOMATCH); + eptr++; + /* No need to skip more bytes - we know it's a 1-byte character */ + } + break; + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } /* End switch(ctype) */ + + else +#endif /* SUPPORT_UTF */ + + /* Code for the non-UTF-8 case for minimum matching of operators other + than OP_PROP and OP_NOTPROP. */ + + switch(ctype) + { + case OP_ANY: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); + if (md->partial != 0 && + eptr + 1 >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + *eptr == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + eptr++; + } + break; + + case OP_ALLANY: + if (eptr > md->end_subject - min) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + eptr += min; + break; + + case OP_ANYBYTE: + if (eptr > md->end_subject - min) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + eptr += min; + break; + + case OP_ANYNL: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + switch(*eptr++) + { + default: RRETURN(MATCH_NOMATCH); + + case CHAR_CR: + if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++; + break; + + case CHAR_LF: + break; + + case CHAR_VT: + case CHAR_FF: + case CHAR_NEL: +#if defined COMPILE_PCRE16 || defined COMPILE_PCRE32 + case 0x2028: + case 0x2029: +#endif + if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH); + break; + } + } + break; + + case OP_NOT_HSPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + switch(*eptr++) + { + default: break; + HSPACE_BYTE_CASES: +#if defined COMPILE_PCRE16 || defined COMPILE_PCRE32 + HSPACE_MULTIBYTE_CASES: +#endif + RRETURN(MATCH_NOMATCH); + } + } + break; + + case OP_HSPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + switch(*eptr++) + { + default: RRETURN(MATCH_NOMATCH); + HSPACE_BYTE_CASES: +#if defined COMPILE_PCRE16 || defined COMPILE_PCRE32 + HSPACE_MULTIBYTE_CASES: +#endif + break; + } + } + break; + + case OP_NOT_VSPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + switch(*eptr++) + { + VSPACE_BYTE_CASES: +#if defined COMPILE_PCRE16 || defined COMPILE_PCRE32 + VSPACE_MULTIBYTE_CASES: +#endif + RRETURN(MATCH_NOMATCH); + default: break; + } + } + break; + + case OP_VSPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + switch(*eptr++) + { + default: RRETURN(MATCH_NOMATCH); + VSPACE_BYTE_CASES: +#if defined COMPILE_PCRE16 || defined COMPILE_PCRE32 + VSPACE_MULTIBYTE_CASES: +#endif + break; + } + } + break; + + case OP_NOT_DIGIT: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0) + RRETURN(MATCH_NOMATCH); + eptr++; + } + break; + + case OP_DIGIT: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0) + RRETURN(MATCH_NOMATCH); + eptr++; + } + break; + + case OP_NOT_WHITESPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0) + RRETURN(MATCH_NOMATCH); + eptr++; + } + break; + + case OP_WHITESPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0) + RRETURN(MATCH_NOMATCH); + eptr++; + } + break; + + case OP_NOT_WORDCHAR: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0) + RRETURN(MATCH_NOMATCH); + eptr++; + } + break; + + case OP_WORDCHAR: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0) + RRETURN(MATCH_NOMATCH); + eptr++; + } + break; + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } + } + + /* If min = max, continue at the same level without recursing */ + + if (min == max) continue; + + /* If minimizing, we have to test the rest of the pattern before each + subsequent match. Again, separate the UTF-8 case for speed, and also + separate the UCP cases. */ + + if (minimize) + { +#ifdef SUPPORT_UCP + if (prop_type >= 0) + { + switch(prop_type) + { + case PT_ANY: + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM36); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if (prop_fail_result) RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + case PT_LAMP: + for (fi = min;; fi++) + { + int chartype; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM37); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + chartype = UCD_CHARTYPE(c); + if ((chartype == ucp_Lu || + chartype == ucp_Ll || + chartype == ucp_Lt) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + case PT_GC: + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM38); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + case PT_PC: + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM39); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + case PT_SC: + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM40); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + case PT_ALNUM: + for (fi = min;; fi++) + { + int category; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM59); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + category = UCD_CATEGORY(c); + if ((category == ucp_L || category == ucp_N) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + /* Perl space used to exclude VT, but from Perl 5.18 it is included, + which means that Perl space and POSIX space are now identical. PCRE + was changed at release 8.34. */ + + case PT_SPACE: /* Perl space */ + case PT_PXSPACE: /* POSIX space */ + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM61); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + switch(c) + { + HSPACE_CASES: + VSPACE_CASES: + if (prop_fail_result) RRETURN(MATCH_NOMATCH); + break; + + default: + if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + break; + } + } + /* Control never gets here */ + + case PT_WORD: + for (fi = min;; fi++) + { + int category; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM62); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + category = UCD_CATEGORY(c); + if ((category == ucp_L || + category == ucp_N || + c == CHAR_UNDERSCORE) + == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + case PT_CLIST: + for (fi = min;; fi++) + { + const pcre_uint32 *cp; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM67); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + cp = PRIV(ucd_caseless_sets) + prop_value; + for (;;) + { + if (c < *cp) + { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } } + if (c == *cp++) + { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; } + } + } + /* Control never gets here */ + + case PT_UCNC: + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM60); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT || + c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) || + c >= 0xe000) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + /* This should never occur */ + default: + RRETURN(PCRE_ERROR_INTERNAL); + } + } + + /* Match extended Unicode sequences. We will get here only if the + support is in the binary; otherwise a compile-time error occurs. */ + + else if (ctype == OP_EXTUNI) + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM41); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + else + { + int lgb, rgb; + GETCHARINCTEST(c, eptr); + lgb = UCD_GRAPHBREAK(c); + while (eptr < md->end_subject) + { + int len = 1; + if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); } + rgb = UCD_GRAPHBREAK(c); + if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break; + lgb = rgb; + eptr += len; + } + } + CHECK_PARTIAL(); + } + } + else +#endif /* SUPPORT_UCP */ + +#ifdef SUPPORT_UTF + if (utf) + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM42); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (ctype == OP_ANY && IS_NEWLINE(eptr)) + RRETURN(MATCH_NOMATCH); + GETCHARINC(c, eptr); + switch(ctype) + { + case OP_ANY: /* This is the non-NL case */ + if (md->partial != 0 && /* Take care with CRLF partial */ + eptr >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + break; + + case OP_ALLANY: + case OP_ANYBYTE: + break; + + case OP_ANYNL: + switch(c) + { + default: RRETURN(MATCH_NOMATCH); + case CHAR_CR: + if (eptr < md->end_subject && UCHAR21(eptr) == CHAR_LF) eptr++; + break; + + case CHAR_LF: + break; + + case CHAR_VT: + case CHAR_FF: + case CHAR_NEL: +#ifndef EBCDIC + case 0x2028: + case 0x2029: +#endif /* Not EBCDIC */ + if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH); + break; + } + break; + + case OP_NOT_HSPACE: + switch(c) + { + HSPACE_CASES: RRETURN(MATCH_NOMATCH); + default: break; + } + break; + + case OP_HSPACE: + switch(c) + { + HSPACE_CASES: break; + default: RRETURN(MATCH_NOMATCH); + } + break; + + case OP_NOT_VSPACE: + switch(c) + { + VSPACE_CASES: RRETURN(MATCH_NOMATCH); + default: break; + } + break; + + case OP_VSPACE: + switch(c) + { + VSPACE_CASES: break; + default: RRETURN(MATCH_NOMATCH); + } + break; + + case OP_NOT_DIGIT: + if (c < 256 && (md->ctypes[c] & ctype_digit) != 0) + RRETURN(MATCH_NOMATCH); + break; + + case OP_DIGIT: + if (c >= 256 || (md->ctypes[c] & ctype_digit) == 0) + RRETURN(MATCH_NOMATCH); + break; + + case OP_NOT_WHITESPACE: + if (c < 256 && (md->ctypes[c] & ctype_space) != 0) + RRETURN(MATCH_NOMATCH); + break; + + case OP_WHITESPACE: + if (c >= 256 || (md->ctypes[c] & ctype_space) == 0) + RRETURN(MATCH_NOMATCH); + break; + + case OP_NOT_WORDCHAR: + if (c < 256 && (md->ctypes[c] & ctype_word) != 0) + RRETURN(MATCH_NOMATCH); + break; + + case OP_WORDCHAR: + if (c >= 256 || (md->ctypes[c] & ctype_word) == 0) + RRETURN(MATCH_NOMATCH); + break; + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } + } + } + else +#endif + /* Not UTF mode */ + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM43); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (ctype == OP_ANY && IS_NEWLINE(eptr)) + RRETURN(MATCH_NOMATCH); + c = *eptr++; + switch(ctype) + { + case OP_ANY: /* This is the non-NL case */ + if (md->partial != 0 && /* Take care with CRLF partial */ + eptr >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + break; + + case OP_ALLANY: + case OP_ANYBYTE: + break; + + case OP_ANYNL: + switch(c) + { + default: RRETURN(MATCH_NOMATCH); + case CHAR_CR: + if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++; + break; + + case CHAR_LF: + break; + + case CHAR_VT: + case CHAR_FF: + case CHAR_NEL: +#if defined COMPILE_PCRE16 || defined COMPILE_PCRE32 + case 0x2028: + case 0x2029: +#endif + if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH); + break; + } + break; + + case OP_NOT_HSPACE: + switch(c) + { + default: break; + HSPACE_BYTE_CASES: +#if defined COMPILE_PCRE16 || defined COMPILE_PCRE32 + HSPACE_MULTIBYTE_CASES: +#endif + RRETURN(MATCH_NOMATCH); + } + break; + + case OP_HSPACE: + switch(c) + { + default: RRETURN(MATCH_NOMATCH); + HSPACE_BYTE_CASES: +#if defined COMPILE_PCRE16 || defined COMPILE_PCRE32 + HSPACE_MULTIBYTE_CASES: +#endif + break; + } + break; + + case OP_NOT_VSPACE: + switch(c) + { + default: break; + VSPACE_BYTE_CASES: +#if defined COMPILE_PCRE16 || defined COMPILE_PCRE32 + VSPACE_MULTIBYTE_CASES: +#endif + RRETURN(MATCH_NOMATCH); + } + break; + + case OP_VSPACE: + switch(c) + { + default: RRETURN(MATCH_NOMATCH); + VSPACE_BYTE_CASES: +#if defined COMPILE_PCRE16 || defined COMPILE_PCRE32 + VSPACE_MULTIBYTE_CASES: +#endif + break; + } + break; + + case OP_NOT_DIGIT: + if (MAX_255(c) && (md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH); + break; + + case OP_DIGIT: + if (!MAX_255(c) || (md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH); + break; + + case OP_NOT_WHITESPACE: + if (MAX_255(c) && (md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH); + break; + + case OP_WHITESPACE: + if (!MAX_255(c) || (md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH); + break; + + case OP_NOT_WORDCHAR: + if (MAX_255(c) && (md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH); + break; + + case OP_WORDCHAR: + if (!MAX_255(c) || (md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH); + break; + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } + } + } + /* Control never gets here */ + } + + /* If maximizing, it is worth using inline code for speed, doing the type + test once at the start (i.e. keep it out of the loop). Again, keep the + UTF-8 and UCP stuff separate. */ + + else + { + pp = eptr; /* Remember where we started */ + +#ifdef SUPPORT_UCP + if (prop_type >= 0) + { + switch(prop_type) + { + case PT_ANY: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + if (prop_fail_result) break; + eptr+= len; + } + break; + + case PT_LAMP: + for (i = min; i < max; i++) + { + int chartype; + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + chartype = UCD_CHARTYPE(c); + if ((chartype == ucp_Lu || + chartype == ucp_Ll || + chartype == ucp_Lt) == prop_fail_result) + break; + eptr+= len; + } + break; + + case PT_GC: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result) break; + eptr+= len; + } + break; + + case PT_PC: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result) break; + eptr+= len; + } + break; + + case PT_SC: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result) break; + eptr+= len; + } + break; + + case PT_ALNUM: + for (i = min; i < max; i++) + { + int category; + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + category = UCD_CATEGORY(c); + if ((category == ucp_L || category == ucp_N) == prop_fail_result) + break; + eptr+= len; + } + break; + + /* Perl space used to exclude VT, but from Perl 5.18 it is included, + which means that Perl space and POSIX space are now identical. PCRE + was changed at release 8.34. */ + + case PT_SPACE: /* Perl space */ + case PT_PXSPACE: /* POSIX space */ + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + switch(c) + { + HSPACE_CASES: + VSPACE_CASES: + if (prop_fail_result) goto ENDLOOP99; /* Break the loop */ + break; + + default: + if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result) + goto ENDLOOP99; /* Break the loop */ + break; + } + eptr+= len; + } + ENDLOOP99: + break; + + case PT_WORD: + for (i = min; i < max; i++) + { + int category; + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + category = UCD_CATEGORY(c); + if ((category == ucp_L || category == ucp_N || + c == CHAR_UNDERSCORE) == prop_fail_result) + break; + eptr+= len; + } + break; + + case PT_CLIST: + for (i = min; i < max; i++) + { + const pcre_uint32 *cp; + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + cp = PRIV(ucd_caseless_sets) + prop_value; + for (;;) + { + if (c < *cp) + { if (prop_fail_result) break; else goto GOT_MAX; } + if (c == *cp++) + { if (prop_fail_result) goto GOT_MAX; else break; } + } + eptr += len; + } + GOT_MAX: + break; + + case PT_UCNC: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT || + c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) || + c >= 0xe000) == prop_fail_result) + break; + eptr += len; + } + break; + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } + + /* eptr is now past the end of the maximum run */ + + if (possessive) continue; /* No backtracking */ + for(;;) + { + if (eptr <= pp) goto TAIL_RECURSE; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM44); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + eptr--; + if (utf) BACKCHAR(eptr); + } + } + + /* Match extended Unicode grapheme clusters. We will get here only if the + support is in the binary; otherwise a compile-time error occurs. */ + + else if (ctype == OP_EXTUNI) + { + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + else + { + int lgb, rgb; + GETCHARINCTEST(c, eptr); + lgb = UCD_GRAPHBREAK(c); + while (eptr < md->end_subject) + { + int len = 1; + if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); } + rgb = UCD_GRAPHBREAK(c); + if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break; + lgb = rgb; + eptr += len; + } + } + CHECK_PARTIAL(); + } + + /* eptr is now past the end of the maximum run */ + + if (possessive) continue; /* No backtracking */ + + /* We use <= pp rather than == pp to detect the start of the run while + backtracking because the use of \C in UTF mode can cause BACKCHAR to + move back past pp. This is just palliative; the use of \C in UTF mode + is fraught with danger. */ + + for(;;) + { + int lgb, rgb; + PCRE_PUCHAR fptr; + + if (eptr <= pp) goto TAIL_RECURSE; /* At start of char run */ + RMATCH(eptr, ecode, offset_top, md, eptrb, RM45); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + + /* Backtracking over an extended grapheme cluster involves inspecting + the previous two characters (if present) to see if a break is + permitted between them. */ + + eptr--; + if (!utf) c = *eptr; else + { + BACKCHAR(eptr); + GETCHAR(c, eptr); + } + rgb = UCD_GRAPHBREAK(c); + + for (;;) + { + if (eptr <= pp) goto TAIL_RECURSE; /* At start of char run */ + fptr = eptr - 1; + if (!utf) c = *fptr; else + { + BACKCHAR(fptr); + GETCHAR(c, fptr); + } + lgb = UCD_GRAPHBREAK(c); + if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break; + eptr = fptr; + rgb = lgb; + } + } + } + + else +#endif /* SUPPORT_UCP */ + +#ifdef SUPPORT_UTF + if (utf) + { + switch(ctype) + { + case OP_ANY: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (IS_NEWLINE(eptr)) break; + if (md->partial != 0 && /* Take care with CRLF partial */ + eptr + 1 >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + UCHAR21(eptr) == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + eptr++; + ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++); + } + break; + + case OP_ALLANY: + if (max < INT_MAX) + { + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + eptr++; + ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++); + } + } + else + { + eptr = md->end_subject; /* Unlimited UTF-8 repeat */ + SCHECK_PARTIAL(); + } + break; + + /* The byte case is the same as non-UTF8 */ + + case OP_ANYBYTE: + c = max - min; + if (c > (unsigned int)(md->end_subject - eptr)) + { + eptr = md->end_subject; + SCHECK_PARTIAL(); + } + else eptr += c; + break; + + case OP_ANYNL: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + if (c == CHAR_CR) + { + if (++eptr >= md->end_subject) break; + if (UCHAR21(eptr) == CHAR_LF) eptr++; + } + else + { + if (c != CHAR_LF && + (md->bsr_anycrlf || + (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL +#ifndef EBCDIC + && c != 0x2028 && c != 0x2029 +#endif /* Not EBCDIC */ + ))) + break; + eptr += len; + } + } + break; + + case OP_NOT_HSPACE: + case OP_HSPACE: + for (i = min; i < max; i++) + { + BOOL gotspace; + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + switch(c) + { + HSPACE_CASES: gotspace = TRUE; break; + default: gotspace = FALSE; break; + } + if (gotspace == (ctype == OP_NOT_HSPACE)) break; + eptr += len; + } + break; + + case OP_NOT_VSPACE: + case OP_VSPACE: + for (i = min; i < max; i++) + { + BOOL gotspace; + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + switch(c) + { + VSPACE_CASES: gotspace = TRUE; break; + default: gotspace = FALSE; break; + } + if (gotspace == (ctype == OP_NOT_VSPACE)) break; + eptr += len; + } + break; + + case OP_NOT_DIGIT: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + if (c < 256 && (md->ctypes[c] & ctype_digit) != 0) break; + eptr+= len; + } + break; + + case OP_DIGIT: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + if (c >= 256 ||(md->ctypes[c] & ctype_digit) == 0) break; + eptr+= len; + } + break; + + case OP_NOT_WHITESPACE: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + if (c < 256 && (md->ctypes[c] & ctype_space) != 0) break; + eptr+= len; + } + break; + + case OP_WHITESPACE: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + if (c >= 256 ||(md->ctypes[c] & ctype_space) == 0) break; + eptr+= len; + } + break; + + case OP_NOT_WORDCHAR: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + if (c < 256 && (md->ctypes[c] & ctype_word) != 0) break; + eptr+= len; + } + break; + + case OP_WORDCHAR: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + if (c >= 256 || (md->ctypes[c] & ctype_word) == 0) break; + eptr+= len; + } + break; + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } + + if (possessive) continue; /* No backtracking */ + for(;;) + { + if (eptr <= pp) goto TAIL_RECURSE; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM46); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + eptr--; + BACKCHAR(eptr); + if (ctype == OP_ANYNL && eptr > pp && UCHAR21(eptr) == CHAR_NL && + UCHAR21(eptr - 1) == CHAR_CR) eptr--; + } + } + else +#endif /* SUPPORT_UTF */ + /* Not UTF mode */ + { + switch(ctype) + { + case OP_ANY: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (IS_NEWLINE(eptr)) break; + if (md->partial != 0 && /* Take care with CRLF partial */ + eptr + 1 >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + *eptr == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + eptr++; + } + break; + + case OP_ALLANY: + case OP_ANYBYTE: + c = max - min; + if (c > (unsigned int)(md->end_subject - eptr)) + { + eptr = md->end_subject; + SCHECK_PARTIAL(); + } + else eptr += c; + break; + + case OP_ANYNL: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + c = *eptr; + if (c == CHAR_CR) + { + if (++eptr >= md->end_subject) break; + if (*eptr == CHAR_LF) eptr++; + } + else + { + if (c != CHAR_LF && (md->bsr_anycrlf || + (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL +#if defined COMPILE_PCRE16 || defined COMPILE_PCRE32 + && c != 0x2028 && c != 0x2029 +#endif + ))) break; + eptr++; + } + } + break; + + case OP_NOT_HSPACE: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + switch(*eptr) + { + default: eptr++; break; + HSPACE_BYTE_CASES: +#if defined COMPILE_PCRE16 || defined COMPILE_PCRE32 + HSPACE_MULTIBYTE_CASES: +#endif + goto ENDLOOP00; + } + } + ENDLOOP00: + break; + + case OP_HSPACE: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + switch(*eptr) + { + default: goto ENDLOOP01; + HSPACE_BYTE_CASES: +#if defined COMPILE_PCRE16 || defined COMPILE_PCRE32 + HSPACE_MULTIBYTE_CASES: +#endif + eptr++; break; + } + } + ENDLOOP01: + break; + + case OP_NOT_VSPACE: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + switch(*eptr) + { + default: eptr++; break; + VSPACE_BYTE_CASES: +#if defined COMPILE_PCRE16 || defined COMPILE_PCRE32 + VSPACE_MULTIBYTE_CASES: +#endif + goto ENDLOOP02; + } + } + ENDLOOP02: + break; + + case OP_VSPACE: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + switch(*eptr) + { + default: goto ENDLOOP03; + VSPACE_BYTE_CASES: +#if defined COMPILE_PCRE16 || defined COMPILE_PCRE32 + VSPACE_MULTIBYTE_CASES: +#endif + eptr++; break; + } + } + ENDLOOP03: + break; + + case OP_NOT_DIGIT: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0) break; + eptr++; + } + break; + + case OP_DIGIT: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0) break; + eptr++; + } + break; + + case OP_NOT_WHITESPACE: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0) break; + eptr++; + } + break; + + case OP_WHITESPACE: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0) break; + eptr++; + } + break; + + case OP_NOT_WORDCHAR: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0) break; + eptr++; + } + break; + + case OP_WORDCHAR: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0) break; + eptr++; + } + break; + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } + + if (possessive) continue; /* No backtracking */ + for (;;) + { + if (eptr == pp) goto TAIL_RECURSE; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM47); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + eptr--; + if (ctype == OP_ANYNL && eptr > pp && *eptr == CHAR_LF && + eptr[-1] == CHAR_CR) eptr--; + } + } + + /* Control never gets here */ + } + + /* There's been some horrible disaster. Arrival here can only mean there is + something seriously wrong in the code above or the OP_xxx definitions. */ + + default: + DPRINTF(("Unknown opcode %d\n", *ecode)); + RRETURN(PCRE_ERROR_UNKNOWN_OPCODE); + } + + /* Do not stick any code in here without much thought; it is assumed + that "continue" in the code above comes out to here to repeat the main + loop. */ + + } /* End of main loop */ +/* Control never reaches here */ + + +/* When compiling to use the heap rather than the stack for recursive calls to +match(), the RRETURN() macro jumps here. The number that is saved in +frame->Xwhere indicates which label we actually want to return to. */ + +#ifdef NO_RECURSE +#define LBL(val) case val: goto L_RM##val; +HEAP_RETURN: +switch (frame->Xwhere) + { + LBL( 1) LBL( 2) LBL( 3) LBL( 4) LBL( 5) LBL( 6) LBL( 7) LBL( 8) + LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(17) + LBL(19) LBL(24) LBL(25) LBL(26) LBL(27) LBL(29) LBL(31) LBL(33) + LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52) + LBL(53) LBL(54) LBL(55) LBL(56) LBL(57) LBL(58) LBL(63) LBL(64) + LBL(65) LBL(66) +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + LBL(20) LBL(21) +#endif +#ifdef SUPPORT_UTF + LBL(16) LBL(18) + LBL(22) LBL(23) LBL(28) LBL(30) + LBL(32) LBL(34) LBL(42) LBL(46) +#ifdef SUPPORT_UCP + LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45) + LBL(59) LBL(60) LBL(61) LBL(62) LBL(67) +#endif /* SUPPORT_UCP */ +#endif /* SUPPORT_UTF */ + default: + DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere)); + return PCRE_ERROR_INTERNAL; + } +#undef LBL +#endif /* NO_RECURSE */ +} + + +/*************************************************************************** +**************************************************************************** + RECURSION IN THE match() FUNCTION + +Undefine all the macros that were defined above to handle this. */ + +#ifdef NO_RECURSE +#undef eptr +#undef ecode +#undef mstart +#undef offset_top +#undef eptrb +#undef flags + +#undef callpat +#undef charptr +#undef data +#undef next +#undef pp +#undef prev +#undef saved_eptr + +#undef new_recursive + +#undef cur_is_word +#undef condition +#undef prev_is_word + +#undef ctype +#undef length +#undef max +#undef min +#undef number +#undef offset +#undef op +#undef save_capture_last +#undef save_offset1 +#undef save_offset2 +#undef save_offset3 +#undef stacksave + +#undef newptrb + +#endif + +/* These two are defined as macros in both cases */ + +#undef fc +#undef fi + +/*************************************************************************** +***************************************************************************/ + + +#ifdef NO_RECURSE +/************************************************* +* Release allocated heap frames * +*************************************************/ + +/* This function releases all the allocated frames. The base frame is on the +machine stack, and so must not be freed. + +Argument: the address of the base frame +Returns: nothing +*/ + +static void +release_match_heapframes (heapframe *frame_base) +{ +heapframe *nextframe = frame_base->Xnextframe; +while (nextframe != NULL) + { + heapframe *oldframe = nextframe; + nextframe = nextframe->Xnextframe; + (PUBL(stack_free))(oldframe); + } +} +#endif + + +/************************************************* +* Execute a Regular Expression * +*************************************************/ + +/* This function applies a compiled re to a subject string and picks out +portions of the string if it matches. Two elements in the vector are set for +each substring: the offsets to the start and end of the substring. + +Arguments: + argument_re points to the compiled expression + extra_data points to extra data or is NULL + subject points to the subject string + length length of subject string (may contain binary zeros) + start_offset where to start in the subject string + options option bits + offsets points to a vector of ints to be filled in with offsets + offsetcount the number of elements in the vector + +Returns: > 0 => success; value is the number of elements filled in + = 0 => success, but offsets is not big enough + -1 => failed to match + < -1 => some kind of unexpected problem +*/ + +#if defined COMPILE_PCRE8 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre_exec(const pcre *argument_re, const pcre_extra *extra_data, + PCRE_SPTR subject, int length, int start_offset, int options, int *offsets, + int offsetcount) +#elif defined COMPILE_PCRE16 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data, + PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets, + int offsetcount) +#elif defined COMPILE_PCRE32 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre32_exec(const pcre32 *argument_re, const pcre32_extra *extra_data, + PCRE_SPTR32 subject, int length, int start_offset, int options, int *offsets, + int offsetcount) +#endif +{ +int rc, ocount, arg_offset_max; +int newline; +BOOL using_temporary_offsets = FALSE; +BOOL anchored; +BOOL startline; +BOOL firstline; +BOOL utf; +BOOL has_first_char = FALSE; +BOOL has_req_char = FALSE; +pcre_uchar first_char = 0; +pcre_uchar first_char2 = 0; +pcre_uchar req_char = 0; +pcre_uchar req_char2 = 0; +match_data match_block; +match_data *md = &match_block; +const pcre_uint8 *tables; +const pcre_uint8 *start_bits = NULL; +PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset; +PCRE_PUCHAR end_subject; +PCRE_PUCHAR start_partial = NULL; +PCRE_PUCHAR match_partial = NULL; +PCRE_PUCHAR req_char_ptr = start_match - 1; + +const pcre_study_data *study; +const REAL_PCRE *re = (const REAL_PCRE *)argument_re; + +#ifdef NO_RECURSE +heapframe frame_zero; +frame_zero.Xprevframe = NULL; /* Marks the top level */ +frame_zero.Xnextframe = NULL; /* None are allocated yet */ +md->match_frames_base = &frame_zero; +#endif + +/* Check for the special magic call that measures the size of the stack used +per recursive call of match(). Without the funny casting for sizeof, a Windows +compiler gave this error: "unary minus operator applied to unsigned type, +result still unsigned". Hopefully the cast fixes that. */ + +if (re == NULL && extra_data == NULL && subject == NULL && length == -999 && + start_offset == -999) +#ifdef NO_RECURSE + return -((int)sizeof(heapframe)); +#else + return match(NULL, NULL, NULL, 0, NULL, NULL, 0); +#endif + +/* Plausibility checks */ + +if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION; +if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0)) + return PCRE_ERROR_NULL; +if (offsetcount < 0) return PCRE_ERROR_BADCOUNT; +if (length < 0) return PCRE_ERROR_BADLENGTH; +if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET; + +/* Check that the first field in the block is the magic number. If it is not, +return with PCRE_ERROR_BADMAGIC. However, if the magic number is equal to +REVERSED_MAGIC_NUMBER we return with PCRE_ERROR_BADENDIANNESS, which +means that the pattern is likely compiled with different endianness. */ + +if (re->magic_number != MAGIC_NUMBER) + return re->magic_number == REVERSED_MAGIC_NUMBER? + PCRE_ERROR_BADENDIANNESS:PCRE_ERROR_BADMAGIC; +if ((re->flags & PCRE_MODE) == 0) return PCRE_ERROR_BADMODE; + +/* These two settings are used in the code for checking a UTF-8 string that +follows immediately afterwards. Other values in the md block are used only +during "normal" pcre_exec() processing, not when the JIT support is in use, +so they are set up later. */ + +/* PCRE_UTF16 has the same value as PCRE_UTF8. */ +utf = md->utf = (re->options & PCRE_UTF8) != 0; +md->partial = ((options & PCRE_PARTIAL_HARD) != 0)? 2 : + ((options & PCRE_PARTIAL_SOFT) != 0)? 1 : 0; + +/* Check a UTF-8 string if required. Pass back the character offset and error +code for an invalid string if a results vector is available. */ + +#ifdef SUPPORT_UTF +if (utf && (options & PCRE_NO_UTF8_CHECK) == 0) + { + int erroroffset; + int errorcode = PRIV(valid_utf)((PCRE_PUCHAR)subject, length, &erroroffset); + if (errorcode != 0) + { + if (offsetcount >= 2) + { + offsets[0] = erroroffset; + offsets[1] = errorcode; + } +#if defined COMPILE_PCRE8 + return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)? + PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8; +#elif defined COMPILE_PCRE16 + return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)? + PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16; +#elif defined COMPILE_PCRE32 + return PCRE_ERROR_BADUTF32; +#endif + } +#if defined COMPILE_PCRE8 || defined COMPILE_PCRE16 + /* Check that a start_offset points to the start of a UTF character. */ + if (start_offset > 0 && start_offset < length && + NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset])) + return PCRE_ERROR_BADUTF8_OFFSET; +#endif + } +#endif + +/* If the pattern was successfully studied with JIT support, run the JIT +executable instead of the rest of this function. Most options must be set at +compile time for the JIT code to be usable. Fallback to the normal code path if +an unsupported flag is set. */ + +#ifdef SUPPORT_JIT +if (extra_data != NULL + && (extra_data->flags & (PCRE_EXTRA_EXECUTABLE_JIT | + PCRE_EXTRA_TABLES)) == PCRE_EXTRA_EXECUTABLE_JIT + && extra_data->executable_jit != NULL + && (options & ~PUBLIC_JIT_EXEC_OPTIONS) == 0) + { + rc = PRIV(jit_exec)(extra_data, (const pcre_uchar *)subject, length, + start_offset, options, offsets, offsetcount); + + /* PCRE_ERROR_NULL means that the selected normal or partial matching + mode is not compiled. In this case we simply fallback to interpreter. */ + + if (rc != PCRE_ERROR_JIT_BADOPTION) return rc; + } +#endif + +/* Carry on with non-JIT matching. This information is for finding all the +numbers associated with a given name, for condition testing. */ + +md->name_table = (pcre_uchar *)re + re->name_table_offset; +md->name_count = re->name_count; +md->name_entry_size = re->name_entry_size; + +/* Fish out the optional data from the extra_data structure, first setting +the default values. */ + +study = NULL; +md->match_limit = MATCH_LIMIT; +md->match_limit_recursion = MATCH_LIMIT_RECURSION; +md->callout_data = NULL; + +/* The table pointer is always in native byte order. */ + +tables = re->tables; + +/* The two limit values override the defaults, whatever their value. */ + +if (extra_data != NULL) + { + unsigned long int flags = extra_data->flags; + if ((flags & PCRE_EXTRA_STUDY_DATA) != 0) + study = (const pcre_study_data *)extra_data->study_data; + if ((flags & PCRE_EXTRA_MATCH_LIMIT) != 0) + md->match_limit = extra_data->match_limit; + if ((flags & PCRE_EXTRA_MATCH_LIMIT_RECURSION) != 0) + md->match_limit_recursion = extra_data->match_limit_recursion; + if ((flags & PCRE_EXTRA_CALLOUT_DATA) != 0) + md->callout_data = extra_data->callout_data; + if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables; + } + +/* Limits in the regex override only if they are smaller. */ + +if ((re->flags & PCRE_MLSET) != 0 && re->limit_match < md->match_limit) + md->match_limit = re->limit_match; + +if ((re->flags & PCRE_RLSET) != 0 && + re->limit_recursion < md->match_limit_recursion) + md->match_limit_recursion = re->limit_recursion; + +/* If the exec call supplied NULL for tables, use the inbuilt ones. This +is a feature that makes it possible to save compiled regex and re-use them +in other programs later. */ + +if (tables == NULL) tables = PRIV(default_tables); + +/* Set up other data */ + +anchored = ((re->options | options) & PCRE_ANCHORED) != 0; +startline = (re->flags & PCRE_STARTLINE) != 0; +firstline = (re->options & PCRE_FIRSTLINE) != 0; + +/* The code starts after the real_pcre block and the capture name table. */ + +md->start_code = (const pcre_uchar *)re + re->name_table_offset + + re->name_count * re->name_entry_size; + +md->start_subject = (PCRE_PUCHAR)subject; +md->start_offset = start_offset; +md->end_subject = md->start_subject + length; +end_subject = md->end_subject; + +md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0; +md->use_ucp = (re->options & PCRE_UCP) != 0; +md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0; +md->ignore_skip_arg = 0; + +/* Some options are unpacked into BOOL variables in the hope that testing +them will be faster than individual option bits. */ + +md->notbol = (options & PCRE_NOTBOL) != 0; +md->noteol = (options & PCRE_NOTEOL) != 0; +md->notempty = (options & PCRE_NOTEMPTY) != 0; +md->notempty_atstart = (options & PCRE_NOTEMPTY_ATSTART) != 0; + +md->hitend = FALSE; +md->mark = md->nomatch_mark = NULL; /* In case never set */ + +md->recursive = NULL; /* No recursion at top level */ +md->hasthen = (re->flags & PCRE_HASTHEN) != 0; + +md->lcc = tables + lcc_offset; +md->fcc = tables + fcc_offset; +md->ctypes = tables + ctypes_offset; + +/* Handle different \R options. */ + +switch (options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) + { + case 0: + if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0) + md->bsr_anycrlf = (re->options & PCRE_BSR_ANYCRLF) != 0; + else +#ifdef BSR_ANYCRLF + md->bsr_anycrlf = TRUE; +#else + md->bsr_anycrlf = FALSE; +#endif + break; + + case PCRE_BSR_ANYCRLF: + md->bsr_anycrlf = TRUE; + break; + + case PCRE_BSR_UNICODE: + md->bsr_anycrlf = FALSE; + break; + + default: return PCRE_ERROR_BADNEWLINE; + } + +/* Handle different types of newline. The three bits give eight cases. If +nothing is set at run time, whatever was used at compile time applies. */ + +switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options : + (pcre_uint32)options) & PCRE_NEWLINE_BITS) + { + case 0: newline = NEWLINE; break; /* Compile-time default */ + case PCRE_NEWLINE_CR: newline = CHAR_CR; break; + case PCRE_NEWLINE_LF: newline = CHAR_NL; break; + case PCRE_NEWLINE_CR+ + PCRE_NEWLINE_LF: newline = (CHAR_CR << 8) | CHAR_NL; break; + case PCRE_NEWLINE_ANY: newline = -1; break; + case PCRE_NEWLINE_ANYCRLF: newline = -2; break; + default: return PCRE_ERROR_BADNEWLINE; + } + +if (newline == -2) + { + md->nltype = NLTYPE_ANYCRLF; + } +else if (newline < 0) + { + md->nltype = NLTYPE_ANY; + } +else + { + md->nltype = NLTYPE_FIXED; + if (newline > 255) + { + md->nllen = 2; + md->nl[0] = (newline >> 8) & 255; + md->nl[1] = newline & 255; + } + else + { + md->nllen = 1; + md->nl[0] = newline; + } + } + +/* Partial matching was originally supported only for a restricted set of +regexes; from release 8.00 there are no restrictions, but the bits are still +defined (though never set). So there's no harm in leaving this code. */ + +if (md->partial && (re->flags & PCRE_NOPARTIAL) != 0) + return PCRE_ERROR_BADPARTIAL; + +/* If the expression has got more back references than the offsets supplied can +hold, we get a temporary chunk of working store to use during the matching. +Otherwise, we can use the vector supplied, rounding down its size to a multiple +of 3. */ + +ocount = offsetcount - (offsetcount % 3); +arg_offset_max = (2*ocount)/3; + +if (re->top_backref > 0 && re->top_backref >= ocount/3) + { + ocount = re->top_backref * 3 + 3; + md->offset_vector = (int *)(PUBL(malloc))(ocount * sizeof(int)); + if (md->offset_vector == NULL) return PCRE_ERROR_NOMEMORY; + using_temporary_offsets = TRUE; + DPRINTF(("Got memory to hold back references\n")); + } +else md->offset_vector = offsets; +md->offset_end = ocount; +md->offset_max = (2*ocount)/3; +md->capture_last = 0; + +/* Reset the working variable associated with each extraction. These should +never be used unless previously set, but they get saved and restored, and so we +initialize them to avoid reading uninitialized locations. Also, unset the +offsets for the matched string. This is really just for tidiness with callouts, +in case they inspect these fields. */ + +if (md->offset_vector != NULL) + { + register int *iptr = md->offset_vector + ocount; + register int *iend = iptr - re->top_bracket; + if (iend < md->offset_vector + 2) iend = md->offset_vector + 2; + while (--iptr >= iend) *iptr = -1; + if (offsetcount > 0) md->offset_vector[0] = -1; + if (offsetcount > 1) md->offset_vector[1] = -1; + } + +/* Set up the first character to match, if available. The first_char value is +never set for an anchored regular expression, but the anchoring may be forced +at run time, so we have to test for anchoring. The first char may be unset for +an unanchored pattern, of course. If there's no first char and the pattern was +studied, there may be a bitmap of possible first characters. */ + +if (!anchored) + { + if ((re->flags & PCRE_FIRSTSET) != 0) + { + has_first_char = TRUE; + first_char = first_char2 = (pcre_uchar)(re->first_char); + if ((re->flags & PCRE_FCH_CASELESS) != 0) + { + first_char2 = TABLE_GET(first_char, md->fcc, first_char); +#if defined SUPPORT_UCP && !(defined COMPILE_PCRE8) + if (utf && first_char > 127) + first_char2 = UCD_OTHERCASE(first_char); +#endif + } + } + else + if (!startline && study != NULL && + (study->flags & PCRE_STUDY_MAPPED) != 0) + start_bits = study->start_bits; + } + +/* For anchored or unanchored matches, there may be a "last known required +character" set. */ + +if ((re->flags & PCRE_REQCHSET) != 0) + { + has_req_char = TRUE; + req_char = req_char2 = (pcre_uchar)(re->req_char); + if ((re->flags & PCRE_RCH_CASELESS) != 0) + { + req_char2 = TABLE_GET(req_char, md->fcc, req_char); +#if defined SUPPORT_UCP && !(defined COMPILE_PCRE8) + if (utf && req_char > 127) + req_char2 = UCD_OTHERCASE(req_char); +#endif + } + } + + +/* ==========================================================================*/ + +/* Loop for handling unanchored repeated matching attempts; for anchored regexs +the loop runs just once. */ + +for(;;) + { + PCRE_PUCHAR save_end_subject = end_subject; + PCRE_PUCHAR new_start_match; + + /* If firstline is TRUE, the start of the match is constrained to the first + line of a multiline string. That is, the match must be before or at the first + newline. Implement this by temporarily adjusting end_subject so that we stop + scanning at a newline. If the match fails at the newline, later code breaks + this loop. */ + + if (firstline) + { + PCRE_PUCHAR t = start_match; +#ifdef SUPPORT_UTF + if (utf) + { + while (t < md->end_subject && !IS_NEWLINE(t)) + { + t++; + ACROSSCHAR(t < end_subject, *t, t++); + } + } + else +#endif + while (t < md->end_subject && !IS_NEWLINE(t)) t++; + end_subject = t; + } + + /* There are some optimizations that avoid running the match if a known + starting point is not found, or if a known later character is not present. + However, there is an option that disables these, for testing and for ensuring + that all callouts do actually occur. The option can be set in the regex by + (*NO_START_OPT) or passed in match-time options. */ + + if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0) + { + /* Advance to a unique first char if there is one. */ + + if (has_first_char) + { + pcre_uchar smc; + + if (first_char != first_char2) + while (start_match < end_subject && + (smc = UCHAR21TEST(start_match)) != first_char && smc != first_char2) + start_match++; + else + while (start_match < end_subject && UCHAR21TEST(start_match) != first_char) + start_match++; + } + + /* Or to just after a linebreak for a multiline match */ + + else if (startline) + { + if (start_match > md->start_subject + start_offset) + { +#ifdef SUPPORT_UTF + if (utf) + { + while (start_match < end_subject && !WAS_NEWLINE(start_match)) + { + start_match++; + ACROSSCHAR(start_match < end_subject, *start_match, + start_match++); + } + } + else +#endif + while (start_match < end_subject && !WAS_NEWLINE(start_match)) + start_match++; + + /* If we have just passed a CR and the newline option is ANY or ANYCRLF, + and we are now at a LF, advance the match position by one more character. + */ + + if (start_match[-1] == CHAR_CR && + (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) && + start_match < end_subject && + UCHAR21TEST(start_match) == CHAR_NL) + start_match++; + } + } + + /* Or to a non-unique first byte after study */ + + else if (start_bits != NULL) + { + while (start_match < end_subject) + { + register pcre_uint32 c = UCHAR21TEST(start_match); +#ifndef COMPILE_PCRE8 + if (c > 255) c = 255; +#endif + if ((start_bits[c/8] & (1 << (c&7))) != 0) break; + start_match++; + } + } + } /* Starting optimizations */ + + /* Restore fudged end_subject */ + + end_subject = save_end_subject; + + /* The following two optimizations are disabled for partial matching or if + disabling is explicitly requested. */ + + if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0 && !md->partial) + { + /* If the pattern was studied, a minimum subject length may be set. This is + a lower bound; no actual string of that length may actually match the + pattern. Although the value is, strictly, in characters, we treat it as + bytes to avoid spending too much time in this optimization. */ + + if (study != NULL && (study->flags & PCRE_STUDY_MINLEN) != 0 && + (pcre_uint32)(end_subject - start_match) < study->minlength) + { + rc = MATCH_NOMATCH; + break; + } + + /* If req_char is set, we know that that character must appear in the + subject for the match to succeed. If the first character is set, req_char + must be later in the subject; otherwise the test starts at the match point. + This optimization can save a huge amount of backtracking in patterns with + nested unlimited repeats that aren't going to match. Writing separate code + for cased/caseless versions makes it go faster, as does using an + autoincrement and backing off on a match. + + HOWEVER: when the subject string is very, very long, searching to its end + can take a long time, and give bad performance on quite ordinary patterns. + This showed up when somebody was matching something like /^\d+C/ on a + 32-megabyte string... so we don't do this when the string is sufficiently + long. */ + + if (has_req_char && end_subject - start_match < REQ_BYTE_MAX) + { + register PCRE_PUCHAR p = start_match + (has_first_char? 1:0); + + /* We don't need to repeat the search if we haven't yet reached the + place we found it at last time. */ + + if (p > req_char_ptr) + { + if (req_char != req_char2) + { + while (p < end_subject) + { + register pcre_uint32 pp = UCHAR21INCTEST(p); + if (pp == req_char || pp == req_char2) { p--; break; } + } + } + else + { + while (p < end_subject) + { + if (UCHAR21INCTEST(p) == req_char) { p--; break; } + } + } + + /* If we can't find the required character, break the matching loop, + forcing a match failure. */ + + if (p >= end_subject) + { + rc = MATCH_NOMATCH; + break; + } + + /* If we have found the required character, save the point where we + found it, so that we don't search again next time round the loop if + the start hasn't passed this character yet. */ + + req_char_ptr = p; + } + } + } + +#ifdef PCRE_DEBUG /* Sigh. Some compilers never learn. */ + printf(">>>> Match against: "); + pchars(start_match, end_subject - start_match, TRUE, md); + printf("\n"); +#endif + + /* OK, we can now run the match. If "hitend" is set afterwards, remember the + first starting point for which a partial match was found. */ + + md->start_match_ptr = start_match; + md->start_used_ptr = start_match; + md->match_call_count = 0; + md->match_function_type = 0; + md->end_offset_top = 0; + md->skip_arg_count = 0; + rc = match(start_match, md->start_code, start_match, 2, md, NULL, 0); + if (md->hitend && start_partial == NULL) + { + start_partial = md->start_used_ptr; + match_partial = start_match; + } + + switch(rc) + { + /* If MATCH_SKIP_ARG reaches this level it means that a MARK that matched + the SKIP's arg was not found. In this circumstance, Perl ignores the SKIP + entirely. The only way we can do that is to re-do the match at the same + point, with a flag to force SKIP with an argument to be ignored. Just + treating this case as NOMATCH does not work because it does not check other + alternatives in patterns such as A(*SKIP:A)B|AC when the subject is AC. */ + + case MATCH_SKIP_ARG: + new_start_match = start_match; + md->ignore_skip_arg = md->skip_arg_count; + break; + + /* SKIP passes back the next starting point explicitly, but if it is no + greater than the match we have just done, treat it as NOMATCH. */ + + case MATCH_SKIP: + if (md->start_match_ptr > start_match) + { + new_start_match = md->start_match_ptr; + break; + } + /* Fall through */ + + /* NOMATCH and PRUNE advance by one character. THEN at this level acts + exactly like PRUNE. Unset ignore SKIP-with-argument. */ + + case MATCH_NOMATCH: + case MATCH_PRUNE: + case MATCH_THEN: + md->ignore_skip_arg = 0; + new_start_match = start_match + 1; +#ifdef SUPPORT_UTF + if (utf) + ACROSSCHAR(new_start_match < end_subject, *new_start_match, + new_start_match++); +#endif + break; + + /* COMMIT disables the bumpalong, but otherwise behaves as NOMATCH. */ + + case MATCH_COMMIT: + rc = MATCH_NOMATCH; + goto ENDLOOP; + + /* Any other return is either a match, or some kind of error. */ + + default: + goto ENDLOOP; + } + + /* Control reaches here for the various types of "no match at this point" + result. Reset the code to MATCH_NOMATCH for subsequent checking. */ + + rc = MATCH_NOMATCH; + + /* If PCRE_FIRSTLINE is set, the match must happen before or at the first + newline in the subject (though it may continue over the newline). Therefore, + if we have just failed to match, starting at a newline, do not continue. */ + + if (firstline && IS_NEWLINE(start_match)) break; + + /* Advance to new matching position */ + + start_match = new_start_match; + + /* Break the loop if the pattern is anchored or if we have passed the end of + the subject. */ + + if (anchored || start_match > end_subject) break; + + /* If we have just passed a CR and we are now at a LF, and the pattern does + not contain any explicit matches for \r or \n, and the newline option is CRLF + or ANY or ANYCRLF, advance the match position by one more character. In + normal matching start_match will aways be greater than the first position at + this stage, but a failed *SKIP can cause a return at the same point, which is + why the first test exists. */ + + if (start_match > (PCRE_PUCHAR)subject + start_offset && + start_match[-1] == CHAR_CR && + start_match < end_subject && + *start_match == CHAR_NL && + (re->flags & PCRE_HASCRORLF) == 0 && + (md->nltype == NLTYPE_ANY || + md->nltype == NLTYPE_ANYCRLF || + md->nllen == 2)) + start_match++; + + md->mark = NULL; /* Reset for start of next match attempt */ + } /* End of for(;;) "bumpalong" loop */ + +/* ==========================================================================*/ + +/* We reach here when rc is not MATCH_NOMATCH, or if one of the stopping +conditions is true: + +(1) The pattern is anchored or the match was failed by (*COMMIT); + +(2) We are past the end of the subject; + +(3) PCRE_FIRSTLINE is set and we have failed to match at a newline, because + this option requests that a match occur at or before the first newline in + the subject. + +When we have a match and the offset vector is big enough to deal with any +backreferences, captured substring offsets will already be set up. In the case +where we had to get some local store to hold offsets for backreference +processing, copy those that we can. In this case there need not be overflow if +certain parts of the pattern were not used, even though there are more +capturing parentheses than vector slots. */ + +ENDLOOP: + +if (rc == MATCH_MATCH || rc == MATCH_ACCEPT) + { + if (using_temporary_offsets) + { + if (arg_offset_max >= 4) + { + memcpy(offsets + 2, md->offset_vector + 2, + (arg_offset_max - 2) * sizeof(int)); + DPRINTF(("Copied offsets from temporary memory\n")); + } + if (md->end_offset_top > arg_offset_max) md->capture_last |= OVFLBIT; + DPRINTF(("Freeing temporary memory\n")); + (PUBL(free))(md->offset_vector); + } + + /* Set the return code to the number of captured strings, or 0 if there were + too many to fit into the vector. */ + + rc = ((md->capture_last & OVFLBIT) != 0 && + md->end_offset_top >= arg_offset_max)? + 0 : md->end_offset_top/2; + + /* If there is space in the offset vector, set any unused pairs at the end of + the pattern to -1 for backwards compatibility. It is documented that this + happens. In earlier versions, the whole set of potential capturing offsets + was set to -1 each time round the loop, but this is handled differently now. + "Gaps" are set to -1 dynamically instead (this fixes a bug). Thus, it is only + those at the end that need unsetting here. We can't just unset them all at + the start of the whole thing because they may get set in one branch that is + not the final matching branch. */ + + if (md->end_offset_top/2 <= re->top_bracket && offsets != NULL) + { + register int *iptr, *iend; + int resetcount = 2 + re->top_bracket * 2; + if (resetcount > offsetcount) resetcount = offsetcount; + iptr = offsets + md->end_offset_top; + iend = offsets + resetcount; + while (iptr < iend) *iptr++ = -1; + } + + /* If there is space, set up the whole thing as substring 0. The value of + md->start_match_ptr might be modified if \K was encountered on the success + matching path. */ + + if (offsetcount < 2) rc = 0; else + { + offsets[0] = (int)(md->start_match_ptr - md->start_subject); + offsets[1] = (int)(md->end_match_ptr - md->start_subject); + } + + /* Return MARK data if requested */ + + if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0) + *(extra_data->mark) = (pcre_uchar *)md->mark; + DPRINTF((">>>> returning %d\n", rc)); +#ifdef NO_RECURSE + release_match_heapframes(&frame_zero); +#endif + return rc; + } + +/* Control gets here if there has been an error, or if the overall match +attempt has failed at all permitted starting positions. */ + +if (using_temporary_offsets) + { + DPRINTF(("Freeing temporary memory\n")); + (PUBL(free))(md->offset_vector); + } + +/* For anything other than nomatch or partial match, just return the code. */ + +if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL) + { + DPRINTF((">>>> error: returning %d\n", rc)); +#ifdef NO_RECURSE + release_match_heapframes(&frame_zero); +#endif + return rc; + } + +/* Handle partial matches - disable any mark data */ + +if (match_partial != NULL) + { + DPRINTF((">>>> returning PCRE_ERROR_PARTIAL\n")); + md->mark = NULL; + if (offsetcount > 1) + { + offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject); + offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject); + if (offsetcount > 2) + offsets[2] = (int)(match_partial - (PCRE_PUCHAR)subject); + } + rc = PCRE_ERROR_PARTIAL; + } + +/* This is the classic nomatch case */ + +else + { + DPRINTF((">>>> returning PCRE_ERROR_NOMATCH\n")); + rc = PCRE_ERROR_NOMATCH; + } + +/* Return the MARK data if it has been requested. */ + +if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0) + *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark; +#ifdef NO_RECURSE + release_match_heapframes(&frame_zero); +#endif +return rc; +} + +/* End of pcre_exec.c */ diff --git a/include/pcre/pcre_fullinfo.c b/include/pcre/pcre_fullinfo.c new file mode 100644 index 000000000..a6c2ece6c --- /dev/null +++ b/include/pcre/pcre_fullinfo.c @@ -0,0 +1,245 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + +/* PCRE is a library of functions to support regular expressions whose syntax +and semantics are as close as possible to those of the Perl 5 language. + + Written by Philip Hazel + Copyright (c) 1997-2013 University of Cambridge + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + + +/* This module contains the external function pcre_fullinfo(), which returns +information about a compiled pattern. */ + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + + +/************************************************* +* Return info about compiled pattern * +*************************************************/ + +/* This is a newer "info" function which has an extensible interface so +that additional items can be added compatibly. + +Arguments: + argument_re points to compiled code + extra_data points extra data, or NULL + what what information is required + where where to put the information + +Returns: 0 if data returned, negative on error +*/ + +#if defined COMPILE_PCRE8 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre_fullinfo(const pcre *argument_re, const pcre_extra *extra_data, + int what, void *where) +#elif defined COMPILE_PCRE16 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre16_fullinfo(const pcre16 *argument_re, const pcre16_extra *extra_data, + int what, void *where) +#elif defined COMPILE_PCRE32 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre32_fullinfo(const pcre32 *argument_re, const pcre32_extra *extra_data, + int what, void *where) +#endif +{ +const REAL_PCRE *re = (const REAL_PCRE *)argument_re; +const pcre_study_data *study = NULL; + +if (re == NULL || where == NULL) return PCRE_ERROR_NULL; + +if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_STUDY_DATA) != 0) + study = (const pcre_study_data *)extra_data->study_data; + +/* Check that the first field in the block is the magic number. If it is not, +return with PCRE_ERROR_BADMAGIC. However, if the magic number is equal to +REVERSED_MAGIC_NUMBER we return with PCRE_ERROR_BADENDIANNESS, which +means that the pattern is likely compiled with different endianness. */ + +if (re->magic_number != MAGIC_NUMBER) + return re->magic_number == REVERSED_MAGIC_NUMBER? + PCRE_ERROR_BADENDIANNESS:PCRE_ERROR_BADMAGIC; + +/* Check that this pattern was compiled in the correct bit mode */ + +if ((re->flags & PCRE_MODE) == 0) return PCRE_ERROR_BADMODE; + +switch (what) + { + case PCRE_INFO_OPTIONS: + *((unsigned long int *)where) = re->options & PUBLIC_COMPILE_OPTIONS; + break; + + case PCRE_INFO_SIZE: + *((size_t *)where) = re->size; + break; + + case PCRE_INFO_STUDYSIZE: + *((size_t *)where) = (study == NULL)? 0 : study->size; + break; + + case PCRE_INFO_JITSIZE: +#ifdef SUPPORT_JIT + *((size_t *)where) = + (extra_data != NULL && + (extra_data->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0 && + extra_data->executable_jit != NULL)? + PRIV(jit_get_size)(extra_data->executable_jit) : 0; +#else + *((size_t *)where) = 0; +#endif + break; + + case PCRE_INFO_CAPTURECOUNT: + *((int *)where) = re->top_bracket; + break; + + case PCRE_INFO_BACKREFMAX: + *((int *)where) = re->top_backref; + break; + + case PCRE_INFO_FIRSTBYTE: + *((int *)where) = + ((re->flags & PCRE_FIRSTSET) != 0)? (int)re->first_char : + ((re->flags & PCRE_STARTLINE) != 0)? -1 : -2; + break; + + case PCRE_INFO_FIRSTCHARACTER: + *((pcre_uint32 *)where) = + (re->flags & PCRE_FIRSTSET) != 0 ? re->first_char : 0; + break; + + case PCRE_INFO_FIRSTCHARACTERFLAGS: + *((int *)where) = + ((re->flags & PCRE_FIRSTSET) != 0) ? 1 : + ((re->flags & PCRE_STARTLINE) != 0) ? 2 : 0; + break; + + /* Make sure we pass back the pointer to the bit vector in the external + block, not the internal copy (with flipped integer fields). */ + + case PCRE_INFO_FIRSTTABLE: + *((const pcre_uint8 **)where) = + (study != NULL && (study->flags & PCRE_STUDY_MAPPED) != 0)? + ((const pcre_study_data *)extra_data->study_data)->start_bits : NULL; + break; + + case PCRE_INFO_MINLENGTH: + *((int *)where) = + (study != NULL && (study->flags & PCRE_STUDY_MINLEN) != 0)? + (int)(study->minlength) : -1; + break; + + case PCRE_INFO_JIT: + *((int *)where) = extra_data != NULL && + (extra_data->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0 && + extra_data->executable_jit != NULL; + break; + + case PCRE_INFO_LASTLITERAL: + *((int *)where) = + ((re->flags & PCRE_REQCHSET) != 0)? (int)re->req_char : -1; + break; + + case PCRE_INFO_REQUIREDCHAR: + *((pcre_uint32 *)where) = + ((re->flags & PCRE_REQCHSET) != 0) ? re->req_char : 0; + break; + + case PCRE_INFO_REQUIREDCHARFLAGS: + *((int *)where) = + ((re->flags & PCRE_REQCHSET) != 0); + break; + + case PCRE_INFO_NAMEENTRYSIZE: + *((int *)where) = re->name_entry_size; + break; + + case PCRE_INFO_NAMECOUNT: + *((int *)where) = re->name_count; + break; + + case PCRE_INFO_NAMETABLE: + *((const pcre_uchar **)where) = (const pcre_uchar *)re + re->name_table_offset; + break; + + case PCRE_INFO_DEFAULT_TABLES: + *((const pcre_uint8 **)where) = (const pcre_uint8 *)(PRIV(default_tables)); + break; + + /* From release 8.00 this will always return TRUE because NOPARTIAL is + no longer ever set (the restrictions have been removed). */ + + case PCRE_INFO_OKPARTIAL: + *((int *)where) = (re->flags & PCRE_NOPARTIAL) == 0; + break; + + case PCRE_INFO_JCHANGED: + *((int *)where) = (re->flags & PCRE_JCHANGED) != 0; + break; + + case PCRE_INFO_HASCRORLF: + *((int *)where) = (re->flags & PCRE_HASCRORLF) != 0; + break; + + case PCRE_INFO_MAXLOOKBEHIND: + *((int *)where) = re->max_lookbehind; + break; + + case PCRE_INFO_MATCHLIMIT: + if ((re->flags & PCRE_MLSET) == 0) return PCRE_ERROR_UNSET; + *((pcre_uint32 *)where) = re->limit_match; + break; + + case PCRE_INFO_RECURSIONLIMIT: + if ((re->flags & PCRE_RLSET) == 0) return PCRE_ERROR_UNSET; + *((pcre_uint32 *)where) = re->limit_recursion; + break; + + case PCRE_INFO_MATCH_EMPTY: + *((int *)where) = (re->flags & PCRE_MATCH_EMPTY) != 0; + break; + + default: return PCRE_ERROR_BADOPTION; + } + +return 0; +} + +/* End of pcre_fullinfo.c */ diff --git a/include/pcre/pcre_globals.c b/include/pcre/pcre_globals.c new file mode 100644 index 000000000..0f106aa90 --- /dev/null +++ b/include/pcre/pcre_globals.c @@ -0,0 +1,86 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + +/* PCRE is a library of functions to support regular expressions whose syntax +and semantics are as close as possible to those of the Perl 5 language. + + Written by Philip Hazel + Copyright (c) 1997-2014 University of Cambridge + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + + +/* This module contains global variables that are exported by the PCRE library. +PCRE is thread-clean and doesn't use any global variables in the normal sense. +However, it calls memory allocation and freeing functions via the four +indirections below, and it can optionally do callouts, using the fifth +indirection. These values can be changed by the caller, but are shared between +all threads. + +For MS Visual Studio and Symbian OS, there are problems in initializing these +variables to non-local functions. In these cases, therefore, an indirection via +a local function is used. + +Also, when compiling for Virtual Pascal, things are done differently, and +global variables are not used. */ + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + +#if defined _MSC_VER || defined __SYMBIAN32__ +static void* LocalPcreMalloc(size_t aSize) + { + return malloc(aSize); + } +static void LocalPcreFree(void* aPtr) + { + free(aPtr); + } +PCRE_EXP_DATA_DEFN void *(*PUBL(malloc))(size_t) = LocalPcreMalloc; +PCRE_EXP_DATA_DEFN void (*PUBL(free))(void *) = LocalPcreFree; +PCRE_EXP_DATA_DEFN void *(*PUBL(stack_malloc))(size_t) = LocalPcreMalloc; +PCRE_EXP_DATA_DEFN void (*PUBL(stack_free))(void *) = LocalPcreFree; +PCRE_EXP_DATA_DEFN int (*PUBL(callout))(PUBL(callout_block) *) = NULL; +PCRE_EXP_DATA_DEFN int (*PUBL(stack_guard))(void) = NULL; + +#elif !defined VPCOMPAT +PCRE_EXP_DATA_DEFN void *(*PUBL(malloc))(size_t) = malloc; +PCRE_EXP_DATA_DEFN void (*PUBL(free))(void *) = free; +PCRE_EXP_DATA_DEFN void *(*PUBL(stack_malloc))(size_t) = malloc; +PCRE_EXP_DATA_DEFN void (*PUBL(stack_free))(void *) = free; +PCRE_EXP_DATA_DEFN int (*PUBL(callout))(PUBL(callout_block) *) = NULL; +PCRE_EXP_DATA_DEFN int (*PUBL(stack_guard))(void) = NULL; +#endif + +/* End of pcre_globals.c */ diff --git a/include/pcre/pcre_internal.h b/include/pcre/pcre_internal.h new file mode 100644 index 000000000..eca7188a3 --- /dev/null +++ b/include/pcre/pcre_internal.h @@ -0,0 +1,2800 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + + +/* PCRE is a library of functions to support regular expressions whose syntax +and semantics are as close as possible to those of the Perl 5 language. + + Written by Philip Hazel + Copyright (c) 1997-2014 University of Cambridge + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + +/* This header contains definitions that are shared between the different +modules, but which are not relevant to the exported API. This includes some +functions whose names all begin with "_pcre_", "_pcre16_" or "_pcre32_" +depending on the PRIV macro. */ + +#ifndef PCRE_INTERNAL_H +#define PCRE_INTERNAL_H + +#include "config.h" + +/* Define PCRE_DEBUG to get debugging output on stdout. */ + +#if 0 +#define PCRE_DEBUG +#endif + +/* PCRE is compiled as an 8 bit library if it is not requested otherwise. */ + +#if !defined COMPILE_PCRE16 && !defined COMPILE_PCRE32 +#define COMPILE_PCRE8 +#endif + +/* If SUPPORT_UCP is defined, SUPPORT_UTF must also be defined. The +"configure" script ensures this, but not everybody uses "configure". */ + +#if defined SUPPORT_UCP && !(defined SUPPORT_UTF) +#define SUPPORT_UTF 1 +#endif + +/* We define SUPPORT_UTF if SUPPORT_UTF8 is enabled for compatibility +reasons with existing code. */ + +#if defined SUPPORT_UTF8 && !(defined SUPPORT_UTF) +#define SUPPORT_UTF 1 +#endif + +/* Fixme: SUPPORT_UTF8 should be eventually disappear from the code. +Until then we define it if SUPPORT_UTF is defined. */ + +#if defined SUPPORT_UTF && !(defined SUPPORT_UTF8) +#define SUPPORT_UTF8 1 +#endif + +/* We do not support both EBCDIC and UTF-8/16/32 at the same time. The "configure" +script prevents both being selected, but not everybody uses "configure". */ + +#if defined EBCDIC && defined SUPPORT_UTF +#error The use of both EBCDIC and SUPPORT_UTF is not supported. +#endif + +/* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef +inline, and there are *still* stupid compilers about that don't like indented +pre-processor statements, or at least there were when I first wrote this. After +all, it had only been about 10 years then... + +It turns out that the Mac Debugging.h header also defines the macro DPRINTF, so +be absolutely sure we get our version. */ + +#undef DPRINTF +#ifdef PCRE_DEBUG +#define DPRINTF(p) printf p +#else +#define DPRINTF(p) /* Nothing */ +#endif + + +/* Standard C headers plus the external interface definition. The only time +setjmp and stdarg are used is when NO_RECURSE is set. */ + +#include +#include +#include +#include +#include +#include + +/* Valgrind (memcheck) support */ + +#ifdef SUPPORT_VALGRIND +#include +#endif + +/* When compiling a DLL for Windows, the exported symbols have to be declared +using some MS magic. I found some useful information on this web page: +http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the +information there, using __declspec(dllexport) without "extern" we have a +definition; with "extern" we have a declaration. The settings here override the +setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL, +which is all that is needed for applications (they just import the symbols). We +use: + + PCRE_EXP_DECL for declarations + PCRE_EXP_DEFN for definitions of exported functions + PCRE_EXP_DATA_DEFN for definitions of exported variables + +The reason for the two DEFN macros is that in non-Windows environments, one +does not want to have "extern" before variable definitions because it leads to +compiler warnings. So we distinguish between functions and variables. In +Windows, the two should always be the same. + +The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest, +which is an application, but needs to import this file in order to "peek" at +internals, can #include pcre.h first to get an application's-eye view. + +In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon, +special-purpose environments) might want to stick other stuff in front of +exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and +PCRE_EXP_DATA_DEFN only if they are not already set. */ + +#ifndef PCRE_EXP_DECL +# ifdef _WIN32 +# ifndef PCRE_STATIC +# define PCRE_EXP_DECL extern __declspec(dllexport) +# define PCRE_EXP_DEFN __declspec(dllexport) +# define PCRE_EXP_DATA_DEFN __declspec(dllexport) +# else +# define PCRE_EXP_DECL extern +# define PCRE_EXP_DEFN +# define PCRE_EXP_DATA_DEFN +# endif +# else +# ifdef __cplusplus +# define PCRE_EXP_DECL extern "C" +# else +# define PCRE_EXP_DECL extern +# endif +# ifndef PCRE_EXP_DEFN +# define PCRE_EXP_DEFN PCRE_EXP_DECL +# endif +# ifndef PCRE_EXP_DATA_DEFN +# define PCRE_EXP_DATA_DEFN +# endif +# endif +#endif + +/* When compiling with the MSVC compiler, it is sometimes necessary to include +a "calling convention" before exported function names. (This is secondhand +information; I know nothing about MSVC myself). For example, something like + + void __cdecl function(....) + +might be needed. In order so make this easy, all the exported functions have +PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not +set, we ensure here that it has no effect. */ + +#ifndef PCRE_CALL_CONVENTION +#define PCRE_CALL_CONVENTION +#endif + +/* We need to have types that specify unsigned 8, 16 and 32-bit integers. We +cannot determine these outside the compilation (e.g. by running a program as +part of "configure") because PCRE is often cross-compiled for use on other +systems. Instead we make use of the maximum sizes that are available at +preprocessor time in standard C environments. */ + +typedef unsigned char pcre_uint8; + +#if USHRT_MAX == 65535 +typedef unsigned short pcre_uint16; +typedef short pcre_int16; +#define PCRE_UINT16_MAX USHRT_MAX +#define PCRE_INT16_MAX SHRT_MAX +#elif UINT_MAX == 65535 +typedef unsigned int pcre_uint16; +typedef int pcre_int16; +#define PCRE_UINT16_MAX UINT_MAX +#define PCRE_INT16_MAX INT_MAX +#else +#error Cannot determine a type for 16-bit integers +#endif + +#if UINT_MAX == 4294967295U +typedef unsigned int pcre_uint32; +typedef int pcre_int32; +#define PCRE_UINT32_MAX UINT_MAX +#define PCRE_INT32_MAX INT_MAX +#elif ULONG_MAX == 4294967295UL +typedef unsigned long int pcre_uint32; +typedef long int pcre_int32; +#define PCRE_UINT32_MAX ULONG_MAX +#define PCRE_INT32_MAX LONG_MAX +#else +#error Cannot determine a type for 32-bit integers +#endif + +/* When checking for integer overflow in pcre_compile(), we need to handle +large integers. If a 64-bit integer type is available, we can use that. +Otherwise we have to cast to double, which of course requires floating point +arithmetic. Handle this by defining a macro for the appropriate type. If +stdint.h is available, include it; it may define INT64_MAX. Systems that do not +have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set +by "configure". */ + +#if defined HAVE_STDINT_H +#include +#elif defined HAVE_INTTYPES_H +#include +#endif + +#if defined INT64_MAX || defined int64_t +#define INT64_OR_DOUBLE int64_t +#else +#define INT64_OR_DOUBLE double +#endif + +/* All character handling must be done as unsigned characters. Otherwise there +are problems with top-bit-set characters and functions such as isspace(). +However, we leave the interface to the outside world as char * or short *, +because that should make things easier for callers. This character type is +called pcre_uchar. + +The IN_UCHARS macro multiply its argument with the byte size of the current +pcre_uchar type. Useful for memcpy and such operations, whose require the +byte size of their input/output buffers. + +The MAX_255 macro checks whether its pcre_uchar input is less than 256. + +The TABLE_GET macro is designed for accessing elements of tables whose contain +exactly 256 items. When the character is able to contain more than 256 +items, some check is needed before accessing these tables. +*/ + +#if defined COMPILE_PCRE8 + +typedef unsigned char pcre_uchar; +#define IN_UCHARS(x) (x) +#define MAX_255(c) 1 +#define TABLE_GET(c, table, default) ((table)[c]) + +#elif defined COMPILE_PCRE16 + +#if USHRT_MAX != 65535 +/* This is a warning message. Change PCRE_UCHAR16 to a 16 bit data type in +pcre.h(.in) and disable (comment out) this message. */ +#error Warning: PCRE_UCHAR16 is not a 16 bit data type. +#endif + +typedef pcre_uint16 pcre_uchar; +#define UCHAR_SHIFT (1) +#define IN_UCHARS(x) ((x) << UCHAR_SHIFT) +#define MAX_255(c) ((c) <= 255u) +#define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default)) + +#elif defined COMPILE_PCRE32 + +typedef pcre_uint32 pcre_uchar; +#define UCHAR_SHIFT (2) +#define IN_UCHARS(x) ((x) << UCHAR_SHIFT) +#define MAX_255(c) ((c) <= 255u) +#define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default)) + +#else +#error Unsupported compiling mode +#endif /* COMPILE_PCRE[8|16|32] */ + +/* This is an unsigned int value that no character can ever have. UTF-8 +characters only go up to 0x7fffffff (though Unicode doesn't go beyond +0x0010ffff). */ + +#define NOTACHAR 0xffffffff + +/* PCRE is able to support several different kinds of newline (CR, LF, CRLF, +"any" and "anycrlf" at present). The following macros are used to package up +testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various +modules to indicate in which datablock the parameters exist, and what the +start/end of string field names are. */ + +#define NLTYPE_FIXED 0 /* Newline is a fixed length string */ +#define NLTYPE_ANY 1 /* Newline is any Unicode line ending */ +#define NLTYPE_ANYCRLF 2 /* Newline is CR, LF, or CRLF */ + +/* This macro checks for a newline at the given position */ + +#define IS_NEWLINE(p) \ + ((NLBLOCK->nltype != NLTYPE_FIXED)? \ + ((p) < NLBLOCK->PSEND && \ + PRIV(is_newline)((p), NLBLOCK->nltype, NLBLOCK->PSEND, \ + &(NLBLOCK->nllen), utf)) \ + : \ + ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \ + UCHAR21TEST(p) == NLBLOCK->nl[0] && \ + (NLBLOCK->nllen == 1 || UCHAR21TEST(p+1) == NLBLOCK->nl[1]) \ + ) \ + ) + +/* This macro checks for a newline immediately preceding the given position */ + +#define WAS_NEWLINE(p) \ + ((NLBLOCK->nltype != NLTYPE_FIXED)? \ + ((p) > NLBLOCK->PSSTART && \ + PRIV(was_newline)((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \ + &(NLBLOCK->nllen), utf)) \ + : \ + ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \ + UCHAR21TEST(p - NLBLOCK->nllen) == NLBLOCK->nl[0] && \ + (NLBLOCK->nllen == 1 || UCHAR21TEST(p - NLBLOCK->nllen + 1) == NLBLOCK->nl[1]) \ + ) \ + ) + +/* When PCRE is compiled as a C++ library, the subject pointer can be replaced +with a custom type. This makes it possible, for example, to allow pcre_exec() +to process subject strings that are discontinuous by using a smart pointer +class. It must always be possible to inspect all of the subject string in +pcre_exec() because of the way it backtracks. Two macros are required in the +normal case, for sign-unspecified and unsigned char pointers. The former is +used for the external interface and appears in pcre.h, which is why its name +must begin with PCRE_. */ + +#ifdef CUSTOM_SUBJECT_PTR +#define PCRE_PUCHAR CUSTOM_SUBJECT_PTR +#else +#define PCRE_PUCHAR const pcre_uchar * +#endif + +/* Include the public PCRE header and the definitions of UCP character property +values. */ + +#include "pcre.h" +#include "ucp.h" + +#ifdef COMPILE_PCRE32 +/* Assert that the public PCRE_UCHAR32 is a 32-bit type */ +typedef int __assert_pcre_uchar32_size[sizeof(PCRE_UCHAR32) == 4 ? 1 : -1]; +#endif + +/* When compiling for use with the Virtual Pascal compiler, these functions +need to have their names changed. PCRE must be compiled with the -DVPCOMPAT +option on the command line. */ + +#ifdef VPCOMPAT +#define strlen(s) _strlen(s) +#define strncmp(s1,s2,m) _strncmp(s1,s2,m) +#define memcmp(s,c,n) _memcmp(s,c,n) +#define memcpy(d,s,n) _memcpy(d,s,n) +#define memmove(d,s,n) _memmove(d,s,n) +#define memset(s,c,n) _memset(s,c,n) +#else /* VPCOMPAT */ + +/* To cope with SunOS4 and other systems that lack memmove() but have bcopy(), +define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY +is set. Otherwise, include an emulating function for those systems that have +neither (there some non-Unix environments where this is the case). */ + +#ifndef HAVE_MEMMOVE +#undef memmove /* some systems may have a macro */ +#ifdef HAVE_BCOPY +#define memmove(a, b, c) bcopy(b, a, c) +#else /* HAVE_BCOPY */ +static void * +pcre_memmove(void *d, const void *s, size_t n) +{ +size_t i; +unsigned char *dest = (unsigned char *)d; +const unsigned char *src = (const unsigned char *)s; +if (dest > src) + { + dest += n; + src += n; + for (i = 0; i < n; ++i) *(--dest) = *(--src); + return (void *)dest; + } +else + { + for (i = 0; i < n; ++i) *dest++ = *src++; + return (void *)(dest - n); + } +} +#define memmove(a, b, c) pcre_memmove(a, b, c) +#endif /* not HAVE_BCOPY */ +#endif /* not HAVE_MEMMOVE */ +#endif /* not VPCOMPAT */ + + +/* PCRE keeps offsets in its compiled code as 2-byte quantities (always stored +in big-endian order) by default. These are used, for example, to link from the +start of a subpattern to its alternatives and its end. The use of 2 bytes per +offset limits the size of the compiled regex to around 64K, which is big enough +for almost everybody. However, I received a request for an even bigger limit. +For this reason, and also to make the code easier to maintain, the storing and +loading of offsets from the byte string is now handled by the macros that are +defined here. + +The macros are controlled by the value of LINK_SIZE. This defaults to 2 in +the config.h file, but can be overridden by using -D on the command line. This +is automated on Unix systems via the "configure" command. */ + +#if defined COMPILE_PCRE8 + +#if LINK_SIZE == 2 + +#define PUT(a,n,d) \ + (a[n] = (d) >> 8), \ + (a[(n)+1] = (d) & 255) + +#define GET(a,n) \ + (((a)[n] << 8) | (a)[(n)+1]) + +#define MAX_PATTERN_SIZE (1 << 16) + + +#elif LINK_SIZE == 3 + +#define PUT(a,n,d) \ + (a[n] = (d) >> 16), \ + (a[(n)+1] = (d) >> 8), \ + (a[(n)+2] = (d) & 255) + +#define GET(a,n) \ + (((a)[n] << 16) | ((a)[(n)+1] << 8) | (a)[(n)+2]) + +#define MAX_PATTERN_SIZE (1 << 24) + + +#elif LINK_SIZE == 4 + +#define PUT(a,n,d) \ + (a[n] = (d) >> 24), \ + (a[(n)+1] = (d) >> 16), \ + (a[(n)+2] = (d) >> 8), \ + (a[(n)+3] = (d) & 255) + +#define GET(a,n) \ + (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3]) + +/* Keep it positive */ +#define MAX_PATTERN_SIZE (1 << 30) + +#else +#error LINK_SIZE must be either 2, 3, or 4 +#endif + +#elif defined COMPILE_PCRE16 + +#if LINK_SIZE == 2 + +/* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */ +#undef LINK_SIZE +#define LINK_SIZE 1 + +#define PUT(a,n,d) \ + (a[n] = (d)) + +#define GET(a,n) \ + (a[n]) + +#define MAX_PATTERN_SIZE (1 << 16) + +#elif LINK_SIZE == 3 || LINK_SIZE == 4 + +/* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */ +#undef LINK_SIZE +#define LINK_SIZE 2 + +#define PUT(a,n,d) \ + (a[n] = (d) >> 16), \ + (a[(n)+1] = (d) & 65535) + +#define GET(a,n) \ + (((a)[n] << 16) | (a)[(n)+1]) + +/* Keep it positive */ +#define MAX_PATTERN_SIZE (1 << 30) + +#else +#error LINK_SIZE must be either 2, 3, or 4 +#endif + +#elif defined COMPILE_PCRE32 + +/* Only supported LINK_SIZE is 4 */ +/* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */ +#undef LINK_SIZE +#define LINK_SIZE 1 + +#define PUT(a,n,d) \ + (a[n] = (d)) + +#define GET(a,n) \ + (a[n]) + +/* Keep it positive */ +#define MAX_PATTERN_SIZE (1 << 30) + +#else +#error Unsupported compiling mode +#endif /* COMPILE_PCRE[8|16|32] */ + +/* Convenience macro defined in terms of the others */ + +#define PUTINC(a,n,d) PUT(a,n,d), a += LINK_SIZE + + +/* PCRE uses some other 2-byte quantities that do not change when the size of +offsets changes. There are used for repeat counts and for other things such as +capturing parenthesis numbers in back references. */ + +#if defined COMPILE_PCRE8 + +#define IMM2_SIZE 2 + +#define PUT2(a,n,d) \ + a[n] = (d) >> 8; \ + a[(n)+1] = (d) & 255 + +/* For reasons that I do not understand, the expression in this GET2 macro is +treated by gcc as a signed expression, even when a is declared as unsigned. It +seems that any kind of arithmetic results in a signed value. */ + +#define GET2(a,n) \ + (unsigned int)(((a)[n] << 8) | (a)[(n)+1]) + +#elif defined COMPILE_PCRE16 + +#define IMM2_SIZE 1 + +#define PUT2(a,n,d) \ + a[n] = d + +#define GET2(a,n) \ + a[n] + +#elif defined COMPILE_PCRE32 + +#define IMM2_SIZE 1 + +#define PUT2(a,n,d) \ + a[n] = d + +#define GET2(a,n) \ + a[n] + +#else +#error Unsupported compiling mode +#endif /* COMPILE_PCRE[8|16|32] */ + +#define PUT2INC(a,n,d) PUT2(a,n,d), a += IMM2_SIZE + +/* The maximum length of a MARK name is currently one data unit; it may be +changed in future to be a fixed number of bytes or to depend on LINK_SIZE. */ + +#if defined COMPILE_PCRE16 || defined COMPILE_PCRE32 +#define MAX_MARK ((1u << 16) - 1) +#else +#define MAX_MARK ((1u << 8) - 1) +#endif + +/* There is a proposed future special "UTF-21" mode, in which only the lowest +21 bits of a 32-bit character are interpreted as UTF, with the remaining 11 +high-order bits available to the application for other uses. In preparation for +the future implementation of this mode, there are macros that load a data item +and, if in this special mode, mask it to 21 bits. These macros all have names +starting with UCHAR21. In all other modes, including the normal 32-bit +library, the macros all have the same simple definitions. When the new mode is +implemented, it is expected that these definitions will be varied appropriately +using #ifdef when compiling the library that supports the special mode. */ + +#define UCHAR21(eptr) (*(eptr)) +#define UCHAR21TEST(eptr) (*(eptr)) +#define UCHAR21INC(eptr) (*(eptr)++) +#define UCHAR21INCTEST(eptr) (*(eptr)++) + +/* When UTF encoding is being used, a character is no longer just a single +byte in 8-bit mode or a single short in 16-bit mode. The macros for character +handling generate simple sequences when used in the basic mode, and more +complicated ones for UTF characters. GETCHARLENTEST and other macros are not +used when UTF is not supported. To make sure they can never even appear when +UTF support is omitted, we don't even define them. */ + +#ifndef SUPPORT_UTF + +/* #define MAX_VALUE_FOR_SINGLE_CHAR */ +/* #define HAS_EXTRALEN(c) */ +/* #define GET_EXTRALEN(c) */ +/* #define NOT_FIRSTCHAR(c) */ +#define GETCHAR(c, eptr) c = *eptr; +#define GETCHARTEST(c, eptr) c = *eptr; +#define GETCHARINC(c, eptr) c = *eptr++; +#define GETCHARINCTEST(c, eptr) c = *eptr++; +#define GETCHARLEN(c, eptr, len) c = *eptr; +/* #define GETCHARLENTEST(c, eptr, len) */ +/* #define BACKCHAR(eptr) */ +/* #define FORWARDCHAR(eptr) */ +/* #define ACROSSCHAR(condition, eptr, action) */ + +#else /* SUPPORT_UTF */ + +/* Tests whether the code point needs extra characters to decode. */ + +#define HASUTF8EXTRALEN(c) ((c) >= 0xc0) + +/* Base macro to pick up the remaining bytes of a UTF-8 character, not +advancing the pointer. */ + +#define GETUTF8(c, eptr) \ + { \ + if ((c & 0x20) == 0) \ + c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \ + else if ((c & 0x10) == 0) \ + c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \ + else if ((c & 0x08) == 0) \ + c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \ + ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \ + else if ((c & 0x04) == 0) \ + c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \ + ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \ + (eptr[4] & 0x3f); \ + else \ + c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \ + ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \ + ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \ + } + +/* Base macro to pick up the remaining bytes of a UTF-8 character, advancing +the pointer. */ + +#define GETUTF8INC(c, eptr) \ + { \ + if ((c & 0x20) == 0) \ + c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \ + else if ((c & 0x10) == 0) \ + { \ + c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \ + eptr += 2; \ + } \ + else if ((c & 0x08) == 0) \ + { \ + c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \ + ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \ + eptr += 3; \ + } \ + else if ((c & 0x04) == 0) \ + { \ + c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \ + ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \ + (eptr[3] & 0x3f); \ + eptr += 4; \ + } \ + else \ + { \ + c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \ + ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \ + ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \ + eptr += 5; \ + } \ + } + +#if defined COMPILE_PCRE8 + +/* These macros were originally written in the form of loops that used data +from the tables whose names start with PRIV(utf8_table). They were rewritten by +a user so as not to use loops, because in some environments this gives a +significant performance advantage, and it seems never to do any harm. */ + +/* Tells the biggest code point which can be encoded as a single character. */ + +#define MAX_VALUE_FOR_SINGLE_CHAR 127 + +/* Tests whether the code point needs extra characters to decode. */ + +#define HAS_EXTRALEN(c) ((c) >= 0xc0) + +/* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE. +Otherwise it has an undefined behaviour. */ + +#define GET_EXTRALEN(c) (PRIV(utf8_table4)[(c) & 0x3f]) + +/* Returns TRUE, if the given character is not the first character +of a UTF sequence. */ + +#define NOT_FIRSTCHAR(c) (((c) & 0xc0) == 0x80) + +/* Get the next UTF-8 character, not advancing the pointer. This is called when +we know we are in UTF-8 mode. */ + +#define GETCHAR(c, eptr) \ + c = *eptr; \ + if (c >= 0xc0) GETUTF8(c, eptr); + +/* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the +pointer. */ + +#define GETCHARTEST(c, eptr) \ + c = *eptr; \ + if (utf && c >= 0xc0) GETUTF8(c, eptr); + +/* Get the next UTF-8 character, advancing the pointer. This is called when we +know we are in UTF-8 mode. */ + +#define GETCHARINC(c, eptr) \ + c = *eptr++; \ + if (c >= 0xc0) GETUTF8INC(c, eptr); + +/* Get the next character, testing for UTF-8 mode, and advancing the pointer. +This is called when we don't know if we are in UTF-8 mode. */ + +#define GETCHARINCTEST(c, eptr) \ + c = *eptr++; \ + if (utf && c >= 0xc0) GETUTF8INC(c, eptr); + +/* Base macro to pick up the remaining bytes of a UTF-8 character, not +advancing the pointer, incrementing the length. */ + +#define GETUTF8LEN(c, eptr, len) \ + { \ + if ((c & 0x20) == 0) \ + { \ + c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \ + len++; \ + } \ + else if ((c & 0x10) == 0) \ + { \ + c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \ + len += 2; \ + } \ + else if ((c & 0x08) == 0) \ + {\ + c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \ + ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \ + len += 3; \ + } \ + else if ((c & 0x04) == 0) \ + { \ + c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \ + ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \ + (eptr[4] & 0x3f); \ + len += 4; \ + } \ + else \ + {\ + c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \ + ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \ + ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \ + len += 5; \ + } \ + } + +/* Get the next UTF-8 character, not advancing the pointer, incrementing length +if there are extra bytes. This is called when we know we are in UTF-8 mode. */ + +#define GETCHARLEN(c, eptr, len) \ + c = *eptr; \ + if (c >= 0xc0) GETUTF8LEN(c, eptr, len); + +/* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the +pointer, incrementing length if there are extra bytes. This is called when we +do not know if we are in UTF-8 mode. */ + +#define GETCHARLENTEST(c, eptr, len) \ + c = *eptr; \ + if (utf && c >= 0xc0) GETUTF8LEN(c, eptr, len); + +/* If the pointer is not at the start of a character, move it back until +it is. This is called only in UTF-8 mode - we don't put a test within the macro +because almost all calls are already within a block of UTF-8 only code. */ + +#define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr-- + +/* Same as above, just in the other direction. */ +#define FORWARDCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr++ + +/* Same as above, but it allows a fully customizable form. */ +#define ACROSSCHAR(condition, eptr, action) \ + while((condition) && ((eptr) & 0xc0) == 0x80) action + +#elif defined COMPILE_PCRE16 + +/* Tells the biggest code point which can be encoded as a single character. */ + +#define MAX_VALUE_FOR_SINGLE_CHAR 65535 + +/* Tests whether the code point needs extra characters to decode. */ + +#define HAS_EXTRALEN(c) (((c) & 0xfc00) == 0xd800) + +/* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE. +Otherwise it has an undefined behaviour. */ + +#define GET_EXTRALEN(c) 1 + +/* Returns TRUE, if the given character is not the first character +of a UTF sequence. */ + +#define NOT_FIRSTCHAR(c) (((c) & 0xfc00) == 0xdc00) + +/* Base macro to pick up the low surrogate of a UTF-16 character, not +advancing the pointer. */ + +#define GETUTF16(c, eptr) \ + { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; } + +/* Get the next UTF-16 character, not advancing the pointer. This is called when +we know we are in UTF-16 mode. */ + +#define GETCHAR(c, eptr) \ + c = *eptr; \ + if ((c & 0xfc00) == 0xd800) GETUTF16(c, eptr); + +/* Get the next UTF-16 character, testing for UTF-16 mode, and not advancing the +pointer. */ + +#define GETCHARTEST(c, eptr) \ + c = *eptr; \ + if (utf && (c & 0xfc00) == 0xd800) GETUTF16(c, eptr); + +/* Base macro to pick up the low surrogate of a UTF-16 character, advancing +the pointer. */ + +#define GETUTF16INC(c, eptr) \ + { c = (((c & 0x3ff) << 10) | (*eptr++ & 0x3ff)) + 0x10000; } + +/* Get the next UTF-16 character, advancing the pointer. This is called when we +know we are in UTF-16 mode. */ + +#define GETCHARINC(c, eptr) \ + c = *eptr++; \ + if ((c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr); + +/* Get the next character, testing for UTF-16 mode, and advancing the pointer. +This is called when we don't know if we are in UTF-16 mode. */ + +#define GETCHARINCTEST(c, eptr) \ + c = *eptr++; \ + if (utf && (c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr); + +/* Base macro to pick up the low surrogate of a UTF-16 character, not +advancing the pointer, incrementing the length. */ + +#define GETUTF16LEN(c, eptr, len) \ + { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; len++; } + +/* Get the next UTF-16 character, not advancing the pointer, incrementing +length if there is a low surrogate. This is called when we know we are in +UTF-16 mode. */ + +#define GETCHARLEN(c, eptr, len) \ + c = *eptr; \ + if ((c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len); + +/* Get the next UTF-816character, testing for UTF-16 mode, not advancing the +pointer, incrementing length if there is a low surrogate. This is called when +we do not know if we are in UTF-16 mode. */ + +#define GETCHARLENTEST(c, eptr, len) \ + c = *eptr; \ + if (utf && (c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len); + +/* If the pointer is not at the start of a character, move it back until +it is. This is called only in UTF-16 mode - we don't put a test within the +macro because almost all calls are already within a block of UTF-16 only +code. */ + +#define BACKCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr-- + +/* Same as above, just in the other direction. */ +#define FORWARDCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr++ + +/* Same as above, but it allows a fully customizable form. */ +#define ACROSSCHAR(condition, eptr, action) \ + if ((condition) && ((eptr) & 0xfc00) == 0xdc00) action + +#elif defined COMPILE_PCRE32 + +/* These are trivial for the 32-bit library, since all UTF-32 characters fit +into one pcre_uchar unit. */ +#define MAX_VALUE_FOR_SINGLE_CHAR (0x10ffffu) +#define HAS_EXTRALEN(c) (0) +#define GET_EXTRALEN(c) (0) +#define NOT_FIRSTCHAR(c) (0) + +/* Get the next UTF-32 character, not advancing the pointer. This is called when +we know we are in UTF-32 mode. */ + +#define GETCHAR(c, eptr) \ + c = *(eptr); + +/* Get the next UTF-32 character, testing for UTF-32 mode, and not advancing the +pointer. */ + +#define GETCHARTEST(c, eptr) \ + c = *(eptr); + +/* Get the next UTF-32 character, advancing the pointer. This is called when we +know we are in UTF-32 mode. */ + +#define GETCHARINC(c, eptr) \ + c = *((eptr)++); + +/* Get the next character, testing for UTF-32 mode, and advancing the pointer. +This is called when we don't know if we are in UTF-32 mode. */ + +#define GETCHARINCTEST(c, eptr) \ + c = *((eptr)++); + +/* Get the next UTF-32 character, not advancing the pointer, not incrementing +length (since all UTF-32 is of length 1). This is called when we know we are in +UTF-32 mode. */ + +#define GETCHARLEN(c, eptr, len) \ + GETCHAR(c, eptr) + +/* Get the next UTF-32character, testing for UTF-32 mode, not advancing the +pointer, not incrementing the length (since all UTF-32 is of length 1). +This is called when we do not know if we are in UTF-32 mode. */ + +#define GETCHARLENTEST(c, eptr, len) \ + GETCHARTEST(c, eptr) + +/* If the pointer is not at the start of a character, move it back until +it is. This is called only in UTF-32 mode - we don't put a test within the +macro because almost all calls are already within a block of UTF-32 only +code. +These are all no-ops since all UTF-32 characters fit into one pcre_uchar. */ + +#define BACKCHAR(eptr) do { } while (0) + +/* Same as above, just in the other direction. */ +#define FORWARDCHAR(eptr) do { } while (0) + +/* Same as above, but it allows a fully customizable form. */ +#define ACROSSCHAR(condition, eptr, action) do { } while (0) + +#else +#error Unsupported compiling mode +#endif /* COMPILE_PCRE[8|16|32] */ + +#endif /* SUPPORT_UTF */ + +/* Tests for Unicode horizontal and vertical whitespace characters must check a +number of different values. Using a switch statement for this generates the +fastest code (no loop, no memory access), and there are several places in the +interpreter code where this happens. In order to ensure that all the case lists +remain in step, we use macros so that there is only one place where the lists +are defined. + +These values are also required as lists in pcre_compile.c when processing \h, +\H, \v and \V in a character class. The lists are defined in pcre_tables.c, but +macros that define the values are here so that all the definitions are +together. The lists must be in ascending character order, terminated by +NOTACHAR (which is 0xffffffff). + +Any changes should ensure that the various macros are kept in step with each +other. NOTE: The values also appear in pcre_jit_compile.c. */ + +/* ------ ASCII/Unicode environments ------ */ + +#ifndef EBCDIC + +#define HSPACE_LIST \ + CHAR_HT, CHAR_SPACE, CHAR_NBSP, \ + 0x1680, 0x180e, 0x2000, 0x2001, 0x2002, 0x2003, 0x2004, 0x2005, \ + 0x2006, 0x2007, 0x2008, 0x2009, 0x200A, 0x202f, 0x205f, 0x3000, \ + NOTACHAR + +#define HSPACE_MULTIBYTE_CASES \ + case 0x1680: /* OGHAM SPACE MARK */ \ + case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ \ + case 0x2000: /* EN QUAD */ \ + case 0x2001: /* EM QUAD */ \ + case 0x2002: /* EN SPACE */ \ + case 0x2003: /* EM SPACE */ \ + case 0x2004: /* THREE-PER-EM SPACE */ \ + case 0x2005: /* FOUR-PER-EM SPACE */ \ + case 0x2006: /* SIX-PER-EM SPACE */ \ + case 0x2007: /* FIGURE SPACE */ \ + case 0x2008: /* PUNCTUATION SPACE */ \ + case 0x2009: /* THIN SPACE */ \ + case 0x200A: /* HAIR SPACE */ \ + case 0x202f: /* NARROW NO-BREAK SPACE */ \ + case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ \ + case 0x3000 /* IDEOGRAPHIC SPACE */ + +#define HSPACE_BYTE_CASES \ + case CHAR_HT: \ + case CHAR_SPACE: \ + case CHAR_NBSP + +#define HSPACE_CASES \ + HSPACE_BYTE_CASES: \ + HSPACE_MULTIBYTE_CASES + +#define VSPACE_LIST \ + CHAR_LF, CHAR_VT, CHAR_FF, CHAR_CR, CHAR_NEL, 0x2028, 0x2029, NOTACHAR + +#define VSPACE_MULTIBYTE_CASES \ + case 0x2028: /* LINE SEPARATOR */ \ + case 0x2029 /* PARAGRAPH SEPARATOR */ + +#define VSPACE_BYTE_CASES \ + case CHAR_LF: \ + case CHAR_VT: \ + case CHAR_FF: \ + case CHAR_CR: \ + case CHAR_NEL + +#define VSPACE_CASES \ + VSPACE_BYTE_CASES: \ + VSPACE_MULTIBYTE_CASES + +/* ------ EBCDIC environments ------ */ + +#else +#define HSPACE_LIST CHAR_HT, CHAR_SPACE, CHAR_NBSP, NOTACHAR + +#define HSPACE_BYTE_CASES \ + case CHAR_HT: \ + case CHAR_SPACE: \ + case CHAR_NBSP + +#define HSPACE_CASES HSPACE_BYTE_CASES + +#ifdef EBCDIC_NL25 +#define VSPACE_LIST \ + CHAR_VT, CHAR_FF, CHAR_CR, CHAR_NEL, CHAR_LF, NOTACHAR +#else +#define VSPACE_LIST \ + CHAR_VT, CHAR_FF, CHAR_CR, CHAR_LF, CHAR_NEL, NOTACHAR +#endif + +#define VSPACE_BYTE_CASES \ + case CHAR_LF: \ + case CHAR_VT: \ + case CHAR_FF: \ + case CHAR_CR: \ + case CHAR_NEL + +#define VSPACE_CASES VSPACE_BYTE_CASES +#endif /* EBCDIC */ + +/* ------ End of whitespace macros ------ */ + + + +/* Private flags containing information about the compiled regex. They used to +live at the top end of the options word, but that got almost full, so they were +moved to a 16-bit flags word - which got almost full, so now they are in a +32-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as the +restrictions on partial matching have been lifted. It remains for backwards +compatibility. */ + +#define PCRE_MODE8 0x00000001 /* compiled in 8 bit mode */ +#define PCRE_MODE16 0x00000002 /* compiled in 16 bit mode */ +#define PCRE_MODE32 0x00000004 /* compiled in 32 bit mode */ +#define PCRE_FIRSTSET 0x00000010 /* first_char is set */ +#define PCRE_FCH_CASELESS 0x00000020 /* caseless first char */ +#define PCRE_REQCHSET 0x00000040 /* req_byte is set */ +#define PCRE_RCH_CASELESS 0x00000080 /* caseless requested char */ +#define PCRE_STARTLINE 0x00000100 /* start after \n for multiline */ +#define PCRE_NOPARTIAL 0x00000200 /* can't use partial with this regex */ +#define PCRE_JCHANGED 0x00000400 /* j option used in regex */ +#define PCRE_HASCRORLF 0x00000800 /* explicit \r or \n in pattern */ +#define PCRE_HASTHEN 0x00001000 /* pattern contains (*THEN) */ +#define PCRE_MLSET 0x00002000 /* match limit set by regex */ +#define PCRE_RLSET 0x00004000 /* recursion limit set by regex */ +#define PCRE_MATCH_EMPTY 0x00008000 /* pattern can match empty string */ + +#if defined COMPILE_PCRE8 +#define PCRE_MODE PCRE_MODE8 +#elif defined COMPILE_PCRE16 +#define PCRE_MODE PCRE_MODE16 +#elif defined COMPILE_PCRE32 +#define PCRE_MODE PCRE_MODE32 +#endif +#define PCRE_MODE_MASK (PCRE_MODE8 | PCRE_MODE16 | PCRE_MODE32) + +/* Flags for the "extra" block produced by pcre_study(). */ + +#define PCRE_STUDY_MAPPED 0x0001 /* a map of starting chars exists */ +#define PCRE_STUDY_MINLEN 0x0002 /* a minimum length field exists */ + +/* Masks for identifying the public options that are permitted at compile +time, run time, or study time, respectively. */ + +#define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \ + PCRE_NEWLINE_ANYCRLF) + +#define PUBLIC_COMPILE_OPTIONS \ + (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \ + PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \ + PCRE_NO_AUTO_CAPTURE|PCRE_NO_AUTO_POSSESS| \ + PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \ + PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \ + PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE|PCRE_NEVER_UTF) + +#define PUBLIC_EXEC_OPTIONS \ + (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \ + PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \ + PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE) + +#define PUBLIC_DFA_EXEC_OPTIONS \ + (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \ + PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \ + PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \ + PCRE_NO_START_OPTIMIZE) + +#define PUBLIC_STUDY_OPTIONS \ + (PCRE_STUDY_JIT_COMPILE|PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE| \ + PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE|PCRE_STUDY_EXTRA_NEEDED) + +#define PUBLIC_JIT_EXEC_OPTIONS \ + (PCRE_NO_UTF8_CHECK|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|\ + PCRE_NOTEMPTY_ATSTART|PCRE_PARTIAL_SOFT|PCRE_PARTIAL_HARD) + +/* Magic number to provide a small check against being handed junk. */ + +#define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */ + +/* This variable is used to detect a loaded regular expression +in different endianness. */ + +#define REVERSED_MAGIC_NUMBER 0x45524350UL /* 'ERCP' */ + +/* The maximum remaining length of subject we are prepared to search for a +req_byte match. */ + +#define REQ_BYTE_MAX 1000 + +/* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in +environments where these macros are defined elsewhere. Unfortunately, there +is no way to do the same for the typedef. */ + +typedef int BOOL; + +#ifndef FALSE +#define FALSE 0 +#define TRUE 1 +#endif + +/* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal +character constants like '*' because the compiler would emit their EBCDIC code, +which is different from their ASCII/UTF-8 code. Instead we define macros for +the characters so that they always use the ASCII/UTF-8 code when UTF-8 support +is enabled. When UTF-8 support is not enabled, the definitions use character +literals. Both character and string versions of each character are needed, and +there are some longer strings as well. + +This means that, on EBCDIC platforms, the PCRE library can handle either +EBCDIC, or UTF-8, but not both. To support both in the same compiled library +would need different lookups depending on whether PCRE_UTF8 was set or not. +This would make it impossible to use characters in switch/case statements, +which would reduce performance. For a theoretical use (which nobody has asked +for) in a minority area (EBCDIC platforms), this is not sensible. Any +application that did need both could compile two versions of the library, using +macros to give the functions distinct names. */ + +#ifndef SUPPORT_UTF + +/* UTF-8 support is not enabled; use the platform-dependent character literals +so that PCRE works in both ASCII and EBCDIC environments, but only in non-UTF +mode. Newline characters are problematic in EBCDIC. Though it has CR and LF +characters, a common practice has been to use its NL (0x15) character as the +line terminator in C-like processing environments. However, sometimes the LF +(0x25) character is used instead, according to this Unicode document: + +http://unicode.org/standard/reports/tr13/tr13-5.html + +PCRE defaults EBCDIC NL to 0x15, but has a build-time option to select 0x25 +instead. Whichever is *not* chosen is defined as NEL. + +In both ASCII and EBCDIC environments, CHAR_NL and CHAR_LF are synonyms for the +same code point. */ + +#ifdef EBCDIC + +#ifndef EBCDIC_NL25 +#define CHAR_NL '\x15' +#define CHAR_NEL '\x25' +#define STR_NL "\x15" +#define STR_NEL "\x25" +#else +#define CHAR_NL '\x25' +#define CHAR_NEL '\x15' +#define STR_NL "\x25" +#define STR_NEL "\x15" +#endif + +#define CHAR_LF CHAR_NL +#define STR_LF STR_NL + +#define CHAR_ESC '\047' +#define CHAR_DEL '\007' +#define CHAR_NBSP '\x41' +#define STR_ESC "\047" +#define STR_DEL "\007" + +#else /* Not EBCDIC */ + +/* In ASCII/Unicode, linefeed is '\n' and we equate this to NL for +compatibility. NEL is the Unicode newline character; make sure it is +a positive value. */ + +#define CHAR_LF '\n' +#define CHAR_NL CHAR_LF +#define CHAR_NEL ((unsigned char)'\x85') +#define CHAR_ESC '\033' +#define CHAR_DEL '\177' +#define CHAR_NBSP ((unsigned char)'\xa0') + +#define STR_LF "\n" +#define STR_NL STR_LF +#define STR_NEL "\x85" +#define STR_ESC "\033" +#define STR_DEL "\177" + +#endif /* EBCDIC */ + +/* The remaining definitions work in both environments. */ + +#define CHAR_NULL '\0' +#define CHAR_HT '\t' +#define CHAR_VT '\v' +#define CHAR_FF '\f' +#define CHAR_CR '\r' +#define CHAR_BS '\b' +#define CHAR_BEL '\a' + +#define CHAR_SPACE ' ' +#define CHAR_EXCLAMATION_MARK '!' +#define CHAR_QUOTATION_MARK '"' +#define CHAR_NUMBER_SIGN '#' +#define CHAR_DOLLAR_SIGN '$' +#define CHAR_PERCENT_SIGN '%' +#define CHAR_AMPERSAND '&' +#define CHAR_APOSTROPHE '\'' +#define CHAR_LEFT_PARENTHESIS '(' +#define CHAR_RIGHT_PARENTHESIS ')' +#define CHAR_ASTERISK '*' +#define CHAR_PLUS '+' +#define CHAR_COMMA ',' +#define CHAR_MINUS '-' +#define CHAR_DOT '.' +#define CHAR_SLASH '/' +#define CHAR_0 '0' +#define CHAR_1 '1' +#define CHAR_2 '2' +#define CHAR_3 '3' +#define CHAR_4 '4' +#define CHAR_5 '5' +#define CHAR_6 '6' +#define CHAR_7 '7' +#define CHAR_8 '8' +#define CHAR_9 '9' +#define CHAR_COLON ':' +#define CHAR_SEMICOLON ';' +#define CHAR_LESS_THAN_SIGN '<' +#define CHAR_EQUALS_SIGN '=' +#define CHAR_GREATER_THAN_SIGN '>' +#define CHAR_QUESTION_MARK '?' +#define CHAR_COMMERCIAL_AT '@' +#define CHAR_A 'A' +#define CHAR_B 'B' +#define CHAR_C 'C' +#define CHAR_D 'D' +#define CHAR_E 'E' +#define CHAR_F 'F' +#define CHAR_G 'G' +#define CHAR_H 'H' +#define CHAR_I 'I' +#define CHAR_J 'J' +#define CHAR_K 'K' +#define CHAR_L 'L' +#define CHAR_M 'M' +#define CHAR_N 'N' +#define CHAR_O 'O' +#define CHAR_P 'P' +#define CHAR_Q 'Q' +#define CHAR_R 'R' +#define CHAR_S 'S' +#define CHAR_T 'T' +#define CHAR_U 'U' +#define CHAR_V 'V' +#define CHAR_W 'W' +#define CHAR_X 'X' +#define CHAR_Y 'Y' +#define CHAR_Z 'Z' +#define CHAR_LEFT_SQUARE_BRACKET '[' +#define CHAR_BACKSLASH '\\' +#define CHAR_RIGHT_SQUARE_BRACKET ']' +#define CHAR_CIRCUMFLEX_ACCENT '^' +#define CHAR_UNDERSCORE '_' +#define CHAR_GRAVE_ACCENT '`' +#define CHAR_a 'a' +#define CHAR_b 'b' +#define CHAR_c 'c' +#define CHAR_d 'd' +#define CHAR_e 'e' +#define CHAR_f 'f' +#define CHAR_g 'g' +#define CHAR_h 'h' +#define CHAR_i 'i' +#define CHAR_j 'j' +#define CHAR_k 'k' +#define CHAR_l 'l' +#define CHAR_m 'm' +#define CHAR_n 'n' +#define CHAR_o 'o' +#define CHAR_p 'p' +#define CHAR_q 'q' +#define CHAR_r 'r' +#define CHAR_s 's' +#define CHAR_t 't' +#define CHAR_u 'u' +#define CHAR_v 'v' +#define CHAR_w 'w' +#define CHAR_x 'x' +#define CHAR_y 'y' +#define CHAR_z 'z' +#define CHAR_LEFT_CURLY_BRACKET '{' +#define CHAR_VERTICAL_LINE '|' +#define CHAR_RIGHT_CURLY_BRACKET '}' +#define CHAR_TILDE '~' + +#define STR_HT "\t" +#define STR_VT "\v" +#define STR_FF "\f" +#define STR_CR "\r" +#define STR_BS "\b" +#define STR_BEL "\a" + +#define STR_SPACE " " +#define STR_EXCLAMATION_MARK "!" +#define STR_QUOTATION_MARK "\"" +#define STR_NUMBER_SIGN "#" +#define STR_DOLLAR_SIGN "$" +#define STR_PERCENT_SIGN "%" +#define STR_AMPERSAND "&" +#define STR_APOSTROPHE "'" +#define STR_LEFT_PARENTHESIS "(" +#define STR_RIGHT_PARENTHESIS ")" +#define STR_ASTERISK "*" +#define STR_PLUS "+" +#define STR_COMMA "," +#define STR_MINUS "-" +#define STR_DOT "." +#define STR_SLASH "/" +#define STR_0 "0" +#define STR_1 "1" +#define STR_2 "2" +#define STR_3 "3" +#define STR_4 "4" +#define STR_5 "5" +#define STR_6 "6" +#define STR_7 "7" +#define STR_8 "8" +#define STR_9 "9" +#define STR_COLON ":" +#define STR_SEMICOLON ";" +#define STR_LESS_THAN_SIGN "<" +#define STR_EQUALS_SIGN "=" +#define STR_GREATER_THAN_SIGN ">" +#define STR_QUESTION_MARK "?" +#define STR_COMMERCIAL_AT "@" +#define STR_A "A" +#define STR_B "B" +#define STR_C "C" +#define STR_D "D" +#define STR_E "E" +#define STR_F "F" +#define STR_G "G" +#define STR_H "H" +#define STR_I "I" +#define STR_J "J" +#define STR_K "K" +#define STR_L "L" +#define STR_M "M" +#define STR_N "N" +#define STR_O "O" +#define STR_P "P" +#define STR_Q "Q" +#define STR_R "R" +#define STR_S "S" +#define STR_T "T" +#define STR_U "U" +#define STR_V "V" +#define STR_W "W" +#define STR_X "X" +#define STR_Y "Y" +#define STR_Z "Z" +#define STR_LEFT_SQUARE_BRACKET "[" +#define STR_BACKSLASH "\\" +#define STR_RIGHT_SQUARE_BRACKET "]" +#define STR_CIRCUMFLEX_ACCENT "^" +#define STR_UNDERSCORE "_" +#define STR_GRAVE_ACCENT "`" +#define STR_a "a" +#define STR_b "b" +#define STR_c "c" +#define STR_d "d" +#define STR_e "e" +#define STR_f "f" +#define STR_g "g" +#define STR_h "h" +#define STR_i "i" +#define STR_j "j" +#define STR_k "k" +#define STR_l "l" +#define STR_m "m" +#define STR_n "n" +#define STR_o "o" +#define STR_p "p" +#define STR_q "q" +#define STR_r "r" +#define STR_s "s" +#define STR_t "t" +#define STR_u "u" +#define STR_v "v" +#define STR_w "w" +#define STR_x "x" +#define STR_y "y" +#define STR_z "z" +#define STR_LEFT_CURLY_BRACKET "{" +#define STR_VERTICAL_LINE "|" +#define STR_RIGHT_CURLY_BRACKET "}" +#define STR_TILDE "~" + +#define STRING_ACCEPT0 "ACCEPT\0" +#define STRING_COMMIT0 "COMMIT\0" +#define STRING_F0 "F\0" +#define STRING_FAIL0 "FAIL\0" +#define STRING_MARK0 "MARK\0" +#define STRING_PRUNE0 "PRUNE\0" +#define STRING_SKIP0 "SKIP\0" +#define STRING_THEN "THEN" + +#define STRING_alpha0 "alpha\0" +#define STRING_lower0 "lower\0" +#define STRING_upper0 "upper\0" +#define STRING_alnum0 "alnum\0" +#define STRING_ascii0 "ascii\0" +#define STRING_blank0 "blank\0" +#define STRING_cntrl0 "cntrl\0" +#define STRING_digit0 "digit\0" +#define STRING_graph0 "graph\0" +#define STRING_print0 "print\0" +#define STRING_punct0 "punct\0" +#define STRING_space0 "space\0" +#define STRING_word0 "word\0" +#define STRING_xdigit "xdigit" + +#define STRING_DEFINE "DEFINE" +#define STRING_WEIRD_STARTWORD "[:<:]]" +#define STRING_WEIRD_ENDWORD "[:>:]]" + +#define STRING_CR_RIGHTPAR "CR)" +#define STRING_LF_RIGHTPAR "LF)" +#define STRING_CRLF_RIGHTPAR "CRLF)" +#define STRING_ANY_RIGHTPAR "ANY)" +#define STRING_ANYCRLF_RIGHTPAR "ANYCRLF)" +#define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)" +#define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)" +#define STRING_UTF8_RIGHTPAR "UTF8)" +#define STRING_UTF16_RIGHTPAR "UTF16)" +#define STRING_UTF32_RIGHTPAR "UTF32)" +#define STRING_UTF_RIGHTPAR "UTF)" +#define STRING_UCP_RIGHTPAR "UCP)" +#define STRING_NO_AUTO_POSSESS_RIGHTPAR "NO_AUTO_POSSESS)" +#define STRING_NO_START_OPT_RIGHTPAR "NO_START_OPT)" +#define STRING_LIMIT_MATCH_EQ "LIMIT_MATCH=" +#define STRING_LIMIT_RECURSION_EQ "LIMIT_RECURSION=" + +#else /* SUPPORT_UTF */ + +/* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This +works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode +only. */ + +#define CHAR_HT '\011' +#define CHAR_VT '\013' +#define CHAR_FF '\014' +#define CHAR_CR '\015' +#define CHAR_LF '\012' +#define CHAR_NL CHAR_LF +#define CHAR_NEL ((unsigned char)'\x85') +#define CHAR_BS '\010' +#define CHAR_BEL '\007' +#define CHAR_ESC '\033' +#define CHAR_DEL '\177' + +#define CHAR_NULL '\0' +#define CHAR_SPACE '\040' +#define CHAR_EXCLAMATION_MARK '\041' +#define CHAR_QUOTATION_MARK '\042' +#define CHAR_NUMBER_SIGN '\043' +#define CHAR_DOLLAR_SIGN '\044' +#define CHAR_PERCENT_SIGN '\045' +#define CHAR_AMPERSAND '\046' +#define CHAR_APOSTROPHE '\047' +#define CHAR_LEFT_PARENTHESIS '\050' +#define CHAR_RIGHT_PARENTHESIS '\051' +#define CHAR_ASTERISK '\052' +#define CHAR_PLUS '\053' +#define CHAR_COMMA '\054' +#define CHAR_MINUS '\055' +#define CHAR_DOT '\056' +#define CHAR_SLASH '\057' +#define CHAR_0 '\060' +#define CHAR_1 '\061' +#define CHAR_2 '\062' +#define CHAR_3 '\063' +#define CHAR_4 '\064' +#define CHAR_5 '\065' +#define CHAR_6 '\066' +#define CHAR_7 '\067' +#define CHAR_8 '\070' +#define CHAR_9 '\071' +#define CHAR_COLON '\072' +#define CHAR_SEMICOLON '\073' +#define CHAR_LESS_THAN_SIGN '\074' +#define CHAR_EQUALS_SIGN '\075' +#define CHAR_GREATER_THAN_SIGN '\076' +#define CHAR_QUESTION_MARK '\077' +#define CHAR_COMMERCIAL_AT '\100' +#define CHAR_A '\101' +#define CHAR_B '\102' +#define CHAR_C '\103' +#define CHAR_D '\104' +#define CHAR_E '\105' +#define CHAR_F '\106' +#define CHAR_G '\107' +#define CHAR_H '\110' +#define CHAR_I '\111' +#define CHAR_J '\112' +#define CHAR_K '\113' +#define CHAR_L '\114' +#define CHAR_M '\115' +#define CHAR_N '\116' +#define CHAR_O '\117' +#define CHAR_P '\120' +#define CHAR_Q '\121' +#define CHAR_R '\122' +#define CHAR_S '\123' +#define CHAR_T '\124' +#define CHAR_U '\125' +#define CHAR_V '\126' +#define CHAR_W '\127' +#define CHAR_X '\130' +#define CHAR_Y '\131' +#define CHAR_Z '\132' +#define CHAR_LEFT_SQUARE_BRACKET '\133' +#define CHAR_BACKSLASH '\134' +#define CHAR_RIGHT_SQUARE_BRACKET '\135' +#define CHAR_CIRCUMFLEX_ACCENT '\136' +#define CHAR_UNDERSCORE '\137' +#define CHAR_GRAVE_ACCENT '\140' +#define CHAR_a '\141' +#define CHAR_b '\142' +#define CHAR_c '\143' +#define CHAR_d '\144' +#define CHAR_e '\145' +#define CHAR_f '\146' +#define CHAR_g '\147' +#define CHAR_h '\150' +#define CHAR_i '\151' +#define CHAR_j '\152' +#define CHAR_k '\153' +#define CHAR_l '\154' +#define CHAR_m '\155' +#define CHAR_n '\156' +#define CHAR_o '\157' +#define CHAR_p '\160' +#define CHAR_q '\161' +#define CHAR_r '\162' +#define CHAR_s '\163' +#define CHAR_t '\164' +#define CHAR_u '\165' +#define CHAR_v '\166' +#define CHAR_w '\167' +#define CHAR_x '\170' +#define CHAR_y '\171' +#define CHAR_z '\172' +#define CHAR_LEFT_CURLY_BRACKET '\173' +#define CHAR_VERTICAL_LINE '\174' +#define CHAR_RIGHT_CURLY_BRACKET '\175' +#define CHAR_TILDE '\176' +#define CHAR_NBSP ((unsigned char)'\xa0') + +#define STR_HT "\011" +#define STR_VT "\013" +#define STR_FF "\014" +#define STR_CR "\015" +#define STR_NL "\012" +#define STR_BS "\010" +#define STR_BEL "\007" +#define STR_ESC "\033" +#define STR_DEL "\177" + +#define STR_SPACE "\040" +#define STR_EXCLAMATION_MARK "\041" +#define STR_QUOTATION_MARK "\042" +#define STR_NUMBER_SIGN "\043" +#define STR_DOLLAR_SIGN "\044" +#define STR_PERCENT_SIGN "\045" +#define STR_AMPERSAND "\046" +#define STR_APOSTROPHE "\047" +#define STR_LEFT_PARENTHESIS "\050" +#define STR_RIGHT_PARENTHESIS "\051" +#define STR_ASTERISK "\052" +#define STR_PLUS "\053" +#define STR_COMMA "\054" +#define STR_MINUS "\055" +#define STR_DOT "\056" +#define STR_SLASH "\057" +#define STR_0 "\060" +#define STR_1 "\061" +#define STR_2 "\062" +#define STR_3 "\063" +#define STR_4 "\064" +#define STR_5 "\065" +#define STR_6 "\066" +#define STR_7 "\067" +#define STR_8 "\070" +#define STR_9 "\071" +#define STR_COLON "\072" +#define STR_SEMICOLON "\073" +#define STR_LESS_THAN_SIGN "\074" +#define STR_EQUALS_SIGN "\075" +#define STR_GREATER_THAN_SIGN "\076" +#define STR_QUESTION_MARK "\077" +#define STR_COMMERCIAL_AT "\100" +#define STR_A "\101" +#define STR_B "\102" +#define STR_C "\103" +#define STR_D "\104" +#define STR_E "\105" +#define STR_F "\106" +#define STR_G "\107" +#define STR_H "\110" +#define STR_I "\111" +#define STR_J "\112" +#define STR_K "\113" +#define STR_L "\114" +#define STR_M "\115" +#define STR_N "\116" +#define STR_O "\117" +#define STR_P "\120" +#define STR_Q "\121" +#define STR_R "\122" +#define STR_S "\123" +#define STR_T "\124" +#define STR_U "\125" +#define STR_V "\126" +#define STR_W "\127" +#define STR_X "\130" +#define STR_Y "\131" +#define STR_Z "\132" +#define STR_LEFT_SQUARE_BRACKET "\133" +#define STR_BACKSLASH "\134" +#define STR_RIGHT_SQUARE_BRACKET "\135" +#define STR_CIRCUMFLEX_ACCENT "\136" +#define STR_UNDERSCORE "\137" +#define STR_GRAVE_ACCENT "\140" +#define STR_a "\141" +#define STR_b "\142" +#define STR_c "\143" +#define STR_d "\144" +#define STR_e "\145" +#define STR_f "\146" +#define STR_g "\147" +#define STR_h "\150" +#define STR_i "\151" +#define STR_j "\152" +#define STR_k "\153" +#define STR_l "\154" +#define STR_m "\155" +#define STR_n "\156" +#define STR_o "\157" +#define STR_p "\160" +#define STR_q "\161" +#define STR_r "\162" +#define STR_s "\163" +#define STR_t "\164" +#define STR_u "\165" +#define STR_v "\166" +#define STR_w "\167" +#define STR_x "\170" +#define STR_y "\171" +#define STR_z "\172" +#define STR_LEFT_CURLY_BRACKET "\173" +#define STR_VERTICAL_LINE "\174" +#define STR_RIGHT_CURLY_BRACKET "\175" +#define STR_TILDE "\176" + +#define STRING_ACCEPT0 STR_A STR_C STR_C STR_E STR_P STR_T "\0" +#define STRING_COMMIT0 STR_C STR_O STR_M STR_M STR_I STR_T "\0" +#define STRING_F0 STR_F "\0" +#define STRING_FAIL0 STR_F STR_A STR_I STR_L "\0" +#define STRING_MARK0 STR_M STR_A STR_R STR_K "\0" +#define STRING_PRUNE0 STR_P STR_R STR_U STR_N STR_E "\0" +#define STRING_SKIP0 STR_S STR_K STR_I STR_P "\0" +#define STRING_THEN STR_T STR_H STR_E STR_N + +#define STRING_alpha0 STR_a STR_l STR_p STR_h STR_a "\0" +#define STRING_lower0 STR_l STR_o STR_w STR_e STR_r "\0" +#define STRING_upper0 STR_u STR_p STR_p STR_e STR_r "\0" +#define STRING_alnum0 STR_a STR_l STR_n STR_u STR_m "\0" +#define STRING_ascii0 STR_a STR_s STR_c STR_i STR_i "\0" +#define STRING_blank0 STR_b STR_l STR_a STR_n STR_k "\0" +#define STRING_cntrl0 STR_c STR_n STR_t STR_r STR_l "\0" +#define STRING_digit0 STR_d STR_i STR_g STR_i STR_t "\0" +#define STRING_graph0 STR_g STR_r STR_a STR_p STR_h "\0" +#define STRING_print0 STR_p STR_r STR_i STR_n STR_t "\0" +#define STRING_punct0 STR_p STR_u STR_n STR_c STR_t "\0" +#define STRING_space0 STR_s STR_p STR_a STR_c STR_e "\0" +#define STRING_word0 STR_w STR_o STR_r STR_d "\0" +#define STRING_xdigit STR_x STR_d STR_i STR_g STR_i STR_t + +#define STRING_DEFINE STR_D STR_E STR_F STR_I STR_N STR_E +#define STRING_WEIRD_STARTWORD STR_LEFT_SQUARE_BRACKET STR_COLON STR_LESS_THAN_SIGN STR_COLON STR_RIGHT_SQUARE_BRACKET STR_RIGHT_SQUARE_BRACKET +#define STRING_WEIRD_ENDWORD STR_LEFT_SQUARE_BRACKET STR_COLON STR_GREATER_THAN_SIGN STR_COLON STR_RIGHT_SQUARE_BRACKET STR_RIGHT_SQUARE_BRACKET + +#define STRING_CR_RIGHTPAR STR_C STR_R STR_RIGHT_PARENTHESIS +#define STRING_LF_RIGHTPAR STR_L STR_F STR_RIGHT_PARENTHESIS +#define STRING_CRLF_RIGHTPAR STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS +#define STRING_ANY_RIGHTPAR STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS +#define STRING_ANYCRLF_RIGHTPAR STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS +#define STRING_BSR_ANYCRLF_RIGHTPAR STR_B STR_S STR_R STR_UNDERSCORE STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS +#define STRING_BSR_UNICODE_RIGHTPAR STR_B STR_S STR_R STR_UNDERSCORE STR_U STR_N STR_I STR_C STR_O STR_D STR_E STR_RIGHT_PARENTHESIS +#define STRING_UTF8_RIGHTPAR STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS +#define STRING_UTF16_RIGHTPAR STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS +#define STRING_UTF32_RIGHTPAR STR_U STR_T STR_F STR_3 STR_2 STR_RIGHT_PARENTHESIS +#define STRING_UTF_RIGHTPAR STR_U STR_T STR_F STR_RIGHT_PARENTHESIS +#define STRING_UCP_RIGHTPAR STR_U STR_C STR_P STR_RIGHT_PARENTHESIS +#define STRING_NO_AUTO_POSSESS_RIGHTPAR STR_N STR_O STR_UNDERSCORE STR_A STR_U STR_T STR_O STR_UNDERSCORE STR_P STR_O STR_S STR_S STR_E STR_S STR_S STR_RIGHT_PARENTHESIS +#define STRING_NO_START_OPT_RIGHTPAR STR_N STR_O STR_UNDERSCORE STR_S STR_T STR_A STR_R STR_T STR_UNDERSCORE STR_O STR_P STR_T STR_RIGHT_PARENTHESIS +#define STRING_LIMIT_MATCH_EQ STR_L STR_I STR_M STR_I STR_T STR_UNDERSCORE STR_M STR_A STR_T STR_C STR_H STR_EQUALS_SIGN +#define STRING_LIMIT_RECURSION_EQ STR_L STR_I STR_M STR_I STR_T STR_UNDERSCORE STR_R STR_E STR_C STR_U STR_R STR_S STR_I STR_O STR_N STR_EQUALS_SIGN + +#endif /* SUPPORT_UTF */ + +/* Escape items that are just an encoding of a particular data value. */ + +#ifndef ESC_a +#define ESC_a CHAR_BEL +#endif + +#ifndef ESC_e +#define ESC_e CHAR_ESC +#endif + +#ifndef ESC_f +#define ESC_f CHAR_FF +#endif + +#ifndef ESC_n +#define ESC_n CHAR_LF +#endif + +#ifndef ESC_r +#define ESC_r CHAR_CR +#endif + +/* We can't officially use ESC_t because it is a POSIX reserved identifier +(presumably because of all the others like size_t). */ + +#ifndef ESC_tee +#define ESC_tee CHAR_HT +#endif + +/* Codes for different types of Unicode property */ + +#define PT_ANY 0 /* Any property - matches all chars */ +#define PT_LAMP 1 /* L& - the union of Lu, Ll, Lt */ +#define PT_GC 2 /* Specified general characteristic (e.g. L) */ +#define PT_PC 3 /* Specified particular characteristic (e.g. Lu) */ +#define PT_SC 4 /* Script (e.g. Han) */ +#define PT_ALNUM 5 /* Alphanumeric - the union of L and N */ +#define PT_SPACE 6 /* Perl space - Z plus 9,10,12,13 */ +#define PT_PXSPACE 7 /* POSIX space - Z plus 9,10,11,12,13 */ +#define PT_WORD 8 /* Word - L plus N plus underscore */ +#define PT_CLIST 9 /* Pseudo-property: match character list */ +#define PT_UCNC 10 /* Universal Character nameable character */ +#define PT_TABSIZE 11 /* Size of square table for autopossessify tests */ + +/* The following special properties are used only in XCLASS items, when POSIX +classes are specified and PCRE_UCP is set - in other words, for Unicode +handling of these classes. They are not available via the \p or \P escapes like +those in the above list, and so they do not take part in the autopossessifying +table. */ + +#define PT_PXGRAPH 11 /* [:graph:] - characters that mark the paper */ +#define PT_PXPRINT 12 /* [:print:] - [:graph:] plus non-control spaces */ +#define PT_PXPUNCT 13 /* [:punct:] - punctuation characters */ + +/* Flag bits and data types for the extended class (OP_XCLASS) for classes that +contain characters with values greater than 255. */ + +#define XCL_NOT 0x01 /* Flag: this is a negative class */ +#define XCL_MAP 0x02 /* Flag: a 32-byte map is present */ +#define XCL_HASPROP 0x04 /* Flag: property checks are present. */ + +#define XCL_END 0 /* Marks end of individual items */ +#define XCL_SINGLE 1 /* Single item (one multibyte char) follows */ +#define XCL_RANGE 2 /* A range (two multibyte chars) follows */ +#define XCL_PROP 3 /* Unicode property (2-byte property code follows) */ +#define XCL_NOTPROP 4 /* Unicode inverted property (ditto) */ + +/* These are escaped items that aren't just an encoding of a particular data +value such as \n. They must have non-zero values, as check_escape() returns 0 +for a data character. Also, they must appear in the same order as in the +opcode definitions below, up to ESC_z. There's a dummy for OP_ALLANY because it +corresponds to "." in DOTALL mode rather than an escape sequence. It is also +used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In +non-DOTALL mode, "." behaves like \N. + +The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc. +when PCRE_UCP is set and replacement of \d etc by \p sequences is required. +They must be contiguous, and remain in order so that the replacements can be +looked up from a table. + +Negative numbers are used to encode a backreference (\1, \2, \3, etc.) in +check_escape(). There are two tests in the code for an escape +greater than ESC_b and less than ESC_Z to detect the types that may be +repeated. These are the types that consume characters. If any new escapes are +put in between that don't consume a character, that code will have to change. +*/ + +enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, + ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H, + ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, + ESC_E, ESC_Q, ESC_g, ESC_k, + ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu }; + + +/********************** Opcode definitions ******************/ + +/****** NOTE NOTE NOTE ****** + +Starting from 1 (i.e. after OP_END), the values up to OP_EOD must correspond in +order to the list of escapes immediately above. Furthermore, values up to +OP_DOLLM must not be changed without adjusting the table called autoposstab in +pcre_compile.c + +Whenever this list is updated, the two macro definitions that follow must be +updated to match. The possessification table called "opcode_possessify" in +pcre_compile.c must also be updated, and also the tables called "coptable" +and "poptable" in pcre_dfa_exec.c. + +****** NOTE NOTE NOTE ******/ + + +/* The values between FIRST_AUTOTAB_OP and LAST_AUTOTAB_RIGHT_OP, inclusive, +are used in a table for deciding whether a repeated character type can be +auto-possessified. */ + +#define FIRST_AUTOTAB_OP OP_NOT_DIGIT +#define LAST_AUTOTAB_LEFT_OP OP_EXTUNI +#define LAST_AUTOTAB_RIGHT_OP OP_DOLLM + +enum { + OP_END, /* 0 End of pattern */ + + /* Values corresponding to backslashed metacharacters */ + + OP_SOD, /* 1 Start of data: \A */ + OP_SOM, /* 2 Start of match (subject + offset): \G */ + OP_SET_SOM, /* 3 Set start of match (\K) */ + OP_NOT_WORD_BOUNDARY, /* 4 \B */ + OP_WORD_BOUNDARY, /* 5 \b */ + OP_NOT_DIGIT, /* 6 \D */ + OP_DIGIT, /* 7 \d */ + OP_NOT_WHITESPACE, /* 8 \S */ + OP_WHITESPACE, /* 9 \s */ + OP_NOT_WORDCHAR, /* 10 \W */ + OP_WORDCHAR, /* 11 \w */ + + OP_ANY, /* 12 Match any character except newline (\N) */ + OP_ALLANY, /* 13 Match any character */ + OP_ANYBYTE, /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */ + OP_NOTPROP, /* 15 \P (not Unicode property) */ + OP_PROP, /* 16 \p (Unicode property) */ + OP_ANYNL, /* 17 \R (any newline sequence) */ + OP_NOT_HSPACE, /* 18 \H (not horizontal whitespace) */ + OP_HSPACE, /* 19 \h (horizontal whitespace) */ + OP_NOT_VSPACE, /* 20 \V (not vertical whitespace) */ + OP_VSPACE, /* 21 \v (vertical whitespace) */ + OP_EXTUNI, /* 22 \X (extended Unicode sequence */ + OP_EODN, /* 23 End of data or \n at end of data (\Z) */ + OP_EOD, /* 24 End of data (\z) */ + + /* Line end assertions */ + + OP_DOLL, /* 25 End of line - not multiline */ + OP_DOLLM, /* 26 End of line - multiline */ + OP_CIRC, /* 27 Start of line - not multiline */ + OP_CIRCM, /* 28 Start of line - multiline */ + + /* Single characters; caseful must precede the caseless ones */ + + OP_CHAR, /* 29 Match one character, casefully */ + OP_CHARI, /* 30 Match one character, caselessly */ + OP_NOT, /* 31 Match one character, not the given one, casefully */ + OP_NOTI, /* 32 Match one character, not the given one, caselessly */ + + /* The following sets of 13 opcodes must always be kept in step because + the offset from the first one is used to generate the others. */ + + /* Repeated characters; caseful must precede the caseless ones */ + + OP_STAR, /* 33 The maximizing and minimizing versions of */ + OP_MINSTAR, /* 34 these six opcodes must come in pairs, with */ + OP_PLUS, /* 35 the minimizing one second. */ + OP_MINPLUS, /* 36 */ + OP_QUERY, /* 37 */ + OP_MINQUERY, /* 38 */ + + OP_UPTO, /* 39 From 0 to n matches of one character, caseful*/ + OP_MINUPTO, /* 40 */ + OP_EXACT, /* 41 Exactly n matches */ + + OP_POSSTAR, /* 42 Possessified star, caseful */ + OP_POSPLUS, /* 43 Possessified plus, caseful */ + OP_POSQUERY, /* 44 Posesssified query, caseful */ + OP_POSUPTO, /* 45 Possessified upto, caseful */ + + /* Repeated characters; caseless must follow the caseful ones */ + + OP_STARI, /* 46 */ + OP_MINSTARI, /* 47 */ + OP_PLUSI, /* 48 */ + OP_MINPLUSI, /* 49 */ + OP_QUERYI, /* 50 */ + OP_MINQUERYI, /* 51 */ + + OP_UPTOI, /* 52 From 0 to n matches of one character, caseless */ + OP_MINUPTOI, /* 53 */ + OP_EXACTI, /* 54 */ + + OP_POSSTARI, /* 55 Possessified star, caseless */ + OP_POSPLUSI, /* 56 Possessified plus, caseless */ + OP_POSQUERYI, /* 57 Posesssified query, caseless */ + OP_POSUPTOI, /* 58 Possessified upto, caseless */ + + /* The negated ones must follow the non-negated ones, and match them */ + /* Negated repeated character, caseful; must precede the caseless ones */ + + OP_NOTSTAR, /* 59 The maximizing and minimizing versions of */ + OP_NOTMINSTAR, /* 60 these six opcodes must come in pairs, with */ + OP_NOTPLUS, /* 61 the minimizing one second. They must be in */ + OP_NOTMINPLUS, /* 62 exactly the same order as those above. */ + OP_NOTQUERY, /* 63 */ + OP_NOTMINQUERY, /* 64 */ + + OP_NOTUPTO, /* 65 From 0 to n matches, caseful */ + OP_NOTMINUPTO, /* 66 */ + OP_NOTEXACT, /* 67 Exactly n matches */ + + OP_NOTPOSSTAR, /* 68 Possessified versions, caseful */ + OP_NOTPOSPLUS, /* 69 */ + OP_NOTPOSQUERY, /* 70 */ + OP_NOTPOSUPTO, /* 71 */ + + /* Negated repeated character, caseless; must follow the caseful ones */ + + OP_NOTSTARI, /* 72 */ + OP_NOTMINSTARI, /* 73 */ + OP_NOTPLUSI, /* 74 */ + OP_NOTMINPLUSI, /* 75 */ + OP_NOTQUERYI, /* 76 */ + OP_NOTMINQUERYI, /* 77 */ + + OP_NOTUPTOI, /* 78 From 0 to n matches, caseless */ + OP_NOTMINUPTOI, /* 79 */ + OP_NOTEXACTI, /* 80 Exactly n matches */ + + OP_NOTPOSSTARI, /* 81 Possessified versions, caseless */ + OP_NOTPOSPLUSI, /* 82 */ + OP_NOTPOSQUERYI, /* 83 */ + OP_NOTPOSUPTOI, /* 84 */ + + /* Character types */ + + OP_TYPESTAR, /* 85 The maximizing and minimizing versions of */ + OP_TYPEMINSTAR, /* 86 these six opcodes must come in pairs, with */ + OP_TYPEPLUS, /* 87 the minimizing one second. These codes must */ + OP_TYPEMINPLUS, /* 88 be in exactly the same order as those above. */ + OP_TYPEQUERY, /* 89 */ + OP_TYPEMINQUERY, /* 90 */ + + OP_TYPEUPTO, /* 91 From 0 to n matches */ + OP_TYPEMINUPTO, /* 92 */ + OP_TYPEEXACT, /* 93 Exactly n matches */ + + OP_TYPEPOSSTAR, /* 94 Possessified versions */ + OP_TYPEPOSPLUS, /* 95 */ + OP_TYPEPOSQUERY, /* 96 */ + OP_TYPEPOSUPTO, /* 97 */ + + /* These are used for character classes and back references; only the + first six are the same as the sets above. */ + + OP_CRSTAR, /* 98 The maximizing and minimizing versions of */ + OP_CRMINSTAR, /* 99 all these opcodes must come in pairs, with */ + OP_CRPLUS, /* 100 the minimizing one second. These codes must */ + OP_CRMINPLUS, /* 101 be in exactly the same order as those above. */ + OP_CRQUERY, /* 102 */ + OP_CRMINQUERY, /* 103 */ + + OP_CRRANGE, /* 104 These are different to the three sets above. */ + OP_CRMINRANGE, /* 105 */ + + OP_CRPOSSTAR, /* 106 Possessified versions */ + OP_CRPOSPLUS, /* 107 */ + OP_CRPOSQUERY, /* 108 */ + OP_CRPOSRANGE, /* 109 */ + + /* End of quantifier opcodes */ + + OP_CLASS, /* 110 Match a character class, chars < 256 only */ + OP_NCLASS, /* 111 Same, but the bitmap was created from a negative + class - the difference is relevant only when a + character > 255 is encountered. */ + OP_XCLASS, /* 112 Extended class for handling > 255 chars within the + class. This does both positive and negative. */ + OP_REF, /* 113 Match a back reference, casefully */ + OP_REFI, /* 114 Match a back reference, caselessly */ + OP_DNREF, /* 115 Match a duplicate name backref, casefully */ + OP_DNREFI, /* 116 Match a duplicate name backref, caselessly */ + OP_RECURSE, /* 117 Match a numbered subpattern (possibly recursive) */ + OP_CALLOUT, /* 118 Call out to external function if provided */ + + OP_ALT, /* 119 Start of alternation */ + OP_KET, /* 120 End of group that doesn't have an unbounded repeat */ + OP_KETRMAX, /* 121 These two must remain together and in this */ + OP_KETRMIN, /* 122 order. They are for groups the repeat for ever. */ + OP_KETRPOS, /* 123 Possessive unlimited repeat. */ + + /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four + asserts must remain in order. */ + + OP_REVERSE, /* 124 Move pointer back - used in lookbehind assertions */ + OP_ASSERT, /* 125 Positive lookahead */ + OP_ASSERT_NOT, /* 126 Negative lookahead */ + OP_ASSERTBACK, /* 127 Positive lookbehind */ + OP_ASSERTBACK_NOT, /* 128 Negative lookbehind */ + + /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately + after the assertions, with ONCE first, as there's a test for >= ONCE for a + subpattern that isn't an assertion. The POS versions must immediately follow + the non-POS versions in each case. */ + + OP_ONCE, /* 129 Atomic group, contains captures */ + OP_ONCE_NC, /* 130 Atomic group containing no captures */ + OP_BRA, /* 131 Start of non-capturing bracket */ + OP_BRAPOS, /* 132 Ditto, with unlimited, possessive repeat */ + OP_CBRA, /* 133 Start of capturing bracket */ + OP_CBRAPOS, /* 134 Ditto, with unlimited, possessive repeat */ + OP_COND, /* 135 Conditional group */ + + /* These five must follow the previous five, in the same order. There's a + check for >= SBRA to distinguish the two sets. */ + + OP_SBRA, /* 136 Start of non-capturing bracket, check empty */ + OP_SBRAPOS, /* 137 Ditto, with unlimited, possessive repeat */ + OP_SCBRA, /* 138 Start of capturing bracket, check empty */ + OP_SCBRAPOS, /* 139 Ditto, with unlimited, possessive repeat */ + OP_SCOND, /* 140 Conditional group, check empty */ + + /* The next two pairs must (respectively) be kept together. */ + + OP_CREF, /* 141 Used to hold a capture number as condition */ + OP_DNCREF, /* 142 Used to point to duplicate names as a condition */ + OP_RREF, /* 143 Used to hold a recursion number as condition */ + OP_DNRREF, /* 144 Used to point to duplicate names as a condition */ + OP_DEF, /* 145 The DEFINE condition */ + + OP_BRAZERO, /* 146 These two must remain together and in this */ + OP_BRAMINZERO, /* 147 order. */ + OP_BRAPOSZERO, /* 148 */ + + /* These are backtracking control verbs */ + + OP_MARK, /* 149 always has an argument */ + OP_PRUNE, /* 150 */ + OP_PRUNE_ARG, /* 151 same, but with argument */ + OP_SKIP, /* 152 */ + OP_SKIP_ARG, /* 153 same, but with argument */ + OP_THEN, /* 154 */ + OP_THEN_ARG, /* 155 same, but with argument */ + OP_COMMIT, /* 156 */ + + /* These are forced failure and success verbs */ + + OP_FAIL, /* 157 */ + OP_ACCEPT, /* 158 */ + OP_ASSERT_ACCEPT, /* 159 Used inside assertions */ + OP_CLOSE, /* 160 Used before OP_ACCEPT to close open captures */ + + /* This is used to skip a subpattern with a {0} quantifier */ + + OP_SKIPZERO, /* 161 */ + + /* This is not an opcode, but is used to check that tables indexed by opcode + are the correct length, in order to catch updating errors - there have been + some in the past. */ + + OP_TABLE_LENGTH +}; + +/* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro +definitions that follow must also be updated to match. There are also tables +called "opcode_possessify" in pcre_compile.c and "coptable" and "poptable" in +pcre_dfa_exec.c that must be updated. */ + + +/* This macro defines textual names for all the opcodes. These are used only +for debugging, and some of them are only partial names. The macro is referenced +only in pcre_printint.c, which fills out the full names in many cases (and in +some cases doesn't actually use these names at all). */ + +#define OP_NAME_LIST \ + "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d", \ + "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte", \ + "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v", \ + "extuni", "\\Z", "\\z", \ + "$", "$", "^", "^", "char", "chari", "not", "noti", \ + "*", "*?", "+", "+?", "?", "??", \ + "{", "{", "{", \ + "*+","++", "?+", "{", \ + "*", "*?", "+", "+?", "?", "??", \ + "{", "{", "{", \ + "*+","++", "?+", "{", \ + "*", "*?", "+", "+?", "?", "??", \ + "{", "{", "{", \ + "*+","++", "?+", "{", \ + "*", "*?", "+", "+?", "?", "??", \ + "{", "{", "{", \ + "*+","++", "?+", "{", \ + "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \ + "*+","++", "?+", "{", \ + "*", "*?", "+", "+?", "?", "??", "{", "{", \ + "*+","++", "?+", "{", \ + "class", "nclass", "xclass", "Ref", "Refi", "DnRef", "DnRefi", \ + "Recurse", "Callout", \ + "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos", \ + "Reverse", "Assert", "Assert not", "AssertB", "AssertB not", \ + "Once", "Once_NC", \ + "Bra", "BraPos", "CBra", "CBraPos", \ + "Cond", \ + "SBra", "SBraPos", "SCBra", "SCBraPos", \ + "SCond", \ + "Cond ref", "Cond dnref", "Cond rec", "Cond dnrec", "Cond def", \ + "Brazero", "Braminzero", "Braposzero", \ + "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP", \ + "*THEN", "*THEN", "*COMMIT", "*FAIL", \ + "*ACCEPT", "*ASSERT_ACCEPT", \ + "Close", "Skip zero" + + +/* This macro defines the length of fixed length operations in the compiled +regex. The lengths are used when searching for specific things, and also in the +debugging printing of a compiled regex. We use a macro so that it can be +defined close to the definitions of the opcodes themselves. + +As things have been extended, some of these are no longer fixed lenths, but are +minima instead. For example, the length of a single-character repeat may vary +in UTF-8 mode. The code that uses this table must know about such things. */ + +#define OP_LENGTHS \ + 1, /* End */ \ + 1, 1, 1, 1, 1, /* \A, \G, \K, \B, \b */ \ + 1, 1, 1, 1, 1, 1, /* \D, \d, \S, \s, \W, \w */ \ + 1, 1, 1, /* Any, AllAny, Anybyte */ \ + 3, 3, /* \P, \p */ \ + 1, 1, 1, 1, 1, /* \R, \H, \h, \V, \v */ \ + 1, /* \X */ \ + 1, 1, 1, 1, 1, 1, /* \Z, \z, $, $M ^, ^M */ \ + 2, /* Char - the minimum length */ \ + 2, /* Chari - the minimum length */ \ + 2, /* not */ \ + 2, /* noti */ \ + /* Positive single-char repeats ** These are */ \ + 2, 2, 2, 2, 2, 2, /* *, *?, +, +?, ?, ?? ** minima in */ \ + 2+IMM2_SIZE, 2+IMM2_SIZE, /* upto, minupto ** mode */ \ + 2+IMM2_SIZE, /* exact */ \ + 2, 2, 2, 2+IMM2_SIZE, /* *+, ++, ?+, upto+ */ \ + 2, 2, 2, 2, 2, 2, /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8 */ \ + 2+IMM2_SIZE, 2+IMM2_SIZE, /* upto I, minupto I */ \ + 2+IMM2_SIZE, /* exact I */ \ + 2, 2, 2, 2+IMM2_SIZE, /* *+I, ++I, ?+I, upto+I */ \ + /* Negative single-char repeats - only for chars < 256 */ \ + 2, 2, 2, 2, 2, 2, /* NOT *, *?, +, +?, ?, ?? */ \ + 2+IMM2_SIZE, 2+IMM2_SIZE, /* NOT upto, minupto */ \ + 2+IMM2_SIZE, /* NOT exact */ \ + 2, 2, 2, 2+IMM2_SIZE, /* Possessive NOT *, +, ?, upto */ \ + 2, 2, 2, 2, 2, 2, /* NOT *I, *?I, +I, +?I, ?I, ??I */ \ + 2+IMM2_SIZE, 2+IMM2_SIZE, /* NOT upto I, minupto I */ \ + 2+IMM2_SIZE, /* NOT exact I */ \ + 2, 2, 2, 2+IMM2_SIZE, /* Possessive NOT *I, +I, ?I, upto I */ \ + /* Positive type repeats */ \ + 2, 2, 2, 2, 2, 2, /* Type *, *?, +, +?, ?, ?? */ \ + 2+IMM2_SIZE, 2+IMM2_SIZE, /* Type upto, minupto */ \ + 2+IMM2_SIZE, /* Type exact */ \ + 2, 2, 2, 2+IMM2_SIZE, /* Possessive *+, ++, ?+, upto+ */ \ + /* Character class & ref repeats */ \ + 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ \ + 1+2*IMM2_SIZE, 1+2*IMM2_SIZE, /* CRRANGE, CRMINRANGE */ \ + 1, 1, 1, 1+2*IMM2_SIZE, /* Possessive *+, ++, ?+, CRPOSRANGE */ \ + 1+(32/sizeof(pcre_uchar)), /* CLASS */ \ + 1+(32/sizeof(pcre_uchar)), /* NCLASS */ \ + 0, /* XCLASS - variable length */ \ + 1+IMM2_SIZE, /* REF */ \ + 1+IMM2_SIZE, /* REFI */ \ + 1+2*IMM2_SIZE, /* DNREF */ \ + 1+2*IMM2_SIZE, /* DNREFI */ \ + 1+LINK_SIZE, /* RECURSE */ \ + 2+2*LINK_SIZE, /* CALLOUT */ \ + 1+LINK_SIZE, /* Alt */ \ + 1+LINK_SIZE, /* Ket */ \ + 1+LINK_SIZE, /* KetRmax */ \ + 1+LINK_SIZE, /* KetRmin */ \ + 1+LINK_SIZE, /* KetRpos */ \ + 1+LINK_SIZE, /* Reverse */ \ + 1+LINK_SIZE, /* Assert */ \ + 1+LINK_SIZE, /* Assert not */ \ + 1+LINK_SIZE, /* Assert behind */ \ + 1+LINK_SIZE, /* Assert behind not */ \ + 1+LINK_SIZE, /* ONCE */ \ + 1+LINK_SIZE, /* ONCE_NC */ \ + 1+LINK_SIZE, /* BRA */ \ + 1+LINK_SIZE, /* BRAPOS */ \ + 1+LINK_SIZE+IMM2_SIZE, /* CBRA */ \ + 1+LINK_SIZE+IMM2_SIZE, /* CBRAPOS */ \ + 1+LINK_SIZE, /* COND */ \ + 1+LINK_SIZE, /* SBRA */ \ + 1+LINK_SIZE, /* SBRAPOS */ \ + 1+LINK_SIZE+IMM2_SIZE, /* SCBRA */ \ + 1+LINK_SIZE+IMM2_SIZE, /* SCBRAPOS */ \ + 1+LINK_SIZE, /* SCOND */ \ + 1+IMM2_SIZE, 1+2*IMM2_SIZE, /* CREF, DNCREF */ \ + 1+IMM2_SIZE, 1+2*IMM2_SIZE, /* RREF, DNRREF */ \ + 1, /* DEF */ \ + 1, 1, 1, /* BRAZERO, BRAMINZERO, BRAPOSZERO */ \ + 3, 1, 3, /* MARK, PRUNE, PRUNE_ARG */ \ + 1, 3, /* SKIP, SKIP_ARG */ \ + 1, 3, /* THEN, THEN_ARG */ \ + 1, 1, 1, 1, /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT */ \ + 1+IMM2_SIZE, 1 /* CLOSE, SKIPZERO */ + +/* A magic value for OP_RREF to indicate the "any recursion" condition. */ + +#define RREF_ANY 0xffff + +/* Compile time error code numbers. They are given names so that they can more +easily be tracked. When a new number is added, the table called eint in +pcreposix.c must be updated. */ + +enum { ERR0, ERR1, ERR2, ERR3, ERR4, ERR5, ERR6, ERR7, ERR8, ERR9, + ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19, + ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29, + ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39, + ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49, + ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59, + ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69, + ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERR78, ERR79, + ERR80, ERR81, ERR82, ERR83, ERR84, ERR85, ERR86, ERRCOUNT }; + +/* JIT compiling modes. The function list is indexed by them. */ + +enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE, + JIT_NUMBER_OF_COMPILE_MODES }; + +/* The real format of the start of the pcre block; the index of names and the +code vector run on as long as necessary after the end. We store an explicit +offset to the name table so that if a regex is compiled on one host, saved, and +then run on another where the size of pointers is different, all might still +be well. + +The size of the structure must be a multiple of 8 bytes. For the case of +compiled-on-4 and run-on-8, we include an extra pointer that is always NULL so +that there are an even number of pointers which therefore are a multiple of 8 +bytes. + +It is necessary to fork the struct for the 32 bit library, since it needs to +use pcre_uint32 for first_char and req_char. We can't put an ifdef inside the +typedef because pcretest needs access to the struct of the 8-, 16- and 32-bit +variants. + +*** WARNING *** +When new fields are added to these structures, remember to adjust the code in +pcre_byte_order.c that is concerned with swapping the byte order of the fields +when a compiled regex is reloaded on a host with different endianness. +*** WARNING *** +There is also similar byte-flipping code in pcretest.c, which is used for +testing the byte-flipping features. It must also be kept in step. +*** WARNING *** +*/ + +typedef struct real_pcre8_or_16 { + pcre_uint32 magic_number; + pcre_uint32 size; /* Total that was malloced */ + pcre_uint32 options; /* Public options */ + pcre_uint32 flags; /* Private flags */ + pcre_uint32 limit_match; /* Limit set from regex */ + pcre_uint32 limit_recursion; /* Limit set from regex */ + pcre_uint16 first_char; /* Starting character */ + pcre_uint16 req_char; /* This character must be seen */ + pcre_uint16 max_lookbehind; /* Longest lookbehind (characters) */ + pcre_uint16 top_bracket; /* Highest numbered group */ + pcre_uint16 top_backref; /* Highest numbered back reference */ + pcre_uint16 name_table_offset; /* Offset to name table that follows */ + pcre_uint16 name_entry_size; /* Size of any name items */ + pcre_uint16 name_count; /* Number of name items */ + pcre_uint16 ref_count; /* Reference count */ + pcre_uint16 dummy1; /* To ensure size is a multiple of 8 */ + pcre_uint16 dummy2; /* To ensure size is a multiple of 8 */ + pcre_uint16 dummy3; /* To ensure size is a multiple of 8 */ + const pcre_uint8 *tables; /* Pointer to tables or NULL for std */ + void *nullpad; /* NULL padding */ +} real_pcre8_or_16; + +typedef struct real_pcre8_or_16 real_pcre; +typedef struct real_pcre8_or_16 real_pcre16; + +typedef struct real_pcre32 { + pcre_uint32 magic_number; + pcre_uint32 size; /* Total that was malloced */ + pcre_uint32 options; /* Public options */ + pcre_uint32 flags; /* Private flags */ + pcre_uint32 limit_match; /* Limit set from regex */ + pcre_uint32 limit_recursion; /* Limit set from regex */ + pcre_uint32 first_char; /* Starting character */ + pcre_uint32 req_char; /* This character must be seen */ + pcre_uint16 max_lookbehind; /* Longest lookbehind (characters) */ + pcre_uint16 top_bracket; /* Highest numbered group */ + pcre_uint16 top_backref; /* Highest numbered back reference */ + pcre_uint16 name_table_offset; /* Offset to name table that follows */ + pcre_uint16 name_entry_size; /* Size of any name items */ + pcre_uint16 name_count; /* Number of name items */ + pcre_uint16 ref_count; /* Reference count */ + pcre_uint16 dummy; /* To ensure size is a multiple of 8 */ + const pcre_uint8 *tables; /* Pointer to tables or NULL for std */ + void *nullpad; /* NULL padding */ +} real_pcre32; + +#if defined COMPILE_PCRE8 +#define REAL_PCRE real_pcre +#elif defined COMPILE_PCRE16 +#define REAL_PCRE real_pcre16 +#elif defined COMPILE_PCRE32 +#define REAL_PCRE real_pcre32 +#endif + +/* Assert that the size of REAL_PCRE is divisible by 8 */ +typedef int __assert_real_pcre_size_divisible_8[(sizeof(REAL_PCRE) % 8) == 0 ? 1 : -1]; + +/* Needed in pcretest to access some fields in the real_pcre* structures + * directly. They're unified for 8/16/32 bits since the structs only differ + * after these fields; if that ever changes, need to fork those defines into + * 8/16 and 32 bit versions. */ +#define REAL_PCRE_MAGIC(re) (((REAL_PCRE*)re)->magic_number) +#define REAL_PCRE_SIZE(re) (((REAL_PCRE*)re)->size) +#define REAL_PCRE_OPTIONS(re) (((REAL_PCRE*)re)->options) +#define REAL_PCRE_FLAGS(re) (((REAL_PCRE*)re)->flags) + +/* The format of the block used to store data from pcre_study(). The same +remark (see NOTE above) about extending this structure applies. */ + +typedef struct pcre_study_data { + pcre_uint32 size; /* Total that was malloced */ + pcre_uint32 flags; /* Private flags */ + pcre_uint8 start_bits[32]; /* Starting char bits */ + pcre_uint32 minlength; /* Minimum subject length */ +} pcre_study_data; + +/* Structure for building a chain of open capturing subpatterns during +compiling, so that instructions to close them can be compiled when (*ACCEPT) is +encountered. This is also used to identify subpatterns that contain recursive +back references to themselves, so that they can be made atomic. */ + +typedef struct open_capitem { + struct open_capitem *next; /* Chain link */ + pcre_uint16 number; /* Capture number */ + pcre_uint16 flag; /* Set TRUE if recursive back ref */ +} open_capitem; + +/* Structure for building a list of named groups during the first pass of +compiling. */ + +typedef struct named_group { + const pcre_uchar *name; /* Points to the name in the pattern */ + int length; /* Length of the name */ + pcre_uint32 number; /* Group number */ +} named_group; + +/* Structure for passing "static" information around between the functions +doing the compiling, so that they are thread-safe. */ + +typedef struct compile_data { + const pcre_uint8 *lcc; /* Points to lower casing table */ + const pcre_uint8 *fcc; /* Points to case-flipping table */ + const pcre_uint8 *cbits; /* Points to character type table */ + const pcre_uint8 *ctypes; /* Points to table of type maps */ + const pcre_uchar *start_workspace;/* The start of working space */ + const pcre_uchar *start_code; /* The start of the compiled code */ + const pcre_uchar *start_pattern; /* The start of the pattern */ + const pcre_uchar *end_pattern; /* The end of the pattern */ + pcre_uchar *hwm; /* High watermark of workspace */ + open_capitem *open_caps; /* Chain of open capture items */ + named_group *named_groups; /* Points to vector in pre-compile */ + pcre_uchar *name_table; /* The name/number table */ + int names_found; /* Number of entries so far */ + int name_entry_size; /* Size of each entry */ + int named_group_list_size; /* Number of entries in the list */ + int workspace_size; /* Size of workspace */ + unsigned int bracount; /* Count of capturing parens as we compile */ + int final_bracount; /* Saved value after first pass */ + int max_lookbehind; /* Maximum lookbehind (characters) */ + int top_backref; /* Maximum back reference */ + unsigned int backref_map; /* Bitmap of low back refs */ + unsigned int namedrefcount; /* Number of backreferences by name */ + int parens_depth; /* Depth of nested parentheses */ + int assert_depth; /* Depth of nested assertions */ + pcre_uint32 external_options; /* External (initial) options */ + pcre_uint32 external_flags; /* External flag bits to be set */ + int req_varyopt; /* "After variable item" flag for reqbyte */ + BOOL had_accept; /* (*ACCEPT) encountered */ + BOOL had_pruneorskip; /* (*PRUNE) or (*SKIP) encountered */ + BOOL check_lookbehind; /* Lookbehinds need later checking */ + BOOL dupnames; /* Duplicate names exist */ + BOOL dupgroups; /* Duplicate groups exist: (?| found */ + BOOL iscondassert; /* Next assert is a condition */ + int nltype; /* Newline type */ + int nllen; /* Newline string length */ + pcre_uchar nl[4]; /* Newline string when fixed length */ +} compile_data; + +/* Structure for maintaining a chain of pointers to the currently incomplete +branches, for testing for left recursion while compiling. */ + +typedef struct branch_chain { + struct branch_chain *outer; + pcre_uchar *current_branch; +} branch_chain; + +/* Structure for mutual recursion detection. */ + +typedef struct recurse_check { + struct recurse_check *prev; + const pcre_uchar *group; +} recurse_check; + +/* Structure for items in a linked list that represents an explicit recursive +call within the pattern; used by pcre_exec(). */ + +typedef struct recursion_info { + struct recursion_info *prevrec; /* Previous recursion record (or NULL) */ + unsigned int group_num; /* Number of group that was called */ + int *offset_save; /* Pointer to start of saved offsets */ + int saved_max; /* Number of saved offsets */ + int saved_capture_last; /* Last capture number */ + PCRE_PUCHAR subject_position; /* Position at start of recursion */ +} recursion_info; + +/* A similar structure for pcre_dfa_exec(). */ + +typedef struct dfa_recursion_info { + struct dfa_recursion_info *prevrec; + int group_num; + PCRE_PUCHAR subject_position; +} dfa_recursion_info; + +/* Structure for building a chain of data for holding the values of the subject +pointer at the start of each subpattern, so as to detect when an empty string +has been matched by a subpattern - to break infinite loops; used by +pcre_exec(). */ + +typedef struct eptrblock { + struct eptrblock *epb_prev; + PCRE_PUCHAR epb_saved_eptr; +} eptrblock; + + +/* Structure for passing "static" information around between the functions +doing traditional NFA matching, so that they are thread-safe. */ + +typedef struct match_data { + unsigned long int match_call_count; /* As it says */ + unsigned long int match_limit; /* As it says */ + unsigned long int match_limit_recursion; /* As it says */ + int *offset_vector; /* Offset vector */ + int offset_end; /* One past the end */ + int offset_max; /* The maximum usable for return data */ + int nltype; /* Newline type */ + int nllen; /* Newline string length */ + int name_count; /* Number of names in name table */ + int name_entry_size; /* Size of entry in names table */ + unsigned int skip_arg_count; /* For counting SKIP_ARGs */ + unsigned int ignore_skip_arg; /* For re-run when SKIP arg name not found */ + pcre_uchar *name_table; /* Table of names */ + pcre_uchar nl[4]; /* Newline string when fixed */ + const pcre_uint8 *lcc; /* Points to lower casing table */ + const pcre_uint8 *fcc; /* Points to case-flipping table */ + const pcre_uint8 *ctypes; /* Points to table of type maps */ + BOOL notbol; /* NOTBOL flag */ + BOOL noteol; /* NOTEOL flag */ + BOOL utf; /* UTF-8 / UTF-16 flag */ + BOOL jscript_compat; /* JAVASCRIPT_COMPAT flag */ + BOOL use_ucp; /* PCRE_UCP flag */ + BOOL endonly; /* Dollar not before final \n */ + BOOL notempty; /* Empty string match not wanted */ + BOOL notempty_atstart; /* Empty string match at start not wanted */ + BOOL hitend; /* Hit the end of the subject at some point */ + BOOL bsr_anycrlf; /* \R is just any CRLF, not full Unicode */ + BOOL hasthen; /* Pattern contains (*THEN) */ + const pcre_uchar *start_code; /* For use when recursing */ + PCRE_PUCHAR start_subject; /* Start of the subject string */ + PCRE_PUCHAR end_subject; /* End of the subject string */ + PCRE_PUCHAR start_match_ptr; /* Start of matched string */ + PCRE_PUCHAR end_match_ptr; /* Subject position at end match */ + PCRE_PUCHAR start_used_ptr; /* Earliest consulted character */ + int partial; /* PARTIAL options */ + int end_offset_top; /* Highwater mark at end of match */ + pcre_int32 capture_last; /* Most recent capture number + overflow flag */ + int start_offset; /* The start offset value */ + int match_function_type; /* Set for certain special calls of MATCH() */ + eptrblock *eptrchain; /* Chain of eptrblocks for tail recursions */ + int eptrn; /* Next free eptrblock */ + recursion_info *recursive; /* Linked list of recursion data */ + void *callout_data; /* To pass back to callouts */ + const pcre_uchar *mark; /* Mark pointer to pass back on success */ + const pcre_uchar *nomatch_mark;/* Mark pointer to pass back on failure */ + const pcre_uchar *once_target; /* Where to back up to for atomic groups */ +#ifdef NO_RECURSE + void *match_frames_base; /* For remembering malloc'd frames */ +#endif +} match_data; + +/* A similar structure is used for the same purpose by the DFA matching +functions. */ + +typedef struct dfa_match_data { + const pcre_uchar *start_code; /* Start of the compiled pattern */ + const pcre_uchar *start_subject ; /* Start of the subject string */ + const pcre_uchar *end_subject; /* End of subject string */ + const pcre_uchar *start_used_ptr; /* Earliest consulted character */ + const pcre_uint8 *tables; /* Character tables */ + int start_offset; /* The start offset value */ + int moptions; /* Match options */ + int poptions; /* Pattern options */ + int nltype; /* Newline type */ + int nllen; /* Newline string length */ + pcre_uchar nl[4]; /* Newline string when fixed */ + void *callout_data; /* To pass back to callouts */ + dfa_recursion_info *recursive; /* Linked list of recursion data */ +} dfa_match_data; + +/* Bit definitions for entries in the pcre_ctypes table. */ + +#define ctype_space 0x01 +#define ctype_letter 0x02 +#define ctype_digit 0x04 +#define ctype_xdigit 0x08 +#define ctype_word 0x10 /* alphanumeric or '_' */ +#define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */ + +/* Offsets for the bitmap tables in pcre_cbits. Each table contains a set +of bits for a class map. Some classes are built by combining these tables. */ + +#define cbit_space 0 /* [:space:] or \s */ +#define cbit_xdigit 32 /* [:xdigit:] */ +#define cbit_digit 64 /* [:digit:] or \d */ +#define cbit_upper 96 /* [:upper:] */ +#define cbit_lower 128 /* [:lower:] */ +#define cbit_word 160 /* [:word:] or \w */ +#define cbit_graph 192 /* [:graph:] */ +#define cbit_print 224 /* [:print:] */ +#define cbit_punct 256 /* [:punct:] */ +#define cbit_cntrl 288 /* [:cntrl:] */ +#define cbit_length 320 /* Length of the cbits table */ + +/* Offsets of the various tables from the base tables pointer, and +total length. */ + +#define lcc_offset 0 +#define fcc_offset 256 +#define cbits_offset 512 +#define ctypes_offset (cbits_offset + cbit_length) +#define tables_length (ctypes_offset + 256) + +/* Internal function and data prefixes. */ + +#if defined COMPILE_PCRE8 +#ifndef PUBL +#define PUBL(name) pcre_##name +#endif +#ifndef PRIV +#define PRIV(name) _pcre_##name +#endif +#elif defined COMPILE_PCRE16 +#ifndef PUBL +#define PUBL(name) pcre16_##name +#endif +#ifndef PRIV +#define PRIV(name) _pcre16_##name +#endif +#elif defined COMPILE_PCRE32 +#ifndef PUBL +#define PUBL(name) pcre32_##name +#endif +#ifndef PRIV +#define PRIV(name) _pcre32_##name +#endif +#else +#error Unsupported compiling mode +#endif /* COMPILE_PCRE[8|16|32] */ + +/* Layout of the UCP type table that translates property names into types and +codes. Each entry used to point directly to a name, but to reduce the number of +relocations in shared libraries, it now has an offset into a single string +instead. */ + +typedef struct { + pcre_uint16 name_offset; + pcre_uint16 type; + pcre_uint16 value; +} ucp_type_table; + + +/* Internal shared data tables. These are tables that are used by more than one +of the exported public functions. They have to be "external" in the C sense, +but are not part of the PCRE public API. The data for these tables is in the +pcre_tables.c module. */ + +#ifdef COMPILE_PCRE8 +extern const int PRIV(utf8_table1)[]; +extern const int PRIV(utf8_table1_size); +extern const int PRIV(utf8_table2)[]; +extern const int PRIV(utf8_table3)[]; +extern const pcre_uint8 PRIV(utf8_table4)[]; +#endif /* COMPILE_PCRE8 */ + +extern const char PRIV(utt_names)[]; +extern const ucp_type_table PRIV(utt)[]; +extern const int PRIV(utt_size); + +extern const pcre_uint8 PRIV(OP_lengths)[]; +extern const pcre_uint8 PRIV(default_tables)[]; + +extern const pcre_uint32 PRIV(hspace_list)[]; +extern const pcre_uint32 PRIV(vspace_list)[]; + + +/* Internal shared functions. These are functions that are used by more than +one of the exported public functions. They have to be "external" in the C +sense, but are not part of the PCRE public API. */ + +/* String comparison functions. */ +#if defined COMPILE_PCRE8 + +#define STRCMP_UC_UC(str1, str2) \ + strcmp((char *)(str1), (char *)(str2)) +#define STRCMP_UC_C8(str1, str2) \ + strcmp((char *)(str1), (str2)) +#define STRNCMP_UC_UC(str1, str2, num) \ + strncmp((char *)(str1), (char *)(str2), (num)) +#define STRNCMP_UC_C8(str1, str2, num) \ + strncmp((char *)(str1), (str2), (num)) +#define STRLEN_UC(str) strlen((const char *)str) + +#elif defined COMPILE_PCRE16 || defined COMPILE_PCRE32 + +extern int PRIV(strcmp_uc_uc)(const pcre_uchar *, + const pcre_uchar *); +extern int PRIV(strcmp_uc_c8)(const pcre_uchar *, + const char *); +extern int PRIV(strncmp_uc_uc)(const pcre_uchar *, + const pcre_uchar *, unsigned int num); +extern int PRIV(strncmp_uc_c8)(const pcre_uchar *, + const char *, unsigned int num); +extern unsigned int PRIV(strlen_uc)(const pcre_uchar *str); + +#define STRCMP_UC_UC(str1, str2) \ + PRIV(strcmp_uc_uc)((str1), (str2)) +#define STRCMP_UC_C8(str1, str2) \ + PRIV(strcmp_uc_c8)((str1), (str2)) +#define STRNCMP_UC_UC(str1, str2, num) \ + PRIV(strncmp_uc_uc)((str1), (str2), (num)) +#define STRNCMP_UC_C8(str1, str2, num) \ + PRIV(strncmp_uc_c8)((str1), (str2), (num)) +#define STRLEN_UC(str) PRIV(strlen_uc)(str) + +#endif /* COMPILE_PCRE[8|16|32] */ + +#if defined COMPILE_PCRE8 || defined COMPILE_PCRE16 + +#define STRCMP_UC_UC_TEST(str1, str2) STRCMP_UC_UC(str1, str2) +#define STRCMP_UC_C8_TEST(str1, str2) STRCMP_UC_C8(str1, str2) + +#elif defined COMPILE_PCRE32 + +extern int PRIV(strcmp_uc_uc_utf)(const pcre_uchar *, + const pcre_uchar *); +extern int PRIV(strcmp_uc_c8_utf)(const pcre_uchar *, + const char *); + +#define STRCMP_UC_UC_TEST(str1, str2) \ + (utf ? PRIV(strcmp_uc_uc_utf)((str1), (str2)) : PRIV(strcmp_uc_uc)((str1), (str2))) +#define STRCMP_UC_C8_TEST(str1, str2) \ + (utf ? PRIV(strcmp_uc_c8_utf)((str1), (str2)) : PRIV(strcmp_uc_c8)((str1), (str2))) + +#endif /* COMPILE_PCRE[8|16|32] */ + +extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int); +extern BOOL PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR, + int *, BOOL); +extern unsigned int PRIV(ord2utf)(pcre_uint32, pcre_uchar *); +extern int PRIV(valid_utf)(PCRE_PUCHAR, int, int *); +extern BOOL PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR, + int *, BOOL); +extern BOOL PRIV(xclass)(pcre_uint32, const pcre_uchar *, BOOL); + +#ifdef SUPPORT_JIT +extern void PRIV(jit_compile)(const REAL_PCRE *, + PUBL(extra) *, int); +extern int PRIV(jit_exec)(const PUBL(extra) *, + const pcre_uchar *, int, int, int, int *, int); +extern void PRIV(jit_free)(void *); +extern int PRIV(jit_get_size)(void *); +extern const char* PRIV(jit_get_target)(void); +#endif + +/* Unicode character database (UCD) */ + +typedef struct { + pcre_uint8 script; /* ucp_Arabic, etc. */ + pcre_uint8 chartype; /* ucp_Cc, etc. (general categories) */ + pcre_uint8 gbprop; /* ucp_gbControl, etc. (grapheme break property) */ + pcre_uint8 caseset; /* offset to multichar other cases or zero */ + pcre_int32 other_case; /* offset to other case, or zero if none */ +} ucd_record; + +extern const pcre_uint32 PRIV(ucd_caseless_sets)[]; +extern const ucd_record PRIV(ucd_records)[]; +extern const pcre_uint8 PRIV(ucd_stage1)[]; +extern const pcre_uint16 PRIV(ucd_stage2)[]; +extern const pcre_uint32 PRIV(ucp_gentype)[]; +extern const pcre_uint32 PRIV(ucp_gbtable)[]; +#ifdef SUPPORT_JIT +extern const int PRIV(ucp_typerange)[]; +#endif + +#ifdef SUPPORT_UCP +/* UCD access macros */ + +#define UCD_BLOCK_SIZE 128 +#define GET_UCD(ch) (PRIV(ucd_records) + \ + PRIV(ucd_stage2)[PRIV(ucd_stage1)[(int)(ch) / UCD_BLOCK_SIZE] * \ + UCD_BLOCK_SIZE + (int)(ch) % UCD_BLOCK_SIZE]) + +#define UCD_CHARTYPE(ch) GET_UCD(ch)->chartype +#define UCD_SCRIPT(ch) GET_UCD(ch)->script +#define UCD_CATEGORY(ch) PRIV(ucp_gentype)[UCD_CHARTYPE(ch)] +#define UCD_GRAPHBREAK(ch) GET_UCD(ch)->gbprop +#define UCD_CASESET(ch) GET_UCD(ch)->caseset +#define UCD_OTHERCASE(ch) ((pcre_uint32)((int)ch + (int)(GET_UCD(ch)->other_case))) + +#endif /* SUPPORT_UCP */ + +#endif + +/* End of pcre_internal.h */ diff --git a/include/pcre/pcre_newline.c b/include/pcre/pcre_newline.c new file mode 100644 index 000000000..b8f5a4de1 --- /dev/null +++ b/include/pcre/pcre_newline.c @@ -0,0 +1,210 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + +/* PCRE is a library of functions to support regular expressions whose syntax +and semantics are as close as possible to those of the Perl 5 language. + + Written by Philip Hazel + Copyright (c) 1997-2012 University of Cambridge + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + + +/* This module contains internal functions for testing newlines when more than +one kind of newline is to be recognized. When a newline is found, its length is +returned. In principle, we could implement several newline "types", each +referring to a different set of newline characters. At present, PCRE supports +only NLTYPE_FIXED, which gets handled without these functions, NLTYPE_ANYCRLF, +and NLTYPE_ANY. The full list of Unicode newline characters is taken from +http://unicode.org/unicode/reports/tr18/. */ + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + + + +/************************************************* +* Check for newline at given position * +*************************************************/ + +/* It is guaranteed that the initial value of ptr is less than the end of the +string that is being processed. + +Arguments: + ptr pointer to possible newline + type the newline type + endptr pointer to the end of the string + lenptr where to return the length + utf TRUE if in utf mode + +Returns: TRUE or FALSE +*/ + +BOOL +PRIV(is_newline)(PCRE_PUCHAR ptr, int type, PCRE_PUCHAR endptr, int *lenptr, + BOOL utf) +{ +pcre_uint32 c; +(void)utf; +#ifdef SUPPORT_UTF +if (utf) + { + GETCHAR(c, ptr); + } +else +#endif /* SUPPORT_UTF */ + c = *ptr; + +/* Note that this function is called only for ANY or ANYCRLF. */ + +if (type == NLTYPE_ANYCRLF) switch(c) + { + case CHAR_LF: *lenptr = 1; return TRUE; + case CHAR_CR: *lenptr = (ptr < endptr - 1 && ptr[1] == CHAR_LF)? 2 : 1; + return TRUE; + default: return FALSE; + } + +/* NLTYPE_ANY */ + +else switch(c) + { +#ifdef EBCDIC + case CHAR_NEL: +#endif + case CHAR_LF: + case CHAR_VT: + case CHAR_FF: *lenptr = 1; return TRUE; + + case CHAR_CR: + *lenptr = (ptr < endptr - 1 && ptr[1] == CHAR_LF)? 2 : 1; + return TRUE; + +#ifndef EBCDIC +#ifdef COMPILE_PCRE8 + case CHAR_NEL: *lenptr = utf? 2 : 1; return TRUE; + case 0x2028: /* LS */ + case 0x2029: *lenptr = 3; return TRUE; /* PS */ +#else /* COMPILE_PCRE16 || COMPILE_PCRE32 */ + case CHAR_NEL: + case 0x2028: /* LS */ + case 0x2029: *lenptr = 1; return TRUE; /* PS */ +#endif /* COMPILE_PCRE8 */ +#endif /* Not EBCDIC */ + + default: return FALSE; + } +} + + + +/************************************************* +* Check for newline at previous position * +*************************************************/ + +/* It is guaranteed that the initial value of ptr is greater than the start of +the string that is being processed. + +Arguments: + ptr pointer to possible newline + type the newline type + startptr pointer to the start of the string + lenptr where to return the length + utf TRUE if in utf mode + +Returns: TRUE or FALSE +*/ + +BOOL +PRIV(was_newline)(PCRE_PUCHAR ptr, int type, PCRE_PUCHAR startptr, int *lenptr, + BOOL utf) +{ +pcre_uint32 c; +(void)utf; +ptr--; +#ifdef SUPPORT_UTF +if (utf) + { + BACKCHAR(ptr); + GETCHAR(c, ptr); + } +else +#endif /* SUPPORT_UTF */ + c = *ptr; + +/* Note that this function is called only for ANY or ANYCRLF. */ + +if (type == NLTYPE_ANYCRLF) switch(c) + { + case CHAR_LF: + *lenptr = (ptr > startptr && ptr[-1] == CHAR_CR)? 2 : 1; + return TRUE; + + case CHAR_CR: *lenptr = 1; return TRUE; + default: return FALSE; + } + +/* NLTYPE_ANY */ + +else switch(c) + { + case CHAR_LF: + *lenptr = (ptr > startptr && ptr[-1] == CHAR_CR)? 2 : 1; + return TRUE; + +#ifdef EBCDIC + case CHAR_NEL: +#endif + case CHAR_VT: + case CHAR_FF: + case CHAR_CR: *lenptr = 1; return TRUE; + +#ifndef EBCDIC +#ifdef COMPILE_PCRE8 + case CHAR_NEL: *lenptr = utf? 2 : 1; return TRUE; + case 0x2028: /* LS */ + case 0x2029: *lenptr = 3; return TRUE; /* PS */ +#else /* COMPILE_PCRE16 || COMPILE_PCRE32 */ + case CHAR_NEL: + case 0x2028: /* LS */ + case 0x2029: *lenptr = 1; return TRUE; /* PS */ +#endif /* COMPILE_PCRE8 */ +#endif /* NotEBCDIC */ + + default: return FALSE; + } +} + +/* End of pcre_newline.c */ diff --git a/include/pcre/pcre_string_utils.c b/include/pcre/pcre_string_utils.c new file mode 100644 index 000000000..25eacc850 --- /dev/null +++ b/include/pcre/pcre_string_utils.c @@ -0,0 +1,211 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + +/* PCRE is a library of functions to support regular expressions whose syntax +and semantics are as close as possible to those of the Perl 5 language. + + Written by Philip Hazel + Copyright (c) 1997-2014 University of Cambridge + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + + +/* This module contains internal functions for comparing and finding the length +of strings for different data item sizes. */ + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + +#ifndef COMPILE_PCRE8 + +/************************************************* +* Compare string utilities * +*************************************************/ + +/* The following two functions compares two strings. Basically a strcmp +for non 8 bit characters. + +Arguments: + str1 first string + str2 second string + +Returns: 0 if both string are equal (like strcmp), 1 otherwise +*/ + +int +PRIV(strcmp_uc_uc)(const pcre_uchar *str1, const pcre_uchar *str2) +{ +pcre_uchar c1; +pcre_uchar c2; + +while (*str1 != '\0' || *str2 != '\0') + { + c1 = *str1++; + c2 = *str2++; + if (c1 != c2) + return ((c1 > c2) << 1) - 1; + } +/* Both length and characters must be equal. */ +return 0; +} + +#ifdef COMPILE_PCRE32 + +int +PRIV(strcmp_uc_uc_utf)(const pcre_uchar *str1, const pcre_uchar *str2) +{ +pcre_uchar c1; +pcre_uchar c2; + +while (*str1 != '\0' || *str2 != '\0') + { + c1 = UCHAR21INC(str1); + c2 = UCHAR21INC(str2); + if (c1 != c2) + return ((c1 > c2) << 1) - 1; + } +/* Both length and characters must be equal. */ +return 0; +} + +#endif /* COMPILE_PCRE32 */ + +int +PRIV(strcmp_uc_c8)(const pcre_uchar *str1, const char *str2) +{ +const pcre_uint8 *ustr2 = (pcre_uint8 *)str2; +pcre_uchar c1; +pcre_uchar c2; + +while (*str1 != '\0' || *ustr2 != '\0') + { + c1 = *str1++; + c2 = (pcre_uchar)*ustr2++; + if (c1 != c2) + return ((c1 > c2) << 1) - 1; + } +/* Both length and characters must be equal. */ +return 0; +} + +#ifdef COMPILE_PCRE32 + +int +PRIV(strcmp_uc_c8_utf)(const pcre_uchar *str1, const char *str2) +{ +const pcre_uint8 *ustr2 = (pcre_uint8 *)str2; +pcre_uchar c1; +pcre_uchar c2; + +while (*str1 != '\0' || *ustr2 != '\0') + { + c1 = UCHAR21INC(str1); + c2 = (pcre_uchar)*ustr2++; + if (c1 != c2) + return ((c1 > c2) << 1) - 1; + } +/* Both length and characters must be equal. */ +return 0; +} + +#endif /* COMPILE_PCRE32 */ + +/* The following two functions compares two, fixed length +strings. Basically an strncmp for non 8 bit characters. + +Arguments: + str1 first string + str2 second string + num size of the string + +Returns: 0 if both string are equal (like strcmp), 1 otherwise +*/ + +int +PRIV(strncmp_uc_uc)(const pcre_uchar *str1, const pcre_uchar *str2, unsigned int num) +{ +pcre_uchar c1; +pcre_uchar c2; + +while (num-- > 0) + { + c1 = *str1++; + c2 = *str2++; + if (c1 != c2) + return ((c1 > c2) << 1) - 1; + } +/* Both length and characters must be equal. */ +return 0; +} + +int +PRIV(strncmp_uc_c8)(const pcre_uchar *str1, const char *str2, unsigned int num) +{ +const pcre_uint8 *ustr2 = (pcre_uint8 *)str2; +pcre_uchar c1; +pcre_uchar c2; + +while (num-- > 0) + { + c1 = *str1++; + c2 = (pcre_uchar)*ustr2++; + if (c1 != c2) + return ((c1 > c2) << 1) - 1; + } +/* Both length and characters must be equal. */ +return 0; +} + +/* The following function returns with the length of +a zero terminated string. Basically an strlen for non 8 bit characters. + +Arguments: + str string + +Returns: length of the string +*/ + +unsigned int +PRIV(strlen_uc)(const pcre_uchar *str) +{ +unsigned int len = 0; +while (*str++ != 0) + len++; +return len; +} + +#endif /* !COMPILE_PCRE8 */ + +/* End of pcre_string_utils.c */ diff --git a/include/pcre/pcre_tables.c b/include/pcre/pcre_tables.c new file mode 100644 index 000000000..4960af57c --- /dev/null +++ b/include/pcre/pcre_tables.c @@ -0,0 +1,727 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + +/* PCRE is a library of functions to support regular expressions whose syntax +and semantics are as close as possible to those of the Perl 5 language. + + Written by Philip Hazel + Copyright (c) 1997-2012 University of Cambridge + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + +#ifndef PCRE_INCLUDED + +/* This module contains some fixed tables that are used by more than one of the +PCRE code modules. The tables are also #included by the pcretest program, which +uses macros to change their names from _pcre_xxx to xxxx, thereby avoiding name +clashes with the library. */ + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + +#endif /* PCRE_INCLUDED */ + +/* Table of sizes for the fixed-length opcodes. It's defined in a macro so that +the definition is next to the definition of the opcodes in pcre_internal.h. */ + +const pcre_uint8 PRIV(OP_lengths)[] = { OP_LENGTHS }; + +/* Tables of horizontal and vertical whitespace characters, suitable for +adding to classes. */ + +const pcre_uint32 PRIV(hspace_list)[] = { HSPACE_LIST }; +const pcre_uint32 PRIV(vspace_list)[] = { VSPACE_LIST }; + + + +/************************************************* +* Tables for UTF-8 support * +*************************************************/ + +/* These are the breakpoints for different numbers of bytes in a UTF-8 +character. */ + +#if (defined SUPPORT_UTF && defined COMPILE_PCRE8) \ + || (defined PCRE_INCLUDED && (defined SUPPORT_PCRE16 || defined SUPPORT_PCRE32)) + +/* These tables are also required by pcretest in 16- or 32-bit mode. */ + +const int PRIV(utf8_table1)[] = + { 0x7f, 0x7ff, 0xffff, 0x1fffff, 0x3ffffff, 0x7fffffff}; + +const int PRIV(utf8_table1_size) = sizeof(PRIV(utf8_table1)) / sizeof(int); + +/* These are the indicator bits and the mask for the data bits to set in the +first byte of a character, indexed by the number of additional bytes. */ + +const int PRIV(utf8_table2)[] = { 0, 0xc0, 0xe0, 0xf0, 0xf8, 0xfc}; +const int PRIV(utf8_table3)[] = { 0xff, 0x1f, 0x0f, 0x07, 0x03, 0x01}; + +/* Table of the number of extra bytes, indexed by the first byte masked with +0x3f. The highest number for a valid UTF-8 first byte is in fact 0x3d. */ + +const pcre_uint8 PRIV(utf8_table4)[] = { + 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, + 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, + 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, + 3,3,3,3,3,3,3,3,4,4,4,4,5,5,5,5 }; + +#endif /* (SUPPORT_UTF && COMPILE_PCRE8) || (PCRE_INCLUDED && SUPPORT_PCRE[16|32])*/ + +#ifdef SUPPORT_UTF + +/* Table to translate from particular type value to the general value. */ + +const pcre_uint32 PRIV(ucp_gentype)[] = { + ucp_C, ucp_C, ucp_C, ucp_C, ucp_C, /* Cc, Cf, Cn, Co, Cs */ + ucp_L, ucp_L, ucp_L, ucp_L, ucp_L, /* Ll, Lu, Lm, Lo, Lt */ + ucp_M, ucp_M, ucp_M, /* Mc, Me, Mn */ + ucp_N, ucp_N, ucp_N, /* Nd, Nl, No */ + ucp_P, ucp_P, ucp_P, ucp_P, ucp_P, /* Pc, Pd, Pe, Pf, Pi */ + ucp_P, ucp_P, /* Ps, Po */ + ucp_S, ucp_S, ucp_S, ucp_S, /* Sc, Sk, Sm, So */ + ucp_Z, ucp_Z, ucp_Z /* Zl, Zp, Zs */ +}; + +/* This table encodes the rules for finding the end of an extended grapheme +cluster. Every code point has a grapheme break property which is one of the +ucp_gbXX values defined in ucp.h. The 2-dimensional table is indexed by the +properties of two adjacent code points. The left property selects a word from +the table, and the right property selects a bit from that word like this: + + ucp_gbtable[left-property] & (1 << right-property) + +The value is non-zero if a grapheme break is NOT permitted between the relevant +two code points. The breaking rules are as follows: + +1. Break at the start and end of text (pretty obviously). + +2. Do not break between a CR and LF; otherwise, break before and after + controls. + +3. Do not break Hangul syllable sequences, the rules for which are: + + L may be followed by L, V, LV or LVT + LV or V may be followed by V or T + LVT or T may be followed by T + +4. Do not break before extending characters. + +The next two rules are only for extended grapheme clusters (but that's what we +are implementing). + +5. Do not break before SpacingMarks. + +6. Do not break after Prepend characters. + +7. Otherwise, break everywhere. +*/ + +const pcre_uint32 PRIV(ucp_gbtable[]) = { + (1< 255 and/or Unicode properties. + +Arguments: + c the character + data points to the flag byte of the XCLASS data + +Returns: TRUE if character matches, else FALSE +*/ + +BOOL +PRIV(xclass)(pcre_uint32 c, const pcre_uchar *data, BOOL utf) +{ +pcre_uchar t; +BOOL negated = (*data & XCL_NOT) != 0; + +(void)utf; +#ifdef COMPILE_PCRE8 +/* In 8 bit mode, this must always be TRUE. Help the compiler to know that. */ +utf = TRUE; +#endif + +/* Character values < 256 are matched against a bitmap, if one is present. If +not, we still carry on, because there may be ranges that start below 256 in the +additional data. */ + +if (c < 256) + { + if ((*data & XCL_HASPROP) == 0) + { + if ((*data & XCL_MAP) == 0) return negated; + return (((pcre_uint8 *)(data + 1))[c/8] & (1 << (c&7))) != 0; + } + if ((*data & XCL_MAP) != 0 && + (((pcre_uint8 *)(data + 1))[c/8] & (1 << (c&7))) != 0) + return !negated; /* char found */ + } + +/* First skip the bit map if present. Then match against the list of Unicode +properties or large chars or ranges that end with a large char. We won't ever +encounter XCL_PROP or XCL_NOTPROP when UCP support is not compiled. */ + +if ((*data++ & XCL_MAP) != 0) data += 32 / sizeof(pcre_uchar); + +while ((t = *data++) != XCL_END) + { + pcre_uint32 x, y; + if (t == XCL_SINGLE) + { +#ifdef SUPPORT_UTF + if (utf) + { + GETCHARINC(x, data); /* macro generates multiple statements */ + } + else +#endif + x = *data++; + if (c == x) return !negated; + } + else if (t == XCL_RANGE) + { +#ifdef SUPPORT_UTF + if (utf) + { + GETCHARINC(x, data); /* macro generates multiple statements */ + GETCHARINC(y, data); /* macro generates multiple statements */ + } + else +#endif + { + x = *data++; + y = *data++; + } + if (c >= x && c <= y) return !negated; + } + +#ifdef SUPPORT_UCP + else /* XCL_PROP & XCL_NOTPROP */ + { + const ucd_record *prop = GET_UCD(c); + BOOL isprop = t == XCL_PROP; + + switch(*data) + { + case PT_ANY: + if (isprop) return !negated; + break; + + case PT_LAMP: + if ((prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || + prop->chartype == ucp_Lt) == isprop) return !negated; + break; + + case PT_GC: + if ((data[1] == PRIV(ucp_gentype)[prop->chartype]) == isprop) + return !negated; + break; + + case PT_PC: + if ((data[1] == prop->chartype) == isprop) return !negated; + break; + + case PT_SC: + if ((data[1] == prop->script) == isprop) return !negated; + break; + + case PT_ALNUM: + if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N) == isprop) + return !negated; + break; + + /* Perl space used to exclude VT, but from Perl 5.18 it is included, + which means that Perl space and POSIX space are now identical. PCRE + was changed at release 8.34. */ + + case PT_SPACE: /* Perl space */ + case PT_PXSPACE: /* POSIX space */ + switch(c) + { + HSPACE_CASES: + VSPACE_CASES: + if (isprop) return !negated; + break; + + default: + if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z) == isprop) + return !negated; + break; + } + break; + + case PT_WORD: + if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N || c == CHAR_UNDERSCORE) + == isprop) + return !negated; + break; + + case PT_UCNC: + if (c < 0xa0) + { + if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT || + c == CHAR_GRAVE_ACCENT) == isprop) + return !negated; + } + else + { + if ((c < 0xd800 || c > 0xdfff) == isprop) + return !negated; + } + break; + + /* The following three properties can occur only in an XCLASS, as there + is no \p or \P coding for them. */ + + /* Graphic character. Implement this as not Z (space or separator) and + not C (other), except for Cf (format) with a few exceptions. This seems + to be what Perl does. The exceptional characters are: + + U+061C Arabic Letter Mark + U+180E Mongolian Vowel Separator + U+2066 - U+2069 Various "isolate"s + */ + + case PT_PXGRAPH: + if ((PRIV(ucp_gentype)[prop->chartype] != ucp_Z && + (PRIV(ucp_gentype)[prop->chartype] != ucp_C || + (prop->chartype == ucp_Cf && + c != 0x061c && c != 0x180e && (c < 0x2066 || c > 0x2069)) + )) == isprop) + return !negated; + break; + + /* Printable character: same as graphic, with the addition of Zs, i.e. + not Zl and not Zp, and U+180E. */ + + case PT_PXPRINT: + if ((prop->chartype != ucp_Zl && + prop->chartype != ucp_Zp && + (PRIV(ucp_gentype)[prop->chartype] != ucp_C || + (prop->chartype == ucp_Cf && + c != 0x061c && (c < 0x2066 || c > 0x2069)) + )) == isprop) + return !negated; + break; + + /* Punctuation: all Unicode punctuation, plus ASCII characters that + Unicode treats as symbols rather than punctuation, for Perl + compatibility (these are $+<=>^`|~). */ + + case PT_PXPUNCT: + if ((PRIV(ucp_gentype)[prop->chartype] == ucp_P || + (c < 128 && PRIV(ucp_gentype)[prop->chartype] == ucp_S)) == isprop) + return !negated; + break; + + /* This should never occur, but compilers may mutter if there is no + default. */ + + default: + return FALSE; + } + + data += 2; + } +#endif /* SUPPORT_UCP */ + } + +return negated; /* char did not match */ +} + +/* End of pcre_xclass.c */ diff --git a/include/pcre/ucp.h b/include/pcre/ucp.h new file mode 100644 index 000000000..2fa00296e --- /dev/null +++ b/include/pcre/ucp.h @@ -0,0 +1,224 @@ +/************************************************* +* Unicode Property Table handler * +*************************************************/ + +#ifndef _UCP_H +#define _UCP_H + +/* This file contains definitions of the property values that are returned by +the UCD access macros. New values that are added for new releases of Unicode +should always be at the end of each enum, for backwards compatibility. + +IMPORTANT: Note also that the specific numeric values of the enums have to be +the same as the values that are generated by the maint/MultiStage2.py script, +where the equivalent property descriptive names are listed in vectors. + +ALSO: The specific values of the first two enums are assumed for the table +called catposstab in pcre_compile.c. */ + +/* These are the general character categories. */ + +enum { + ucp_C, /* Other */ + ucp_L, /* Letter */ + ucp_M, /* Mark */ + ucp_N, /* Number */ + ucp_P, /* Punctuation */ + ucp_S, /* Symbol */ + ucp_Z /* Separator */ +}; + +/* These are the particular character categories. */ + +enum { + ucp_Cc, /* Control */ + ucp_Cf, /* Format */ + ucp_Cn, /* Unassigned */ + ucp_Co, /* Private use */ + ucp_Cs, /* Surrogate */ + ucp_Ll, /* Lower case letter */ + ucp_Lm, /* Modifier letter */ + ucp_Lo, /* Other letter */ + ucp_Lt, /* Title case letter */ + ucp_Lu, /* Upper case letter */ + ucp_Mc, /* Spacing mark */ + ucp_Me, /* Enclosing mark */ + ucp_Mn, /* Non-spacing mark */ + ucp_Nd, /* Decimal number */ + ucp_Nl, /* Letter number */ + ucp_No, /* Other number */ + ucp_Pc, /* Connector punctuation */ + ucp_Pd, /* Dash punctuation */ + ucp_Pe, /* Close punctuation */ + ucp_Pf, /* Final punctuation */ + ucp_Pi, /* Initial punctuation */ + ucp_Po, /* Other punctuation */ + ucp_Ps, /* Open punctuation */ + ucp_Sc, /* Currency symbol */ + ucp_Sk, /* Modifier symbol */ + ucp_Sm, /* Mathematical symbol */ + ucp_So, /* Other symbol */ + ucp_Zl, /* Line separator */ + ucp_Zp, /* Paragraph separator */ + ucp_Zs /* Space separator */ +}; + +/* These are grapheme break properties. Note that the code for processing them +assumes that the values are less than 16. If more values are added that take +the number to 16 or more, the code will have to be rewritten. */ + +enum { + ucp_gbCR, /* 0 */ + ucp_gbLF, /* 1 */ + ucp_gbControl, /* 2 */ + ucp_gbExtend, /* 3 */ + ucp_gbPrepend, /* 4 */ + ucp_gbSpacingMark, /* 5 */ + ucp_gbL, /* 6 Hangul syllable type L */ + ucp_gbV, /* 7 Hangul syllable type V */ + ucp_gbT, /* 8 Hangul syllable type T */ + ucp_gbLV, /* 9 Hangul syllable type LV */ + ucp_gbLVT, /* 10 Hangul syllable type LVT */ + ucp_gbRegionalIndicator, /* 11 */ + ucp_gbOther /* 12 */ +}; + +/* These are the script identifications. */ + +enum { + ucp_Arabic, + ucp_Armenian, + ucp_Bengali, + ucp_Bopomofo, + ucp_Braille, + ucp_Buginese, + ucp_Buhid, + ucp_Canadian_Aboriginal, + ucp_Cherokee, + ucp_Common, + ucp_Coptic, + ucp_Cypriot, + ucp_Cyrillic, + ucp_Deseret, + ucp_Devanagari, + ucp_Ethiopic, + ucp_Georgian, + ucp_Glagolitic, + ucp_Gothic, + ucp_Greek, + ucp_Gujarati, + ucp_Gurmukhi, + ucp_Han, + ucp_Hangul, + ucp_Hanunoo, + ucp_Hebrew, + ucp_Hiragana, + ucp_Inherited, + ucp_Kannada, + ucp_Katakana, + ucp_Kharoshthi, + ucp_Khmer, + ucp_Lao, + ucp_Latin, + ucp_Limbu, + ucp_Linear_B, + ucp_Malayalam, + ucp_Mongolian, + ucp_Myanmar, + ucp_New_Tai_Lue, + ucp_Ogham, + ucp_Old_Italic, + ucp_Old_Persian, + ucp_Oriya, + ucp_Osmanya, + ucp_Runic, + ucp_Shavian, + ucp_Sinhala, + ucp_Syloti_Nagri, + ucp_Syriac, + ucp_Tagalog, + ucp_Tagbanwa, + ucp_Tai_Le, + ucp_Tamil, + ucp_Telugu, + ucp_Thaana, + ucp_Thai, + ucp_Tibetan, + ucp_Tifinagh, + ucp_Ugaritic, + ucp_Yi, + /* New for Unicode 5.0: */ + ucp_Balinese, + ucp_Cuneiform, + ucp_Nko, + ucp_Phags_Pa, + ucp_Phoenician, + /* New for Unicode 5.1: */ + ucp_Carian, + ucp_Cham, + ucp_Kayah_Li, + ucp_Lepcha, + ucp_Lycian, + ucp_Lydian, + ucp_Ol_Chiki, + ucp_Rejang, + ucp_Saurashtra, + ucp_Sundanese, + ucp_Vai, + /* New for Unicode 5.2: */ + ucp_Avestan, + ucp_Bamum, + ucp_Egyptian_Hieroglyphs, + ucp_Imperial_Aramaic, + ucp_Inscriptional_Pahlavi, + ucp_Inscriptional_Parthian, + ucp_Javanese, + ucp_Kaithi, + ucp_Lisu, + ucp_Meetei_Mayek, + ucp_Old_South_Arabian, + ucp_Old_Turkic, + ucp_Samaritan, + ucp_Tai_Tham, + ucp_Tai_Viet, + /* New for Unicode 6.0.0: */ + ucp_Batak, + ucp_Brahmi, + ucp_Mandaic, + /* New for Unicode 6.1.0: */ + ucp_Chakma, + ucp_Meroitic_Cursive, + ucp_Meroitic_Hieroglyphs, + ucp_Miao, + ucp_Sharada, + ucp_Sora_Sompeng, + ucp_Takri, + /* New for Unicode 7.0.0: */ + ucp_Bassa_Vah, + ucp_Caucasian_Albanian, + ucp_Duployan, + ucp_Elbasan, + ucp_Grantha, + ucp_Khojki, + ucp_Khudawadi, + ucp_Linear_A, + ucp_Mahajani, + ucp_Manichaean, + ucp_Mende_Kikakui, + ucp_Modi, + ucp_Mro, + ucp_Nabataean, + ucp_Old_North_Arabian, + ucp_Old_Permic, + ucp_Pahawh_Hmong, + ucp_Palmyrene, + ucp_Psalter_Pahlavi, + ucp_Pau_Cin_Hau, + ucp_Siddham, + ucp_Tirhuta, + ucp_Warang_Citi +}; + +#endif + +/* End of ucp.h */ diff --git a/src/alloc.c b/src/alloc.c index 362232593..5837fc56b 100644 --- a/src/alloc.c +++ b/src/alloc.c @@ -111,10 +111,14 @@ void *hl_gc_alloc( int size ) { return malloc(size); } -char *hl_gc_alloc_noptr( int size ) { +void *hl_gc_alloc_noptr( int size ) { return (char*)malloc(size); } +void *hl_gc_alloc_finalizer( int size ) { + return malloc(size); +} + vdynamic *hl_alloc_dynamic( hl_type *t ) { vdynamic *d = (vdynamic*)hl_gc_alloc(sizeof(vdynamic)); d->t = t; diff --git a/src/hl.h b/src/hl.h index 4afa9155a..a19144d6e 100644 --- a/src/hl.h +++ b/src/hl.h @@ -56,8 +56,16 @@ #if defined(_MSC_VER) # define HL_VCC -// remove deprecated C API usage warnings -# pragma warning( disable : 4996 ) +# pragma warning(disable:4996) // remove deprecated C API usage warnings +# pragma warning(disable:4055) // void* - to - function cast +# pragma warning(disable:4152) // void* - to - function cast +# pragma warning(disable:4201) // anonymous struct +# pragma warning(disable:4127) // while( true ) +# pragma warning(disable:4710) // inline disabled +# pragma warning(disable:4711) // inline activated +# pragma warning(disable:4255) // windows include +# pragma warning(disable:4820) // windows include +# pragma warning(disable:4668) // windows include #endif #if defined(HL_VCC) || defined(HL_MINGW) || defined(HL_CYGWIN) @@ -128,6 +136,7 @@ typedef wchar_t uchar; # define utod(s,end) wcstod(s,end) # define utoi(s,end) wcstol(s,end,10) # define ucmp(a,b) wcscmp(a,b) +# define strtou(out,size,str) mbstowcs(out,str,size) #else typedef unsigned short uchar; # undef USTR @@ -404,7 +413,8 @@ vclosure *hl_alloc_closure_wrapper( hl_type *t, void *fvalue, void *ptr ); // ----------------------- ALLOC -------------------------------------------------- void *hl_gc_alloc( int size ); -char *hl_gc_alloc_noptr( int size ); +void *hl_gc_alloc_noptr( int size ); +void *hl_gc_alloc_finalizer( int size ); void hl_alloc_init( hl_alloc *a ); void *hl_malloc( hl_alloc *a, int size ); @@ -422,6 +432,7 @@ hl_buffer *hl_alloc_buffer(); void hl_buffer_val( hl_buffer *b, vdynamic *v ); void hl_buffer_char( hl_buffer *b, uchar c ); void hl_buffer_str( hl_buffer *b, const uchar *str ); +void hl_buffer_cstr( hl_buffer *b, const char *str ); void hl_buffer_str_sub( hl_buffer *b, const uchar *str, int len ); int hl_buffer_length( hl_buffer *b ); uchar *hl_buffer_content( hl_buffer *b, int *len ); diff --git a/src/hlc.h b/src/hlc.h index 96f1455a6..37d25a106 100644 --- a/src/hlc.h +++ b/src/hlc.h @@ -25,15 +25,19 @@ #include #include -#pragma warning(disable:4101) -#pragma warning(disable:4700) - #ifdef HL_64 # define PAD_64_VAL ,0 #else # define PAD_64_VAL #endif +#ifdef HL_VCC +# pragma warning(disable:4702) // unreachable code +# pragma warning(disable:4100) // unreferenced param +# pragma warning(disable:4101) // unreferenced local var +# pragma warning(disable:4723) // potential divide by 0 +#endif + #undef CONST static void hl_null_access() { diff --git a/src/objdata.c b/src/objdata.c deleted file mode 100644 index db53b1081..000000000 --- a/src/objdata.c +++ /dev/null @@ -1,415 +0,0 @@ -#include "hl.h" -#include - -static void hl_lookup_insert( hl_field_lookup *l, int size, int hash, hl_type *t, int index ) { - int min = 0; - int max = size; - int pos; - while( min < max ) { - int mid = (min + max) >> 1; - int h = l[mid].hashed_name; - if( h < hash ) min = mid + 1; else max = mid; - } - pos = (min + max) >> 1; - memcpy(l + pos + 1, l + pos, (size - pos) * sizeof(hl_field_lookup)); - l[pos].field_index = index; - l[pos].hashed_name = hash; - l[pos].t = t; -} - -static hl_field_lookup *hl_lookup_find( hl_field_lookup *l, int size, int hash ) { - int min = 0; - int max = size; - while( min < max ) { - int mid = (min + max) >> 1; - int h = l[mid].hashed_name; - if( h < hash ) min = mid + 1; else if( h > hash ) max = mid; else return l + mid; - } - return NULL; -} - -static int hl_lookup_find_index( hl_field_lookup *l, int size, int hash ) { - int min = 0; - int max = size; - while( min < max ) { - int mid = (min + max) >> 1; - int h = l[mid].hashed_name; - if( h < hash ) min = mid + 1; else if( h > hash ) max = mid; else return mid; - } - return (min + max) >> 1; -} - -static int hl_cache_count = 0; -static int hl_cache_size = 0; -static hl_field_lookup *hl_cache = NULL; - -int hl_hash( vbytes *b ) { - return hl_hash_gen((uchar*)b,true); -} - -int hl_hash_gen( const uchar *name, bool cache_name ) { - int h = 0; - const uchar *oname = name; - while( *name ) { - h = 223 * h + (unsigned)*name; - name++; - } - h %= 0x1FFFFF7B; - if( cache_name ) { - hl_field_lookup *l = hl_lookup_find(hl_cache, hl_cache_count, h); - if( l == NULL ) { - if( hl_cache_size == hl_cache_count ) { - // resize - int newsize = hl_cache_size ? (hl_cache_size * 3) >> 1 : 16; - hl_field_lookup *cache = (hl_field_lookup*)hl_gc_alloc(sizeof(hl_field_lookup) * newsize); - memcpy(cache,hl_cache,sizeof(hl_field_lookup) * hl_cache_count); - free(hl_cache); - hl_cache = cache; - hl_cache_size = newsize; - } - hl_lookup_insert(hl_cache,hl_cache_count++,h,(hl_type*)ustrdup(oname),0); - } - } - return h; -} - -const uchar *hl_field_name( int hash ) { - hl_field_lookup *l = hl_lookup_find(hl_cache, hl_cache_count, hash); - return l ? (uchar*)l->t : USTR("???"); -} - -void hl_cache_free() { - int i; - for(i=0;iobj; - hl_module_context *m = o->m; - hl_alloc *alloc = &m->alloc; - hl_runtime_obj *p = NULL, *t; - int i, size, start; - if( o->rt ) return o->rt; - if( o->super ) p = hl_get_obj_rt(o->super); - t = (hl_runtime_obj*)hl_malloc(alloc,sizeof(hl_runtime_obj)); - t->t = ot; - t->nfields = o->nfields + (p ? p->nfields : 0); - t->nproto = p ? p->nproto : 0; - t->nlookup = o->nfields + o->nproto; - t->lookup = (hl_field_lookup*)hl_malloc(alloc,sizeof(hl_field_lookup) * t->nlookup); - t->fields_indexes = (int*)hl_malloc(alloc,sizeof(int)*t->nfields); - t->toStringFun = NULL; - t->compareFun = NULL; - t->castFun = NULL; - t->parent = p; - - // fields indexes - start = 0; - if( p ) { - start = p->nfields; - memcpy(t->fields_indexes, p->fields_indexes, sizeof(int)*p->nfields); - } - size = p ? p->size : HL_WSIZE; // hl_type* - for(i=0;infields;i++) { - hl_type *ft = o->fields[i].t; - size += hl_pad_size(size,ft); - t->fields_indexes[i+start] = size; - hl_lookup_insert(t->lookup,i,o->fields[i].hashed_name,o->fields[i].t,size); - size += hl_type_size(ft); - } - t->size = size; - t->methods = NULL; - o->rt = t; - ot->vobj_proto = NULL; - - // fields lookup - size = 0; - for(i=0;inproto;i++) { - hl_obj_proto *p = o->proto + i; - if( p->pindex >= t->nproto ) t->nproto = p->pindex + 1; - hl_lookup_insert(t->lookup,i + o->nfields,p->hashed_name,m->functions_types[p->findex],-(i+1)); - } - return t; -} - -/** - Fill class prototype with method pointers. - Requires method table to be finalized -**/ -hl_runtime_obj *hl_get_obj_proto( hl_type *ot ) { - hl_type_obj *o = ot->obj; - hl_module_context *m = o->m; - hl_alloc *alloc = &m->alloc; - hl_runtime_obj *p = NULL, *t = hl_get_obj_rt(ot); - hl_field_lookup *strField, *cmpField, *castField; - int i; - if( ot->vobj_proto ) return t; - if( o->super ) p = hl_get_obj_proto(o->super); - - strField = hl_lookup_find(t->lookup,t->nlookup,hl_hash_gen(USTR("__string"),false)); - cmpField = hl_lookup_find(t->lookup,t->nlookup,hl_hash_gen(USTR("__compare"),false)); - castField = hl_lookup_find(t->lookup,t->nlookup,hl_hash_gen(USTR("__cast"),false)); - t->toStringFun = strField ? m->functions_ptrs[o->proto[-(strField->field_index+1)].findex] : (p ? p->toStringFun : NULL); - t->compareFun = cmpField ? m->functions_ptrs[o->proto[-(cmpField->field_index+1)].findex] : (p ? p->compareFun : NULL); - t->castFun = castField ? m->functions_ptrs[o->proto[-(castField->field_index+1)].findex] : (p ? p->castFun : NULL); - - if( t->nproto ) { - void **fptr = (void**)hl_malloc(alloc, sizeof(void*) * t->nproto); - ot->vobj_proto = fptr; - if( p ) - memcpy(fptr, p->t->vobj_proto, p->nproto * sizeof(void*)); - for(i=0;inproto;i++) { - hl_obj_proto *p = o->proto + i; - if( p->pindex >= 0 ) fptr[p->pindex] = m->functions_ptrs[p->findex]; - } - } - - t->methods = (void**)hl_malloc(alloc, sizeof(void*) * o->nproto); - for(i=0;inproto;i++) { - hl_obj_proto *p = o->proto + i; - t->methods[i] = m->functions_ptrs[p->findex]; - } - - return t; -} - -/** - Allocate a virtual fields mapping to a given value. -**/ -vvirtual *hl_to_virtual( hl_type *vt, vdynamic *obj ) { - vvirtual *v; - if( obj == NULL ) return NULL; - switch( obj->t->kind ) { - case HOBJ: - { - int i; - hl_runtime_obj *rt = obj->t->obj->rt; - v = (vvirtual*)hl_gc_alloc(sizeof(vvirtual)); - v->t = vt; - v->fields_data = (char*)obj; - v->indexes = (int*)hl_gc_alloc(sizeof(int)*vt->virt->nfields); - v->value = obj; - v->next = NULL; - for(i=0;ivirt->nfields;i++) { - hl_runtime_obj *rtt = rt; - hl_field_lookup *f = NULL; - while( rtt ) { - f = hl_lookup_find(rtt->lookup,rtt->nlookup,vt->virt->fields[i].hashed_name); - if( f != NULL ) break; - rtt = rtt->parent; - } - v->indexes[i] = f == NULL || f->field_index <= 0 || (f->t != vt->virt->fields[i].t) ? 0 : f->field_index; - } - } - break; - case HDYNOBJ: - { - int i; - vdynobj *o = (vdynobj*)obj; - v = o->virtuals; - while( v ) { - if( v->t->virt == vt->virt ) - return v; - v = v->next; - } - // allocate a new virtual mapping - v = (vvirtual*)hl_gc_alloc(sizeof(vvirtual)); - v->t = vt; - v->fields_data = o->fields_data - sizeof(void*); - v->indexes = hl_gc_alloc(sizeof(int)*vt->virt->nfields); - v->value = obj; - for(i=0;ivirt->nfields;i++) { - hl_field_lookup *f = hl_lookup_find(&o->dproto->fields,o->nfields,vt->virt->fields[i].hashed_name); - v->indexes[i] = f == NULL || f->t != vt->virt->fields[i].t ? 0 : f->field_index + sizeof(void*); - } - // add it to the list - v->next = o->virtuals; - o->virtuals = v; - } - break; - default: - hl_fatal_fmt("Don't know how to virtual %d",obj->t->kind); - } - return v; -} - -/* -void *hl_fetch_method( vvirtual *v, int fid ) { - hl_obj_field *f = v->t->virt->fields + fid; - switch( v->value->t->kind ) { - case HOBJ: - { - hl_type_obj *o = ((vobj*)v->value)->t->obj; - hl_module_context *m = o->m; - hl_runtime_obj *rt = o->rt; - while( rt ) { - hl_field_lookup *found = hl_lookup_find(rt->lookup,rt->nlookup,f->hashed_name); - if( found ) { - int fid; - if( found->field_index > 0 ) return NULL; - fid = rt->obj->proto[-found->field_index].findex; - return hl_alloc_closure_ptr(m->functions_types[fid], m->functions_ptrs[fid], o); - } - rt = rt->parent; - } - } - break; - } - return NULL; -}*/ - -#define B2(t1,t2) ((t1) + ((t2) * HLAST)) -#define fetch_i(data,src,dst) src == dst ? *(int*)(data) : hl_dyn_casti(data,src,dst) -#define fetch_p(data,src,dst) src == dst ? *(void**)(data) : hl_dyn_castp(data,src,dst) - -static hl_field_lookup *hl_dyn_alloc_field( vdynobj *o, int hfield, hl_type *t ) { - int pad = hl_pad_size(o->dataSize, t); - int size = hl_type_size(t); - int index; - char *newData = (char*)hl_gc_alloc(o->dataSize + pad + size); - vdynobj_proto *proto = (vdynobj_proto*)hl_gc_alloc(sizeof(vdynobj_proto) + sizeof(hl_field_lookup) * (o->nfields + 1 - 1)); - int field_pos = hl_lookup_find_index(&o->dproto->fields, o->nfields, hfield); - hl_field_lookup *f; - // update data - memcpy(newData,o->fields_data,o->dataSize); - o->fields_data = newData; - o->dataSize += pad; - index = o->dataSize; - o->dataSize += size; - // update field table - proto->t = o->dproto->t; - memcpy(&proto->fields,&o->dproto->fields,field_pos * sizeof(hl_field_lookup)); - f = (&proto->fields) + field_pos; - f->t = t; - f->hashed_name = hfield; - f->field_index = index; - memcpy(&proto->fields + (field_pos + 1),&o->dproto->fields + field_pos, (o->nfields - field_pos) * sizeof(hl_field_lookup)); - o->nfields++; - o->dproto = proto; - return f; -} - -int hl_dyn_geti( vdynamic *d, int hfield, hl_type *t ) { - switch( d->t->kind ) { - case HDYNOBJ: - { - vdynobj *o = (vdynobj*)d; - hl_field_lookup *f = hl_lookup_find(&o->dproto->fields,o->nfields,hfield); - if( f == NULL ) return 0; - return fetch_i(o->fields_data + f->field_index,f->t, t); - } - break; - case HOBJ: - { - hl_runtime_obj *rt = d->t->obj->rt; - hl_field_lookup *f = NULL; - do { - f = hl_lookup_find(rt->lookup,rt->nlookup,hfield); - if( f != NULL ) break; - rt = rt->parent; - } while( rt ); - if( f == NULL ) return 0; - return fetch_i((char*)d + f->field_index,f->t,t); - } - break; - case HVIRTUAL: - return hl_dyn_geti(((vvirtual*)d)->value, hfield, t); - default: - hl_error("Invalid field access"); - break; - } - return 0; -} - -void *hl_dyn_getp( vdynamic *d, int hfield, hl_type *t ) { - switch( d->t->kind ) { - case HDYNOBJ: - { - vdynobj *o = (vdynobj*)d; - hl_field_lookup *f = hl_lookup_find(&o->dproto->fields,o->nfields,hfield); - if( f == NULL ) return 0; - return fetch_p(o->fields_data + f->field_index,f->t,t); - } - break; - case HOBJ: - { - hl_runtime_obj *rt = d->t->obj->rt; - hl_field_lookup *f = NULL; - do { - f = hl_lookup_find(rt->lookup,rt->nlookup,hfield); - if( f != NULL ) break; - rt = rt->parent; - } while( rt ); - if( f == NULL ) return NULL; - if( f->field_index < 0 ) { - vclosure *c = hl_alloc_closure_ptr(f->t,rt->methods[-f->field_index-1],d); - return hl_dyn_castp(&c,c->t,t); - } - return fetch_p((char*)d + f->field_index,f->t,t); - } - break; - case HVIRTUAL: - return hl_dyn_getp(((vvirtual*)d)->value, hfield, t); - default: - hl_error("Invalid field access"); - } - return 0; -} - -void hl_dyn_seti( vdynamic *d, int hfield, hl_type *t, int value ) { - if( d == NULL ) hl_error("Invalid field access"); - switch( d->t->kind ) { - case HDYNOBJ: - { - vdynobj *o = (vdynobj*)d; - hl_field_lookup *f = hl_lookup_find(&o->dproto->fields,o->nfields,hfield); - if( f == NULL ) - f = hl_dyn_alloc_field(o,hfield,t); - else if( f->t != t ) - hl_error("Invalid dynset cast"); - *(int*)(o->fields_data + f->field_index) = value; - } - break; - case HOBJ: - hl_error("TODO"); - break; - case HVIRTUAL: - hl_dyn_seti(((vvirtual*)d)->value, hfield, t, value); - break; - default: - hl_error("Invalid field access"); - } -} - -void hl_dyn_setp( vdynamic *d, int hfield, hl_type *t, void *value ) { - if( d == NULL ) hl_error("Invalid field access"); - switch( d->t->kind ) { - case HDYNOBJ: - { - vdynobj *o = (vdynobj*)d; - hl_field_lookup *f = hl_lookup_find(&o->dproto->fields,o->nfields,hfield); - if( f == NULL ) - f = hl_dyn_alloc_field(o,hfield,t); - else if( f->t != t ) - hl_error("Invalid dynset cast"); - *(void**)(o->fields_data + f->field_index) = value; - } - break; - case HVIRTUAL: - hl_dyn_setp(((vvirtual*)d)->value, hfield, t, value); - break; - case HOBJ: - hl_error("TODO"); - break; - default: - hl_error("Invalid field access"); - } -} diff --git a/src/std/buffer.c b/src/std/buffer.c index 1e7ce0777..fcd6a3c1b 100644 --- a/src/std/buffer.c +++ b/src/std/buffer.c @@ -59,7 +59,17 @@ void hl_buffer_str_sub( hl_buffer *b, const uchar *s, int len ) { } void hl_buffer_str( hl_buffer *b, const uchar *s ) { - if( s ) hl_buffer_str_sub(b,s,(int)ustrlen(s)); + if( s ) hl_buffer_str_sub(b,s,(int)ustrlen(s)); else hl_buffer_str_sub(b,USTR("NULL"),4); +} + +void hl_buffer_cstr( hl_buffer *b, const char *s ) { + if( s ) { + int len = strlen(s); + uchar *out = (uchar*)malloc(sizeof(uchar)*(len+1)); + strtou(out,len,s); + hl_buffer_str_sub(b,out,len); + free(out); + } else hl_buffer_str_sub(b,USTR("NULL"),4); } void hl_buffer_char( hl_buffer *b, uchar c ) { diff --git a/src/std/obj.c b/src/std/obj.c index 222a1d8ed..681435bb0 100644 --- a/src/std/obj.c +++ b/src/std/obj.c @@ -1,4 +1,418 @@ -#include +#include "hl.h" +#include + +static void hl_lookup_insert( hl_field_lookup *l, int size, int hash, hl_type *t, int index ) { + int min = 0; + int max = size; + int pos; + while( min < max ) { + int mid = (min + max) >> 1; + int h = l[mid].hashed_name; + if( h < hash ) min = mid + 1; else max = mid; + } + pos = (min + max) >> 1; + memcpy(l + pos + 1, l + pos, (size - pos) * sizeof(hl_field_lookup)); + l[pos].field_index = index; + l[pos].hashed_name = hash; + l[pos].t = t; +} + +static hl_field_lookup *hl_lookup_find( hl_field_lookup *l, int size, int hash ) { + int min = 0; + int max = size; + while( min < max ) { + int mid = (min + max) >> 1; + int h = l[mid].hashed_name; + if( h < hash ) min = mid + 1; else if( h > hash ) max = mid; else return l + mid; + } + return NULL; +} + +static int hl_lookup_find_index( hl_field_lookup *l, int size, int hash ) { + int min = 0; + int max = size; + while( min < max ) { + int mid = (min + max) >> 1; + int h = l[mid].hashed_name; + if( h < hash ) min = mid + 1; else if( h > hash ) max = mid; else return mid; + } + return (min + max) >> 1; +} + +static int hl_cache_count = 0; +static int hl_cache_size = 0; +static hl_field_lookup *hl_cache = NULL; + +int hl_hash( vbytes *b ) { + return hl_hash_gen((uchar*)b,true); +} + +int hl_hash_gen( const uchar *name, bool cache_name ) { + int h = 0; + const uchar *oname = name; + while( *name ) { + h = 223 * h + (unsigned)*name; + name++; + } + h %= 0x1FFFFF7B; + if( cache_name ) { + hl_field_lookup *l = hl_lookup_find(hl_cache, hl_cache_count, h); + if( l == NULL ) { + if( hl_cache_size == hl_cache_count ) { + // resize + int newsize = hl_cache_size ? (hl_cache_size * 3) >> 1 : 16; + hl_field_lookup *cache = (hl_field_lookup*)hl_gc_alloc(sizeof(hl_field_lookup) * newsize); + memcpy(cache,hl_cache,sizeof(hl_field_lookup) * hl_cache_count); + free(hl_cache); + hl_cache = cache; + hl_cache_size = newsize; + } + hl_lookup_insert(hl_cache,hl_cache_count++,h,(hl_type*)ustrdup(oname),0); + } + } + return h; +} + +const uchar *hl_field_name( int hash ) { + hl_field_lookup *l = hl_lookup_find(hl_cache, hl_cache_count, hash); + return l ? (uchar*)l->t : USTR("???"); +} + +void hl_cache_free() { + int i; + for(i=0;iobj; + hl_module_context *m = o->m; + hl_alloc *alloc = &m->alloc; + hl_runtime_obj *p = NULL, *t; + int i, size, start; + if( o->rt ) return o->rt; + if( o->super ) p = hl_get_obj_rt(o->super); + t = (hl_runtime_obj*)hl_malloc(alloc,sizeof(hl_runtime_obj)); + t->t = ot; + t->nfields = o->nfields + (p ? p->nfields : 0); + t->nproto = p ? p->nproto : 0; + t->nlookup = o->nfields + o->nproto; + t->lookup = (hl_field_lookup*)hl_malloc(alloc,sizeof(hl_field_lookup) * t->nlookup); + t->fields_indexes = (int*)hl_malloc(alloc,sizeof(int)*t->nfields); + t->toStringFun = NULL; + t->compareFun = NULL; + t->castFun = NULL; + t->parent = p; + + // fields indexes + start = 0; + if( p ) { + start = p->nfields; + memcpy(t->fields_indexes, p->fields_indexes, sizeof(int)*p->nfields); + } + size = p ? p->size : HL_WSIZE; // hl_type* + for(i=0;infields;i++) { + hl_type *ft = o->fields[i].t; + size += hl_pad_size(size,ft); + t->fields_indexes[i+start] = size; + hl_lookup_insert(t->lookup,i,o->fields[i].hashed_name,o->fields[i].t,size); + size += hl_type_size(ft); + } + t->size = size; + t->methods = NULL; + o->rt = t; + ot->vobj_proto = NULL; + + // fields lookup + size = 0; + for(i=0;inproto;i++) { + hl_obj_proto *p = o->proto + i; + if( p->pindex >= t->nproto ) t->nproto = p->pindex + 1; + hl_lookup_insert(t->lookup,i + o->nfields,p->hashed_name,m->functions_types[p->findex],-(i+1)); + } + return t; +} + +/** + Fill class prototype with method pointers. + Requires method table to be finalized +**/ +hl_runtime_obj *hl_get_obj_proto( hl_type *ot ) { + hl_type_obj *o = ot->obj; + hl_module_context *m = o->m; + hl_alloc *alloc = &m->alloc; + hl_runtime_obj *p = NULL, *t = hl_get_obj_rt(ot); + hl_field_lookup *strField, *cmpField, *castField; + int i; + if( ot->vobj_proto ) return t; + if( o->super ) p = hl_get_obj_proto(o->super); + + strField = hl_lookup_find(t->lookup,t->nlookup,hl_hash_gen(USTR("__string"),false)); + cmpField = hl_lookup_find(t->lookup,t->nlookup,hl_hash_gen(USTR("__compare"),false)); + castField = hl_lookup_find(t->lookup,t->nlookup,hl_hash_gen(USTR("__cast"),false)); + t->toStringFun = strField ? m->functions_ptrs[o->proto[-(strField->field_index+1)].findex] : (p ? p->toStringFun : NULL); + t->compareFun = cmpField ? m->functions_ptrs[o->proto[-(cmpField->field_index+1)].findex] : (p ? p->compareFun : NULL); + t->castFun = castField ? m->functions_ptrs[o->proto[-(castField->field_index+1)].findex] : (p ? p->castFun : NULL); + + if( t->nproto ) { + void **fptr = (void**)hl_malloc(alloc, sizeof(void*) * t->nproto); + ot->vobj_proto = fptr; + if( p ) + memcpy(fptr, p->t->vobj_proto, p->nproto * sizeof(void*)); + for(i=0;inproto;i++) { + hl_obj_proto *p = o->proto + i; + if( p->pindex >= 0 ) fptr[p->pindex] = m->functions_ptrs[p->findex]; + } + } + + t->methods = (void**)hl_malloc(alloc, sizeof(void*) * o->nproto); + for(i=0;inproto;i++) { + hl_obj_proto *p = o->proto + i; + t->methods[i] = m->functions_ptrs[p->findex]; + } + + return t; +} + +/** + Allocate a virtual fields mapping to a given value. +**/ +vvirtual *hl_to_virtual( hl_type *vt, vdynamic *obj ) { + vvirtual *v; + if( obj == NULL ) return NULL; + switch( obj->t->kind ) { + case HOBJ: + { + int i; + hl_runtime_obj *rt = obj->t->obj->rt; + v = (vvirtual*)hl_gc_alloc(sizeof(vvirtual)); + v->t = vt; + v->fields_data = (char*)obj; + v->indexes = (int*)hl_gc_alloc(sizeof(int)*vt->virt->nfields); + v->value = obj; + v->next = NULL; + for(i=0;ivirt->nfields;i++) { + hl_runtime_obj *rtt = rt; + hl_field_lookup *f = NULL; + while( rtt ) { + f = hl_lookup_find(rtt->lookup,rtt->nlookup,vt->virt->fields[i].hashed_name); + if( f != NULL ) break; + rtt = rtt->parent; + } + v->indexes[i] = f == NULL || f->field_index <= 0 || (f->t != vt->virt->fields[i].t) ? 0 : f->field_index; + } + } + break; + case HDYNOBJ: + { + int i; + vdynobj *o = (vdynobj*)obj; + v = o->virtuals; + while( v ) { + if( v->t->virt == vt->virt ) + return v; + v = v->next; + } + // allocate a new virtual mapping + v = (vvirtual*)hl_gc_alloc(sizeof(vvirtual)); + v->t = vt; + v->fields_data = o->fields_data - sizeof(void*); + v->indexes = hl_gc_alloc(sizeof(int)*vt->virt->nfields); + v->value = obj; + for(i=0;ivirt->nfields;i++) { + hl_field_lookup *f = hl_lookup_find(&o->dproto->fields,o->nfields,vt->virt->fields[i].hashed_name); + v->indexes[i] = f == NULL || f->t != vt->virt->fields[i].t ? 0 : f->field_index + sizeof(void*); + } + // add it to the list + v->next = o->virtuals; + o->virtuals = v; + } + break; + default: + hl_fatal_fmt("Don't know how to virtual %d",obj->t->kind); + } + return v; +} + +/* +void *hl_fetch_method( vvirtual *v, int fid ) { + hl_obj_field *f = v->t->virt->fields + fid; + switch( v->value->t->kind ) { + case HOBJ: + { + hl_type_obj *o = ((vobj*)v->value)->t->obj; + hl_module_context *m = o->m; + hl_runtime_obj *rt = o->rt; + while( rt ) { + hl_field_lookup *found = hl_lookup_find(rt->lookup,rt->nlookup,f->hashed_name); + if( found ) { + int fid; + if( found->field_index > 0 ) return NULL; + fid = rt->obj->proto[-found->field_index].findex; + return hl_alloc_closure_ptr(m->functions_types[fid], m->functions_ptrs[fid], o); + } + rt = rt->parent; + } + } + break; + } + return NULL; +}*/ + +#define B2(t1,t2) ((t1) + ((t2) * HLAST)) +#define fetch_i(data,src,dst) src == dst ? *(int*)(data) : hl_dyn_casti(data,src,dst) +#define fetch_p(data,src,dst) src == dst ? *(void**)(data) : hl_dyn_castp(data,src,dst) + +static hl_field_lookup *hl_dyn_alloc_field( vdynobj *o, int hfield, hl_type *t ) { + int pad = hl_pad_size(o->dataSize, t); + int size = hl_type_size(t); + int index; + char *newData = (char*)hl_gc_alloc(o->dataSize + pad + size); + vdynobj_proto *proto = (vdynobj_proto*)hl_gc_alloc(sizeof(vdynobj_proto) + sizeof(hl_field_lookup) * (o->nfields + 1 - 1)); + int field_pos = hl_lookup_find_index(&o->dproto->fields, o->nfields, hfield); + hl_field_lookup *f; + // update data + memcpy(newData,o->fields_data,o->dataSize); + o->fields_data = newData; + o->dataSize += pad; + index = o->dataSize; + o->dataSize += size; + // update field table + proto->t = o->dproto->t; + memcpy(&proto->fields,&o->dproto->fields,field_pos * sizeof(hl_field_lookup)); + f = (&proto->fields) + field_pos; + f->t = t; + f->hashed_name = hfield; + f->field_index = index; + memcpy(&proto->fields + (field_pos + 1),&o->dproto->fields + field_pos, (o->nfields - field_pos) * sizeof(hl_field_lookup)); + o->nfields++; + o->dproto = proto; + return f; +} + +int hl_dyn_geti( vdynamic *d, int hfield, hl_type *t ) { + switch( d->t->kind ) { + case HDYNOBJ: + { + vdynobj *o = (vdynobj*)d; + hl_field_lookup *f = hl_lookup_find(&o->dproto->fields,o->nfields,hfield); + if( f == NULL ) return 0; + return fetch_i(o->fields_data + f->field_index,f->t, t); + } + break; + case HOBJ: + { + hl_runtime_obj *rt = d->t->obj->rt; + hl_field_lookup *f = NULL; + do { + f = hl_lookup_find(rt->lookup,rt->nlookup,hfield); + if( f != NULL ) break; + rt = rt->parent; + } while( rt ); + if( f == NULL ) return 0; + return fetch_i((char*)d + f->field_index,f->t,t); + } + break; + case HVIRTUAL: + return hl_dyn_geti(((vvirtual*)d)->value, hfield, t); + default: + hl_error("Invalid field access"); + break; + } + return 0; +} + +void *hl_dyn_getp( vdynamic *d, int hfield, hl_type *t ) { + switch( d->t->kind ) { + case HDYNOBJ: + { + vdynobj *o = (vdynobj*)d; + hl_field_lookup *f = hl_lookup_find(&o->dproto->fields,o->nfields,hfield); + if( f == NULL ) return 0; + return fetch_p(o->fields_data + f->field_index,f->t,t); + } + break; + case HOBJ: + { + hl_runtime_obj *rt = d->t->obj->rt; + hl_field_lookup *f = NULL; + do { + f = hl_lookup_find(rt->lookup,rt->nlookup,hfield); + if( f != NULL ) break; + rt = rt->parent; + } while( rt ); + if( f == NULL ) return NULL; + if( f->field_index < 0 ) { + vclosure *c = hl_alloc_closure_ptr(f->t,rt->methods[-f->field_index-1],d); + return hl_dyn_castp(&c,c->t,t); + } + return fetch_p((char*)d + f->field_index,f->t,t); + } + break; + case HVIRTUAL: + return hl_dyn_getp(((vvirtual*)d)->value, hfield, t); + default: + hl_error("Invalid field access"); + } + return 0; +} + +void hl_dyn_seti( vdynamic *d, int hfield, hl_type *t, int value ) { + if( d == NULL ) hl_error("Invalid field access"); + switch( d->t->kind ) { + case HDYNOBJ: + { + vdynobj *o = (vdynobj*)d; + hl_field_lookup *f = hl_lookup_find(&o->dproto->fields,o->nfields,hfield); + if( f == NULL ) + f = hl_dyn_alloc_field(o,hfield,t); + else if( f->t != t ) + hl_error("Invalid dynset cast"); + *(int*)(o->fields_data + f->field_index) = value; + } + break; + case HOBJ: + hl_error("TODO"); + break; + case HVIRTUAL: + hl_dyn_seti(((vvirtual*)d)->value, hfield, t, value); + break; + default: + hl_error("Invalid field access"); + } +} + +void hl_dyn_setp( vdynamic *d, int hfield, hl_type *t, void *value ) { + if( d == NULL ) hl_error("Invalid field access"); + switch( d->t->kind ) { + case HDYNOBJ: + { + vdynobj *o = (vdynobj*)d; + hl_field_lookup *f = hl_lookup_find(&o->dproto->fields,o->nfields,hfield); + if( f == NULL ) + f = hl_dyn_alloc_field(o,hfield,t); + else if( f->t != t ) + hl_error("Invalid dynset cast"); + *(void**)(o->fields_data + f->field_index) = value; + } + break; + case HVIRTUAL: + hl_dyn_setp(((vvirtual*)d)->value, hfield, t, value); + break; + case HOBJ: + hl_error("TODO"); + break; + default: + hl_error("Invalid field access"); + } +} HL_PRIM vdynamic *hl_obj_get_field( vdynamic *obj, int hfield ) { return (vdynamic*)hl_dyn_getp(obj,hfield,&hlt_dyn); @@ -9,7 +423,28 @@ HL_PRIM void hl_obj_set_field( vdynamic *obj, int hfield, vdynamic *v ) { } HL_PRIM bool hl_obj_has_field( vdynamic *obj, int hfield ) { - hl_fatal("TODO"); + if( obj == NULL ) return false; + switch( obj->t->kind ) { + case HOBJ: + { + hl_runtime_obj *rt = obj->t->obj->rt; + do { + hl_field_lookup *f = hl_lookup_find(rt->lookup,rt->nlookup,hfield); + if( f ) return f->field_index >= 0; + rt = rt->parent; + } while( rt ); + } + break; + case HDYNOBJ: + { + vdynobj *d = (vdynobj*)obj; + hl_field_lookup *f = hl_lookup_find(&d->dproto->fields,d->nfields,hfield); + return f != NULL; + } + break; + case HVIRTUAL: + return hl_obj_has_field(((vvirtual*)obj)->value,hfield); + } return false; } diff --git a/src/std/regexp.c b/src/std/regexp.c index 8db955c6b..4d806db44 100644 --- a/src/std/regexp.c +++ b/src/std/regexp.c @@ -1,22 +1,101 @@ #include +#define PCRE_STATIC +#include -typedef struct _ereg { - int _; -} ereg; +typedef struct _ereg ereg; + +static pcre16_extra limit; + +struct _ereg { + void (*finalize)( ereg * ); + pcre16 *p; + int *matches; + int nmatches; + bool matched; +}; + +static void regexp_finalize( ereg *e ) { + pcre16_free(e->p); + free(e->matches); +} HL_PRIM ereg *regexp_regexp_new_options( vbytes *str, vbytes *opts ) { - return NULL; + ereg *r; + const char *error; + int err_offset; + int errorcode; + pcre16 *p; + uchar *o = (uchar*)opts; + int options = 0; + while( *o ) { + switch( *o++ ) { + case 'i': + options |= PCRE_CASELESS; + break; + case 's': + options |= PCRE_DOTALL; + break; + case 'm': + options |= PCRE_MULTILINE; + break; + case 'u': + options |= PCRE_UTF8; + break; + case 'g': + options |= PCRE_UNGREEDY; + break; + default: + return NULL; + } + } + p = pcre16_compile2((PCRE_SPTR16)str,options,&errorcode,&error,&err_offset,NULL); + if( p == NULL ) { + hl_buffer *b = hl_alloc_buffer(); + hl_buffer_str(b,USTR("Regexp compilation error : ")); + hl_buffer_cstr(b,"abcde"); + hl_buffer_str(b,USTR(" in ")); + hl_buffer_str(b,(uchar*)str); + hl_error_msg(USTR("%s"),hl_buffer_content(b,NULL)); + } + r = (ereg*)hl_gc_alloc_finalizer(sizeof(ereg)); + r->finalize = regexp_finalize; + r->p = p; + r->nmatches = 0; + pcre16_fullinfo(p,NULL,PCRE_INFO_CAPTURECOUNT,&r->nmatches); + r->nmatches++; + r->matches = (int*)malloc(sizeof(int) * 3 * r->nmatches); + limit.flags = PCRE_EXTRA_MATCH_LIMIT_RECURSION; + limit.match_limit_recursion = 3500; // adapted based on Windows 1MB stack size + return r; } -HL_PRIM int regexp_regexp_matched_pos( ereg *e, int pos, int *len ) { - return 0; +static bool do_exec( ereg *e, vbytes *str, int len, int pos ) { + int res = pcre16_exec(e->p,&limit,(PCRE_SPTR16)str,len,pos,0,e->matches,e->nmatches * 3); + e->matched = res >= 0; + if( res >= 0 ) + return true; + if( res != PCRE_ERROR_NOMATCH ) + hl_error("An error occured while running pcre_exec"); + return false; +} + +HL_PRIM int regexp_regexp_matched_pos( ereg *e, int m, int *len ) { + int start; + if( !e->matched ) + hl_error("Calling matchedPos() on an unmatched regexp"); + if( m < 0 || m >= e->nmatches ) + hl_error_msg(USTR("Matched index %d outside bounds"),m); + start = e->matches[m*2]; + *len = e->matches[m*2+1] - start; + return start; } HL_PRIM vbytes *regexp_regexp_matched( ereg *e, int pos, int *len ) { + hl_fatal("TODO"); return NULL; } HL_PRIM bool regexp_regexp_match( ereg *e, vbytes *s, int pos, int len ) { - return false; + return do_exec(e,s,pos+len,pos); }