Input refactoring in progress.

taocpp · Dec 16, 2023 · 708fc94 · 708fc94
1 parent 3fec7aa
commit 708fc94
Show file tree

Hide file tree

Showing 106 changed files with 778 additions and 682 deletions.
diff --git a/doc/Changelog.md b/doc/Changelog.md
@@ -29,6 +29,7 @@
 * Added rule [`try_catch_raise_nested`](Rule-Reference.md#try_catch_raise_nested-r-).
 * Added rule [`try_catch_std_raise_nested`](Rule-Reference.md#try_catch_std_raise_nested-r-).
 * Added rule [`try_catch_type_raise_nested`](Rule-Reference.md#try_catch_type_raise_nested-e-r-).
+* Added rules for matching signed integers mirroring the existing ones for unsigned integers.
 * Moved depth counter to adapter class `input_with_depth` in [contrib](Contrib-and-Examples#contrib).
 * Changed default top-level `rewind_mode` to ~~`dontcare`~~ `optional`.
 * Replaced `rewind_mode` values `dontcare` and `active` with new value `optional`.

diff --git a/doc/Rule-Reference.md b/doc/Rule-Reference.md
@@ -1466,20 +1466,30 @@ Convenience wrappers for enumerated properties that return a value instead of an
 
 These rules are available in multiple versions,
 
-* in namespace `tao::pegtl::uint8` for 8-bit integer values,
+* in namespace `tao::pegtl::int8` for signed 8-bit integer values,
+* in namespace `tao::pegtl::uint8` for unsigned 8-bit integer values,
+* in namespace `tao::pegtl::int16_be` for big-endian 16-bit integer values,
+* in namespace `tao::pegtl::int16_le` for little-endian 16-bit integer values,
+* in namespace `tao::pegtl::int32_be` for big-endian 32-bit integer values,
+* in namespace `tao::pegtl::int32_le` for little-endian 32-bit integer values,
+* in namespace `tao::pegtl::int64_be` for big-endian 64-bit integer values, and
+* in namespace `tao::pegtl::int64_le` for little-endian 64-bit integer values.
 * in namespace `tao::pegtl::uint16_be` for big-endian 16-bit integer values,
 * in namespace `tao::pegtl::uint16_le` for little-endian 16-bit integer values,
 * in namespace `tao::pegtl::uint32_be` for big-endian 32-bit integer values,
 * in namespace `tao::pegtl::uint32_le` for little-endian 32-bit integer values,
 * in namespace `tao::pegtl::uint64_be` for big-endian 64-bit integer values, and
-* in namespace `tao::pegtl::uint64_le` for little-endian 64-bit integer values.
+* in namespace `tao::pegtl::uint64_le` for little-endian 64-bit integer values,
 
-The binary rules need to be manually included from their corresponding headers in the `contrib` section.
+however please not that the masked rules are available only for unsigned integers.
 
-These rules read one or more bytes from the input to form (and match) an 8, 16, 32 or 64-bit value, respectively, and corresponding template parameters are given as either `std::uint8_t`, `std::uint16_t`, `std::uint32_t` or `std::uin64_t`.
+The binary rules need to be manually included from their corresponding headers.
+
+These rules read one or more bytes from the input to form (and match) an 8, 16, 32 or 64-bit value, respectively, and corresponding template parameters are given as either `std::int8_t`, `std::uint8_t`, `std::int16_t`, `std::uint16_t`, `std::int32_t`, `std::uint32_t`, `std::int64_t` or `std::uin64_t`.
 
 In the following descriptions, the parameter N is the size of a single value in bytes, i.e. either 1, 2, 4 or 8.
-The term *input value* indicates a correspondingly sized integer value read from successive bytes of the input.
+The term *input value* indicates a correspondingly sized integer value read from the input.
+For inputs of values of size 1 like `char` or `std::byte` all integer rules can be used, for inputs of values of size greater than 1 like `int` or `long` only integer rules of matching size are possible.
 
 Binary rules do not rely on other rules.
 

diff --git a/include/tao/pegtl/ascii.hpp b/include/tao/pegtl/ascii.hpp
@@ -7,7 +7,7 @@
 
 #include "config.hpp"
 
-#include "internal/peeks.hpp"
+#include "internal/peek_integer.hpp"
 #include "internal/result_on_found.hpp"
 #include "internal/rules.hpp"
 

diff --git a/include/tao/pegtl/contrib/http.hpp b/include/tao/pegtl/contrib/http.hpp
@@ -12,12 +12,11 @@
 #include "../ascii.hpp"
 #include "../config.hpp"
 #include "../nothing.hpp"
+#include "../remove_first_state.hpp"
 #include "../rules.hpp"
 #include "../utf8.hpp"
 
 #include "abnf.hpp"
-#include "forward.hpp"
-#include "remove_first_state.hpp"
 #include "uri.hpp"
 
 namespace TAO_PEGTL_NAMESPACE::http
@@ -167,7 +166,7 @@ namespace TAO_PEGTL_NAMESPACE::http
  }
  break;
  }
- in.bump_in_this_line( i );
+ in.template consume< internal::eol_exclude_tag >( i );
  return i > 0;
  }
  };
@@ -193,7 +192,7 @@ namespace TAO_PEGTL_NAMESPACE::http
  [[nodiscard]] static bool match( ParseInput& in, const std::size_t size, States&&... /*unused*/ )
  {
  if( in.size( size ) >= size ) {
- in.bump( size );
+ in.template consume< internal::eol_unknown_tag >( size );
  return true;
  }
  return false;

diff --git a/include/tao/pegtl/contrib/unescape.hpp b/include/tao/pegtl/contrib/unescape.hpp
@@ -17,7 +17,7 @@ namespace TAO_PEGTL_NAMESPACE::unescape
 {
  // Utility functions for the unescape actions.
 
- [[nodiscard]] inline bool utf8_append_utf32( std::string& string, const unsigned utf32 )
+ [[nodiscard]] inline bool utf8_append_utf32( std::string& string, const char32_t utf32 )
  {
  if( utf32 <= 0x7f ) {
  string += static_cast< char >( utf32 & 0xff );
@@ -150,7 +150,7 @@ namespace TAO_PEGTL_NAMESPACE::unescape
  static void apply( const ActionInput& in, std::string& s )
  {
  assert( !in.empty() ); // First character MUST be present, usually 'u' or 'U'.
- if( !utf8_append_utf32( s, unhex_string< unsigned >( in.begin() + 1, in.end() ) ) ) {
+ if( !utf8_append_utf32( s, unhex_string< char32_t >( in.begin() + 1, in.end() ) ) ) {
  throw parse_error( "invalid escaped unicode code point", in );
  }
  }
@@ -159,7 +159,7 @@ namespace TAO_PEGTL_NAMESPACE::unescape
  static bool apply( const ActionInput& in, std::string& s )
  {
  assert( !in.empty() ); // First character MUST be present, usually 'u' or 'U'.
- return utf8_append_utf32( s, unhex_string< unsigned >( in.begin() + 1, in.end() ) );
+ return utf8_append_utf32( s, unhex_string< char32_t >( in.begin() + 1, in.end() ) );
  }
 #endif
  };
@@ -189,9 +189,9 @@ namespace TAO_PEGTL_NAMESPACE::unescape
  {
  assert( ( ( in.size() + 1 ) % 6 ) == 0 ); // Expects multiple "\\u1234", starting with the first "u".
  for( const char* b = in.begin() + 1; b < in.end(); b += 6 ) {
- const auto c = unhex_string< unsigned >( b, b + 4 );
+ const auto c = unhex_string< char32_t >( b, b + 4 );
  if( ( 0xd800 <= c ) && ( c <= 0xdbff ) && ( b + 6 < in.end() ) ) {
- const auto d = unhex_string< unsigned >( b + 6, b + 10 );
+ const auto d = unhex_string< char32_t >( b + 6, b + 10 );
  if( ( 0xdc00 <= d ) && ( d <= 0xdfff ) ) {
  b += 6;
  (void)utf8_append_utf32( s, ( ( ( c & 0x03ff ) << 10 ) | ( d & 0x03ff ) ) + 0x10000 );

diff --git a/include/tao/pegtl/enums.hpp b/include/tao/pegtl/enums.hpp
@@ -0,0 +1,45 @@
+// Copyright (c) 2022-2023 Dr. Colin Hirsch and Daniel Frey
+// Distributed under the Boost Software License, Version 1.0.
+// (See accompanying file LICENSE_1_0.txt or copy at https://www.boost.org/LICENSE_1_0.txt)
+
+#ifndef TAO_PEGTL_ENUMS_HPP
+#define TAO_PEGTL_ENUMS_HPP
+
+#include "config.hpp"
+
+#include "internal/endian.hpp"
+#include "internal/peek_endian.hpp"
+#include "internal/result_on_found.hpp"
+#include "internal/rules.hpp"
+
+namespace TAO_PEGTL_NAMESPACE
+{
+ namespace enums_be
+ {
+ // clang-format off
+ template< auto E, decltype( E )... Es > struct not_one : internal::one< internal::result_on_found::failure, internal::peek_endian< decltype( E ), internal::big_endian >, E, Es... > {};
+ template< auto Lo, decltype( Lo ) Hi > struct not_range : internal::range< internal::result_on_found::failure, internal::peek_endian< decltype( Lo ), internal::big_endian >, Lo, Hi > {};
+ template< auto E, decltype( E )... Es > struct one : internal::one< internal::result_on_found::success, internal::peek_endian< decltype( E ), internal::big_endian >, E, Es... > {};
+ template< auto Lo, decltype( Lo ) Hi > struct range : internal::range< internal::result_on_found::success, internal::peek_endian< decltype( Lo ), internal::big_endian >, Lo, Hi > {};
+ template< auto E, decltype( E )... Es > struct ranges : internal::ranges< internal::peek_endian< decltype( E ), internal::big_endian >, E, Es... > {};
+ template< auto E, decltype( E )... Es > struct string : internal::seq< internal::one< internal::result_on_found::success, internal::peek_endian< decltype( E ), internal::big_endian >, E >, internal::one< internal::result_on_found::success, internal::peek_endian< decltype( E ), internal::big_endian >, Es >... > {};
+ // clang-format on
+
+ } // namespace enums_be
+
+ namespace enums_le
+ {
+ // clang-format off
+ template< auto E, decltype( E )... Es > struct not_one : internal::one< internal::result_on_found::failure, internal::peek_endian< decltype( E ), internal::little_endian >, E, Es... > {};
+ template< auto Lo, decltype( Lo ) Hi > struct not_range : internal::range< internal::result_on_found::failure, internal::peek_endian< decltype( Lo ), internal::little_endian >, Lo, Hi > {};
+ template< auto E, decltype( E )... Es > struct one : internal::one< internal::result_on_found::success, internal::peek_endian< decltype( E ), internal::little_endian >, E, Es... > {};
+ template< auto Lo, decltype( Lo ) Hi > struct range : internal::range< internal::result_on_found::success, internal::peek_endian< decltype( Lo ), internal::little_endian >, Lo, Hi > {};
+ template< auto E, decltype( E )... Es > struct ranges : internal::ranges< internal::peek_endian< decltype( E ), internal::little_endian >, E, Es... > {};
+ template< auto E, decltype( E )... Es > struct string : internal::seq< internal::one< internal::result_on_found::success, internal::peek_endian< decltype( E ), internal::little_endian >, E >, internal::one< internal::result_on_found::success, internal::peek_endian< decltype( E ), internal::little_endian >, Es >... > {};
+ // clang-format on
+
+ } // namespace enums_le
+
+} // namespace TAO_PEGTL_NAMESPACE
+
+#endif
diff --git a/include/tao/pegtl/inputs.hpp b/include/tao/pegtl/inputs.hpp
@@ -24,7 +24,8 @@ namespace TAO_PEGTL_NAMESPACE
  using argv_input_with_source = internal::input_with_fakes< internal::input_with_peeks< internal::argv_input_with_source > >; // TODO: Add input_with_start?
  template< typename Container >
  using copy_input_with_source = internal::input_with_fakes< internal::input_with_peeks< internal::input_with_source< std::string, internal::copy_input< Container > > > >;
- using file_input_with_source = internal::input_with_fakes< internal::input_with_peeks< internal::file_input_with_source > >;
+ using file_input_with_source = internal::input_with_fakes< internal::input_with_peeks< internal::file_input_with_source< file_input > > >;
+ using read_input_with_source = internal::input_with_fakes< internal::input_with_peeks< internal::file_input_with_source< read_input > > >;
  template< typename Data >
  using view_input_with_source = internal::input_with_fakes< internal::input_with_peeks< internal::input_with_source< std::string, internal::view_input< Data > > > >; // TODO: Add input_with_start?
 
@@ -37,6 +38,10 @@ namespace TAO_PEGTL_NAMESPACE
  template< typename Eol >
  using text_file_input = internal::input_with_fakes< internal::input_with_peeks< internal::text_input< Eol, internal::file_input > > >;
  template< typename Eol >
+ using lazy_read_input = internal::input_with_fakes< internal::input_with_peeks< internal::lazy_input< Eol, internal::read_input > > >;
+ template< typename Eol >
+ using text_read_input = internal::input_with_fakes< internal::input_with_peeks< internal::text_input< Eol, internal::read_input > > >;
+ template< typename Eol >
  using lazy_view_input = internal::input_with_fakes< internal::input_with_peeks< internal::lazy_input< Eol, internal::view_input< char > > > >; // TODO: Add input_with_start?
  template< typename Eol >
  using text_view_input = internal::input_with_fakes< internal::input_with_peeks< internal::text_input< Eol, internal::view_input< char > > > >; // TODO: Add input_with_start?
@@ -46,14 +51,31 @@ namespace TAO_PEGTL_NAMESPACE
  template< typename Eol >
  using text_copy_input_with_source = internal::input_with_fakes< internal::input_with_peeks< internal::text_input_with_source< Eol, std::string, internal::copy_input< std::string > > > >;
  template< typename Eol >
- using lazy_file_input_with_source = internal::input_with_fakes< internal::input_with_peeks< internal::lazy_file_input_with_source< Eol > > >;
+ using lazy_file_input_with_source = internal::input_with_fakes< internal::input_with_peeks< internal::input_double_path< internal::lazy_input_with_source< Eol, std::filesystem::path, internal::file_input > > > >;
+ template< typename Eol >
+ using text_file_input_with_source = internal::input_with_fakes< internal::input_with_peeks< internal::input_double_path< internal::text_input_with_source< Eol, std::filesystem::path, internal::file_input > > > >;
  template< typename Eol >
- using text_file_input_with_source = internal::input_with_fakes< internal::input_with_peeks< internal::text_file_input_with_source< Eol > > >;
+ using lazy_read_input_with_source = internal::input_with_fakes< internal::input_with_peeks< internal::input_double_path< internal::lazy_input_with_source< Eol, std::filesystem::path, internal::read_input > > > >;
+ template< typename Eol >
+ using text_read_input_with_source = internal::input_with_fakes< internal::input_with_peeks< internal::input_double_path< internal::text_input_with_source< Eol, std::filesystem::path, internal::read_input > > > >;
  template< typename Eol >
  using lazy_view_input_with_source = internal::input_with_fakes< internal::input_with_peeks< internal::lazy_input_with_source< Eol, std::string, internal::view_input< char > > > >; // TODO: Add input_with_start?
  template< typename Eol >
  using text_view_input_with_source = internal::input_with_fakes< internal::input_with_peeks< internal::text_input_with_source< Eol, std::string, internal::view_input< char > > > >; // TODO: Add input_with_start?
 
+#if defined( TAO_PEGTL_MMAP_AVAILABLE )
+ using mmap_input = internal::input_with_fakes< internal::input_with_peeks< internal::mmap_input< char > > >;
+ using mmap_input_with_source = internal::input_with_fakes< internal::input_with_peeks< internal::file_input_with_source< internal::mmap_input< char > > > >;
+ template< typename Eol >
+ using lazy_mmap_input = internal::input_with_fakes< internal::input_with_peeks< internal::lazy_input< Eol, internal::mmap_input< char > > > >;
+ template< typename Eol >
+ using text_mmap_input = internal::input_with_fakes< internal::input_with_peeks< internal::text_input< Eol, internal::mmap_input< char > > > >;
+ template< typename Eol >
+ using lazy_mmap_input_with_source = internal::input_with_fakes< internal::input_with_peeks< internal::input_double_path< internal::lazy_input_with_source< Eol, std::filesystem::path, internal::mmap_input< char > > > > >;
+ template< typename Eol >
+ using text_mmap_input_with_source = internal::input_with_fakes< internal::input_with_peeks< internal::input_double_path< internal::text_input_with_source< Eol, std::filesystem::path, internal::mmap_input< char > > > > >;
+#endif
+
 } // namespace TAO_PEGTL_NAMESPACE
 
 #endif
diff --git a/include/tao/pegtl/int16.hpp b/include/tao/pegtl/int16.hpp
@@ -0,0 +1,50 @@
+// Copyright (c) 2018-2023 Dr. Colin Hirsch and Daniel Frey
+// Distributed under the Boost Software License, Version 1.0.
+// (See accompanying file LICENSE_1_0.txt or copy at https://www.boost.org/LICENSE_1_0.txt)
+
+#ifndef TAO_PEGTL_INT16_HPP
+#define TAO_PEGTL_INT16_HPP
+
+#include "config.hpp"
+
+#include "internal/peek_integer.hpp"
+#include "internal/result_on_found.hpp"
+#include "internal/rules.hpp"
+
+namespace TAO_PEGTL_NAMESPACE
+{
+ namespace int16_be
+ {
+ // clang-format off
+ struct any : internal::any< internal::peek_int16_be > {};
+ template< unsigned Count > struct many : internal::many< Count, internal::peek_int16_be > {};
+
+ template< std::int16_t... Cs > struct not_one : internal::one< internal::result_on_found::failure, internal::peek_int16_be, Cs... > {};
+ template< std::int16_t Lo, std::int16_t Hi > struct not_range : internal::range< internal::result_on_found::failure, internal::peek_int16_be, Lo, Hi > {};
+ template< std::int16_t... Cs > struct one : internal::one< internal::result_on_found::success, internal::peek_int16_be, Cs... > {};
+ template< std::int16_t Lo, std::int16_t Hi > struct range : internal::range< internal::result_on_found::success, internal::peek_int16_be, Lo, Hi > {};
+ template< std::int16_t... Cs > struct ranges : internal::ranges< internal::peek_int16_be, Cs... > {};
+ template< std::int16_t... Cs > struct string : internal::seq< internal::one< internal::result_on_found::success, internal::peek_int16_be, Cs >... > {};
+ // clang-format on
+
+ } // namespace int16_be
+
+ namespace int16_le
+ {
+ // clang-format off
+ struct any : internal::any< internal::peek_int16_le > {};
+ template< unsigned Count > struct many : internal::many< Count, internal::peek_int16_le > {};
+
+ template< std::int16_t... Cs > struct not_one : internal::one< internal::result_on_found::failure, internal::peek_int16_le, Cs... > {};
+ template< std::int16_t Lo, std::int16_t Hi > struct not_range : internal::range< internal::result_on_found::failure, internal::peek_int16_le, Lo, Hi > {};
+ template< std::int16_t... Cs > struct one : internal::one< internal::result_on_found::success, internal::peek_int16_le, Cs... > {};
+ template< std::int16_t Lo, std::int16_t Hi > struct range : internal::range< internal::result_on_found::success, internal::peek_int16_le, Lo, Hi > {};
+ template< std::int16_t... Cs > struct ranges : internal::ranges< internal::peek_int16_le, Cs... > {};
+ template< std::int16_t... Cs > struct string : internal::seq< internal::one< internal::result_on_found::success, internal::peek_int16_le, Cs >... > {};
+ // clang-format on
+
+ } // namespace int16_le
+
+} // namespace TAO_PEGTL_NAMESPACE
+
+#endif