fix: only disallow xml, xml-stylesheet, xml-model in PITarget (#3)

Wilfred · Aug 9, 2023 · 148ad3d · 148ad3d
1 parent 9fcc73c
commit 148ad3d
Show file tree

Hide file tree

Showing 7 changed files with 88 additions and 33 deletions.
diff --git a/common/scanner.h b/common/scanner.h
@@ -8,13 +8,20 @@ enum TokenType {
     PIContent,
     Comment,
     CharData,
+    XmlModel,
+    XmlStylesheet,
 };
 
 /// Advance the lexer to the next token
 static inline void advance(TSLexer *lexer) { lexer->advance(lexer, false); }
 
+static inline bool is_valid_pi_char(int32_t chr) {
+    return isalnum(chr) || chr == '_' || chr == ':' || chr == '.' ||
+           chr == '-' || chr == L'·';
+}
+
 /// Scan for the target of a PI node
-static bool scan_pi_target(TSLexer *lexer) {
+static bool scan_pi_target(TSLexer *lexer, const bool *valid_symbols) {
     bool advanced_once = false, found_x_first = false;
 
     if (isalpha(lexer->lookahead) || lexer->lookahead == '_') {
@@ -27,27 +34,50 @@ static bool scan_pi_target(TSLexer *lexer) {
     }
 
     if (advanced_once) {
-        while (isalnum(lexer->lookahead) || lexer->lookahead == '_' ||
-               lexer->lookahead == ':' || lexer->lookahead == '.' ||
-               lexer->lookahead == L'·' || lexer->lookahead == '-') {
-            if (lexer->lookahead == 'x' || lexer->lookahead == 'X') {
-                lexer->mark_end(lexer);
-                advance(lexer);
-                if (lexer->lookahead == 'm' || lexer->lookahead == 'M') {
-                    advance(lexer);
-                    if (lexer->lookahead == 'l' || lexer->lookahead == 'L') {
-                        advance(lexer);
-                        return false;
-                    }
-                }
-            }
-
+        while (is_valid_pi_char(lexer->lookahead)) {
             if (found_x_first &&
                 (lexer->lookahead == 'm' || lexer->lookahead == 'M')) {
                 advance(lexer);
                 if (lexer->lookahead == 'l' || lexer->lookahead == 'L') {
                     advance(lexer);
-                    return false;
+                    if (is_valid_pi_char(lexer->lookahead)) {
+                        found_x_first = false;
+                        bool last_char_hyphen = lexer->lookahead == '-';
+                        advance(lexer);
+                        if (last_char_hyphen) {
+                            // scan for stylesheet/model and disallow that
+                            if (valid_symbols[XmlModel]) {
+                                const char *const word = "model";
+                                int j = 0;
+                                while (word[j] != '\0') {
+                                    if (word[j] != lexer->lookahead) {
+                                        break;
+                                    }
+                                    j++;
+                                    advance(lexer);
+                                }
+                                if (word[j] == '\0') {
+                                    return false;
+                                }
+                            }
+                            if (valid_symbols[XmlStylesheet]) {
+                                const char *const word = "stylesheet";
+                                int j = 0;
+                                while (word[j] != '\0') {
+                                    if (word[j] != lexer->lookahead) {
+                                        break;
+                                    }
+                                    j++;
+                                    advance(lexer);
+                                }
+                                if (word[j] == '\0') {
+                                    return false;
+                                }
+                            }
+                        }
+                    } else {
+                        return false;
+                    }
                 }
             }
 

diff --git a/tree-sitter-dtd/src/scanner.c b/tree-sitter-dtd/src/scanner.c
@@ -8,7 +8,7 @@ static inline bool in_error_recovery(const bool *valid_symbols) {
 bool tree_sitter_dtd_external_scanner_scan(void *payload, TSLexer *lexer, const bool *valid_symbols) {
     if (in_error_recovery(valid_symbols)) return false;
 
-    if (valid_symbols[PITarget]) return scan_pi_target(lexer);
+    if (valid_symbols[PITarget]) return scan_pi_target(lexer, valid_symbols);
 
     if (valid_symbols[PIContent]) return scan_pi_content(lexer);
 

diff --git a/tree-sitter-dtd/src/tree_sitter/parser.h b/tree-sitter-dtd/src/tree_sitter/parser.h
@@ -13,9 +13,8 @@ extern "C" {
 #define ts_builtin_sym_end 0
 #define TREE_SITTER_SERIALIZATION_BUFFER_SIZE 1024
 
-typedef uint16_t TSStateId;
-
 #ifndef TREE_SITTER_API_H_
+typedef uint16_t TSStateId;
 typedef uint16_t TSSymbol;
 typedef uint16_t TSFieldId;
 typedef struct TSLanguage TSLanguage;
@@ -130,9 +129,16 @@ struct TSLanguage {
  *  Lexer Macros
  */
 
+#ifdef _MSC_VER
+#define UNUSED __pragma(warning(suppress : 4101))
+#else
+#define UNUSED __attribute__((unused))
+#endif
+
 #define START_LEXER()           \
   bool result = false;          \
   bool skip = false;            \
+  UNUSED                        \
   bool eof = false;             \
   int32_t lookahead;            \
   goto start;                   \
@@ -166,7 +172,7 @@ struct TSLanguage {
  *  Parse Table Macros
  */
 
-#define SMALL_STATE(id) id - LARGE_STATE_COUNT
+#define SMALL_STATE(id) ((id) - LARGE_STATE_COUNT)
 
 #define STATE(id) id
 
@@ -176,15 +182,15 @@ struct TSLanguage {
   {{                                  \
     .shift = {                        \
       .type = TSParseActionTypeShift, \
-      .state = state_value            \
+      .state = (state_value)          \
     }                                 \
   }}
 
 #define SHIFT_REPEAT(state_value)     \
   {{                                  \
     .shift = {                        \
       .type = TSParseActionTypeShift, \
-      .state = state_value,           \
+      .state = (state_value),         \
       .repetition = true              \
     }                                 \
   }}

diff --git a/tree-sitter-xml/grammar.js b/tree-sitter-xml/grammar.js
@@ -22,6 +22,8 @@ module.exports = grammar(DTD, {
     $._pi_content,
     $.Comment,
     $.CharData,
+    'xml-model',
+    'xml-stylesheet',
   ],
 
   inline: $ => [

diff --git a/tree-sitter-xml/src/parser.c b/tree-sitter-xml/src/parser.c
diff --git a/tree-sitter-xml/src/scanner.c b/tree-sitter-xml/src/scanner.c
@@ -43,7 +43,7 @@ bool tree_sitter_xml_external_scanner_scan(void *payload, TSLexer *lexer,
                                            const bool *valid_symbols) {
     if (in_error_recovery(valid_symbols)) return false;
 
-    if (valid_symbols[PITarget]) return scan_pi_target(lexer);
+    if (valid_symbols[PITarget]) return scan_pi_target(lexer, valid_symbols);
 
     if (valid_symbols[PIContent]) return scan_pi_content(lexer);
 

diff --git a/tree-sitter-xml/src/tree_sitter/parser.h b/tree-sitter-xml/src/tree_sitter/parser.h
@@ -13,9 +13,8 @@ extern "C" {
 #define ts_builtin_sym_end 0
 #define TREE_SITTER_SERIALIZATION_BUFFER_SIZE 1024
 
-typedef uint16_t TSStateId;
-
 #ifndef TREE_SITTER_API_H_
+typedef uint16_t TSStateId;
 typedef uint16_t TSSymbol;
 typedef uint16_t TSFieldId;
 typedef struct TSLanguage TSLanguage;
@@ -130,9 +129,16 @@ struct TSLanguage {
  *  Lexer Macros
  */
 
+#ifdef _MSC_VER
+#define UNUSED __pragma(warning(suppress : 4101))
+#else
+#define UNUSED __attribute__((unused))
+#endif
+
 #define START_LEXER()           \
   bool result = false;          \
   bool skip = false;            \
+  UNUSED                        \
   bool eof = false;             \
   int32_t lookahead;            \
   goto start;                   \
@@ -166,7 +172,7 @@ struct TSLanguage {
  *  Parse Table Macros
  */
 
-#define SMALL_STATE(id) id - LARGE_STATE_COUNT
+#define SMALL_STATE(id) ((id) - LARGE_STATE_COUNT)
 
 #define STATE(id) id
 
@@ -176,15 +182,15 @@ struct TSLanguage {
   {{                                  \
     .shift = {                        \
       .type = TSParseActionTypeShift, \
-      .state = state_value            \
+      .state = (state_value)          \
     }                                 \
   }}
 
 #define SHIFT_REPEAT(state_value)     \
   {{                                  \
     .shift = {                        \
       .type = TSParseActionTypeShift, \
-      .state = state_value,           \
+      .state = (state_value),         \
       .repetition = true              \
     }                                 \
   }}