fix(xml): don't advance after char data end

tree-sitter-grammars · Jun 30, 2024 · 809266e · 809266e
1 parent 648183d
commit 809266e
Show file tree

Hide file tree

Showing 6 changed files with 32 additions and 6 deletions.
diff --git a/Cargo.toml b/Cargo.toml
@@ -1,7 +1,7 @@
 [package]
 name = "tree-sitter-xml"
 description = "XML & DTD grammars for tree-sitter"
-version = "0.6.3"
+version = "0.6.4"
 license = "MIT"
 readme = "README.md"
 keywords = ["incremental", "parsing", "tree-sitter", "dtd", "xml"]

diff --git a/common/common.mak b/common/common.mak
@@ -1,4 +1,4 @@
-VERSION := 0.6.3
+VERSION := 0.6.4
 
 # repository
 SRC_DIR := src

diff --git a/package.json b/package.json
@@ -1,6 +1,6 @@
 {
   "name": "@tree-sitter-grammars/tree-sitter-xml",
-  "version": "0.6.3",
+  "version": "0.6.4",
   "license": "MIT",
   "description": "XML & DTD grammars for tree-sitter",
   "repository": "tree-sitter-grammars/tree-sitter-xml",

diff --git a/pyproject.toml b/pyproject.toml
@@ -5,7 +5,7 @@ build-backend = "setuptools.build_meta"
 [project]
 name = "tree-sitter-xml"
 description = "XML & DTD grammars for tree-sitter"
-version = "0.6.3"
+version = "0.6.4"
 keywords = ["incremental", "parsing", "tree-sitter", "xml"]
 classifiers = [
   "Intended Audience :: Developers",

diff --git a/test/corpus/issues.txt b/test/corpus/issues.txt
@@ -283,3 +283,22 @@ Parameter-entity references (#10)
       (DefaultDecl))
     (PEReference
       (Name))))
+
+================================================================================
+Closing bracket in CharData (#22)
+
+:language(xml)
+================================================================================
+
+<test>]</test>
+
+--------------------------------------------------------------------------------
+
+(document
+  (element
+    (STag
+      (Name))
+    (content
+      (CharData))
+    (ETag
+      (Name))))
diff --git a/xml/src/scanner.c b/xml/src/scanner.c
@@ -72,11 +72,16 @@ static inline bool in_error_recovery(const bool *valid_symbols) {
            valid_symbols[CHAR_DATA] && valid_symbols[CDATA];
 }
 
+/// Check if the lexer is in a char data node
+static inline bool in_char_data(TSLexer *lexer) {
+    return !lexer->eof(lexer) && lexer->lookahead != '<' && lexer->lookahead != '&';
+}
+
 /// Scan for a CharData node
 static bool scan_char_data(TSLexer *lexer) {
     bool advanced_once = false;
 
-    while (!lexer->eof(lexer) && lexer->lookahead != '<' && lexer->lookahead != '&') {
+    while (in_char_data(lexer)) {
         if (lexer->lookahead == ']') {
             lexer->mark_end(lexer);
             advance(lexer);
@@ -92,7 +97,9 @@ static bool scan_char_data(TSLexer *lexer) {
             }
         }
         advanced_once = true;
-        advance(lexer);
+        if (in_char_data(lexer)) {
+            advance(lexer);
+        }
     }
 
     if (advanced_once) {