@willbooster/tree-sitter-javascript 1.0.0 → 1.0.2

package/grammar.js

@@ -573,12 +573,24 @@ module.exports = grammar({
       choice(
         seq(
           '"',
-          repeat(choice(alias($.unescaped_double_jsx_string_fragment, $.string_fragment), $.html_character_reference)),
+          repeat(
+            choice(
+              alias($.unescaped_double_jsx_string_fragment, $.string_fragment),
+              $.html_character_reference,
+              alias($._jsx_string_ampersand, $.string_fragment)
+            )
+          ),
           '"'
         ),
         seq(
           "'",
-          repeat(choice(alias($.unescaped_single_jsx_string_fragment, $.string_fragment), $.html_character_reference)),
+          repeat(
+            choice(
+              alias($.unescaped_single_jsx_string_fragment, $.string_fragment),
+              $.html_character_reference,
+              alias($._jsx_string_ampersand, $.string_fragment)
+            )
+          ),
           "'"
         )
       ),
@@ -587,10 +599,14 @@ module.exports = grammar({
     // We give names to the token() constructs containing a regexp
     // so as to obtain a node in the CST.
     //
-    unescaped_double_jsx_string_fragment: () => token.immediate(prec(1, /([^"&]|&[^#A-Za-z])+/)),
+    unescaped_double_jsx_string_fragment: () => token.immediate(prec(1, /([^"&]|&[^#A-Za-z"&])+/)),
 
     // same here
-    unescaped_single_jsx_string_fragment: () => token.immediate(prec(1, /([^'&]|&[^#A-Za-z])+/)),
+    unescaped_single_jsx_string_fragment: () => token.immediate(prec(1, /([^'&]|&[^#A-Za-z'&])+/)),
+
+    // An `&` that does not start a complete html_character_reference is a literal character, as in HTML. The lexer
+    // prefers the longer html_character_reference when one follows.
+    _jsx_string_ampersand: () => token.immediate(/&/),
 
     _jsx_attribute_value: ($) => choice(alias($._jsx_string, $.string), $.jsx_expression, $._jsx_element),
 
@@ -896,10 +912,22 @@ module.exports = grammar({
     template_string: ($) =>
       seq(
         '`',
-        repeat(choice(alias($._template_chars, $.string_fragment), $.escape_sequence, $.template_substitution)),
+        repeat(
+          choice(
+            alias($._template_chars, $.string_fragment),
+            $.escape_sequence,
+            alias($._invalid_template_escape, $.escape_sequence),
+            $.template_substitution
+          )
+        ),
         '`'
       ),
 
+    // Tagged templates allow `\x` and `\u` without the hex digits an escape_sequence needs (the cooked string is
+    // undefined). The lexer prefers the longer escape_sequence when the digits follow. Untagged templates reject these
+    // escapes, but the lexer cannot tell the two apart.
+    _invalid_template_escape: () => token.immediate(seq('\\', /[xu]/)),
+
     template_substitution: ($) => seq('${', $._expressions, '}'),
 
     regex: ($) =>

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@willbooster/tree-sitter-javascript",
-  "version": "1.0.0",
+  "version": "1.0.2",
   "description": "JavaScript grammar for tree-sitter",
   "keywords": [
     "incremental",

package/tree-sitter.json

@@ -16,7 +16,7 @@
     }
   ],
   "metadata": {
-    "version": "1.0.0",
+    "version": "1.0.2",
     "license": "MIT",
     "description": "JavaScript grammar for tree-sitter",
     "authors": [