Fix handling of astral character reference

Closes GH-158. Closes remarkjs/remark#1251. Reviewed-by: Titus Wormer <tituswormer@gmail.com>
micromark · Oct 27, 2023 · 6db3b42 · 6db3b42
1 parent e0dce3d
commit 6db3b42
Show file tree

Hide file tree

Showing 3 changed files with 9 additions and 3 deletions.
diff --git a/packages/micromark-util-decode-numeric-character-reference/dev/index.js b/packages/micromark-util-decode-numeric-character-reference/dev/index.js
@@ -4,7 +4,7 @@ import {codes, values} from 'micromark-util-symbol'
  * Turn the number (in string form as either hexa- or plain decimal) coming from
  * a numeric character reference into a character.
  *
- * Sort of like `String.fromCharCode(Number.parseInt(value, base))`, but makes
+ * Sort of like `String.fromCodePoint(Number.parseInt(value, base))`, but makes
  * non-characters and control characters safe.
  *
  * @param {string} value
@@ -38,5 +38,5 @@ export function decodeNumericCharacterReference(value, base) {
     return values.replacementCharacter
   }
 
-  return String.fromCharCode(code)
+  return String.fromCodePoint(code)
 }
diff --git a/packages/micromark-util-decode-numeric-character-reference/readme.md b/packages/micromark-util-decode-numeric-character-reference/readme.md
@@ -78,7 +78,7 @@ There is no default export.
 Turn the number (in string form as either hexa- or plain decimal) coming from
 a numeric character reference into a character.
 
-Sort of like `String.fromCharCode(Number.parseInt(value, base))`, but makes
+Sort of like `String.fromCodePoint(Number.parseInt(value, base))`, but makes
 non-characters and control characters safe.
 
 ###### Parameters

diff --git a/test/io/text/character-reference.js b/test/io/text/character-reference.js
@@ -27,6 +27,12 @@ test('character-reference', function () {
     'should support hexadecimal character references'
   )
 
+  assert.equal(
+    micromark('片&#xE0103;'),
+    '<p>片\u{E0103}</p>',
+    'should support astral character references'
+  )
+
   assert.equal(
     micromark(
       [