XRegExp.escape: Escape whitespace in a way that works with ES6 flag u (…

…fixes #197)
slevithan · Feb 6, 2021 · e22a52b · e22a52b
1 parent a514be2
commit e22a52b
Show file tree

Hide file tree

Showing 2 changed files with 25 additions and 5 deletions.
diff --git a/src/xregexp.js b/src/xregexp.js
@@ -788,7 +788,13 @@ XRegExp.cache.flush = (cacheName) => {
  * XRegExp.escape('Escaped? <.>');
  * // -> 'Escaped\?\ <\.>'
  */
-XRegExp.escape = (str) => String(nullThrows(str)).replace(/[-\[\]{}()*+?.,\\^$|#\s]/g, '\\$&');
+XRegExp.escape = (str) => String(nullThrows(str)).replace(/[-\[\]{}()*+?.,\\^$|#\s]/g, (match) => {
+    if (/\s/.test(match)) {
+        // Converting to \uNNNN since whitespace can't be escaped when used with ES6 flag `u`
+        return `\\u${pad4(hex(match.charCodeAt(0)))}`;
+    }
+    return `\\${match}`;
+});
 
 /**
  * Executes a regex search in a specified string. Returns a match array or `null`. If the provided

diff --git a/tests/spec/s-xregexp-methods.js b/tests/spec/s-xregexp-methods.js
@@ -219,20 +219,34 @@ describe('XRegExp.cache()', function() {
 
 describe('XRegExp.escape()', function() {
 
-    it('should escape metacharacters', function() {
+    it('should escape standard metacharacters', function() {
         expect(XRegExp.escape('[()*+?.\\^$|')).toBe('\\[\\(\\)\\*\\+\\?\\.\\\\\\^\\$\\|');
     });
 
-    it('should escape context-aware metacharacters', function() {
-        expect(XRegExp.escape(']{}-, \n#')).toBe('\\]\\{\\}\\-\\,\\ \\\n\\#');
+    it('should escape context-sensitive metacharacters, excluding whitespace', function() {
+        expect(XRegExp.escape(']{}-,#')).toBe('\\]\\{\\}\\-\\,\\#');
+    });
+
+    it('should escape context-sensitive whitespace', function() {
+        // Don't want to explicitly check the escape pattern like this since there are multiple
+        // valid representations of escaped whitespace
+        //expect(XRegExp.escape('\t\n\f\r\x20\u2028\u2029')).toBe('\\u0009\\u000a\\u000c\\u000d\\u0020\\u2028\\u2029');
+
+        var whitespace = '\t\n\f\r\x20\u2028\u2029';
+        expect(new RegExp('^' + XRegExp.escape(whitespace) + '$').test(whitespace)).toBe(true);
+        expect(XRegExp.escape(whitespace)).not.toBe(whitespace);
     });
 
     it('should not escape nonmetacharacters', function() {
         expect(XRegExp.escape('abc_<123>!\0\uFFFF')).toBe('abc_<123>!\0\uFFFF');
     });
 
     it('should escape a nonstring pattern after type converting to a string', function() {
-        expect(XRegExp.escape({})).toBe('\\[object\\ Object\\]');
+        // Don't want to explicitly check the escape pattern like this since there are multiple
+        // valid representations of escaped whitespace
+        //expect(XRegExp.escape({})).toBe('\\[object\\ Object\\]');
+
+        expect(new RegExp('^' + XRegExp.escape({}) + '$').test('[object Object]')).toBe(true);
     });
 
     it('should throw an exception when given a null or undefined pattern', function() {