Expand unesc handling to correctly handle spec edgecase for lone

samccone · samccone · commit 5c6c98885484 · 2021-04-12T18:23:36.000-07:00
surrogates and out of bound codepoint values.
diff --git a/src/__tests__/util/unesc.js b/src/__tests__/util/unesc.js
@@ -48,3 +48,15 @@ test('class selector with escaping with more chars', '.\\1D306k', (t, tree) => {
 test('class selector with escaping with more chars with whitespace', '.wow\\1D306 k', (t, tree) => {
     t.deepEqual(tree.nodes[0].nodes[0].value, 'wow𝌆k');
 });
+
+test('handles 0 value hex', '\\0', (t, tree) => {
+    t.deepEqual(tree.nodes[0].nodes[0].value, String.fromCodePoint(0xFFFD));
+});
+
+test('handles lone surrogate value hex', '\\DBFF', (t, tree) => {
+    t.deepEqual(tree.nodes[0].nodes[0].value, String.fromCodePoint(0xFFFD));
+});
+
+test('handles out of bound values', '\\110000', (t, tree) => {
+    t.deepEqual(tree.nodes[0].nodes[0].value, String.fromCodePoint(0xFFFD));
+});
diff --git a/src/util/unesc.js b/src/util/unesc.js
@@ -25,9 +25,18 @@ function gobbleHex (str) {
     if (hex.length === 0) {
         return undefined;
     }
+    const codePoint = parseInt(hex, 16);
+
+    const isSurrogate = codePoint >= 0xD800 && codePoint <= 0xDFFF;
+    // Add special case for
+    // "If this number is zero, or is for a surrogate, or is greater than the maximum allowed code point"
+    // https://drafts.csswg.org/css-syntax/#maximum-allowed-code-point
+    if (isSurrogate || codePoint === 0x0000 || codePoint > 0x10FFFF) {
+        return ['\uFFFD', hex.length + (spaceTerminated ? 1 : 0)];
+    }
 
     return [
-        String.fromCodePoint(parseInt(hex, 16)),
+        String.fromCodePoint(codePoint),
         hex.length + (spaceTerminated ? 1 : 0),
     ];
 }
