fix: consume whitespace after hexadecimal escapes

src/__tests__/escapes.mjs

@@ -17,3 +17,29 @@ test("bare parens capture contents as a string", "(h1)", (t, tree) => {
   t.deepEqual(tree.nodes[0].nodes[0].value, "(h1)");
   t.deepEqual(tree.nodes[0].nodes[0].type, "string");
 });
+
+for (const hex of ["61", "000061"]) {
+  for (const whitespace of [" ", "\t", "\n", "\r", "\f", "\r\n"]) {
+    test(
+      `hex escape ${hex} terminated by ${JSON.stringify(whitespace)}`,
+      `[x="\\${hex}${whitespace}b"]`,
+      (t, tree) => {
+        t.is(tree.first.first.value, "ab");
+      },
+    );
+  }
+}
+
+for (const [value, expected] of [
+  ["\\000061  b", "a b"],
+  ["\\000000 b", "\uFFFDb"],
+  ["\\00D800 b", "\uFFFDb"],
+  ["\\110000 b", "\uFFFDb"],
+  ["\\000061b", "ab"],
+  ["\\000061\u00a0b", "a\u00a0b"],
+  ["\\000061\vb", "a\vb"],
+]) {
+  test(`hex escape boundary ${JSON.stringify(value)}`, `[x="${value}"]`, (t, tree) => {
+    t.is(tree.first.first.value, expected);
+  });
+}

src/util/unesc.js

@@ -9,13 +9,11 @@
 function gobbleHex(str) {
   const lower = str.toLowerCase();
   let hex = "";
-  let spaceTerminated = false;
   for (let i = 0; i < 6 && lower[i] !== undefined; i++) {
     const code = lower.charCodeAt(i);
     // check to see if we are dealing with a valid hex char [a-f|0-9]
     const valid = (code >= 97 && code <= 102) || (code >= 48 && code <= 57);
     // https://drafts.csswg.org/css-syntax/#consume-escaped-code-point
-    spaceTerminated = code === 32;
     if (!valid) {
       break;
     }
@@ -25,17 +23,23 @@ function gobbleHex(str) {
   if (hex.length === 0) {
     return undefined;
   }
+  let consumed = hex.length;
+  const next = str[consumed];
+  if (next === " " || next === "\t" || next === "\n" || next === "\r" || next === "\f") {
+    // CSS preprocessing treats CRLF as a single newline.
+    consumed += next === "\r" && str[consumed + 1] === "\n" ? 2 : 1;
+  }
   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 ["\uFFFD", consumed];
   }
 
-  return [String.fromCodePoint(codePoint), hex.length + (spaceTerminated ? 1 : 0)];
+  return [String.fromCodePoint(codePoint), consumed];
 }
 
 const CONTAINS_ESCAPE = /\\/;
@@ -49,7 +53,7 @@ export default function unesc(str) {
 
   for (let i = 0; i < str.length; i++) {
     if (str[i] === "\\") {
-      const gobbled = gobbleHex(str.slice(i + 1, i + 7));
+      const gobbled = gobbleHex(str.slice(i + 1, i + 9));
       if (gobbled !== undefined) {
         ret += gobbled[0];
         i += gobbled[1];