[regexp] Updates for unicode escapes in capture names

Update docs and tests for recent changes in the spec for unicode escapes
in capture group names.

https://github.com/tc39/proposal-regexp-named-groups/issues/23

BUG=v8:5437

Review-Url: https://codereview.chromium.org/2788423003
Cr-Commit-Position: refs/heads/master@{#44474}

src/regexp/regexp-parser.cc

@@ -700,10 +700,6 @@ void RegExpParser::ScanForCaptures() {
           if (current() != '<') break;
 
           if (FLAG_harmony_regexp_lookbehind) {
-            // TODO(jgruber): To be more future-proof we could test for
-            // IdentifierStart here once it becomes clear whether group names
-            // allow unicode escapes.
-            // https://github.com/tc39/proposal-regexp-named-groups/issues/23
             Advance();
             if (current() == '=' || current() == '!') break;
           }
@@ -783,8 +779,6 @@ const ZoneVector<uc16>* RegExpParser::ParseCaptureGroupName() {
 
     // Convert unicode escapes.
     if (c == '\\' && current() == 'u') {
-      // TODO(jgruber): Reconsider this once the spec has settled.
-      // https://github.com/tc39/proposal-regexp-named-groups/issues/23
       Advance(scan_mode);
       if (!ParseUnicodeEscape(&c)) {
         ReportError(CStrVector("Invalid Unicode escape sequence"));

test/mjsunit/harmony/regexp-named-captures.js

@@ -187,6 +187,31 @@ assertThrows("/\\1(?<=a)./u", SyntaxError);
 assertThrows("/\\1(?<!a)./u", SyntaxError);
 assertEquals(["a", "a"], /\1(?<a>.)/u.exec("abcd"));
 
+// Unicode escapes in capture names.
+assertTrue(/(?<a\uD801\uDCA4>.)/u.test("a"));  // \u Lead \u Trail
+assertThrows("/(?<a\\uD801>.)/u", SyntaxError);  // \u Lead
+assertThrows("/(?<a\\uDCA4>.)/u", SyntaxError);  // \u Trail
+assertTrue(/(?<\u0041>.)/u.test("a"));  // \u NonSurrogate
+assertTrue(/(?<\u{0041}>.)/u.test("a"));  // \u{ Non-surrogate }
+assertTrue(/(?<a\u{104A4}>.)/u.test("a"));  // \u{ Surrogate, ID_Continue }
+assertThrows("/(?<a\\u{110000}>.)/u", SyntaxError);  // \u{ Out-of-bounds }
+assertThrows("/(?<a\uD801>.)/u", SyntaxError);  // Lead
+assertThrows("/(?<a\uDCA4>.)/u", SyntaxError);  // Trail
+assertTrue(RegExp("(?<\u{0041}>.)", "u").test("a"));  // Non-surrogate
+assertTrue(RegExp("(?<a\u{104A4}>.)", "u").test("a"));  // Surrogate,ID_Continue
+
+assertThrows("/(?<a\\uD801\uDCA4>.)/", SyntaxError);
+assertThrows("/(?<a\\uD801>.)/", SyntaxError);
+assertThrows("/(?<a\\uDCA4>.)/", SyntaxError);
+assertTrue(/(?<\u0041>.)/.test("a"));
+assertThrows("/(?<\\u{0041}>.)/", SyntaxError);
+assertThrows("/(?<a\\u{104A4}>.)/", SyntaxError);
+assertThrows("/(?<a\\u{10FFFF}>.)/", SyntaxError);
+assertThrows("/(?<a\uD801>.)/", SyntaxError);  // Lead
+assertThrows("/(?<a\uDCA4>.)/", SyntaxError);  // Trail
+assertTrue(RegExp("(?<\u{0041}>.)").test("a"));  // Non-surrogate
+assertTrue(RegExp("(?<a\u{104A4}>.)").test("a"));  // Surrogate, ID_Continue
+
 // @@replace with a callable replacement argument (no named captures).
 {
   let result = "abcd".replace(/(.)(.)/u, (match, fst, snd, offset, str) => {