[regexp] implement character classes for unicode regexps.

test/mjsunit/harmony/unicode-character-ranges.js

+// Copyright 2016 the V8 project authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+// Flags: --harmony-unicode-regexps --harmony-regexp-lookbehind
+
+function testl(expectation, regexp, subject) {
+  if (regexp instanceof String) regexp = new RegExp(regexp, "u");
+  assertEquals(expectation, regexp.exec(subject));
+}
+
+function tests(expectation, string, subject) {

[erikcorry, 2016/01/20 13:48:51]

string -> source or regexp_source

[Yang, 2016/01/20 14:04:14]

Done.

+  testl(expectation, new RegExp(string, "u"), subject);
+}
+
+// Character ranges.
+testl(["A"], /[A-D]/u, "A");
+tests(["A"], "[A-D]", "A");
+testl(["ABCD"], /[A-D]+/u, "ZABCDEF");
+tests(["ABCD"], "[A-D]+", "ZABCDEF");
+
+testl(["\u{12345}"], /[\u1234-\u{12345}]/u, "\u{12345}");
+tests(["\u{12345}"], "[\u1234-\u{12345}]", "\u{12345}");
+testl(null, /[^\u1234-\u{12345}]/u, "\u{12345}");
+tests(null, "[^\u1234-\u{12345}]", "\u{12345}");
+
+testl(["\u{1234}"], /[\u1234-\u{12345}]/u, "\u{1234}");
+tests(["\u{1234}"], "[\u1234-\u{12345}]", "\u{1234}");
+testl(null, /[^\u1234-\u{12345}]/u, "\u{1234}");
+tests(null, "[^\u1234-\u{12345}]", "\u{1234}");
+
+testl(null, /[\u1234-\u{12345}]/u, "\u{1233}");
+tests(null, "[\u1234-\u{12345}]", "\u{1233}");
+testl(["\u{1233}"], /[^\u1234-\u{12345}]/u, "\u{1233}");
+tests(["\u{1233}"], "[^\u1234-\u{12345}]", "\u{1233}");
+
+testl(["\u{12346}"], /[^\u1234-\u{12345}]/u, "\u{12346}");
+tests(["\u{12346}"], "[^\u1234-\u{12345}]", "\u{12346}");
+testl(null, /[\u1234-\u{12345}]/u, "\u{12346}");
+tests(null, "[\u1234-\u{12345}]", "\u{12346}");
+
+testl(["\u{12342}"], /[\u{12340}-\u{12345}]/u, "\u{12342}");
+tests(["\u{12342}"], "[\u{12340}-\u{12345}]", "\u{12342}");
+testl(null, /[^\u{12340}-\u{12345}]/u, "\u{12342}");
+tests(null, "[^\u{12340}-\u{12345}]", "\u{12342}");
+
+testl(["\u{ffff}"], /[\u{ff80}-\u{12345}]/u, "\u{ffff}");
+tests(["\u{ffff}"], "[\u{ff80}-\u{12345}]", "\u{ffff}");
+testl(null, /[^\u{ff80}-\u{12345}]/u, "\u{ffff}");
+tests(null, "[^\u{ff80}-\u{12345}]", "\u{ffff}");
+
+// Lone surrogate
+testl(["\ud800"], /[^\u{ff80}-\u{12345}]/u, "\uff99\u{d800}A");
+tests(["\udc00"], "[^\u{ff80}-\u{12345}]", "\uff99\u{dc00}A");
+testl(["\udc01"], /[\u0100-\u{10ffff}]/u, "A\udc01");
+testl(["\udc03"], /[\udc01-\udc03]/u, "\ud801\udc02\udc03");
+testl(["\ud801"], /[\ud801-\ud803]/u, "\ud802\udc01\ud801");
+
+// Paired sorrogate.
+testl(null, /[^\u{ff80}-\u{12345}]/u, "\u{d800}\u{dc00}");
+tests(null, "[^\u{ff80}-\u{12345}]", "\u{d800}\u{dc00}");
+testl(["\ud800\udc00"], /[\u{ff80}-\u{12345}]/u, "\u{d800}\u{dc00}");
+tests(["\ud800\udc00"], "[\u{ff80}-\u{12345}]", "\u{d800}\u{dc00}");
+testl(["foo\u{10e6d}bar"], /foo\ud803\ude6dbar/u, "foo\u{10e6d}bar");
+
+// Lone surrogates
+testl(["\ud801\ud801"], /\ud801+/u, "\ud801\udc01\ud801\ud801");
+testl(["\udc01\udc01"], /\udc01+/u, "\ud801\ud801\udc01\udc01\udc01");
+
+testl(["\udc02\udc03A"], /\W\WA/u, "\ud801\udc01A\udc02\udc03A");
+testl(["\ud801\ud802"], /\ud801./u, "\ud801\udc01\ud801\ud802");
+testl(["\udc02\udc03A"], /[\ud800-\udfff][\ud800-\udfff]A/u,
+      "\ud801\udc01A\udc02\udc03A");
+
+// Character classes
+testl(null, /\w/u, "\ud801\udc01");
+testl(["\ud801"], /[^\w]/, "\ud801\udc01");
+testl(["\ud801\udc01"], /[^\w]/u, "\ud801\udc01");
+testl(["\ud801"], /\W/, "\ud801\udc01");
+testl(["\ud801\udc01"], /\W/u, "\ud801\udc01");
+
+testl(["\ud800X"], /.X/u, "\ud800XaX");
+testl(["aX"], /.(?<!\ud800)X/u, "\ud800XaX");
+testl(["aX"], /.(?<![\ud800-\ud900])X/u, "\ud800XaX");
+
+testl(null, /[]/u, "\u1234");
+testl(["0abc"], /[^]abc/u, "0abc");
+testl(["\u1234abc"], /[^]abc/u, "\u1234abc");
+testl(["\u{12345}abc"], /[^]abc/u, "\u{12345}abc");
+
+// Backward matches of lone surrogates.
+testl(["B", "\ud803A"], /(?<=([\ud800-\ud900]A))B/u,
+      "\ud801\udc00AB\udc00AB\ud802\ud803AB");

[erikcorry, 2016/01/20 13:48:51]

This is saying that if we try to match lone-surrogate-followed by A it should
match the \udc00 even though it is not lone?

[Yang, 2016/01/20 14:04:14]

This is just (a random test) to check that character ranges inside a lookbehind
works. I added another one.

+testl(["B", "\udc11A"], /(?<=([\udc00-\udd00]A))B/u,
+      "\ud801\udc00AB\udc11AB\ud802\ud803AB");
+testl(["X", "\ud800C"], /(?<=(\ud800\w))X/u,
+      "\ud800\udc00AX\udc11BX\ud800\ud800CX");
+testl(["C", "\ud800\ud800"], /(?<=(\ud800.))\w/u,
+      "\ud800\udc00AX\udc11BX\ud800\ud800CX");
+testl(["X", "\udc01C"], /(?<=(\udc01\w))X/u,
+      "\ud800\udc01AX\udc11BX\udc01\udc01CX");
+testl(["C", "\udc01\udc01"], /(?<=(\udc01.))./u,
+      "\ud800\udc01AX\udc11BX\udc01\udc01CX");