Extend big-disjunction optimization to case-independent regexps

test/mjsunit/regexp-sort.js

Index: test/mjsunit/regexp-sort.js
diff --git a/test/mjsunit/regexp-sort.js b/test/mjsunit/regexp-sort.js
new file mode 100644
index 0000000000000000000000000000000000000000..9304fb4802b32d7cf760b953ae6b91851fd79c00
--- /dev/null
+++ b/test/mjsunit/regexp-sort.js
@@ -0,0 +1,40 @@
+// Copyright 2015 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.
+
+function Test(lower, upper) {
+  assertEquals(lower + "x", new RegExp(upper + "x|" + lower + "|" + lower + "cat", "i").exec(lower + "x") + "");
+  assertEquals(upper + "x", new RegExp(upper + "x|" + lower + "|" + lower + "cat", "i").exec(upper + "x") + "");
+  assertEquals(lower, new RegExp(lower + "|" + upper + "x|" + lower + "cat", "i").exec(lower + "x") + "");
+  assertEquals(upper, new RegExp(lower + "|" + upper + "x|" + lower + "cat", "i").exec(upper + "x") + "");
+}
+
+function TestFail(lower, upper) {
+  assertEquals(lower, new RegExp(upper + "x|" + lower + "|" + lower + "cat", "i").exec(lower + "x") + "");
+  assertEquals(upper + "x", new RegExp(upper + "x|" + lower + "|" + lower + "cat", "i").exec(upper + "x") + "");
+  assertEquals(lower, new RegExp(lower + "|" + upper + "x|" + lower + "cat", "i").exec(lower + "x") + "");
+  assertEquals(upper + "x", new RegExp(lower + "|" + upper + "x|" + lower + "cat", "i").exec(upper + "x") + "");
+}
+
+Test("a", "A");
+Test("0", "0");
+TestFail("a", "b");
+// Small and capital o-umlaut
+Test(String.fromCharCode(0xf6), String.fromCharCode(0xd6));
+// Small and capital kha.
+Test(String.fromCharCode(0x445), String.fromCharCode(0x425));
+// Small and capital y-umlaut.
+Test(String.fromCharCode(0xff), String.fromCharCode(0x178));
+// Small and large Greek mu.
+Test(String.fromCharCode(0x3bc), String.fromCharCode(0x39c));
+// Micron and large Greek mu.
+Test(String.fromCharCode(0xb5), String.fromCharCode(0x39c));
+// Micron and small Greek mu.
+Test(String.fromCharCode(0xb5), String.fromCharCode(0x3bc));

[Erik Corry Chromium.org, 2015/06/17 07:21:13]

Firefox fails this.  I think we are doing it right.
See https://bugzilla.mozilla.org/show_bug.cgi?id=1175416

+// German double s and capital S. These are not equivalent since one is double.
+TestFail(String.fromCharCode(0xdf), "S");
+// Small i and Turkish capital dotted I. These are not equivalent due to
+// 21.2.2.8.2 section 3g.  One is below 128 and the other is above 127.
+TestFail("i", String.fromCharCode(0x130));
+// Small dotless i and I. These are not equivalent either.
+TestFail(String.fromCharCode(0x131), "I");