Use ICU case conversion/transliterator for case conversion behind a flag

test/intl/general/case-mapping.js

Index: test/intl/general/case-mapping.js
diff --git a/test/intl/general/case-mapping.js b/test/intl/general/case-mapping.js
new file mode 100644
index 0000000000000000000000000000000000000000..a49a2b8846d919d7971f84e1807f0db84263a59f
--- /dev/null
+++ b/test/intl/general/case-mapping.js
@@ -0,0 +1,119 @@
+// 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: --icu_case_mapping

[jungshik at Google, 2016/04/29 06:23:16]

This does not work. --icu_case_mapping is not used by tools/run_tests.py. 
I have to pass --icu_case_mapping via '--extra-flags' option. 

./tools/run-tests.py --buildbot --extra-flags=--icu_case_mapping
--arch-and-mode=x64.Release test
262/intl402/String/*

+
+// Some edge cases that unibrow got wrong
+
+assertEquals("𐐘", "𐑀".toUpperCase());
+assertEquals("𐑀", "𐐘".toLowerCase());
+assertEquals("σ", "Σ".toLowerCase());
+
+// Some different paths in the ICU case conversion fastpath
+
+assertEquals("σς", "\u03A3\u03A3".toLowerCase());
+// Expand sharp s in latin1 fastpath
+assertEquals("ASSB", "A\u00DFB".toUpperCase());
+assertEquals("AB", "Ab".toUpperCase());
+// Find first upper case in fastpath
+assertEquals("ab", "aB".toLowerCase());
+assertEquals("AÜ", "aü".toUpperCase());
+assertEquals("AÜ", "AÜ".toUpperCase());
+assertEquals("aü", "aü".toLowerCase());
+assertEquals("aü", "AÜ".toLowerCase());
+assertEquals("aü", "AÜ".toLowerCase());
+
+// Starts with fastpath, but switches to full Unicode path
+// U+00FF is uppercased to U+0178.
+assertEquals("AŸ", "aÿ".toUpperCase());
+// U+00B5 (µ) is uppercased to U+039C (Μ)
+assertEquals("AΜ", "aµ".toUpperCase());
+
+// Buffer size increase
+assertEquals("CSSBẶ", "cßbặ".toUpperCase());
+assertEquals("FIFLFFIFFL", "\uFB01\uFB02\uFB03\uFB04".toUpperCase());
+
+// More comprehensive tests for "tr", "az" and "lt" are in
+// test262/intl402/Strings/*
+
+// Buffer size decrease with a single locale or locale list.
+// In Turkic (tr, az), U+0307 preceeded by Capital Letter I is dropped.
+assertEquals("abci", "aBcI\u0307".toLocaleLowerCase("tr"));
+assertEquals("abci", "aBcI\u0307".toLocaleLowerCase("az"));
+assertEquals("abci", "aBcI\u0307".toLocaleLowerCase(["tr", "en"]));
+
+// Cons string
+assertEquals("abcijkl", ("aBcI" + "\u0307jkl").toLocaleLowerCase("tr"));
+assertEquals("abcijkl",
+             ("aB" + "cI" + "\u0307j" + "kl").toLocaleLowerCase("tr"));
+assertEquals("abci\u0307jkl", ("aBcI" + "\u0307jkl").toLocaleLowerCase("en"));
+assertEquals("abci\u0307jkl",
+             ("aB" + "cI" + "\u0307j" + "kl").toLocaleLowerCase("en"));
+assertEquals("abci\u0307jkl", ("aBcI" + "\u0307jkl").toLowerCase());
+assertEquals("abci\u0307jkl",
+             ("aB" + "cI" + "\u0307j" + "kl").toLowerCase());
+
+// "tr" and "az" should behave identically.
+assertEquals("aBcI\u0307".toLocaleLowerCase("tr"),
+             "aBcI\u0307".toLocaleLowerCase("az"));
+// What matters is the first locale in the locale list.
+assertEquals("aBcI\u0307".toLocaleLowerCase(["tr", "en", "fr"]),
+             "aBcI\u0307".toLocaleLowerCase("tr"));
+assertEquals("aBcI\u0307".toLocaleLowerCase(["en", "tr", "az"]),
+             "aBcI\u0307".toLocaleLowerCase("en"));
+assertEquals("aBcI\u0307".toLocaleLowerCase(["en", "tr", "az"]),
+             "aBcI\u0307".toLowerCase());
+
+// An empty locale list is the same as the default locale. Try these tests
+// under Turkish and Greek locale.
+assertEquals("aBcI\u0307".toLocaleLowerCase([]),
+             "aBcI\u0307".toLocaleLowerCase());
+assertEquals("aBcI\u0307".toLocaleLowerCase([]),
+             "aBcI\u0307".toLocaleLowerCase(Intl.GetDefaultLocale));
+assertEquals("άόύώ".toLocaleUpperCase([]), "άόύώ".toLocaleUpperCase());
+assertEquals("άόύώ".toLocaleUpperCase([]),
+             "άόύώ".toLocaleUpperCase(Intl.GetDefaultLocale));
+
+
+// English/root locale keeps U+0307 (combining dot above).
+assertEquals("abci\u0307", "aBcI\u0307".toLocaleLowerCase("en"));
+assertEquals("abci\u0307", "aBcI\u0307".toLocaleLowerCase(["en", "tr"]));
+assertEquals("abci\u0307", "aBcI\u0307".toLowerCase());
+
+// Greek uppercasing: not covered by intl402/String/*, yet. Tonos (U+0301) and
+// other diacritic marks are dropped. This rule is based on the current CLDR's
+// el-Upper transformation, but Greek uppercasing rules are more sophisticated
+// than this. See http://bugs.icu-project.org/trac/ticket/10582 and
+// http://unicode.org/cldr/trac/ticket/7905 .
+assertEquals("Α", "α\u0301".toLocaleUpperCase("el"));
+assertEquals("Α", "α\u0301".toLocaleUpperCase("el-GR"));
+assertEquals("Α", "α\u0301".toLocaleUpperCase("el-Grek"));
+assertEquals("Α", "α\u0301".toLocaleUpperCase("el-Grek-GR"));
+assertEquals("Α", "ά".toLocaleUpperCase("el"));
+assertEquals("ΑΟΥΩ", "άόύώ".toLocaleUpperCase("el"));
+assertEquals("ΑΟΥΩ", "α\u0301ο\u0301υ\u0301ω\u0301".toLocaleUpperCase("el"));
+assertEquals("ΑΟΥΩ", "άόύώ".toLocaleUpperCase("el"));
+assertEquals("ΟΕ", "Ό\u1f15".toLocaleUpperCase("el"));
+assertEquals("ΟΕ", "Ο\u0301ε\u0314\u0301".toLocaleUpperCase("el"));
+
+// In other locales, U+0301 is preserved.
+assertEquals("Α\u0301Ο\u0301Υ\u0301Ω\u0301",
+             "α\u0301ο\u0301υ\u0301ω\u0301".toLocaleUpperCase("en"));
+assertEquals("Α\u0301Ο\u0301Υ\u0301Ω\u0301",
+             "α\u0301ο\u0301υ\u0301ω\u0301".toUpperCase());
+
+// Plane 1; Deseret and Warang Citi Script.
+assertEquals("\u{10400}\u{118A0}", "\u{10428}\u{118C0}".toUpperCase());
+assertEquals("\u{10428}\u{118C0}", "\u{10400}\u{118A0}".toLowerCase());
+// Mathematical Bold {Capital, Small} Letter A do not change.
+assertEquals("\u{1D400}\u{1D41A}", "\u{1D400}\u{1D41A}".toUpperCase());
+assertEquals("\u{1D400}\u{1D41A}", "\u{1D400}\u{1D41A}".toLowerCase());
+// Plane 1; New characters in Unicode 8.0
+assertEquals("\u{10C80}", "\u{10CC0}".toUpperCase());
+assertEquals("\u{10CC0}", "\u{10C80}".toLowerCase());
+assertEquals("\u{10C80}", "\u{10CC0}".toLocaleUpperCase());
+assertEquals("\u{10CC0}", "\u{10C80}".toLocaleLowerCase());
+assertEquals("\u{10C80}", "\u{10CC0}".toLocaleUpperCase(["tr"]));
+assertEquals("\u{10C80}", "\u{10CC0}".toLocaleUpperCase(["tr"]));
+assertEquals("\u{10CC0}", "\u{10C80}".toLocaleLowerCase());