Optimize case conversion with icu_case_mapping

test/intl/general/case-mapping.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: --icu_case_mapping
 
 // 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
+// Find first uppercase in fastpath
+// Input length < a machine word size
+assertEquals("ab", "ab".toLowerCase());
 assertEquals("ab", "aB".toLowerCase());
 assertEquals("AÜ", "aü".toUpperCase());
 assertEquals("AÜ", "AÜ".toUpperCase());
 assertEquals("aü", "aü".toLowerCase());
+assertEquals("aü", "aÜ".toLowerCase());
 assertEquals("aü", "AÜ".toLowerCase());
 assertEquals("aü", "AÜ".toLowerCase());
 
+// Input length >= a machine word size
+assertEquals("abcdefghij", "abcdefghij".toLowerCase());
+assertEquals("abcdefghij", "abcdefghiJ".toLowerCase());
+assertEquals("abçdefghij", "abçdefghiJ".toLowerCase());
+assertEquals("abçdefghij", "abÇdefghiJ".toLowerCase());
+assertEquals("abcdefghiá", "abcdeFghiá".toLowerCase());
+assertEquals("abcdefghiá", "abcdeFghiÁ".toLowerCase());
+
+assertEquals("ABCDEFGHIJ", "ABCDEFGHIJ".toUpperCase());
+assertEquals("ABCDEFGHIJ", "ABCDEFGHIj".toUpperCase());
+assertEquals("ABÇDEFGHIJ", "ABÇDEFGHIj".toUpperCase());
+assertEquals("ABÇDEFGHIJ", "ABçDEFGHIj".toUpperCase());
+assertEquals("ABCDEFGHIÁ", "ABCDEfGHIÁ".toUpperCase());
+assertEquals("ABCDEFGHIÁ", "ABCDEfGHIá".toUpperCase());
+
+
 // 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());
+assertEquals("ABCÀCSSA", "abcàcßa".toUpperCase());
+assertEquals("ABCDEFGHIÀCSSA", "ABCDEFGHIàcßa".toUpperCase());
+assertEquals("ABCDEFGHIÀCSSA", "abcdeFghiàcßa".toUpperCase());
+
 // OneByte input with buffer size increase: non-fast path
 assertEquals("ABCSS", "abCß".toLocaleUpperCase("tr"));
 
 // 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"));
@@ skipping 111 matching lines @@
 assertEquals(latin1SupplLowercased, latin1Suppl.toLocaleLowerCase("tr"));
 assertEquals(latin1SupplUppercased, latin1Suppl.toLocaleUpperCase("az"));
 assertEquals(latin1SupplLowercased, latin1Suppl.toLocaleLowerCase("az"));
 assertEquals(latin1SupplUppercased, latin1Suppl.toLocaleUpperCase("lt"));
 // Lithuanian need to have a dot-above for U+00CC(Ì) and U+00CD(Í) when
 // lowercasing.
 assertEquals("\u00A0¡¢£¤¥¦§¨©ª«¬\u00AD®°±²³´µ¶·¸¹º»¼½¾¿" +
     "àáâãäåæçèéêëi\u0307\u0300i\u0307\u0301îïðñòóôõö×øùúûüýþß" +
     "àáâãäåæçèéêëìíîïðñòóôõö÷øùúûüýþÿ",
     latin1Suppl.toLocaleLowerCase("lt"));