mirror of https://github.com/lukechilds/node.git
You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
200 lines
8.2 KiB
200 lines
8.2 KiB
15 years ago
|
// Copyright 2009 the V8 project authors. All rights reserved.
|
||
|
// Redistribution and use in source and binary forms, with or without
|
||
|
// modification, are permitted provided that the following conditions are
|
||
|
// met:
|
||
|
//
|
||
|
// * Redistributions of source code must retain the above copyright
|
||
|
// notice, this list of conditions and the following disclaimer.
|
||
|
// * Redistributions in binary form must reproduce the above
|
||
|
// copyright notice, this list of conditions and the following
|
||
|
// disclaimer in the documentation and/or other materials provided
|
||
|
// with the distribution.
|
||
|
// * Neither the name of Google Inc. nor the names of its
|
||
|
// contributors may be used to endorse or promote products derived
|
||
|
// from this software without specific prior written permission.
|
||
|
//
|
||
|
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
|
||
|
// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
|
||
|
// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
|
||
|
// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
|
||
|
// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
|
||
|
// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
|
||
|
// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
|
||
|
// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
|
||
|
// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
|
||
|
// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
|
||
|
// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
||
|
|
||
|
// Test Unicode character ranges in regexps.
|
||
|
|
||
|
|
||
|
// Cyrillic.
|
||
|
var cyrillic = {
|
||
|
FIRST: "\u0410", // A
|
||
|
first: "\u0430", // a
|
||
|
LAST: "\u042f", // YA
|
||
|
last: "\u044f", // ya
|
||
|
MIDDLE: "\u0427", // CHE
|
||
|
middle: "\u0447", // che
|
||
|
// Actually no characters are between the cases in Cyrillic.
|
||
|
BetweenCases: false};
|
||
|
|
||
|
var SIGMA = "\u03a3";
|
||
|
var sigma = "\u03c3";
|
||
|
var alternative_sigma = "\u03c2";
|
||
|
|
||
|
// Greek.
|
||
|
var greek = {
|
||
|
FIRST: "\u0391", // ALPHA
|
||
|
first: "\u03b1", // alpha
|
||
|
LAST: "\u03a9", // OMEGA
|
||
|
last: "\u03c9", // omega
|
||
|
MIDDLE: SIGMA, // SIGMA
|
||
|
middle: sigma, // sigma
|
||
|
// Epsilon acute is between ALPHA-OMEGA and alpha-omega, ie it
|
||
|
// is between OMEGA and alpha.
|
||
|
BetweenCases: "\u03ad"};
|
||
|
|
||
|
|
||
|
function Range(from, to, flags) {
|
||
|
return new RegExp("[" + from + "-" + to + "]", flags);
|
||
|
}
|
||
|
|
||
|
// Test Cyrillic and Greek separately.
|
||
|
for (var lang = 0; lang < 2; lang++) {
|
||
|
var chars = (lang == 0) ? cyrillic : greek;
|
||
|
|
||
|
for (var i = 0; i < 2; i++) {
|
||
|
var lc = (i == 0); // Lower case.
|
||
|
var first = lc ? chars.first : chars.FIRST;
|
||
|
var middle = lc ? chars.middle : chars.MIDDLE;
|
||
|
var last = lc ? chars.last : chars.LAST;
|
||
|
var first_other_case = lc ? chars.FIRST : chars.first;
|
||
|
var middle_other_case = lc ? chars.MIDDLE : chars.middle;
|
||
|
var last_other_case = lc ? chars.LAST : chars.last;
|
||
|
|
||
|
assertTrue(Range(first, last).test(first), 1);
|
||
|
assertTrue(Range(first, last).test(middle), 2);
|
||
|
assertTrue(Range(first, last).test(last), 3);
|
||
|
|
||
|
assertFalse(Range(first, last).test(first_other_case), 4);
|
||
|
assertFalse(Range(first, last).test(middle_other_case), 5);
|
||
|
assertFalse(Range(first, last).test(last_other_case), 6);
|
||
|
|
||
|
assertTrue(Range(first, last, "i").test(first), 7);
|
||
|
assertTrue(Range(first, last, "i").test(middle), 8);
|
||
|
assertTrue(Range(first, last, "i").test(last), 9);
|
||
|
|
||
|
assertTrue(Range(first, last, "i").test(first_other_case), 10);
|
||
|
assertTrue(Range(first, last, "i").test(middle_other_case), 11);
|
||
|
assertTrue(Range(first, last, "i").test(last_other_case), 12);
|
||
|
|
||
|
if (chars.BetweenCases) {
|
||
|
assertFalse(Range(first, last).test(chars.BetweenCases), 13);
|
||
|
assertFalse(Range(first, last, "i").test(chars.BetweenCases), 14);
|
||
|
}
|
||
|
}
|
||
|
if (chars.BetweenCases) {
|
||
|
assertTrue(Range(chars.FIRST, chars.last).test(chars.BetweenCases), 15);
|
||
|
assertTrue(Range(chars.FIRST, chars.last, "i").test(chars.BetweenCases), 16);
|
||
|
}
|
||
|
}
|
||
|
|
||
|
// Test range that covers both greek and cyrillic characters.
|
||
|
for (key in greek) {
|
||
|
assertTrue(Range(greek.FIRST, cyrillic.last).test(greek[key]), 17 + key);
|
||
|
if (cyrillic[key]) {
|
||
|
assertTrue(Range(greek.FIRST, cyrillic.last).test(cyrillic[key]), 18 + key);
|
||
|
}
|
||
|
}
|
||
|
|
||
|
for (var i = 0; i < 2; i++) {
|
||
|
var ignore_case = (i == 0);
|
||
|
var flag = ignore_case ? "i" : "";
|
||
|
assertTrue(Range(greek.first, cyrillic.LAST, flag).test(greek.first), 19);
|
||
|
assertTrue(Range(greek.first, cyrillic.LAST, flag).test(greek.middle), 20);
|
||
|
assertTrue(Range(greek.first, cyrillic.LAST, flag).test(greek.last), 21);
|
||
|
|
||
|
assertTrue(Range(greek.first, cyrillic.LAST, flag).test(cyrillic.FIRST), 22);
|
||
|
assertTrue(Range(greek.first, cyrillic.LAST, flag).test(cyrillic.MIDDLE), 23);
|
||
|
assertTrue(Range(greek.first, cyrillic.LAST, flag).test(cyrillic.LAST), 24);
|
||
|
|
||
|
// A range that covers the lower case greek letters and the upper case cyrillic
|
||
|
// letters.
|
||
|
assertEquals(ignore_case, Range(greek.first, cyrillic.LAST, flag).test(greek.FIRST), 25);
|
||
|
assertEquals(ignore_case, Range(greek.first, cyrillic.LAST, flag).test(greek.MIDDLE), 26);
|
||
|
assertEquals(ignore_case, Range(greek.first, cyrillic.LAST, flag).test(greek.LAST), 27);
|
||
|
|
||
|
assertEquals(ignore_case, Range(greek.first, cyrillic.LAST, flag).test(cyrillic.first), 28);
|
||
|
assertEquals(ignore_case, Range(greek.first, cyrillic.LAST, flag).test(cyrillic.middle), 29);
|
||
|
assertEquals(ignore_case, Range(greek.first, cyrillic.LAST, flag).test(cyrillic.last), 30);
|
||
|
}
|
||
|
|
||
|
|
||
|
// Sigma is special because there are two lower case versions of the same upper
|
||
|
// case character. JS requires that case independece means that you should
|
||
|
// convert everything to upper case, so the two sigma variants are equal to each
|
||
|
// other in a case independt comparison.
|
||
|
for (var i = 0; i < 2; i++) {
|
||
|
var simple = (i != 0);
|
||
|
var name = simple ? "" : "[]";
|
||
|
var regex = simple ? SIGMA : "[" + SIGMA + "]";
|
||
|
|
||
|
assertFalse(new RegExp(regex).test(sigma), 31 + name);
|
||
|
assertFalse(new RegExp(regex).test(alternative_sigma), 32 + name);
|
||
|
assertTrue(new RegExp(regex).test(SIGMA), 33 + name);
|
||
|
|
||
|
assertTrue(new RegExp(regex, "i").test(sigma), 34 + name);
|
||
|
// JSC and Tracemonkey fail this one.
|
||
|
assertTrue(new RegExp(regex, "i").test(alternative_sigma), 35 + name);
|
||
|
assertTrue(new RegExp(regex, "i").test(SIGMA), 36 + name);
|
||
|
|
||
|
regex = simple ? sigma : "[" + sigma + "]";
|
||
|
|
||
|
assertTrue(new RegExp(regex).test(sigma), 41 + name);
|
||
|
assertFalse(new RegExp(regex).test(alternative_sigma), 42 + name);
|
||
|
assertFalse(new RegExp(regex).test(SIGMA), 43 + name);
|
||
|
|
||
|
assertTrue(new RegExp(regex, "i").test(sigma), 44 + name);
|
||
|
// JSC and Tracemonkey fail this one.
|
||
|
assertTrue(new RegExp(regex, "i").test(alternative_sigma), 45 + name);
|
||
|
assertTrue(new RegExp(regex, "i").test(SIGMA), 46 + name);
|
||
|
|
||
|
regex = simple ? alternative_sigma : "[" + alternative_sigma + "]";
|
||
|
|
||
|
assertFalse(new RegExp(regex).test(sigma), 51 + name);
|
||
|
assertTrue(new RegExp(regex).test(alternative_sigma), 52 + name);
|
||
|
assertFalse(new RegExp(regex).test(SIGMA), 53 + name);
|
||
|
|
||
|
// JSC and Tracemonkey fail this one.
|
||
|
assertTrue(new RegExp(regex, "i").test(sigma), 54 + name);
|
||
|
assertTrue(new RegExp(regex, "i").test(alternative_sigma), 55 + name);
|
||
|
// JSC and Tracemonkey fail this one.
|
||
|
assertTrue(new RegExp(regex, "i").test(SIGMA), 56 + name);
|
||
|
}
|
||
|
|
||
|
|
||
|
for (var add_non_ascii_character_to_subject = 0;
|
||
|
add_non_ascii_character_to_subject < 2;
|
||
|
add_non_ascii_character_to_subject++) {
|
||
|
var suffix = add_non_ascii_character_to_subject ? "\ufffe" : "";
|
||
|
// A range that covers both ASCII and non-ASCII.
|
||
|
for (var i = 0; i < 2; i++) {
|
||
|
var full = (i != 0);
|
||
|
var mixed = full ? "[a-\uffff]" : "[a-" + cyrillic.LAST + "]";
|
||
|
var f = full ? "f" : "c";
|
||
|
for (var j = 0; j < 2; j++) {
|
||
|
var ignore_case = (j == 0);
|
||
|
var flag = ignore_case ? "i" : "";
|
||
|
var re = new RegExp(mixed, flag);
|
||
|
assertEquals(ignore_case || (full && add_non_ascii_character_to_subject),
|
||
|
re.test("A" + suffix),
|
||
|
58 + flag + f);
|
||
|
assertTrue(re.test("a" + suffix), 59 + flag + f);
|
||
|
assertTrue(re.test("~" + suffix), 60 + flag + f);
|
||
|
assertTrue(re.test(cyrillic.MIDDLE), 61 + flag + f);
|
||
|
assertEquals(ignore_case || full, re.test(cyrillic.middle), 62 + flag + f);
|
||
|
}
|
||
|
}
|
||
|
}
|