v8/test/mjsunit/regexp-capture-3.js

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

219 lines
7.8 KiB
JavaScript
Raw Normal View History

// Copyright 2012 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.
function oneMatch(re) {
"abcd".replace(re, function() { });
assertEquals("abcd", RegExp.input);
assertEquals("a", RegExp.leftContext);
assertEquals("b", RegExp.lastMatch);
assertEquals("", RegExp.lastParen);
assertEquals(undefined, RegExp.lastIndex);
assertEquals(undefined, RegExp.index);
assertEquals("cd", RegExp.rightContext);
for (var i = 1; i < 10; i++) {
assertEquals("", RegExp['$' + i]);
}
}
oneMatch(/b/);
oneMatch(/b/g);
"abcdabcd".replace(/b/g, function() { });
assertEquals("abcdabcd", RegExp.input);
assertEquals("abcda", RegExp.leftContext);
assertEquals("b", RegExp.lastMatch);
assertEquals("", RegExp.lastParen);
assertEquals(undefined, RegExp.lastIndex);
assertEquals(undefined, RegExp.index);
assertEquals("cd", RegExp.rightContext);
for (var i = 1; i < 10; i++) {
assertEquals("", RegExp['$' + i]);
}
function captureMatch(re) {
"abcd".replace(re, function() { });
assertEquals("abcd", RegExp.input);
assertEquals("a", RegExp.leftContext);
assertEquals("bc", RegExp.lastMatch);
assertEquals("c", RegExp.lastParen);
assertEquals(undefined, RegExp.lastIndex);
assertEquals(undefined, RegExp.index);
assertEquals("d", RegExp.rightContext);
assertEquals('b', RegExp.$1);
assertEquals('c', RegExp.$2);
for (var i = 3; i < 10; i++) {
assertEquals("", RegExp['$' + i]);
}
}
captureMatch(/(b)(c)/);
captureMatch(/(b)(c)/g);
"abcdabcd".replace(/(b)(c)/g, function() { });
assertEquals("abcdabcd", RegExp.input);
assertEquals("abcda", RegExp.leftContext);
assertEquals("bc", RegExp.lastMatch);
assertEquals("c", RegExp.lastParen);
assertEquals(undefined, RegExp.lastIndex);
assertEquals(undefined, RegExp.index);
assertEquals("d", RegExp.rightContext);
assertEquals('b', RegExp.$1);
assertEquals('c', RegExp.$2);
for (var i = 3; i < 10; i++) {
assertEquals("", RegExp['$' + i]);
}
function Override() {
// Set the internal lastMatchInfoOverride. After calling this we do a normal
// match and verify the override was cleared and that we record the new
// captures.
"abcdabcd".replace(/(b)(c)/g, function() { });
}
function TestOverride(input, expect, property, re_src) {
var re = new RegExp(re_src);
var re_g = new RegExp(re_src, "g");
function OverrideCase(fn) {
Override();
fn();
assertEquals(expect, RegExp[property]);
}
OverrideCase(function() { return input.replace(re, "x"); });
OverrideCase(function() { return input.replace(re_g, "x"); });
OverrideCase(function() { return input.replace(re, ""); });
OverrideCase(function() { return input.replace(re_g, ""); });
OverrideCase(function() { return input.match(re); });
OverrideCase(function() { return input.match(re_g); });
OverrideCase(function() { return re.test(input); });
OverrideCase(function() { return re_g.test(input); });
}
var input = "bar.foo baz......";
var re_str = "(ba.).*?f";
TestOverride(input, "bar", "$1", re_str);
input = "foo bar baz";
var re_str = "bar";
TestOverride(input, "bar", "$&", re_str);
function no_last_match(fn) {
fn();
assertEquals("hestfisk", RegExp.$1);
}
/(hestfisk)/.test("There's no such thing as a hestfisk!");
no_last_match(function() { "foo".replace("f", ""); });
no_last_match(function() { "foo".replace("f", "f"); });
no_last_match(function() { "foo".split("o"); });
var base = "In the music. In the music. ";
var cons = base + base + base + base;
no_last_match(function() { cons.replace("x", "y"); });
no_last_match(function() { cons.replace("e", "E"); });
// Here's one that matches once, then tries to match again, but fails.
// Verify that the last match info is from the last match, not from the
// failure that came after.
"bar.foo baz......".replace(/(ba.).*?f/g, function() { return "x";});
assertEquals("bar", RegExp.$1);
// A test that initially does a zero width match, but later does a non-zero
// width match.
var a = "foo bar baz".replace(/^|bar/g, "");
assertEquals("foo baz", a);
a = "foo bar baz".replace(/^|bar/g, "*");
assertEquals("*foo * baz", a);
// We test FilterASCII using regexps that will backtrack forever. Since
// a regexp with a non-ASCII character in it can never match an ASCII
// string we can test that the relevant node is removed by verifying that
// there is no hang.
function NoHang(re) {
"This is an ASCII string that could take forever".match(re);
}
NoHang(/(((.*)*)*x)Ā/); // Continuation after loop is filtered, so is loop.
NoHang(/(((.*)*)*Ā)foo/); // Body of loop filtered.
NoHang(/Ā(((.*)*)*x)/); // Everything after a filtered character is filtered.
NoHang(/(((.*)*)*x)Ā/); // Everything before a filtered character is filtered.
NoHang(/[ćăĀ](((.*)*)*x)/); // Everything after a filtered class is filtered.
NoHang(/(((.*)*)*x)[ćăĀ]/); // Everything before a filtered class is filtered.
NoHang(/[^\x00-\xff](((.*)*)*x)/); // After negated class.
NoHang(/(((.*)*)*x)[^\x00-\xff]/); // Before negated class.
NoHang(/(?!(((.*)*)*x)Ā)foo/); // Negative lookahead is filtered.
NoHang(/(?!(((.*)*)*x))Ā/); // Continuation branch of negative lookahead.
NoHang(/(?=(((.*)*)*x)Ā)foo/); // Positive lookahead is filtered.
NoHang(/(?=(((.*)*)*x))Ā/); // Continuation branch of positive lookahead.
NoHang(/(?=Ā)(((.*)*)*x)/); // Positive lookahead also prunes continuation.
NoHang(/(æ|ø|Ā)(((.*)*)*x)/); // All branches of alternation are filtered.
NoHang(/(a|b|(((.*)*)*x))Ā/); // 1 out of 3 branches pruned.
NoHang(/(a|(((.*)*)*x)ă|(((.*)*)*x)Ā)/); // 2 out of 3 branches pruned.
var s = "Don't prune based on a repetition of length 0";
assertEquals(null, s.match(/å{1,1}prune/));
assertEquals("prune", (s.match(/å{0,0}prune/)[0]));
// Some very deep regexps where FilterASCII gives up in order not to make the
// stack overflow.
var regex6 = /a*\u0100*\w/;
var input0 = "a";
regex6.exec(input0);
var re = "\u0100*\\w";
for (var i = 0; i < 200; i++) re = "a*" + re;
var regex7 = new RegExp(re);
regex7.exec(input0);
var regex8 = new RegExp(re, "i");
regex8.exec(input0);
re = "[\u0100]*\\w";
for (var i = 0; i < 200; i++) re = "a*" + re;
var regex9 = new RegExp(re);
regex9.exec(input0);
var regex10 = new RegExp(re, "i");
regex10.exec(input0);
var regex11 = /^(?:[^\u0000-\u0080]|[0-9a-z?,.!&\s#()])+$/i;
regex11.exec(input0);
var regex12 = /u(\xf0{8}?\D*?|( ? !)$h??(|)*?(||)+?\6((?:\W\B|--\d-*-|)?$){0, }?|^Y( ? !1)\d+)+a/;
regex12.exec("");