mirror of
https://github.com/LadybirdBrowser/ladybird
synced 2026-04-26 09:45:06 +02:00
535d2476a7
This introduces a new mechanism for evaluating lookbehind assertions by adding four new bytecode opcodes: SetStepBack, IncStepBack, CheckStepBack, and CheckSavedPosition. These opcodes replace the previous GoBack-based approach and enables correct handling of variable-length lookbehind patterns, where the match length cannot be known statically. Track lookbehind greediness in the parser and propagate it to bytecode generation. Allow controlled backtracking in lookbehind bodies while avoiding incorrect captures during step-back execution. Partially fix issue: #3459
207 lines
12 KiB
HTML
207 lines
12 KiB
HTML
<!DOCTYPE html>
|
|
<meta charset="utf-8">
|
|
<!--
|
|
Copyright (C) 2017 the V8 project authors. All rights reserved.
|
|
This code is governed by the BSD license found in the LICENSE file.
|
|
-->
|
|
<title>JavaScript RegExp lookbehind assertions: alternations</title>
|
|
<script src="../resources/testharness.js"></script>
|
|
<script src="../resources/testharnessreport.js"></script>
|
|
|
|
<script>
|
|
test(t => {
|
|
// https://github.com/tc39/test262/blob/main/test/built-ins/RegExp/lookBehind/alternations.js
|
|
assert_array_equals("xabcd".match(/.*(?<=(..|...|....))(.*)/), ["xabcd", "cd", ""], "#1");
|
|
assert_array_equals("xabcd".match(/.*(?<=(xx|...|....))(.*)/), ["xabcd", "bcd", ""], "#2");
|
|
assert_array_equals("xxabcd".match(/.*(?<=(xx|...))(.*)/), ["xxabcd", "bcd", ""], "#3");
|
|
assert_array_equals("xxabcd".match(/.*(?<=(xx|xxx))(.*)/), ["xxabcd", "xx", "abcd"], "#4");
|
|
}, "Alternations are tried left to right, with no backtracking into a lookbehind");
|
|
|
|
test(t => {
|
|
// https://github.com/tc39/test262/blob/main/test/built-ins/RegExp/lookBehind/back-references.js
|
|
assert_array_equals("abb".match(/(.)(?<=(\1\1))/), ["b", "b", "bb"], "#1");
|
|
assert_array_equals("abB".match(/(.)(?<=(\1\1))/i), ["B", "B", "bB"], "#2");
|
|
assert_array_equals("aabAaBa".match(/((\w)\w)(?<=\1\2\1)/i), ["aB", "aB", "a"], "#3");
|
|
assert_array_equals("aabAaBa".match(/(\w(\w))(?<=\1\2\1)/i), ["Ba", "Ba", "a"], "#4");
|
|
assert_array_equals("abaBbAa".match(/(?=(\w))(?<=(\1))./i), ["b", "b", "B"], "#5");
|
|
assert_array_equals(" 'foo' ".match(/(?<=(.))(\w+)(?=\1)/), ["foo", "'", "foo"], "#6");
|
|
assert_array_equals(" \"foo\" ".match(/(?<=(.))(\w+)(?=\1)/), ["foo", "\"", "foo"], "#7");
|
|
assert_array_equals("abbb".match(/(.)(?<=\1\1\1)/), ["b", "b"], "#8");
|
|
assert_array_equals("fababab".match(/(..)(?<=\1\1\1)/), ["ab", "ab"], "#9");
|
|
assert_equals(" .foo\" ".match(/(?<=(.))(\w+)(?=\1)/), null, "#10");
|
|
assert_equals("ab".match(/(.)(?<=\1\1\1)/), null, "#11");
|
|
assert_equals("abb".match(/(.)(?<=\1\1\1)/), null, "#12");
|
|
assert_equals("ab".match(/(..)(?<=\1\1\1)/), null, "#13");
|
|
assert_equals("abb".match(/(..)(?<=\1\1\1)/), null, "#14");
|
|
assert_equals("aabb".match(/(..)(?<=\1\1\1)/), null, "#15");
|
|
assert_equals("abab".match(/(..)(?<=\1\1\1)/), null, "#16");
|
|
assert_equals("fabxbab".match(/(..)(?<=\1\1\1)/), null, "#17");
|
|
assert_equals("faxabab".match(/(..)(?<=\1\1\1)/), null, "#18");
|
|
}, "Back-references");
|
|
|
|
test(t => {
|
|
// https://github.com/tc39/test262/blob/main/test/built-ins/RegExp/lookBehind/back-references-to-captures.js
|
|
assert_array_equals("abcCd".match(/(?<=\1(\w))d/i), ["d", "C"], "#1");
|
|
assert_array_equals("abxxd".match(/(?<=\1([abx]))d/), ["d", "x"], "#2");
|
|
assert_array_equals("ababc".match(/(?<=\1(\w+))c/), ["c", "ab"], "#3");
|
|
assert_array_equals("ababbc".match(/(?<=\1(\w+))c/), ["c", "b"], "#4");
|
|
assert_equals("ababdc".match(/(?<=\1(\w+))c/), null, "#5");
|
|
assert_array_equals("ababc".match(/(?<=(\w+)\1)c/), ["c", "abab"], "#6");
|
|
}, "Back-references to captures inside the lookbehind");
|
|
|
|
test(t => {
|
|
// https://github.com/tc39/test262/blob/main/test/built-ins/RegExp/lookBehind/captures.js
|
|
const str = "abcdef";
|
|
assert_array_equals(str.match(/(?<=(c))def/), ["def", "c"], "#1");
|
|
assert_array_equals(str.match(/(?<=(\w{2}))def/), ["def", "bc"], "#2");
|
|
assert_array_equals(str.match(/(?<=(\w(\w)))def/), ["def", "bc", "c"], "#3");
|
|
assert_array_equals(str.match(/(?<=(\w){3})def/), ["def", "a"], "#4");
|
|
assert_array_equals(str.match(/(?<=(bc)|(cd))./), ["d", "bc", undefined], "#5");
|
|
assert_array_equals(str.match(/(?<=([ab]{1,2})\D|(abc))\w/), ["c", "a", undefined], "#6");
|
|
assert_array_equals(str.match(/\D(?<=([ab]+))(\w)/), ["ab", "a", "b"], "#7");
|
|
assert_array_equals(str.match(/(?<=b|c)\w/g), ["c", "d"], "#8");
|
|
assert_array_equals(str.match(/(?<=[b-e])\w{2}/g), ["cd", "ef"], "#9");
|
|
}, "Capturing matches");
|
|
|
|
test(t => {
|
|
// https://github.com/tc39/test262/blob/main/test/built-ins/RegExp/lookBehind/captures-negative.js
|
|
assert_array_equals("abcdef".match(/(?<!(^|[ab]))\w{2}/), ["de", undefined]);
|
|
}, "Captures inside negative lookbehind");
|
|
|
|
test(t => {
|
|
// https://github.com/tc39/test262/blob/main/test/built-ins/RegExp/lookBehind/do-not-backtrack.js
|
|
// The lookbehind captures "abc" so that \1 does not match. We do not backtrack
|
|
// to capture only "bc" in the lookbehind.
|
|
assert_equals("abcdbc".match(/(?<=([abc]+)).\1/), null);
|
|
}, "Do not backtrack into a lookbehind");
|
|
|
|
test(t => {
|
|
// https://github.com/tc39/test262/blob/main/test/built-ins/RegExp/lookBehind/greedy-loop.js
|
|
assert_array_equals("abbbbbbc".match(/(?<=(b+))c/), ["c", "bbbbbb"], "#1");
|
|
assert_array_equals("ab1234c".match(/(?<=(b\d+))c/), ["c", "b1234"], "#2");
|
|
assert_array_equals("ab12b23b34c".match(/(?<=((?:b\d{2})+))c/), ["c", "b12b23b34"], "#3");
|
|
}, "Greedy loop");
|
|
|
|
test(t => {
|
|
// https://github.com/tc39/test262/blob/main/test/built-ins/RegExp/lookBehind/misc.js
|
|
assert_equals("abcdef".match(/(?<=$abc)def/), null, "#1");
|
|
assert_equals("fno".match(/^f.o(?<=foo)$/), null, "#2");
|
|
assert_equals("foo".match(/^foo(?<!foo)$/), null, "#3");
|
|
assert_equals("foo".match(/^f.o(?<!foo)$/), null, "#4");
|
|
assert_array_equals("foo".match(/^foo(?<=foo)$/), ["foo"], "#5");
|
|
assert_array_equals("foo".match(/^f.o(?<=foo)$/), ["foo"], "#6");
|
|
assert_array_equals("fno".match(/^f.o(?<!foo)$/), ["fno"], "#7");
|
|
assert_array_equals("foooo".match(/^foooo(?<=fo+)$/), ["foooo"], "#8");
|
|
assert_array_equals("foooo".match(/^foooo(?<=fo*)$/), ["foooo"], "#9");
|
|
assert_array_equals(/(abc\1)/.exec("abc"), ["abc", "abc"], "#10");
|
|
assert_array_equals(/(abc\1)/.exec("abc\u1234"), ["abc", "abc"], "#11");
|
|
assert_array_equals(/(abc\1)/i.exec("abc"), ["abc", "abc"], "#12");
|
|
assert_array_equals(/(abc\1)/i.exec("abc\u1234"), ["abc", "abc"], "#13");
|
|
}, "Miscellaneous");
|
|
|
|
test(t => {
|
|
// https://github.com/tc39/test262/blob/main/test/built-ins/RegExp/lookBehind/mutual-recursive.js
|
|
assert_array_equals(/(?<=a(.\2)b(\1)).{4}/.exec("aabcacbc"), ["cacb", "a", ""], "#1");
|
|
assert_array_equals(/(?<=a(\2)b(..\1))b/.exec("aacbacb"), ["b", "ac", "ac"], "#2");
|
|
assert_array_equals(/(?<=(?:\1b)(aa))./.exec("aabaax"), ["x", "aa"], "#3");
|
|
assert_array_equals(/(?<=(?:\1|b)(aa))./.exec("aaaax"), ["x", "aa"], "#4");
|
|
}, "Mutual recursive capture/back references");
|
|
|
|
test(t => {
|
|
// https://github.com/tc39/test262/blob/main/test/built-ins/RegExp/lookBehind/negative.js
|
|
assert_array_equals("abcdef".match(/(?<!abc)\w\w\w/), ["abc"], "#1");
|
|
assert_array_equals("abcdef".match(/(?<!a.c)\w\w\w/), ["abc"], "#2");
|
|
assert_array_equals("abcdef".match(/(?<!a\wc)\w\w\w/), ["abc"], "#3");
|
|
assert_array_equals("abcdef".match(/(?<!a[a-z])\w\w\w/), ["abc"], "#4");
|
|
assert_array_equals("abcdef".match(/(?<!a[a-z]{2})\w\w\w/), ["abc"], "#5");
|
|
assert_equals("abcdef".match(/(?<!abc)def/), null, "#6");
|
|
assert_equals("abcdef".match(/(?<!a.c)def/), null, "#7");
|
|
assert_equals("abcdef".match(/(?<!a\wc)def/), null, "#8");
|
|
assert_equals("abcdef".match(/(?<!a[a-z][a-z])def/), null, "#9");
|
|
assert_equals("abcdef".match(/(?<!a[a-z]{2})def/), null, "#10");
|
|
assert_equals("abcdef".match(/(?<!a{1}b{1})cde/), null, "#11");
|
|
assert_equals("abcdef".match(/(?<!a{1}[a-z]{2})def/), null, "#12");
|
|
}, "Negative lookbehinds");
|
|
|
|
test(t => {
|
|
// https://github.com/tc39/test262/blob/main/test/built-ins/RegExp/lookBehind/nested-lookaround.js
|
|
assert_array_equals("abcdef".match(/(?<=ab(?=c)\wd)\w\w/), ["ef"], "#1");
|
|
assert_array_equals("abcdef".match(/(?<=a(?=([^a]{2})d)\w{3})\w\w/), ["ef", "bc"], "#2");
|
|
assert_array_equals("abcdef".match(/(?<=a(?=([bc]{2}(?<!a{2}))d)\w{3})\w\w/), ["ef", "bc"], "#3");
|
|
assert_array_equals("faaao".match(/^faaao?(?<=^f[oa]+(?=o))/), ["faaa"], "#4");
|
|
assert_equals("abcdef".match(/(?<=a(?=([bc]{2}(?<!a*))d)\w{3})\w\w/), null, "#5");
|
|
}, "Nested lookaround");
|
|
|
|
test(t => {
|
|
// https://github.com/tc39/test262/blob/main/test/built-ins/RegExp/lookBehind/simple-fixed-length.js
|
|
assert_equals("b".match(/^.(?<=a)/), null, "#1");
|
|
assert_equals("boo".match(/^f\w\w(?<=\woo)/), null, "#2");
|
|
assert_equals("fao".match(/^f\w\w(?<=\woo)/), null, "#3");
|
|
assert_equals("foa".match(/^f\w\w(?<=\woo)/), null, "#4");
|
|
assert_array_equals("a".match(/^.(?<=a)/), ["a"], "#5");
|
|
assert_array_equals("foo1".match(/^f..(?<=.oo)/), ["foo"], "#6");
|
|
assert_array_equals("foo2".match(/^f\w\w(?<=\woo)/), ["foo"], "#7");
|
|
assert_array_equals("abcdef".match(/(?<=abc)\w\w\w/), ["def"], "#8");
|
|
assert_array_equals("abcdef".match(/(?<=a.c)\w\w\w/), ["def"], "#9");
|
|
assert_array_equals("abcdef".match(/(?<=a\wc)\w\w\w/), ["def"], "#10");
|
|
assert_array_equals("abcdef".match(/(?<=a[a-z])\w\w\w/), ["cde"], "#11");
|
|
assert_array_equals("abcdef".match(/(?<=a[a-z][a-z])\w\w\w/), ["def"], "#12");
|
|
assert_array_equals("abcdef".match(/(?<=a[a-z]{2})\w\w\w/), ["def"], "#13");
|
|
assert_array_equals("abcdef".match(/(?<=a{1})\w\w\w/), ["bcd"], "#14");
|
|
assert_array_equals("abcdef".match(/(?<=a{1}b{1})\w\w\w/), ["cde"], "#15");
|
|
assert_array_equals("abcdef".match(/(?<=a{1}[a-z]{2})\w\w\w/), ["def"], "#16");
|
|
}, "Simple fixed-length matches");
|
|
|
|
test(t => {
|
|
// https://github.com/tc39/test262/blob/main/test/built-ins/RegExp/lookBehind/sliced-strings.js
|
|
const oob_subject = "abcdefghijklmnabcdefghijklmn".slice(14);
|
|
assert_equals(oob_subject.match(/(?=(abcdefghijklmn))(?<=\1)a/i), null, "");
|
|
assert_equals(oob_subject.match(/(?=(abcdefghijklmn))(?<=\1)a/), null, "");
|
|
assert_equals("abcdefgabcdefg".slice(1).match(/(?=(abcdefg))(?<=\1)/), null, "");
|
|
}, "Sliced strings");
|
|
|
|
test(t => {
|
|
// https://github.com/tc39/test262/blob/main/test/built-ins/RegExp/lookBehind/start-of-line.js
|
|
assert_equals("abcdef".match(/(?<=^[^a-c]{3})def/), null, "#1");
|
|
assert_equals("foooo".match(/"^foooo(?<=^o+)$/), null, "#2");
|
|
assert_equals("foooo".match(/"^foooo(?<=^o*)$/), null, "#3");
|
|
assert_array_equals("abcdef".match(/(?<=^abc)def/), ["def"], "#4");
|
|
assert_array_equals("abcdef".match(/(?<=^[a-c]{3})def/), ["def"], "#5");
|
|
assert_array_equals("xyz\nabcdef".match(/(?<=^[a-c]{3})def/m), ["def"], "#6");
|
|
assert_array_equals("ab\ncd\nefg".match(/(?<=^)\w+/gm), ["ab", "cd", "efg"], "#7");
|
|
assert_array_equals("ab\ncd\nefg".match(/\w+(?<=$)/gm), ["ab", "cd", "efg"], "#8");
|
|
assert_array_equals("ab\ncd\nefg".match(/(?<=^)\w+(?<=$)/gm), ["ab", "cd", "efg"], "#9");
|
|
assert_array_equals("foo".match(/^foo(?<=^fo+)$/), ["foo"], "#10");
|
|
assert_array_equals("foooo".match(/^foooo(?<=^fo*)/), ["foooo"], "#11");
|
|
assert_array_equals("foo".match(/^(f)oo(?<=^\1o+)$/), ["foo", "f"], "#12");
|
|
assert_array_equals("foo".match(/^(f)oo(?<=^\1o+)$/i), ["foo", "f"], "#13");
|
|
assert_array_equals("foo\u1234".match(/^(f)oo(?<=^\1o+).$/i), ["foo\u1234", "f"], "#14");
|
|
assert_array_equals("abcdefdef".match(/(?<=^\w+)def/), ["def"], "#15");
|
|
assert_array_equals("abcdefdef".match(/(?<=^\w+)def/g), ["def", "def"], "#16");
|
|
}, "Start of line matches");
|
|
|
|
test(t => {
|
|
// https://github.com/tc39/test262/blob/main/test/built-ins/RegExp/lookBehind/sticky.js
|
|
const re1 = /(?<=^(\w+))def/g;
|
|
assert_array_equals(re1.exec("abcdefdef"), ["def", "abc"], "#1");
|
|
assert_array_equals(re1.exec("abcdefdef"), ["def", "abcdef"], "#2");
|
|
const re2 = /\Bdef/g;
|
|
assert_array_equals(re2.exec("abcdefdef"), ["def"], "#3");
|
|
assert_array_equals(re2.exec("abcdefdef"), ["def"], "#4");
|
|
}, "Sticky matches");
|
|
|
|
test(t => {
|
|
// https://github.com/tc39/test262/blob/main/test/built-ins/RegExp/lookBehind/variable-length.js
|
|
assert_array_equals("abcdef".match(/(?<=[a|b|c]*)[^a|b|c]{3}/), ["def"], "#1");
|
|
assert_array_equals("abcdef".match(/(?<=\w*)[^a|b|c]{3}/), ["def"], "#2");
|
|
}, "Variable-length matches");
|
|
|
|
test(t => {
|
|
// https://github.com/tc39/test262/blob/main/test/built-ins/RegExp/lookBehind/word-boundary.js
|
|
assert_array_equals("abc def".match(/(?<=\b)[d-f]{3}/), ["def"], "#1");
|
|
assert_array_equals("ab cdef".match(/(?<=\B)\w{3}/), ["def"], "#2");
|
|
assert_array_equals("ab cdef".match(/(?<=\B)(?<=c(?<=\w))\w{3}/), ["def"], "#3");
|
|
assert_equals("abcdef".match(/(?<=\b)[d-f]{3}/), null, "#4");
|
|
}, "Word boundary matches");
|
|
</script>
|