[webkit-changes] [WebKit/WebKit] cbde56: REGRESSION(257823 at main): named-groups/lookbehind.j...

[Michael Saboff]

  Branch: refs/heads/main
  Home:   https://github.com/WebKit/WebKit
  Commit: cbde560aa655d9f1759f23336aa506ccaaef8ee2
      https://github.com/WebKit/WebKit/commit/cbde560aa655d9f1759f23336aa506ccaaef8ee2
  Author: Michael Saboff <msaboff at apple.com>
  Date:   2023-01-04 (Wed, 04 Jan 2023)

  Changed paths:
    A JSTests/stress/regexp-lookaround-captures.js
    M Source/JavaScriptCore/yarr/YarrInterpreter.cpp
    M Source/JavaScriptCore/yarr/YarrInterpreter.h
    M Source/JavaScriptCore/yarr/YarrJIT.cpp
    M Source/JavaScriptCore/yarr/YarrParser.h
    M Source/JavaScriptCore/yarr/YarrPattern.cpp
    M Source/JavaScriptCore/yarr/YarrPattern.h
    M Source/JavaScriptCore/yarr/YarrSyntaxChecker.cpp
    M Source/WebCore/contentextensions/URLFilterParser.cpp

  Log Message:
  -----------
  REGRESSION(257823 at main): named-groups/lookbehind.js Test262-test is failing
https://bugs.webkit.org/show_bug.cgi?id=249330
rdar://103367993

Reviewed by Mark Lam.

Relanding 258310 at main with a fix for the issue found in https://bugs.webkit.org/show_bug.cgi?id=249855.
That issue was that we were hitting an ASSERT in clearSubpatternStart() where the subpatternId passed in was 0,
which is the whole match start index and not a subpattern.  The fix is to make sure that we always start with
a subpatternId of no less than 1.  This is handled in the new method PatternAlternative::firstCleanupSubpatternId().

Fixed the case where a nested capturing group within a lookahead / lookbehind doesn't get its capture cleared if
the lookaround fails.  This has been a long standing issue for lookaheads.

The fix is slightly different in the Yarr interpreter than in the Yarr JIT.  For the interpreter, the Parenthetical
Assertion processing will clear the nested captures within the assertion for the case where an inverted assertion
succeeds or a normal assertion fails.  For the JIT, the clearing is done at the end of a failed top level alternative
for all the contained captures for that alternative.

Updated the test accordingly.

* JSTests/stress/regexp-lookaround-captures.js: Added.
(arrayToString):
(dumpValue):
(compareArray):
(testRegExp):
(testRegExp.c):
* Source/JavaScriptCore/yarr/YarrInterpreter.cpp:
(JSC::Yarr::Interpreter::ParenthesesDisjunctionContext::ParenthesesDisjunctionContext):
(JSC::Yarr::Interpreter::matchBackReference):
(JSC::Yarr::Interpreter::backtrackBackReference):
(JSC::Yarr::Interpreter::recordParenthesesMatch):
(JSC::Yarr::Interpreter::resetMatches):
(JSC::Yarr::Interpreter::matchParenthesesOnceBegin):
(JSC::Yarr::Interpreter::matchParenthesesOnceEnd):
(JSC::Yarr::Interpreter::backtrackParenthesesOnceBegin):
(JSC::Yarr::Interpreter::backtrackParenthesesOnceEnd):
(JSC::Yarr::Interpreter::matchParentheticalAssertionEnd):
(JSC::Yarr::Interpreter::backtrackParentheticalAssertionEnd):
(JSC::Yarr::Interpreter::matchDisjunction):
(JSC::Yarr::ByteCompiler::atomParentheticalAssertionEnd):
(JSC::Yarr::ByteCompiler::atomParenthesesSubpatternEnd):
(JSC::Yarr::ByteCompiler::atomParenthesesOnceEnd):
(JSC::Yarr::ByteCompiler::atomParenthesesTerminalEnd):
(JSC::Yarr::ByteCompiler::emitDisjunction):
(JSC::Yarr::ByteTermDumper::dumpTerm):
* Source/JavaScriptCore/yarr/YarrInterpreter.h:
(JSC::Yarr::ByteTerm::ByteTerm):
(JSC::Yarr::ByteTerm::containsAnyCaptures):
(JSC::Yarr::ByteTerm::subpatternId):
(JSC::Yarr::ByteTerm::lastSubpatternId):
* Source/JavaScriptCore/yarr/YarrJIT.cpp:
* Source/JavaScriptCore/yarr/YarrParser.h:
(JSC::Yarr::Parser::parseTokens):
* Source/JavaScriptCore/yarr/YarrPattern.cpp:
(JSC::Yarr::YarrPatternConstructor::atomParenthesesSubpatternBegin):
(JSC::Yarr::YarrPatternConstructor::atomParentheticalAssertionBegin):
(JSC::Yarr::YarrPatternConstructor::copyDisjunction):
(JSC::Yarr::YarrPatternConstructor::disjunction):
* Source/JavaScriptCore/yarr/YarrPattern.h:
(JSC::Yarr::PatternTerm::containsAnyCaptures):
(JSC::Yarr::PatternAlternative::PatternAlternative):
(JSC::Yarr::PatternAlternative::needToCleanupCaptures const):
(JSC::Yarr::PatternAlternative::firstCleanupSubpatternId):
(JSC::Yarr::PatternDisjunction::addNewAlternative):
* Source/JavaScriptCore/yarr/YarrSyntaxChecker.cpp:
(JSC::Yarr::SyntaxChecker::disjunction):
* Source/WebCore/contentextensions/URLFilterParser.cpp:
(WebCore::ContentExtensions::PatternParser::disjunction):

Canonical link: https://commits.webkit.org/258441@main