Update RegExp and related classes to use 8 bit strings when available
https://bugs.webkit.org/show_bug.cgi?id=67337
Source/JavaScriptCore:
Modified both the Yarr interpreter and JIT to handle 8 bit subject strings.
The code paths are triggered by the UString::is8bit() method which currently
returns false. Implemented JIT changes for all current architectures.
Tested X86_64 and ARM v7.
This includes some code that will likely change as we complete the
8 bit string changes. This includes the way the raw buffer pointers
are accessed as well as replacing the CharAccess class with a
string interator returned from UString.
Fixed build breakage in testRegExp.cpp due to globalObject construction
changes.
Reviewed by Gavin Barraclough.
(GlobalObject::finishCreation):
(GlobalObject::GlobalObject):
- assembler/ARMAssembler.cpp:
(JSC::ARMAssembler::baseIndexTransfer32):
- assembler/ARMAssembler.h:
- assembler/ARMv7Assembler.h:
(JSC::ARMv7Assembler::ubfx):
(JSC::ARMv7Assembler::ARMInstructionFormatter::twoWordOp12Reg40Imm3Reg4Imm20Imm5):
- assembler/MacroAssemblerARM.h:
(JSC::MacroAssemblerARM::load8):
(JSC::MacroAssemblerARM::branch8):
(JSC::MacroAssemblerARM::branch16):
- assembler/MacroAssemblerARMv7.h:
(JSC::MacroAssemblerARMv7::load8):
(JSC::MacroAssemblerARMv7::branch16):
(JSC::MacroAssemblerARMv7::branch8):
- assembler/MacroAssemblerMIPS.h:
(JSC::MacroAssemblerMIPS::load8):
(JSC::MacroAssemblerMIPS::branch8):
(JSC::MacroAssemblerMIPS::branch16):
- assembler/MacroAssemblerSH4.h:
(JSC::MacroAssemblerSH4::load8):
(JSC::MacroAssemblerSH4::branch8):
(JSC::MacroAssemblerSH4::branch16):
- assembler/MacroAssemblerX86Common.h:
(JSC::MacroAssemblerX86Common::load8):
(JSC::MacroAssemblerX86Common::branch16):
(JSC::MacroAssemblerX86Common::branch8):
- assembler/SH4Assembler.h:
(JSC::SH4Assembler::extub):
(JSC::SH4Assembler::printInstr):
- assembler/X86Assembler.h:
(JSC::X86Assembler::cmpw_ir):
(JSC::X86Assembler::movzbl_mr):
(JSC::RegExp::compile):
(JSC::RegExp::compileIfNecessary):
(JSC::RegExp::match):
(JSC::RegExp::matchCompareWithInterpreter):
- runtime/RegExp.h:
- runtime/UString.h:
(JSC::UString::is8Bit):
- yarr/Yarr.h:
- yarr/YarrInterpreter.cpp:
(JSC::Yarr::Interpreter::CharAccess::CharAccess):
(JSC::Yarr::Interpreter::CharAccess::~CharAccess):
(JSC::Yarr::Interpreter::CharAccess::operator[]):
(JSC::Yarr::Interpreter::InputStream::InputStream):
(JSC::Yarr::Interpreter::Interpreter):
(JSC::Yarr::interpret):
(JSC::Yarr::YarrGenerator::jumpIfCharNotEquals):
(JSC::Yarr::YarrGenerator::readCharacter):
(JSC::Yarr::YarrGenerator::generatePatternCharacterOnce):
(JSC::Yarr::YarrGenerator::generatePatternCharacterFixed):
(JSC::Yarr::YarrGenerator::generatePatternCharacterGreedy):
(JSC::Yarr::YarrGenerator::backtrackPatternCharacterNonGreedy):
(JSC::Yarr::YarrGenerator::generateCharacterClassFixed):
(JSC::Yarr::YarrGenerator::generateDotStarEnclosure):
(JSC::Yarr::YarrGenerator::YarrGenerator):
(JSC::Yarr::YarrGenerator::compile):
(JSC::Yarr::jitCompile):
(JSC::Yarr::execute):
(JSC::Yarr::YarrCodeBlock::has8BitCode):
(JSC::Yarr::YarrCodeBlock::has16BitCode):
(JSC::Yarr::YarrCodeBlock::set8BitCode):
(JSC::Yarr::YarrCodeBlock::set16BitCode):
(JSC::Yarr::YarrCodeBlock::execute):
(JSC::Yarr::Parser::Parser):
Source/WebCore:
Updated call to match to use UString& instead of UChar*.
Reviewed by Gavin Barraclough.
No new tests, Covered by existing tests.
- platform/text/RegularExpression.cpp:
(WebCore::RegularExpression::match):