Google Git
Sign in
webkit / WebKit / 770a8a809515625b0f9127e8dd126b81698b6fc4 / . / JavaScriptCore / ChangeLog
blob: 307c44dfbbbd49178d777cde2e90390b2b92bad1 [file] [log] [blame]
2008-11-14 Cameron Zwarich <zwarich@apple.com>
Reviewed by Darin Adler.
Bug 22245: Move wtf/dtoa.h into the WTF namespace
<https://bugs.webkit.org/show_bug.cgi?id=22245>
Move wtf/dtoa.h into the WTF namespace from the JSC namespace. This
introduces some ambiguities in name lookups, so I changed all uses of
the functions in wtf/dtoa.h to explicitly state the namespace.
* JavaScriptCore.exp:
* parser/Lexer.cpp:
(JSC::Lexer::lex):
* runtime/InitializeThreading.cpp:
* runtime/JSGlobalObjectFunctions.cpp:
(JSC::parseInt):
* runtime/NumberPrototype.cpp:
(JSC::integerPartNoExp):
(JSC::numberProtoFuncToExponential):
* runtime/UString.cpp:
(JSC::concatenate):
(JSC::UString::from):
(JSC::UString::toDouble):
* wtf/dtoa.cpp:
* wtf/dtoa.h:
2008-11-14 Cameron Zwarich <zwarich@apple.com>
Reviewed by Maciej Stachowiak.
Bug 22257: Enable redundant read optimizations for results generated by compileBinaryArithOp()
<https://bugs.webkit.org/show_bug.cgi?id=22257>
This shows no change in performance on either SunSpider or the V8
benchmark suite, but it removes an ugly special case and allows for
future optimizations to be implemented in a cleaner fashion.
This patch was essentially given to me by Gavin Barraclough upon my
request, but I did regression and performance testing so that he could
work on something else.
* VM/CTI.cpp:
(JSC::CTI::putDoubleResultToJSNumberCellOrJSImmediate): Move the final
result to eax if it is not already there.
(JSC::CTI::compileBinaryArithOp): Remove the killing of the final result
register that disables the optimization.
2008-11-13 Eric Seidel <eric@webkit.org>
Reviewed by Adam Roben.
Add a Scons-based build system for building
the Chromium-Mac build of JavaScriptCore.
https://bugs.webkit.org/show_bug.cgi?id=21991
* JavaScriptCore.scons: Added.
* SConstruct: Added.
2008-11-13 Eric Seidel <eric@webkit.org>
Reviewed by Adam Roben.
Add PLATFORM(CHROMIUM) to the "we don't use cairo" blacklist
until https://bugs.webkit.org/show_bug.cgi?id=22250 is fixed.
* wtf/Platform.h:
2008-11-13 Cameron Zwarich <zwarich@apple.com>
Reviewed by Sam Weinig.
In r38375 the 'jsc' shell was changed to improve teardown on quit. The
main() function in jsc.cpp uses Structured Exception Handling, so Visual
C++ emits a warning when destructors are used.
In order to speculatively fix the Windows build, this patch changes that
code to use explicit pointer manipulation and locking rather than smart
pointers and RAII.
* jsc.cpp:
(main):
2008-11-13 Cameron Zwarich <zwarich@apple.com>
Reviewed by Darin Adler.
Bug 22246: Get arguments for opcodes together to eliminate more redundant memory reads
<https://bugs.webkit.org/show_bug.cgi?id=22246>
It is common for opcodes to read their first operand into eax and their
second operand into edx. If the value intended for the second operand is
in eax, we should first move eax to the register for the second operand
and then read the first operand into eax.
This is a 0.5% speedup on SunSpider and a 2.0% speedup on the V8
benchmark suite when measured using the V8 harness.
* VM/CTI.cpp:
(JSC::CTI::emitGetArgs):
(JSC::CTI::compileOpStrictEq):
(JSC::CTI::compileBinaryArithOp):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
* VM/CTI.h:
2008-11-13 Cameron Zwarich <zwarich@apple.com>
Reviewed by Darin Adler.
Bug 22238: Avoid unnecessary reads of temporaries when the target machine register is not eax
<https://bugs.webkit.org/show_bug.cgi?id=22238>
Enable the optimization of not reading a value back from memory that we
just wrote when the target machine register is not eax. In order to do
this, the code generation for op_put_global_var must be changed to
read its argument into a register before overwriting eax.
This is a 0.5% speedup on SunSpider and shows no change on the V8
benchmark suite when run in either harness.
* VM/CTI.cpp:
(JSC::CTI::emitGetArg):
(JSC::CTI::privateCompileMainPass):
2008-11-13 Cameron Zwarich <zwarich@apple.com>
Reviewed by Alexey Proskuryakov.
Perform teardown in the 'jsc' shell in order to suppress annoying and
misleading leak messages. There is still a lone JSC::Node leaking when
quit() is called, but hopefully that can be fixed as well.
* jsc.cpp:
(functionQuit):
(main):
2008-11-13 Mike Pinkerton <pinkerton@chromium.org>
Reviewed by Sam Weinig.
Fix for https://bugs.webkit.org/show_bug.cgi?id=22087
Need correct platform defines for Mac Chromium
Set the appropriate platform defines for Mac Chromium, which is
similar to PLATFORM(MAC), but isn't.
* wtf/Platform.h:
2008-11-13 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich.
- remove immediate checks from native codegen for known non-immediate cases like "this"
~.5% speedup on v8 benchmarks
In the future we can extend this model to remove all sorts of
typechecks based on local type info or type inference.
I also added an assertion to verify that all slow cases linked as
many slow case jumps as the corresponding fast case generated, and
fixed the pre-existing cases where this was not true.
* VM/CTI.cpp:
(JSC::CTI::emitJumpSlowCaseIfNotJSCell):
(JSC::CTI::linkSlowCaseIfNotJSCell):
(JSC::CTI::compileBinaryArithOp):
(JSC::CTI::compileBinaryArithOpSlowCase):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
* VM/CTI.h:
* VM/CodeBlock.h:
(JSC::CodeBlock::isKnownNotImmediate):
2008-11-13 Cameron Zwarich <zwarich@apple.com>
Reviewed by Maciej Stachowiak.
Bug 21943: Avoid needless reads of temporary values in CTI code
<https://bugs.webkit.org/show_bug.cgi?id=21943>
If an opcode needs to load a virtual register and a previous opcode left
the contents of that virtual register in a machine register, use the
value in the machine register rather than getting it from memory.
In order to perform this optimization, it is necessary to know the
jump tagets in the CodeBlock. For temporaries, the only problematic
jump targets are binary logical operators and the ternary conditional
operator. However, if this optimization were to be extended to local
variable registers as well, other jump targets would need to be
included, like switch statement cases and the beginnings of catch
blocks.
This optimization also requires that the fast case and the slow case
of an opcode use emitPutResult() on the same register, which was chosen
to be eax, as that is the register into which we read the first operand
of opcodes. In order to make this the case, we needed to add some mov
instructions to the slow cases of some instructions.
This optimizaton is not applied whenever compileBinaryArithOp() is used
to compile an opcode, because different machine registers may be used to
store the final result. It seems possible to rewrite the code generation
in compileBinaryArithOp() to allow for this optimization.
This optimization is also not applied when generating slow cases,
because some fast cases overwrite the value of eax before jumping to the
slow case. In the future, it may be possible to apply this optimization
to slow cases as well, but it did not seem to be a speedup when testing
an early version of this patch.
This is a 1.0% speedup on SunSpider and a 6.3% speedup on the V8
benchmark suite.
* VM/CTI.cpp:
(JSC::CTI::killLastResultRegister):
(JSC::CTI::emitGetArg):
(JSC::CTI::emitGetPutArg):
(JSC::CTI::emitGetCTIParam):
(JSC::CTI::emitGetFromCallFrameHeader):
(JSC::CTI::emitPutResult):
(JSC::CTI::emitCTICall):
(JSC::CTI::CTI):
(JSC::CTI::compileOpCall):
(JSC::CTI::compileOpStrictEq):
(JSC::CTI::emitSlowScriptCheck):
(JSC::CTI::compileBinaryArithOp):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
(JSC::CTI::privateCompileGetByIdProto):
(JSC::CTI::privateCompilePatchGetArrayLength):
* VM/CTI.h:
* VM/CodeBlock.h:
(JSC::CodeBlock::isTemporaryRegisterIndex):
* bytecompiler/CodeGenerator.cpp:
(JSC::CodeGenerator::emitLabel):
2008-11-12 Alp Toker <alp@nuanti.com>
autotools build system fix-up only. Add FloatQuad.h to the source
lists and sort them.
* GNUmakefile.am:
2008-11-12 Geoffrey Garen <ggaren@apple.com>
Reviewed by Sam Weinig.
Fixed https://bugs.webkit.org/show_bug.cgi?id=22192
+37 failures in fast/profiler
along with Darin's review comments in
https://bugs.webkit.org/show_bug.cgi?id=22174
Simplified op_call by nixing its responsibility for moving the value of
"this" into the first argument slot
* VM/Machine.cpp:
(JSC::returnToThrowTrampoline):
(JSC::throwStackOverflowError):
(JSC::Machine::cti_register_file_check):
(JSC::Machine::cti_op_call_arityCheck):
(JSC::Machine::cti_vm_throw): Moved the throw logic into a function, since
functions are better than macros.
* bytecompiler/CodeGenerator.cpp:
(JSC::CodeGenerator::emitCall):
(JSC::CodeGenerator::emitConstruct): Ensure that the function register
is preserved if profiling is enabled, since the profiler uses that
register.
* runtime/JSGlobalData.h: Renamed throwReturnAddress to exceptionLocation,
because I had a hard time understanding what "throwReturnAddress" meant.
2008-11-12 Geoffrey Garen <ggaren@apple.com>
Reviewed by Sam Weinig.
Roll in r38322, now that test failures have been fixed.
* VM/CTI.cpp:
(JSC::CTI::compileOpCallSetupArgs):
(JSC::CTI::compileOpCallEvalSetupArgs):
(JSC::CTI::compileOpConstructSetupArgs):
(JSC::CTI::compileOpCall):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
* VM/CTI.h:
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
* VM/Machine.cpp:
(JSC::Machine::callEval):
(JSC::Machine::dumpCallFrame):
(JSC::Machine::dumpRegisters):
(JSC::Machine::execute):
(JSC::Machine::privateExecute):
(JSC::Machine::cti_register_file_check):
(JSC::Machine::cti_op_call_arityCheck):
(JSC::Machine::cti_op_call_NotJSFunction):
(JSC::Machine::cti_op_construct_JSConstruct):
(JSC::Machine::cti_op_construct_NotJSConstruct):
(JSC::Machine::cti_op_call_eval):
(JSC::Machine::cti_vm_throw):
* VM/Machine.h:
* bytecompiler/CodeGenerator.cpp:
(JSC::CodeGenerator::emitCall):
(JSC::CodeGenerator::emitCallEval):
(JSC::CodeGenerator::emitConstruct):
* bytecompiler/CodeGenerator.h:
* parser/Nodes.cpp:
(JSC::EvalFunctionCallNode::emitCode):
(JSC::FunctionCallValueNode::emitCode):
(JSC::FunctionCallResolveNode::emitCode):
(JSC::FunctionCallBracketNode::emitCode):
(JSC::FunctionCallDotNode::emitCode):
* parser/Nodes.h:
(JSC::ScopeNode::neededConstants):
2008-11-12 Gavin Barraclough <barraclough@apple.com>
Reviewed by Cameron Zwarich.
Fix for https://bugs.webkit.org/show_bug.cgi?id=22201
Integer conversion in array.length was safe signed values,
but the length is unsigned.
* VM/CTI.cpp:
(JSC::CTI::privateCompilePatchGetArrayLength):
2008-11-12 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Mark Rowe.
Roll out r38322 due to test failures on the bots.
* VM/CTI.cpp:
(JSC::CTI::compileOpCallSetupArgs):
(JSC::CTI::compileOpCall):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
* VM/CTI.h:
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
* VM/Machine.cpp:
(JSC::Machine::callEval):
(JSC::Machine::dumpCallFrame):
(JSC::Machine::dumpRegisters):
(JSC::Machine::execute):
(JSC::Machine::privateExecute):
(JSC::Machine::throwStackOverflowPreviousFrame):
(JSC::Machine::cti_register_file_check):
(JSC::Machine::cti_op_call_arityCheck):
(JSC::Machine::cti_op_call_NotJSFunction):
(JSC::Machine::cti_op_construct_JSConstruct):
(JSC::Machine::cti_op_construct_NotJSConstruct):
(JSC::Machine::cti_op_call_eval):
(JSC::Machine::cti_vm_throw):
* VM/Machine.h:
* bytecompiler/CodeGenerator.cpp:
(JSC::CodeGenerator::emitCall):
(JSC::CodeGenerator::emitCallEval):
(JSC::CodeGenerator::emitConstruct):
* bytecompiler/CodeGenerator.h:
* parser/Nodes.cpp:
(JSC::EvalFunctionCallNode::emitCode):
(JSC::FunctionCallValueNode::emitCode):
(JSC::FunctionCallResolveNode::emitCode):
(JSC::FunctionCallBracketNode::emitCode):
(JSC::FunctionCallDotNode::emitCode):
* parser/Nodes.h:
(JSC::ScopeNode::neededConstants):
2008-11-11 Geoffrey Garen <ggaren@apple.com>
Reviewed by Darin Adler.
Fixed https://bugs.webkit.org/show_bug.cgi?id=22174
Simplified op_call by nixing its responsibility for moving the value of
"this" into the first argument slot.
Instead, the caller emits an explicit load or mov instruction, or relies
on implicit knowledge that "this" is already in the first argument slot.
As a result, two operands to op_call are gone: firstArg and thisVal.
SunSpider and v8 tests show no change in bytecode or CTI.
* VM/CTI.cpp:
(JSC::CTI::compileOpCallSetupArgs):
(JSC::CTI::compileOpCallEvalSetupArgs):
(JSC::CTI::compileOpConstructSetupArgs): Split apart these three versions
of setting up arguments to op_call, because they're more different than
they are the same -- even more so with this patch.
(JSC::CTI::compileOpCall): Updated for the fact that op_construct doesn't
match op_call anymore.
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases): Merged a few call cases. Updated
for changes mentioned above.
* VM/CTI.h:
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump): Updated for new bytecode format of call / construct.
* VM/Machine.cpp:
(JSC::Machine::callEval): Updated for new bytecode format of call / construct.
(JSC::Machine::dumpCallFrame):
(JSC::Machine::dumpRegisters): Simplified these debugging functions,
taking advantage of the new call frame layout.
(JSC::Machine::execute): Fixed up the eval version of execute to be
friendlier to calls in the new format.
(JSC::Machine::privateExecute): Implemented the new call format in
bytecode.
(JSC::Machine::cti_op_call_NotJSFunction):
(JSC::Machine::cti_op_construct_JSConstruct):
(JSC::Machine::cti_op_construct_NotJSConstruct):
(JSC::Machine::cti_op_call_eval): Updated CTI helpers to match the new
call format.
Fixed a latent bug in stack overflow checking that is now hit because
the register layout has changed a bit -- namely: when throwing a stack
overflow exception inside an op_call helper, we need to account for the
fact that the current call frame is only half-constructed, and use the
parent call frame instead.
* VM/Machine.h:
* bytecompiler/CodeGenerator.cpp:
(JSC::CodeGenerator::emitCall):
(JSC::CodeGenerator::emitCallEval):
(JSC::CodeGenerator::emitConstruct):
* bytecompiler/CodeGenerator.h: Updated codegen to match the new call
format.
* parser/Nodes.cpp:
(JSC::EvalFunctionCallNode::emitCode):
(JSC::FunctionCallValueNode::emitCode):
(JSC::FunctionCallResolveNode::emitCode):
(JSC::FunctionCallBracketNode::emitCode):
(JSC::FunctionCallDotNode::emitCode):
* parser/Nodes.h:
(JSC::ScopeNode::neededConstants): ditto
2008-11-11 Cameron Zwarich <zwarich@apple.com>
Reviewed by Geoff Garen.
Remove an unused forwarding header for a file that no longer exists.
* ForwardingHeaders/JavaScriptCore/JSLock.h: Removed.
2008-11-11 Mark Rowe <mrowe@apple.com>
Fix broken dependencies building JavaScriptCore on a freezing cold cat, caused
by failure to update all instances of "kjs" to their new locations.
* JavaScriptCore.xcodeproj/project.pbxproj:
2008-11-11 Alexey Proskuryakov <ap@webkit.org>
Rubber-stamped by Adam Roben.
* wtf/AVLTree.h: (WTF::AVLTree::Iterator::start_iter):
Fix indentation a little more.
2008-11-11 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Sam Weinig.
Clean up EvalCodeCache to match our coding style a bit more.
* VM/EvalCodeCache.h:
(JSC::EvalCodeCache::get):
2008-11-11 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Sam Weinig.
Bug 22179: Move EvalCodeCache from CodeBlock.h into its own file
<https://bugs.webkit.org/show_bug.cgi?id=22179>
* GNUmakefile.am:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* VM/CodeBlock.h:
* VM/EvalCodeCache.h: Copied from VM/CodeBlock.h.
* VM/Machine.cpp:
2008-11-11 Cameron Zwarich <zwarich@apple.com>
Reviewed by Sam Weinig.
Remove the 'm_' prefix from the fields of the SwitchRecord struct.
* VM/CTI.cpp:
(JSC::CTI::privateCompile):
* VM/CTI.h:
(JSC::SwitchRecord):
(JSC::SwitchRecord::SwitchRecord):
2008-11-11 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Sam Weinig.
Make asInteger() a static function so that it has internal linkage.
* VM/CTI.cpp:
(JSC::asInteger):
2008-11-11 Maciej Stachowiak <mjs@apple.com>
Reviewed by Mark Rowe.
- shrink CodeBlock and AST related Vectors to exact fit (5-10M savings on membuster test)
No perf regression combined with the last patch (each seems like a small regression individually)
* bytecompiler/CodeGenerator.cpp:
(JSC::CodeGenerator::generate):
* parser/Nodes.h:
(JSC::SourceElements::releaseContentsIntoVector):
* wtf/Vector.h:
(WTF::Vector::shrinkToFit):
2008-11-11 Maciej Stachowiak <mjs@apple.com>
Reviewed by Mark Rowe.
- remove inline capacity from declaration stacks (15M savings on membuster test)
No perf regression on SunSpider or V8 test combined with other upcoming memory improvement patch.
* JavaScriptCore.exp:
* parser/Nodes.h:
2008-11-11 Cameron Zwarich <zwarich@apple.com>
Reviewed by Oliver Hunt.
While r38286 removed the need for the m_callFrame member variable of
CTI, it should be also be removed.
* VM/CTI.h:
2008-11-10 Cameron Zwarich <zwarich@apple.com>
Reviewed by Oliver Hunt.
Make CTI::asInteger() a non-member function, since it needs no access to
any of CTI's member variables.
* VM/CTI.cpp:
(JSC::asInteger):
* VM/CTI.h:
2008-11-10 Cameron Zwarich <zwarich@apple.com>
Reviewed by Maciej Stachowiak.
Use 'value' instead of 'js' in CTI as a name for JSValue* to match our
usual convention elsewhere.
* VM/CTI.cpp:
(JSC::CTI::emitGetArg):
(JSC::CTI::emitGetPutArg):
(JSC::CTI::getConstantImmediateNumericArg):
(JSC::CTI::printOpcodeOperandTypes):
2008-11-10 Cameron Zwarich <zwarich@apple.com>
Reviewed by Maciej Stachowiak.
Make CTI::getConstant() a member function of CodeBlock instead.
* VM/CTI.cpp:
(JSC::CTI::emitGetArg):
(JSC::CTI::emitGetPutArg):
(JSC::CTI::getConstantImmediateNumericArg):
(JSC::CTI::printOpcodeOperandTypes):
(JSC::CTI::privateCompileMainPass):
* VM/CTI.h:
* VM/CodeBlock.h:
(JSC::CodeBlock::getConstant):
2008-11-10 Cameron Zwarich <zwarich@apple.com>
Reviewed by Sam Weinig.
Rename CodeBlock::isConstant() to isConstantRegisterIndex().
* VM/CTI.cpp:
(JSC::CTI::emitGetArg):
(JSC::CTI::emitGetPutArg):
(JSC::CTI::getConstantImmediateNumericArg):
(JSC::CTI::printOpcodeOperandTypes):
(JSC::CTI::privateCompileMainPass):
* VM/CodeBlock.h:
(JSC::CodeBlock::isConstantRegisterIndex):
* bytecompiler/CodeGenerator.cpp:
(JSC::CodeGenerator::emitEqualityOp):
2008-11-10 Gavin Barraclough <barraclough@apple.com>
Build fix for non-CTI builds.
* VM/Machine.cpp:
(JSC::Machine::initialize):
2008-11-10 Cameron Zwarich <zwarich@apple.com>
Reviewed by Sam Weinig.
Remove the unused labels member variable of CodeBlock.
* VM/CodeBlock.h:
* VM/LabelID.h:
(JSC::LabelID::setLocation):
2008-11-10 Gavin Barraclough <barraclough@apple.com>
Reviewed by Camron Zwarich.
Batch compile the set of static trampolines at the point Machine is constructed, using a single allocation.
Refactor out m_callFrame from CTI, since this is only needed to access the global data (instead store a
pointer to the global data directly, since this is available at the point the Machine is constructed).
Add a method to align the code buffer, to allow JIT generation for multiple trampolines in one block.
* VM/CTI.cpp:
(JSC::CTI::getConstant):
(JSC::CTI::emitGetArg):
(JSC::CTI::emitGetPutArg):
(JSC::CTI::getConstantImmediateNumericArg):
(JSC::CTI::printOpcodeOperandTypes):
(JSC::CTI::CTI):
(JSC::CTI::compileBinaryArithOp):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileGetByIdProto):
(JSC::CTI::privateCompileGetByIdChain):
(JSC::CTI::privateCompileCTIMachineTrampolines):
(JSC::CTI::freeCTIMachineTrampolines):
* VM/CTI.h:
(JSC::CTI::compile):
(JSC::CTI::compileGetByIdSelf):
(JSC::CTI::compileGetByIdProto):
(JSC::CTI::compileGetByIdChain):
(JSC::CTI::compilePutByIdReplace):
(JSC::CTI::compilePutByIdTransition):
(JSC::CTI::compileCTIMachineTrampolines):
(JSC::CTI::compilePatchGetArrayLength):
* VM/Machine.cpp:
(JSC::Machine::initialize):
(JSC::Machine::~Machine):
(JSC::Machine::execute):
(JSC::Machine::tryCTICachePutByID):
(JSC::Machine::tryCTICacheGetByID):
(JSC::Machine::cti_op_call_JSFunction):
(JSC::Machine::cti_vm_lazyLinkCall):
* VM/Machine.h:
* masm/X86Assembler.h:
(JSC::JITCodeBuffer::isAligned):
(JSC::X86Assembler::):
(JSC::X86Assembler::align):
* runtime/JSGlobalData.cpp:
(JSC::JSGlobalData::JSGlobalData):
2008-11-10 Maciej Stachowiak <mjs@apple.com>
Reviewed by Antti Koivisto.
- Make Vector::clear() release the Vector's memory (1MB savings on membuster)
https://bugs.webkit.org/show_bug.cgi?id=22170
* wtf/Vector.h:
(WTF::VectorBufferBase::deallocateBuffer): Set capacity to 0 as
well as size, otherwise shrinking capacity to 0 can fail to reset
the capacity and thus cause a future crash.
(WTF::Vector::~Vector): Shrink size not capacity; we only need
to call destructors, the buffer will be freed anyway.
(WTF::Vector::clear): Change this to shrinkCapacity(0), not just shrink(0).
(WTF::::shrinkCapacity): Use shrink() instead of resize() for case where
the size is greater than the new capacity, to work with types that have no
default constructor.
2008-11-10 Cameron Zwarich <zwarich@apple.com>
Reviewed by Maciej Stachowiak.
Split multiple definitions into separate lines.
* VM/CTI.cpp:
(JSC::CTI::compileBinaryArithOp):
2008-11-10 Cameron Zwarich <zwarich@apple.com>
Reviewed by Geoff Garen.
Bug 22162: Remove cachedValueGetter from the JavaScriptCore API implementation
<https://bugs.webkit.org/show_bug.cgi?id=22162>
There is no more need for the cachedValueGetter hack now that we have
PropertySlot::setValue(), so we should remove it.
* API/JSCallbackObject.h:
* API/JSCallbackObjectFunctions.h:
(JSC::::getOwnPropertySlot):
2008-11-10 Cameron Zwarich <zwarich@apple.com>
Reviewed by Darin Adler.
Bug 22152: Remove asObject() call from JSCallbackObject::getOwnPropertySlot()
<https://bugs.webkit.org/show_bug.cgi?id=22152>
With the recent change to adopt asType() style cast functions with
assertions instead of static_casts in many places, the assertion for
the asObject() call in JSCallbackObject::getOwnPropertySlot() has been
failing when using any nontrivial client of the JavaScriptCore API.
The cast isn't even necessary to call slot.setCustom(), so it should
be removed.
* API/JSCallbackObjectFunctions.h:
(JSC::JSCallbackObject::getOwnPropertySlot):
2008-11-10 Alexey Proskuryakov <ap@webkit.org>
Reviewed by Adam Roben.
A few coding style fixes for AVLTree.
* wtf/AVLTree.h: Moved to WTF namespace, Removed "KJS_" from include guards.
(WTF::AVLTree::Iterator::start_iter): Fixed indentation
* runtime/JSArray.cpp: Added "using namepace WTF".
2008-11-09 Cameron Zwarich <zwarich@apple.com>
Not reviewed.
Speculatively fix the non-AllInOne build.
* runtime/NativeErrorConstructor.cpp:
2008-11-09 Darin Adler <darin@apple.com>
Reviewed by Tim Hatcher.
- https://bugs.webkit.org/show_bug.cgi?id=22149
remove unused code from the parser
* AllInOneFile.cpp: Removed nodes2string.cpp.
* GNUmakefile.am: Ditto.
* JavaScriptCore.exp: Ditto.
* JavaScriptCore.pri: Ditto.
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Ditto.
* JavaScriptCore.xcodeproj/project.pbxproj: Ditto.
* JavaScriptCoreSources.bkl: Ditto.
* VM/CodeBlock.h: Added include.
* VM/Machine.cpp: (JSC::Machine::execute): Use the types from
DeclarationStacks as DeclarationStacks:: rather than Node:: since
"Node" really has little to do with it.
* bytecompiler/CodeGenerator.cpp:
(JSC::CodeGenerator::CodeGenerator): Ditto.
* jsc.cpp:
(Options::Options): Removed prettyPrint option.
(runWithScripts): Ditto.
(printUsageStatement): Ditto.
(parseArguments): Ditto.
(jscmain): Ditto.
* parser/Grammar.y: Removed use of obsolete ImmediateNumberNode.
* parser/Nodes.cpp:
(JSC::ThrowableExpressionData::emitThrowError): Use inline functions
instead of direct member access for ThrowableExpressionData values.
(JSC::BracketAccessorNode::emitCode): Ditto.
(JSC::DotAccessorNode::emitCode): Ditto.
(JSC::NewExprNode::emitCode): Ditto.
(JSC::EvalFunctionCallNode::emitCode): Ditto.
(JSC::FunctionCallValueNode::emitCode): Ditto.
(JSC::FunctionCallResolveNode::emitCode): Ditto.
(JSC::FunctionCallBracketNode::emitCode): Ditto.
(JSC::FunctionCallDotNode::emitCode): Ditto.
(JSC::PostfixResolveNode::emitCode): Ditto.
(JSC::PostfixBracketNode::emitCode): Ditto.
(JSC::PostfixDotNode::emitCode): Ditto.
(JSC::DeleteResolveNode::emitCode): Ditto.
(JSC::DeleteBracketNode::emitCode): Ditto.
(JSC::DeleteDotNode::emitCode): Ditto.
(JSC::PrefixResolveNode::emitCode): Ditto.
(JSC::PrefixBracketNode::emitCode): Ditto.
(JSC::PrefixDotNode::emitCode): Ditto.
(JSC::ThrowableBinaryOpNode::emitCode): Ditto.
(JSC::InstanceOfNode::emitCode): Ditto.
(JSC::ReadModifyResolveNode::emitCode): Ditto.
(JSC::AssignResolveNode::emitCode): Ditto.
(JSC::AssignDotNode::emitCode): Ditto.
(JSC::ReadModifyDotNode::emitCode): Ditto.
(JSC::AssignBracketNode::emitCode): Ditto.
(JSC::ReadModifyBracketNode::emitCode): Ditto.
(JSC::statementListEmitCode): Take a const StatementVector instead
of a non-const one. Also removed unused statementListPushFIFO.
(JSC::ForInNode::emitCode): Inline functions instead of member access.
(JSC::ThrowNode::emitCode): Ditto.
(JSC::EvalNode::emitCode): Ditto.
(JSC::FunctionBodyNode::emitCode): Ditto.
(JSC::ProgramNode::emitCode): Ditto.
* parser/Nodes.h: Removed unused includes and forward declarations.
Removed Precedence enum. Made many more members private instead of
protected or public. Removed unused NodeStack typedef. Moved the
VarStack and FunctionStack typedefs from Node to ScopeNode. Made
Node::emitCode pure virtual and changed classes that don't emit
any code to inherit from ParserRefCounted rather than Node.
Moved isReturnNode from Node to StatementNode. Removed the
streamTo, precedence, and needsParensIfLeftmost functions from
all classes. Removed the ImmediateNumberNode class and make
NumberNode::setValue nonvirtual.
* parser/nodes2string.cpp: Removed.
2008-11-09 Darin Adler <darin@apple.com>
Reviewed by Sam Weinig and Maciej Stachowiak.
Includes some work done by Chris Brichford.
- fix https://bugs.webkit.org/show_bug.cgi?id=14886
Stack overflow due to deeply nested parse tree doing repeated string concatentation
Test: fast/js/large-expressions.html
1) Code generation is recursive, so takes stack proportional to the complexity
of the source code expression. Fixed by setting an arbitrary recursion limit
of 10,000 nodes.
2) Destruction of the syntax tree was recursive. Fixed by introducing a
non-recursive mechanism for destroying the tree.
* bytecompiler/CodeGenerator.cpp:
(JSC::CodeGenerator::CodeGenerator): Initialize depth to 0.
(JSC::CodeGenerator::emitThrowExpressionTooDeepException): Added. Emits the code
to throw a "too deep" exception.
* bytecompiler/CodeGenerator.h:
(JSC::CodeGenerator::emitNode): Check depth and emit an exception if we exceed
the maximum depth.
* parser/Nodes.cpp:
(JSC::NodeReleaser::releaseAllNodes): Added. To be called inside node destructors
to avoid recursive calls to destructors for nodes inside this one.
(JSC::NodeReleaser::release): Added. To be called inside releaseNodes functions.
Also added releaseNodes functions and calls to releaseAllNodes inside destructors
for each class derived from Node that has RefPtr to other nodes.
(JSC::NodeReleaser::adopt): Added. Used by the release function.
(JSC::NodeReleaser::adoptFunctionBodyNode): Added.
* parser/Nodes.h: Added declarations of releaseNodes and destructors in all classes
that needed it. Eliminated use of ListRefPtr and releaseNext, which are the two parts
of an older solution to the non-recursive destruction problem that works only for
lists, whereas the new solution works for other graphs. Changed ReverseBinaryOpNode
to use BinaryOpNode as a base class to avoid some duplicated code.
2008-11-08 Kevin Ollivier <kevino@theolliviers.com>
wx build fixes after addition of JSCore parser and bycompiler dirs. Also cleanup
the JSCore Bakefile's group names to be consistent.
* JavaScriptCoreSources.bkl:
* jscore.bkl:
2008-11-07 Cameron Zwarich <zwarich@apple.com>
Reviewed by Geoff Garen.
Bug 21801: REGRESSION (r37821): YUI date formatting JavaScript puts the letter 'd' in place of the day
<https://bugs.webkit.org/show_bug.cgi?id=21801>
Fix the constant register check in the 'typeof' optimization in
CodeGenerator, which was completely broken after r37821.
* bytecompiler/CodeGenerator.cpp:
(JSC::CodeGenerator::emitEqualityOp):
2008-11-07 Cameron Zwarich <zwarich@apple.com>
Reviewed by Geoff Garen.
Bug 22129: Move CTI::isConstant() to CodeBlock
<https://bugs.webkit.org/show_bug.cgi?id=22129>
* VM/CTI.cpp:
(JSC::CTI::emitGetArg):
(JSC::CTI::emitGetPutArg):
(JSC::CTI::getConstantImmediateNumericArg):
(JSC::CTI::printOpcodeOperandTypes):
(JSC::CTI::privateCompileMainPass):
* VM/CTI.h:
* VM/CodeBlock.h:
(JSC::CodeBlock::isConstant):
2008-11-07 Alp Toker <alp@nuanti.com>
autotools fix. Always use the configured perl binary (which may be
different to the one in $PATH) when generating sources.
* GNUmakefile.am:
2008-11-07 Cameron Zwarich <zwarich@apple.com>
Not reviewed.
Change grammar.cpp to Grammar.cpp and grammar.h to Grammar.h in several
build scripts.
* DerivedSources.make:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCoreSources.bkl:
2008-11-07 Alp Toker <alp@nuanti.com>
More grammar.cpp -> Grammar.cpp build fixes.
* AllInOneFile.cpp:
* GNUmakefile.am:
2008-11-07 Simon Hausmann <hausmann@webkit.org>
Fix the build on case-sensitive file systems. grammar.y was renamed to
Grammar.y but Lexer.cpp includes grammar.h. The build bots didn't
notice this change because of stale files.
* parser/Lexer.cpp:
2008-11-07 Cameron Zwarich <zwarich@apple.com>
Reviewed by Alexey Proskuryakov.
Rename the m_nextGlobal, m_nextParameter, and m_nextConstant member
variables of CodeGenerator to m_nextGlobalIndex, m_nextParameterIndex,
and m_nextConstantIndex respectively. This is to distinguish these from
member variables like m_lastConstant, which are actually RefPtrs to
Registers.
* bytecompiler/CodeGenerator.cpp:
(JSC::CodeGenerator::addGlobalVar):
(JSC::CodeGenerator::allocateConstants):
(JSC::CodeGenerator::CodeGenerator):
(JSC::CodeGenerator::addParameter):
(JSC::CodeGenerator::addConstant):
* bytecompiler/CodeGenerator.h:
2008-11-06 Gavin Barraclough barraclough@apple.com
Reviewed by Oliver Hunt.
Do not make a cti_* call to perform an op_call unless either:
(1) The codeblock for the function body has not been generated.
(2) The number of arguments passed does not match the callee arity.
~1% progression on sunspider --v8
* VM/CTI.cpp:
(JSC::CTI::compileOpCallInitializeCallFrame):
(JSC::CTI::compileOpCall):
(JSC::CTI::privateCompileSlowCases):
* VM/CTI.h:
* VM/Machine.cpp:
(JSC::Machine::cti_op_call_JSFunction):
(JSC::Machine::cti_op_call_arityCheck):
(JSC::Machine::cti_op_construct_JSConstruct):
* VM/Machine.h:
* kjs/nodes.h:
2008-11-06 Cameron Zwarich <zwarich@apple.com>
Reviewed by Geoff Garen.
Move the remaining files in the kjs subdirectory of JavaScriptCore to
a new parser subdirectory, and remove the kjs subdirectory entirely.
* AllInOneFile.cpp:
* DerivedSources.make:
* GNUmakefile.am:
* JavaScriptCore.pri:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.vcproj/WTF/WTF.vcproj:
* JavaScriptCore.vcproj/jsc/jsc.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* JavaScriptCoreSources.bkl:
* VM/CodeBlock.h:
* VM/ExceptionHelpers.cpp:
* VM/SamplingTool.h:
* bytecompiler/CodeGenerator.h:
* jsc.pro:
* jscore.bkl:
* kjs: Removed.
* kjs/NodeInfo.h: Removed.
* kjs/Parser.cpp: Removed.
* kjs/Parser.h: Removed.
* kjs/ResultType.h: Removed.
* kjs/SourceCode.h: Removed.
* kjs/SourceProvider.h: Removed.
* kjs/grammar.y: Removed.
* kjs/keywords.table: Removed.
* kjs/lexer.cpp: Removed.
* kjs/lexer.h: Removed.
* kjs/nodes.cpp: Removed.
* kjs/nodes.h: Removed.
* kjs/nodes2string.cpp: Removed.
* parser: Added.
* parser/Grammar.y: Copied from kjs/grammar.y.
* parser/Keywords.table: Copied from kjs/keywords.table.
* parser/Lexer.cpp: Copied from kjs/lexer.cpp.
* parser/Lexer.h: Copied from kjs/lexer.h.
* parser/NodeInfo.h: Copied from kjs/NodeInfo.h.
* parser/Nodes.cpp: Copied from kjs/nodes.cpp.
* parser/Nodes.h: Copied from kjs/nodes.h.
* parser/Parser.cpp: Copied from kjs/Parser.cpp.
* parser/Parser.h: Copied from kjs/Parser.h.
* parser/ResultType.h: Copied from kjs/ResultType.h.
* parser/SourceCode.h: Copied from kjs/SourceCode.h.
* parser/SourceProvider.h: Copied from kjs/SourceProvider.h.
* parser/nodes2string.cpp: Copied from kjs/nodes2string.cpp.
* pcre/pcre.pri:
* pcre/pcre_exec.cpp:
* runtime/FunctionConstructor.cpp:
* runtime/JSActivation.h:
* runtime/JSFunction.h:
* runtime/JSGlobalData.cpp:
* runtime/JSGlobalObjectFunctions.cpp:
* runtime/JSObject.cpp:
(JSC::JSObject::toNumber):
* runtime/RegExp.cpp:
2008-11-06 Adam Roben <aroben@apple.com>
Windows build fix after r38196
* JavaScriptCore.vcproj/jsc/jsc.vcproj: Added bytecompiler/ to the
include path.
2008-11-06 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Sam Weinig.
Create a new bytecompiler subdirectory of JavaScriptCore and move some
relevant files to it.
* AllInOneFile.cpp:
* GNUmakefile.am:
* JavaScriptCore.pri:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* JavaScriptCoreSources.bkl:
* VM/CodeGenerator.cpp: Removed.
* VM/CodeGenerator.h: Removed.
* bytecompiler: Added.
* bytecompiler/CodeGenerator.cpp: Copied from VM/CodeGenerator.cpp.
* bytecompiler/CodeGenerator.h: Copied from VM/CodeGenerator.h.
* bytecompiler/LabelScope.h: Copied from kjs/LabelScope.h.
* jscore.bkl:
* kjs/LabelScope.h: Removed.
2008-11-06 Adam Roben <aroben@apple.com>
Windows clean build fix after r38155
Rubberstamped by Cameron Zwarich.
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Update
the post-build event for the move of create_hash_table out of kjs/.
2008-11-06 Laszlo Gombos <laszlo.1.gombos@nokia.com>
Reviewed by Darin Adler.
https://bugs.webkit.org/show_bug.cgi?id=22107
Bug uncovered during RVCT port in functions not used. get_lt() and
get_gt() takes only one argument - remove second argument where
applicable.
* wtf/AVLTree.h:
(JSC::AVLTree::remove): Remove second argument of get_lt/get_gt().
(JSC::AVLTree::subst): Ditto.
2008-11-06 Alp Toker <alp@nuanti.com>
Reviewed by Cameron Zwarich.
https://bugs.webkit.org/show_bug.cgi?id=22033
[GTK] CTI/Linux r38064 crashes; JIT requires executable memory
Mark pages allocated by the FastMalloc mmap code path executable with
PROT_EXEC. This fixes crashes seen on CPUs and kernels that enforce
non-executable memory (like ExecShield on Fedora Linux) when the JIT
is enabled.
This patch does not resolve the issue on debug builds so affected
developers may still need to pass --disable-jit to configure.
* wtf/TCSystemAlloc.cpp:
(TryMmap):
(TryDevMem):
(TCMalloc_SystemRelease):
2008-11-06 Peter Gal <galpeter@inf.u-szeged.hu>
Reviewed by Cameron Zwarich.
Bug 22099: Make the Qt port build the JSC shell in the correct place
<https://bugs.webkit.org/show_bug.cgi?id=22099>
Adjust include paths and build destination dir for the 'jsc' executable
in the Qt build.
* jsc.pro:
2008-11-06 Kristian Amlie <kristian.amlie@nokia.com>
Reviewed by Simon Hausmann.
Implemented the block allocation on Symbian through heap allocation.
Unfortunately there is no way to allocate virtual memory. The Posix
layer provides mmap() but no anonymous mapping. So this is a very slow
solution but it should work as a start.
* runtime/Collector.cpp:
(JSC::allocateBlock):
(JSC::freeBlock):
2008-11-06 Laszlo Gombos <laszlo.1.gombos@nokia.com>
Reviewed by Simon Hausmann.
Borrow some math functions from the MSVC port to the build with the
RVCT compiler.
* wtf/MathExtras.h:
(isinf):
(isnan):
(signbit):
2008-11-06 Laszlo Gombos <laszlo.1.gombos@nokia.com>
Reviewed by Simon Hausmann.
Include strings.h for strncasecmp().
This is needed for compilation inside Symbian and it is also
confirmed by the man-page on Linux.
* runtime/DateMath.cpp:
2008-11-06 Norbert Leser <norbert.leser@nokia.com>
Reviewed by Simon Hausmann.
Implemented currentThreadStackBase for Symbian.
* runtime/Collector.cpp:
(JSC::currentThreadStackBase):
2008-11-06 Laszlo Gombos <laszlo.1.gombos@nokia.com>
Reviewed by Simon Hausmann.
RVCT does not support tm_gmtoff field, so disable that code just like
for MSVC.
* runtime/DateMath.h:
(JSC::GregorianDateTime::GregorianDateTime):
(JSC::GregorianDateTime::operator tm):
2008-11-06 Kristian Amlie <kristian.amlie@nokia.com>
Reviewed by Simon Hausmann.
Define PLATFORM(UNIX) for S60. Effectively WebKit on S60 is compiled
on top of the Posix layer.
* wtf/Platform.h:
2008-11-06 Norbert Leser <norbert.leser@nokia.com>
Reviewed by Simon Hausmann.
Added __SYMBIAN32__ condition for defining PLATFORM(SYMBIAN).
* wtf/Platform.h:
2008-11-06 Ariya Hidayat <ariya.hidayat@trolltech.com>
Reviewed by Simon Hausmann.
Added WINSCW compiler define for Symbian S60.
* wtf/Platform.h:
2008-11-06 Kristian Amlie <kristian.amlie@nokia.com>
Reviewed by Simon Hausmann.
Use the GCC defines of the WTF_ALIGN* macros for the RVCT and the
MINSCW compiler.
* wtf/Vector.h:
2008-11-06 Kristian Amlie <kristian.amlie@nokia.com>
Reviewed by Simon Hausmann.
Define capabilities of the SYMBIAN platform. Some of the system
headers are actually dependent on RVCT.
* wtf/Platform.h:
2008-11-06 Kristian Amlie <kristian.amlie@nokia.com>
Reviewed by Simon Hausmann.
Add missing stddef.h header needed for compilation in Symbian.
* runtime/Collector.h:
2008-11-06 Kristian Amlie <kristian.amlie@nokia.com>
Reviewed by Simon Hausmann.
Added COMPILER(RVCT) to detect the ARM RVCT compiler used in the Symbian environment.
* wtf/Platform.h:
2008-11-06 Simon Hausmann <hausmann@webkit.org>
Fix the Qt build, adjust include paths after move of jsc.pro.
* jsc.pro:
2008-11-06 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Sam Weinig.
Move kjs/Shell.cpp to the top level of the JavaScriptCore directory and
rename it to jsc.cpp to reflect the name of the binary compiled from it.
* GNUmakefile.am:
* JavaScriptCore.vcproj/jsc/jsc.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* jsc.cpp: Copied from kjs/Shell.cpp.
* jsc.pro:
* jscore.bkl:
* kjs/Shell.cpp: Removed.
2008-11-06 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Sam Weinig.
Move create_hash_table and jsc.pro out of the kjs directory and into the
root directory of JavaScriptCore.
* DerivedSources.make:
* JavaScriptCore.pri:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* create_hash_table: Copied from kjs/create_hash_table.
* jsc.pro: Copied from kjs/jsc.pro.
* kjs/create_hash_table: Removed.
* kjs/jsc.pro: Removed.
* make-generated-sources.sh:
2008-11-05 Gavin Barraclough <barraclough@apple.com>
Reviewed by Maciej Stachowiak.
https://bugs.webkit.org/show_bug.cgi?id=22094
Fix for bug where the callee incorrectly recieves the caller's lexical
global object as this, rather than its own. Implementation closely
follows the spec, passing jsNull, checking in the callee and replacing
with the global object where necessary.
* VM/CTI.cpp:
(JSC::CTI::compileOpCall):
* VM/Machine.cpp:
(JSC::Machine::cti_op_call_NotJSFunction):
(JSC::Machine::cti_op_call_eval):
* runtime/JSCell.h:
(JSC::JSValue::toThisObject):
* runtime/JSImmediate.cpp:
(JSC::JSImmediate::toThisObject):
* runtime/JSImmediate.h:
2008-11-05 Kevin Ollivier <kevino@theolliviers.com>
wx build fix after Operations.cpp move.
* JavaScriptCoreSources.bkl:
2008-11-05 Cameron Zwarich <zwarich@apple.com>
Not reviewed.
Fix the build for case-sensitive build systems and wxWindows.
* JavaScriptCoreSources.bkl:
* kjs/create_hash_table:
2008-11-05 Cameron Zwarich <zwarich@apple.com>
Not reviewed.
Fix the build for case-sensitive build systems.
* JavaScriptCoreSources.bkl:
* kjs/Shell.cpp:
* runtime/Interpreter.cpp:
* runtime/JSArray.cpp:
2008-11-05 Cameron Zwarich <zwarich@apple.com>
Not reviewed.
Fix the build for case-sensitive build systems.
* API/JSBase.cpp:
* API/JSObjectRef.cpp:
* runtime/CommonIdentifiers.h:
* runtime/Identifier.cpp:
* runtime/InitializeThreading.cpp:
* runtime/InternalFunction.h:
* runtime/JSString.h:
* runtime/Lookup.h:
* runtime/PropertyNameArray.h:
* runtime/PropertySlot.h:
* runtime/StructureID.cpp:
* runtime/StructureID.h:
* runtime/UString.cpp:
2008-11-05 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Sam Weinig.
Move more files to the runtime subdirectory of JavaScriptCore.
* API/APICast.h:
* API/JSBase.cpp:
* API/JSCallbackObject.cpp:
* API/JSClassRef.cpp:
* API/JSClassRef.h:
* API/JSStringRefCF.cpp:
* API/JSValueRef.cpp:
* API/OpaqueJSString.cpp:
* API/OpaqueJSString.h:
* AllInOneFile.cpp:
* GNUmakefile.am:
* JavaScriptCore.pri:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* JavaScriptCoreSources.bkl:
* VM/CodeBlock.h:
* VM/CodeGenerator.cpp:
* VM/Machine.cpp:
* VM/RegisterFile.h:
* debugger/Debugger.h:
* kjs/SourceProvider.h:
* kjs/TypeInfo.h: Removed.
* kjs/collector.cpp: Removed.
* kjs/collector.h: Removed.
* kjs/completion.h: Removed.
* kjs/create_hash_table:
* kjs/identifier.cpp: Removed.
* kjs/identifier.h: Removed.
* kjs/interpreter.cpp: Removed.
* kjs/interpreter.h: Removed.
* kjs/lexer.cpp:
* kjs/lexer.h:
* kjs/lookup.cpp: Removed.
* kjs/lookup.h: Removed.
* kjs/nodes.cpp:
* kjs/nodes.h:
* kjs/operations.cpp: Removed.
* kjs/operations.h: Removed.
* kjs/protect.h: Removed.
* kjs/regexp.cpp: Removed.
* kjs/regexp.h: Removed.
* kjs/ustring.cpp: Removed.
* kjs/ustring.h: Removed.
* pcre/pcre_exec.cpp:
* profiler/CallIdentifier.h:
* profiler/Profile.h:
* runtime/ArrayConstructor.cpp:
* runtime/ArrayPrototype.cpp:
* runtime/ArrayPrototype.h:
* runtime/Collector.cpp: Copied from kjs/collector.cpp.
* runtime/Collector.h: Copied from kjs/collector.h.
* runtime/CollectorHeapIterator.h:
* runtime/Completion.h: Copied from kjs/completion.h.
* runtime/ErrorPrototype.cpp:
* runtime/Identifier.cpp: Copied from kjs/identifier.cpp.
* runtime/Identifier.h: Copied from kjs/identifier.h.
* runtime/InitializeThreading.cpp:
* runtime/Interpreter.cpp: Copied from kjs/interpreter.cpp.
* runtime/Interpreter.h: Copied from kjs/interpreter.h.
* runtime/JSCell.h:
* runtime/JSGlobalData.cpp:
* runtime/JSGlobalData.h:
* runtime/JSLock.cpp:
* runtime/JSNumberCell.cpp:
* runtime/JSNumberCell.h:
* runtime/JSObject.cpp:
* runtime/JSValue.h:
* runtime/Lookup.cpp: Copied from kjs/lookup.cpp.
* runtime/Lookup.h: Copied from kjs/lookup.h.
* runtime/MathObject.cpp:
* runtime/NativeErrorPrototype.cpp:
* runtime/NumberPrototype.cpp:
* runtime/Operations.cpp: Copied from kjs/operations.cpp.
* runtime/Operations.h: Copied from kjs/operations.h.
* runtime/PropertyMapHashTable.h:
* runtime/Protect.h: Copied from kjs/protect.h.
* runtime/RegExp.cpp: Copied from kjs/regexp.cpp.
* runtime/RegExp.h: Copied from kjs/regexp.h.
* runtime/RegExpConstructor.cpp:
* runtime/RegExpObject.h:
* runtime/RegExpPrototype.cpp:
* runtime/SmallStrings.h:
* runtime/StringObjectThatMasqueradesAsUndefined.h:
* runtime/StructureID.cpp:
* runtime/StructureID.h:
* runtime/StructureIDTransitionTable.h:
* runtime/SymbolTable.h:
* runtime/TypeInfo.h: Copied from kjs/TypeInfo.h.
* runtime/UString.cpp: Copied from kjs/ustring.cpp.
* runtime/UString.h: Copied from kjs/ustring.h.
* wrec/CharacterClassConstructor.h:
* wrec/WREC.h:
2008-11-05 Geoffrey Garen <ggaren@apple.com>
Suggested by Darin Adler.
Removed two copy constructors that the compiler can generate for us
automatically.
* VM/LabelID.h:
(JSC::LabelID::setLocation):
(JSC::LabelID::offsetFrom):
(JSC::LabelID::ref):
(JSC::LabelID::refCount):
* kjs/LabelScope.h:
2008-11-05 Anders Carlsson <andersca@apple.com>
Fix Snow Leopard build.
* JavaScriptCore.xcodeproj/project.pbxproj:
2008-11-04 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Steve Falkenburg.
Move dtoa.cpp and dtoa.h to the WTF Visual Studio project to reflect
their movement in the filesystem.
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.vcproj/WTF/WTF.vcproj:
2008-11-04 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Sam Weinig.
Move kjs/dtoa.h to the wtf subdirectory of JavaScriptCore.
* AllInOneFile.cpp:
* GNUmakefile.am:
* JavaScriptCore.pri:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* JavaScriptCoreSources.bkl:
* kjs/dtoa.cpp: Removed.
* kjs/dtoa.h: Removed.
* wtf/dtoa.cpp: Copied from kjs/dtoa.cpp.
* wtf/dtoa.h: Copied from kjs/dtoa.h.
2008-11-04 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Sam Weinig.
Move kjs/config.h to the top level of JavaScriptCore.
* GNUmakefile.am:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* config.h: Copied from kjs/config.h.
* kjs/config.h: Removed.
2008-11-04 Darin Adler <darin@apple.com>
Reviewed by Tim Hatcher.
* wtf/ThreadingNone.cpp: Tweak formatting.
2008-11-03 Darin Adler <darin@apple.com>
Reviewed by Tim Hatcher.
- https://bugs.webkit.org/show_bug.cgi?id=22061
create script to check for exit-time destructors
* JavaScriptCore.exp: Changed to export functions rather than
a global for the atomically initialized static mutex.
* JavaScriptCore.xcodeproj/project.pbxproj: Added a script
phase that runs the check-for-exit-time-destructors script.
* wtf/MainThread.cpp:
(WTF::mainThreadFunctionQueueMutex): Changed to leak an object
rather than using an exit time destructor.
(WTF::functionQueue): Ditto.
* wtf/unicode/icu/CollatorICU.cpp:
(WTF::cachedCollatorMutex): Ditto.
* wtf/Threading.h: Changed other platforms to share the Windows
approach where the mutex is internal and the functions are exported.
* wtf/ThreadingGtk.cpp:
(WTF::lockAtomicallyInitializedStaticMutex): Ditto.
(WTF::unlockAtomicallyInitializedStaticMutex): Ditto.
* wtf/ThreadingNone.cpp:
(WTF::lockAtomicallyInitializedStaticMutex): Ditto.
(WTF::unlockAtomicallyInitializedStaticMutex): Ditto.
* wtf/ThreadingPthreads.cpp:
(WTF::threadMapMutex): Changed to leak an object rather than using
an exit time destructor.
(WTF::lockAtomicallyInitializedStaticMutex): Mutex change.
(WTF::unlockAtomicallyInitializedStaticMutex): Ditto.
(WTF::threadMap): Changed to leak an object rather than using
an exit time destructor.
* wtf/ThreadingQt.cpp:
(WTF::lockAtomicallyInitializedStaticMutex): Mutex change.
(WTF::unlockAtomicallyInitializedStaticMutex): Ditto.
* wtf/ThreadingWin.cpp:
(WTF::lockAtomicallyInitializedStaticMutex): Added an assertion.
2008-11-04 Adam Roben <aroben@apple.com>
Windows build fix
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Update
the location of JSStaticScopeObject.{cpp,h}.
2008-11-04 Cameron Zwarich <zwarich@apple.com>
Reviewed by Alexey Proskuryakov.
Move AllInOneFile.cpp to the top level of JavaScriptCore.
* AllInOneFile.cpp: Copied from kjs/AllInOneFile.cpp.
* GNUmakefile.am:
* JavaScriptCore.xcodeproj/project.pbxproj:
* kjs/AllInOneFile.cpp: Removed.
2008-11-04 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Alexey Proskuryakov.
Add NodeInfo.h to the JavaScriptCore Xcode project.
* JavaScriptCore.xcodeproj/project.pbxproj:
2008-11-03 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Maciej Stachowiak.
Move more files into the runtime subdirectory of JavaScriptCore.
* API/JSBase.cpp:
* API/JSCallbackConstructor.cpp:
* API/JSCallbackFunction.cpp:
* API/JSClassRef.cpp:
* API/OpaqueJSString.cpp:
* GNUmakefile.am:
* JavaScriptCore.pri:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* JavaScriptCoreSources.bkl:
* kjs/AllInOneFile.cpp:
* kjs/ArgList.cpp: Removed.
* kjs/ArgList.h: Removed.
* kjs/Arguments.cpp: Removed.
* kjs/Arguments.h: Removed.
* kjs/BatchedTransitionOptimizer.h: Removed.
* kjs/CollectorHeapIterator.h: Removed.
* kjs/CommonIdentifiers.cpp: Removed.
* kjs/CommonIdentifiers.h: Removed.
* kjs/ExecState.cpp: Removed.
* kjs/ExecState.h: Removed.
* kjs/GetterSetter.cpp: Removed.
* kjs/GetterSetter.h: Removed.
* kjs/InitializeThreading.cpp: Removed.
* kjs/InitializeThreading.h: Removed.
* kjs/JSActivation.cpp: Removed.
* kjs/JSActivation.h: Removed.
* kjs/JSGlobalData.cpp: Removed.
* kjs/JSGlobalData.h: Removed.
* kjs/JSLock.cpp: Removed.
* kjs/JSLock.h: Removed.
* kjs/JSStaticScopeObject.cpp: Removed.
* kjs/JSStaticScopeObject.h: Removed.
* kjs/JSType.h: Removed.
* kjs/PropertyNameArray.cpp: Removed.
* kjs/PropertyNameArray.h: Removed.
* kjs/ScopeChain.cpp: Removed.
* kjs/ScopeChain.h: Removed.
* kjs/ScopeChainMark.h: Removed.
* kjs/SymbolTable.h: Removed.
* kjs/Tracing.d: Removed.
* kjs/Tracing.h: Removed.
* runtime/ArgList.cpp: Copied from kjs/ArgList.cpp.
* runtime/ArgList.h: Copied from kjs/ArgList.h.
* runtime/Arguments.cpp: Copied from kjs/Arguments.cpp.
* runtime/Arguments.h: Copied from kjs/Arguments.h.
* runtime/BatchedTransitionOptimizer.h: Copied from kjs/BatchedTransitionOptimizer.h.
* runtime/CollectorHeapIterator.h: Copied from kjs/CollectorHeapIterator.h.
* runtime/CommonIdentifiers.cpp: Copied from kjs/CommonIdentifiers.cpp.
* runtime/CommonIdentifiers.h: Copied from kjs/CommonIdentifiers.h.
* runtime/ExecState.cpp: Copied from kjs/ExecState.cpp.
* runtime/ExecState.h: Copied from kjs/ExecState.h.
* runtime/GetterSetter.cpp: Copied from kjs/GetterSetter.cpp.
* runtime/GetterSetter.h: Copied from kjs/GetterSetter.h.
* runtime/InitializeThreading.cpp: Copied from kjs/InitializeThreading.cpp.
* runtime/InitializeThreading.h: Copied from kjs/InitializeThreading.h.
* runtime/JSActivation.cpp: Copied from kjs/JSActivation.cpp.
* runtime/JSActivation.h: Copied from kjs/JSActivation.h.
* runtime/JSGlobalData.cpp: Copied from kjs/JSGlobalData.cpp.
* runtime/JSGlobalData.h: Copied from kjs/JSGlobalData.h.
* runtime/JSLock.cpp: Copied from kjs/JSLock.cpp.
* runtime/JSLock.h: Copied from kjs/JSLock.h.
* runtime/JSStaticScopeObject.cpp: Copied from kjs/JSStaticScopeObject.cpp.
* runtime/JSStaticScopeObject.h: Copied from kjs/JSStaticScopeObject.h.
* runtime/JSType.h: Copied from kjs/JSType.h.
* runtime/PropertyNameArray.cpp: Copied from kjs/PropertyNameArray.cpp.
* runtime/PropertyNameArray.h: Copied from kjs/PropertyNameArray.h.
* runtime/ScopeChain.cpp: Copied from kjs/ScopeChain.cpp.
* runtime/ScopeChain.h: Copied from kjs/ScopeChain.h.
* runtime/ScopeChainMark.h: Copied from kjs/ScopeChainMark.h.
* runtime/SymbolTable.h: Copied from kjs/SymbolTable.h.
* runtime/Tracing.d: Copied from kjs/Tracing.d.
* runtime/Tracing.h: Copied from kjs/Tracing.h.
2008-11-03 Sam Weinig <sam@webkit.org>
Reviewed by Mark Rowe.
Move #define to turn on dumping StructureID statistics to StructureID.cpp so that
turning it on does not require a full rebuild.
* runtime/StructureID.cpp:
(JSC::StructureID::dumpStatistics):
* runtime/StructureID.h:
2008-11-03 Alp Toker <alp@nuanti.com>
Reviewed by Geoffrey Garen.
Fix warning when building on Darwin without JSC_MULTIPLE_THREADS
enabled.
* kjs/InitializeThreading.cpp:
2008-11-02 Matt Lilek <webkit@mattlilek.com>
Reviewed by Cameron Zwarich.
Bug 22042: REGRESSION(r38066): ASSERTION FAILED: source in CodeBlock
<https://bugs.webkit.org/show_bug.cgi?id=22042>
Rename parameter name to avoid ASSERT.
* VM/CodeBlock.h:
(JSC::CodeBlock::CodeBlock):
(JSC::ProgramCodeBlock::ProgramCodeBlock):
(JSC::EvalCodeBlock::EvalCodeBlock):
2008-11-02 Cameron Zwarich <zwarich@apple.com>
Reviewed by Oliver Hunt.
Bug 22035: Remove the '_' suffix on constructor parameter names for structs
<https://bugs.webkit.org/show_bug.cgi?id=22035>
* API/JSCallbackObject.h:
(JSC::JSCallbackObject::JSCallbackObjectData::JSCallbackObjectData):
* VM/CodeBlock.h:
(JSC::CodeBlock::CodeBlock):
(JSC::ProgramCodeBlock::ProgramCodeBlock):
(JSC::EvalCodeBlock::EvalCodeBlock):
* wrec/WREC.h:
(JSC::Quantifier::Quantifier):
2008-10-31 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Geoff Garen.
Rename SourceRange.h to SourceCode.h.
* API/JSBase.cpp:
* GNUmakefile.am:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* VM/CodeBlock.h:
* kjs/SourceCode.h: Copied from kjs/SourceRange.h.
* kjs/SourceRange.h: Removed.
* kjs/grammar.y:
* kjs/lexer.h:
* kjs/nodes.cpp:
(JSC::ForInNode::ForInNode):
* kjs/nodes.h:
(JSC::ThrowableExpressionData::setExceptionSourceCode):
2008-10-31 Cameron Zwarich <zwarich@apple.com>
Reviewed by Darin Adler.
Bug 22019: Move JSC::Interpreter::shouldPrintExceptions() to WebCore::Console
<https://bugs.webkit.org/show_bug.cgi?id=22019>
The JSC::Interpreter::shouldPrintExceptions() function is not used at
all in JavaScriptCore, so it should be moved to WebCore::Console, its
only user.
* JavaScriptCore.exp:
* kjs/interpreter.cpp:
* kjs/interpreter.h:
2008-10-31 Cameron Zwarich <zwarich@apple.com>
Not reviewed.
Windows build fix.
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
2008-10-31 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Sam Weinig.
Remove the call to Interpreter::setShouldPrintExceptions() from the
GlobalObject constructor in the shell. The shouldPrintExceptions()
information is not used anywhere in JavaScriptCore, only in WebCore.
* kjs/Shell.cpp:
(GlobalObject::GlobalObject):
2008-10-31 Kevin Ollivier <kevino@theolliviers.com>
wxMSW build fix.
* wtf/Threading.h:
2008-10-31 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Sam Weinig.
Move more files from the kjs subdirectory of JavaScriptCore to the
runtime subdirectory.
* GNUmakefile.am:
* JavaScriptCore.pri:
* JavaScriptCore.xcodeproj/project.pbxproj:
* JavaScriptCoreSources.bkl:
* kjs/AllInOneFile.cpp:
* kjs/RegExpConstructor.cpp: Removed.
* kjs/RegExpConstructor.h: Removed.
* kjs/RegExpMatchesArray.h: Removed.
* kjs/RegExpObject.cpp: Removed.
* kjs/RegExpObject.h: Removed.
* kjs/RegExpPrototype.cpp: Removed.
* kjs/RegExpPrototype.h: Removed.
* runtime/RegExpConstructor.cpp: Copied from kjs/RegExpConstructor.cpp.
* runtime/RegExpConstructor.h: Copied from kjs/RegExpConstructor.h.
* runtime/RegExpMatchesArray.h: Copied from kjs/RegExpMatchesArray.h.
* runtime/RegExpObject.cpp: Copied from kjs/RegExpObject.cpp.
* runtime/RegExpObject.h: Copied from kjs/RegExpObject.h.
* runtime/RegExpPrototype.cpp: Copied from kjs/RegExpPrototype.cpp.
* runtime/RegExpPrototype.h: Copied from kjs/RegExpPrototype.h.
2008-10-31 Mark Rowe <mrowe@apple.com>
Revert an incorrect portion of r38034.
* profiler/ProfilerServer.mm:
2008-10-31 Mark Rowe <mrowe@apple.com>
Fix the 64-bit build.
Disable strict aliasing in ProfilerServer.mm as it leads to the compiler being unhappy
with the common Obj-C idiom self = [super init];
* JavaScriptCore.xcodeproj/project.pbxproj:
2008-10-31 Cameron Zwarich <zwarich@apple.com>
Reviewed by Alexey Proskuryakov.
Change a header guard to match our coding style.
* kjs/InitializeThreading.h:
2008-10-30 Geoffrey Garen <ggaren@apple.com>
Reviewed by Oliver Hunt.
Fixed a small bit of https://bugs.webkit.org/show_bug.cgi?id=21962
AST uses way too much memory
Removed a word from StatementNode by nixing LabelStack and turning it
into a compile-time data structure managed by CodeGenerator.
v8 tests and SunSpider, run by Gavin, report no change.
* GNUmakefile.am:
* JavaScriptCore.order:
* JavaScriptCore.pri:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* kjs/AllInOneFile.cpp:
* JavaScriptCoreSources.bkl: I sure hope this builds!
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::CodeGenerator):
(JSC::CodeGenerator::newLabelScope):
(JSC::CodeGenerator::breakTarget):
(JSC::CodeGenerator::continueTarget):
* VM/CodeGenerator.h: Nixed the JumpContext system because it depended
on a LabelStack in the AST, and it was a little cumbersome on the client
side. Replaced with LabelScope, which tracks all break / continue
information in the CodeGenerator, just like we track LabelIDs and other
stacks of compile-time data.
* kjs/LabelScope.h: Added.
(JSC::LabelScope::):
(JSC::LabelScope::LabelScope):
(JSC::LabelScope::ref):
(JSC::LabelScope::deref):
(JSC::LabelScope::refCount):
(JSC::LabelScope::breakTarget):
(JSC::LabelScope::continueTarget):
(JSC::LabelScope::type):
(JSC::LabelScope::name):
(JSC::LabelScope::scopeDepth): Simple abstraction for holding everything
you might want to know about a break-able / continue-able scope.
* kjs/LabelStack.cpp: Removed.
* kjs/LabelStack.h: Removed.
* kjs/grammar.y: No need to push labels at parse time -- we don't store
LabelStacks in the AST anymore.
* kjs/nodes.cpp:
(JSC::DoWhileNode::emitCode):
(JSC::WhileNode::emitCode):
(JSC::ForNode::emitCode):
(JSC::ForInNode::emitCode):
(JSC::ContinueNode::emitCode):
(JSC::BreakNode::emitCode):
(JSC::SwitchNode::emitCode):
(JSC::LabelNode::emitCode):
* kjs/nodes.h:
(JSC::StatementNode::):
(JSC::LabelNode::): Use LabelScope where we used to use JumpContext.
Simplified a bunch of code. Touched up label-related error messages a
bit.
* kjs/nodes2string.cpp:
(JSC::LabelNode::streamTo): Updated for rename.
2008-10-31 Cameron Zwarich <zwarich@apple.com>
Reviewed by Darin Adler.
Bug 22005: Move StructureIDChain into its own file
<https://bugs.webkit.org/show_bug.cgi?id=22005>
* GNUmakefile.am:
* JavaScriptCore.pri:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* JavaScriptCoreSources.bkl:
* runtime/StructureID.cpp:
* runtime/StructureID.h:
* runtime/StructureIDChain.cpp: Copied from runtime/StructureID.cpp.
* runtime/StructureIDChain.h: Copied from runtime/StructureID.h.
2008-10-31 Steve Falkenburg <sfalken@apple.com>
Build fix.
* JavaScriptCore.vcproj/jsc/jsc.vcproj:
2008-10-31 Steve Falkenburg <sfalken@apple.com>
Build fix.
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
2008-10-31 Darin Adler <darin@apple.com>
Reviewed by Dan Bernstein.
- fix storage leak seen on buildbot
Some other cleanup too. The storage leak was caused by the fact
that HashTraits<CallIdentifier>::needsDestruction was false, so
the call identifier objects didn't get deleted.
* profiler/CallIdentifier.h:
Added a default constructor to create empty call identifiers.
Changed the normal constructor to use const UString&
to avoid extra copying and reference count thrash.
Removed the explicit copy constructor definition, since it's what
the compiler will automatically generate. (Rule of thumb: Either
you need both a custom copy constructor and a custom assignment
operator, or neither.)
Moved the CallIdentifier hash function out of the WTF namespace;
there's no reason to put it there.
Changed the CallIdentifier hash function to be a struct rather than
a specialization of the IntHash struct template. Having it be
a specialization made no sense, since CallIdentifier is not an integer,
and did no good.
Removed explicit definition of emptyValueIsZero in the hash traits,
since inheriting from GenericHashTraits already makes that false.
Removed explicit definition of emptyValue, instead relying on the
default constructor and GenericHashTraits.
Removed explicit definition of needsDestruction, because we want it
to have its default value: true, not false. This fixes the leak!
Changed constructDeletedValue and isDeletedValue to use a line number
of numeric_limits<unsigned>::max() to indicate a value is deleted.
Previously this used empty strings for the empty value and null strings
for the deleted value, but it's more efficient to use null for both.
2008-10-31 Timothy Hatcher <timothy@apple.com>
Emit the WillExecuteStatement debugger hook before the for loop body
when the statement node for the body isn't a block. This allows
breakpoints on those statements in the Web Inspector.
https://bugs.webkit.org/show_bug.cgi?id=22004
Reviewed by Darin Adler.
* kjs/nodes.cpp:
(JSC::ForNode::emitCode): Emit the WillExecuteStatement
debugger hook before the statement node if isn't a block.
Also emit the WillExecuteStatement debugger hook for the
loop as the first op-code.
(JSC::ForInNode::emitCode): Ditto.
2008-10-31 Timothy Hatcher <timothy@apple.com>
Fixes console warnings about not having an autorelease pool.
Also fixes the build for Snow Leopard, by including individual
Foundation headers instead of Foundation.h.
https://bugs.webkit.org/show_bug.cgi?id=21995
Reviewed by Oliver Hunt.
* profiler/ProfilerServer.mm:
(-[ProfilerServer init]): Create a NSAutoreleasePool and drain it.
2008-10-31 Cameron Zwarich <zwarich@apple.com>
Not reviewed.
Speculative wxWindows build fix.
* JavaScriptCoreSources.bkl:
* jscore.bkl:
2008-10-31 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Maciej Stachowiak.
Move VM/JSPropertyNameIterator.cpp and VM/JSPropertyNameIterator.h to
the runtime directory.
* GNUmakefile.am:
* JavaScriptCore.pri:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* JavaScriptCoreSources.bkl:
* VM/JSPropertyNameIterator.cpp: Removed.
* VM/JSPropertyNameIterator.h: Removed.
* runtime/JSPropertyNameIterator.cpp: Copied from VM/JSPropertyNameIterator.cpp.
* runtime/JSPropertyNameIterator.h: Copied from VM/JSPropertyNameIterator.h.
2008-10-31 Cameron Zwarich <zwarich@apple.com>
Not reviewed.
Speculative wxWindows build fix.
* jscore.bkl:
2008-10-30 Mark Rowe <mrowe@apple.com>
Reviewed by Jon Homeycutt.
Explicitly default to building for only the native architecture in debug and release builds.
* Configurations/DebugRelease.xcconfig:
2008-10-30 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Sam Weinig.
Create a debugger directory in JavaScriptCore and move the relevant
files to it.
* GNUmakefile.am:
* JavaScriptCore.pri:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* VM/CodeBlock.cpp:
* VM/CodeGenerator.h:
* VM/Machine.cpp:
* debugger: Added.
* debugger/Debugger.cpp: Copied from kjs/debugger.cpp.
* debugger/Debugger.h: Copied from kjs/debugger.h.
* debugger/DebuggerCallFrame.cpp: Copied from kjs/DebuggerCallFrame.cpp.
* debugger/DebuggerCallFrame.h: Copied from kjs/DebuggerCallFrame.h.
* kjs/AllInOneFile.cpp:
* kjs/DebuggerCallFrame.cpp: Removed.
* kjs/DebuggerCallFrame.h: Removed.
* kjs/Parser.cpp:
* kjs/Parser.h:
* kjs/debugger.cpp: Removed.
* kjs/debugger.h: Removed.
* kjs/interpreter.cpp:
* kjs/nodes.cpp:
* runtime/FunctionConstructor.cpp:
* runtime/JSGlobalObject.cpp:
2008-10-30 Benjamin K. Stuhl <bks24@cornell.edu>
gcc 4.3.3/linux-x86 generates "suggest parentheses around && within ||"
warnings; add some parentheses to disambiguate things. No functional
changes, so no tests.
https://bugs.webkit.org/show_bug.cgi?id=21973
Add parentheses to clean up some gcc warnings
Reviewed by Dan Bernstein.
* wtf/ASCIICType.h:
(WTF::isASCIIAlphanumeric):
(WTF::isASCIIHexDigit):
2008-10-30 Kevin Lindeman <klindeman@apple.com>
Adds ProfilerServer, which is a distributed notification listener
that allows starting and stopping the profiler remotely for use
in conjunction with the profiler's DTace probes.
https://bugs.webkit.org/show_bug.cgi?id=21719
Reviewed by Timothy Hatcher.
* JavaScriptCore.xcodeproj/project.pbxproj:
* kjs/JSGlobalData.cpp:
(JSC::JSGlobalData::JSGlobalData): Calls startProfilerServerIfNeeded.
* profiler/ProfilerServer.h: Added.
* profiler/ProfilerServer.mm: Added.
(+[ProfilerServer sharedProfileServer]):
(-[ProfilerServer init]):
(-[ProfilerServer startProfiling]):
(-[ProfilerServer stopProfiling]):
(JSC::startProfilerServerIfNeeded):
2008-10-30 Kevin Ollivier <kevino@theolliviers.com>
wx build fix after PropertyMap and StructureID merge.
* JavaScriptCoreSources.bkl:
2008-10-30 Cameron Zwarich <zwarich@apple.com>
Reviewed by Mark Rowe.
Change the JavaScriptCore Xcode project to use relative paths for the
PCRE source files.
* JavaScriptCore.xcodeproj/project.pbxproj:
2008-10-30 Sam Weinig <sam@webkit.org>
Reviewed by Cameron Zwarich and Geoffrey Garen.
Fix for https://bugs.webkit.org/show_bug.cgi?id=21989
Merge PropertyMap and StructureID
- Move PropertyMap code into StructureID in preparation for lazily
creating the map on gets.
- Make remove with transition explicit by adding removePropertyTransition.
- Make the put/remove without transition explicit.
- Make cache invalidation part of put/remove without transition.
1% speedup on SunSpider; 0.5% speedup on v8 suite.
* GNUmakefile.am:
* JavaScriptCore.exp:
* JavaScriptCore.pri:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* JavaScriptCoreSources.bkl:
* kjs/AllInOneFile.cpp:
* kjs/identifier.h:
* runtime/JSObject.cpp:
(JSC::JSObject::removeDirect):
* runtime/JSObject.h:
(JSC::JSObject::putDirect):
* runtime/PropertyMap.cpp: Removed.
* runtime/PropertyMap.h: Removed.
* runtime/PropertyMapHashTable.h: Copied from runtime/PropertyMap.h.
* runtime/StructureID.cpp:
(JSC::StructureID::dumpStatistics):
(JSC::StructureID::StructureID):
(JSC::StructureID::~StructureID):
(JSC::StructureID::getEnumerablePropertyNames):
(JSC::StructureID::addPropertyTransition):
(JSC::StructureID::removePropertyTransition):
(JSC::StructureID::toDictionaryTransition):
(JSC::StructureID::changePrototypeTransition):
(JSC::StructureID::getterSetterTransition):
(JSC::StructureID::addPropertyWithoutTransition):
(JSC::StructureID::removePropertyWithoutTransition):
(JSC::PropertyMapStatisticsExitLogger::~PropertyMapStatisticsExitLogger):
(JSC::StructureID::checkConsistency):
(JSC::StructureID::copyPropertyTable):
(JSC::StructureID::get):
(JSC::StructureID::put):
(JSC::StructureID::remove):
(JSC::StructureID::insertIntoPropertyMapHashTable):
(JSC::StructureID::expandPropertyMapHashTable):
(JSC::StructureID::createPropertyMapHashTable):
(JSC::StructureID::rehashPropertyMapHashTable):
(JSC::comparePropertyMapEntryIndices):
(JSC::StructureID::getEnumerablePropertyNamesInternal):
* runtime/StructureID.h:
(JSC::StructureID::propertyStorageSize):
(JSC::StructureID::isEmpty):
(JSC::StructureID::get):
2008-10-30 Cameron Zwarich <zwarich@apple.com>
Reviewed by Oliver Hunt.
Bug 21987: CTI::putDoubleResultToJSNumberCellOrJSImmediate() hardcodes its result register
<https://bugs.webkit.org/show_bug.cgi?id=21987>
CTI::putDoubleResultToJSNumberCellOrJSImmediate() hardcodes its result
register as ecx, but it should be tempReg1, which is ecx at all of its
callsites.
* VM/CTI.cpp:
(JSC::CTI::putDoubleResultToJSNumberCellOrJSImmediate):
2008-10-30 Cameron Zwarich <zwarich@apple.com>
Reviewed by Sam Weinig.
Bug 21985: Opcodes should use eax as their destination register whenever possible
<https://bugs.webkit.org/show_bug.cgi?id=21985>
Change more opcodes to use eax as the register for their final result,
and change calls to emitPutResult() that pass eax to rely on the default
value of eax.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
2008-10-30 Alp Toker <alp@nuanti.com>
Build fix attempt for older gcc on the trunk-mac-intel build bot
(error: initializer for scalar variable requires one element).
Modify the initializer syntax slightly with an additional comma.
* VM/Machine.cpp:
(JSC::Machine::cti_op_call_JSFunction):
(JSC::Machine::cti_op_construct_JSConstruct):
(JSC::Machine::cti_op_resolve_func):
(JSC::Machine::cti_op_post_inc):
(JSC::Machine::cti_op_resolve_with_base):
(JSC::Machine::cti_op_post_dec):
2008-10-30 Alp Toker <alp@nuanti.com>
Reviewed by Alexey Proskuryakov.
https://bugs.webkit.org/show_bug.cgi?id=21571
VoidPtrPair breaks CTI on Linux
The VoidPtrPair return change made in r37457 does not work on Linux
since POD structs aren't passed in registers.
This patch uses a union to vectorize VoidPtrPair to a uint64_t and
matches Darwin/MSVC fixing CTI/WREC on Linux.
Alexey reports no measurable change in Mac performance with this fix.
* VM/Machine.cpp:
(JSC::Machine::cti_op_call_JSFunction):
(JSC::Machine::cti_op_construct_JSConstruct):
(JSC::Machine::cti_op_resolve_func):
(JSC::Machine::cti_op_post_inc):
(JSC::Machine::cti_op_resolve_with_base):
(JSC::Machine::cti_op_post_dec):
* VM/Machine.h:
(JSC::):
2008-10-29 Oliver Hunt <oliver@apple.com>
Reviewed by Geoff Garen.
Initial work to reduce cost of JSNumberCell allocation
This does the initial work needed to bring more of number
allocation into CTI code directly, rather than just falling
back onto the slow paths if we can't guarantee that a number
cell can be reused.
Initial implementation only used by op_negate to make sure
it all works. In a negate heavy (though not dominated) test
it results in a 10% win in the non-reusable cell case.
* VM/CTI.cpp:
(JSC::):
(JSC::CTI::emitAllocateNumber):
(JSC::CTI::emitNakedFastCall):
(JSC::CTI::emitArithIntToImmWithJump):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
* VM/CTI.h:
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitUnaryOp):
* VM/CodeGenerator.h:
(JSC::CodeGenerator::emitToJSNumber):
(JSC::CodeGenerator::emitTypeOf):
(JSC::CodeGenerator::emitGetPropertyNames):
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
* VM/Machine.h:
* kjs/ResultType.h:
(JSC::ResultType::isReusableNumber):
(JSC::ResultType::toInt):
* kjs/nodes.cpp:
(JSC::UnaryOpNode::emitCode):
(JSC::BinaryOpNode::emitCode):
(JSC::EqualNode::emitCode):
* masm/X86Assembler.h:
(JSC::X86Assembler::):
(JSC::X86Assembler::negl_r):
(JSC::X86Assembler::xorpd_mr):
* runtime/JSNumberCell.h:
(JSC::JSNumberCell::JSNumberCell):
2008-10-29 Steve Falkenburg <sfalken@apple.com>
<rdar://problem/6326563> Crash on launch
For Windows, export explicit functions rather than exporting data for atomicallyInitializedStaticMutex.
Exporting data from a DLL on Windows requires specifying __declspec(dllimport) in the header used by
callers, but __declspec(dllexport) when defined in the DLL implementation. By instead exporting
the explicit lock/unlock functions, we can avoid this.
Fixes a crash on launch, since we were previously erroneously exporting atomicallyInitializedStaticMutex as a function.
Reviewed by Darin Adler.
* wtf/Threading.h:
(WTF::lockAtomicallyInitializedStaticMutex):
(WTF::unlockAtomicallyInitializedStaticMutex):
* wtf/ThreadingWin.cpp:
(WTF::lockAtomicallyInitializedStaticMutex):
(WTF::unlockAtomicallyInitializedStaticMutex):
2008-10-29 Sam Weinig <sam@webkit.org>
Reviewed by Oliver Hunt.
Remove direct use of PropertyMap.
* JavaScriptCore.exp:
* runtime/JSObject.cpp:
(JSC::JSObject::mark):
(JSC::JSObject::put):
(JSC::JSObject::deleteProperty):
(JSC::JSObject::getPropertyAttributes):
(JSC::JSObject::removeDirect):
* runtime/JSObject.h:
(JSC::JSObject::getDirect):
(JSC::JSObject::getDirectLocation):
(JSC::JSObject::hasCustomProperties):
(JSC::JSObject::JSObject):
(JSC::JSObject::putDirect):
* runtime/PropertyMap.cpp:
(JSC::PropertyMap::get):
* runtime/PropertyMap.h:
(JSC::PropertyMap::isEmpty):
(JSC::PropertyMap::get):
* runtime/StructureID.cpp:
(JSC::StructureID::dumpStatistics):
* runtime/StructureID.h:
(JSC::StructureID::propertyStorageSize):
(JSC::StructureID::get):
(JSC::StructureID::put):
(JSC::StructureID::remove):
(JSC::StructureID::isEmpty):
2008-10-29 Sam Weinig <sam@webkit.org>
Reviewed by Geoffrey Garen.
Rename and move the StructureID transition table to its own file.
* GNUmakefile.am:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* runtime/StructureID.cpp:
(JSC::StructureID::addPropertyTransition):
* runtime/StructureID.h:
(JSC::StructureID::):
* runtime/StructureIDTransitionTable.h: Copied from runtime/StructureID.h.
(JSC::StructureIDTransitionTableHash::hash):
(JSC::StructureIDTransitionTableHash::equal):
2008-10-29 Sam Weinig <sam@webkit.org>
Reviewed by Cameron Zwarich.
Fix for https://bugs.webkit.org/show_bug.cgi?id=21958
Pack bits in StructureID to reduce the size of each StructureID by 2 words.
* runtime/PropertyMap.h:
(JSC::PropertyMap::propertyMapSize):
* runtime/StructureID.cpp:
(JSC::StructureID::dumpStatistics): Add additional size statistics when dumping.
(JSC::StructureID::StructureID):
* runtime/StructureID.h:
2008-10-29 Kevin Ollivier <kevino@theolliviers.com>
wx build fixes after addition of runtime and ImageBuffer changes.
* JavaScriptCoreSources.bkl:
* jscore.bkl:
2008-10-29 Timothy Hatcher <timothy@apple.com>
Emit the WillExecuteStatement debugger hook before the "else" body
when there is no block for the "else" body. This allows breakpoints
on those statements in the Web Inspector.
https://bugs.webkit.org/show_bug.cgi?id=21944
Reviewed by Maciej Stachowiak.
* kjs/nodes.cpp:
(JSC::IfElseNode::emitCode): Emit the WillExecuteStatement
debugger hook before the else node if isn't a block.
2008-10-29 Alexey Proskuryakov <ap@webkit.org>
Build fix.
* JavaScriptCore.exp: Export HashTable::deleteTable().
2008-10-28 Alp Toker <alp@nuanti.com>
Fix builddir != srcdir builds after kjs -> runtime breakage. Sources
may now be generated in both kjs/ and runtime/.
Also sort the sources list for readability.
* GNUmakefile.am:
2008-10-28 Alp Toker <alp@nuanti.com>
Reviewed by Cameron Zwarich.
Build fix attempt after kjs -> runtime rename.
* GNUmakefile.am:
2008-10-28 Cameron Zwarich <zwarich@apple.com>
Not reviewed.
Remove a duplicate includes directory.
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
2008-10-28 Cameron Zwarich <zwarich@apple.com>
Not reviewed.
Attempt to fix the Windows build.
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.vcproj/jsc/jsc.vcproj:
2008-10-28 Dan Bernstein <mitz@apple.com>
Reviewed by Mark Rowe.
- export WTF::atomicallyInitializedStaticMutex
* JavaScriptCore.exp:
2008-10-28 Geoffrey Garen <ggaren@apple.com>
Reviewed by Cameron Zwarich.
Fixed CodeBlock dumping to accurately report constant register indices.
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
2008-10-28 Cameron Zwarich <zwarich@apple.com>
Not reviewed.
More Qt build fixes.
* JavaScriptCore.pri:
2008-10-28 Cameron Zwarich <zwarich@apple.com>
Not reviewed.
Fix the Qt build, hopefully for real this time.
* JavaScriptCore.pri:
2008-10-28 Cameron Zwarich <zwarich@apple.com>
Not reviewed.
Fix the Qt build.
* JavaScriptCore.pri:
2008-10-28 Cameron Zwarich <zwarich@apple.com>
Not reviewed.
Fix the Windows build.
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
2008-10-28 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Sam Weinig.
Create a runtime directory in JavaScriptCore and begin moving files to
it. This is the first step towards removing the kjs directory and
placing files in more meaningful subdirectories of JavaScriptCore.
* API/JSBase.cpp:
* API/JSCallbackConstructor.cpp:
* API/JSCallbackConstructor.h:
* API/JSCallbackFunction.cpp:
* API/JSClassRef.cpp:
* API/JSClassRef.h:
* API/JSStringRefCF.cpp:
* API/JSValueRef.cpp:
* API/OpaqueJSString.cpp:
* DerivedSources.make:
* GNUmakefile.am:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* kjs/AllInOneFile.cpp:
* kjs/ArrayConstructor.cpp: Removed.
* kjs/ArrayConstructor.h: Removed.
* kjs/ArrayPrototype.cpp: Removed.
* kjs/ArrayPrototype.h: Removed.
* kjs/BooleanConstructor.cpp: Removed.
* kjs/BooleanConstructor.h: Removed.
* kjs/BooleanObject.cpp: Removed.
* kjs/BooleanObject.h: Removed.
* kjs/BooleanPrototype.cpp: Removed.
* kjs/BooleanPrototype.h: Removed.
* kjs/CallData.cpp: Removed.
* kjs/CallData.h: Removed.
* kjs/ClassInfo.h: Removed.
* kjs/ConstructData.cpp: Removed.
* kjs/ConstructData.h: Removed.
* kjs/DateConstructor.cpp: Removed.
* kjs/DateConstructor.h: Removed.
* kjs/DateInstance.cpp: Removed.
* kjs/DateInstance.h: Removed.
* kjs/DateMath.cpp: Removed.
* kjs/DateMath.h: Removed.
* kjs/DatePrototype.cpp: Removed.
* kjs/DatePrototype.h: Removed.
* kjs/Error.cpp: Removed.
* kjs/Error.h: Removed.
* kjs/ErrorConstructor.cpp: Removed.
* kjs/ErrorConstructor.h: Removed.
* kjs/ErrorInstance.cpp: Removed.
* kjs/ErrorInstance.h: Removed.
* kjs/ErrorPrototype.cpp: Removed.
* kjs/ErrorPrototype.h: Removed.
* kjs/FunctionConstructor.cpp: Removed.
* kjs/FunctionConstructor.h: Removed.
* kjs/FunctionPrototype.cpp: Removed.
* kjs/FunctionPrototype.h: Removed.
* kjs/GlobalEvalFunction.cpp: Removed.
* kjs/GlobalEvalFunction.h: Removed.
* kjs/InternalFunction.cpp: Removed.
* kjs/InternalFunction.h: Removed.
* kjs/JSArray.cpp: Removed.
* kjs/JSArray.h: Removed.
* kjs/JSCell.cpp: Removed.
* kjs/JSCell.h: Removed.
* kjs/JSFunction.cpp: Removed.
* kjs/JSFunction.h: Removed.
* kjs/JSGlobalObject.cpp: Removed.
* kjs/JSGlobalObject.h: Removed.
* kjs/JSGlobalObjectFunctions.cpp: Removed.
* kjs/JSGlobalObjectFunctions.h: Removed.
* kjs/JSImmediate.cpp: Removed.
* kjs/JSImmediate.h: Removed.
* kjs/JSNotAnObject.cpp: Removed.
* kjs/JSNotAnObject.h: Removed.
* kjs/JSNumberCell.cpp: Removed.
* kjs/JSNumberCell.h: Removed.
* kjs/JSObject.cpp: Removed.
* kjs/JSObject.h: Removed.
* kjs/JSString.cpp: Removed.
* kjs/JSString.h: Removed.
* kjs/JSValue.cpp: Removed.
* kjs/JSValue.h: Removed.
* kjs/JSVariableObject.cpp: Removed.
* kjs/JSVariableObject.h: Removed.
* kjs/JSWrapperObject.cpp: Removed.
* kjs/JSWrapperObject.h: Removed.
* kjs/MathObject.cpp: Removed.
* kjs/MathObject.h: Removed.
* kjs/NativeErrorConstructor.cpp: Removed.
* kjs/NativeErrorConstructor.h: Removed.
* kjs/NativeErrorPrototype.cpp: Removed.
* kjs/NativeErrorPrototype.h: Removed.
* kjs/NumberConstructor.cpp: Removed.
* kjs/NumberConstructor.h: Removed.
* kjs/NumberObject.cpp: Removed.
* kjs/NumberObject.h: Removed.
* kjs/NumberPrototype.cpp: Removed.
* kjs/NumberPrototype.h: Removed.
* kjs/ObjectConstructor.cpp: Removed.
* kjs/ObjectConstructor.h: Removed.
* kjs/ObjectPrototype.cpp: Removed.
* kjs/ObjectPrototype.h: Removed.
* kjs/PropertyMap.cpp: Removed.
* kjs/PropertyMap.h: Removed.
* kjs/PropertySlot.cpp: Removed.
* kjs/PropertySlot.h: Removed.
* kjs/PrototypeFunction.cpp: Removed.
* kjs/PrototypeFunction.h: Removed.
* kjs/PutPropertySlot.h: Removed.
* kjs/SmallStrings.cpp: Removed.
* kjs/SmallStrings.h: Removed.
* kjs/StringConstructor.cpp: Removed.
* kjs/StringConstructor.h: Removed.
* kjs/StringObject.cpp: Removed.
* kjs/StringObject.h: Removed.
* kjs/StringObjectThatMasqueradesAsUndefined.h: Removed.
* kjs/StringPrototype.cpp: Removed.
* kjs/StringPrototype.h: Removed.
* kjs/StructureID.cpp: Removed.
* kjs/StructureID.h: Removed.
* kjs/completion.h:
* kjs/interpreter.h:
* runtime: Added.
* runtime/ArrayConstructor.cpp: Copied from kjs/ArrayConstructor.cpp.
* runtime/ArrayConstructor.h: Copied from kjs/ArrayConstructor.h.
* runtime/ArrayPrototype.cpp: Copied from kjs/ArrayPrototype.cpp.
* runtime/ArrayPrototype.h: Copied from kjs/ArrayPrototype.h.
* runtime/BooleanConstructor.cpp: Copied from kjs/BooleanConstructor.cpp.
* runtime/BooleanConstructor.h: Copied from kjs/BooleanConstructor.h.
* runtime/BooleanObject.cpp: Copied from kjs/BooleanObject.cpp.
* runtime/BooleanObject.h: Copied from kjs/BooleanObject.h.
* runtime/BooleanPrototype.cpp: Copied from kjs/BooleanPrototype.cpp.
* runtime/BooleanPrototype.h: Copied from kjs/BooleanPrototype.h.
* runtime/CallData.cpp: Copied from kjs/CallData.cpp.
* runtime/CallData.h: Copied from kjs/CallData.h.
* runtime/ClassInfo.h: Copied from kjs/ClassInfo.h.
* runtime/ConstructData.cpp: Copied from kjs/ConstructData.cpp.
* runtime/ConstructData.h: Copied from kjs/ConstructData.h.
* runtime/DateConstructor.cpp: Copied from kjs/DateConstructor.cpp.
* runtime/DateConstructor.h: Copied from kjs/DateConstructor.h.
* runtime/DateInstance.cpp: Copied from kjs/DateInstance.cpp.
* runtime/DateInstance.h: Copied from kjs/DateInstance.h.
* runtime/DateMath.cpp: Copied from kjs/DateMath.cpp.
* runtime/DateMath.h: Copied from kjs/DateMath.h.
* runtime/DatePrototype.cpp: Copied from kjs/DatePrototype.cpp.
* runtime/DatePrototype.h: Copied from kjs/DatePrototype.h.
* runtime/Error.cpp: Copied from kjs/Error.cpp.
* runtime/Error.h: Copied from kjs/Error.h.
* runtime/ErrorConstructor.cpp: Copied from kjs/ErrorConstructor.cpp.
* runtime/ErrorConstructor.h: Copied from kjs/ErrorConstructor.h.
* runtime/ErrorInstance.cpp: Copied from kjs/ErrorInstance.cpp.
* runtime/ErrorInstance.h: Copied from kjs/ErrorInstance.h.
* runtime/ErrorPrototype.cpp: Copied from kjs/ErrorPrototype.cpp.
* runtime/ErrorPrototype.h: Copied from kjs/ErrorPrototype.h.
* runtime/FunctionConstructor.cpp: Copied from kjs/FunctionConstructor.cpp.
* runtime/FunctionConstructor.h: Copied from kjs/FunctionConstructor.h.
* runtime/FunctionPrototype.cpp: Copied from kjs/FunctionPrototype.cpp.
* runtime/FunctionPrototype.h: Copied from kjs/FunctionPrototype.h.
* runtime/GlobalEvalFunction.cpp: Copied from kjs/GlobalEvalFunction.cpp.
* runtime/GlobalEvalFunction.h: Copied from kjs/GlobalEvalFunction.h.
* runtime/InternalFunction.cpp: Copied from kjs/InternalFunction.cpp.
* runtime/InternalFunction.h: Copied from kjs/InternalFunction.h.
* runtime/JSArray.cpp: Copied from kjs/JSArray.cpp.
* runtime/JSArray.h: Copied from kjs/JSArray.h.
* runtime/JSCell.cpp: Copied from kjs/JSCell.cpp.
* runtime/JSCell.h: Copied from kjs/JSCell.h.
* runtime/JSFunction.cpp: Copied from kjs/JSFunction.cpp.
* runtime/JSFunction.h: Copied from kjs/JSFunction.h.
* runtime/JSGlobalObject.cpp: Copied from kjs/JSGlobalObject.cpp.
* runtime/JSGlobalObject.h: Copied from kjs/JSGlobalObject.h.
* runtime/JSGlobalObjectFunctions.cpp: Copied from kjs/JSGlobalObjectFunctions.cpp.
* runtime/JSGlobalObjectFunctions.h: Copied from kjs/JSGlobalObjectFunctions.h.
* runtime/JSImmediate.cpp: Copied from kjs/JSImmediate.cpp.
* runtime/JSImmediate.h: Copied from kjs/JSImmediate.h.
* runtime/JSNotAnObject.cpp: Copied from kjs/JSNotAnObject.cpp.
* runtime/JSNotAnObject.h: Copied from kjs/JSNotAnObject.h.
* runtime/JSNumberCell.cpp: Copied from kjs/JSNumberCell.cpp.
* runtime/JSNumberCell.h: Copied from kjs/JSNumberCell.h.
* runtime/JSObject.cpp: Copied from kjs/JSObject.cpp.
* runtime/JSObject.h: Copied from kjs/JSObject.h.
* runtime/JSString.cpp: Copied from kjs/JSString.cpp.
* runtime/JSString.h: Copied from kjs/JSString.h.
* runtime/JSValue.cpp: Copied from kjs/JSValue.cpp.
* runtime/JSValue.h: Copied from kjs/JSValue.h.
* runtime/JSVariableObject.cpp: Copied from kjs/JSVariableObject.cpp.
* runtime/JSVariableObject.h: Copied from kjs/JSVariableObject.h.
* runtime/JSWrapperObject.cpp: Copied from kjs/JSWrapperObject.cpp.
* runtime/JSWrapperObject.h: Copied from kjs/JSWrapperObject.h.
* runtime/MathObject.cpp: Copied from kjs/MathObject.cpp.
* runtime/MathObject.h: Copied from kjs/MathObject.h.
* runtime/NativeErrorConstructor.cpp: Copied from kjs/NativeErrorConstructor.cpp.
* runtime/NativeErrorConstructor.h: Copied from kjs/NativeErrorConstructor.h.
* runtime/NativeErrorPrototype.cpp: Copied from kjs/NativeErrorPrototype.cpp.
* runtime/NativeErrorPrototype.h: Copied from kjs/NativeErrorPrototype.h.
* runtime/NumberConstructor.cpp: Copied from kjs/NumberConstructor.cpp.
* runtime/NumberConstructor.h: Copied from kjs/NumberConstructor.h.
* runtime/NumberObject.cpp: Copied from kjs/NumberObject.cpp.
* runtime/NumberObject.h: Copied from kjs/NumberObject.h.
* runtime/NumberPrototype.cpp: Copied from kjs/NumberPrototype.cpp.
* runtime/NumberPrototype.h: Copied from kjs/NumberPrototype.h.
* runtime/ObjectConstructor.cpp: Copied from kjs/ObjectConstructor.cpp.
* runtime/ObjectConstructor.h: Copied from kjs/ObjectConstructor.h.
* runtime/ObjectPrototype.cpp: Copied from kjs/ObjectPrototype.cpp.
* runtime/ObjectPrototype.h: Copied from kjs/ObjectPrototype.h.
* runtime/PropertyMap.cpp: Copied from kjs/PropertyMap.cpp.
* runtime/PropertyMap.h: Copied from kjs/PropertyMap.h.
* runtime/PropertySlot.cpp: Copied from kjs/PropertySlot.cpp.
* runtime/PropertySlot.h: Copied from kjs/PropertySlot.h.
* runtime/PrototypeFunction.cpp: Copied from kjs/PrototypeFunction.cpp.
* runtime/PrototypeFunction.h: Copied from kjs/PrototypeFunction.h.
* runtime/PutPropertySlot.h: Copied from kjs/PutPropertySlot.h.
* runtime/SmallStrings.cpp: Copied from kjs/SmallStrings.cpp.
* runtime/SmallStrings.h: Copied from kjs/SmallStrings.h.
* runtime/StringConstructor.cpp: Copied from kjs/StringConstructor.cpp.
* runtime/StringConstructor.h: Copied from kjs/StringConstructor.h.
* runtime/StringObject.cpp: Copied from kjs/StringObject.cpp.
* runtime/StringObject.h: Copied from kjs/StringObject.h.
* runtime/StringObjectThatMasqueradesAsUndefined.h: Copied from kjs/StringObjectThatMasqueradesAsUndefined.h.
* runtime/StringPrototype.cpp: Copied from kjs/StringPrototype.cpp.
* runtime/StringPrototype.h: Copied from kjs/StringPrototype.h.
* runtime/StructureID.cpp: Copied from kjs/StructureID.cpp.
* runtime/StructureID.h: Copied from kjs/StructureID.h.
2008-10-28 Geoffrey Garen <ggaren@apple.com>
Reviewed by Sam Weinig.
Fixed https://bugs.webkit.org/show_bug.cgi?id=21919
Sampler reports bogus time in op_enter during 3d-raytrace.js
Fixed a bug where we would pass the incorrect Instruction* during some
parts of CTI codegen.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
* VM/SamplingTool.cpp:
(JSC::SamplingTool::run):
* wtf/Platform.h:
2008-10-28 Kevin McCullough <kmccullough@apple.com>
Reviewed by Dan Bernstein.
-Removed unused includes.
Apparent .4% speedup in Sunspider
* kjs/JSObject.cpp:
* kjs/interpreter.cpp:
2008-10-28 Alp Toker <alp@nuanti.com>
Include copyright license files in the autotools dist target.
Change suggested by Mike Hommey.
* GNUmakefile.am:
2008-10-27 Geoffrey Garen <ggaren@apple.com>
Reviewed by Maciej Stachowiak.
Stop discarding CodeBlock samples that can't be charged to a specific
opcode. Instead, charge the relevant CodeBlock, and provide a footnote
explaining the situation.
This will help us tell which CodeBlocks are hot, even if we can't
identify specific lines of code within the CodeBlocks.
* VM/SamplingTool.cpp:
(JSC::ScopeSampleRecord::sample):
(JSC::compareScopeSampleRecords):
(JSC::SamplingTool::dump):
* VM/SamplingTool.h:
(JSC::ScopeSampleRecord::ScopeSampleRecord):
(JSC::ScopeSampleRecord::~ScopeSampleRecord):
2008-10-27 Geoffrey Garen <ggaren@apple.com>
Reviewed by Sam Weinig.
Added a mutex around the SamplingTool's ScopeNode* map, to solve a crash
when sampling the v8 tests.
* VM/SamplingTool.cpp:
(JSC::SamplingTool::run):
(JSC::SamplingTool::notifyOfScope):
* VM/SamplingTool.h: Since new ScopeNodes can be created after
the SamplingTools has begun sampling, reads and writes to / from the
map need to be synchronized. Shark says this doesn't measurably increase
sampling overhead.
2008-10-25 Geoffrey Garen <ggaren@apple.com>
Not reviewed.
Try to fix Windows build.
* VM/Machine.cpp:
(JSC::Machine::privateExecute): Provide a dummy value to the
HostCallRecord in CTI non-sampling builds, to silence compiler warning.
2008-10-25 Geoffrey Garen <ggaren@apple.com>
Not reviewed.
Try to fix Windows build.
* VM/SamplingTool.h:
(JSC::SamplingTool::encodeSample): Explicitly cast bool to int, to
silence compiler warning.
2008-10-25 Geoffrey Garen <ggaren@apple.com>
Reviewed by Sam Weinig, with Gavin Barraclough's help.
Fixed Sampling Tool:
- Made CodeBlock sampling work with CTI
- Improved accuracy by unifying most sampling data into a single
32bit word, which can be written / read atomically.
- Split out three different #ifdefs for modularity: OPCODE_SAMPLING;
CODEBLOCK_SAMPLING; OPCODE_STATS.
- Improved reporting clarity
- Refactored for code clarity
* JavaScriptCore.exp: Exported another symbol.
* VM/CTI.cpp:
(JSC::CTI::emitCTICall):
(JSC::CTI::compileOpCall):
(JSC::CTI::emitSlowScriptCheck):
(JSC::CTI::compileBinaryArithOpSlowCase):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
(JSC::CTI::privateCompile):
* VM/CTI.h: Updated CTI codegen to use the unified SamplingTool interface
for encoding samples. (This required passing the current vPC to a lot
more functions, since the unified interface samples the current vPC.)
Added hooks for writing the current CodeBlock* on function entry and
after a function call, for the sake of the CodeBlock sampler. Removed
obsolete hook for clearing the current sample inside op_end. Also removed
the custom enum used to differentiate flavors of op_call, since the
OpcodeID enum works just as well. (This was important in an earlier
version of the patch, but now it's just cleanup.)
* VM/CodeBlock.cpp:
(JSC::CodeBlock::lineNumberForVPC):
* VM/CodeBlock.h: Upated for refactored #ifdefs. Changed lineNumberForVPC
to be robust against vPCs not recorded for exception handling, since
the Sampler may ask for an arbitrary vPC.
* VM/Machine.cpp:
(JSC::Machine::execute):
(JSC::Machine::privateExecute):
(JSC::Machine::cti_op_call_NotJSFunction):
(JSC::Machine::cti_op_construct_NotJSConstruct):
* VM/Machine.h:
(JSC::Machine::setSampler):
(JSC::Machine::sampler):
(JSC::Machine::jitCodeBuffer): Upated for refactored #ifdefs. Changed
Machine to use SamplingTool helper objects to record movement in and
out of host code. This makes samples a bit more precise.
* VM/Opcode.cpp:
(JSC::OpcodeStats::~OpcodeStats):
* VM/Opcode.h: Upated for refactored #ifdefs. Added a little more padding,
to accomodate our more verbose opcode names.
* VM/SamplingTool.cpp:
(JSC::ScopeSampleRecord::sample): Only count a sample toward our total
if we actually record it. This solves cases where a CodeBlock will
claim to have been sampled many times, with reported samples that don't
match.
(JSC::SamplingTool::run): Read the current sample into a Sample helper
object, to ensure that the data doesn't change while we're analyzing it,
and to help decode the data. Only access the CodeBlock sampling hash
table if CodeBlock sampling has been enabled, so non-CodeBlock sampling
runs can operate with even less overhead.
(JSC::SamplingTool::dump): I reorganized this code a lot to print the
most important info at the top, print as a table, annotate and document
the stuff I didn't understand when I started, etc.
* VM/SamplingTool.h: New helper classes, described above.
* kjs/Parser.h:
* kjs/Shell.cpp:
(runWithScripts):
* kjs/nodes.cpp:
(JSC::ScopeNode::ScopeNode): Updated for new sampling APIs.
* wtf/Platform.h: Moved sampling #defines here, since our custom is to
put ENABLE #defines into Platform.h. Made explicit the fact that
CODEBLOCK_SAMPLING depends on OPCODE_SAMPLING.
2008-10-25 Jan Michael Alonzo <jmalonzo@webkit.org>
JSC Build fix, not reviewed.
* VM/CTI.cpp: add missing include stdio.h for debug builds
2008-10-24 Eric Seidel <eric@webkit.org>
Reviewed by Darin Adler.
Get rid of a bonus ASSERT when using a null string as a regexp.
Specifically calling: RegularExpression::match() with String::empty()
will hit this ASSERT.
Chromium hits this, but I don't know of any way to make a layout test.
* pcre/pcre_exec.cpp:
(jsRegExpExecute):
2008-10-24 Alexey Proskuryakov <ap@webkit.org>
Suggested and rubber-stamped by Geoff Garen.
Fix a crash when opening Font Picker.
The change also hopefully fixes this bug, which I could never reproduce:
https://bugs.webkit.org/show_bug.cgi?id=20241
<rdar://problem/6290576> Safari crashes at JSValueUnprotect() when fontpicker view close
* API/JSContextRef.cpp: (JSContextGetGlobalObject): Use lexical global object instead of
dynamic one.
2008-10-24 Cameron Zwarich <zwarich@apple.com>
Reviewed by Geoff Garen.
Remove ScopeChainNode::bottom() and inline it into its only caller,
ScopeChainnode::globalObject().
* kjs/JSGlobalObject.h:
(JSC::ScopeChainNode::globalObject):
* kjs/ScopeChain.h:
(JSC::ScopeChain::bottom):
2008-10-24 Cameron Zwarich <zwarich@apple.com>
Reviewed by Maciej Stachowiak.
Bug 21862: Create JSFunction prototype property lazily
<https://bugs.webkit.org/show_bug.cgi?id=21862>
This is a 1.5% speedup on SunSpider and a 1.4% speedup on the V8
benchmark suite, including a 3.8% speedup on Earley-Boyer.
* kjs/JSFunction.cpp:
(JSC::JSFunction::getOwnPropertySlot):
* kjs/nodes.cpp:
(JSC::FuncDeclNode::makeFunction):
(JSC::FuncExprNode::makeFunction):
2008-10-24 Greg Bolsinga <bolsinga@apple.com>
Reviewed by Sam Weinig.
https://bugs.webkit.org/show_bug.cgi?id=21475
Provide support for the Geolocation API
http://dev.w3.org/geo/api/spec-source.html
* wtf/Platform.h: ENABLE_GEOLOCATION defaults to 0
2008-10-24 Darin Adler <darin@apple.com>
- finish rolling out https://bugs.webkit.org/show_bug.cgi?id=21732
* API/APICast.h:
* API/JSCallbackConstructor.h:
* API/JSCallbackFunction.cpp:
* API/JSCallbackFunction.h:
* API/JSCallbackObject.h:
* API/JSCallbackObjectFunctions.h:
* API/JSContextRef.cpp:
* API/JSObjectRef.cpp:
* API/JSValueRef.cpp:
* VM/CTI.cpp:
* VM/CTI.h:
* VM/CodeBlock.cpp:
* VM/CodeBlock.h:
* VM/CodeGenerator.cpp:
* VM/CodeGenerator.h:
* VM/ExceptionHelpers.cpp:
* VM/ExceptionHelpers.h:
* VM/JSPropertyNameIterator.cpp:
* VM/JSPropertyNameIterator.h:
* VM/Machine.cpp:
* VM/Machine.h:
* VM/Register.h:
* kjs/ArgList.cpp:
* kjs/ArgList.h:
* kjs/Arguments.cpp:
* kjs/Arguments.h:
* kjs/ArrayConstructor.cpp:
* kjs/ArrayPrototype.cpp:
* kjs/BooleanConstructor.cpp:
* kjs/BooleanConstructor.h:
* kjs/BooleanObject.h:
* kjs/BooleanPrototype.cpp:
* kjs/CallData.cpp:
* kjs/CallData.h:
* kjs/ConstructData.cpp:
* kjs/ConstructData.h:
* kjs/DateConstructor.cpp:
* kjs/DateInstance.h:
* kjs/DatePrototype.cpp:
* kjs/DatePrototype.h:
* kjs/DebuggerCallFrame.cpp:
* kjs/DebuggerCallFrame.h:
* kjs/ErrorConstructor.cpp:
* kjs/ErrorPrototype.cpp:
* kjs/ExecState.cpp:
* kjs/ExecState.h:
* kjs/FunctionConstructor.cpp:
* kjs/FunctionPrototype.cpp:
* kjs/FunctionPrototype.h:
* kjs/GetterSetter.cpp:
* kjs/GetterSetter.h:
* kjs/InternalFunction.h:
* kjs/JSActivation.cpp:
* kjs/JSActivation.h:
* kjs/JSArray.cpp:
* kjs/JSArray.h:
* kjs/JSCell.cpp:
* kjs/JSCell.h:
* kjs/JSFunction.cpp:
* kjs/JSFunction.h:
* kjs/JSGlobalData.h:
* kjs/JSGlobalObject.cpp:
* kjs/JSGlobalObject.h:
* kjs/JSGlobalObjectFunctions.cpp:
* kjs/JSGlobalObjectFunctions.h:
* kjs/JSImmediate.cpp:
* kjs/JSImmediate.h:
* kjs/JSNotAnObject.cpp:
* kjs/JSNotAnObject.h:
* kjs/JSNumberCell.cpp:
* kjs/JSNumberCell.h:
* kjs/JSObject.cpp:
* kjs/JSObject.h:
* kjs/JSStaticScopeObject.cpp:
* kjs/JSStaticScopeObject.h:
* kjs/JSString.cpp:
* kjs/JSString.h:
* kjs/JSValue.h:
* kjs/JSVariableObject.h:
* kjs/JSWrapperObject.h:
* kjs/MathObject.cpp:
* kjs/MathObject.h:
* kjs/NativeErrorConstructor.cpp:
* kjs/NumberConstructor.cpp:
* kjs/NumberConstructor.h:
* kjs/NumberObject.cpp:
* kjs/NumberObject.h:
* kjs/NumberPrototype.cpp:
* kjs/ObjectConstructor.cpp:
* kjs/ObjectPrototype.cpp:
* kjs/ObjectPrototype.h:
* kjs/PropertyMap.h:
* kjs/PropertySlot.cpp:
* kjs/PropertySlot.h:
* kjs/RegExpConstructor.cpp:
* kjs/RegExpConstructor.h:
* kjs/RegExpMatchesArray.h:
* kjs/RegExpObject.cpp:
* kjs/RegExpObject.h:
* kjs/RegExpPrototype.cpp:
* kjs/Shell.cpp:
* kjs/StringConstructor.cpp:
* kjs/StringObject.cpp:
* kjs/StringObject.h:
* kjs/StringObjectThatMasqueradesAsUndefined.h:
* kjs/StringPrototype.cpp:
* kjs/StructureID.cpp:
* kjs/StructureID.h:
* kjs/collector.cpp:
* kjs/collector.h:
* kjs/completion.h:
* kjs/grammar.y:
* kjs/interpreter.cpp:
* kjs/interpreter.h:
* kjs/lookup.cpp:
* kjs/lookup.h:
* kjs/nodes.h:
* kjs/operations.cpp:
* kjs/operations.h:
* kjs/protect.h:
* profiler/ProfileGenerator.cpp:
* profiler/Profiler.cpp:
* profiler/Profiler.h:
Use JSValue* instead of JSValuePtr.
2008-10-24 David Kilzer <ddkilzer@apple.com>
Rolled out r37840.
* wtf/Platform.h:
2008-10-23 Greg Bolsinga <bolsinga@apple.com>
Reviewed by Sam Weinig.
https://bugs.webkit.org/show_bug.cgi?id=21475
Provide support for the Geolocation API
http://dev.w3.org/geo/api/spec-source.html
* wtf/Platform.h: ENABLE_GEOLOCATION defaults to 0
2008-10-23 David Kilzer <ddkilzer@apple.com>
Bug 21832: Fix scripts using 'new File::Temp' for Perl 5.10
<https://bugs.webkit.org/show_bug.cgi?id=21832>
Reviewed by Sam Weinig.
* pcre/dftables: Use imported tempfile() from File::Temp instead of
'new File::Temp' to make the script work with Perl 5.10.
2008-10-23 Gavin Barraclough <barraclough@apple.com>
Reviewed by Oliver Hunt.
Fix hideous pathological case performance when looking up repatch info, bug #21727.
When repatching JIT code to optimize we look up records providing information about
the generated code (also used to track recsources used in linking to be later released).
The lookup was being performed using a linear scan of all such records.
(1) Split up the different types of reptach information. This means we can search them
separately, and in some cases should reduce their size.
(2) In the case of property accesses, search with a binary chop over the data.
(3) In the case of calls, pass a pointer to the repatch info into the relink function.
* VM/CTI.cpp:
(JSC::CTI::CTI):
(JSC::CTI::compileOpCall):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
(JSC::CTI::privateCompile):
(JSC::CTI::unlinkCall):
(JSC::CTI::linkCall):
* VM/CTI.h:
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
(JSC::CodeBlock::~CodeBlock):
(JSC::CodeBlock::unlinkCallers):
(JSC::CodeBlock::derefStructureIDs):
* VM/CodeBlock.h:
(JSC::StructureStubInfo::StructureStubInfo):
(JSC::CallLinkInfo::CallLinkInfo):
(JSC::CallLinkInfo::setUnlinked):
(JSC::CallLinkInfo::isLinked):
(JSC::getStructureStubInfoReturnLocation):
(JSC::binaryChop):
(JSC::CodeBlock::addCaller):
(JSC::CodeBlock::getStubInfo):
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitResolve):
(JSC::CodeGenerator::emitGetById):
(JSC::CodeGenerator::emitPutById):
(JSC::CodeGenerator::emitCall):
(JSC::CodeGenerator::emitConstruct):
* VM/Machine.cpp:
(JSC::Machine::cti_vm_lazyLinkCall):
2008-10-23 Peter Kasting <pkasting@google.com>
Reviewed by Adam Roben.
https://bugs.webkit.org/show_bug.cgi?id=21833
Place JavaScript Debugger hooks under #if ENABLE(JAVASCRIPT_DEBUGGER).
* wtf/Platform.h:
2008-10-23 David Kilzer <ddkilzer@apple.com>
Bug 21831: Fix create_hash_table for Perl 5.10
<https://bugs.webkit.org/show_bug.cgi?id=21831>
Reviewed by Sam Weinig.
* kjs/create_hash_table: Escaped square brackets so that Perl 5.10
doesn't try to use @nameEntries.
2008-10-23 Darin Adler <darin@apple.com>
- roll out https://bugs.webkit.org/show_bug.cgi?id=21732
to remove the JSValuePtr class, to fix two problems
1) slowness under MSVC, since it doesn't handle a
class with a single pointer in it as efficiently
as a pointer
2) uninitialized pointers in Vector
* JavaScriptCore.exp: Updated.
* API/APICast.h:
(toRef):
* VM/CTI.cpp:
(JSC::CTI::asInteger):
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::addConstant):
* VM/CodeGenerator.h:
(JSC::CodeGenerator::JSValueHashTraits::constructDeletedValue):
(JSC::CodeGenerator::JSValueHashTraits::isDeletedValue):
* VM/Machine.cpp:
(JSC::Machine::cti_op_add):
(JSC::Machine::cti_op_pre_inc):
(JSC::Machine::cti_op_get_by_id):
(JSC::Machine::cti_op_get_by_id_second):
(JSC::Machine::cti_op_get_by_id_generic):
(JSC::Machine::cti_op_get_by_id_fail):
(JSC::Machine::cti_op_instanceof):
(JSC::Machine::cti_op_del_by_id):
(JSC::Machine::cti_op_mul):
(JSC::Machine::cti_op_call_NotJSFunction):
(JSC::Machine::cti_op_resolve):
(JSC::Machine::cti_op_construct_NotJSConstruct):
(JSC::Machine::cti_op_get_by_val):
(JSC::Machine::cti_op_sub):
(JSC::Machine::cti_op_lesseq):
(JSC::Machine::cti_op_negate):
(JSC::Machine::cti_op_resolve_base):
(JSC::Machine::cti_op_resolve_skip):
(JSC::Machine::cti_op_resolve_global):
(JSC::Machine::cti_op_div):
(JSC::Machine::cti_op_pre_dec):
(JSC::Machine::cti_op_not):
(JSC::Machine::cti_op_eq):
(JSC::Machine::cti_op_lshift):
(JSC::Machine::cti_op_bitand):
(JSC::Machine::cti_op_rshift):
(JSC::Machine::cti_op_bitnot):
(JSC::Machine::cti_op_mod):
(JSC::Machine::cti_op_less):
(JSC::Machine::cti_op_neq):
(JSC::Machine::cti_op_urshift):
(JSC::Machine::cti_op_bitxor):
(JSC::Machine::cti_op_bitor):
(JSC::Machine::cti_op_call_eval):
(JSC::Machine::cti_op_throw):
(JSC::Machine::cti_op_next_pname):
(JSC::Machine::cti_op_typeof):
(JSC::Machine::cti_op_is_undefined):
(JSC::Machine::cti_op_is_boolean):
(JSC::Machine::cti_op_is_number):
(JSC::Machine::cti_op_is_string):
(JSC::Machine::cti_op_is_object):
(JSC::Machine::cti_op_is_function):
(JSC::Machine::cti_op_stricteq):
(JSC::Machine::cti_op_nstricteq):
(JSC::Machine::cti_op_to_jsnumber):
(JSC::Machine::cti_op_in):
(JSC::Machine::cti_op_del_by_val):
(JSC::Machine::cti_vm_throw):
Removed calls to payload functions.
* VM/Register.h:
(JSC::Register::Register): Removed overload for JSCell and call
to payload function.
* kjs/JSCell.h: Changed JSCell to derive from JSValue again.
Removed JSValuePtr constructor.
(JSC::asCell): Changed cast from reinterpret_cast to static_cast.
* kjs/JSImmediate.h: Removed JSValuePtr class. Added typedef back.
* kjs/JSValue.h:
(JSC::JSValue::JSValue): Added empty protected inline constructor back.
(JSC::JSValue::~JSValue): Same for destructor.
Removed == and != operator for JSValuePtr.
* kjs/PropertySlot.h:
(JSC::PropertySlot::PropertySlot): Chnaged argument to const JSValue*
and added a const_cast.
* kjs/protect.h: Removed overloads and specialization for JSValuePtr.
2008-10-22 Oliver Hunt <oliver@apple.com>
Reviewed by Maciej Stachowiak.
Really "fix" CTI mode on windows 2k3.
This adds new methods fastMallocExecutable and fastFreeExecutable
to wrap allocation for cti code. This still just makes fastMalloc
return executable memory all the time, which will be fixed in a
later patch.
However in windows debug builds all executable allocations will be
allocated on separate executable pages, which should resolve any
remaining 2k3 issues. Conveniently the 2k3 bot will now also fail
if there are any fastFree vs. fastFreeExecutable errors.
* ChangeLog:
* VM/CodeBlock.cpp:
(JSC::CodeBlock::~CodeBlock):
* kjs/regexp.cpp:
(JSC::RegExp::~RegExp):
* masm/X86Assembler.h:
(JSC::JITCodeBuffer::copy):
* wtf/FastMalloc.cpp:
(WTF::fastMallocExecutable):
(WTF::fastFreeExecutable):
(WTF::TCMallocStats::fastMallocExecutable):
(WTF::TCMallocStats::fastFreeExecutable):
* wtf/FastMalloc.h:
2008-10-22 Darin Adler <darin@apple.com>
Reviewed by Sam Weinig.
- fix https://bugs.webkit.org/show_bug.cgi?id=21294
Bug 21294: Devirtualize getOwnPropertySlot()
A bit over 3% faster on V8 tests.
* JavascriptCore.exp: Export leak-related functions..
* API/JSCallbackConstructor.h:
(JSC::JSCallbackConstructor::createStructureID): Set HasStandardGetOwnPropertySlot
since this class doesn't override getPropertySlot.
* API/JSCallbackFunction.h:
(JSC::JSCallbackFunction::createStructureID): Ditto.
* VM/ExceptionHelpers.cpp:
(JSC::InterruptedExecutionError::InterruptedExecutionError): Use a structure
that's created just for this class instead of trying to share a single "null
prototype" structure.
* VM/Machine.cpp:
(JSC::Machine::cti_op_create_arguments_no_params): Rename
Arguments::ArgumentsNoParameters to Arguments::NoParameters.
* kjs/Arguments.h: Rename the enum from Arguments::ArgumentsParameters to
Arguments::NoParametersType and the value from Arguments::ArgumentsNoParameters
to Arguments::NoParameters.
(JSC::Arguments::createStructureID): Added. Returns a structure without
HasStandardGetOwnPropertySlot since this class overrides getOwnPropertySlot.
(JSC::Arguments::Arguments): Added an assertion that there are no parameters.
* kjs/DatePrototype.h:
(JSC::DatePrototype::createStructureID): Added. Returns a structure without
HasStandardGetOwnPropertySlot since this class overrides getOwnPropertySlot.
* kjs/FunctionPrototype.h:
(JSC::FunctionPrototype::createStructureID): Set HasStandardGetOwnPropertySlot
since this class doesn't override getPropertySlot.
* kjs/InternalFunction.h:
(JSC::InternalFunction::createStructureID): Ditto.
* kjs/JSArray.h:
(JSC::JSArray::createStructureID): Added. Returns a structure without
HasStandardGetOwnPropertySlot since this class overrides getOwnPropertySlot.
* kjs/JSCell.h: Added declaration of fastGetOwnPropertySlot; a non-virtual
version that uses the structure bit to decide whether to call the virtual
version.
* kjs/JSFunction.h:
(JSC::JSFunction::createStructureID): Added. Returns a structure without
HasStandardGetOwnPropertySlot since this class overrides getOwnPropertySlot.
* kjs/JSGlobalData.cpp:
(JSC::JSGlobalData::JSGlobalData): Initialize new structures; removed
nullProtoStructureID.
* kjs/JSGlobalData.h: Added new structures. Removed nullProtoStructureID.
* kjs/JSGlobalObject.h:
(JSC::JSGlobalObject::createStructureID): Added. Returns a structure without
HasStandardGetOwnPropertySlot since this class overrides getOwnPropertySlot.
* kjs/JSNotAnObject.h:
(JSC::JSNotAnObjectErrorStub::JSNotAnObjectErrorStub): Use a structure
that's created just for this class instead of trying to share a single "null
prototype" structure.
(JSC::JSNotAnObjectErrorStub::isNotAnObjectErrorStub): Marked this function
virtual for clarity and made it private since no one should call it if they
already have a pointer to this specific type.
(JSC::JSNotAnObject::JSNotAnObject): Use a structure that's created just
for this class instead of trying to share a single "null prototype" structure.
(JSC::JSNotAnObject::createStructureID): Added. Returns a structure without
HasStandardGetOwnPropertySlot since this class overrides getOwnPropertySlot.
* kjs/JSObject.h:
(JSC::JSObject::createStructureID): Added HasStandardGetOwnPropertySlot.
(JSC::JSObject::inlineGetOwnPropertySlot): Added. Used so we can share code
between getOwnPropertySlot and fastGetOwnPropertySlot.
(JSC::JSObject::getOwnPropertySlot): Moved so that functions are above the
functions that call them. Moved the guts of this function into
inlineGetOwnPropertySlot.
(JSC::JSCell::fastGetOwnPropertySlot): Added. Checks the
HasStandardGetOwnPropertySlot bit and if it's set, calls
inlineGetOwnPropertySlot, otherwise calls getOwnPropertySlot.
(JSC::JSObject::getPropertySlot): Changed to call fastGetOwnPropertySlot.
(JSC::JSValue::get): Changed to call fastGetOwnPropertySlot.
* kjs/JSWrapperObject.h: Made constructor protected to emphasize that
this class is only a base class and never instantiated.
* kjs/MathObject.h:
(JSC::MathObject::createStructureID): Added. Returns a structure without
HasStandardGetOwnPropertySlot since this class overrides getOwnPropertySlot.
* kjs/NumberConstructor.h:
(JSC::NumberConstructor::createStructureID): Ditto.
* kjs/RegExpConstructor.h:
(JSC::RegExpConstructor::createStructureID): Ditto.
* kjs/RegExpObject.h:
(JSC::RegExpObject::createStructureID): Ditto.
* kjs/StringObject.h:
(JSC::StringObject::createStructureID): Ditto.
* kjs/TypeInfo.h: Added HasStandardGetOwnPropertySlot flag and
hasStandardGetOwnPropertySlot accessor function.
2008-10-22 Cameron Zwarich <zwarich@apple.com>
Reviewed by Geoff Garen.
Bug 21803: Fuse op_jfalse with op_eq_null and op_neq_null
<https://bugs.webkit.org/show_bug.cgi?id=21803>
Fuse op_jfalse with op_eq_null and op_neq_null to make the new opcodes
op_jeq_null and op_jneq_null.
This is a 2.6% speedup on the V8 Raytrace benchmark, and strangely also
a 4.7% speedup on the V8 Arguments benchmark, even though it uses
neither of the two new opcodes.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitJumpIfTrue):
(JSC::CodeGenerator::emitJumpIfFalse):
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
* VM/Opcode.h:
2008-10-22 Darin Fisher <darin@chromium.org>
Reviewed by Eric Seidel.
Should not define PLATFORM(WIN,MAC,GTK) when PLATFORM(CHROMIUM) is defined
https://bugs.webkit.org/show_bug.cgi?id=21757
PLATFORM(CHROMIUM) implies HAVE_ACCESSIBILITY
* wtf/Platform.h:
2008-10-22 Cameron Zwarich <zwarich@apple.com>
Reviewed by Alexey Proskuryakov.
Correct opcode names in documentation.
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
2008-10-21 Oliver Hunt <oliver@apple.com>
RS=Maciej Stachowiak.
Force FastMalloc to make all allocated pages executable in
a vague hope this will allow the Win2k3 bot to be able to
run tests.
Filed Bug 21783: Need more granular control over allocation of executable memory
to cover a more granular version of this patch.
* wtf/TCSystemAlloc.cpp:
(TryVirtualAlloc):
2008-10-21 Alexey Proskuryakov <ap@webkit.org>
Reviewed by Darin Adler.
https://bugs.webkit.org/show_bug.cgi?id=21769
MessagePort should be GC protected if there are messages to be delivered
* wtf/MessageQueue.h:
(WTF::::isEmpty): Added. Also added a warning for methods that return a snapshot of queue
state, thus likely to cause race conditions.
2008-10-21 Darin Adler <darin@apple.com>
Reviewed by Maciej Stachowiak.
- convert post-increment to pre-increment in a couple more places for speed
Speeds up V8 benchmarks a little on most computers. (But, strangely, slows
them down a little on my computer.)
* kjs/nodes.cpp:
(JSC::statementListEmitCode): Removed default argument, since we always want
to specify this explicitly.
(JSC::ForNode::emitCode): Tolerate ignoredResult() as the dst -- means the
same thing as 0.
(JSC::ReturnNode::emitCode): Ditto.
(JSC::ThrowNode::emitCode): Ditto.
(JSC::FunctionBodyNode::emitCode): Pass ignoredResult() so that we know we
don't have to compute the result of function statements.
2008-10-21 Peter Kasting <pkasting@google.com>
Reviewed by Maciej Stachowiak.
Fix an include of a non-public header to use "" instead of <>.
* API/JSProfilerPrivate.cpp:
2008-10-20 Sam Weinig <sam@webkit.org>
Reviewed by Cameron Zwarich.
Fix for https://bugs.webkit.org/show_bug.cgi?id=21766
REGRESSION: 12 JSC tests fail
The JSGlobalObject was mutating the shared nullProtoStructureID when
used in jsc. Instead of using nullProtoStructureID, use a new StructureID.
* API/JSCallbackObject.h:
* API/JSCallbackObjectFunctions.h:
(JSC::::JSCallbackObject):
* API/JSContextRef.cpp:
(JSGlobalContextCreateInGroup):
* kjs/JSGlobalObject.h:
(JSC::JSGlobalObject::JSGlobalObject):
* kjs/Shell.cpp:
(GlobalObject::GlobalObject):
(jscmain):
2008-10-20 Cameron Zwarich <zwarich@apple.com>
Reviewed by Maciej Stachowiak.
Remove an untaken branch in CodeGenerator::emitJumpIfFalse(). This
function is never called with a backwards target LabelID, and there is
even an assertion to this effect at the top of the function body.
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitJumpIfFalse):
2008-10-20 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Sam Weinig.
Add opcode documentation for undocumented opcodes.
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
2008-10-16 Sam Weinig <sam@webkit.org>
Reviewed by Cameron Zwarich.
Fix for https://bugs.webkit.org/show_bug.cgi?id=21683
Don't create intermediate StructureIDs for builtin objects
Second stage in reduce number of StructureIDs created when initializing the
JSGlobalObject.
- Use putDirectWithoutTransition for the remaining singleton objects to reduce
the number of StructureIDs create for about:blank from 132 to 73.
* kjs/ArrayConstructor.cpp:
(JSC::ArrayConstructor::ArrayConstructor):
* kjs/BooleanConstructor.cpp:
(JSC::BooleanConstructor::BooleanConstructor):
* kjs/BooleanPrototype.cpp:
(JSC::BooleanPrototype::BooleanPrototype):
* kjs/DateConstructor.cpp:
(JSC::DateConstructor::DateConstructor):
* kjs/ErrorConstructor.cpp:
(JSC::ErrorConstructor::ErrorConstructor):
* kjs/ErrorPrototype.cpp:
(JSC::ErrorPrototype::ErrorPrototype):
* kjs/FunctionConstructor.cpp:
(JSC::FunctionConstructor::FunctionConstructor):
* kjs/FunctionPrototype.cpp:
(JSC::FunctionPrototype::FunctionPrototype):
(JSC::FunctionPrototype::addFunctionProperties):
* kjs/FunctionPrototype.h:
(JSC::FunctionPrototype::createStructureID):
* kjs/InternalFunction.cpp:
* kjs/InternalFunction.h:
(JSC::InternalFunction::InternalFunction):
* kjs/JSGlobalObject.cpp:
(JSC::JSGlobalObject::reset):
* kjs/JSObject.h:
* kjs/MathObject.cpp:
(JSC::MathObject::MathObject):
* kjs/NumberConstructor.cpp:
(JSC::NumberConstructor::NumberConstructor):
* kjs/NumberPrototype.cpp:
(JSC::NumberPrototype::NumberPrototype):
* kjs/ObjectConstructor.cpp:
(JSC::ObjectConstructor::ObjectConstructor):
* kjs/RegExpConstructor.cpp:
(JSC::RegExpConstructor::RegExpConstructor):
* kjs/RegExpPrototype.cpp:
(JSC::RegExpPrototype::RegExpPrototype):
* kjs/StringConstructor.cpp:
(JSC::StringConstructor::StringConstructor):
* kjs/StringPrototype.cpp:
(JSC::StringPrototype::StringPrototype):
* kjs/StructureID.cpp:
(JSC::StructureID::dumpStatistics):
* kjs/StructureID.h:
(JSC::StructureID::setPrototypeWithoutTransition):
2008-10-20 Alp Toker <alp@nuanti.com>
Fix autotools dist build target by listing recently added header
files only. Not reviewed.
* GNUmakefile.am:
2008-10-20 Geoffrey Garen <ggaren@apple.com>
Reviewed by Anders Carlsson.
* VM/Machine.cpp:
(JSC::Machine::tryCacheGetByID): Removed a redundant and sometimes
incorrect cast, which started ASSERTing after Darin's last checkin.
2008-10-20 Geoffrey Garen <ggaren@apple.com>
Not reviewed.
Re-enable CTI, which I accidentally disabled while checking in fixes
to bytecode.
* wtf/Platform.h:
2008-10-20 Alp Toker <alp@nuanti.com>
Rubber-stamped by Mark Rowe.
Typo fix in function name: mimimum -> minimum.
* kjs/DateMath.cpp:
(JSC::minimumYearForDST):
(JSC::equivalentYearForDST):
2008-10-20 Alp Toker <alp@nuanti.com>
Reviewed by Mark Rowe.
Use pthread instead of GThread where possible in the GTK+ port. This
fixes issues with global initialisation, particularly on GTK+/Win32
where a late g_thread_init() will cause hangs.
* GNUmakefile.am:
* wtf/Platform.h:
* wtf/Threading.h:
* wtf/ThreadingGtk.cpp:
* wtf/ThreadingPthreads.cpp:
2008-10-20 Geoffrey Garen <ggaren@apple.com>
Reviewed by Darin Adler.
Fixed https://bugs.webkit.org/show_bug.cgi?id=21735
Emit profiling instrumentation only if the Web Inspector's profiling
feature is enabled
22.2% speedup on empty function call benchmark.
2.9% speedup on v8 benchmark.
0.7% speedup on SunSpider.
Lesser but similar speedups in bytecode.
* VM/CTI.cpp:
(JSC::CTI::compileOpCall):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases): Nixed JITed profiler hooks. Profiler
hooks now have their own opcodes. Added support for compiling profiler
hook opcodes.
(JSC::CodeBlock::dump): Dump support for the new profiling opcodes.
* VM/CodeGenerator.h:
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::CodeGenerator):
(JSC::CodeGenerator::emitCall):
(JSC::CodeGenerator::emitConstruct): Conditionally emit profiling hooks
around call and construct, at the call site. (It's easier to get things
right this way, if you have profiled code calling non-profiled code.
Also, you get a slightly more accurate profile, since you charge the full
cost of the call / construct operation to the callee.)
Also, fixed a bug where construct would fetch the ".prototype" property
from the constructor before evaluating the arguments to the constructor,
incorrectly allowing an "invalid constructor" exception to short-circuit
argument evaluation. I encountered this bug when trying to make
constructor exceptions work with profiling.
* VM/Machine.cpp:
(JSC::Machine::callEval): Removed obsolete profiler hooks.
(JSC::Machine::throwException): Added a check for an exception thrown
within a call instruction. We didn't need this before because the call
instruction would check for a valid call before involing the profiler.
(JSC::Machine::execute): Added a didExecute hook at the end of top-level
function invocation, since op_ret no longer does this for us.
(JSC::Machine::privateExecute): Removed obsolete profiler hooks. Added
profiler opcodes. Changed some ++vPC to vPC[x] notation, since the
latter is better for performance, and it makes reasoning about the
current opcode in exception handling much simpler.
(JSC::Machine::cti_op_call_NotJSFunction): Removed obsolete profiler
hooks.
(JSC::Machine::cti_op_create_arguments_no_params): Added missing
CTI_STACK_HACK that I noticed when adding CTI_STACK_HACK to the new
profiler opcode functions.
(JSC::Machine::cti_op_profile_will_call):
(JSC::Machine::cti_op_profile_did_call): The new profiler opcode
functions.
(JSC::Machine::cti_op_construct_NotJSConstruct): Removed obsolete profiler
hooks.
* VM/Machine.h:
(JSC::Machine::isCallOpcode): Helper for exception handling.
* VM/Opcode.h: Declare new opcodes.
* kjs/JSGlobalObject.h:
(JSC::JSGlobalObject::supportsProfiling): Added virtual interface that
allows WebCore to specify whether the target global object has the Web
Inspector's profiling feature enabled.
* profiler/Profiler.cpp:
(JSC::Profiler::willExecute):
(JSC::Profiler::didExecute):
(JSC::Profiler::createCallIdentifier):
* profiler/Profiler.h: Added support for invoking the profiler with
an arbitrary JSValue*, and not a known object. We didn't need this
before because the call instruction would check for a valid call before
involing the profiler.
2008-10-20 Darin Adler <darin@apple.com>
Reviewed by Geoff Garen.
- get CTI working on Windows again
* VM/CTI.cpp:
(JSC::CTI::emitCTICall): Add an overload for functions that
return JSObject*.
* VM/CTI.h: Use JSValue* and JSObject* as return types for
cti_op functions. Apparently, MSVC doesn't handle returning
the JSValuePtr struct in a register. We'll have to look into
this more.
* VM/Machine.cpp:
(JSC::Machine::cti_op_convert_this):
(JSC::Machine::cti_op_add):
(JSC::Machine::cti_op_pre_inc):
(JSC::Machine::cti_op_new_object):
(JSC::Machine::cti_op_get_by_id):
(JSC::Machine::cti_op_get_by_id_second):
(JSC::Machine::cti_op_get_by_id_generic):
(JSC::Machine::cti_op_get_by_id_fail):
(JSC::Machine::cti_op_instanceof):
(JSC::Machine::cti_op_del_by_id):
(JSC::Machine::cti_op_mul):
(JSC::Machine::cti_op_new_func):
(JSC::Machine::cti_op_push_activation):
(JSC::Machine::cti_op_call_NotJSFunction):
(JSC::Machine::cti_op_new_array):
(JSC::Machine::cti_op_resolve):
(JSC::Machine::cti_op_construct_JSConstructFast):
(JSC::Machine::cti_op_construct_NotJSConstruct):
(JSC::Machine::cti_op_get_by_val):
(JSC::Machine::cti_op_sub):
(JSC::Machine::cti_op_lesseq):
(JSC::Machine::cti_op_negate):
(JSC::Machine::cti_op_resolve_base):
(JSC::Machine::cti_op_resolve_skip):
(JSC::Machine::cti_op_resolve_global):
(JSC::Machine::cti_op_div):
(JSC::Machine::cti_op_pre_dec):
(JSC::Machine::cti_op_not):
(JSC::Machine::cti_op_eq):
(JSC::Machine::cti_op_lshift):
(JSC::Machine::cti_op_bitand):
(JSC::Machine::cti_op_rshift):
(JSC::Machine::cti_op_bitnot):
(JSC::Machine::cti_op_new_func_exp):
(JSC::Machine::cti_op_mod):
(JSC::Machine::cti_op_less):
(JSC::Machine::cti_op_neq):
(JSC::Machine::cti_op_urshift):
(JSC::Machine::cti_op_bitxor):
(JSC::Machine::cti_op_new_regexp):
(JSC::Machine::cti_op_bitor):
(JSC::Machine::cti_op_call_eval):
(JSC::Machine::cti_op_throw):
(JSC::Machine::cti_op_next_pname):
(JSC::Machine::cti_op_typeof):
(JSC::Machine::cti_op_is_undefined):
(JSC::Machine::cti_op_is_boolean):
(JSC::Machine::cti_op_is_number):
(JSC::Machine::cti_op_is_string):
(JSC::Machine::cti_op_is_object):
(JSC::Machine::cti_op_is_function):
(JSC::Machine::cti_op_stricteq):
(JSC::Machine::cti_op_nstricteq):
(JSC::Machine::cti_op_to_jsnumber):
(JSC::Machine::cti_op_in):
(JSC::Machine::cti_op_push_new_scope):
(JSC::Machine::cti_op_del_by_val):
(JSC::Machine::cti_op_new_error):
(JSC::Machine::cti_vm_throw):
Change these functions to return pointer types, and never
JSValuePtr.
* VM/Machine.h: Ditto.
2008-10-20 Geoffrey Garen <ggaren@apple.com>
Reviewed by Darin Adler.
Fixed some recent break-age in bytecode mode.
* VM/CodeBlock.cpp:
(JSC::CodeBlock::printStructureIDs): Fixed up an ASSERT caused by
Gavin's last checkin. This is a temporary fix so I can keep on moving.
I'll send email about what I think is an underlying problem soon.
* VM/Machine.cpp:
(JSC::Machine::privateExecute): Removed a redundant and sometimes
incorrect cast, which started ASSERTing after Darin's last checkin.
2008-10-20 Darin Adler <darin@apple.com>
- another similar Windows build fix
* VM/CTI.cpp: Changed return type to JSObject* instead of JSValuePtr.
2008-10-20 Darin Adler <darin@apple.com>
- try to fix Windows build
* VM/CTI.cpp: Use JSValue* instead of JSValuePtr for ctiTrampoline.
* VM/CTI.h: Ditto.
2008-10-19 Darin Adler <darin@apple.com>
Reviewed by Cameron Zwarich.
- finish https://bugs.webkit.org/show_bug.cgi?id=21732
improve performance by eliminating JSValue as a base class for JSCell
* VM/Machine.cpp:
(JSC::Machine::cti_op_call_profiler): Use asFunction.
(JSC::Machine::cti_vm_lazyLinkCall): Ditto.
(JSC::Machine::cti_op_construct_JSConstructFast): Use asObject.
* kjs/JSCell.h: Re-sort friend classes. Eliminate inheritance from
JSValue. Changed cast in asCell from static_cast to reinterpret_cast.
Removed JSValue::getNumber(double&) and one of JSValue::getObject
overloads.
* kjs/JSValue.h: Made the private constructor and destructor both
non-virtual and also remove the definitions. This class can never
be instantiated or derived.
2008-10-19 Darin Adler <darin@apple.com>
Reviewed by Cameron Zwarich.
- next step of https://bugs.webkit.org/show_bug.cgi?id=21732
improve performance by eliminating JSValue as a base class for JSCell
Change JSValuePtr from a typedef into a class. This allows us to support
conversion from JSCell* to JSValuePtr even if JSCell isn't derived from
JSValue.
* JavaScriptCore.exp: Updated symbols that involve JSValuePtr, since
it's now a distinct type.
* API/APICast.h:
(toRef): Extract the JSValuePtr payload explicitly since we can't just
cast any more.
* VM/CTI.cpp:
(JSC::CTI::asInteger): Ditto.
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::addConstant): Get at the payload directly.
(JSC::CodeGenerator::emitLoad): Added an overload of JSCell* because
otherwise classes derived from JSValue end up calling the bool
overload instead of JSValuePtr.
* VM/CodeGenerator.h: Ditto. Also update traits to use JSValue*
and the payload functions.
* VM/Register.h: Added a JSCell* overload and use of payload functions.
* kjs/JSCell.h:
(JSC::asCell): Use payload function.
(JSC::JSValue::asCell): Use JSValue* instead of JSValuePtr.
(JSC::JSValuePtr::JSValuePtr): Added. Constructor that takes JSCell*
and creates a JSValuePtr.
* kjs/JSImmediate.h: Added JSValuePtr class. Also updated makeValue
and makeInt to work with JSValue* and the payload function.
* kjs/JSValue.h: Added == and != operators for JSValuePtr. Put them
here because eventually all the JSValue functions should go here
except what's needed by JSImmediate. Also fix asValue to use
JSValue* instead of JSValuePtr.
* kjs/PropertySlot.h: Change constructor to take JSValuePtr.
* kjs/protect.h: Update gcProtect functions to work with JSCell*
as well as JSValuePtr. Also updated the ProtectedPtr<JSValuePtr>
specialization to work more directly. Also changed all the call
sites to use gcProtectNullTolerant.
2008-10-19 Darin Adler <darin@apple.com>
Reviewed by Oliver Hunt.
- next step of https://bugs.webkit.org/show_bug.cgi?id=21732
improve performance by eliminating JSValue as a base class for JSCell
Remove most uses of JSValue, which will be removed in a future patch.
* VM/Machine.cpp:
(JSC::fastToUInt32): Call toUInt32SlowCase function; no longer a member
of JSValue.
* kjs/JSNumberCell.h:
(JSC::JSNumberCell::toInt32): Ditto.
(JSC::JSNumberCell::toUInt32): Ditto.
* kjs/JSValue.cpp:
(JSC::toInt32SlowCase): Made a non-member function.
(JSC::JSValue::toInt32SlowCase): Changed to call non-member function.
(JSC::toUInt32SlowCase): More of the same.
(JSC::JSValue::toUInt32SlowCase): Ditto.
* kjs/JSValue.h: Moved static member function so they are no longer
member functions at all.
* VM/CTI.h: Removed forward declaration of JSValue.
* VM/ExceptionHelpers.h: Ditto.
* kjs/CallData.h: Ditto.
* kjs/ConstructData.h: Ditto.
* kjs/JSGlobalObjectFunctions.h: Ditto.
* kjs/PropertyMap.h: Ditto.
* kjs/StructureID.h: Ditto.
* kjs/collector.h: Ditto.
* kjs/completion.h: Ditto.
* kjs/grammar.y:
(JSC::makeBitwiseNotNode): Call new non-member toInt32 function.
(JSC::makeLeftShiftNode): More of the same.
(JSC::makeRightShiftNode): Ditto.
* kjs/protect.h: Added a specialization for ProtectedPtr<JSValuePtr>
so this can be used with JSValuePtr.
2008-10-18 Darin Adler <darin@apple.com>
Reviewed by Oliver Hunt.
- next step of https://bugs.webkit.org/show_bug.cgi?id=21732
improve performance by eliminating JSValue as a base class for JSCell
Tweak a little more to get closer to where we can make JSValuePtr a class.
* API/APICast.h:
(toJS): Change back to JSValue* here, since we're converting the
pointer type.
* VM/CTI.cpp:
(JSC::CTI::unlinkCall): Call asPointer.
* VM/CTI.h: Cast to JSValue* here, since it's a pointer cast.
* kjs/DebuggerCallFrame.h:
(JSC::DebuggerCallFrame::DebuggerCallFrame): Call noValue.
* kjs/JSGlobalData.cpp:
(JSC::JSGlobalData::JSGlobalData): Call noValue.
* kjs/JSImmediate.cpp:
(JSC::JSImmediate::toObject): Remove unneeded const_cast.
* kjs/JSWrapperObject.h:
(JSC::JSWrapperObject::JSWrapperObject): Call noValue.
2008-10-18 Darin Adler <darin@apple.com>
- fix non-all-in-one build
* kjs/completion.h:
(JSC::Completion::Completion): Add include of JSValue.h.
2008-10-18 Darin Adler <darin@apple.com>
Reviewed by Oliver Hunt.
- fix assertions I introduced with my casting changes
These were showing up as failures in the JavaScriptCore tests.
* VM/Machine.cpp:
(JSC::Machine::cti_op_instanceof): Remove the bogus asCell casting that
was at the top of the function, and instead cast at the point of use.
(JSC::Machine::cti_op_construct_NotJSConstruct): Moved the cast to
object after checking the construct type.
2008-10-18 Darin Adler <darin@apple.com>
- fix non-all-in-one build
* kjs/JSGlobalObjectFunctions.h: Add include of JSImmedate.h (for now).
2008-10-18 Darin Adler <darin@apple.com>
- fix build
* kjs/interpreter.h: Include JSValue.h instead of JSImmediate.h.
2008-10-18 Darin Adler <darin@apple.com>
* kjs/interpreter.h: Fix include of JSImmediate.h.
2008-10-18 Darin Adler <darin@apple.com>
- fix non-all-in-one build
* kjs/interpreter.h: Add include of JSImmediate.h.
2008-10-18 Darin Adler <darin@apple.com>
- fix non-all-in-one build
* kjs/ConstructData.h: Add include of JSImmedate.h (for now).
2008-10-18 Darin Adler <darin@apple.com>
- try to fix Windows build
* VM/Machine.cpp:
(JSC::Machine::Machine): Use JSCell* type since MSVC seems to only allow
calling ~JSCell directly if it's a JSCell*.
2008-10-18 Darin Adler <darin@apple.com>
Reviewed by Cameron Zwarich.
- next step on https://bugs.webkit.org/show_bug.cgi?id=21732
improve performance by eliminating JSValue as a base class for JSCell
Use JSValuePtr everywhere instead of JSValue*. In the future, we'll be
changing JSValuePtr to be a class, and then eventually renaming it
to JSValue once that's done.
* JavaScriptCore.exp: Update entry points, since some now take JSValue*
instead of const JSValue*.
* API/APICast.h:
* API/JSCallbackConstructor.h:
* API/JSCallbackFunction.cpp:
* API/JSCallbackFunction.h:
* API/JSCallbackObject.h:
* API/JSCallbackObjectFunctions.h:
* API/JSContextRef.cpp:
* API/JSObjectRef.cpp:
* API/JSValueRef.cpp:
* VM/CTI.cpp:
* VM/CTI.h:
* VM/CodeBlock.cpp:
* VM/CodeBlock.h:
* VM/CodeGenerator.cpp:
* VM/CodeGenerator.h:
* VM/ExceptionHelpers.cpp:
* VM/ExceptionHelpers.h:
* VM/JSPropertyNameIterator.cpp:
* VM/JSPropertyNameIterator.h:
* VM/Machine.cpp:
* VM/Machine.h:
* VM/Register.h:
* kjs/ArgList.cpp:
* kjs/ArgList.h:
* kjs/Arguments.cpp:
* kjs/Arguments.h:
* kjs/ArrayConstructor.cpp:
* kjs/ArrayPrototype.cpp:
* kjs/BooleanConstructor.cpp:
* kjs/BooleanConstructor.h:
* kjs/BooleanObject.h:
* kjs/BooleanPrototype.cpp:
* kjs/CallData.cpp:
* kjs/CallData.h:
* kjs/ConstructData.cpp:
* kjs/ConstructData.h:
* kjs/DateConstructor.cpp:
* kjs/DateInstance.h:
* kjs/DatePrototype.cpp:
* kjs/DebuggerCallFrame.cpp:
* kjs/DebuggerCallFrame.h:
* kjs/ErrorConstructor.cpp:
* kjs/ErrorPrototype.cpp:
* kjs/ExecState.cpp:
* kjs/ExecState.h:
* kjs/FunctionConstructor.cpp:
* kjs/FunctionPrototype.cpp:
* kjs/GetterSetter.cpp:
* kjs/GetterSetter.h:
* kjs/InternalFunction.h:
* kjs/JSActivation.cpp:
* kjs/JSActivation.h:
* kjs/JSArray.cpp:
* kjs/JSArray.h:
* kjs/JSCell.cpp:
* kjs/JSCell.h:
* kjs/JSFunction.cpp:
* kjs/JSFunction.h:
* kjs/JSGlobalData.h:
* kjs/JSGlobalObject.cpp:
* kjs/JSGlobalObject.h:
* kjs/JSGlobalObjectFunctions.cpp:
* kjs/JSGlobalObjectFunctions.h:
* kjs/JSImmediate.cpp:
* kjs/JSImmediate.h:
* kjs/JSNotAnObject.cpp:
* kjs/JSNotAnObject.h:
* kjs/JSNumberCell.cpp:
* kjs/JSNumberCell.h:
* kjs/JSObject.cpp:
* kjs/JSObject.h:
* kjs/JSStaticScopeObject.cpp:
* kjs/JSStaticScopeObject.h:
* kjs/JSString.cpp:
* kjs/JSString.h:
* kjs/JSValue.h:
* kjs/JSVariableObject.h:
* kjs/JSWrapperObject.h:
* kjs/MathObject.cpp:
* kjs/NativeErrorConstructor.cpp:
* kjs/NumberConstructor.cpp:
* kjs/NumberConstructor.h:
* kjs/NumberObject.cpp:
* kjs/NumberObject.h:
* kjs/NumberPrototype.cpp:
* kjs/ObjectConstructor.cpp:
* kjs/ObjectPrototype.cpp:
* kjs/ObjectPrototype.h:
* kjs/PropertyMap.h:
* kjs/PropertySlot.cpp:
* kjs/PropertySlot.h:
* kjs/RegExpConstructor.cpp:
* kjs/RegExpConstructor.h:
* kjs/RegExpMatchesArray.h:
* kjs/RegExpObject.cpp:
* kjs/RegExpObject.h:
* kjs/RegExpPrototype.cpp:
* kjs/Shell.cpp:
* kjs/StringConstructor.cpp:
* kjs/StringObject.cpp:
* kjs/StringObject.h:
* kjs/StringObjectThatMasqueradesAsUndefined.h:
* kjs/StringPrototype.cpp:
* kjs/StructureID.cpp:
* kjs/StructureID.h:
* kjs/collector.cpp:
* kjs/collector.h:
* kjs/completion.h:
* kjs/grammar.y:
* kjs/interpreter.cpp:
* kjs/interpreter.h:
* kjs/lookup.cpp:
* kjs/lookup.h:
* kjs/nodes.h:
* kjs/operations.cpp:
* kjs/operations.h:
* kjs/protect.h:
* profiler/ProfileGenerator.cpp:
Replace JSValue* with JSValuePtr.
2008-10-18 Darin Adler <darin@apple.com>
* VM/Machine.cpp:
(JSC::Machine::cti_op_call_eval): Removed stray parentheses from my
last check-in.
2008-10-18 Darin Adler <darin@apple.com>
Reviewed by Oliver Hunt.
- first step of https://bugs.webkit.org/show_bug.cgi?id=21732
improve performance by eliminating JSValue as a base class for JSCell
Remove casts from JSValue* to derived classes, replacing them with
calls to inline casting functions. These functions are also a bit
better than aidrect cast because they also do a runtime assertion.
Removed use of 0 as for JSValue*, changing call sites to use a
noValue() function instead.
Move things needed by classes derived from JSValue out of the class,
since the classes won't be deriving from JSValue any more soon.
I did most of these changes by changing JSValue to not be JSValue* any
more, then fixing a lot of the compilation problems, then rolling out
the JSValue change.
1.011x as fast on SunSpider (presumably due to some of the Machine.cpp changes)
* API/APICast.h: Removed unneeded forward declarations.
* API/JSCallbackObject.h: Added an asCallbackObject function for casting.
* API/JSCallbackObjectFunctions.h:
(JSC::JSCallbackObject::asCallbackObject): Added.
(JSC::JSCallbackObject::getOwnPropertySlot): Use asObject.
(JSC::JSCallbackObject::call): Use noValue.
(JSC::JSCallbackObject::staticValueGetter): Use asCallbackObject.
(JSC::JSCallbackObject::staticFunctionGetter): Ditto.
(JSC::JSCallbackObject::callbackGetter): Ditto.
* JavaScriptCore.exp: Updated.
* JavaScriptCore.xcodeproj/project.pbxproj: Added RegExpMatchesArray.h.
* VM/CTI.cpp:
(JSC::CTI::asInteger): Added. For use casting a JSValue to an integer.
(JSC::CTI::emitGetArg): Use asInteger.
(JSC::CTI::emitGetPutArg): Ditto.
(JSC::CTI::getConstantImmediateNumericArg): Ditto. Also use noValue.
(JSC::CTI::emitInitRegister): Use asInteger.
(JSC::CTI::getDeTaggedConstantImmediate): Ditto.
(JSC::CTI::compileOpCallInitializeCallFrame): Ditto.
(JSC::CTI::compileOpCall): Ditto.
(JSC::CTI::compileOpStrictEq): Ditto.
(JSC::CTI::privateCompileMainPass): Ditto.
(JSC::CTI::privateCompileGetByIdProto): Ditto.
(JSC::CTI::privateCompileGetByIdChain): Ditto.
(JSC::CTI::privateCompilePutByIdTransition): Ditto.
* VM/CTI.h: Rewrite the ARG-related macros to use C++ casts instead of
C casts and get rid of some extra parentheses. Addd declaration of
asInteger.
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitEqualityOp): Use asString.
(JSC::CodeGenerator::emitLoad): Use noValue.
(JSC::CodeGenerator::findScopedProperty): Change globalObject argument
to JSObject* instead of JSValue*.
(JSC::CodeGenerator::emitResolve): Remove unneeded cast.
(JSC::CodeGenerator::emitGetScopedVar): Use asCell.
(JSC::CodeGenerator::emitPutScopedVar): Ditto.
* VM/CodeGenerator.h: Changed out argument of findScopedProperty.
Also change the JSValueMap to use PtrHash explicitly instead of
getting it from DefaultHash.
* VM/JSPropertyNameIterator.cpp:
(JSC::JSPropertyNameIterator::toPrimitive): Use noValue.
* VM/JSPropertyNameIterator.h:
(JSC::JSPropertyNameIterator::next): Ditto.
* VM/Machine.cpp:
(JSC::fastIsNumber): Moved isImmediate check here instead of
checking for 0 inside Heap::isNumber. Use asCell and asNumberCell.
(JSC::fastToInt32): Ditto.
(JSC::fastToUInt32): Ditto.
(JSC::jsLess): Use asString.
(JSC::jsLessEq): Ditto.
(JSC::jsAdd): Ditto.
(JSC::jsTypeStringForValue): Use asObject.
(JSC::jsIsObjectType): Ditto.
(JSC::jsIsFunctionType): Ditto.
(JSC::inlineResolveBase): Use noValue.
(JSC::Machine::callEval): Use asString. Initialize result to
undefined, not 0.
(JSC::Machine::Machine): Remove unneeded casts to JSCell*.
(JSC::Machine::throwException): Use asObject.
(JSC::Machine::debug): Remove explicit calls to the DebuggerCallFrame
constructor.
(JSC::Machine::checkTimeout): Use noValue.
(JSC::cachePrototypeChain): Use asObject.
(JSC::Machine::tryCachePutByID): Use asCell.
(JSC::Machine::tryCacheGetByID): Use aCell and asObject.
(JSC::Machine::privateExecute): Use noValue, asCell, asObject, asString,
asArray, asActivation, asFunction. Changed code that creates call frames
for host functions to pass 0 for the function pointer -- the call frame
needs a JSFunction* and a host function object is not one. This was
caught by the assertions in the casting functions. Also remove some
unneeded casts in cases where two values are compared.
(JSC::Machine::retrieveLastCaller): Use noValue.
(JSC::Machine::tryCTICachePutByID): Use asCell.
(JSC::Machine::tryCTICacheGetByID): Use aCell and asObject.
(JSC::setUpThrowTrampolineReturnAddress): Added this function to restore
the PIC-branch-avoidance that was recently lost.
(JSC::Machine::cti_op_add): Use asString.
(JSC::Machine::cti_op_instanceof): Use asCell and asObject.
(JSC::Machine::cti_op_call_JSFunction): Use asFunction.
(JSC::Machine::cti_op_call_NotJSFunction): Changed code to pass 0 for
the function pointer, since we don't have a JSFunction. Use asObject.
(JSC::Machine::cti_op_tear_off_activation): Use asActivation.
(JSC::Machine::cti_op_construct_JSConstruct): Use asFunction and asObject.
(JSC::Machine::cti_op_construct_NotJSConstruct): use asObject.
(JSC::Machine::cti_op_get_by_val): Use asArray and asString.
(JSC::Machine::cti_op_resolve_func): Use asPointer; this helps prepare
us for a situation where JSValue is not a pointer.
(JSC::Machine::cti_op_put_by_val): Use asArray.
(JSC::Machine::cti_op_put_by_val_array): Ditto.
(JSC::Machine::cti_op_resolve_global): Use asGlobalObject.
(JSC::Machine::cti_op_post_inc): Change VM_CHECK_EXCEPTION_2 to
VM_CHECK_EXCEPTION_AT_END, since there's no observable work done after
that point. Also use asPointer.
(JSC::Machine::cti_op_resolve_with_base): Use asPointer.
(JSC::Machine::cti_op_post_dec): Change VM_CHECK_EXCEPTION_2 to
VM_CHECK_EXCEPTION_AT_END, since there's no observable work done after
that point. Also use asPointer.
(JSC::Machine::cti_op_call_eval): Use asObject, noValue, and change
VM_CHECK_EXCEPTION_ARG to VM_THROW_EXCEPTION_AT_END.
(JSC::Machine::cti_op_throw): Change return value to a JSValue*.
(JSC::Machine::cti_op_in): Use asObject.
(JSC::Machine::cti_op_switch_char): Use asString.
(JSC::Machine::cti_op_switch_string): Ditto.
(JSC::Machine::cti_op_put_getter): Use asObject.
(JSC::Machine::cti_op_put_setter): Ditto.
(JSC::Machine::cti_vm_throw): Change return value to a JSValue*.
Use noValue.
* VM/Machine.h: Change return values of both cti_op_throw and
cti_vm_throw to JSValue*.
* VM/Register.h: Remove nullJSValue, which is the same thing
as noValue(). Also removed unneeded definition of JSValue.
* kjs/ArgList.h: Removed unneeded definition of JSValue.
* kjs/Arguments.h:
(JSC::asArguments): Added.
* kjs/ArrayPrototype.cpp:
(JSC::getProperty): Use noValue.
(JSC::arrayProtoFuncToString): Use asArray.
(JSC::arrayProtoFuncToLocaleString): Ditto.
(JSC::arrayProtoFuncConcat): Ditto.
(JSC::arrayProtoFuncPop): Ditto. Also removed unneeded initialization
of the result, which is set in both sides of the branch.
(JSC::arrayProtoFuncPush): Ditto.
(JSC::arrayProtoFuncShift): Removed unneeded initialization
of the result, which is set in both sides of the branch.
(JSC::arrayProtoFuncSort): Use asArray.
* kjs/BooleanObject.h:
(JSC::asBooleanObject): Added.
* kjs/BooleanPrototype.cpp:
(JSC::booleanProtoFuncToString): Use asBooleanObject.
(JSC::booleanProtoFuncValueOf): Ditto.
* kjs/CallData.cpp:
(JSC::call): Use asObject and asFunction.
* kjs/ConstructData.cpp:
(JSC::construct): Ditto.
* kjs/DateConstructor.cpp:
(JSC::constructDate): Use asDateInstance.
* kjs/DateInstance.h:
(JSC::asDateInstance): Added.
* kjs/DatePrototype.cpp:
(JSC::dateProtoFuncToString): Use asDateInstance.
(JSC::dateProtoFuncToUTCString): Ditto.
(JSC::dateProtoFuncToDateString): Ditto.
(JSC::dateProtoFuncToTimeString): Ditto.
(JSC::dateProtoFuncToLocaleString): Ditto.
(JSC::dateProtoFuncToLocaleDateString): Ditto.
(JSC::dateProtoFuncToLocaleTimeString): Ditto.
(JSC::dateProtoFuncValueOf): Ditto.
(JSC::dateProtoFuncGetTime): Ditto.
(JSC::dateProtoFuncGetFullYear): Ditto.
(JSC::dateProtoFuncGetUTCFullYear): Ditto.
(JSC::dateProtoFuncToGMTString): Ditto.
(JSC::dateProtoFuncGetMonth): Ditto.
(JSC::dateProtoFuncGetUTCMonth): Ditto.
(JSC::dateProtoFuncGetDate): Ditto.
(JSC::dateProtoFuncGetUTCDate): Ditto.
(JSC::dateProtoFuncGetDay): Ditto.
(JSC::dateProtoFuncGetUTCDay): Ditto.
(JSC::dateProtoFuncGetHours): Ditto.
(JSC::dateProtoFuncGetUTCHours): Ditto.
(JSC::dateProtoFuncGetMinutes): Ditto.
(JSC::dateProtoFuncGetUTCMinutes): Ditto.
(JSC::dateProtoFuncGetSeconds): Ditto.
(JSC::dateProtoFuncGetUTCSeconds): Ditto.
(JSC::dateProtoFuncGetMilliSeconds): Ditto.
(JSC::dateProtoFuncGetUTCMilliseconds): Ditto.
(JSC::dateProtoFuncGetTimezoneOffset): Ditto.
(JSC::dateProtoFuncSetTime): Ditto.
(JSC::setNewValueFromTimeArgs): Ditto.
(JSC::setNewValueFromDateArgs): Ditto.
(JSC::dateProtoFuncSetYear): Ditto.
(JSC::dateProtoFuncGetYear): Ditto.
* kjs/DebuggerCallFrame.cpp:
(JSC::DebuggerCallFrame::thisObject): Use asObject.
(JSC::DebuggerCallFrame::evaluate): Use noValue.
* kjs/DebuggerCallFrame.h: Added a constructor that
takes only a callFrame.
* kjs/ExecState.h:
(JSC::ExecState::clearException): Use noValue.
* kjs/FunctionPrototype.cpp:
(JSC::functionProtoFuncToString): Use asFunction.
(JSC::functionProtoFuncApply): Use asArguments and asArray.
* kjs/GetterSetter.cpp:
(JSC::GetterSetter::getPrimitiveNumber): Use noValue.
* kjs/GetterSetter.h:
(JSC::asGetterSetter): Added.
* kjs/InternalFunction.cpp:
(JSC::InternalFunction::name): Use asString.
* kjs/InternalFunction.h:
(JSC::asInternalFunction): Added.
* kjs/JSActivation.cpp:
(JSC::JSActivation::argumentsGetter): Use asActivation.
* kjs/JSActivation.h:
(JSC::asActivation): Added.
* kjs/JSArray.cpp:
(JSC::JSArray::putSlowCase): Use noValue.
(JSC::JSArray::deleteProperty): Ditto.
(JSC::JSArray::increaseVectorLength): Ditto.
(JSC::JSArray::setLength): Ditto.
(JSC::JSArray::pop): Ditto.
(JSC::JSArray::sort): Ditto.
(JSC::JSArray::compactForSorting): Ditto.
* kjs/JSArray.h:
(JSC::asArray): Added.
* kjs/JSCell.cpp:
(JSC::JSCell::getJSNumber): Use noValue.
* kjs/JSCell.h:
(JSC::asCell): Added.
(JSC::JSValue::asCell): Changed to not preserve const.
Given the wide use of JSValue* and JSCell*, it's not
really useful to use const.
(JSC::JSValue::isNumber): Use asValue.
(JSC::JSValue::isString): Ditto.
(JSC::JSValue::isGetterSetter): Ditto.
(JSC::JSValue::isObject): Ditto.
(JSC::JSValue::getNumber): Ditto.
(JSC::JSValue::getString): Ditto.
(JSC::JSValue::getObject): Ditto.
(JSC::JSValue::getCallData): Ditto.
(JSC::JSValue::getConstructData): Ditto.
(JSC::JSValue::getUInt32): Ditto.
(JSC::JSValue::getTruncatedInt32): Ditto.
(JSC::JSValue::getTruncatedUInt32): Ditto.
(JSC::JSValue::mark): Ditto.
(JSC::JSValue::marked): Ditto.
(JSC::JSValue::toPrimitive): Ditto.
(JSC::JSValue::getPrimitiveNumber): Ditto.
(JSC::JSValue::toBoolean): Ditto.
(JSC::JSValue::toNumber): Ditto.
(JSC::JSValue::toString): Ditto.
(JSC::JSValue::toObject): Ditto.
(JSC::JSValue::toThisObject): Ditto.
(JSC::JSValue::needsThisConversion): Ditto.
(JSC::JSValue::toThisString): Ditto.
(JSC::JSValue::getJSNumber): Ditto.
* kjs/JSFunction.cpp:
(JSC::JSFunction::argumentsGetter): Use asFunction.
(JSC::JSFunction::callerGetter): Ditto.
(JSC::JSFunction::lengthGetter): Ditto.
(JSC::JSFunction::construct): Use asObject.
* kjs/JSFunction.h:
(JSC::asFunction): Added.
* kjs/JSGlobalObject.cpp:
(JSC::lastInPrototypeChain): Use asObject.
* kjs/JSGlobalObject.h:
(JSC::asGlobalObject): Added.
(JSC::ScopeChainNode::globalObject): Use asGlobalObject.
* kjs/JSImmediate.h: Added noValue, asPointer, and makeValue
functions. Use rawValue, makeValue, and noValue consistently
instead of doing reinterpret_cast in various functions.
* kjs/JSNumberCell.h:
(JSC::asNumberCell): Added.
(JSC::JSValue::uncheckedGetNumber): Use asValue and asNumberCell.
(JSC::JSValue::toJSNumber): Use asValue.
* kjs/JSObject.cpp:
(JSC::JSObject::put): Use asObject and asGetterSetter.
(JSC::callDefaultValueFunction): Use noValue.
(JSC::JSObject::defineGetter): Use asGetterSetter.
(JSC::JSObject::defineSetter): Ditto.
(JSC::JSObject::lookupGetter): Ditto. Also use asObject.
(JSC::JSObject::lookupSetter): Ditto.
(JSC::JSObject::hasInstance): Use asObject.
(JSC::JSObject::fillGetterPropertySlot): Use asGetterSetter.
* kjs/JSObject.h:
(JSC::JSObject::getDirect): Use noValue.
(JSC::asObject): Added.
(JSC::JSValue::isObject): Use asValue.
(JSC::JSObject::get): Removed unneeded const_cast.
(JSC::JSObject::getPropertySlot): Use asObject.
(JSC::JSValue::get): Removed unneeded const_cast.
Use asValue, asCell, and asObject.
(JSC::JSValue::put): Ditto.
(JSC::JSObject::allocatePropertyStorageInline): Fixed spelling
of "oldPropertStorage".
* kjs/JSString.cpp:
(JSC::JSString::getOwnPropertySlot): Use asObject.
* kjs/JSString.h:
(JSC::asString): Added.
(JSC::JSValue::toThisJSString): Use asValue.
* kjs/JSValue.h: Make PreferredPrimitiveType a top level enum
instead of a member of JSValue. Added an asValue function that
returns this. Removed overload of asCell for const. Use asValue
instead of getting right at this.
* kjs/ObjectPrototype.cpp:
(JSC::objectProtoFuncIsPrototypeOf): Use asObject.
(JSC::objectProtoFuncDefineGetter): Ditto.
(JSC::objectProtoFuncDefineSetter): Ditto.
* kjs/PropertySlot.h:
(JSC::PropertySlot::PropertySlot): Take a const JSValue* so the
callers don't have to worry about const.
(JSC::PropertySlot::clearBase): Use noValue.
(JSC::PropertySlot::clearValue): Ditto.
* kjs/RegExpConstructor.cpp:
(JSC::regExpConstructorDollar1): Use asRegExpConstructor.
(JSC::regExpConstructorDollar2): Ditto.
(JSC::regExpConstructorDollar3): Ditto.
(JSC::regExpConstructorDollar4): Ditto.
(JSC::regExpConstructorDollar5): Ditto.
(JSC::regExpConstructorDollar6): Ditto.
(JSC::regExpConstructorDollar7): Ditto.
(JSC::regExpConstructorDollar8): Ditto.
(JSC::regExpConstructorDollar9): Ditto.
(JSC::regExpConstructorInput): Ditto.
(JSC::regExpConstructorMultiline): Ditto.
(JSC::regExpConstructorLastMatch): Ditto.
(JSC::regExpConstructorLastParen): Ditto.
(JSC::regExpConstructorLeftContext): Ditto.
(JSC::regExpConstructorRightContext): Ditto.
(JSC::setRegExpConstructorInput): Ditto.
(JSC::setRegExpConstructorMultiline): Ditto.
(JSC::constructRegExp): Use asObject.
* kjs/RegExpConstructor.h:
(JSC::asRegExpConstructor): Added.
* kjs/RegExpObject.cpp:
(JSC::regExpObjectGlobal): Use asRegExpObject.
(JSC::regExpObjectIgnoreCase): Ditto.
(JSC::regExpObjectMultiline): Ditto.
(JSC::regExpObjectSource): Ditto.
(JSC::regExpObjectLastIndex): Ditto.
(JSC::setRegExpObjectLastIndex): Ditto.
(JSC::callRegExpObject): Ditto.
* kjs/RegExpObject.h:
(JSC::asRegExpObject): Added.
* kjs/RegExpPrototype.cpp:
(JSC::regExpProtoFuncTest): Use asRegExpObject.
(JSC::regExpProtoFuncExec): Ditto.
(JSC::regExpProtoFuncCompile): Ditto.
(JSC::regExpProtoFuncToString): Ditto.
* kjs/StringObject.h:
(JSC::StringObject::internalValue): Use asString.
(JSC::asStringObject): Added.
* kjs/StringPrototype.cpp:
(JSC::stringProtoFuncReplace): Use asRegExpObject.
(JSC::stringProtoFuncToString): Ue asStringObject.
(JSC::stringProtoFuncMatch): Use asRegExpObject.
(JSC::stringProtoFuncSearch): Ditto.
(JSC::stringProtoFuncSplit): Ditto.
* kjs/StructureID.cpp:
(JSC::StructureID::getEnumerablePropertyNames): Use asObject.
(JSC::StructureID::createCachedPrototypeChain): Ditto.
(JSC::StructureIDChain::StructureIDChain): Use asCell and asObject.
* kjs/collector.h:
(JSC::Heap::isNumber): Removed null handling. This can only be called
on valid cells.
(JSC::Heap::cellBlock): Removed overload for const and non-const.
Whether the JSCell* is const or not really should have no effect on
whether you can modify the collector block it's in.
* kjs/interpreter.cpp:
(JSC::Interpreter::evaluate): Use noValue and noObject.
* kjs/nodes.cpp:
(JSC::FunctionCallResolveNode::emitCode): Use JSObject for the global
object rather than JSValue.
(JSC::PostfixResolveNode::emitCode): Ditto.
(JSC::PrefixResolveNode::emitCode): Ditto.
(JSC::ReadModifyResolveNode::emitCode): Ditto.
(JSC::AssignResolveNode::emitCode): Ditto.
* kjs/operations.h:
(JSC::equalSlowCaseInline): Use asString, asCell, asNumberCell,
(JSC::strictEqualSlowCaseInline): Ditto.
2008-10-18 Cameron Zwarich <zwarich@apple.com>
Reviewed by Oliver Hunt.
Bug 21702: Special op_create_activation for the case where there are no named parameters
<https://bugs.webkit.org/show_bug.cgi?id=21702>
This is a 2.5% speedup on the V8 Raytrace benchmark and a 1.1% speedup
on the V8 Earley-Boyer benchmark.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
* VM/Machine.cpp:
(JSC::Machine::cti_op_create_arguments_no_params):
* VM/Machine.h:
* kjs/Arguments.h:
(JSC::Arguments::):
(JSC::Arguments::Arguments):
2008-10-17 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich.
- in debug builds, alter the stack to avoid blowing out MallocStackLogging
(In essence, while executing a CTI function we alter the return
address to jscGeneratedNativeCode so that a single consistent
function is on the stack instead of many random functions without
symbols.)
* VM/CTI.h:
* VM/Machine.cpp:
(JSC::doSetReturnAddress):
(JSC::):
(JSC::StackHack::StackHack):
(JSC::StackHack::~StackHack):
(JSC::Machine::cti_op_convert_this):
(JSC::Machine::cti_op_end):
(JSC::Machine::cti_op_add):
(JSC::Machine::cti_op_pre_inc):
(JSC::Machine::cti_timeout_check):
(JSC::Machine::cti_register_file_check):
(JSC::Machine::cti_op_loop_if_less):
(JSC::Machine::cti_op_loop_if_lesseq):
(JSC::Machine::cti_op_new_object):
(JSC::Machine::cti_op_put_by_id):
(JSC::Machine::cti_op_put_by_id_second):
(JSC::Machine::cti_op_put_by_id_generic):
(JSC::Machine::cti_op_put_by_id_fail):
(JSC::Machine::cti_op_get_by_id):
(JSC::Machine::cti_op_get_by_id_second):
(JSC::Machine::cti_op_get_by_id_generic):
(JSC::Machine::cti_op_get_by_id_fail):
(JSC::Machine::cti_op_instanceof):
(JSC::Machine::cti_op_del_by_id):
(JSC::Machine::cti_op_mul):
(JSC::Machine::cti_op_new_func):
(JSC::Machine::cti_op_call_profiler):
(JSC::Machine::cti_op_call_JSFunction):
(JSC::Machine::cti_vm_lazyLinkCall):
(JSC::Machine::cti_vm_compile):
(JSC::Machine::cti_op_push_activation):
(JSC::Machine::cti_op_call_NotJSFunction):
(JSC::Machine::cti_op_create_arguments):
(JSC::Machine::cti_op_tear_off_activation):
(JSC::Machine::cti_op_tear_off_arguments):
(JSC::Machine::cti_op_ret_profiler):
(JSC::Machine::cti_op_ret_scopeChain):
(JSC::Machine::cti_op_new_array):
(JSC::Machine::cti_op_resolve):
(JSC::Machine::cti_op_construct_JSConstructFast):
(JSC::Machine::cti_op_construct_JSConstruct):
(JSC::Machine::cti_op_construct_NotJSConstruct):
(JSC::Machine::cti_op_get_by_val):
(JSC::Machine::cti_op_resolve_func):
(JSC::Machine::cti_op_sub):
(JSC::Machine::cti_op_put_by_val):
(JSC::Machine::cti_op_put_by_val_array):
(JSC::Machine::cti_op_lesseq):
(JSC::Machine::cti_op_loop_if_true):
(JSC::Machine::cti_op_negate):
(JSC::Machine::cti_op_resolve_base):
(JSC::Machine::cti_op_resolve_skip):
(JSC::Machine::cti_op_resolve_global):
(JSC::Machine::cti_op_div):
(JSC::Machine::cti_op_pre_dec):
(JSC::Machine::cti_op_jless):
(JSC::Machine::cti_op_not):
(JSC::Machine::cti_op_jtrue):
(JSC::Machine::cti_op_post_inc):
(JSC::Machine::cti_op_eq):
(JSC::Machine::cti_op_lshift):
(JSC::Machine::cti_op_bitand):
(JSC::Machine::cti_op_rshift):
(JSC::Machine::cti_op_bitnot):
(JSC::Machine::cti_op_resolve_with_base):
(JSC::Machine::cti_op_new_func_exp):
(JSC::Machine::cti_op_mod):
(JSC::Machine::cti_op_less):
(JSC::Machine::cti_op_neq):
(JSC::Machine::cti_op_post_dec):
(JSC::Machine::cti_op_urshift):
(JSC::Machine::cti_op_bitxor):
(JSC::Machine::cti_op_new_regexp):
(JSC::Machine::cti_op_bitor):
(JSC::Machine::cti_op_call_eval):
(JSC::Machine::cti_op_throw):
(JSC::Machine::cti_op_get_pnames):
(JSC::Machine::cti_op_next_pname):
(JSC::Machine::cti_op_push_scope):
(JSC::Machine::cti_op_pop_scope):
(JSC::Machine::cti_op_typeof):
(JSC::Machine::cti_op_is_undefined):
(JSC::Machine::cti_op_is_boolean):
(JSC::Machine::cti_op_is_number):
(JSC::Machine::cti_op_is_string):
(JSC::Machine::cti_op_is_object):
(JSC::Machine::cti_op_is_function):
(JSC::Machine::cti_op_stricteq):
(JSC::Machine::cti_op_nstricteq):
(JSC::Machine::cti_op_to_jsnumber):
(JSC::Machine::cti_op_in):
(JSC::Machine::cti_op_push_new_scope):
(JSC::Machine::cti_op_jmp_scopes):
(JSC::Machine::cti_op_put_by_index):
(JSC::Machine::cti_op_switch_imm):
(JSC::Machine::cti_op_switch_char):
(JSC::Machine::cti_op_switch_string):
(JSC::Machine::cti_op_del_by_val):
(JSC::Machine::cti_op_put_getter):
(JSC::Machine::cti_op_put_setter):
(JSC::Machine::cti_op_new_error):
(JSC::Machine::cti_op_debug):
(JSC::Machine::cti_vm_throw):
2008-10-17 Gavin Barraclough <barraclough@apple.com>
Optimize op_call by allowing call sites to be directly linked to callees.
For the hot path of op_call, CTI now generates a check (initially for an impossible
value), and the first time the call is executed we attempt to link the call directly
to the callee. We can currently only do so if the arity of the caller and callee
match. The (optimized) setup for the call on the hot path is linked directly to
the ctiCode for the callee, without indirection.
Two forms of the slow case of the call are generated, the first will be executed the
first time the call is reached. As well as this path attempting to link the call to
a callee, it also relinks the slow case to a second slow case, which will not continue
to attempt relinking the call. (This policy could be changed in future, but for not
this is intended to prevent thrashing).
If a callee that the caller has been linked to is garbage collected, then the link
in the caller's JIt code will be reset back to a value that cannot match - to prevent
any false positive matches.
~20% progression on deltablue & richards, >12% overall reduction in v8-tests
runtime, one or two percent progression on sunspider.
Reviewed by Oliver Hunt.
* VM/CTI.cpp:
(JSC::):
(JSC::CTI::emitNakedCall):
(JSC::unreachable):
(JSC::CTI::compileOpCallInitializeCallFrame):
(JSC::CTI::compileOpCallSetupArgs):
(JSC::CTI::compileOpCall):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
(JSC::CTI::privateCompile):
(JSC::CTI::unlinkCall):
(JSC::CTI::linkCall):
* VM/CTI.h:
* VM/CodeBlock.cpp:
(JSC::CodeBlock::~CodeBlock):
(JSC::CodeBlock::unlinkCallers):
(JSC::CodeBlock::derefStructureIDs):
* VM/CodeBlock.h:
(JSC::StructureStubInfo::StructureStubInfo):
(JSC::CallLinkInfo::CallLinkInfo):
(JSC::CodeBlock::addCaller):
(JSC::CodeBlock::removeCaller):
(JSC::CodeBlock::getStubInfo):
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitCall):
(JSC::CodeGenerator::emitConstruct):
* VM/Machine.cpp:
(JSC::Machine::cti_op_call_profiler):
(JSC::Machine::cti_op_call_JSFunction):
(JSC::Machine::cti_vm_lazyLinkCall):
(JSC::Machine::cti_op_construct_JSConstructFast):
(JSC::Machine::cti_op_construct_JSConstruct):
(JSC::Machine::cti_op_construct_NotJSConstruct):
* VM/Machine.h:
* kjs/JSFunction.cpp:
(JSC::JSFunction::~JSFunction):
* kjs/JSFunction.h:
* kjs/nodes.h:
(JSC::FunctionBodyNode::):
* masm/X86Assembler.h:
(JSC::X86Assembler::getDifferenceBetweenLabels):
2008-10-17 Maciej Stachowiak <mjs@apple.com>
Reviewed by Geoff Garen.
- remove ASSERT that makes the leaks buildbot cry
* kjs/JSFunction.cpp:
(JSC::JSFunction::JSFunction):
2008-10-17 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich
- don't bother to do arguments tearoff when it will have no effect
~1% on v8 raytrace
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitReturn):
2008-10-17 Marco Barisione <marco.barisione@collabora.co.uk>
Reviewed by Sam Weinig. Landed by Jan Alonzo.
https://bugs.webkit.org/show_bug.cgi?id=21603
[GTK] Minor fixes to GOwnPtr
* wtf/GOwnPtr.cpp:
(WTF::GError):
(WTF::GList):
(WTF::GCond):
(WTF::GMutex):
(WTF::GPatternSpec):
(WTF::GDir):
* wtf/GOwnPtr.h:
(WTF::freeOwnedGPtr):
(WTF::GOwnPtr::~GOwnPtr):
(WTF::GOwnPtr::outPtr):
(WTF::GOwnPtr::set):
(WTF::GOwnPtr::clear):
* wtf/Threading.h:
2008-10-17 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich.
- speed up transitions that resize the property storage a fair bit
~3% speedup on v8 RayTrace benchmark, ~1% on DeltaBlue
* VM/CTI.cpp:
(JSC::resizePropertyStorage): renamed from transitionObject, and reduced to just resize
the object's property storage with one inline call.
(JSC::CTI::privateCompilePutByIdTransition): Use a separate function for property storage
resize, but still do all the rest of the work in assembly in that case, and pass the known
compile-time constants of old and new size rather than structureIDs, saving a bunch of
redundant memory access.
* kjs/JSObject.cpp:
(JSC::JSObject::allocatePropertyStorage): Just call the inline version.
* kjs/JSObject.h:
(JSC::JSObject::allocatePropertyStorageInline): Inline version of allocatePropertyStorage
* masm/X86Assembler.h:
(JSC::X86Assembler::):
(JSC::X86Assembler::pushl_i32): Add code to assmeble push of a constant; code originally by Cameron Zwarich.
2008-10-17 Cameron Zwarich <zwarich@apple.com>
Reviewed by Maciej Stachowiak.
Remove some C style casts.
* masm/X86Assembler.h:
(JSC::JITCodeBuffer::putIntUnchecked):
(JSC::X86Assembler::link):
(JSC::X86Assembler::linkAbsoluteAddress):
(JSC::X86Assembler::getRelocatedAddress):
2008-10-17 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Maciej Stachowiak.
Remove some C style casts.
* VM/CTI.cpp:
(JSC::CTI::patchGetByIdSelf):
(JSC::CTI::patchPutByIdReplace):
* VM/Machine.cpp:
(JSC::Machine::tryCTICachePutByID):
(JSC::Machine::tryCTICacheGetByID):
(JSC::Machine::cti_op_put_by_id):
(JSC::Machine::cti_op_put_by_id_fail):
(JSC::Machine::cti_op_get_by_id):
(JSC::Machine::cti_op_get_by_id_fail):
2008-10-17 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich.
- Avoid restoring the caller's 'r' value in op_ret
https://bugs.webkit.org/show_bug.cgi?id=21319
This patch stops writing the call frame at call and return points;
instead it does so immediately before any CTI call.
0.5% speedup or so on the v8 benchmark
* VM/CTI.cpp:
(JSC::CTI::emitCTICall):
(JSC::CTI::compileOpCall):
(JSC::CTI::emitSlowScriptCheck):
(JSC::CTI::compileBinaryArithOpSlowCase):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
(JSC::CTI::privateCompile):
* VM/CTI.h:
2008-10-17 Cameron Zwarich <zwarich@apple.com>
Reviewed by Sam Weinig.
Make WREC require CTI because it won't actually compile otherwise.
* wtf/Platform.h:
2008-10-16 Maciej Stachowiak <mjs@apple.com>
Reviewed by Geoff Garen.
- fixed <rdar://problem/5806316> JavaScriptCore should not force building with gcc 4.0
- use gcc 4.2 when building with Xcode 3.1 or newer on Leopard, even though this is not the default
This time there is no performance regression; we can avoid having
to use the fastcall calling convention for CTI functions by using
varargs to prevent the compiler from moving things around on the
stack.
* Configurations/DebugRelease.xcconfig:
* JavaScriptCore.xcodeproj/project.pbxproj:
* VM/CTI.cpp:
* VM/Machine.h:
* wtf/Platform.h:
2008-10-16 Maciej Stachowiak <mjs@apple.com>
Reviewed by Oliver Hunt.
- fix for REGRESSION: r37631 causing crashes on buildbot
https://bugs.webkit.org/show_bug.cgi?id=21682
* kjs/collector.cpp:
(JSC::Heap::collect): Avoid crashing when a GC occurs while no global objects are live.
2008-10-16 Sam Weinig <sam@webkit.org>
Reviewed by Maciej Stachowiak.
Fix for https://bugs.webkit.org/show_bug.cgi?id=21683
Don't create intermediate StructureIDs for builtin objects
First step in reduce number of StructureIDs created when initializing the
JSGlobalObject.
- In order to avoid creating the intermediate StructureIDs use the new putDirectWithoutTransition
and putDirectFunctionWithoutTransition to add properties to JSObjects without transitioning
the StructureID. This patch just implements this strategy for ObjectPrototype but alone
reduces the number of StructureIDs create for about:blank by 10, from 142 to 132.
* kjs/JSGlobalObject.cpp:
(JSC::JSGlobalObject::reset):
* kjs/JSObject.cpp:
(JSC::JSObject::putDirectFunctionWithoutTransition):
* kjs/JSObject.h:
(JSC::JSObject::putDirectWithoutTransition):
* kjs/ObjectPrototype.cpp:
(JSC::ObjectPrototype::ObjectPrototype):
* kjs/ObjectPrototype.h:
* kjs/StructureID.cpp:
(JSC::StructureID::addPropertyWithoutTransition):
* kjs/StructureID.h:
2008-10-16 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich.
- fix for: REGRESSION: over 100 StructureIDs leak loading about:blank (result of fix for bug 21633)
Apparent slight progression (< 0.5%) on v8 benchmarks and SunSpider.
* kjs/StructureID.cpp:
(JSC::StructureID::~StructureID): Don't deref this object's parent's pointer to
itself from the destructor; that doesn't even make sense.
(JSC::StructureID::addPropertyTransition): Don't refer the single transition;
the rule is that parent StructureIDs are ref'd but child ones are not. Refing
the child creates a cycle.
2008-10-15 Alexey Proskuryakov <ap@webkit.org>
Reviewed by Darin Adler.
https://bugs.webkit.org/show_bug.cgi?id=21609
Make MessagePorts protect their peers across heaps
* JavaScriptCore.exp:
* kjs/JSGlobalObject.cpp:
(JSC::JSGlobalObject::markCrossHeapDependentObjects):
* kjs/JSGlobalObject.h:
* kjs/collector.cpp:
(JSC::Heap::collect):
Before GC sweep phase, a function supplied by global object is now called for all global
objects in the heap, making it possible to implement cross-heap dependencies.
2008-10-15 Alexey Proskuryakov <ap@webkit.org>
Reviewed by Darin Adler.
https://bugs.webkit.org/show_bug.cgi?id=21610
run-webkit-threads --threaded crashes in StructureID destructor
* kjs/StructureID.cpp:
(JSC::StructureID::StructureID):
(JSC::StructureID::~StructureID):
Protect access to a static (debug-only) HashSet with a lock.
2008-10-15 Sam Weinig <sam@webkit.org>
Reviewed by Goeffrey Garen.
Add function to dump statistics for StructureIDs.
* kjs/StructureID.cpp:
(JSC::StructureID::dumpStatistics):
(JSC::StructureID::StructureID):
(JSC::StructureID::~StructureID):
* kjs/StructureID.h:
2008-10-15 Cameron Zwarich <zwarich@apple.com>
Reviewed by Maciej Stachowiak.
Bug 21633: Avoid using a HashMap when there is only a single transition
<https://bugs.webkit.org/show_bug.cgi?id=21633>
This is a 0.8% speedup on SunSpider and between a 0.5% and 1.0% speedup
on the V8 benchmark suite, depending on which harness we use. It will
also slightly reduce the memory footprint of a StructureID.
* kjs/StructureID.cpp:
(JSC::StructureID::StructureID):
(JSC::StructureID::~StructureID):
(JSC::StructureID::addPropertyTransition):
* kjs/StructureID.h:
(JSC::StructureID::):
2008-10-15 Csaba Osztrogonac <oszi@inf.u-szeged.hu>
Reviewed by Geoffrey Garen.
1.40% speedup on SunSpider, 1.44% speedup on V8. (Linux)
No change on Mac.
* VM/Machine.cpp:
(JSC::fastIsNumber): ALWAYS_INLINE modifier added.
2008-10-15 Geoffrey Garen <ggaren@apple.com>
Reviewed by Cameron Zwarich.
Fixed https://bugs.webkit.org/show_bug.cgi?id=21345
Start the debugger without reloading the inspected page
* JavaScriptCore.exp: New symbols.
* JavaScriptCore.xcodeproj/project.pbxproj: New files.
* VM/CodeBlock.h:
(JSC::EvalCodeCache::get): Updated for tweak to parsing API.
* kjs/CollectorHeapIterator.h: Added. An iterator for the object heap,
which we use to find all the live functions and recompile them.
* kjs/DebuggerCallFrame.cpp:
(JSC::DebuggerCallFrame::evaluate): Updated for tweak to parsing API.
* kjs/FunctionConstructor.cpp:
(JSC::constructFunction): Updated for tweak to parsing API.
* kjs/JSFunction.cpp:
(JSC::JSFunction::JSFunction): Try to validate our SourceCode in debug
builds by ASSERTing that it's syntactically valid. This doesn't catch
all SourceCode bugs, but it catches a lot of them.
* kjs/JSGlobalObjectFunctions.cpp:
(JSC::globalFuncEval): Updated for tweak to parsing API.
* kjs/Parser.cpp:
(JSC::Parser::parse):
* kjs/Parser.h:
(JSC::Parser::parse): Tweaked the parser to make it possible to parse
without an ExecState, and to allow the client to specify a debugger to
notify (or not) about the source we parse. This allows the inspector
to recompile even though no JavaScript is executing, then notify the
debugger about all source code when it's done.
* kjs/Shell.cpp:
(prettyPrintScript): Updated for tweak to parsing API.
* kjs/SourceRange.h:
(JSC::SourceCode::isNull): Added to help with ASSERTs.
* kjs/collector.cpp:
(JSC::Heap::heapAllocate):
(JSC::Heap::sweep):
(JSC::Heap::primaryHeapBegin):
(JSC::Heap::primaryHeapEnd):
* kjs/collector.h:
(JSC::): Moved a bunch of declarations around to enable compilation of
CollectorHeapIterator.
* kjs/interpreter.cpp:
(JSC::Interpreter::checkSyntax):
(JSC::Interpreter::evaluate): Updated for tweak to parsing API.
* kjs/lexer.h:
(JSC::Lexer::sourceCode): BUG FIX: Calculate SourceCode ranges relative
to the SourceCode range in which we're lexing, otherwise nested functions
that are compiled individually get SourceCode ranges that don't reflect
their nesting.
* kjs/nodes.cpp:
(JSC::FunctionBodyNode::FunctionBodyNode):
(JSC::FunctionBodyNode::finishParsing):
(JSC::FunctionBodyNode::create):
(JSC::FunctionBodyNode::copyParameters):
* kjs/nodes.h:
(JSC::ScopeNode::setSource):
(JSC::FunctionBodyNode::parameterCount): Added some helper functions for
copying one FunctionBodyNode's parameters to another. The recompiler uses
these when calling "finishParsing".
2008-10-15 Joerg Bornemann <joerg.bornemann@trolltech.com>
Reviewed by Darin Adler.
- part of https://bugs.webkit.org/show_bug.cgi?id=20746
Fix compilation on Windows CE.
str(n)icmp, strdup and vsnprintf are not available on Windows CE,
they are called _str(n)icmp, etc. instead
* wtf/StringExtras.h: Added inline function implementations.
2008-10-15 Gabor Loki <loki@inf.u-szeged.hu>
Reviewed by Cameron Zwarich.
<https://bugs.webkit.org/show_bug.cgi?id=20912>
Use simple uint32_t multiplication on op_mul if both operands are
immediate number and they are between zero and 0x7FFF.
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
2008-10-09 Darin Fisher <darin@chromium.org>
Reviewed by Sam Weinig.
Make pan scrolling a platform configurable option.
https://bugs.webkit.org/show_bug.cgi?id=21515
* wtf/Platform.h: Add ENABLE_PAN_SCROLLING
2008-10-14 Maciej Stachowiak <mjs@apple.com>
Rubber stamped by Sam Weinig.
- revert r37572 and r37581 for now
Turns out GCC 4.2 is still a (small) regression, we'll have to do
more work to turn it on.
* Configurations/DebugRelease.xcconfig:
* JavaScriptCore.xcodeproj/project.pbxproj:
* VM/CTI.cpp:
* VM/CTI.h:
* VM/Machine.cpp:
(JSC::Machine::cti_op_convert_this):
(JSC::Machine::cti_op_end):
(JSC::Machine::cti_op_add):
(JSC::Machine::cti_op_pre_inc):
(JSC::Machine::cti_timeout_check):
(JSC::Machine::cti_register_file_check):
(JSC::Machine::cti_op_loop_if_less):
(JSC::Machine::cti_op_loop_if_lesseq):
(JSC::Machine::cti_op_new_object):
(JSC::Machine::cti_op_put_by_id):
(JSC::Machine::cti_op_put_by_id_second):
(JSC::Machine::cti_op_put_by_id_generic):
(JSC::Machine::cti_op_put_by_id_fail):
(JSC::Machine::cti_op_get_by_id):
(JSC::Machine::cti_op_get_by_id_second):
(JSC::Machine::cti_op_get_by_id_generic):
(JSC::Machine::cti_op_get_by_id_fail):
(JSC::Machine::cti_op_instanceof):
(JSC::Machine::cti_op_del_by_id):
(JSC::Machine::cti_op_mul):
(JSC::Machine::cti_op_new_func):
(JSC::Machine::cti_op_call_JSFunction):
(JSC::Machine::cti_vm_compile):
(JSC::Machine::cti_op_push_activation):
(JSC::Machine::cti_op_call_NotJSFunction):
(JSC::Machine::cti_op_create_arguments):
(JSC::Machine::cti_op_tear_off_activation):
(JSC::Machine::cti_op_tear_off_arguments):
(JSC::Machine::cti_op_ret_profiler):
(JSC::Machine::cti_op_ret_scopeChain):
(JSC::Machine::cti_op_new_array):
(JSC::Machine::cti_op_resolve):
(JSC::Machine::cti_op_construct_JSConstruct):
(JSC::Machine::cti_op_construct_NotJSConstruct):
(JSC::Machine::cti_op_get_by_val):
(JSC::Machine::cti_op_resolve_func):
(JSC::Machine::cti_op_sub):
(JSC::Machine::cti_op_put_by_val):
(JSC::Machine::cti_op_put_by_val_array):
(JSC::Machine::cti_op_lesseq):
(JSC::Machine::cti_op_loop_if_true):
(JSC::Machine::cti_op_negate):
(JSC::Machine::cti_op_resolve_base):
(JSC::Machine::cti_op_resolve_skip):
(JSC::Machine::cti_op_resolve_global):
(JSC::Machine::cti_op_div):
(JSC::Machine::cti_op_pre_dec):
(JSC::Machine::cti_op_jless):
(JSC::Machine::cti_op_not):
(JSC::Machine::cti_op_jtrue):
(JSC::Machine::cti_op_post_inc):
(JSC::Machine::cti_op_eq):
(JSC::Machine::cti_op_lshift):
(JSC::Machine::cti_op_bitand):
(JSC::Machine::cti_op_rshift):
(JSC::Machine::cti_op_bitnot):
(JSC::Machine::cti_op_resolve_with_base):
(JSC::Machine::cti_op_new_func_exp):
(JSC::Machine::cti_op_mod):
(JSC::Machine::cti_op_less):
(JSC::Machine::cti_op_neq):
(JSC::Machine::cti_op_post_dec):
(JSC::Machine::cti_op_urshift):
(JSC::Machine::cti_op_bitxor):
(JSC::Machine::cti_op_new_regexp):
(JSC::Machine::cti_op_bitor):
(JSC::Machine::cti_op_call_eval):
(JSC::Machine::cti_op_throw):
(JSC::Machine::cti_op_get_pnames):
(JSC::Machine::cti_op_next_pname):
(JSC::Machine::cti_op_push_scope):
(JSC::Machine::cti_op_pop_scope):
(JSC::Machine::cti_op_typeof):
(JSC::Machine::cti_op_is_undefined):
(JSC::Machine::cti_op_is_boolean):
(JSC::Machine::cti_op_is_number):
(JSC::Machine::cti_op_is_string):
(JSC::Machine::cti_op_is_object):
(JSC::Machine::cti_op_is_function):
(JSC::Machine::cti_op_stricteq):
(JSC::Machine::cti_op_nstricteq):
(JSC::Machine::cti_op_to_jsnumber):
(JSC::Machine::cti_op_in):
(JSC::Machine::cti_op_push_new_scope):
(JSC::Machine::cti_op_jmp_scopes):
(JSC::Machine::cti_op_put_by_index):
(JSC::Machine::cti_op_switch_imm):
(JSC::Machine::cti_op_switch_char):
(JSC::Machine::cti_op_switch_string):
(JSC::Machine::cti_op_del_by_val):
(JSC::Machine::cti_op_put_getter):
(JSC::Machine::cti_op_put_setter):
(JSC::Machine::cti_op_new_error):
(JSC::Machine::cti_op_debug):
(JSC::Machine::cti_vm_throw):
* VM/Machine.h:
* masm/X86Assembler.h:
(JSC::X86Assembler::emitRestoreArgumentReference):
(JSC::X86Assembler::emitRestoreArgumentReferenceForTrampoline):
* wtf/Platform.h:
2008-10-14 Alexey Proskuryakov <ap@webkit.org>
Reviewed by Darin Adler.
https://bugs.webkit.org/show_bug.cgi?id=20256
Array.push and other standard methods disappear
* kjs/JSGlobalData.cpp:
(JSC::JSGlobalData::JSGlobalData):
(JSC::JSGlobalData::~JSGlobalData):
Don't use static hash tables even on platforms that don't enable JSC_MULTIPLE_THREADS -
these tables reference IdentifierTable, which is always per-GlobalData.
2008-10-14 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich.
- always use CTI_ARGUMENTS and CTI_ARGUMENTS_FASTCALL
This is a small regression for GCC 4.0, but simplifies the code
for future improvements and lets us focus on GCC 4.2+ and MSVC.
* VM/CTI.cpp:
* VM/CTI.h:
* VM/Machine.cpp:
(JSC::Machine::cti_op_convert_this):
(JSC::Machine::cti_op_end):
(JSC::Machine::cti_op_add):
(JSC::Machine::cti_op_pre_inc):
(JSC::Machine::cti_timeout_check):
(JSC::Machine::cti_register_file_check):
(JSC::Machine::cti_op_loop_if_less):
(JSC::Machine::cti_op_loop_if_lesseq):
(JSC::Machine::cti_op_new_object):
(JSC::Machine::cti_op_put_by_id):
(JSC::Machine::cti_op_put_by_id_second):
(JSC::Machine::cti_op_put_by_id_generic):
(JSC::Machine::cti_op_put_by_id_fail):
(JSC::Machine::cti_op_get_by_id):
(JSC::Machine::cti_op_get_by_id_second):
(JSC::Machine::cti_op_get_by_id_generic):
(JSC::Machine::cti_op_get_by_id_fail):
(JSC::Machine::cti_op_instanceof):
(JSC::Machine::cti_op_del_by_id):
(JSC::Machine::cti_op_mul):
(JSC::Machine::cti_op_new_func):
(JSC::Machine::cti_op_call_JSFunction):
(JSC::Machine::cti_vm_compile):
(JSC::Machine::cti_op_push_activation):
(JSC::Machine::cti_op_call_NotJSFunction):
(JSC::Machine::cti_op_create_arguments):
(JSC::Machine::cti_op_tear_off_activation):
(JSC::Machine::cti_op_tear_off_arguments):
(JSC::Machine::cti_op_ret_profiler):
(JSC::Machine::cti_op_ret_scopeChain):
(JSC::Machine::cti_op_new_array):
(JSC::Machine::cti_op_resolve):
(JSC::Machine::cti_op_construct_JSConstruct):
(JSC::Machine::cti_op_construct_NotJSConstruct):
(JSC::Machine::cti_op_get_by_val):
(JSC::Machine::cti_op_resolve_func):
(JSC::Machine::cti_op_sub):
(JSC::Machine::cti_op_put_by_val):
(JSC::Machine::cti_op_put_by_val_array):
(JSC::Machine::cti_op_lesseq):
(JSC::Machine::cti_op_loop_if_true):
(JSC::Machine::cti_op_negate):
(JSC::Machine::cti_op_resolve_base):
(JSC::Machine::cti_op_resolve_skip):
(JSC::Machine::cti_op_resolve_global):
(JSC::Machine::cti_op_div):
(JSC::Machine::cti_op_pre_dec):
(JSC::Machine::cti_op_jless):
(JSC::Machine::cti_op_not):
(JSC::Machine::cti_op_jtrue):
(JSC::Machine::cti_op_post_inc):
(JSC::Machine::cti_op_eq):
(JSC::Machine::cti_op_lshift):
(JSC::Machine::cti_op_bitand):
(JSC::Machine::cti_op_rshift):
(JSC::Machine::cti_op_bitnot):
(JSC::Machine::cti_op_resolve_with_base):
(JSC::Machine::cti_op_new_func_exp):
(JSC::Machine::cti_op_mod):
(JSC::Machine::cti_op_less):
(JSC::Machine::cti_op_neq):
(JSC::Machine::cti_op_post_dec):
(JSC::Machine::cti_op_urshift):
(JSC::Machine::cti_op_bitxor):
(JSC::Machine::cti_op_new_regexp):
(JSC::Machine::cti_op_bitor):
(JSC::Machine::cti_op_call_eval):
(JSC::Machine::cti_op_throw):
(JSC::Machine::cti_op_get_pnames):
(JSC::Machine::cti_op_next_pname):
(JSC::Machine::cti_op_push_scope):
(JSC::Machine::cti_op_pop_scope):
(JSC::Machine::cti_op_typeof):
(JSC::Machine::cti_op_is_undefined):
(JSC::Machine::cti_op_is_boolean):
(JSC::Machine::cti_op_is_number):
(JSC::Machine::cti_op_is_string):
(JSC::Machine::cti_op_is_object):
(JSC::Machine::cti_op_is_function):
(JSC::Machine::cti_op_stricteq):
(JSC::Machine::cti_op_nstricteq):
(JSC::Machine::cti_op_to_jsnumber):
(JSC::Machine::cti_op_in):
(JSC::Machine::cti_op_push_new_scope):
(JSC::Machine::cti_op_jmp_scopes):
(JSC::Machine::cti_op_put_by_index):
(JSC::Machine::cti_op_switch_imm):
(JSC::Machine::cti_op_switch_char):
(JSC::Machine::cti_op_switch_string):
(JSC::Machine::cti_op_del_by_val):
(JSC::Machine::cti_op_put_getter):
(JSC::Machine::cti_op_put_setter):
(JSC::Machine::cti_op_new_error):
(JSC::Machine::cti_op_debug):
(JSC::Machine::cti_vm_throw):
* VM/Machine.h:
* masm/X86Assembler.h:
(JSC::X86Assembler::emitRestoreArgumentReference):
(JSC::X86Assembler::emitRestoreArgumentReferenceForTrampoline):
* wtf/Platform.h:
2008-10-13 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich.
- make Machine::getArgumentsData an Arguments method and inline it
~2% on v8 raytrace
* VM/Machine.cpp:
* kjs/Arguments.h:
(JSC::Machine::getArgumentsData):
2008-10-13 Alp Toker <alp@nuanti.com>
Fix autotools dist build target by listing recently added header
files only. Not reviewed.
* GNUmakefile.am:
2008-10-13 Maciej Stachowiak <mjs@apple.com>
Rubber stamped by Mark Rowe.
- fixed <rdar://problem/5806316> JavaScriptCore should not force building with gcc 4.0
- use gcc 4.2 when building with Xcode 3.1 or newer on Leopard, even though this is not the default
* Configurations/DebugRelease.xcconfig:
* JavaScriptCore.xcodeproj/project.pbxproj:
2008-10-13 Cameron Zwarich <zwarich@apple.com>
Reviewed by Geoff Garen.
Bug 21541: Move RegisterFile growth check to callee
<https://bugs.webkit.org/show_bug.cgi?id=21541>
Move the RegisterFile growth check to the callee in the common case,
where some of the information is known statically at JIT time. There is
still a check in the caller in the case where the caller provides too
few arguments.
This is a 2.1% speedup on the V8 benchmark, including a 5.1% speedup on
the Richards benchmark, a 4.1% speedup on the DeltaBlue benchmark, and a
1.4% speedup on the Earley-Boyer benchmark. It is also a 0.5% speedup on
SunSpider.
* VM/CTI.cpp:
(JSC::CTI::privateCompile):
* VM/Machine.cpp:
(JSC::Machine::cti_register_file_check):
(JSC::Machine::cti_op_call_JSFunction):
(JSC::Machine::cti_op_construct_JSConstruct):
* VM/Machine.h:
* VM/RegisterFile.h:
* masm/X86Assembler.h:
(JSC::X86Assembler::):
(JSC::X86Assembler::cmpl_mr):
(JSC::X86Assembler::emitUnlinkedJg):
2008-10-13 Sam Weinig <sam@webkit.org>
Reviewed by Dan Bernstein.
Fix for https://bugs.webkit.org/show_bug.cgi?id=21577
5 false positive StructureID leaks
- Add leak ignore set to StructureID to selectively ignore leaking some StructureIDs.
- Add create method to JSGlolalData to be used when the data will be intentionally
leaked and ignore all leaks caused the StructureIDs stored in it.
* JavaScriptCore.exp:
* kjs/JSGlobalData.cpp:
(JSC::JSGlobalData::createLeaked):
* kjs/JSGlobalData.h:
* kjs/StructureID.cpp:
(JSC::StructureID::StructureID):
(JSC::StructureID::~StructureID):
(JSC::StructureID::startIgnoringLeaks):
(JSC::StructureID::stopIgnoringLeaks):
* kjs/StructureID.h:
2008-10-13 Marco Barisione <marco.barisione@collabora.co.uk>
Reviewed by Darin Adler. Landed by Jan Alonzo.
WebKit GTK Port needs a smartpointer to handle g_free (GFreePtr?)
http://bugs.webkit.org/show_bug.cgi?id=20483
Add a GOwnPtr smart pointer (similar to OwnPtr) to handle memory
allocated by GLib and start the conversion to use it.
* GNUmakefile.am:
* wtf/GOwnPtr.cpp: Added.
(WTF::GError):
(WTF::GList):
(WTF::GCond):
(WTF::GMutex):
(WTF::GPatternSpec):
(WTF::GDir):
* wtf/GOwnPtr.h: Added.
(WTF::freeOwnedPtr):
(WTF::GOwnPtr::GOwnPtr):
(WTF::GOwnPtr::~GOwnPtr):
(WTF::GOwnPtr::get):
(WTF::GOwnPtr::release):
(WTF::GOwnPtr::rawPtr):
(WTF::GOwnPtr::set):
(WTF::GOwnPtr::clear):
(WTF::GOwnPtr::operator*):
(WTF::GOwnPtr::operator->):
(WTF::GOwnPtr::operator!):
(WTF::GOwnPtr::operator UnspecifiedBoolType):
(WTF::GOwnPtr::swap):
(WTF::swap):
(WTF::operator==):
(WTF::operator!=):
(WTF::getPtr):
* wtf/Threading.h:
* wtf/ThreadingGtk.cpp:
(WTF::Mutex::~Mutex):
(WTF::Mutex::lock):
(WTF::Mutex::tryLock):
(WTF::Mutex::unlock):
(WTF::ThreadCondition::~ThreadCondition):
(WTF::ThreadCondition::wait):
(WTF::ThreadCondition::timedWait):
(WTF::ThreadCondition::signal):
(WTF::ThreadCondition::broadcast):
2008-10-12 Gabriella Toth <gtoth@inf.u-szeged.hu>
Reviewed by Darin Adler.
- part of https://bugs.webkit.org/show_bug.cgi?id=21055
Bug 21055: not invoked functions
* kjs/nodes.cpp: Deleted a function that is not invoked:
statementListInitializeVariableAccessStack.
2008-10-12 Darin Adler <darin@apple.com>
Reviewed by Sam Weinig.
* wtf/unicode/icu/UnicodeIcu.h: Fixed indentation to match WebKit coding style.
* wtf/unicode/qt4/UnicodeQt4.h: Ditto.
2008-10-12 Darin Adler <darin@apple.com>
Reviewed by Sam Weinig.
- https://bugs.webkit.org/show_bug.cgi?id=21556
Bug 21556: non-ASCII digits are allowed in places where only ASCII should be
* wtf/unicode/icu/UnicodeIcu.h: Removed isDigit, digitValue, and isFormatChar.
* wtf/unicode/qt4/UnicodeQt4.h: Ditto.
2008-10-12 Anders Carlsson <andersca@apple.com>
Reviewed by Darin Adler.
Make the append method that takes a Vector more strict - it now requires the elements
of the vector to be appended same type as the elements of the Vector they're being appended to.
This would cause problems when dealing with Vectors containing other Vectors.
* wtf/Vector.h:
(WTF::::append):
2008-10-11 Cameron Zwarich <zwarich@apple.com>
Reviewed by Sam Weinig.
Clean up RegExpMatchesArray.h to match our coding style.
* kjs/RegExpMatchesArray.h:
(JSC::RegExpMatchesArray::getOwnPropertySlot):
(JSC::RegExpMatchesArray::put):
(JSC::RegExpMatchesArray::deleteProperty):
(JSC::RegExpMatchesArray::getPropertyNames):
2008-10-11 Cameron Zwarich <zwarich@apple.com>
Reviewed by Sam Weinig.
Bug 21525: 55 StructureID leaks on Wikitravel's main page
<https://bugs.webkit.org/show_bug.cgi?id=21525>
Bug 21533: Simple JavaScript code leaks StructureIDs
<https://bugs.webkit.org/show_bug.cgi?id=21533>
StructureID::getEnumerablePropertyNames() ends up calling back to itself
via JSObject::getPropertyNames(), which causes the PropertyNameArray to
be cached twice. This leads to a memory leak in almost every use of
JSObject::getPropertyNames() on an object. The fix here is based on a
suggestion of Sam Weinig.
This patch also fixes every StructureID leaks that occurs while running
the Mozilla MemBuster test.
* kjs/PropertyNameArray.h:
(JSC::PropertyNameArray::PropertyNameArray):
(JSC::PropertyNameArray::setCacheable):
(JSC::PropertyNameArray::cacheable):
* kjs/StructureID.cpp:
(JSC::StructureID::getEnumerablePropertyNames):
2008-10-10 Oliver Hunt <oliver@apple.com>
Reviewed by Cameron Zwarich.
Use fastcall calling convention on GCC > 4.0
Results in a 2-3% improvement in GCC 4.2 performance, so
that it is no longer a regression vs. GCC 4.0
* VM/CTI.cpp:
* VM/Machine.h:
* wtf/Platform.h:
2008-10-10 Sam Weinig <sam@webkit.org>
Reviewed by Darin Adler.
- Add a workaround for a bug in ceil in Darwin libc.
- Remove old workarounds for JS math functions that are not needed
anymore.
The math functions are heavily tested by fast/js/math.html.
* kjs/MathObject.cpp:
(JSC::mathProtoFuncAbs): Remove workaround.
(JSC::mathProtoFuncCeil): Ditto.
(JSC::mathProtoFuncFloor): Ditto.
* wtf/MathExtras.h:
(wtf_ceil): Add ceil workaround for darwin.
2008-10-10 Sam Weinig <sam@webkit.org>
Reviewed by Darin Adler
Add Assertions to JSObject constructor.
* kjs/JSObject.h:
(JSC::JSObject::JSObject):
2008-10-10 Sam Weinig <sam@webkit.org>
Reviewed by Cameron Zwarich.
Remove now unused m_getterSetterFlag variable from PropertyMap.
* kjs/PropertyMap.cpp:
(JSC::PropertyMap::operator=):
* kjs/PropertyMap.h:
(JSC::PropertyMap::PropertyMap):
2008-10-09 Sam Weinig <sam@webkit.org>
Reviewed by Maciej Stachowiak.
Add leaks checking to StructureID.
* kjs/StructureID.cpp:
(JSC::StructureID::StructureID):
(JSC::StructureID::~StructureID):
2008-10-09 Alp Toker <alp@nuanti.com>
Reviewed by Mark Rowe.
https://bugs.webkit.org/show_bug.cgi?id=20760
Implement support for x86 Linux in CTI
Prepare to enable CTI/WREC on supported architectures.
Make it possible to use the CTI_ARGUMENT workaround with GCC as well
as MSVC by fixing some preprocessor conditionals.
Note that CTI/WREC no longer requires CTI_ARGUMENT on Linux so we
don't actually enable it except when building with MSVC. GCC on Win32
remains untested.
Adapt inline ASM code to use the global symbol underscore prefix only
on Darwin and to call the properly mangled Machine::cti_vm_throw
symbol name depending on CTI_ARGUMENT.
Also avoid global inclusion of the JIT infrastructure headers
throughout WebCore and WebKit causing recompilation of about ~1500
source files after modification to X86Assembler.h, CTI.h, WREC.h,
which are only used deep inside JavaScriptCore.
* GNUmakefile.am:
* VM/CTI.cpp:
* VM/CTI.h:
* VM/Machine.cpp:
* VM/Machine.h:
* kjs/regexp.cpp:
(JSC::RegExp::RegExp):
(JSC::RegExp::~RegExp):
(JSC::RegExp::match):
* kjs/regexp.h:
* masm/X86Assembler.h:
(JSC::X86Assembler::emitConvertToFastCall):
(JSC::X86Assembler::emitRestoreArgumentReferenceForTrampoline):
(JSC::X86Assembler::emitRestoreArgumentReference):
2008-10-09 Gavin Barraclough <barraclough@apple.com>
Reviewed by Cameron Zwarich.
Fix for bug #21160, x=0;1/(x*-1) == -Infinity
* ChangeLog:
* VM/CTI.cpp:
(JSC::CTI::emitFastArithDeTagImmediate):
(JSC::CTI::emitFastArithDeTagImmediateJumpIfZero):
(JSC::CTI::compileBinaryArithOp):
(JSC::CTI::compileBinaryArithOpSlowCase):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
* VM/CTI.h:
* masm/X86Assembler.h:
(JSC::X86Assembler::):
(JSC::X86Assembler::emitUnlinkedJs):
2008-10-09 Cameron Zwarich <zwarich@apple.com>
Reviewed by Oliver Hunt.
Bug 21459: REGRESSION (r37324): Safari crashes inside JavaScriptCore while browsing hulu.com
<https://bugs.webkit.org/show_bug.cgi?id=21459>
After r37324, an Arguments object does not mark an associated activation
object. This change was made because Arguments no longer directly used
the activation object in any way. However, if an activation is torn off,
then the backing store of Arguments becomes the register array of the
activation object. Arguments directly marks all of the arguments, but
the activation object is being collected, which causes its register
array to be freed and new memory to be allocated in its place.
Unfortunately, it does not seem possible to reproduce this issue in a
layout test.
* kjs/Arguments.cpp:
(JSC::Arguments::mark):
* kjs/Arguments.h:
(JSC::Arguments::setActivation):
(JSC::Arguments::Arguments):
(JSC::JSActivation::copyRegisters):
2008-10-09 Ariya Hidayat <ariya.hidayat@trolltech.com>
Reviewed by Simon.
Build fix for MinGW.
* wtf/AlwaysInline.h:
2008-10-08 Cameron Zwarich <zwarich@apple.com>
Reviewed by Maciej Stachowiak.
Bug 21497: REGRESSION (r37433): Bytecode JSC tests are severely broken
<https://bugs.webkit.org/show_bug.cgi?id=21497>
Fix a typo in r37433 that causes the failure of a large number of JSC
tests with the bytecode interpreter enabled.
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
2008-10-08 Mark Rowe <mrowe@apple.com>
Windows build fix.
* VM/CTI.cpp:
(JSC::): Update type of argument to ctiTrampoline.
2008-10-08 Darin Adler <darin@apple.com>
Reviewed by Cameron Zwarich.
- https://bugs.webkit.org/show_bug.cgi?id=21403
Bug 21403: use new CallFrame class rather than Register* for call frame manipulation
Add CallFrame as a synonym for ExecState. Arguably, some day we should switch every
client over to the new name.
Use CallFrame* consistently rather than Register* or ExecState* in low-level code such
as Machine.cpp and CTI.cpp. Similarly, use callFrame rather than r as its name and use
accessor functions to get at things in the frame.
Eliminate other uses of ExecState* that aren't needed, replacing in some cases with
JSGlobalData* and in other cases eliminating them entirely.
* API/JSObjectRef.cpp:
(JSObjectMakeFunctionWithCallback):
(JSObjectMakeFunction):
(JSObjectHasProperty):
(JSObjectGetProperty):
(JSObjectSetProperty):
(JSObjectDeleteProperty):
* API/OpaqueJSString.cpp:
* API/OpaqueJSString.h:
* VM/CTI.cpp:
(JSC::CTI::getConstant):
(JSC::CTI::emitGetArg):
(JSC::CTI::emitGetPutArg):
(JSC::CTI::getConstantImmediateNumericArg):
(JSC::CTI::printOpcodeOperandTypes):
(JSC::CTI::CTI):
(JSC::CTI::compileOpCall):
(JSC::CTI::compileBinaryArithOp):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompile):
(JSC::CTI::privateCompileGetByIdProto):
(JSC::CTI::privateCompileGetByIdChain):
(JSC::CTI::compileRegExp):
* VM/CTI.h:
* VM/CodeBlock.h:
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitEqualityOp):
(JSC::CodeGenerator::emitLoad):
(JSC::CodeGenerator::emitUnexpectedLoad):
(JSC::CodeGenerator::emitConstruct):
* VM/CodeGenerator.h:
* VM/Machine.cpp:
(JSC::jsLess):
(JSC::jsLessEq):
(JSC::jsAddSlowCase):
(JSC::jsAdd):
(JSC::jsTypeStringForValue):
(JSC::Machine::resolve):
(JSC::Machine::resolveSkip):
(JSC::Machine::resolveGlobal):
(JSC::inlineResolveBase):
(JSC::Machine::resolveBase):
(JSC::Machine::resolveBaseAndProperty):
(JSC::Machine::resolveBaseAndFunc):
(JSC::Machine::slideRegisterWindowForCall):
(JSC::isNotObject):
(JSC::Machine::callEval):
(JSC::Machine::dumpCallFrame):
(JSC::Machine::dumpRegisters):
(JSC::Machine::unwindCallFrame):
(JSC::Machine::throwException):
(JSC::DynamicGlobalObjectScope::DynamicGlobalObjectScope):
(JSC::DynamicGlobalObjectScope::~DynamicGlobalObjectScope):
(JSC::Machine::execute):
(JSC::Machine::debug):
(JSC::Machine::createExceptionScope):
(JSC::cachePrototypeChain):
(JSC::Machine::tryCachePutByID):
(JSC::Machine::tryCacheGetByID):
(JSC::Machine::privateExecute):
(JSC::Machine::retrieveArguments):
(JSC::Machine::retrieveCaller):
(JSC::Machine::retrieveLastCaller):
(JSC::Machine::findFunctionCallFrame):
(JSC::Machine::getArgumentsData):
(JSC::Machine::tryCTICachePutByID):
(JSC::Machine::getCTIArrayLengthTrampoline):
(JSC::Machine::getCTIStringLengthTrampoline):
(JSC::Machine::tryCTICacheGetByID):
(JSC::Machine::cti_op_convert_this):
(JSC::Machine::cti_op_end):
(JSC::Machine::cti_op_add):
(JSC::Machine::cti_op_pre_inc):
(JSC::Machine::cti_timeout_check):
(JSC::Machine::cti_op_loop_if_less):
(JSC::Machine::cti_op_loop_if_lesseq):
(JSC::Machine::cti_op_new_object):
(JSC::Machine::cti_op_put_by_id):
(JSC::Machine::cti_op_put_by_id_second):
(JSC::Machine::cti_op_put_by_id_generic):
(JSC::Machine::cti_op_put_by_id_fail):
(JSC::Machine::cti_op_get_by_id):
(JSC::Machine::cti_op_get_by_id_second):
(JSC::Machine::cti_op_get_by_id_generic):
(JSC::Machine::cti_op_get_by_id_fail):
(JSC::Machine::cti_op_instanceof):
(JSC::Machine::cti_op_del_by_id):
(JSC::Machine::cti_op_mul):
(JSC::Machine::cti_op_new_func):
(JSC::Machine::cti_op_call_JSFunction):
(JSC::Machine::cti_vm_compile):
(JSC::Machine::cti_op_push_activation):
(JSC::Machine::cti_op_call_NotJSFunction):
(JSC::Machine::cti_op_create_arguments):
(JSC::Machine::cti_op_tear_off_activation):
(JSC::Machine::cti_op_tear_off_arguments):
(JSC::Machine::cti_op_ret_profiler):
(JSC::Machine::cti_op_ret_scopeChain):
(JSC::Machine::cti_op_new_array):
(JSC::Machine::cti_op_resolve):
(JSC::Machine::cti_op_construct_JSConstruct):
(JSC::Machine::cti_op_construct_NotJSConstruct):
(JSC::Machine::cti_op_get_by_val):
(JSC::Machine::cti_op_resolve_func):
(JSC::Machine::cti_op_sub):
(JSC::Machine::cti_op_put_by_val):
(JSC::Machine::cti_op_put_by_val_array):
(JSC::Machine::cti_op_lesseq):
(JSC::Machine::cti_op_loop_if_true):
(JSC::Machine::cti_op_negate):
(JSC::Machine::cti_op_resolve_base):
(JSC::Machine::cti_op_resolve_skip):
(JSC::Machine::cti_op_resolve_global):
(JSC::Machine::cti_op_div):
(JSC::Machine::cti_op_pre_dec):
(JSC::Machine::cti_op_jless):
(JSC::Machine::cti_op_not):
(JSC::Machine::cti_op_jtrue):
(JSC::Machine::cti_op_post_inc):
(JSC::Machine::cti_op_eq):
(JSC::Machine::cti_op_lshift):
(JSC::Machine::cti_op_bitand):
(JSC::Machine::cti_op_rshift):
(JSC::Machine::cti_op_bitnot):
(JSC::Machine::cti_op_resolve_with_base):
(JSC::Machine::cti_op_new_func_exp):
(JSC::Machine::cti_op_mod):
(JSC::Machine::cti_op_less):
(JSC::Machine::cti_op_neq):
(JSC::Machine::cti_op_post_dec):
(JSC::Machine::cti_op_urshift):
(JSC::Machine::cti_op_bitxor):
(JSC::Machine::cti_op_new_regexp):
(JSC::Machine::cti_op_bitor):
(JSC::Machine::cti_op_call_eval):
(JSC::Machine::cti_op_throw):
(JSC::Machine::cti_op_get_pnames):
(JSC::Machine::cti_op_next_pname):
(JSC::Machine::cti_op_push_scope):
(JSC::Machine::cti_op_pop_scope):
(JSC::Machine::cti_op_typeof):
(JSC::Machine::cti_op_to_jsnumber):
(JSC::Machine::cti_op_in):
(JSC::Machine::cti_op_push_new_scope):
(JSC::Machine::cti_op_jmp_scopes):
(JSC::Machine::cti_op_put_by_index):
(JSC::Machine::cti_op_switch_imm):
(JSC::Machine::cti_op_switch_char):
(JSC::Machine::cti_op_switch_string):
(JSC::Machine::cti_op_del_by_val):
(JSC::Machine::cti_op_put_getter):
(JSC::Machine::cti_op_put_setter):
(JSC::Machine::cti_op_new_error):
(JSC::Machine::cti_op_debug):
(JSC::Machine::cti_vm_throw):
* VM/Machine.h:
* VM/Register.h:
* VM/RegisterFile.h:
* kjs/Arguments.h:
* kjs/DebuggerCallFrame.cpp:
(JSC::DebuggerCallFrame::functionName):
(JSC::DebuggerCallFrame::type):
(JSC::DebuggerCallFrame::thisObject):
(JSC::DebuggerCallFrame::evaluate):
* kjs/DebuggerCallFrame.h:
* kjs/ExecState.cpp:
(JSC::CallFrame::thisValue):
* kjs/ExecState.h:
* kjs/FunctionConstructor.cpp:
(JSC::constructFunction):
* kjs/JSActivation.cpp:
(JSC::JSActivation::JSActivation):
(JSC::JSActivation::argumentsGetter):
* kjs/JSActivation.h:
* kjs/JSGlobalObject.cpp:
(JSC::JSGlobalObject::init):
* kjs/JSGlobalObjectFunctions.cpp:
(JSC::globalFuncEval):
* kjs/JSVariableObject.h:
* kjs/Parser.cpp:
(JSC::Parser::parse):
* kjs/RegExpConstructor.cpp:
(JSC::constructRegExp):
* kjs/RegExpPrototype.cpp:
(JSC::regExpProtoFuncCompile):
* kjs/Shell.cpp:
(prettyPrintScript):
* kjs/StringPrototype.cpp:
(JSC::stringProtoFuncMatch):
(JSC::stringProtoFuncSearch):
* kjs/identifier.cpp:
(JSC::Identifier::checkSameIdentifierTable):
* kjs/interpreter.cpp:
(JSC::Interpreter::checkSyntax):
(JSC::Interpreter::evaluate):
* kjs/nodes.cpp:
(JSC::ThrowableExpressionData::emitThrowError):
(JSC::RegExpNode::emitCode):
(JSC::ArrayNode::emitCode):
(JSC::InstanceOfNode::emitCode):
* kjs/nodes.h:
* kjs/regexp.cpp:
(JSC::RegExp::RegExp):
(JSC::RegExp::create):
* kjs/regexp.h:
* profiler/HeavyProfile.h:
* profiler/Profile.h:
* wrec/WREC.cpp:
* wrec/WREC.h:
2008-10-08 Mark Rowe <mrowe@apple.com>
Typed by Maciej Stachowiak, reviewed by Mark Rowe.
Fix crash in fast/js/constant-folding.html with CTI disabled.
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
2008-10-08 Timothy Hatcher <timothy@apple.com>
Roll out r37427 because it causes an infinite recursion loading about:blank.
https://bugs.webkit.org/show_bug.cgi?id=21476
2008-10-08 Darin Adler <darin@apple.com>
Reviewed by Cameron Zwarich.
- https://bugs.webkit.org/show_bug.cgi?id=21403
Bug 21403: use new CallFrame class rather than Register* for call frame manipulation
Add CallFrame as a synonym for ExecState. Arguably, some day we should switch every
client over to the new name.
Use CallFrame* consistently rather than Register* or ExecState* in low-level code such
as Machine.cpp and CTI.cpp. Similarly, use callFrame rather than r as its name and use
accessor functions to get at things in the frame.
Eliminate other uses of ExecState* that aren't needed, replacing in some cases with
JSGlobalData* and in other cases eliminating them entirely.
* API/JSObjectRef.cpp:
(JSObjectMakeFunctionWithCallback):
(JSObjectMakeFunction):
(JSObjectHasProperty):
(JSObjectGetProperty):
(JSObjectSetProperty):
(JSObjectDeleteProperty):
* API/OpaqueJSString.cpp:
* API/OpaqueJSString.h:
* VM/CTI.cpp:
(JSC::CTI::getConstant):
(JSC::CTI::emitGetArg):
(JSC::CTI::emitGetPutArg):
(JSC::CTI::getConstantImmediateNumericArg):
(JSC::CTI::printOpcodeOperandTypes):
(JSC::CTI::CTI):
(JSC::CTI::compileOpCall):
(JSC::CTI::compileBinaryArithOp):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompile):
(JSC::CTI::privateCompileGetByIdProto):
(JSC::CTI::privateCompileGetByIdChain):
(JSC::CTI::compileRegExp):
* VM/CTI.h:
* VM/CodeBlock.h:
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitEqualityOp):
(JSC::CodeGenerator::emitLoad):
(JSC::CodeGenerator::emitUnexpectedLoad):
(JSC::CodeGenerator::emitConstruct):
* VM/CodeGenerator.h:
* VM/Machine.cpp:
(JSC::jsLess):
(JSC::jsLessEq):
(JSC::jsAddSlowCase):
(JSC::jsAdd):
(JSC::jsTypeStringForValue):
(JSC::Machine::resolve):
(JSC::Machine::resolveSkip):
(JSC::Machine::resolveGlobal):
(JSC::inlineResolveBase):
(JSC::Machine::resolveBase):
(JSC::Machine::resolveBaseAndProperty):
(JSC::Machine::resolveBaseAndFunc):
(JSC::Machine::slideRegisterWindowForCall):
(JSC::isNotObject):
(JSC::Machine::callEval):
(JSC::Machine::dumpCallFrame):
(JSC::Machine::dumpRegisters):
(JSC::Machine::unwindCallFrame):
(JSC::Machine::throwException):
(JSC::DynamicGlobalObjectScope::DynamicGlobalObjectScope):
(JSC::DynamicGlobalObjectScope::~DynamicGlobalObjectScope):
(JSC::Machine::execute):
(JSC::Machine::debug):
(JSC::Machine::createExceptionScope):
(JSC::cachePrototypeChain):
(JSC::Machine::tryCachePutByID):
(JSC::Machine::tryCacheGetByID):
(JSC::Machine::privateExecute):
(JSC::Machine::retrieveArguments):
(JSC::Machine::retrieveCaller):
(JSC::Machine::retrieveLastCaller):
(JSC::Machine::findFunctionCallFrame):
(JSC::Machine::getArgumentsData):
(JSC::Machine::tryCTICachePutByID):
(JSC::Machine::getCTIArrayLengthTrampoline):
(JSC::Machine::getCTIStringLengthTrampoline):
(JSC::Machine::tryCTICacheGetByID):
(JSC::Machine::cti_op_convert_this):
(JSC::Machine::cti_op_end):
(JSC::Machine::cti_op_add):
(JSC::Machine::cti_op_pre_inc):
(JSC::Machine::cti_timeout_check):
(JSC::Machine::cti_op_loop_if_less):
(JSC::Machine::cti_op_loop_if_lesseq):
(JSC::Machine::cti_op_new_object):
(JSC::Machine::cti_op_put_by_id):
(JSC::Machine::cti_op_put_by_id_second):
(JSC::Machine::cti_op_put_by_id_generic):
(JSC::Machine::cti_op_put_by_id_fail):
(JSC::Machine::cti_op_get_by_id):
(JSC::Machine::cti_op_get_by_id_second):
(JSC::Machine::cti_op_get_by_id_generic):
(JSC::Machine::cti_op_get_by_id_fail):
(JSC::Machine::cti_op_instanceof):
(JSC::Machine::cti_op_del_by_id):
(JSC::Machine::cti_op_mul):
(JSC::Machine::cti_op_new_func):
(JSC::Machine::cti_op_call_JSFunction):
(JSC::Machine::cti_vm_compile):
(JSC::Machine::cti_op_push_activation):
(JSC::Machine::cti_op_call_NotJSFunction):
(JSC::Machine::cti_op_create_arguments):
(JSC::Machine::cti_op_tear_off_activation):
(JSC::Machine::cti_op_tear_off_arguments):
(JSC::Machine::cti_op_ret_profiler):
(JSC::Machine::cti_op_ret_scopeChain):
(JSC::Machine::cti_op_new_array):
(JSC::Machine::cti_op_resolve):
(JSC::Machine::cti_op_construct_JSConstruct):
(JSC::Machine::cti_op_construct_NotJSConstruct):
(JSC::Machine::cti_op_get_by_val):
(JSC::Machine::cti_op_resolve_func):
(JSC::Machine::cti_op_sub):
(JSC::Machine::cti_op_put_by_val):
(JSC::Machine::cti_op_put_by_val_array):
(JSC::Machine::cti_op_lesseq):
(JSC::Machine::cti_op_loop_if_true):
(JSC::Machine::cti_op_negate):
(JSC::Machine::cti_op_resolve_base):
(JSC::Machine::cti_op_resolve_skip):
(JSC::Machine::cti_op_resolve_global):
(JSC::Machine::cti_op_div):
(JSC::Machine::cti_op_pre_dec):
(JSC::Machine::cti_op_jless):
(JSC::Machine::cti_op_not):
(JSC::Machine::cti_op_jtrue):
(JSC::Machine::cti_op_post_inc):
(JSC::Machine::cti_op_eq):
(JSC::Machine::cti_op_lshift):
(JSC::Machine::cti_op_bitand):
(JSC::Machine::cti_op_rshift):
(JSC::Machine::cti_op_bitnot):
(JSC::Machine::cti_op_resolve_with_base):
(JSC::Machine::cti_op_new_func_exp):
(JSC::Machine::cti_op_mod):
(JSC::Machine::cti_op_less):
(JSC::Machine::cti_op_neq):
(JSC::Machine::cti_op_post_dec):
(JSC::Machine::cti_op_urshift):
(JSC::Machine::cti_op_bitxor):
(JSC::Machine::cti_op_new_regexp):
(JSC::Machine::cti_op_bitor):
(JSC::Machine::cti_op_call_eval):
(JSC::Machine::cti_op_throw):
(JSC::Machine::cti_op_get_pnames):
(JSC::Machine::cti_op_next_pname):
(JSC::Machine::cti_op_push_scope):
(JSC::Machine::cti_op_pop_scope):
(JSC::Machine::cti_op_typeof):
(JSC::Machine::cti_op_to_jsnumber):
(JSC::Machine::cti_op_in):
(JSC::Machine::cti_op_push_new_scope):
(JSC::Machine::cti_op_jmp_scopes):
(JSC::Machine::cti_op_put_by_index):
(JSC::Machine::cti_op_switch_imm):
(JSC::Machine::cti_op_switch_char):
(JSC::Machine::cti_op_switch_string):
(JSC::Machine::cti_op_del_by_val):
(JSC::Machine::cti_op_put_getter):
(JSC::Machine::cti_op_put_setter):
(JSC::Machine::cti_op_new_error):
(JSC::Machine::cti_op_debug):
(JSC::Machine::cti_vm_throw):
* VM/Machine.h:
* VM/Register.h:
* VM/RegisterFile.h:
* kjs/Arguments.h:
* kjs/DebuggerCallFrame.cpp:
(JSC::DebuggerCallFrame::functionName):
(JSC::DebuggerCallFrame::type):
(JSC::DebuggerCallFrame::thisObject):
(JSC::DebuggerCallFrame::evaluate):
* kjs/DebuggerCallFrame.h:
* kjs/ExecState.cpp:
(JSC::CallFrame::thisValue):
* kjs/ExecState.h:
* kjs/FunctionConstructor.cpp:
(JSC::constructFunction):
* kjs/JSActivation.cpp:
(JSC::JSActivation::JSActivation):
(JSC::JSActivation::argumentsGetter):
* kjs/JSActivation.h:
* kjs/JSGlobalObject.cpp:
(JSC::JSGlobalObject::init):
* kjs/JSGlobalObjectFunctions.cpp:
(JSC::globalFuncEval):
* kjs/JSVariableObject.h:
* kjs/Parser.cpp:
(JSC::Parser::parse):
* kjs/RegExpConstructor.cpp:
(JSC::constructRegExp):
* kjs/RegExpPrototype.cpp:
(JSC::regExpProtoFuncCompile):
* kjs/Shell.cpp:
(prettyPrintScript):
* kjs/StringPrototype.cpp:
(JSC::stringProtoFuncMatch):
(JSC::stringProtoFuncSearch):
* kjs/identifier.cpp:
(JSC::Identifier::checkSameIdentifierTable):
* kjs/interpreter.cpp:
(JSC::Interpreter::checkSyntax):
(JSC::Interpreter::evaluate):
* kjs/nodes.cpp:
(JSC::ThrowableExpressionData::emitThrowError):
(JSC::RegExpNode::emitCode):
(JSC::ArrayNode::emitCode):
(JSC::InstanceOfNode::emitCode):
* kjs/nodes.h:
* kjs/regexp.cpp:
(JSC::RegExp::RegExp):
(JSC::RegExp::create):
* kjs/regexp.h:
* profiler/HeavyProfile.h:
* profiler/Profile.h:
* wrec/WREC.cpp:
* wrec/WREC.h:
2008-10-08 Prasanth Ullattil <pullatti@trolltech.com>
Reviewed by Oliver Hunt.
Avoid endless loops when compiling without the computed goto
optimization.
NEXT_OPCODE expands to "continue", which will not work inside
loops.
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
2008-10-08 Maciej Stachowiak <mjs@apple.com>
Reviewed by Oliver Hunt.
Re-landing the following fix with the crashing bug in it fixed (r37405):
- optimize away multiplication by constant 1.0
2.3% speedup on v8 RayTrace benchmark
Apparently it's not uncommon for JavaScript code to multiply by
constant 1.0 in the mistaken belief that this converts integer to
floating point and that there is any operational difference.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass): Optimize to_jsnumber for
case where parameter is already number.
(JSC::CTI::privateCompileSlowCases): ditto
* VM/Machine.cpp:
(JSC::Machine::privateExecute): ditto
* kjs/grammar.y:
(makeMultNode): Transform as follows:
+FOO * BAR ==> FOO * BAR
FOO * +BAR ==> FOO * BAR
FOO * 1 ==> +FOO
1 * FOO ==> +FOO
(makeDivNode): Transform as follows:
+FOO / BAR ==> FOO / BAR
FOO / +BAR ==> FOO / BAR
(makeSubNode): Transform as follows:
+FOO - BAR ==> FOO - BAR
FOO - +BAR ==> FOO - BAR
* kjs/nodes.h:
(JSC::ExpressionNode::stripUnaryPlus): Helper for above
grammar.y changes
(JSC::UnaryPlusNode::stripUnaryPlus): ditto
2008-10-08 Maciej Stachowiak <mjs@apple.com>
Reviewed by Oliver Hunt.
- correctly handle appending -0 to a string, it should stringify as just 0
* kjs/ustring.cpp:
(JSC::concatenate):
2008-10-08 Prasanth Ullattil <pullatti@trolltech.com>
Reviewed by Simon.
Fix WebKit compilation with VC2008SP1
Apply the TR1 workaround for JavaScriptCore, too.
* JavaScriptCore.pro:
2008-10-08 Prasanth Ullattil <pullatti@trolltech.com>
Reviewed by Simon.
Fix compilation errors on VS2008 64Bit
* kjs/collector.cpp:
(JSC::currentThreadStackBase):
2008-10-08 André Pönitz <apoenitz@trolltech.com>
Reviewed by Simon.
Fix compilation with Qt namespaces.
* wtf/Threading.h:
2008-10-07 Sam Weinig <sam@webkit.org>
Roll out r37405.
2008-10-07 Oliver Hunt <oliver@apple.com>
Reviewed by Cameron Zwarich.
Switch CTI runtime calls to the fastcall calling convention
Basically this means that we get to store the argument for CTI
calls in the ECX register, which saves a register->memory write
and subsequent memory->register read.
This is a 1.7% progression in SunSpider and 2.4% on commandline
v8 tests on Windows
* VM/CTI.cpp:
(JSC::):
(JSC::CTI::privateCompilePutByIdTransition):
(JSC::CTI::privateCompilePatchGetArrayLength):
* VM/CTI.h:
* VM/Machine.h:
* masm/X86Assembler.h:
(JSC::X86Assembler::emitRestoreArgumentReference):
(JSC::X86Assembler::emitRestoreArgumentReferenceForTrampoline):
We need this to correctly reload ecx from inside certain property access
trampolines.
* wtf/Platform.h:
2008-10-07 Maciej Stachowiak <mjs@apple.com>
Reviewed by Mark Rowe.
- optimize away multiplication by constant 1.0
2.3% speedup on v8 RayTrace benchmark
Apparently it's not uncommon for JavaScript code to multiply by
constant 1.0 in the mistaken belief that this converts integer to
floating point and that there is any operational difference.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass): Optimize to_jsnumber for
case where parameter is already number.
(JSC::CTI::privateCompileSlowCases): ditto
* VM/Machine.cpp:
(JSC::Machine::privateExecute): ditto
* kjs/grammar.y:
(makeMultNode): Transform as follows:
+FOO * BAR ==> FOO * BAR
FOO * +BAR ==> FOO * BAR
FOO * 1 ==> +FOO
1 * FOO ==> +FOO
(makeDivNode): Transform as follows:
+FOO / BAR ==> FOO / BAR
FOO / +BAR ==> FOO / BAR
(makeSubNode): Transform as follows:
+FOO - BAR ==> FOO - BAR
FOO - +BAR ==> FOO - BAR
* kjs/nodes.h:
(JSC::ExpressionNode::stripUnaryPlus): Helper for above
grammar.y changes
(JSC::UnaryPlusNode::stripUnaryPlus): ditto
2008-10-07 Maciej Stachowiak <mjs@apple.com>
Reviewed by Oliver Hunt.
- make constant folding code more consistent
Added a makeSubNode to match add, mult and div; use the makeFooNode functions always,
instead of allocating nodes directly in other places in the grammar.
* kjs/grammar.y:
2008-10-07 Sam Weinig <sam@webkit.org>
Reviewed by Cameron Zwarich.
Move hasGetterSetterProperties flag from PropertyMap to StructureID.
* kjs/JSObject.cpp:
(JSC::JSObject::put):
(JSC::JSObject::defineGetter):
(JSC::JSObject::defineSetter):
* kjs/JSObject.h:
(JSC::JSObject::hasGetterSetterProperties):
(JSC::JSObject::getOwnPropertySlotForWrite):
(JSC::JSObject::getOwnPropertySlot):
* kjs/PropertyMap.h:
* kjs/StructureID.cpp:
(JSC::StructureID::StructureID):
(JSC::StructureID::addPropertyTransition):
(JSC::StructureID::toDictionaryTransition):
(JSC::StructureID::changePrototypeTransition):
(JSC::StructureID::getterSetterTransition):
* kjs/StructureID.h:
(JSC::StructureID::hasGetterSetterProperties):
(JSC::StructureID::setHasGetterSetterProperties):
2008-10-07 Sam Weinig <sam@webkit.org>
Reviewed by Cameron Zwarich.
Roll r37370 back in with bug fixes.
- PropertyMap::storageSize() should reflect the number of keys + deletedOffsets
and has nothing to do with the internal deletedSentinel count anymore.
2008-10-07 Gavin Barraclough <barraclough@apple.com>
Reviewed by Oliver Hunt.
Move callframe initialization into JIT code, again.
As a part of the restructuring the second result from functions is now
returned in edx, allowing the new value of 'r' to be returned via a
register, and stored to the stack from JIT code, too.
4.5% progression on v8-tests. (3% in their harness)
* VM/CTI.cpp:
(JSC::):
(JSC::CTI::emitCall):
(JSC::CTI::compileOpCall):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
(JSC::CTI::privateCompile):
* VM/CTI.h:
(JSC::CallRecord::CallRecord):
* VM/Machine.cpp:
(JSC::Machine::cti_op_call_JSFunction):
(JSC::Machine::cti_op_construct_JSConstruct):
(JSC::Machine::cti_op_resolve_func):
(JSC::Machine::cti_op_post_inc):
(JSC::Machine::cti_op_resolve_with_base):
(JSC::Machine::cti_op_post_dec):
* VM/Machine.h:
* kjs/JSFunction.h:
* kjs/ScopeChain.h:
2008-10-07 Mark Rowe <mrowe@apple.com>
Fix typo in method name.
* wrec/WREC.cpp:
* wrec/WREC.h:
2008-10-07 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Mark Rowe.
Roll out r37370.
2008-10-06 Sam Weinig <sam@webkit.org>
Reviewed by Cameron Zwarich.
Fix for https://bugs.webkit.org/show_bug.cgi?id=21415
Improve the division between PropertyStorageArray and PropertyMap
- Rework ProperyMap to store offsets in the value so that they don't
change when rehashing. This allows us not to have to keep the
PropertyStorageArray in sync and thus not have to pass it in.
- Rename PropertyMap::getOffset -> PropertyMap::get since put/remove
now also return offsets.
- A Vector of deleted offsets is now needed since the storage is out of
band.
1% win on SunSpider. Wash on V8 suite.
* JavaScriptCore.exp:
* VM/CTI.cpp:
(JSC::transitionWillNeedStorageRealloc):
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
Transition logic can be greatly simplified by the fact that
the storage capacity is always known, and is correct for the
inline case.
* kjs/JSObject.cpp:
(JSC::JSObject::put): Rename getOffset -> get.
(JSC::JSObject::deleteProperty): Ditto.
(JSC::JSObject::getPropertyAttributes): Ditto.
(JSC::JSObject::removeDirect): Use returned offset to
clear the value in the PropertyNameArray.
(JSC::JSObject::allocatePropertyStorage): Add assert.
* kjs/JSObject.h:
(JSC::JSObject::getDirect): Rename getOffset -> get
(JSC::JSObject::getDirectLocation): Rename getOffset -> get
(JSC::JSObject::putDirect): Use propertyStorageCapacity to determine whether
or not to resize. Also, since put now returns an offset (and thus
addPropertyTransition does also) setting of the PropertyStorageArray is
now done here.
(JSC::JSObject::transitionTo):
* kjs/PropertyMap.cpp:
(JSC::PropertyMap::checkConsistency): PropertyStorageArray is no longer
passed in.
(JSC::PropertyMap::operator=): Copy the delete offsets vector.
(JSC::PropertyMap::put): Instead of setting the PropertyNameArray
explicitly, return the offset where the value should go.
(JSC::PropertyMap::remove): Instead of removing from the PropertyNameArray
explicitly, return the offset where the value should be removed.
(JSC::PropertyMap::get): Switch to using the stored offset, instead
of the implicit one.
(JSC::PropertyMap::insert):
(JSC::PropertyMap::expand): This is never called when m_table is null,
so remove that branch and add it as an assertion.
(JSC::PropertyMap::createTable): Consistency checks no longer take
a PropertyNameArray.
(JSC::PropertyMap::rehash): No need to rehash the PropertyNameArray
now that it is completely out of band.
* kjs/PropertyMap.h:
(JSC::PropertyMapEntry::PropertyMapEntry): Store offset into PropertyNameArray.
(JSC::PropertyMap::get): Switch to using the stored offset, instead
of the implicit one.
* kjs/StructureID.cpp:
(JSC::StructureID::StructureID): Initialize the propertyStorageCapacity to
JSObject::inlineStorageCapacity.
(JSC::StructureID::growPropertyStorageCapacity): Grow the storage capacity as
described below.
(JSC::StructureID::addPropertyTransition): Copy the storage capacity.
(JSC::StructureID::toDictionaryTransition): Ditto.
(JSC::StructureID::changePrototypeTransition): Ditto.
(JSC::StructureID::getterSetterTransition): Ditto.
* kjs/StructureID.h:
(JSC::StructureID::propertyStorageCapacity): Add propertyStorageCapacity
which is the current capacity for the JSObjects PropertyStorageArray.
It starts at the JSObject::inlineStorageCapacity (currently 2), then
when it first needs to be resized moves to the JSObject::nonInlineBaseStorageCapacity
(currently 16), and after that doubles each time.
2008-10-06 Cameron Zwarich <zwarich@apple.com>
Reviewed by Oliver Hunt.
Bug 21396: Remove the OptionalCalleeActivation call frame slot
<https://bugs.webkit.org/show_bug.cgi?id=21396>
Remove the OptionalCalleeActivation call frame slot. We have to be
careful to store the activation object in a register, because objects
in the scope chain do not get marked.
This is a 0.3% speedup on both SunSpider and the V8 benchmark.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::CodeGenerator):
(JSC::CodeGenerator::emitReturn):
* VM/CodeGenerator.h:
* VM/Machine.cpp:
(JSC::Machine::dumpRegisters):
(JSC::Machine::unwindCallFrame):
(JSC::Machine::privateExecute):
(JSC::Machine::cti_op_call_JSFunction):
(JSC::Machine::cti_op_push_activation):
(JSC::Machine::cti_op_tear_off_activation):
(JSC::Machine::cti_op_construct_JSConstruct):
* VM/Machine.h:
(JSC::Machine::initializeCallFrame):
* VM/RegisterFile.h:
(JSC::RegisterFile::):
2008-10-06 Tony Chang <tony@chromium.org>
Reviewed by Alexey Proskuryakov.
Chromium doesn't use pthreads on windows, so make its use conditional.
Also convert a WORD to a DWORD to avoid a compiler warning. This
matches the other methods around it.
* wtf/ThreadingWin.cpp:
(WTF::wtfThreadEntryPoint):
(WTF::ThreadCondition::broadcast):
2008-10-06 Mark Mentovai <mark@moxienet.com>
Reviewed by Tim Hatcher.
Allow ENABLE_DASHBOARD_SUPPORT and ENABLE_MAC_JAVA_BRIDGE to be
disabled on the Mac.
https://bugs.webkit.org/show_bug.cgi?id=21333
* wtf/Platform.h:
2008-10-06 Steve Falkenburg <sfalken@apple.com>
https://bugs.webkit.org/show_bug.cgi?id=21416
Pass 0 for size to VirtualAlloc, as documented by MSDN.
Identified by Application Verifier.
Reviewed by Darin Adler.
* kjs/collector.cpp:
(KJS::freeBlock):
2008-10-06 Kevin McCullough <kmccullough@apple.com>
Reviewed by Tim Hatcheri and Oliver Hunt.
https://bugs.webkit.org/show_bug.cgi?id=21412
Bug 21412: Refactor user initiated profile count to be more stable
- Export UString::from for use with creating the profile title.
* JavaScriptCore.exp:
2008-10-06 Maciej Stachowiak <mjs@apple.com>
Not reviewed. Build fix.
- revert toBoolean changes (r37333 and r37335); need to make WebCore work with these
* API/JSValueRef.cpp:
(JSValueToBoolean):
* ChangeLog:
* JavaScriptCore.exp:
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
(JSC::Machine::cti_op_loop_if_true):
(JSC::Machine::cti_op_not):
(JSC::Machine::cti_op_jtrue):
* kjs/ArrayPrototype.cpp:
(JSC::arrayProtoFuncFilter):
(JSC::arrayProtoFuncEvery):
(JSC::arrayProtoFuncSome):
* kjs/BooleanConstructor.cpp:
(JSC::constructBoolean):
(JSC::callBooleanConstructor):
* kjs/GetterSetter.h:
* kjs/JSCell.h:
(JSC::JSValue::toBoolean):
* kjs/JSNumberCell.cpp:
(JSC::JSNumberCell::toBoolean):
* kjs/JSNumberCell.h:
* kjs/JSObject.cpp:
(JSC::JSObject::toBoolean):
* kjs/JSObject.h:
* kjs/JSString.cpp:
(JSC::JSString::toBoolean):
* kjs/JSString.h:
* kjs/JSValue.h:
* kjs/RegExpConstructor.cpp:
(JSC::setRegExpConstructorMultiline):
* kjs/RegExpObject.cpp:
(JSC::RegExpObject::match):
* kjs/RegExpPrototype.cpp:
(JSC::regExpProtoFuncToString):
2008-10-06 Maciej Stachowiak <mjs@apple.com>
Reviewed by Sam Weinig.
- optimize op_jtrue, op_loop_if_true and op_not in various ways
https://bugs.webkit.org/show_bug.cgi?id=21404
1) Make JSValue::toBoolean nonvirtual and completely inline by
making use of the StructureID type field.
2) Make JSValue::toBoolean not take an ExecState; doesn't need it.
3) Make op_not, op_loop_if_true and op_jtrue not read the
ExecState (toBoolean doesn't need it any more) and not check
exceptions (toBoolean can't throw).
* API/JSValueRef.cpp:
(JSValueToBoolean):
* JavaScriptCore.exp:
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
(JSC::Machine::cti_op_loop_if_true):
(JSC::Machine::cti_op_not):
(JSC::Machine::cti_op_jtrue):
* kjs/ArrayPrototype.cpp:
(JSC::arrayProtoFuncFilter):
(JSC::arrayProtoFuncEvery):
(JSC::arrayProtoFuncSome):
* kjs/BooleanConstructor.cpp:
(JSC::constructBoolean):
(JSC::callBooleanConstructor):
* kjs/GetterSetter.h:
* kjs/JSCell.h:
(JSC::JSValue::toBoolean):
* kjs/JSNumberCell.cpp:
* kjs/JSNumberCell.h:
(JSC::JSNumberCell::toBoolean):
* kjs/JSObject.cpp:
* kjs/JSObject.h:
(JSC::JSObject::toBoolean):
(JSC::JSCell::toBoolean):
* kjs/JSString.cpp:
* kjs/JSString.h:
(JSC::JSString::toBoolean):
* kjs/JSValue.h:
* kjs/RegExpConstructor.cpp:
(JSC::setRegExpConstructorMultiline):
* kjs/RegExpObject.cpp:
(JSC::RegExpObject::match):
* kjs/RegExpPrototype.cpp:
(JSC::regExpProtoFuncToString):
2008-10-06 Ariya Hidayat <ariya.hidayat@trolltech.com>
Reviewed by Simon.
Build fix for MinGW.
* JavaScriptCore.pri:
* kjs/DateMath.cpp:
(JSC::highResUpTime):
2008-10-05 Cameron Zwarich <zwarich@apple.com>
Reviewed by Oliver Hunt.
Remove ScopeNode::containsClosures() now that it is unused.
* kjs/nodes.h:
(JSC::ScopeNode::containsClosures):
2008-10-05 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich.
- fix releas-only test failures caused by the fix to bug 21375
* VM/Machine.cpp:
(JSC::Machine::unwindCallFrame): Update ExecState while unwinding call frames;
it now matters more to have a still-valid ExecState, since dynamicGlobalObject
will make use of the ExecState's scope chain.
* VM/Machine.h:
2008-10-05 Cameron Zwarich <zwarich@apple.com>
Reviewed by Oliver Hunt.
Bug 21364: Remove the branch in op_ret for OptionalCalleeActivation and OptionalCalleeArguments
<https://bugs.webkit.org/show_bug.cgi?id=21364>
Use information from the parser to detect whether an activation is
needed or 'arguments' is used, and emit explicit instructions to tear
them off before op_ret. This allows a branch to be removed from op_ret
and simplifies some other code. This does cause a small change in the
behaviour of 'f.arguments'; it is no longer live when 'arguments' is not
mentioned in the lexical scope of the function.
It should now be easy to remove the OptionaCalleeActivation slot in the
call frame, but this will be done in a later patch.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitReturn):
* VM/CodeGenerator.h:
* VM/Machine.cpp:
(JSC::Machine::unwindCallFrame):
(JSC::Machine::privateExecute):
(JSC::Machine::retrieveArguments):
(JSC::Machine::cti_op_create_arguments):
(JSC::Machine::cti_op_tear_off_activation):
(JSC::Machine::cti_op_tear_off_arguments):
* VM/Machine.h:
* VM/Opcode.h:
* kjs/Arguments.cpp:
(JSC::Arguments::mark):
* kjs/Arguments.h:
(JSC::Arguments::isTornOff):
(JSC::Arguments::Arguments):
(JSC::Arguments::copyRegisters):
(JSC::JSActivation::copyRegisters):
* kjs/JSActivation.cpp:
(JSC::JSActivation::argumentsGetter):
* kjs/JSActivation.h:
2008-10-05 Maciej Stachowiak <mjs@apple.com>
Reviewed by Oliver Hunt.
- fixed "REGRESSION (r37297): fast/js/deep-recursion-test takes too long and times out"
https://bugs.webkit.org/show_bug.cgi?id=21375
The problem is that dynamicGlobalObject had become O(N) in number
of call frames, but unwinding the stack for an exception called it
for every call frame, resulting in O(N^2) behavior for an
exception thrown from inside deep recursion.
Instead of doing it that way, stash the dynamic global object in JSGlobalData.
* JavaScriptCore.exp:
* VM/Machine.cpp:
(JSC::DynamicGlobalObjectScope::DynamicGlobalObjectScope): Helper class to temporarily
store and later restore a dynamicGlobalObject in JSGlobalData.
(JSC::DynamicGlobalObjectScope::~DynamicGlobalObjectScope):
(JSC::Machine::execute): In each version, establish a DynamicGlobalObjectScope.
For ProgramNode, always establish set new dynamicGlobalObject, for FunctionBody and Eval,
only if none is currently set.
* VM/Machine.h:
* kjs/ExecState.h:
* kjs/JSGlobalData.cpp:
(JSC::JSGlobalData::JSGlobalData): Ininitalize new dynamicGlobalObject field to 0.
* kjs/JSGlobalData.h:
* kjs/JSGlobalObject.h:
(JSC::ExecState::dynamicGlobalObject): Moved here from ExecState for benefit of inlining.
Return lexical global object if this is a globalExec(), otherwise look in JSGlobalData
for the one stashed there.
2008-10-05 Sam Weinig <sam@webkit.org>
Reviewed by Maciej Stachowiak.
Avoid an extra lookup when transitioning to an existing StructureID
by caching the offset of property that caused the transition.
1% win on V8 suite. Wash on SunSpider.
* kjs/PropertyMap.cpp:
(JSC::PropertyMap::put):
* kjs/PropertyMap.h:
* kjs/StructureID.cpp:
(JSC::StructureID::StructureID):
(JSC::StructureID::addPropertyTransition):
* kjs/StructureID.h:
(JSC::StructureID::setCachedTransistionOffset):
(JSC::StructureID::cachedTransistionOffset):
2008-10-05 Cameron Zwarich <zwarich@apple.com>
Reviewed by Maciej Stachowiak.
Bug 21364: Remove the branch in op_ret for OptionalCalleeActivation and OptionalCalleeArguments
<https://bugs.webkit.org/show_bug.cgi?id=21364>
This patch does not yet remove the branch, but it does a bit of refactoring
so that a CodeGenerator now knows whether the associated CodeBlock will need
a full scope before doing any code generation. This makes it possible to emit
explicit tear-off instructions before every op_ret.
* VM/CodeBlock.h:
(JSC::CodeBlock::CodeBlock):
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::generate):
(JSC::CodeGenerator::CodeGenerator):
(JSC::CodeGenerator::emitPushScope):
(JSC::CodeGenerator::emitPushNewScope):
* kjs/nodes.h:
(JSC::ScopeNode::needsActivation):
2008-10-05 Gavin Barraclough <barraclough@apple.com>
Reviewed by Cameron Zwarich.
Fix for bug #21387 - using SamplingTool with CTI.
(1) A repatch offset offset changes due to an additional instruction to update SamplingTool state.
(2) Fix an incusion order problem due to ExecState changes.
(3) Change to a MACHINE_SAMPLING macro, use of exec should now be accessing global data.
* VM/CTI.h:
(JSC::CTI::execute):
* VM/SamplingTool.h:
(JSC::SamplingTool::privateExecuteReturned):
* kjs/Shell.cpp:
2008-10-04 Mark Rowe <mrowe@apple.com>
Reviewed by Tim Hatcher.
Add a 'Check For Weak VTables' build phase to catch weak vtables as early as possible.
* JavaScriptCore.xcodeproj/project.pbxproj:
2008-10-04 Sam Weinig <sam@webkit.org>
Reviewed by Oliver Hunt.
Fix https://bugs.webkit.org/show_bug.cgi?id=21320
leaks of PropertyNameArrayData seen on buildbot
- Fix RefPtr cycle by making PropertyNameArrayData's pointer back
to the StructureID a weak pointer.
* kjs/PropertyNameArray.h:
(JSC::PropertyNameArrayData::setCachedStructureID):
(JSC::PropertyNameArrayData::cachedStructureID):
* kjs/StructureID.cpp:
(JSC::StructureID::getEnumerablePropertyNames):
(JSC::StructureID::clearEnumerationCache):
(JSC::StructureID::~StructureID):
2008-10-04 Darin Adler <darin@apple.com>
Reviewed by Cameron Zwarich.
- https://bugs.webkit.org/show_bug.cgi?id=21295
Bug 21295: Replace ExecState with a call frame Register pointer
10% faster on Richards; other v8 benchmarks faster too.
A wash on SunSpider.
This does the minimum necessary to get the speedup. Next step in
cleaning this up is to replace ExecState with a CallFrame class,
and be more judicious about when to pass a call frame and when
to pass a global data pointer, global object pointer, or perhaps
something else entirely.
* VM/CTI.cpp: Remove the debug-only check of the exception in
ctiVMThrowTrampoline -- already checked in the code the trampoline
jumps to, so not all that useful. Removed the exec argument from
ctiTrampoline. Removed emitDebugExceptionCheck -- no longer needed.
(JSC::CTI::emitCall): Removed code to set ExecState::m_callFrame.
(JSC::CTI::privateCompileMainPass): Removed code in catch to extract
the exception from ExecState::m_exception; instead, the code that
jumps into catch will make sure the exception is already in eax.
* VM/CTI.h: Removed exec from the ctiTrampoline. Also removed the
non-helpful "volatile". Temporarily left ARG_exec in as a synonym
for ARG_r; I'll change that on a future cleanup pass when introducing
more use of the CallFrame type.
(JSC::CTI::execute): Removed the ExecState* argument.
* VM/ExceptionHelpers.cpp:
(JSC::InterruptedExecutionError::InterruptedExecutionError): Take
JSGlobalData* instead of ExecState*.
(JSC::createInterruptedExecutionException): Ditto.
* VM/ExceptionHelpers.h: Ditto. Also removed an unneeded include.
* VM/Machine.cpp:
(JSC::slideRegisterWindowForCall): Removed the exec and
exceptionValue arguments. Changed to return 0 when there's a stack
overflow rather than using a separate exception argument to cut
down on memory accesses in the calling convention.
(JSC::Machine::unwindCallFrame): Removed the exec argument when
constructing a DebuggerCallFrame. Also removed code to set
ExecState::m_callFrame.
(JSC::Machine::throwException): Removed the exec argument when
construction a DebuggerCallFrame.
(JSC::Machine::execute): Updated to use the register instead of
ExecState and also removed various uses of ExecState.
(JSC::Machine::debug):
(JSC::Machine::privateExecute): Put globalData into a local
variable so it can be used throughout the interpreter. Changed
the VM_CHECK_EXCEPTION to get the exception in globalData instead
of through ExecState.
(JSC::Machine::retrieveLastCaller): Turn exec into a registers
pointer by calling registers() instead of by getting m_callFrame.
(JSC::Machine::callFrame): Ditto.
Tweaked exception macros. Made new versions for when you know
you have an exception. Get at global exception with ARG_globalData.
Got rid of the need to pass in the return value type.
(JSC::Machine::cti_op_add): Update to use new version of exception
macros.
(JSC::Machine::cti_op_pre_inc): Ditto.
(JSC::Machine::cti_timeout_check): Ditto.
(JSC::Machine::cti_op_instanceof): Ditto.
(JSC::Machine::cti_op_new_func): Ditto.
(JSC::Machine::cti_op_call_JSFunction): Optimized by using the
ARG values directly instead of through local variables -- this gets
rid of code that just shuffles things around in the stack frame.
Also get rid of ExecState and update for the new way exceptions are
handled in slideRegisterWindowForCall.
(JSC::Machine::cti_vm_compile): Update to make exec out of r since
they are both the same thing now.
(JSC::Machine::cti_op_call_NotJSFunction): Ditto.
(JSC::Machine::cti_op_init_arguments): Ditto.
(JSC::Machine::cti_op_resolve): Ditto.
(JSC::Machine::cti_op_construct_JSConstruct): Ditto.
(JSC::Machine::cti_op_construct_NotJSConstruct): Ditto.
(JSC::Machine::cti_op_resolve_func): Ditto.
(JSC::Machine::cti_op_put_by_val): Ditto.
(JSC::Machine::cti_op_put_by_val_array): Ditto.
(JSC::Machine::cti_op_resolve_skip): Ditto.
(JSC::Machine::cti_op_resolve_global): Ditto.
(JSC::Machine::cti_op_post_inc): Ditto.
(JSC::Machine::cti_op_resolve_with_base): Ditto.
(JSC::Machine::cti_op_post_dec): Ditto.
(JSC::Machine::cti_op_call_eval): Ditto.
(JSC::Machine::cti_op_throw): Ditto. Also rearranged to return
the exception value as the return value so it can be used by
op_catch.
(JSC::Machine::cti_op_push_scope): Ditto.
(JSC::Machine::cti_op_in): Ditto.
(JSC::Machine::cti_op_del_by_val): Ditto.
(JSC::Machine::cti_vm_throw): Ditto. Also rearranged to return
the exception value as the return value so it can be used by
op_catch.
* kjs/DebuggerCallFrame.cpp:
(JSC::DebuggerCallFrame::functionName): Pass globalData.
(JSC::DebuggerCallFrame::evaluate): Eliminated code to make a
new ExecState.
* kjs/DebuggerCallFrame.h: Removed ExecState argument from
constructor.
* kjs/ExecState.h: Eliminated all data members and made ExecState
inherit privately from Register instead. Also added a typedef to
the future name for this class, which is CallFrame. It's just a
Register* that knows it's a pointer at a call frame. The new class
can't be constructed or copied. Changed all functions to use
the this pointer instead of m_callFrame. Changed exception-related
functions to access an exception in JSGlobalData. Removed functions
used by CTI to pass the return address to the throw machinery --
this is now done directly with a global in the global data.
* kjs/FunctionPrototype.cpp:
(JSC::functionProtoFuncToString): Pass globalData instead of exec.
* kjs/InternalFunction.cpp:
(JSC::InternalFunction::name): Take globalData instead of exec.
* kjs/InternalFunction.h: Ditto.
* kjs/JSGlobalData.cpp: Initialize the new exception global to 0.
* kjs/JSGlobalData.h: Declare two new globals. One for the current
exception and another for the return address used by CTI to
implement the throw operation.
* kjs/JSGlobalObject.cpp:
(JSC::JSGlobalObject::init): Removed code to set up globalExec,
which is now the same thing as globalCallFrame.
(JSC::JSGlobalObject::reset): Get globalExec from our globalExec
function so we don't have to repeat the logic twice.
(JSC::JSGlobalObject::mark): Removed code to mark the exception;
the exception is now stored in JSGlobalData and marked there.
(JSC::JSGlobalObject::globalExec): Return a pointer to the end
of the global call frame.
* kjs/JSGlobalObject.h: Removed the globalExec data member.
* kjs/JSObject.cpp:
(JSC::JSObject::putDirectFunction): Pass globalData instead of exec.
* kjs/collector.cpp:
(JSC::Heap::collect): Mark the global exception.
* profiler/ProfileGenerator.cpp:
(JSC::ProfileGenerator::addParentForConsoleStart): Pass globalData
instead of exec to createCallIdentifier.
* profiler/Profiler.cpp:
(JSC::Profiler::willExecute): Pass globalData instead of exec to
createCallIdentifier.
(JSC::Profiler::didExecute): Ditto.
(JSC::Profiler::createCallIdentifier): Take globalData instead of
exec.
(JSC::createCallIdentifierFromFunctionImp): Ditto.
* profiler/Profiler.h: Change interface to take a JSGlobalData
instead of an ExecState.
2008-10-04 Cameron Zwarich <zwarich@apple.com>
Reviewed by Darin Adler.
Bug 21369: Add opcode documentation for all undocumented opcodes
<https://bugs.webkit.org/show_bug.cgi?id=21369>
This patch adds opcode documentation for all undocumented opcodes, and
it also renames op_init_arguments to op_create_arguments.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::CodeGenerator):
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
(JSC::Machine::cti_op_create_arguments):
* VM/Machine.h:
* VM/Opcode.h:
2008-10-03 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich.
- "this" object in methods called on primitives should be wrapper object
https://bugs.webkit.org/show_bug.cgi?id=21362
I changed things so that functions which use "this" do a fast
version of toThisObject conversion if needed. Currently we miss
the conversion entirely, at least for primitive types. Using
TypeInfo and the primitive check, I made the fast case bail out
pretty fast.
This is inexplicably an 1.007x SunSpider speedup (and a wash on V8 benchmarks).
Also renamed some opcodes for clarity:
init ==> enter
init_activation ==> enter_with_activation
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::generate):
(JSC::CodeGenerator::CodeGenerator):
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
(JSC::Machine::cti_op_convert_this):
* VM/Machine.h:
* VM/Opcode.h:
* kjs/JSActivation.cpp:
(JSC::JSActivation::JSActivation):
* kjs/JSActivation.h:
(JSC::JSActivation::createStructureID):
* kjs/JSCell.h:
(JSC::JSValue::needsThisConversion):
* kjs/JSGlobalData.cpp:
(JSC::JSGlobalData::JSGlobalData):
* kjs/JSGlobalData.h:
* kjs/JSNumberCell.h:
(JSC::JSNumberCell::createStructureID):
* kjs/JSStaticScopeObject.h:
(JSC::JSStaticScopeObject::JSStaticScopeObject):
(JSC::JSStaticScopeObject::createStructureID):
* kjs/JSString.h:
(JSC::JSString::createStructureID):
* kjs/JSValue.h:
* kjs/TypeInfo.h:
(JSC::TypeInfo::needsThisConversion):
* kjs/nodes.h:
(JSC::ScopeNode::usesThis):
2008-10-03 Cameron Zwarich <zwarich@apple.com>
Reviewed by Maciej Stachowiak.
Bug 21356: The size of the RegisterFile differs depending on 32-bit / 64-bit and Debug / Release
<https://bugs.webkit.org/show_bug.cgi?id=21356>
The RegisterFile decreases in size (measured in terms of numbers of
Registers) as the size of a Register increases. This causes
js1_5/Regress/regress-159334.js
to fail in 64-bit debug builds. This fix makes the RegisterFile on all
platforms the same size that it is in 32-bit Release builds.
* VM/RegisterFile.h:
(JSC::RegisterFile::RegisterFile):
2008-10-03 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich.
- Some code cleanup to how we handle code features.
1) Rename FeatureInfo typedef to CodeFeatures.
2) Rename NodeFeatureInfo template to NodeInfo.
3) Keep CodeFeature bitmask in ScopeNode instead of trying to break it out into individual bools.
4) Rename misleadingly named "needsClosure" method to "containsClosures", which better describes the meaning
of ClosureFeature.
5) Make setUsersArguments() not take an argument since it only goes one way.
* JavaScriptCore.exp:
* VM/CodeBlock.h:
(JSC::CodeBlock::CodeBlock):
* kjs/NodeInfo.h:
* kjs/Parser.cpp:
(JSC::Parser::didFinishParsing):
* kjs/Parser.h:
(JSC::Parser::parse):
* kjs/grammar.y:
* kjs/nodes.cpp:
(JSC::ScopeNode::ScopeNode):
(JSC::ProgramNode::ProgramNode):
(JSC::ProgramNode::create):
(JSC::EvalNode::EvalNode):
(JSC::EvalNode::create):
(JSC::FunctionBodyNode::FunctionBodyNode):
(JSC::FunctionBodyNode::create):
* kjs/nodes.h:
(JSC::ScopeNode::usesEval):
(JSC::ScopeNode::containsClosures):
(JSC::ScopeNode::usesArguments):
(JSC::ScopeNode::setUsesArguments):
2008-10-03 Cameron Zwarich <zwarich@apple.com>
Reviewed by Maciej Stachowiak.
Bug 21343: REGRESSSION (r37160): ecma_3/ExecutionContexts/10.1.3-1.js and js1_4/Functions/function-001.js fail on 64-bit
<https://bugs.webkit.org/show_bug.cgi?id=21343>
A fix was landed for this issue in r37253, and the ChangeLog assumes
that it is a compiler bug, but it turns out that it is a subtle issue
with mixing signed and unsigned 32-bit values in a 64-bit environment.
In order to properly fix this bug, we should convert our signed offsets
into the register file to use ptrdiff_t.
This may not be the only instance of this issue, but I will land this
fix first and look for more later.
* VM/Machine.cpp:
(JSC::Machine::getArgumentsData):
* VM/Machine.h:
* kjs/Arguments.cpp:
(JSC::Arguments::getOwnPropertySlot):
* kjs/Arguments.h:
(JSC::Arguments::init):
2008-10-03 Darin Adler <darin@apple.com>
* VM/CTI.cpp: Another Windows build fix. Change the args of ctiTrampoline.
* kjs/JSNumberCell.h: A build fix for newer versions of gcc. Added
declarations of JSGlobalData overloads of jsNumberCell.
2008-10-03 Darin Adler <darin@apple.com>
- try to fix Windows build
* kjs/ScopeChain.h: Add forward declaration of JSGlobalData.
2008-10-03 Darin Adler <darin@apple.com>
Reviewed by Geoff Garen.
- next step of https://bugs.webkit.org/show_bug.cgi?id=21295
Turn ExecState into a call frame pointer.
Remove m_globalObject and m_globalData from ExecState.
SunSpider says this is a wash (slightly faster but not statistically
significant); which is good enough since it's a preparation step and
not supposed to be a spedup.
* API/JSCallbackFunction.cpp:
(JSC::JSCallbackFunction::JSCallbackFunction):
* kjs/ArrayConstructor.cpp:
(JSC::ArrayConstructor::ArrayConstructor):
* kjs/BooleanConstructor.cpp:
(JSC::BooleanConstructor::BooleanConstructor):
* kjs/DateConstructor.cpp:
(JSC::DateConstructor::DateConstructor):
* kjs/ErrorConstructor.cpp:
(JSC::ErrorConstructor::ErrorConstructor):
* kjs/FunctionPrototype.cpp:
(JSC::FunctionPrototype::FunctionPrototype):
* kjs/JSFunction.cpp:
(JSC::JSFunction::JSFunction):
* kjs/NativeErrorConstructor.cpp:
(JSC::NativeErrorConstructor::NativeErrorConstructor):
* kjs/NumberConstructor.cpp:
(JSC::NumberConstructor::NumberConstructor):
* kjs/ObjectConstructor.cpp:
(JSC::ObjectConstructor::ObjectConstructor):
* kjs/PrototypeFunction.cpp:
(JSC::PrototypeFunction::PrototypeFunction):
* kjs/RegExpConstructor.cpp:
(JSC::RegExpConstructor::RegExpConstructor):
* kjs/StringConstructor.cpp:
(JSC::StringConstructor::StringConstructor):
Pass JSGlobalData* instead of ExecState* to the InternalFunction
constructor.
* API/OpaqueJSString.cpp: Added now-needed include.
* JavaScriptCore.exp: Updated.
* VM/CTI.cpp:
(JSC::CTI::emitSlowScriptCheck): Changed to use ARGS_globalData
instead of ARGS_exec.
* VM/CTI.h: Added a new argument to the CTI, the global data pointer.
While it's possible to get to the global data pointer using the
ExecState pointer, it's slow enough that it's better to just keep
it around in the CTI arguments.
* VM/CodeBlock.h: Moved the CodeType enum here from ExecState.h.
* VM/Machine.cpp:
(JSC::Machine::execute): Pass fewer arguments when constructing
ExecState, and pass the global data pointer when invoking CTI.
(JSC::Machine::firstCallFrame): Added. Used to get the dynamic global
object, which is in the scope chain of the first call frame.
(JSC::Machine::cti_op_add): Use globalData instead of exec when
possible, to keep fast cases fast, since it's now more expensive to
get to it through the exec pointer.
(JSC::Machine::cti_timeout_check): Ditto.
(JSC::Machine::cti_op_put_by_id_second): Ditto.
(JSC::Machine::cti_op_get_by_id_second): Ditto.
(JSC::Machine::cti_op_mul): Ditto.
(JSC::Machine::cti_vm_compile): Ditto.
(JSC::Machine::cti_op_get_by_val): Ditto.
(JSC::Machine::cti_op_sub): Ditto.
(JSC::Machine::cti_op_put_by_val): Ditto.
(JSC::Machine::cti_op_put_by_val_array): Ditto.
(JSC::Machine::cti_op_negate): Ditto.
(JSC::Machine::cti_op_div): Ditto.
(JSC::Machine::cti_op_pre_dec): Ditto.
(JSC::Machine::cti_op_post_inc): Ditto.
(JSC::Machine::cti_op_lshift): Ditto.
(JSC::Machine::cti_op_bitand): Ditto.
(JSC::Machine::cti_op_rshift): Ditto.
(JSC::Machine::cti_op_bitnot): Ditto.
(JSC::Machine::cti_op_mod): Ditto.
(JSC::Machine::cti_op_post_dec): Ditto.
(JSC::Machine::cti_op_urshift): Ditto.
(JSC::Machine::cti_op_bitxor): Ditto.
(JSC::Machine::cti_op_bitor): Ditto.
(JSC::Machine::cti_op_call_eval): Ditto.
(JSC::Machine::cti_op_throw): Ditto.
(JSC::Machine::cti_op_is_string): Ditto.
(JSC::Machine::cti_op_debug): Ditto.
(JSC::Machine::cti_vm_throw): Ditto.
* VM/Machine.h: Added firstCallFrame.
* kjs/DebuggerCallFrame.cpp:
(JSC::DebuggerCallFrame::evaluate): Pass fewer arguments when
constructing ExecState.
* kjs/ExecState.cpp: Deleted contents. Later we'll remove the
file altogether.
* kjs/ExecState.h: Removed m_globalObject and m_globalData.
Moved CodeType into another header.
(JSC::ExecState::ExecState): Take only a single argument, a
call frame pointer.
(JSC::ExecState::dynamicGlobalObject): Get the object from
the first call frame since it's no longer stored.
(JSC::ExecState::globalData): Get the global data from the
scope chain, since we no longer store a pointer to it here.
(JSC::ExecState::identifierTable): Ditto.
(JSC::ExecState::propertyNames): Ditto.
(JSC::ExecState::emptyList): Ditto.
(JSC::ExecState::lexer): Ditto.
(JSC::ExecState::parser): Ditto.
(JSC::ExecState::machine): Ditto.
(JSC::ExecState::arrayTable): Ditto.
(JSC::ExecState::dateTable): Ditto.
(JSC::ExecState::mathTable): Ditto.
(JSC::ExecState::numberTable): Ditto.
(JSC::ExecState::regExpTable): Ditto.
(JSC::ExecState::regExpConstructorTable): Ditto.
(JSC::ExecState::stringTable): Ditto.
(JSC::ExecState::heap): Ditto.
* kjs/FunctionConstructor.cpp:
(JSC::FunctionConstructor::FunctionConstructor): Pass
JSGlobalData* instead of ExecState* to the InternalFunction
constructor.
(JSC::constructFunction): Pass the global data pointer when
constructing a new scope chain.
* kjs/InternalFunction.cpp:
(JSC::InternalFunction::InternalFunction): Take a JSGlobalData*
instead of an ExecState*. Later we can change more places to
work this way -- it's more efficient to take the type you need
since the caller might already have it.
* kjs/InternalFunction.h: Ditto.
* kjs/JSCell.h:
(JSC::JSCell::operator new): Added an overload that takes a
JSGlobalData* so you can construct without an ExecState*.
* kjs/JSGlobalObject.cpp:
(JSC::JSGlobalObject::init): Moved creation of the global scope
chain in here, since it now requires a pointer to the global data.
Moved the initialization of the call frame in here since it requires
the global scope chain node. Removed the extra argument to ExecState
when creating the global ExecState*.
* kjs/JSGlobalObject.h: Removed initialization of globalScopeChain
and the call frame from the JSGlobalObjectData constructor. Added
a thisValue argument to the init function.
* kjs/JSNumberCell.cpp: Added versions of jsNumberCell that take
JSGlobalData* rather than ExecState*.
* kjs/JSNumberCell.h:
(JSC::JSNumberCell::operator new): Added a version that takes
JSGlobalData*.
(JSC::JSNumberCell::JSNumberCell): Ditto.
(JSC::jsNumber): Ditto.
* kjs/JSString.cpp:
(JSC::jsString): Ditto.
(JSC::jsSubstring): Ditto.
(JSC::jsOwnedString): Ditto.
* kjs/JSString.h:
(JSC::JSString::JSString): Changed to take JSGlobalData*.
(JSC::jsEmptyString): Added a version that takes JSGlobalData*.
(JSC::jsSingleCharacterString): Ditto.
(JSC::jsSingleCharacterSubstring): Ditto.
(JSC::jsNontrivialString): Ditto.
(JSC::JSString::getIndex): Ditto.
(JSC::jsString): Ditto.
(JSC::jsSubstring): Ditto.
(JSC::jsOwnedString): Ditto.
* kjs/ScopeChain.h: Added a globalData pointer to each node.
(JSC::ScopeChainNode::ScopeChainNode): Initialize the globalData
pointer.
(JSC::ScopeChainNode::push): Set the global data pointer in the
new node.
(JSC::ScopeChain::ScopeChain): Take a globalData argument.
* kjs/SmallStrings.cpp:
(JSC::SmallStrings::createEmptyString): Take JSGlobalData* instead of
ExecState*.
(JSC::SmallStrings::createSingleCharacterString): Ditto.
* kjs/SmallStrings.h:
(JSC::SmallStrings::emptyString): Ditto.
(JSC::SmallStrings::singleCharacterString): Ditto.
2008-10-03 Cameron Zwarich <zwarich@apple.com>
Reviewed by Geoff Garen.
Bug 21343: REGRESSSION (r37160): ecma_3/ExecutionContexts/10.1.3-1.js and js1_4/Functions/function-001.js fail on 64-bit
<https://bugs.webkit.org/show_bug.cgi?id=21343>
Add a workaround for a bug in GCC, which affects GCC 4.0, GCC 4.2, and
llvm-gcc 4.2. I put it in an #ifdef because it was a slight regression
on SunSpider in 32-bit, although that might be entirely random.
* kjs/Arguments.cpp:
(JSC::Arguments::getOwnPropertySlot):
2008-10-03 Darin Adler <darin@apple.com>
Rubber stamped by Alexey Proskuryakov.
* kjs/Shell.cpp: (main): Don't delete JSGlobalData. Later, we need to change
this tool to use public JavaScriptCore API instead.
2008-10-03 Darin Adler <darin@apple.com>
Suggested by Alexey Proskuryakov.
* kjs/JSGlobalData.cpp:
(JSC::JSGlobalData::~JSGlobalData): Remove call to heap.destroy() because
it's too late to ref the JSGlobalData object once it's already being
destroyed. In practice this is not a problem because WebCore's JSGlobalData
is never destroyed and JSGlobalContextRelease takes care of calling
heap.destroy() in advance.
2008-10-02 Oliver Hunt <oliver@apple.com>
Reviewed by Maciej Stachowiak.
Replace SSE3 check with an SSE2 check, and implement SSE2 check on windows.
5.6% win on SunSpider on windows.
* VM/CTI.cpp:
(JSC::isSSE2Present):
(JSC::CTI::compileBinaryArithOp):
(JSC::CTI::compileBinaryArithOpSlowCase):
2008-10-03 Maciej Stachowiak <mjs@apple.com>
Rubber stamped by Cameron Zwarich.
- fix mistaken change of | to || which caused a big perf regression on EarleyBoyer
* kjs/grammar.y:
2008-10-02 Darin Adler <darin@apple.com>
Reviewed by Geoff Garen.
- https://bugs.webkit.org/show_bug.cgi?id=21321
Bug 21321: speed up JavaScriptCore by inlining Heap in JSGlobalData
1.019x as fast on SunSpider.
* API/JSBase.cpp:
(JSEvaluateScript): Use heap. instead of heap-> to work with the heap.
(JSCheckScriptSyntax): Ditto.
(JSGarbageCollect): Ditto.
(JSReportExtraMemoryCost): Ditto.
* API/JSContextRef.cpp:
(JSGlobalContextRetain): Ditto.
(JSGlobalContextRelease): Destroy the heap with the destroy function instead
of the delete operator.
(JSContextGetGlobalObject): Use heap. instead of heap-> to work with the heap.
* API/JSObjectRef.cpp:
(JSObjectMake): Use heap. instead of heap-> to work with the heap.
(JSObjectMakeFunctionWithCallback): Ditto.
(JSObjectMakeConstructor): Ditto.
(JSObjectMakeFunction): Ditto.
(JSObjectMakeArray): Ditto.
(JSObjectMakeDate): Ditto.
(JSObjectMakeError): Ditto.
(JSObjectMakeRegExp): Ditto.
(JSObjectHasProperty): Ditto.
(JSObjectGetProperty): Ditto.
(JSObjectSetProperty): Ditto.
(JSObjectGetPropertyAtIndex): Ditto.
(JSObjectSetPropertyAtIndex): Ditto.
(JSObjectDeleteProperty): Ditto.
(JSObjectCallAsFunction): Ditto.
(JSObjectCallAsConstructor): Ditto.
(JSObjectCopyPropertyNames): Ditto.
(JSPropertyNameAccumulatorAddName): Ditto.
* API/JSValueRef.cpp:
(JSValueIsEqual): Ditto.
(JSValueIsInstanceOfConstructor): Ditto.
(JSValueMakeNumber): Ditto.
(JSValueMakeString): Ditto.
(JSValueToNumber): Ditto.
(JSValueToStringCopy): Ditto.
(JSValueToObject): Ditto.
(JSValueProtect): Ditto.
(JSValueUnprotect): Ditto.
* kjs/ExecState.h:
(JSC::ExecState::heap): Update to use the & operator.
* kjs/JSGlobalData.cpp:
(JSC::JSGlobalData::JSGlobalData): Update to initialize a heap member
instead of calling new to make a heap.
(JSC::JSGlobalData::~JSGlobalData): Destroy the heap with the destroy
function instead of the delete operator.
* kjs/JSGlobalData.h: Change from Heap* to a Heap.
* kjs/JSGlobalObject.cpp:
(JSC::JSGlobalObject::mark): Use the & operator here.
(JSC::JSGlobalObject::operator new): Use heap. instead of heap-> to work
with the heap.
2008-10-02 Cameron Zwarich <zwarich@apple.com>
Reviewed by Geoff Garen.
Bug 21317: Replace RegisterFile size and capacity information with Register pointers
<https://bugs.webkit.org/show_bug.cgi?id=21317>
This is a 2.3% speedup on the V8 DeltaBlue benchmark, a 3.3% speedup on
the V8 Raytrace benchmark, and a 1.0% speedup on SunSpider.
* VM/Machine.cpp:
(JSC::slideRegisterWindowForCall):
(JSC::Machine::callEval):
(JSC::Machine::execute):
(JSC::Machine::privateExecute):
(JSC::Machine::cti_op_call_JSFunction):
(JSC::Machine::cti_op_construct_JSConstruct):
* VM/RegisterFile.cpp:
(JSC::RegisterFile::~RegisterFile):
* VM/RegisterFile.h:
(JSC::RegisterFile::RegisterFile):
(JSC::RegisterFile::start):
(JSC::RegisterFile::end):
(JSC::RegisterFile::size):
(JSC::RegisterFile::shrink):
(JSC::RegisterFile::grow):
(JSC::RegisterFile::lastGlobal):
(JSC::RegisterFile::markGlobals):
(JSC::RegisterFile::markCallFrames):
* kjs/JSGlobalObject.cpp:
(JSC::JSGlobalObject::copyGlobalsTo):
2008-10-02 Cameron Zwarich <zwarich@apple.com>
Rubber-stamped by Darin Adler.
Change bitwise operations introduced in r37166 to boolean operations. We
only use bitwise operations over boolean operations for increasing
performance in extremely hot code, but that does not apply to anything
in the parser.
* kjs/grammar.y:
2008-10-02 Gavin Barraclough <barraclough@apple.com>
Reviewed by Darin Adler.
Fix for bug #21232 - should reset m_isPendingDash on flush,
and should allow '\-' as beginning or end of a range (though
not to specifiy a range itself).
* ChangeLog:
* wrec/CharacterClassConstructor.cpp:
(JSC::CharacterClassConstructor::put):
(JSC::CharacterClassConstructor::flush):
* wrec/CharacterClassConstructor.h:
(JSC::CharacterClassConstructor::flushBeforeEscapedHyphen):
* wrec/WREC.cpp:
(JSC::WRECGenerator::generateDisjunction):
(JSC::WRECParser::parseCharacterClass):
(JSC::WRECParser::parseDisjunction):
* wrec/WREC.h:
2008-10-02 Darin Adler <darin@apple.com>
Reviewed by Sam Weinig.
- remove the "static" from declarations in a header file, since we
don't want them to have internal linkage
* VM/Machine.h: Remove the static keyword from the constant and the
three inline functions that Geoff just moved here.
2008-10-02 Geoffrey Garen <ggaren@apple.com>
Reviewed by Sam Weinig.
Fixed https://bugs.webkit.org/show_bug.cgi?id=21283.
Profiler Crashes When Started
* VM/Machine.cpp:
* VM/Machine.h:
(JSC::makeHostCallFramePointer):
(JSC::isHostCallFrame):
(JSC::stripHostCallFrameBit): Moved some things to the header so
JSGlobalObject could use them.
* kjs/JSGlobalObject.h:
(JSC::JSGlobalObject::JSGlobalObjectData::JSGlobalObjectData): Call the
new makeHostCallFramePointer API, since 0 no longer indicates a host
call frame.
2008-10-02 Alexey Proskuryakov <ap@webkit.org>
Reviewed by Darin Adler.
https://bugs.webkit.org/show_bug.cgi?id=21304
Stop using a static wrapper map for WebCore JS bindings
* kjs/JSGlobalData.cpp:
(JSC::JSGlobalData::JSGlobalData):
(JSC::JSGlobalData::~JSGlobalData):
(JSC::JSGlobalData::ClientData::~ClientData):
* kjs/JSGlobalData.h:
Added a client data member to JSGlobalData. WebCore will use it to store bindings-related
global data.
* JavaScriptCore.exp: Export virtual ClientData destructor.
2008-10-02 Geoffrey Garen <ggaren@apple.com>
Not reviewed.
Try to fix Qt build.
* kjs/Error.h:
2008-10-01 Geoffrey Garen <ggaren@apple.com>
Reviewed by Darin Adler and Cameron Zwarich.
Preliminary step toward dynamic recompilation: Standardized and
simplified the parsing interface.
The main goal in this patch is to make it easy to ask for a duplicate
compilation, and get back a duplicate result -- same source URL, same
debugger / profiler ID, same toString behavior, etc.
The basic unit of compilation and evaluation is now SourceCode, which
encompasses a SourceProvider, a range in that provider, and a starting
line number.
A SourceProvider now encompasses a source URL, and *is* a source ID,
since a pointer is a unique identifier.
* API/JSBase.cpp:
(JSEvaluateScript):
(JSCheckScriptSyntax): Provide a SourceCode to the Interpreter, since
other APIs are no longer supported.
* VM/CodeBlock.h:
(JSC::EvalCodeCache::get): Provide a SourceCode to the Interpreter, since
other APIs are no longer supported.
(JSC::CodeBlock::CodeBlock): ASSERT something that used to be ASSERTed
by our caller -- this is a better bottleneck.
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::CodeGenerator): Updated for the fact that
FunctionBodyNode's parameters are no longer a WTF::Vector.
* kjs/Arguments.cpp:
(JSC::Arguments::Arguments): ditto
* kjs/DebuggerCallFrame.cpp:
(JSC::DebuggerCallFrame::evaluate): Provide a SourceCode to the Parser,
since other APIs are no longer supported.
* kjs/FunctionConstructor.cpp:
(JSC::constructFunction): Provide a SourceCode to the Parser, since
other APIs are no longer supported. Adopt FunctionBodyNode's new
"finishParsing" API.
* kjs/JSFunction.cpp:
(JSC::JSFunction::lengthGetter):
(JSC::JSFunction::getParameterName): Updated for the fact that
FunctionBodyNode's parameters are no longer a wtf::Vector.
* kjs/JSFunction.h: Nixed some cruft.
* kjs/JSGlobalObjectFunctions.cpp:
(JSC::globalFuncEval): Provide a SourceCode to the Parser, since
other APIs are no longer supported.
* kjs/Parser.cpp:
(JSC::Parser::parse): Require a SourceCode argument, instead of a bunch
of broken out parameters. Stop tracking sourceId as an integer, since we
use the SourceProvider pointer for this now. Don't clamp the
startingLineNumber, since SourceCode does that now.
* kjs/Parser.h:
(JSC::Parser::parse): Standardized the parsing interface to require a
SourceCode.
* kjs/Shell.cpp:
(functionRun):
(functionLoad):
(prettyPrintScript):
(runWithScripts):
(runInteractive): Provide a SourceCode to the Interpreter, since
other APIs are no longer supported.
* kjs/SourceProvider.h:
(JSC::SourceProvider::SourceProvider):
(JSC::SourceProvider::url):
(JSC::SourceProvider::asId):
(JSC::UStringSourceProvider::create):
(JSC::UStringSourceProvider::UStringSourceProvider): Added new
responsibilities described above.
* kjs/SourceRange.h:
(JSC::SourceCode::SourceCode):
(JSC::SourceCode::toString):
(JSC::SourceCode::provider):
(JSC::SourceCode::firstLine):
(JSC::SourceCode::data):
(JSC::SourceCode::length): Added new responsibilities described above.
Renamed SourceRange to SourceCode, based on review feedback. Added
a makeSource function for convenience.
* kjs/debugger.h: Provide a SourceCode to the client, since other APIs
are no longer supported.
* kjs/grammar.y: Provide startingLineNumber when creating a SourceCode.
* kjs/debugger.h: Treat sourceId as intptr_t to avoid loss of precision
on 64bit platforms.
* kjs/interpreter.cpp:
(JSC::Interpreter::checkSyntax):
(JSC::Interpreter::evaluate):
* kjs/interpreter.h: Require a SourceCode instead of broken out arguments.
* kjs/lexer.cpp:
(JSC::Lexer::setCode):
* kjs/lexer.h:
(JSC::Lexer::sourceRange): Fold together the SourceProvider and line number
into a SourceCode. Fixed a bug where the Lexer would accidentally keep
alive the last SourceProvider forever.
* kjs/nodes.cpp:
(JSC::ScopeNode::ScopeNode):
(JSC::ProgramNode::ProgramNode):
(JSC::ProgramNode::create):
(JSC::EvalNode::EvalNode):
(JSC::EvalNode::generateCode):
(JSC::EvalNode::create):
(JSC::FunctionBodyNode::FunctionBodyNode):
(JSC::FunctionBodyNode::finishParsing):
(JSC::FunctionBodyNode::create):
(JSC::FunctionBodyNode::generateCode):
(JSC::ProgramNode::generateCode):
(JSC::FunctionBodyNode::paramString):
* kjs/nodes.h:
(JSC::ScopeNode::):
(JSC::ScopeNode::sourceId):
(JSC::FunctionBodyNode::):
(JSC::FunctionBodyNode::parameterCount):
(JSC::FuncExprNode::):
(JSC::FuncDeclNode::): Store a SourceCode in all ScopeNodes, since
SourceCode is now responsible for tracking URL, ID, etc. Streamlined
some ad hoc FunctionBodyNode fixups into a "finishParsing" function, to
help make clear what you need to do in order to finish parsing a
FunctionBodyNode.
* wtf/Vector.h:
(WTF::::releaseBuffer): Don't ASSERT that releaseBuffer() is only called
when buffer is not 0, since FunctionBodyNode is more than happy
to get back a 0 buffer, and other functions like RefPtr::release() allow
for 0, too.
2008-10-01 Cameron Zwarich <zwarich@apple.com>
Reviewed by Maciej Stachowiak.
Bug 21289: REGRESSION (r37160): Inspector crashes on load
<https://bugs.webkit.org/show_bug.cgi?id=21289>
The code in Arguments::mark() in r37160 was wrong. It marks indices in
d->registers, but that makes no sense (they are local variables, not
arguments). It should mark those indices in d->registerArray instead.
This patch also changes Arguments::copyRegisters() to use d->numParameters
instead of recomputing it.
* kjs/Arguments.cpp:
(JSC::Arguments::mark):
* kjs/Arguments.h:
(JSC::Arguments::copyRegisters):
2008-09-30 Darin Adler <darin@apple.com>
Reviewed by Eric Seidel.
- https://bugs.webkit.org/show_bug.cgi?id=21214
work on getting rid of ExecState
Eliminate some unneeded uses of dynamicGlobalObject.
* API/JSClassRef.cpp:
(OpaqueJSClass::contextData): Changed to use a map in the global data instead
of on the global object. Also fixed to use only a single hash table lookup.
* API/JSObjectRef.cpp:
(JSObjectMakeConstructor): Use lexicalGlobalObject rather than dynamicGlobalObject
to get the object prototype.
* kjs/ArrayPrototype.cpp:
(JSC::arrayProtoFuncToString): Use arrayVisitedElements set in global data rather
than in the global object.
(JSC::arrayProtoFuncToLocaleString): Ditto.
(JSC::arrayProtoFuncJoin): Ditto.
* kjs/JSGlobalData.cpp:
(JSC::JSGlobalData::JSGlobalData): Don't initialize opaqueJSClassData, since
it's no longer a pointer.
(JSC::JSGlobalData::~JSGlobalData): We still need to delete all the values, but
we don't need to delete the map since it's no longer a pointer.
* kjs/JSGlobalData.h: Made opaqueJSClassData a map instead of a pointer to a map.
Also added arrayVisitedElements.
* kjs/JSGlobalObject.h: Removed arrayVisitedElements.
* kjs/Shell.cpp:
(functionRun): Use lexicalGlobalObject instead of dynamicGlobalObject.
(functionLoad): Ditto.
2008-10-01 Cameron Zwarich <zwarich@apple.com>
Not reviewed.
Speculative Windows build fix.
* kjs/grammar.y:
2008-10-01 Cameron Zwarich <zwarich@apple.com>
Reviewed by Darin Adler.
Bug 21123: using "arguments" in a function should not force creation of an activation object
<https://bugs.webkit.org/show_bug.cgi?id=21123>
Make the 'arguments' object not require a JSActivation. We store the
'arguments' object in the OptionalCalleeArguments call frame slot. We
need to be able to get the original 'arguments' object to tear it off
when returning from a function, but 'arguments' may be assigned to in a
number of ways.
Therefore, we use the OptionalCalleeArguments slot when we want to get
the original activation or we know that 'arguments' was not assigned a
different value. When 'arguments' may have been assigned a new value,
we use a new local variable that is initialized with 'arguments'. Since
a function parameter named 'arguments' may overwrite the value of
'arguments', we also need to be careful to look up 'arguments' in the
symbol table, so we get the parameter named 'arguments' instead of the
local variable that we have added for holding the 'arguments' object.
This is a 19.1% win on the V8 Raytrace benchmark using the SunSpider
harness, and a 20.7% win using the V8 harness. This amounts to a 6.5%
total speedup on the V8 benchmark suite using the V8 harness.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
* VM/CodeBlock.h:
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::CodeGenerator):
* VM/Machine.cpp:
(JSC::Machine::unwindCallFrame):
(JSC::Machine::privateExecute):
(JSC::Machine::retrieveArguments):
(JSC::Machine::cti_op_init_arguments):
(JSC::Machine::cti_op_ret_activation_arguments):
* VM/Machine.h:
* VM/RegisterFile.h:
(JSC::RegisterFile::):
* kjs/Arguments.cpp:
(JSC::Arguments::mark):
(JSC::Arguments::fillArgList):
(JSC::Arguments::getOwnPropertySlot):
(JSC::Arguments::put):
* kjs/Arguments.h:
(JSC::Arguments::setRegisters):
(JSC::Arguments::init):
(JSC::Arguments::Arguments):
(JSC::Arguments::copyRegisters):
(JSC::JSActivation::copyRegisters):
* kjs/JSActivation.cpp:
(JSC::JSActivation::argumentsGetter):
* kjs/JSActivation.h:
(JSC::JSActivation::JSActivationData::JSActivationData):
* kjs/grammar.y:
* kjs/nodes.h:
(JSC::ScopeNode::setUsesArguments):
* masm/X86Assembler.h:
(JSC::X86Assembler::):
(JSC::X86Assembler::orl_mr):
2008-10-01 Kevin McCullough <kmccullough@apple.com>
Rubberstamped by Geoff Garen.
Remove BreakpointCheckStatement because it's not used anymore.
No effect on sunspider or the jsc tests.
* kjs/nodes.cpp:
* kjs/nodes.h:
2008-09-30 Oliver Hunt <oliver@apple.com>
Reviewed by Geoff Garen.
Improve performance of CTI on windows.
Currently on platforms where the compiler doesn't allow us to safely
index relative to the address of a parameter we need to actually
provide a pointer to CTI runtime call arguments. This patch improves
performance in this case by making the CTI logic for restoring this
parameter much less conservative by only resetting it before we actually
make a call, rather than between each and every SF bytecode we generate
code for.
This results in a 3.6% progression on the v8 benchmark when compiled with MSVC.
* VM/CTI.cpp:
(JSC::CTI::emitCall):
(JSC::CTI::compileOpCall):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
(JSC::CTI::privateCompilePutByIdTransition):
* VM/CTI.h:
* masm/X86Assembler.h:
* wtf/Platform.h:
2008-09-30 Maciej Stachowiak <mjs@apple.com>
Reviewed by Oliver Hunt.
- track uses of "this", "with" and "catch" in the parser
Knowing this up front will be useful for future optimizations.
Perf and correctness remain the same.
* kjs/NodeInfo.h:
* kjs/grammar.y:
2008-09-30 Sam Weinig <sam@webkit.org>
Reviewed by Mark Rowe.
Add WebKitAvailability macros for JSObjectMakeArray, JSObjectMakeDate, JSObjectMakeError,
and JSObjectMakeRegExp
* API/JSObjectRef.h:
2008-09-30 Darin Adler <darin@apple.com>
Reviewed by Geoff Garen.
- https://bugs.webkit.org/show_bug.cgi?id=21214
work on getting rid of ExecState
Replaced the m_prev field of ExecState with a bit in the
call frame pointer to indicate "host" call frames.
* VM/Machine.cpp:
(JSC::makeHostCallFramePointer): Added. Sets low bit.
(JSC::isHostCallFrame): Added. Checks low bit.
(JSC::stripHostCallFrameBit): Added. Clears low bit.
(JSC::Machine::unwindCallFrame): Replaced null check that was
formerly used to detect host call frames with an isHostCallFrame check.
(JSC::Machine::execute): Pass in a host call frame pointer rather than
always passing 0 when starting execution from the host. This allows us
to follow the entire call frame pointer chain when desired, or to stop
at the host calls when that's desired.
(JSC::Machine::privateExecute): Replaced null check that was
formerly used to detect host call frames with an isHostCallFrame check.
(JSC::Machine::retrieveCaller): Ditto.
(JSC::Machine::retrieveLastCaller): Ditto.
(JSC::Machine::callFrame): Removed the code to walk up m_prev pointers
and replaced it with code that uses the caller pointer and uses the
stripHostCallFrameBit function.
* kjs/ExecState.cpp: Removed m_prev.
* kjs/ExecState.h: Ditto.
2008-09-30 Cameron Zwarich <zwarich@apple.com>
Reviewed by Geoff Garen.
Move all detection of 'arguments' in a lexical scope to the parser, in
preparation for fixing
Bug 21123: using "arguments" in a function should not force creation of an activation object
<https://bugs.webkit.org/show_bug.cgi?id=21123>
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::CodeGenerator):
* kjs/NodeInfo.h:
* kjs/grammar.y:
2008-09-30 Geoffrey Garen <ggaren@apple.com>
Not reviewed.
* kjs/Shell.cpp:
(runWithScripts): Fixed indentation.
2008-09-30 Mark Rowe <mrowe@apple.com>
Rubber-stamped by Sam Weinig.
Build fix. Move InternalFunction::classInfo implementation into the .cpp
file to prevent the vtable for InternalFunction being generated as a weak symbol.
Has no effect on SunSpider.
* kjs/InternalFunction.cpp:
(JSC::InternalFunction::classInfo):
* kjs/InternalFunction.h:
2008-09-29 Maciej Stachowiak <mjs@apple.com>
Reviewed by Darin Adler.
- optimize appending a number to a string
https://bugs.webkit.org/show_bug.cgi?id=21203
It's pretty common in real-world code (and on some of the v8
benchmarks) to append a number to a string, so I made this one of
the fast cases, and also added support to UString to do it
directly without allocating a temporary UString.
~1% speedup on v8 benchmark.
* VM/Machine.cpp:
(JSC::jsAddSlowCase): Make this NEVER_INLINE because somehow otherwise
the change is a regression.
(JSC::jsAdd): Handle number + string special case.
(JSC::Machine::cti_op_add): Integrate much of the logic of jsAdd to
avoid exception check in the str + str, num + num and str + num cases.
* kjs/ustring.cpp:
(JSC::expandedSize): Make this a non-member function, since it needs to be
called in non-member functions but not outside this file.
(JSC::expandCapacity): Ditto.
(JSC::UString::expandCapacity): Call the non-member version.
(JSC::createRep): Helper to make a rep from a char*.
(JSC::UString::UString): Use above helper.
(JSC::concatenate): Guts of concatenating constructor for cases where first
item is a UString::Rep, and second is a UChar* and length, or a char*.
(JSC::UString::append): Implement for cases where first item is a UString::Rep,
and second is an int or double. Sadly duplicates logic of UString::from(int)
and UString::from(double).
* kjs/ustring.h:
2008-09-29 Darin Adler <darin@apple.com>
Reviewed by Sam Weinig.
- https://bugs.webkit.org/show_bug.cgi?id=21214
work on getting rid of ExecState
* JavaScriptCore.exp: Updated since JSGlobalObject::init
no longer takes a parameter.
* VM/Machine.cpp:
(JSC::Machine::execute): Removed m_registerFile argument
for ExecState constructors.
* kjs/DebuggerCallFrame.cpp:
(JSC::DebuggerCallFrame::evaluate): Removed globalThisValue
argument for ExecState constructor.
* kjs/ExecState.cpp:
(JSC::ExecState::ExecState): Removed globalThisValue and
registerFile arguments to constructors.
* kjs/ExecState.h: Removed m_globalThisValue and
m_registerFile data members.
* kjs/JSGlobalObject.cpp:
(JSC::JSGlobalObject::init): Removed globalThisValue
argument for ExecState constructor.
* kjs/JSGlobalObject.h:
(JSC::JSGlobalObject::JSGlobalObject): Got rid of parameter
for the init function.
2008-09-29 Geoffrey Garen <ggaren@apple.com>
Rubber-stamped by Cameron Zwarich.
Fixed https://bugs.webkit.org/show_bug.cgi?id=21225
Machine::retrieveLastCaller should check for a NULL codeBlock
In order to crash, you would need to call retrieveCaller in a situation
where you had two host call frames in a row in the register file. I
don't know how to make that happen, or if it's even possible, so I don't
have a test case -- but better safe than sorry!
* VM/Machine.cpp:
(JSC::Machine::retrieveLastCaller):
2008-09-29 Geoffrey Garen <ggaren@apple.com>
Reviewed by Cameron Zwarich.
Store the callee ScopeChain, not the caller ScopeChain, in the call frame
header. Nix the "scopeChain" local variable and ExecState::m_scopeChain, and
access the callee ScopeChain through the call frame header instead.
Profit: call + return are simpler, because they don't have to update the
"scopeChain" local variable, or ExecState::m_scopeChain.
Because CTI keeps "r" in a register, reading the callee ScopeChain relative
to "r" can be very fast, in any cases we care to optimize.
0% speedup on empty function call benchmark. (5.5% speedup in bytecode.)
0% speedup on SunSpider. (7.5% speedup on controlflow-recursive.)
2% speedup on SunSpider --v8.
2% speedup on v8 benchmark.
* VM/CTI.cpp: Changed scope chain access to read the scope chain from
the call frame header. Sped up op_ret by changing it not to fuss with
the "scopeChain" local variable or ExecState::m_scopeChain.
* VM/CTI.h: Updated CTI trampolines not to take a ScopeChainNode*
argument, since that's stored in the call frame header now.
* VM/Machine.cpp: Access "scopeChain" and "codeBlock" through new helper
functions that read from the call frame header. Updated functions operating
on ExecState::m_callFrame to account for / take advantage of the fact that
Exec:m_callFrame is now never NULL.
Fixed a bug in op_construct, where it would use the caller's default
object prototype, rather than the callee's, when constructing a new object.
* VM/Machine.h: Made some helper functions available. Removed
ScopeChainNode* arguments to a lot of functions, since the ScopeChainNode*
is now stored in the call frame header.
* VM/RegisterFile.h: Renamed "CallerScopeChain" to "ScopeChain", since
that's what it is now.
* kjs/DebuggerCallFrame.cpp: Updated for change to ExecState signature.
* kjs/ExecState.cpp:
* kjs/ExecState.h: Nixed ExecState::m_callFrame, along with the unused
isGlobalObject function.
* kjs/JSGlobalObject.cpp:
* kjs/JSGlobalObject.h: Gave the global object a fake call frame in
which to store the global scope chain, since our code now assumes that
it can always read the scope chain out of the ExecState's call frame.
2008-09-29 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Sam Weinig.
Remove the isActivationObject() virtual method on JSObject and use
StructureID information instead. This should be slightly faster, but
isActivationObject() is only used in assertions and unwinding the stack
for exceptions.
* VM/Machine.cpp:
(JSC::depth):
(JSC::Machine::unwindCallFrame):
(JSC::Machine::privateExecute):
(JSC::Machine::cti_op_ret_activation):
* kjs/JSActivation.cpp:
* kjs/JSActivation.h:
* kjs/JSObject.h:
2008-09-29 Peter Gal <galpeter@inf.u-szeged.hu>
Reviewed and tweaked by Darin Adler.
Fix build for non-all-in-one platforms.
* kjs/StringPrototype.cpp: Added missing ASCIICType.h include.
2008-09-29 Bradley T. Hughes <bradley.hughes@nokia.com>
Reviewed by Simon Hausmann.
Fix compilation with icpc
* wtf/HashSet.h:
(WTF::::find):
(WTF::::contains):
2008-09-29 Thiago Macieira <thiago.macieira@nokia.com>
Reviewed by Simon Hausmann.
Changed copyright from Trolltech ASA to Nokia.
Nokia acquired Trolltech ASA, assets were transferred on September 26th 2008.
* wtf/qt/MainThreadQt.cpp:
2008-09-29 Simon Hausmann <hausmann@webkit.org>
Reviewed by Lars Knoll.
Don't accidentially install libJavaScriptCore.a for the build inside
Qt.
* JavaScriptCore.pro:
2008-09-28 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Bug 21200: Allow direct access to 'arguments' without using op_resolve
<https://bugs.webkit.org/show_bug.cgi?id=21200>
Allow fast access to the 'arguments' object by adding an extra slot to
the callframe to store it.
This is a 3.0% speedup on the V8 Raytrace benchmark.
* JavaScriptCore.exp:
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::CodeGenerator):
(JSC::CodeGenerator::registerFor):
* VM/CodeGenerator.h:
(JSC::CodeGenerator::registerFor):
* VM/Machine.cpp:
(JSC::Machine::initializeCallFrame):
(JSC::Machine::dumpRegisters):
(JSC::Machine::privateExecute):
(JSC::Machine::retrieveArguments):
(JSC::Machine::cti_op_call_JSFunction):
(JSC::Machine::cti_op_create_arguments):
(JSC::Machine::cti_op_construct_JSConstruct):
* VM/Machine.h:
* VM/Opcode.h:
* VM/RegisterFile.h:
(JSC::RegisterFile::):
* kjs/JSActivation.cpp:
(JSC::JSActivation::mark):
(JSC::JSActivation::argumentsGetter):
* kjs/JSActivation.h:
(JSC::JSActivation::JSActivationData::JSActivationData):
* kjs/NodeInfo.h:
* kjs/Parser.cpp:
(JSC::Parser::didFinishParsing):
* kjs/Parser.h:
(JSC::Parser::parse):
* kjs/grammar.y:
* kjs/nodes.cpp:
(JSC::ScopeNode::ScopeNode):
(JSC::ProgramNode::ProgramNode):
(JSC::ProgramNode::create):
(JSC::EvalNode::EvalNode):
(JSC::EvalNode::create):
(JSC::FunctionBodyNode::FunctionBodyNode):
(JSC::FunctionBodyNode::create):
* kjs/nodes.h:
(JSC::ScopeNode::usesArguments):
2008-09-28 Mark Rowe <mrowe@apple.com>
Reviewed by Sam Weinig.
Add an ASCII fast-path to toLowerCase and toUpperCase.
The fast path speeds up the common case of an ASCII-only string by up to 60% while adding a less than 5% penalty
to the less common non-ASCII case.
This also removes stringProtoFuncToLocaleLowerCase and stringProtoFuncToLocaleUpperCase, which were identical
to the non-locale variants of the functions. toLocaleLowerCase and toLocaleUpperCase now use the non-locale
variants of the functions directly.
* kjs/StringPrototype.cpp:
(JSC::stringProtoFuncToLowerCase):
(JSC::stringProtoFuncToUpperCase):
2008-09-28 Mark Rowe <mrowe@apple.com>
Reviewed by Cameron Zwarich.
Speed up parseInt and parseFloat.
Repeatedly indexing into a UString is slow, so retrieve a pointer into the underlying buffer once up front
and use that instead. This is a 7% win on a parseInt/parseFloat micro-benchmark.
* kjs/JSGlobalObjectFunctions.cpp:
(JSC::parseInt):
(JSC::parseFloat):
2008-09-28 Simon Hausmann <hausmann@webkit.org>
Reviewed by David Hyatt.
In Qt's initializeThreading re-use an existing thread identifier for the main
thread if it exists.
currentThread() implicitly creates new identifiers and it could be that
it is called before initializeThreading().
* wtf/ThreadingQt.cpp:
(WTF::initializeThreading):
2008-09-27 Keishi Hattori <casey.hattori@gmail.com>
Added Machine::retrieveCaller to the export list.
Reviewed by Kevin McCullough and Tim Hatcher.
* JavaScriptCore.exp: Added Machine::retrieveCaller.
2008-09-27 Anders Carlsson <andersca@apple.com>
Fix build.
* VM/CTI.cpp:
(JSC::):
2008-09-27 Geoffrey Garen <ggaren@apple.com>
Reviewed by Cameron Zwarich.
https://bugs.webkit.org/show_bug.cgi?id=21175
Store the callee CodeBlock, not the caller CodeBlock, in the call frame
header. Nix the "codeBlock" local variable, and access the callee
CodeBlock through the call frame header instead.
Profit: call + return are simpler, because they don't have to update the
"codeBlock" local variable.
Because CTI keeps "r" in a register, reading the callee CodeBlock relative
to "r" can be very fast, in any cases we care to optimize. Presently,
no such cases seem important.
Also, stop writing "dst" to the call frame header. CTI doesn't use it.
21.6% speedup on empty function call benchmark.
3.8% speedup on SunSpider --v8.
2.1% speedup on v8 benchmark.
0.7% speedup on SunSpider (6% speedup on controlflow-recursive).
Small regression in bytecode, because currently every op_ret reads the
callee CodeBlock to check needsFullScopeChain, and bytecode does not
keep "r" in a register. On-balance, this is probably OK, since CTI is
our high-performance execution model. Also, this should go away once
we make needsFullScopeChain statically determinable at parse time.
* VM/CTI.cpp:
(JSC::CTI::compileOpCall): The speedup!
(JSC::CTI::privateCompileSlowCases): ditto
* VM/CTI.h:
(JSC::): Fixed up magic trampoline constants to account for the nixed
"codeBlock" argument.
(JSC::CTI::execute): Changed trampoline function not to take a "codeBlock"
argument, since codeBlock is now stored in the call frame header.
* VM/Machine.cpp: Read the callee CodeBlock from the register file. Use
a NULL CallerRegisters in the call frame header to signal a built-in
caller, since CodeBlock is now never NULL.
* VM/Machine.h: Made some stand-alone functions Machine member functions
so they could call the private codeBlock() accessor in the Register
class, of which Machine is a friend. Renamed "CallerCodeBlock" to
"CodeBlock", since it's no longer the caller's CodeBlock.
* VM/RegisterFile.h: Marked some methods const to accommodate a
const RegisterFile* being passed around in Machine.cpp.
2008-09-26 Jan Michael Alonzo <jmalonzo@webkit.org>
Gtk build fix. Not reviewed.
Narrow-down the target of the JavaScriptCore .lut.h generator so
it won't try to create the WebCore .lut.hs.
* GNUmakefile.am:
2008-09-26 Matt Lilek <webkit@mattlilek.com>
Reviewed by Tim Hatcher.
Update FEATURE_DEFINES after ENABLE_CROSS_DOCUMENT_MESSAGING was removed.
* Configurations/JavaScriptCore.xcconfig:
2008-09-26 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Rubber-stamped by Anders Carlson.
Change the name 'sc' to 'scopeChainNode' in a few places.
* kjs/nodes.cpp:
(JSC::EvalNode::generateCode):
(JSC::FunctionBodyNode::generateCode):
(JSC::ProgramNode::generateCode):
2008-09-26 Sam Weinig <sam@webkit.org>
Reviewed by Darin Adler.
Patch for https://bugs.webkit.org/show_bug.cgi?id=21152
Speedup static property get/put
Convert getting/setting static property values to use static functions
instead of storing an integer and switching in getValueProperty/putValueProperty.
* kjs/JSObject.cpp:
(JSC::JSObject::deleteProperty):
(JSC::JSObject::getPropertyAttributes):
* kjs/MathObject.cpp:
(JSC::MathObject::getOwnPropertySlot):
* kjs/NumberConstructor.cpp:
(JSC::numberConstructorNaNValue):
(JSC::numberConstructorNegInfinity):
(JSC::numberConstructorPosInfinity):
(JSC::numberConstructorMaxValue):
(JSC::numberConstructorMinValue):
* kjs/PropertySlot.h:
(JSC::PropertySlot::):
* kjs/RegExpConstructor.cpp:
(JSC::regExpConstructorDollar1):
(JSC::regExpConstructorDollar2):
(JSC::regExpConstructorDollar3):
(JSC::regExpConstructorDollar4):
(JSC::regExpConstructorDollar5):
(JSC::regExpConstructorDollar6):
(JSC::regExpConstructorDollar7):
(JSC::regExpConstructorDollar8):
(JSC::regExpConstructorDollar9):
(JSC::regExpConstructorInput):
(JSC::regExpConstructorMultiline):
(JSC::regExpConstructorLastMatch):
(JSC::regExpConstructorLastParen):
(JSC::regExpConstructorLeftContext):
(JSC::regExpConstructorRightContext):
(JSC::setRegExpConstructorInput):
(JSC::setRegExpConstructorMultiline):
(JSC::RegExpConstructor::setInput):
(JSC::RegExpConstructor::setMultiline):
(JSC::RegExpConstructor::multiline):
* kjs/RegExpConstructor.h:
* kjs/RegExpObject.cpp:
(JSC::regExpObjectGlobal):
(JSC::regExpObjectIgnoreCase):
(JSC::regExpObjectMultiline):
(JSC::regExpObjectSource):
(JSC::regExpObjectLastIndex):
(JSC::setRegExpObjectLastIndex):
* kjs/RegExpObject.h:
(JSC::RegExpObject::setLastIndex):
(JSC::RegExpObject::lastIndex):
(JSC::RegExpObject::RegExpObjectData::RegExpObjectData):
* kjs/StructureID.cpp:
(JSC::StructureID::getEnumerablePropertyNames):
* kjs/create_hash_table:
* kjs/lexer.cpp:
(JSC::Lexer::lex):
* kjs/lookup.cpp:
(JSC::HashTable::createTable):
(JSC::HashTable::deleteTable):
(JSC::setUpStaticFunctionSlot):
* kjs/lookup.h:
(JSC::HashEntry::initialize):
(JSC::HashEntry::setKey):
(JSC::HashEntry::key):
(JSC::HashEntry::attributes):
(JSC::HashEntry::function):
(JSC::HashEntry::functionLength):
(JSC::HashEntry::propertyGetter):
(JSC::HashEntry::propertyPutter):
(JSC::HashEntry::lexerValue):
(JSC::HashEntry::):
(JSC::HashTable::entry):
(JSC::getStaticPropertySlot):
(JSC::getStaticValueSlot):
(JSC::lookupPut):
2008-09-26 Gavin Barraclough <barraclough@apple.com>
Reviewed by Maciej Stachowiak & Oliver Hunt.
Add support for reusing temporary JSNumberCells. This change is based on the observation
that if the result of certain operations is a JSNumberCell and is consumed by a subsequent
operation that would produce a JSNumberCell, we can reuse the object rather than allocating
a fresh one. E.g. given the expression ((a * b) * c), we can statically determine that
(a * b) will have a numeric result (or else it will have thrown an exception), so the result
will either be a JSNumberCell or a JSImmediate.
This patch changes three areas of JSC:
* The AST now tracks type information about the result of each node.
* This information is consumed in bytecode compilation, and certain bytecode operations
now carry the statically determined type information about their operands.
* CTI uses the information in a number of fashions:
* Where an operand to certain arithmetic operations is reusable, it will plant code
to try to perform the operation in JIT code & reuse the cell, where appropriate.
* Where it can be statically determined that an operand can only be numeric (typically
the result of another arithmetic operation) the code will not redundantly check that
the JSCell is a JSNumberCell.
* Where either of the operands to an add are non-numeric do not plant an optimized
arithmetic code path, just call straight out to the C function.
+6% Sunspider (10% progression on 3D, 16% progression on math, 60% progression on access-nbody),
+1% v8-tests (improvements in raytrace & crypto)
* VM/CTI.cpp: Add optimized code generation with reuse of temporary JSNumberCells.
* VM/CTI.h:
* kjs/JSNumberCell.h:
* masm/X86Assembler.h:
* VM/CodeBlock.cpp: Add type information to specific bytecodes.
* VM/CodeGenerator.cpp:
* VM/CodeGenerator.h:
* VM/Machine.cpp:
* kjs/nodes.cpp: Track static type information for nodes.
* kjs/nodes.h:
* kjs/ResultDescriptor.h: (Added)
* JavaScriptCore.xcodeproj/project.pbxproj:
2008-09-26 Yichao Yin <yichao.yin@torchmobile.com.cn>
Reviewed by George Staikos, Maciej Stachowiak.
Add utility functions needed for upcoming WML code.
* wtf/ASCIICType.h:
(WTF::isASCIIPrintable):
2008-09-26 Geoffrey Garen <ggaren@apple.com>
Reviewed by Darin Adler.
Reverted the part of r36614 that used static data because static data
is not thread-safe.
2008-09-26 Geoffrey Garen <ggaren@apple.com>
Reviewed by Maciej Stachowiak.
Removed dynamic check for whether the callee needs an activation object.
Replaced with callee code to create the activation object.
0.5% speedup on SunSpider.
No change on v8 benchmark. (Might be a speedup, but it's in range of the
variance.)
0.7% speedup on v8 benchmark in bytecode.
1.3% speedup on empty call benchmark in bytecode.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass): Added support for op_init_activation,
the new opcode that specifies that the callee's initialization should
create an activation object.
(JSC::CTI::privateCompile): Removed previous code that did a similar
thing in an ad-hoc way.
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump): Added a case for dumping op_init_activation.
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::generate): Added fixup code to change op_init to
op_init_activation if necessary. (With a better parser, we would know
which to use from the beginning.)
* VM/Instruction.h:
(JSC::Instruction::Instruction):
(WTF::): Faster traits for the instruction vector. An earlier version
of this patch relied on inserting at the beginning of the vector, and
depended on this change for speed.
* VM/Machine.cpp:
(JSC::Machine::execute): Removed clients of setScopeChain, the old
abstraction for dynamically checking for whether an activation object
needed to be created.
(JSC::Machine::privateExecute): ditto
(JSC::Machine::cti_op_push_activation): Renamed this function from
cti_vm_updateScopeChain, and made it faster by removing the call to
setScopeChain.
* VM/Machine.h:
* VM/Opcode.h: Declared op_init_activation.
2008-09-24 Geoffrey Garen <ggaren@apple.com>
Reviewed by Maciej Stachowiak.
Move most of the return code back into the callee, now that the callee
doesn't have to calculate anything dynamically.
11.5% speedup on empty function call benchmark.
SunSpider says 0.3% faster. SunSpider --v8 says no change.
* VM/CTI.cpp:
(JSC::CTI::compileOpCall):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
2008-09-24 Sam Weinig <sam@webkit.org>
Reviewed by Maciej Stachowiak.
Remove staticFunctionGetter. There is only one remaining user of
staticFunctionGetter and it can be converted to use setUpStaticFunctionSlot.
* JavaScriptCore.exp:
* kjs/lookup.cpp:
* kjs/lookup.h:
2008-09-24 Maciej Stachowiak <mjs@apple.com>
Reviewed by Oliver Hunt.
- inline JIT fast case of op_neq
- remove extra level of function call indirection from slow cases of eq and neq
1% speedup on Richards
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
(JSC::Machine::cti_op_eq):
(JSC::Machine::cti_op_neq):
* kjs/operations.cpp:
(JSC::equal):
(JSC::equalSlowCase):
* kjs/operations.h:
(JSC::equalSlowCaseInline):
2008-09-24 Sam Weinig <sam@webkit.org>
Reviewed by Darin Adler.
Fix for https://bugs.webkit.org/show_bug.cgi?id=21080
<rdar://problem/6243534>
Crash below Function.apply when using a runtime array as the argument list
Test: plugins/bindings-array-apply-crash.html
* kjs/FunctionPrototype.cpp:
(JSC::functionProtoFuncApply): Revert to the slow case if the object inherits from
JSArray (via ClassInfo) but is not a JSArray.
2008-09-24 Kevin McCullough <kmccullough@apple.com>
Style change.
* kjs/nodes.cpp:
(JSC::statementListEmitCode):
2008-09-24 Kevin McCullough <kmccullough@apple.com>
Reviewed by Geoff.
Bug 21031: Breakpoints in the condition of loops only breaks the first
time
- Now when setting breakpoints in the condition of a loop (for, while,
for in, and do while) will successfully break each time throught the
loop.
- For 'for' loops we need a little more complicated behavior that cannot
be accomplished without some more significant changes:
https://bugs.webkit.org/show_bug.cgi?id=21073
* kjs/nodes.cpp:
(JSC::statementListEmitCode): We don't want to blindly emit a debug hook
at the first line of loops, instead let the loop emit the debug hooks.
(JSC::DoWhileNode::emitCode):
(JSC::WhileNode::emitCode):
(JSC::ForNode::emitCode):
(JSC::ForInNode::emitCode):
* kjs/nodes.h:
(JSC::StatementNode::):
(JSC::DoWhileNode::):
(JSC::WhileNode::):
(JSC::ForInNode::):
2008-09-24 Geoffrey Garen <ggaren@apple.com>
Reviewed by Darin Adler.
Fixed <rdar://problem/5605532> Need a SPI for telling JS the size of
the objects it retains
* API/tests/testapi.c: Test the new SPI a little.
* API/JSSPI.cpp: Add the new SPI.
* API/JSSPI.h: Add the new SPI.
* JavaScriptCore.exp: Add the new SPI.
* JavaScriptCore.xcodeproj/project.pbxproj: Add the new SPI.
2008-09-24 Geoffrey Garen <ggaren@apple.com>
Reviewed by Darin Adler.
* API/JSBase.h: Filled in some missing function names.
2008-09-24 Geoffrey Garen <ggaren@apple.com>
Reviewed by Cameron Zwarich.
Fixed https://bugs.webkit.org/show_bug.cgi?id=21057
Crash in RegisterID::deref() running fast/canvas/canvas-putImageData.html
* VM/CodeGenerator.h: Changed declaration order to ensure the
m_lastConstant, which is a RefPtr that points into m_calleeRegisters,
has its destructor called before the destructor for m_calleeRegisters.
2008-09-24 Darin Adler <darin@apple.com>
Reviewed by Sam Weinig.
- https://bugs.webkit.org/show_bug.cgi?id=21047
speed up ret_activation with inlining
About 1% on v8-raytrace.
* JavaScriptCore.exp: Removed JSVariableObject::setRegisters.
* kjs/JSActivation.cpp: Moved copyRegisters to the header to make it inline.
* kjs/JSActivation.h:
(JSC::JSActivation::copyRegisters): Moved here. Also removed the registerArraySize
argument to setRegisters, since the object doesn't need to store the number of
registers.
* kjs/JSGlobalObject.cpp:
(JSC::JSGlobalObject::reset): Removed unnecessary clearing left over from when we
used this on objects that weren't brand new. These days, this function is really
just part of the constructor.
* kjs/JSGlobalObject.h: Added registerArraySize to JSGlobalObjectData, since
JSVariableObjectData no longer needs it. Added a setRegisters override here
that handles storing the size.
* kjs/JSStaticScopeObject.h: Removed code to set registerArraySize, since it
no longer exists.
* kjs/JSVariableObject.cpp: Moved copyRegisterArray and setRegisters to the
header to make them inline.
* kjs/JSVariableObject.h: Removed registerArraySize from JSVariableObjectData,
since it was only used for the global object.
(JSC::JSVariableObject::copyRegisterArray): Moved here ot make it inline.
(JSC::JSVariableObject::setRegisters): Moved here to make it inline. Also
removed the code to set registerArraySize and changed an if statement into
an assert to save an unnnecessary branch.
2008-09-24 Maciej Stachowiak <mjs@apple.com>
Reviewed by Oliver Hunt.
- inline PropertyMap::getOffset to speed up polymorphic lookups
~1.5% speedup on v8 benchmark
no effect on SunSpider
* JavaScriptCore.exp:
* kjs/PropertyMap.cpp:
* kjs/PropertyMap.h:
(JSC::PropertyMap::getOffset):
2008-09-24 Jan Michael Alonzo <jmalonzo@webkit.org>
Reviewed by Alp Toker.
https://bugs.webkit.org/show_bug.cgi?id=20992
Build fails on GTK+ Mac OS
* wtf/ThreadingGtk.cpp: Remove platform ifdef as suggested by
Richard Hult.
(WTF::initializeThreading):
2008-09-23 Oliver Hunt <oliver@apple.com>
Reviewed by Maciej Stachowiak.
Bug 19968: Slow Script at www.huffingtonpost.com
<https://bugs.webkit.org/show_bug.cgi?id=19968>
Finally found the cause of this accursed issue. It is triggered
by synchronous creation of a new global object from JS. The new
global object resets the timer state in this execution group's
Machine, taking timerCheckCount to 0. Then when JS returns the
timerCheckCount is decremented making it non-zero. The next time
we execute JS we will start the timeout counter, however the non-zero
timeoutCheckCount means we don't reset the timer information. This
means that the timeout check is now checking the cumulative time
since the creation of the global object rather than the time since
JS was last entered. At this point the slow script dialog is guaranteed
to eventually be displayed incorrectly unless a page is loaded
asynchronously (which will reset everything into a sane state).
The fix for this is rather trivial -- the JSGlobalObject constructor
should not be resetting the machine timer state.
* VM/Machine.cpp:
(JSC::Machine::Machine):
Now that we can't rely on the GlobalObject initialising the timeout
state, we do it in the Machine constructor.
* VM/Machine.h:
(JSC::Machine::stopTimeoutCheck):
Add assertions to guard against this happening.
* kjs/JSGlobalObject.cpp:
(JSC::JSGlobalObject::init):
Don't reset the timeout state.
2008-09-23 Geoffrey Garen <ggaren@apple.com>
Reviewed by Oliver Hunt.
Fixed https://bugs.webkit.org/show_bug.cgi?id=21038 | <rdar://problem/6240812>
Uncaught exceptions in regex replace callbacks crash webkit
This was a combination of two problems:
(1) the replace function would continue execution after an exception
had been thrown.
(2) In some cases, the Machine would return 0 in the case of an exception,
despite the fact that a few clients dereference the Machine's return
value without first checking for an exception.
* VM/Machine.cpp:
(JSC::Machine::execute):
^ Return jsNull() instead of 0 in the case of an exception, since some
clients depend on using our return value.
^ ASSERT that execution does not continue after an exception has been
thrown, to help catch problems like this in the future.
* kjs/StringPrototype.cpp:
(JSC::stringProtoFuncReplace):
^ Stop execution if an exception has been thrown.
2008-09-23 Geoffrey Garen <ggaren@apple.com>
Try to fix the windows build.
* VM/CTI.cpp:
(JSC::CTI::compileOpCall):
(JSC::CTI::privateCompileMainPass):
2008-09-23 Alp Toker <alp@nuanti.com>
Build fix.
* VM/CTI.h:
2008-09-23 Geoffrey Garen <ggaren@apple.com>
Reviewed by Darin Adler.
* wtf/Platform.h: Removed duplicate #if.
2008-09-23 Geoffrey Garen <ggaren@apple.com>
Reviewed by Darin Adler.
Changed the layout of the call frame from
{ header, parameters, locals | constants, temporaries }
to
{ parameters, header | locals, constants, temporaries }
This simplifies function entry+exit, and enables a number of future
optimizations.
13.5% speedup on empty call benchmark for bytecode; 23.6% speedup on
empty call benchmark for CTI.
SunSpider says no change. SunSpider --v8 says 1% faster.
* VM/CTI.cpp:
Added a bit of abstraction for calculating whether a register is a
constant, since this patch changes that calculation:
(JSC::CTI::isConstant):
(JSC::CTI::getConstant):
(JSC::CTI::emitGetArg):
(JSC::CTI::emitGetPutArg):
(JSC::CTI::getConstantImmediateNumericArg):
Updated for changes to callframe header location:
(JSC::CTI::emitPutToCallFrameHeader):
(JSC::CTI::emitGetFromCallFrameHeader):
(JSC::CTI::printOpcodeOperandTypes):
Renamed to spite Oliver:
(JSC::CTI::emitInitRegister):
Added an abstraction for emitting a call through a register, so that
calls through registers generate exception info, too:
(JSC::CTI::emitCall):
Updated to match the new callframe header layout, and to support calls
through registers, which have no destination address:
(JSC::CTI::compileOpCall):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
(JSC::CTI::privateCompile):
* VM/CTI.h:
More of the above:
(JSC::CallRecord::CallRecord):
* VM/CodeBlock.cpp:
Updated for new register layout:
(JSC::registerName):
(JSC::CodeBlock::dump):
* VM/CodeBlock.h:
Updated CodeBlock to track slightly different information about the
register frame, and tweaked the style of an ASSERT_NOT_REACHED.
(JSC::CodeBlock::CodeBlock):
(JSC::CodeBlock::getStubInfo):
* VM/CodeGenerator.cpp:
Added some abstraction around constant register allocation, since this
patch changes it, changed codegen to account for the new callframe
layout, and added abstraction around register fetching code
that used to assume that all local registers lived at negative indices,
since vars now live at positive indices:
(JSC::CodeGenerator::generate):
(JSC::CodeGenerator::addVar):
(JSC::CodeGenerator::addGlobalVar):
(JSC::CodeGenerator::allocateConstants):
(JSC::CodeGenerator::CodeGenerator):
(JSC::CodeGenerator::addParameter):
(JSC::CodeGenerator::registerFor):
(JSC::CodeGenerator::constRegisterFor):
(JSC::CodeGenerator::newRegister):
(JSC::CodeGenerator::newTemporary):
(JSC::CodeGenerator::highestUsedRegister):
(JSC::CodeGenerator::addConstant):
ASSERT that our caller referenced the registers it passed to us.
Otherwise, we might overwrite them with parameters:
(JSC::CodeGenerator::emitCall):
(JSC::CodeGenerator::emitConstruct):
* VM/CodeGenerator.h:
Added some abstraction for getting a RegisterID for a given index,
since the rules are a little weird:
(JSC::CodeGenerator::registerFor):
* VM/Machine.cpp:
Utility function to transform a machine return PC to a virtual machine
return VPC, for the sake of stack unwinding, since both PCs are stored
in the same location now:
(JSC::vPCForPC):
Tweaked to account for new call frame:
(JSC::Machine::initializeCallFrame):
Tweaked to account for registerOffset supplied by caller:
(JSC::slideRegisterWindowForCall):
Tweaked to account for new register layout:
(JSC::scopeChainForCall):
(JSC::Machine::callEval):
(JSC::Machine::dumpRegisters):
(JSC::Machine::unwindCallFrame):
(JSC::Machine::execute):
Changed op_call and op_construct to implement the new calling convention:
(JSC::Machine::privateExecute):
Tweaked to account for the new register layout:
(JSC::Machine::retrieveArguments):
(JSC::Machine::retrieveCaller):
(JSC::Machine::retrieveLastCaller):
(JSC::Machine::callFrame):
(JSC::Machine::getArgumentsData):
Changed CTI call helpers to implement the new calling convention:
(JSC::Machine::cti_op_call_JSFunction):
(JSC::Machine::cti_op_call_NotJSFunction):
(JSC::Machine::cti_op_ret_activation):
(JSC::Machine::cti_op_ret_profiler):
(JSC::Machine::cti_op_construct_JSConstruct):
(JSC::Machine::cti_op_construct_NotJSConstruct):
(JSC::Machine::cti_op_call_eval):
* VM/Machine.h:
* VM/Opcode.h:
Renamed op_initialise_locals to op_init, because this opcode
doesn't initialize all locals, and it doesn't initialize only locals.
Also, to spite Oliver.
* VM/RegisterFile.h:
New call frame enumeration values:
(JSC::RegisterFile::):
Simplified the calculation of whether a RegisterID is a temporary,
since we can no longer assume that all positive non-constant registers
are temporaries:
* VM/RegisterID.h:
(JSC::RegisterID::RegisterID):
(JSC::RegisterID::setTemporary):
(JSC::RegisterID::isTemporary):
Renamed firstArgumentIndex to firstParameterIndex because the assumption
that this variable pertained to the actual arguments supplied by the
caller caused me to write some buggy code:
* kjs/Arguments.cpp:
(JSC::ArgumentsData::ArgumentsData):
(JSC::Arguments::Arguments):
(JSC::Arguments::fillArgList):
(JSC::Arguments::getOwnPropertySlot):
(JSC::Arguments::put):
Updated for new call frame layout:
* kjs/DebuggerCallFrame.cpp:
(JSC::DebuggerCallFrame::functionName):
(JSC::DebuggerCallFrame::type):
* kjs/DebuggerCallFrame.h:
Changed the activation object to account for the fact that a call frame
header now sits between parameters and local variables. This change
requires all variable objects to do their own marking, since they
now use their register storage differently:
* kjs/JSActivation.cpp:
(JSC::JSActivation::mark):
(JSC::JSActivation::copyRegisters):
(JSC::JSActivation::createArgumentsObject):
* kjs/JSActivation.h:
Updated global object to use the new interfaces required by the change
to JSActivation above:
* kjs/JSGlobalObject.cpp:
(JSC::JSGlobalObject::reset):
(JSC::JSGlobalObject::mark):
(JSC::JSGlobalObject::copyGlobalsFrom):
(JSC::JSGlobalObject::copyGlobalsTo):
* kjs/JSGlobalObject.h:
(JSC::JSGlobalObject::addStaticGlobals):
Updated static scope object to use the new interfaces required by the
change to JSActivation above:
* kjs/JSStaticScopeObject.cpp:
(JSC::JSStaticScopeObject::mark):
(JSC::JSStaticScopeObject::~JSStaticScopeObject):
* kjs/JSStaticScopeObject.h:
(JSC::JSStaticScopeObject::JSStaticScopeObject):
(JSC::JSStaticScopeObject::d):
Updated variable object to use the new interfaces required by the
change to JSActivation above:
* kjs/JSVariableObject.cpp:
(JSC::JSVariableObject::copyRegisterArray):
(JSC::JSVariableObject::setRegisters):
* kjs/JSVariableObject.h:
Changed the bit twiddling in symbol table not to assume that all indices
are negative, since they can be positive now:
* kjs/SymbolTable.h:
(JSC::SymbolTableEntry::SymbolTableEntry):
(JSC::SymbolTableEntry::isNull):
(JSC::SymbolTableEntry::getIndex):
(JSC::SymbolTableEntry::getAttributes):
(JSC::SymbolTableEntry::setAttributes):
(JSC::SymbolTableEntry::isReadOnly):
(JSC::SymbolTableEntry::pack):
(JSC::SymbolTableEntry::isValidIndex):
Changed call and construct nodes to ref their functions and/or bases,
so that emitCall/emitConstruct doesn't overwrite them with parameters.
Also, updated for rename to registerFor:
* kjs/nodes.cpp:
(JSC::ResolveNode::emitCode):
(JSC::NewExprNode::emitCode):
(JSC::EvalFunctionCallNode::emitCode):
(JSC::FunctionCallValueNode::emitCode):
(JSC::FunctionCallResolveNode::emitCode):
(JSC::FunctionCallBracketNode::emitCode):
(JSC::FunctionCallDotNode::emitCode):
(JSC::PostfixResolveNode::emitCode):
(JSC::DeleteResolveNode::emitCode):
(JSC::TypeOfResolveNode::emitCode):
(JSC::PrefixResolveNode::emitCode):
(JSC::ReadModifyResolveNode::emitCode):
(JSC::AssignResolveNode::emitCode):
(JSC::ConstDeclNode::emitCodeSingle):
(JSC::ForInNode::emitCode):
Added abstraction for getting exception info out of a call through a
register:
* masm/X86Assembler.h:
(JSC::X86Assembler::emitCall):
Removed duplicate #if:
* wtf/Platform.h:
2008-09-23 Kevin McCullough <kmccullough@apple.com>
Reviewed by Darin.
Bug 21030: The JS debugger breaks on the do of a do-while not the while
(where the conditional statement is)
https://bugs.webkit.org/show_bug.cgi?id=21030
Now the statementListEmitCode detects if a do-while node is being
emited and emits the debug hook on the last line instead of the first.
This change had no effect on sunspider.
* kjs/nodes.cpp:
(JSC::statementListEmitCode):
* kjs/nodes.h:
(JSC::StatementNode::isDoWhile):
(JSC::DoWhileNode::isDoWhile):
2008-09-23 Maciej Stachowiak <mjs@apple.com>
Reviewed by Camron Zwarich.
- inline the fast case of instanceof
https://bugs.webkit.org/show_bug.cgi?id=20818
~2% speedup on EarleyBoyer test.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
* VM/Machine.cpp:
(JSC::Machine::cti_op_instanceof):
2008-09-23 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich.
- add forgotten slow case logic for !==
* VM/CTI.cpp:
(JSC::CTI::privateCompileSlowCases):
2008-09-23 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich.
- inline the fast cases of !==, same as for ===
2.9% speedup on EarleyBoyer benchmark
* VM/CTI.cpp:
(JSC::CTI::compileOpStrictEq): Factored stricteq codegen into this function,
and parameterized so it can do the reverse version as well.
(JSC::CTI::privateCompileMainPass): Use the above for stricteq and nstricteq.
* VM/CTI.h:
(JSC::CTI::): Declare above stuff.
* VM/Machine.cpp:
(JSC::Machine::cti_op_nstricteq): Removed fast cases, now handled inline.
2008-09-23 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Oliver Hunt.
Bug 20989: Aguments constructor should put 'callee' and 'length' properties in a more efficient way
<https://bugs.webkit.org/show_bug.cgi?id=20989>
Make special cases for the 'callee' and 'length' properties in the
Arguments object.
This is somewhere between a 7.8% speedup and a 10% speedup on the V8
Raytrace benchmark, depending on whether it is run alone or with the
other V8 benchmarks.
* kjs/Arguments.cpp:
(JSC::ArgumentsData::ArgumentsData):
(JSC::Arguments::Arguments):
(JSC::Arguments::mark):
(JSC::Arguments::getOwnPropertySlot):
(JSC::Arguments::put):
(JSC::Arguments::deleteProperty):
2008-09-23 Maciej Stachowiak <mjs@apple.com>
Reviewed by Darin.
- speed up instanceof some more
https://bugs.webkit.org/show_bug.cgi?id=20818
~2% speedup on EarleyBoyer
The idea here is to record in the StructureID whether the class
needs a special hasInstance or if it can use the normal logic from
JSObject.
Based on this I inlined the real work directly into
cti_op_instanceof and put the fastest checks up front and the
error handling at the end (so it should be fairly straightforward
to split off the beginning to be inlined if desired).
I only did this for CTI, not the bytecode interpreter.
* API/JSCallbackObject.h:
(JSC::JSCallbackObject::createStructureID):
* ChangeLog:
* VM/Machine.cpp:
(JSC::Machine::cti_op_instanceof):
* kjs/JSImmediate.h:
(JSC::JSImmediate::isAnyImmediate):
* kjs/TypeInfo.h:
(JSC::TypeInfo::overridesHasInstance):
(JSC::TypeInfo::flags):
2008-09-22 Darin Adler <darin@apple.com>
Reviewed by Sam Weinig.
- https://bugs.webkit.org/show_bug.cgi?id=21019
make FunctionBodyNode::ref/deref fast
Speeds up v8-raytrace by 7.2%.
* kjs/nodes.cpp:
(JSC::FunctionBodyNode::FunctionBodyNode): Initialize m_refCount to 0.
* kjs/nodes.h:
(JSC::FunctionBodyNode::ref): Call base class ref once, and thereafter use
m_refCount.
(JSC::FunctionBodyNode::deref): Ditto, but the deref side.
2008-09-22 Darin Adler <darin@apple.com>
Pointed out by Sam Weinig.
* kjs/Arguments.cpp:
(JSC::Arguments::fillArgList): Fix bad copy and paste. Oops!
2008-09-22 Darin Adler <darin@apple.com>
Reviewed by Cameron Zwarich.
- https://bugs.webkit.org/show_bug.cgi?id=20983
ArgumentsData should have some room to allocate some extra arguments inline
Speeds up v8-raytrace by 5%.
* kjs/Arguments.cpp:
(JSC::ArgumentsData::ArgumentsData): Use a fixed buffer if there are 4 or fewer
extra arguments.
(JSC::Arguments::Arguments): Use a fixed buffer if there are 4 or fewer
extra arguments.
(JSC::Arguments::~Arguments): Delete the buffer if necessary.
(JSC::Arguments::mark): Update since extraArguments are now Register.
(JSC::Arguments::fillArgList): Added special case for the only case that's
actually used in the practice, when there are no parameters. There are some
other special cases in there too, but that's the only one that matters.
(JSC::Arguments::getOwnPropertySlot): Updated to use setValueSlot since there's
no operation to get you at the JSValue* inside a Register as a "slot".
2008-09-22 Sam Weinig <sam@webkit.org>
Reviewed by Maciej Stachowiak.
Patch for https://bugs.webkit.org/show_bug.cgi?id=21014
Speed up for..in by using StructureID to avoid calls to hasProperty
Speeds up fasta by 8%.
* VM/JSPropertyNameIterator.cpp:
(JSC::JSPropertyNameIterator::invalidate):
* VM/JSPropertyNameIterator.h:
(JSC::JSPropertyNameIterator::next):
* kjs/PropertyNameArray.h:
(JSC::PropertyNameArrayData::begin):
(JSC::PropertyNameArrayData::end):
(JSC::PropertyNameArrayData::setCachedStructureID):
(JSC::PropertyNameArrayData::cachedStructureID):
* kjs/StructureID.cpp:
(JSC::StructureID::getEnumerablePropertyNames):
(JSC::structureIDChainsAreEqual):
* kjs/StructureID.h:
2008-09-22 Kelvin Sherlock <ksherlock@gmail.com>
Updated and tweaked by Sam Weinig.
Reviewed by Geoffrey Garen.
Bug 20020: Proposed enhancement to JavaScriptCore API
<https://bugs.webkit.org/show_bug.cgi?id=20020>
Add JSObjectMakeArray, JSObjectMakeDate, JSObjectMakeError, and JSObjectMakeRegExp
functions to create JavaScript Array, Date, Error, and RegExp objects, respectively.
* API/JSObjectRef.cpp: The functions
* API/JSObjectRef.h: Function prototype and documentation
* JavaScriptCore.exp: Added functions to exported function list
* API/tests/testapi.c: Added basic functionality tests.
* kjs/DateConstructor.cpp:
Replaced static JSObject* constructDate(ExecState* exec, JSObject*, const ArgList& args)
with JSObject* constructDate(ExecState* exec, const ArgList& args).
Added static JSObject* constructWithDateConstructor(ExecState* exec, JSObject*, const ArgList& args) function
* kjs/DateConstructor.h:
added prototype for JSObject* constructDate(ExecState* exec, const ArgList& args)
* kjs/ErrorConstructor.cpp:
removed static qualifier from ErrorInstance* constructError(ExecState* exec, const ArgList& args)
* kjs/ErrorConstructor.h:
added prototype for ErrorInstance* constructError(ExecState* exec, const ArgList& args)
* kjs/RegExpConstructor.cpp:
removed static qualifier from JSObject* constructRegExp(ExecState* exec, const ArgList& args)
* kjs/RegExpConstructor.h:
added prototype for JSObject* constructRegExp(ExecState* exec, const ArgList& args)
2008-09-22 Matt Lilek <webkit@mattlilek.com>
Not reviewed, Windows build fix.
* kjs/Arguments.cpp:
* kjs/FunctionPrototype.cpp:
2008-09-22 Sam Weinig <sam@webkit.org>
Reviewed by Darin Adler.
Patch for https://bugs.webkit.org/show_bug.cgi?id=20982
Speed up the apply method of functions by special-casing array and 'arguments' objects
1% speedup on v8-raytrace.
Test: fast/js/function-apply.html
* kjs/Arguments.cpp:
(JSC::Arguments::fillArgList):
* kjs/Arguments.h:
* kjs/FunctionPrototype.cpp:
(JSC::functionProtoFuncApply):
* kjs/JSArray.cpp:
(JSC::JSArray::fillArgList):
* kjs/JSArray.h:
2008-09-22 Darin Adler <darin@apple.com>
Reviewed by Sam Weinig.
- https://bugs.webkit.org/show_bug.cgi?id=20993
Array.push/pop need optimized cases for JSArray
3% or so speedup on DeltaBlue benchmark.
* kjs/ArrayPrototype.cpp:
(JSC::arrayProtoFuncPop): Call JSArray::pop when appropriate.
(JSC::arrayProtoFuncPush): Call JSArray::push when appropriate.
* kjs/JSArray.cpp:
(JSC::JSArray::putSlowCase): Set m_fastAccessCutoff when appropriate, getting
us into the fast code path.
(JSC::JSArray::pop): Added.
(JSC::JSArray::push): Added.
* kjs/JSArray.h: Added push and pop.
* kjs/operations.cpp:
(JSC::throwOutOfMemoryError): Don't inline this. Helps us avoid PIC branches.
2008-09-22 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich.
- speed up instanceof operator by replacing implementsHasInstance method with a TypeInfo flag
Partial work towards <https://bugs.webkit.org/show_bug.cgi?id=20818>
2.2% speedup on EarleyBoyer benchmark.
* API/JSCallbackConstructor.cpp:
* API/JSCallbackConstructor.h:
(JSC::JSCallbackConstructor::createStructureID):
* API/JSCallbackFunction.cpp:
* API/JSCallbackFunction.h:
(JSC::JSCallbackFunction::createStructureID):
* API/JSCallbackObject.h:
(JSC::JSCallbackObject::createStructureID):
* API/JSCallbackObjectFunctions.h:
(JSC::::hasInstance):
* API/JSValueRef.cpp:
(JSValueIsInstanceOfConstructor):
* JavaScriptCore.exp:
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
(JSC::Machine::cti_op_instanceof):
* kjs/InternalFunction.cpp:
* kjs/InternalFunction.h:
(JSC::InternalFunction::createStructureID):
* kjs/JSObject.cpp:
* kjs/JSObject.h:
* kjs/TypeInfo.h:
(JSC::TypeInfo::implementsHasInstance):
2008-09-22 Maciej Stachowiak <mjs@apple.com>
Reviewed by Dave Hyatt.
Based on initial work by Darin Adler.
- replace masqueradesAsUndefined virtual method with a flag in TypeInfo
- use this to JIT inline code for eq_null and neq_null
https://bugs.webkit.org/show_bug.cgi?id=20823
0.5% speedup on SunSpider
~4% speedup on Richards benchmark
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
* VM/Machine.cpp:
(JSC::jsTypeStringForValue):
(JSC::jsIsObjectType):
(JSC::Machine::privateExecute):
(JSC::Machine::cti_op_is_undefined):
* VM/Machine.h:
* kjs/JSCell.h:
* kjs/JSValue.h:
* kjs/StringObjectThatMasqueradesAsUndefined.h:
(JSC::StringObjectThatMasqueradesAsUndefined::create):
(JSC::StringObjectThatMasqueradesAsUndefined::createStructureID):
* kjs/StructureID.h:
(JSC::StructureID::mutableTypeInfo):
* kjs/TypeInfo.h:
(JSC::TypeInfo::TypeInfo):
(JSC::TypeInfo::masqueradesAsUndefined):
* kjs/operations.cpp:
(JSC::equal):
* masm/X86Assembler.h:
(JSC::X86Assembler::):
(JSC::X86Assembler::setne_r):
(JSC::X86Assembler::setnz_r):
(JSC::X86Assembler::testl_i32m):
2008-09-22 Tor Arne Vestbø <tavestbo@trolltech.com>
Reviewed by Simon.
Initialize QCoreApplication in kjs binary/Shell.cpp
This allows us to use QCoreApplication::instance() to
get the main thread in ThreadingQt.cpp
* kjs/Shell.cpp:
(main):
* wtf/ThreadingQt.cpp:
(WTF::initializeThreading):
2008-09-21 Darin Adler <darin@apple.com>
- blind attempt to fix non-all-in-one builds
* kjs/JSGlobalObject.cpp: Added includes of Arguments.h and RegExpObject.h.
2008-09-21 Darin Adler <darin@apple.com>
- fix debug build
* kjs/StructureID.cpp:
(JSC::StructureID::addPropertyTransition): Use typeInfo().type() instead of m_type.
(JSC::StructureID::createCachedPrototypeChain): Ditto.
2008-09-21 Maciej Stachowiak <mjs@apple.com>
Reviewed by Darin Adler.
- introduce a TypeInfo class, for holding per-type (in the C++ class sense) date in StructureID
https://bugs.webkit.org/show_bug.cgi?id=20981
* JavaScriptCore.exp:
* JavaScriptCore.xcodeproj/project.pbxproj:
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompilePutByIdTransition):
* VM/Machine.cpp:
(JSC::jsIsObjectType):
(JSC::Machine::Machine):
* kjs/AllInOneFile.cpp:
* kjs/JSCell.h:
(JSC::JSCell::isObject):
(JSC::JSCell::isString):
* kjs/JSGlobalData.cpp:
(JSC::JSGlobalData::JSGlobalData):
* kjs/JSGlobalObject.cpp:
(JSC::JSGlobalObject::reset):
* kjs/JSGlobalObject.h:
(JSC::StructureID::prototypeForLookup):
* kjs/JSNumberCell.h:
(JSC::JSNumberCell::createStructureID):
* kjs/JSObject.cpp:
(JSC::JSObject::createInheritorID):
* kjs/JSObject.h:
(JSC::JSObject::createStructureID):
* kjs/JSString.h:
(JSC::JSString::createStructureID):
* kjs/NativeErrorConstructor.cpp:
(JSC::NativeErrorConstructor::NativeErrorConstructor):
* kjs/RegExpConstructor.cpp:
* kjs/RegExpMatchesArray.h: Added.
(JSC::RegExpMatchesArray::getOwnPropertySlot):
(JSC::RegExpMatchesArray::put):
(JSC::RegExpMatchesArray::deleteProperty):
(JSC::RegExpMatchesArray::getPropertyNames):
* kjs/StructureID.cpp:
(JSC::StructureID::StructureID):
(JSC::StructureID::addPropertyTransition):
(JSC::StructureID::toDictionaryTransition):
(JSC::StructureID::changePrototypeTransition):
(JSC::StructureID::getterSetterTransition):
* kjs/StructureID.h:
(JSC::StructureID::create):
(JSC::StructureID::typeInfo):
* kjs/TypeInfo.h: Added.
(JSC::TypeInfo::TypeInfo):
(JSC::TypeInfo::type):
2008-09-21 Darin Adler <darin@apple.com>
Reviewed by Cameron Zwarich.
- fix crash logging into Gmail due to recent Arguments change
* kjs/Arguments.cpp:
(JSC::Arguments::Arguments): Fix window where mark() function could
see d->extraArguments with uninitialized contents.
(JSC::Arguments::mark): Check d->extraArguments for 0 to handle two
cases: 1) Inside the constructor before it's initialized.
2) numArguments <= numParameters.
2008-09-21 Darin Adler <darin@apple.com>
- fix loose end from the "duplicate constant values" patch
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitLoad): Add a special case for values the
hash table can't handle.
2008-09-21 Mark Rowe <mrowe@apple.com>
Fix the non-AllInOneFile build.
* kjs/Arguments.cpp: Add missing #include.
2008-09-21 Darin Adler <darin@apple.com>
Reviewed by Cameron Zwarich and Mark Rowe.
- fix test failure caused by my recent IndexToNameMap patch
* kjs/Arguments.cpp:
(JSC::Arguments::deleteProperty): Added the accidentally-omitted
check of the boolean result from toArrayIndex.
2008-09-21 Darin Adler <darin@apple.com>
Reviewed by Maciej Stachowiak.
- https://bugs.webkit.org/show_bug.cgi?id=20975
inline immediate-number case of ==
* VM/CTI.h: Renamed emitJumpSlowCaseIfNotImm to
emitJumpSlowCaseIfNotImmNum, since the old name was incorrect.
* VM/CTI.cpp: Updated for new name.
(JSC::CTI::privateCompileMainPass): Added op_eq.
(JSC::CTI::privateCompileSlowCases): Added op_eq.
* VM/Machine.cpp:
(JSC::Machine::cti_op_eq): Removed fast case, since it's now
compiled.
2008-09-21 Peter Gal <galpter@inf.u-szeged.hu>
Reviewed by Tim Hatcher and Eric Seidel.
Fix the QT/Linux JavaScriptCore segmentation fault.
https://bugs.webkit.org/show_bug.cgi?id=20914
* wtf/ThreadingQt.cpp:
(WTF::initializeThreading): Use currentThread() if
platform is not a MAC (like in pre 36541 revisions)
2008-09-21 Darin Adler <darin@apple.com>
Reviewed by Sam Weinig.
* kjs/debugger.h: Removed some unneeded includes and declarations.
2008-09-21 Darin Adler <darin@apple.com>
Reviewed by Sam Weinig.
- https://bugs.webkit.org/show_bug.cgi?id=20972
speed up Arguments further by eliminating the IndexToNameMap
No change on SunSpider. 1.29x as fast on V8 Raytrace.
* kjs/Arguments.cpp: Moved ArgumentsData in here. Eliminated the
indexToNameMap and hadDeletes data members. Changed extraArguments into
an OwnArrayPtr and added deletedArguments, another OwnArrayPtr.
Replaced numExtraArguments with numParameters, since that's what's
used more directly in hot code paths.
(JSC::Arguments::Arguments): Pass in argument count instead of ArgList.
Initialize ArgumentsData the new way.
(JSC::Arguments::mark): Updated.
(JSC::Arguments::getOwnPropertySlot): Overload for the integer form so
we don't have to convert integers to identifiers just to get an argument.
Integrated the deleted case with the fast case.
(JSC::Arguments::put): Ditto.
(JSC::Arguments::deleteProperty): Ditto.
* kjs/Arguments.h: Minimized includes. Made everything private. Added
overloads for the integral property name case. Eliminated mappedIndexSetter.
Moved ArgumentsData into the .cpp file.
* kjs/IndexToNameMap.cpp: Emptied out and prepared for deletion.
* kjs/IndexToNameMap.h: Ditto.
* kjs/JSActivation.cpp:
(JSC::JSActivation::createArgumentsObject): Elminated ArgList.
* GNUmakefile.am:
* JavaScriptCore.pri:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* JavaScriptCoreSources.bkl:
* kjs/AllInOneFile.cpp:
Removed IndexToNameMap.
2008-09-21 Darin Adler <darin@apple.com>
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitLoad): One more tweak: Wrote this in a slightly
clearer style.
2008-09-21 Judit Jasz <jasy@inf.u-szeged.hu>
Reviewed and tweaked by Darin Adler.
- https://bugs.webkit.org/show_bug.cgi?id=20645
Elminate duplicate constant values in CodeBlocks.
Seems to be a wash on SunSpider.
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitLoad): Use m_numberMap and m_stringMap to guarantee
we emit the same JSValue* for identical numbers and strings.
* VM/CodeGenerator.h: Added overload of emitLoad for const Identifier&.
Add NumberMap and IdentifierStringMap types and m_numberMap and m_stringMap.
* kjs/nodes.cpp:
(JSC::StringNode::emitCode): Call the new emitLoad and let it do the
JSString creation.
2008-09-21 Paul Pedriana <webkit@pedriana.com>
Reviewed and tweaked by Darin Adler.
- https://bugs.webkit.org/show_bug.cgi?id=16925
Fixed lack of Vector buffer alignment for both GCC and MSVC.
Since there's no portable way to do this, for now we don't support
other compilers.
* wtf/Vector.h: Added WTF_ALIGH_ON, WTF_ALIGNED, AlignedBufferChar, and AlignedBuffer.
Use AlignedBuffer insteadof an array of char in VectorBuffer.
2008-09-21 Gabor Loki <loki@inf.u-szeged.hu>
Reviewed by Darin Adler.
- https://bugs.webkit.org/show_bug.cgi?id=19408
Add lightweight constant folding to the parser for *, /, + (only for numbers), <<, >>, ~ operators.
1.008x as fast on SunSpider.
* kjs/grammar.y:
(makeNegateNode): Fold if expression is a number > 0.
(makeBitwiseNotNode): Fold if expression is a number.
(makeMultNode): Fold if expressions are both numbers.
(makeDivNode): Fold if expressions are both numbers.
(makeAddNode): Fold if expressions are both numbers.
(makeLeftShiftNode): Fold if expressions are both numbers.
(makeRightShiftNode): Fold if expressions are both numbers.
2008-09-21 Maciej Stachowiak <mjs@apple.com>
Reviewed by Oliver.
- speed up === operator by generating inline machine code for the fast paths
https://bugs.webkit.org/show_bug.cgi?id=20820
* VM/CTI.cpp:
(JSC::CTI::emitJumpSlowCaseIfNotImmediateNumber):
(JSC::CTI::emitJumpSlowCaseIfNotImmediateNumbers):
(JSC::CTI::emitJumpSlowCaseIfNotImmediates):
(JSC::CTI::emitTagAsBoolImmediate):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
* VM/CTI.h:
* VM/Machine.cpp:
(JSC::Machine::cti_op_stricteq):
* masm/X86Assembler.h:
(JSC::X86Assembler::):
(JSC::X86Assembler::sete_r):
(JSC::X86Assembler::setz_r):
(JSC::X86Assembler::movzbl_rr):
(JSC::X86Assembler::emitUnlinkedJnz):
2008-09-21 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Free memory allocated for extra arguments in the destructor of the
Arguments object.
* kjs/Arguments.cpp:
(JSC::Arguments::~Arguments):
* kjs/Arguments.h:
2008-09-21 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Bug 20815: 'arguments' object creation is non-optimal
<https://bugs.webkit.org/show_bug.cgi?id=20815>
Fix our inefficient way of creating the arguments object by only
creating named properties for each of the arguments after a use of the
'delete' statement. This patch also speeds up access to the 'arguments'
object slightly, but it still does not use the array fast path for
indexed access that exists for many opcodes.
This is about a 20% improvement on the V8 Raytrace benchmark, and a 1.5%
improvement on the Earley-Boyer benchmark, which gives a 4% improvement
overall.
* kjs/Arguments.cpp:
(JSC::Arguments::Arguments):
(JSC::Arguments::mark):
(JSC::Arguments::getOwnPropertySlot):
(JSC::Arguments::put):
(JSC::Arguments::deleteProperty):
* kjs/Arguments.h:
(JSC::Arguments::ArgumentsData::ArgumentsData):
* kjs/IndexToNameMap.h:
(JSC::IndexToNameMap::size):
* kjs/JSActivation.cpp:
(JSC::JSActivation::createArgumentsObject):
* kjs/JSActivation.h:
(JSC::JSActivation::uncheckedSymbolTableGet):
(JSC::JSActivation::uncheckedSymbolTableGetValue):
(JSC::JSActivation::uncheckedSymbolTablePut):
* kjs/JSFunction.h:
(JSC::JSFunction::numParameters):
2008-09-20 Darin Adler <darin@apple.com>
Reviewed by Mark Rowe.
- fix crash seen on buildbot
* kjs/JSGlobalObject.cpp:
(JSC::JSGlobalObject::mark): Add back mark of arrayPrototype,
deleted by accident in my recent check-in.
2008-09-20 Maciej Stachowiak <mjs@apple.com>
Not reviewed, build fix.
- speculative fix for non-AllInOne builds
* kjs/operations.h:
2008-09-20 Maciej Stachowiak <mjs@apple.com>
Reviewed by Darin Adler.
- assorted optimizations to === and !== operators
(work towards <https://bugs.webkit.org/show_bug.cgi?id=20820>)
2.5% speedup on earley-boyer test
* VM/Machine.cpp:
(JSC::Machine::cti_op_stricteq): Use inline version of
strictEqualSlowCase; remove unneeded exception check.
(JSC::Machine::cti_op_nstricteq): ditto
* kjs/operations.cpp:
(JSC::strictEqual): Use strictEqualSlowCaseInline
(JSC::strictEqualSlowCase): ditto
* kjs/operations.h:
(JSC::strictEqualSlowCaseInline): Version of strictEqualSlowCase that can be inlined,
since the extra function call indirection is a lose for CTI.
2008-09-20 Darin Adler <darin@apple.com>
Reviewed by Maciej Stachowiak.
- finish https://bugs.webkit.org/show_bug.cgi?id=20858
make each distinct C++ class get a distinct JSC::Structure
This also includes some optimizations that make the change an overall
small speedup. Without those it was a bit of a slowdown.
* API/JSCallbackConstructor.cpp:
(JSC::JSCallbackConstructor::JSCallbackConstructor): Take a structure.
* API/JSCallbackConstructor.h: Ditto.
* API/JSCallbackFunction.cpp:
(JSC::JSCallbackFunction::JSCallbackFunction): Pass a structure.
* API/JSCallbackObject.h: Take a structure.
* API/JSCallbackObjectFunctions.h:
(JSC::JSCallbackObject::JSCallbackObject): Ditto.
* API/JSClassRef.cpp:
(OpaqueJSClass::prototype): Pass in a structure. Call setPrototype
if there's a custom prototype involved.
* API/JSObjectRef.cpp:
(JSObjectMake): Ditto.
(JSObjectMakeConstructor): Pass in a structure.
* JavaScriptCore.exp: Updated.
* VM/Machine.cpp:
(JSC::jsLess): Added a special case for when both arguments are strings.
This avoids converting both strings to with UString::toDouble.
(JSC::jsLessEq): Ditto.
(JSC::Machine::privateExecute): Pass in a structure.
(JSC::Machine::cti_op_construct_JSConstruct): Ditto.
(JSC::Machine::cti_op_new_regexp): Ditto.
(JSC::Machine::cti_op_is_string): Ditto.
* VM/Machine.h: Made isJSString public so it can be used in the CTI.
* kjs/Arguments.cpp:
(JSC::Arguments::Arguments): Pass in a structure.
* kjs/JSCell.h: Mark constructor explicit.
* kjs/JSGlobalObject.cpp:
(JSC::markIfNeeded): Added an overload for marking structures.
(JSC::JSGlobalObject::reset): Eliminate code to set data members to
zero. We now do that in the constructor, and we no longer use this
anywhere except in the constructor. Added code to create structures.
Pass structures rather than prototypes when creating objects.
(JSC::JSGlobalObject::mark): Mark the structures.
* kjs/JSGlobalObject.h: Removed unneeded class declarations.
Added initializers for raw pointers in JSGlobalObjectData so
everything starts with a 0. Added structure data and accessor
functions.
* kjs/JSImmediate.cpp:
(JSC::JSImmediate::nonInlineNaN): Added.
* kjs/JSImmediate.h:
(JSC::JSImmediate::toDouble): Rewrote to avoid PIC branches.
* kjs/JSNumberCell.cpp:
(JSC::jsNumberCell): Made non-inline to avoid PIC branches
in functions that call this one.
(JSC::jsNaN): Ditto.
* kjs/JSNumberCell.h: Ditto.
* kjs/JSObject.h: Removed constructor that takes a prototype.
All callers now pass structures.
* kjs/ArrayConstructor.cpp:
(JSC::ArrayConstructor::ArrayConstructor):
(JSC::constructArrayWithSizeQuirk):
* kjs/ArrayConstructor.h:
* kjs/ArrayPrototype.cpp:
(JSC::ArrayPrototype::ArrayPrototype):
* kjs/ArrayPrototype.h:
* kjs/BooleanConstructor.cpp:
(JSC::BooleanConstructor::BooleanConstructor):
(JSC::constructBoolean):
(JSC::constructBooleanFromImmediateBoolean):
* kjs/BooleanConstructor.h:
* kjs/BooleanObject.cpp:
(JSC::BooleanObject::BooleanObject):
* kjs/BooleanObject.h:
* kjs/BooleanPrototype.cpp:
(JSC::BooleanPrototype::BooleanPrototype):
* kjs/BooleanPrototype.h:
* kjs/DateConstructor.cpp:
(JSC::DateConstructor::DateConstructor):
(JSC::constructDate):
* kjs/DateConstructor.h:
* kjs/DateInstance.cpp:
(JSC::DateInstance::DateInstance):
* kjs/DateInstance.h:
* kjs/DatePrototype.cpp:
(JSC::DatePrototype::DatePrototype):
* kjs/DatePrototype.h:
* kjs/ErrorConstructor.cpp:
(JSC::ErrorConstructor::ErrorConstructor):
(JSC::constructError):
* kjs/ErrorConstructor.h:
* kjs/ErrorInstance.cpp:
(JSC::ErrorInstance::ErrorInstance):
* kjs/ErrorInstance.h:
* kjs/ErrorPrototype.cpp:
(JSC::ErrorPrototype::ErrorPrototype):
* kjs/ErrorPrototype.h:
* kjs/FunctionConstructor.cpp:
(JSC::FunctionConstructor::FunctionConstructor):
* kjs/FunctionConstructor.h:
* kjs/FunctionPrototype.cpp:
(JSC::FunctionPrototype::FunctionPrototype):
(JSC::FunctionPrototype::addFunctionProperties):
* kjs/FunctionPrototype.h:
* kjs/GlobalEvalFunction.cpp:
(JSC::GlobalEvalFunction::GlobalEvalFunction):
* kjs/GlobalEvalFunction.h:
* kjs/InternalFunction.cpp:
(JSC::InternalFunction::InternalFunction):
* kjs/InternalFunction.h:
(JSC::InternalFunction::InternalFunction):
* kjs/JSArray.cpp:
(JSC::JSArray::JSArray):
(JSC::constructEmptyArray):
(JSC::constructArray):
* kjs/JSArray.h:
* kjs/JSFunction.cpp:
(JSC::JSFunction::JSFunction):
(JSC::JSFunction::construct):
* kjs/JSObject.cpp:
(JSC::constructEmptyObject):
* kjs/JSString.cpp:
(JSC::StringObject::create):
* kjs/JSWrapperObject.h:
* kjs/MathObject.cpp:
(JSC::MathObject::MathObject):
* kjs/MathObject.h:
* kjs/NativeErrorConstructor.cpp:
(JSC::NativeErrorConstructor::NativeErrorConstructor):
(JSC::NativeErrorConstructor::construct):
* kjs/NativeErrorConstructor.h:
* kjs/NativeErrorPrototype.cpp:
(JSC::NativeErrorPrototype::NativeErrorPrototype):
* kjs/NativeErrorPrototype.h:
* kjs/NumberConstructor.cpp:
(JSC::NumberConstructor::NumberConstructor):
(JSC::constructWithNumberConstructor):
* kjs/NumberConstructor.h:
* kjs/NumberObject.cpp:
(JSC::NumberObject::NumberObject):
(JSC::constructNumber):
(JSC::constructNumberFromImmediateNumber):
* kjs/NumberObject.h:
* kjs/NumberPrototype.cpp:
(JSC::NumberPrototype::NumberPrototype):
* kjs/NumberPrototype.h:
* kjs/ObjectConstructor.cpp:
(JSC::ObjectConstructor::ObjectConstructor):
(JSC::constructObject):
* kjs/ObjectConstructor.h:
* kjs/ObjectPrototype.cpp:
(JSC::ObjectPrototype::ObjectPrototype):
* kjs/ObjectPrototype.h:
* kjs/PrototypeFunction.cpp:
(JSC::PrototypeFunction::PrototypeFunction):
* kjs/PrototypeFunction.h:
* kjs/RegExpConstructor.cpp:
(JSC::RegExpConstructor::RegExpConstructor):
(JSC::RegExpMatchesArray::RegExpMatchesArray):
(JSC::constructRegExp):
* kjs/RegExpConstructor.h:
* kjs/RegExpObject.cpp:
(JSC::RegExpObject::RegExpObject):
* kjs/RegExpObject.h:
* kjs/RegExpPrototype.cpp:
(JSC::RegExpPrototype::RegExpPrototype):
* kjs/RegExpPrototype.h:
* kjs/Shell.cpp:
(GlobalObject::GlobalObject):
* kjs/StringConstructor.cpp:
(JSC::StringConstructor::StringConstructor):
(JSC::constructWithStringConstructor):
* kjs/StringConstructor.h:
* kjs/StringObject.cpp:
(JSC::StringObject::StringObject):
* kjs/StringObject.h:
* kjs/StringObjectThatMasqueradesAsUndefined.h:
(JSC::StringObjectThatMasqueradesAsUndefined::StringObjectThatMasqueradesAsUndefined):
* kjs/StringPrototype.cpp:
(JSC::StringPrototype::StringPrototype):
* kjs/StringPrototype.h:
Take and pass structures.
2008-09-19 Alp Toker <alp@nuanti.com>
Build fix for the 'gold' linker and recent binutils. New behaviour
requires that we link to used libraries explicitly.
* GNUmakefile.am:
2008-09-19 Sam Weinig <sam@webkit.org>
Roll r36694 back in. It did not cause the crash.
* JavaScriptCore.exp:
* VM/JSPropertyNameIterator.cpp:
(JSC::JSPropertyNameIterator::~JSPropertyNameIterator):
(JSC::JSPropertyNameIterator::invalidate):
* VM/JSPropertyNameIterator.h:
(JSC::JSPropertyNameIterator::JSPropertyNameIterator):
(JSC::JSPropertyNameIterator::create):
* kjs/JSObject.cpp:
(JSC::JSObject::getPropertyNames):
* kjs/PropertyMap.cpp:
(JSC::PropertyMap::getEnumerablePropertyNames):
* kjs/PropertyMap.h:
* kjs/PropertyNameArray.cpp:
(JSC::PropertyNameArray::add):
* kjs/PropertyNameArray.h:
(JSC::PropertyNameArrayData::create):
(JSC::PropertyNameArrayData::propertyNameVector):
(JSC::PropertyNameArrayData::setCachedPrototypeChain):
(JSC::PropertyNameArrayData::cachedPrototypeChain):
(JSC::PropertyNameArrayData::begin):
(JSC::PropertyNameArrayData::end):
(JSC::PropertyNameArrayData::PropertyNameArrayData):
(JSC::PropertyNameArray::PropertyNameArray):
(JSC::PropertyNameArray::addKnownUnique):
(JSC::PropertyNameArray::size):
(JSC::PropertyNameArray::operator[]):
(JSC::PropertyNameArray::begin):
(JSC::PropertyNameArray::end):
(JSC::PropertyNameArray::setData):
(JSC::PropertyNameArray::data):
(JSC::PropertyNameArray::releaseData):
* kjs/StructureID.cpp:
(JSC::structureIDChainsAreEqual):
(JSC::StructureID::getEnumerablePropertyNames):
(JSC::StructureID::clearEnumerationCache):
(JSC::StructureID::createCachedPrototypeChain):
* kjs/StructureID.h:
2008-09-19 Sam Weinig <sam@webkit.org>
Roll out r36694.
* JavaScriptCore.exp:
* VM/JSPropertyNameIterator.cpp:
(JSC::JSPropertyNameIterator::~JSPropertyNameIterator):
(JSC::JSPropertyNameIterator::invalidate):
* VM/JSPropertyNameIterator.h:
(JSC::JSPropertyNameIterator::JSPropertyNameIterator):
(JSC::JSPropertyNameIterator::create):
* kjs/JSObject.cpp:
(JSC::JSObject::getPropertyNames):
* kjs/PropertyMap.cpp:
(JSC::PropertyMap::getEnumerablePropertyNames):
* kjs/PropertyMap.h:
* kjs/PropertyNameArray.cpp:
(JSC::PropertyNameArray::add):
* kjs/PropertyNameArray.h:
(JSC::PropertyNameArray::PropertyNameArray):
(JSC::PropertyNameArray::addKnownUnique):
(JSC::PropertyNameArray::begin):
(JSC::PropertyNameArray::end):
(JSC::PropertyNameArray::size):
(JSC::PropertyNameArray::operator[]):
(JSC::PropertyNameArray::releaseIdentifiers):
* kjs/StructureID.cpp:
(JSC::StructureID::getEnumerablePropertyNames):
* kjs/StructureID.h:
(JSC::StructureID::clearEnumerationCache):
2008-09-19 Oliver Hunt <oliver@apple.com>
Reviewed by Maciej Stachowiak.
Improve peformance of local variable initialisation.
Pull local and constant initialisation out of slideRegisterWindowForCall
and into its own opcode. This allows the JIT to generate the initialisation
code for a function directly into the instruction stream and so avoids a few
branches on function entry.
Results a 1% progression in SunSpider, particularly in a number of the bitop
tests where the called functions are very fast.
* VM/CTI.cpp:
(JSC::CTI::emitInitialiseRegister):
(JSC::CTI::privateCompileMainPass):
* VM/CTI.h:
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::CodeGenerator):
* VM/Machine.cpp:
(JSC::slideRegisterWindowForCall):
(JSC::Machine::privateExecute):
* VM/Opcode.h:
2008-09-19 Sam Weinig <sam@webkit.org>
Reviewed by Darin Adler.
Patch for https://bugs.webkit.org/show_bug.cgi?id=20928
Speed up JS property enumeration by caching entire PropertyNameArray
1.3% speedup on Sunspider, 30% on string-fasta.
* JavaScriptCore.exp:
* VM/JSPropertyNameIterator.cpp:
(JSC::JSPropertyNameIterator::~JSPropertyNameIterator):
(JSC::JSPropertyNameIterator::invalidate):
* VM/JSPropertyNameIterator.h:
(JSC::JSPropertyNameIterator::JSPropertyNameIterator):
(JSC::JSPropertyNameIterator::create):
* kjs/JSObject.cpp:
(JSC::JSObject::getPropertyNames):
* kjs/PropertyMap.cpp:
(JSC::PropertyMap::getEnumerablePropertyNames):
* kjs/PropertyMap.h:
* kjs/PropertyNameArray.cpp:
(JSC::PropertyNameArray::add):
* kjs/PropertyNameArray.h:
(JSC::PropertyNameArrayData::create):
(JSC::PropertyNameArrayData::propertyNameVector):
(JSC::PropertyNameArrayData::setCachedPrototypeChain):
(JSC::PropertyNameArrayData::cachedPrototypeChain):
(JSC::PropertyNameArrayData::begin):
(JSC::PropertyNameArrayData::end):
(JSC::PropertyNameArrayData::PropertyNameArrayData):
(JSC::PropertyNameArray::PropertyNameArray):
(JSC::PropertyNameArray::addKnownUnique):
(JSC::PropertyNameArray::size):
(JSC::PropertyNameArray::operator[]):
(JSC::PropertyNameArray::begin):
(JSC::PropertyNameArray::end):
(JSC::PropertyNameArray::setData):
(JSC::PropertyNameArray::data):
(JSC::PropertyNameArray::releaseData):
* kjs/ScopeChain.cpp:
(JSC::ScopeChainNode::print):
* kjs/StructureID.cpp:
(JSC::structureIDChainsAreEqual):
(JSC::StructureID::getEnumerablePropertyNames):
(JSC::StructureID::clearEnumerationCache):
(JSC::StructureID::createCachedPrototypeChain):
* kjs/StructureID.h:
2008-09-19 Holger Hans Peter Freyther <zecke@selfish.org>
Reviewed by Maciej Stachowiak.
Fix a mismatched new[]/delete in JSObject::allocatePropertyStorage
* kjs/JSObject.cpp:
(JSC::JSObject::allocatePropertyStorage): Spotted by valgrind.
2008-09-19 Darin Adler <darin@apple.com>
Reviewed by Sam Weinig.
- part 2 of https://bugs.webkit.org/show_bug.cgi?id=20858
make each distinct C++ class get a distinct JSC::Structure
* JavaScriptCore.exp: Exported constructEmptyObject for use in WebCore.
* kjs/JSGlobalObject.h: Changed the protected constructor to take a
structure instead of a prototype.
* kjs/JSVariableObject.h: Removed constructor that takes a prototype.
2008-09-19 Julien Chaffraix <jchaffraix@pleyo.com>
Reviewed by Alexey Proskuryakov.
Use the template hoisting technique on the RefCounted class. This reduces the code bloat due to
non-template methods' code been copied for each instance of the template.
The patch splits RefCounted between a base class that holds non-template methods and attributes
and the template RefCounted class that keeps the same functionnality.
On my Linux with gcc 4.3 for the Gtk port, this is:
- a ~600KB save on libwebkit.so in release.
- a ~1.6MB save on libwebkit.so in debug.
It is a wash on Sunspider and a small win on Dromaeo (not sure it is relevant).
On the whole, it should be a small win as we reduce the compiled code size and the only
new function call should be inlined by the compiler.
* wtf/RefCounted.h:
(WTF::RefCountedBase::ref): Copied from RefCounted.
(WTF::RefCountedBase::hasOneRef): Ditto.
(WTF::RefCountedBase::refCount): Ditto.
(WTF::RefCountedBase::RefCountedBase): Ditto.
(WTF::RefCountedBase::~RefCountedBase): Ditto.
(WTF::RefCountedBase::derefBase): Tweaked from the RefCounted version to remove
template section.
(WTF::RefCounted::RefCounted):
(WTF::RefCounted::deref): Small wrapper around RefCountedBase::derefBase().
(WTF::RefCounted::~RefCounted): Keep private destructor.
2008-09-18 Darin Adler <darin@apple.com>
Reviewed by Maciej Stachowiak.
- part 1 of https://bugs.webkit.org/show_bug.cgi?id=20858
make each distinct C++ class get a distinct JSC::Structure
* kjs/lookup.h: Removed things here that were used only in WebCore:
cacheGlobalObject, JSC_DEFINE_PROTOTYPE, JSC_DEFINE_PROTOTYPE_WITH_PROTOTYPE,
and JSC_IMPLEMENT_PROTOTYPE.
2008-09-18 Darin Adler <darin@apple.com>
Reviewed by Maciej Stachowiak.
- https://bugs.webkit.org/show_bug.cgi?id=20927
simplify/streamline the code to turn strings into identifiers while parsing
* kjs/grammar.y: Get rid of string from the union, and use ident for STRING as
well as for IDENT.
* kjs/lexer.cpp:
(JSC::Lexer::lex): Use makeIdentifier instead of makeUString for String.
* kjs/lexer.h: Remove makeUString.
* kjs/nodes.h: Changed StringNode to hold an Identifier instead of UString.
* VM/CodeGenerator.cpp:
(JSC::keyForCharacterSwitch): Updated since StringNode now holds an Identifier.
(JSC::prepareJumpTableForStringSwitch): Ditto.
* kjs/nodes.cpp:
(JSC::StringNode::emitCode): Ditto. The comment from here is now in the lexer.
(JSC::processClauseList): Ditto.
* kjs/nodes2string.cpp:
(JSC::StringNode::streamTo): Ditto.
2008-09-18 Sam Weinig <sam@webkit.org>
Fix style.
* VM/Instruction.h:
(JSC::Instruction::Instruction):
2008-09-18 Oliver Hunt <oliver@apple.com>
Reviewed by Maciej Stachowiak.
Bug 20911: REGRESSION(r36480?): Reproducible assertion failure below derefStructureIDs 64-bit JavaScriptCore
<https://bugs.webkit.org/show_bug.cgi?id=20911>
The problem was simply caused by the int constructor for Instruction
failing to initialise the full struct in 64bit builds.
* VM/Instruction.h:
(JSC::Instruction::Instruction):
2008-09-18 Darin Adler <darin@apple.com>
- fix release build
* wtf/RefCountedLeakCounter.cpp: Removed stray "static".
2008-09-18 Darin Adler <darin@apple.com>
Reviewed by Sam Weinig.
* kjs/JSGlobalObject.h: Tiny style guideline tweak.
2008-09-18 Darin Adler <darin@apple.com>
Reviewed by Sam Weinig.
- fix https://bugs.webkit.org/show_bug.cgi?id=20925
LEAK messages appear every time I quit
* JavaScriptCore.exp: Updated, and also added an export
needed for future WebCore use of JSC::StructureID.
* wtf/RefCountedLeakCounter.cpp:
(WTF::RefCountedLeakCounter::suppressMessages): Added.
(WTF::RefCountedLeakCounter::cancelMessageSuppression): Added.
(WTF::RefCountedLeakCounter::RefCountedLeakCounter): Tweaked a bit.
(WTF::RefCountedLeakCounter::~RefCountedLeakCounter): Added code to
log the reason there was no leak checking done.
(WTF::RefCountedLeakCounter::increment): Tweaked a bit.
(WTF::RefCountedLeakCounter::decrement): Ditto.
* wtf/RefCountedLeakCounter.h: Replaced setLogLeakMessages with two
new functions, suppressMessages and cancelMessageSuppression. Also
added m_ prefixes to the data member names.
2008-09-18 Holger Hans Peter Freyther <zecke@selfish.org>
Reviewed by Mark Rowe.
https://bugs.webkit.org/show_bug.cgi?id=20437
Add a proper #define to define which XML Parser implementation to use. Client
code can use #if USE(QXMLSTREAM) to decide if the Qt XML StreamReader
implementation is going to be used.
* wtf/Platform.h:
2008-09-18 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Make a Unicode non-breaking space count as a whitespace character in
PCRE. This change was already made in WREC, and it fixes one of the
Mozilla JS tests. Since it is now fixed in PCRE as well, we can check
in a new set of expected test results.
* pcre/pcre_internal.h:
(isSpaceChar):
* tests/mozilla/expected.html:
2008-09-18 Stephanie Lewis <slewis@apple.com>
Reviewed by Mark Rowe and Maciej Stachowiak.
add an option use arch to specify which architecture to run.
* tests/mozilla/jsDriver.pl:
2008-09-17 Oliver Hunt <oliver@apple.com>
Correctly restore argument reference prior to SFX runtime calls.
Reviewed by Steve Falkenburg.
* VM/CTI.cpp:
(JSC::CTI::privateCompileSlowCases):
(JSC::CTI::privateCompile):
2008-09-17 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Bug 20876: REGRESSION (r36417, r36427): fast/js/exception-expression-offset.html fails
<https://bugs.webkit.org/show_bug.cgi?id=20876>
r36417 and r36427 caused an get_by_id opcode to be emitted before the
instanceof and construct opcodes, in order to enable inline caching of
the prototype property. Unfortunately, this regressed some tests dealing
with exceptions thrown by 'instanceof' and the 'new' operator. We fix
these problems by detecting whether an "is not an object" exception is
thrown before op_instanceof or op_construct, and emit the proper
exception in those cases.
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitConstruct):
* VM/CodeGenerator.h:
* VM/ExceptionHelpers.cpp:
(JSC::createInvalidParamError):
(JSC::createNotAConstructorError):
(JSC::createNotAnObjectError):
* VM/ExceptionHelpers.h:
* VM/Machine.cpp:
(JSC::Machine::getOpcode):
(JSC::Machine::privateExecute):
* VM/Machine.h:
* kjs/nodes.cpp:
(JSC::NewExprNode::emitCode):
(JSC::InstanceOfNode::emitCode):
2008-09-17 Gavin Barraclough <barraclough@apple.com>
Reviewed by Oliver Hunt.
JIT generation cti_op_construct_verify.
Quarter to half percent progression on v8-tests.
Roughly not change on SunSpider (possible minor progression).
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
* VM/Machine.cpp:
* VM/Machine.h:
2008-09-15 Steve Falkenburg <sfalken@apple.com>
Improve timer accuracy for JavaScript Date object on Windows.
Use a combination of ftime and QueryPerformanceCounter.
ftime returns the information we want, but doesn't have sufficient resolution.
QueryPerformanceCounter has high resolution, but is only usable to measure time intervals.
To combine them, we call ftime and QueryPerformanceCounter initially. Later calls will use
QueryPerformanceCounter by itself, adding the delta to the saved ftime. We re-sync to
correct for drift if the low-res and high-res elapsed time between calls differs by more
than twice the low-resolution timer resolution.
QueryPerformanceCounter may be inaccurate due to a problems with:
- some PCI bridge chipsets (http://support.microsoft.com/kb/274323)
- BIOS bugs (http://support.microsoft.com/kb/895980/)
- BIOS/HAL bugs on multiprocessor/multicore systems (http://msdn.microsoft.com/en-us/library/ms644904.aspx)
Reviewed by Darin Adler.
* kjs/DateMath.cpp:
(JSC::highResUpTime):
(JSC::lowResUTCTime):
(JSC::qpcAvailable):
(JSC::getCurrentUTCTimeWithMicroseconds):
2008-09-17 Gavin Barraclough <barraclough@apple.com>
Reviewed by Geoff Garen.
Implement JIT generation of CallFrame initialization, for op_call.
1% sunspider 2.5% v8-tests.
* VM/CTI.cpp:
(JSC::CTI::compileOpCall):
* VM/Machine.cpp:
(JSC::Machine::cti_op_call_JSFunction):
(JSC::Machine::cti_op_call_NotJSFunction):
2008-09-17 Gavin Barraclough <barraclough@apple.com>
Reviewed by Geoff Garen.
Optimizations for op_call in CTI. Move check for (ctiCode == 0) into JIT code,
move copying of scopeChain for CodeBlocks that needFullScopeChain into head of
functions, instead of checking prior to making the call.
3% on v8-tests (4% on richards, 6% in delta-blue)
* VM/CTI.cpp:
(JSC::CTI::compileOpCall):
(JSC::CTI::privateCompileSlowCases):
(JSC::CTI::privateCompile):
* VM/Machine.cpp:
(JSC::Machine::execute):
(JSC::Machine::cti_op_call_JSFunction):
(JSC::Machine::cti_vm_compile):
(JSC::Machine::cti_vm_updateScopeChain):
(JSC::Machine::cti_op_construct_JSConstruct):
* VM/Machine.h:
2008-09-17 Tor Arne Vestbø <tavestbo@trolltech.com>
Fix the QtWebKit/Mac build
* wtf/ThreadingQt.cpp:
(WTF::initializeThreading): use QCoreApplication to get the main thread
2008-09-16 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Bug 20857: REGRESSION (r36427): ASSERTION FAILED: m_refCount >= 0 in RegisterID::deref()
<https://bugs.webkit.org/show_bug.cgi?id=20857>
Fix a problem stemming from the slightly unsafe behaviour of the
CodeGenerator::finalDestination() method by putting the "func" argument
of the emitConstruct() method in a RefPtr in its caller. Also, add an
assertion guaranteeing that this is always the case.
CodeGenerator::finalDestination() is still incorrect and can cause
problems with a different allocator; see bug 20340 for more details.
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitConstruct):
* kjs/nodes.cpp:
(JSC::NewExprNode::emitCode):
2008-09-16 Alice Liu <alice.liu@apple.com>
build fix.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
2008-09-16 Gavin Barraclough <barraclough@apple.com>
Reviewed by Geoff Garen.
CTI code generation for op_ret. The majority of the work
(updating variables on the stack & on exec) can be performed
directly in generated code.
We still need to check, & to call out to C-code to handle
activation records, profiling, and full scope chains.
+1.5% Sunspider, +5/6% v8 tests.
* VM/CTI.cpp:
(JSC::CTI::emitPutCTIParam):
(JSC::CTI::compileOpCall):
(JSC::CTI::privateCompileMainPass):
* VM/CTI.h:
* VM/Machine.cpp:
(JSC::Machine::cti_op_ret_activation):
(JSC::Machine::cti_op_ret_profiler):
(JSC::Machine::cti_op_ret_scopeChain):
* VM/Machine.h:
2008-09-16 Dimitri Glazkov <dglazkov@chromium.org>
Fix the Windows build.
Add some extra parentheses to stop MSVC from complaining so much.
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
(JSC::Machine::cti_op_stricteq):
(JSC::Machine::cti_op_nstricteq):
* kjs/operations.cpp:
(JSC::strictEqual):
2008-09-15 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich.
- speed up the === and !== operators by choosing the fast cases better
No effect on SunSpider but speeds up the V8 EarlyBoyer benchmark about 4%.
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
(JSC::Machine::cti_op_stricteq):
(JSC::Machine::cti_op_nstricteq):
* kjs/JSImmediate.h:
(JSC::JSImmediate::areBothImmediate):
* kjs/operations.cpp:
(JSC::strictEqual):
(JSC::strictEqualSlowCase):
* kjs/operations.h:
2008-09-15 Oliver Hunt <oliver@apple.com>
RS=Sam Weinig.
Coding style cleanup.
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
2008-09-15 Oliver Hunt <oliver@apple.com>
Reviewed by Cameron Zwarich.
Bug 20874: op_resolve does not do any form of caching
<https://bugs.webkit.org/show_bug.cgi?id=20874>
This patch adds an op_resolve_global opcode to handle (and cache)
property lookup we can statically determine must occur on the global
object (if at all).
3% progression on sunspider, 3.2x improvement to bitops-bitwise-and, and
10% in math-partial-sums
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
* VM/CTI.h:
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::findScopedProperty):
(JSC::CodeGenerator::emitResolve):
* VM/Machine.cpp:
(JSC::resolveGlobal):
(JSC::Machine::privateExecute):
(JSC::Machine::cti_op_resolve_global):
* VM/Machine.h:
* VM/Opcode.h:
2008-09-15 Sam Weinig <sam@webkit.org>
Roll out r36462. It broke document.all.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
* VM/CTI.h:
* VM/Machine.cpp:
(JSC::Machine::Machine):
(JSC::Machine::cti_op_eq_null):
(JSC::Machine::cti_op_neq_null):
* VM/Machine.h:
(JSC::Machine::isJSString):
* kjs/JSCell.h:
* kjs/JSWrapperObject.h:
* kjs/StringObject.h:
* kjs/StringObjectThatMasqueradesAsUndefined.h:
2008-09-15 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Bug 20863: ASSERTION FAILED: addressOffset < instructions.size() in CodeBlock::getHandlerForVPC
<https://bugs.webkit.org/show_bug.cgi?id=20863>
r36427 changed the number of arguments to op_construct without changing
the argument index for the vPC in the call to initializeCallFrame() in
the CTI case. This caused a JSC test failure. Correcting the argument
index fixes the test failure.
* VM/Machine.cpp:
(JSC::Machine::cti_op_construct_JSConstruct):
2008-09-15 Mark Rowe <mrowe@apple.com>
Fix GCC 4.2 build.
* VM/CTI.h:
2008-09-15 Geoffrey Garen <ggaren@apple.com>
Reviewed by Maciej Stachowiak.
Fixed a typo in op_get_by_id_chain that caused it to miss every time
in the interpreter.
Also, a little cleanup.
* VM/Machine.cpp:
(JSC::Machine::privateExecute): Set up baseObject before entering the
loop, so we compare against the right values.
2008-09-15 Geoffrey Garen <ggaren@apple.com>
Reviewed by Sam Weinig.
Removed the CalledAsConstructor flag from the call frame header. Now,
we use an explicit opcode at the call site to fix up constructor results.
SunSpider says 0.4% faster.
cti_op_construct_verify is an out-of-line function call for now, but we
can fix that once StructureID holds type information like isObject.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass): Codegen for the new opcode.
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
* VM/CodeGenerator.cpp: Codegen for the new opcode. Also...
(JSC::CodeGenerator::emitCall): ... don't test for known non-zero value.
(JSC::CodeGenerator::emitConstruct): ... ditto.
* VM/Machine.cpp: No more CalledAsConstructor
(JSC::Machine::privateExecute): Implementation for the new opcode.
(JSC::Machine::cti_op_ret): The speedup: no need to check whether we were
called as a constructor.
(JSC::Machine::cti_op_construct_verify): Implementation for the new opcode.
* VM/Machine.h:
* VM/Opcode.h: Declare new opcode.
* VM/RegisterFile.h:
(JSC::RegisterFile::): No more CalledAsConstructor
2008-09-15 Gavin Barraclough <barraclough@apple.com>
Reviewed by Geoff Garen.
Inline code generation of eq_null/neq_null for CTI. Uses vptr checking for
StringObjectsThatAreMasqueradingAsBeingUndefined. In the long run, the
masquerading may be handled differently (through the StructureIDs - see bug
#20823).
>1% on v8-tests.
* VM/CTI.cpp:
(JSC::CTI::emitJumpSlowCaseIfIsJSCell):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
* VM/CTI.h:
* VM/Machine.cpp:
(JSC::Machine::Machine):
(JSC::Machine::cti_op_eq_null):
(JSC::Machine::cti_op_neq_null):
* VM/Machine.h:
(JSC::Machine::doesMasqueradesAsUndefined):
* kjs/JSWrapperObject.h:
(JSC::JSWrapperObject::):
(JSC::JSWrapperObject::JSWrapperObject):
* kjs/StringObject.h:
(JSC::StringObject::StringObject):
* kjs/StringObjectThatMasqueradesAsUndefined.h:
(JSC::StringObjectThatMasqueradesAsUndefined::StringObjectThatMasqueradesAsUndefined):
2008-09-15 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Rubber-stamped by Oliver Hunt.
r36427 broke CodeBlock::dump() by changing the number of arguments to
op_construct without changing the code that prints it. This patch fixes
it by printing the additional argument.
* JavaScriptCore.xcodeproj/project.pbxproj:
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
2008-09-15 Adam Roben <aroben@apple.com>
Build fix
* kjs/StructureID.cpp: Removed a stray semicolon.
2008-09-15 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Fix a crash in fast/js/exception-expression-offset.html caused by not
updating all mentions of the length of op_construct in r36427.
* VM/Machine.cpp:
(JSC::Machine::cti_op_construct_NotJSConstruct):
2008-09-15 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich.
- fix layout test failure introduced by fix for 20849
(The failing test was fast/js/delete-then-put.html)
* kjs/JSObject.cpp:
(JSC::JSObject::removeDirect): Clear enumeration cache
in the dictionary case.
* kjs/JSObject.h:
(JSC::JSObject::putDirect): Ditto.
* kjs/StructureID.h:
(JSC::StructureID::clearEnumerationCache): Inline to handle the
clear.
2008-09-15 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich.
- fix JSC test failures introduced by fix for 20849
* kjs/PropertyMap.cpp:
(JSC::PropertyMap::getEnumerablePropertyNames): Use the correct count.
2008-09-15 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Bug 20851: REGRESSION (r36410): fast/js/kde/GlobalObject.html fails
<https://bugs.webkit.org/show_bug.cgi?id=20851>
r36410 introduced an optimization for parseInt() that is incorrect when
its argument is larger than the range of a 32-bit integer. If the
argument is a number that is not an immediate integer, then the correct
behaviour is to return the floor of its value, unless it is an infinite
value, in which case the correct behaviour is to return 0.
* kjs/JSGlobalObjectFunctions.cpp:
(JSC::globalFuncParseInt):
2008-09-15 Sam Weinig <sam@webkit.org>
Reviewed by Maciej Stachowiak.
Patch for https://bugs.webkit.org/show_bug.cgi?id=20849
Cache property names for getEnumerablePropertyNames in the StructureID.
~0.5% speedup on Sunspider overall (9.7% speedup on string-fasta). ~1% speedup
on the v8 test suite.
* kjs/JSObject.cpp:
(JSC::JSObject::getPropertyNames):
* kjs/PropertyMap.cpp:
(JSC::PropertyMap::getEnumerablePropertyNames):
* kjs/PropertyMap.h:
* kjs/StructureID.cpp:
(JSC::StructureID::StructureID):
(JSC::StructureID::getEnumerablePropertyNames):
* kjs/StructureID.h:
2008-09-14 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich.
- speed up JS construction by extracting "prototype" lookup so PIC applies.
~0.5% speedup on SunSpider
Speeds up some of the V8 tests as well, most notably earley-boyer.
* VM/CTI.cpp:
(JSC::CTI::compileOpCall): Account for extra arg for prototype.
(JSC::CTI::privateCompileMainPass): Account for increased size of op_construct.
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitConstruct): Emit separate lookup to get prototype property.
* VM/Machine.cpp:
(JSC::Machine::privateExecute): Expect prototype arg in op_construct.
(JSC::Machine::cti_op_construct_JSConstruct): ditto
(JSC::Machine::cti_op_construct_NotJSConstruct): ditto
2008-09-10 Alexey Proskuryakov <ap@webkit.org>
Reviewed by Eric Seidel.
Add a protected destructor for RefCounted.
It is wrong to call its destructor directly, because (1) this should be taken care of by
deref(), and (2) many classes that use RefCounted have non-virtual destructors.
No change in behavior.
* wtf/RefCounted.h: (WTF::RefCounted::~RefCounted):
2008-09-14 Gavin Barraclough <barraclough@apple.com>
Reviewed by Sam Weinig.
Accelerated property accesses.
Inline more of the array access code into the JIT code for get/put_by_val.
Accelerate get/put_by_id by speculatively inlining a disable direct access
into the hot path of the code, and repatch this with the correct StructureID
and property map offset once these are known. In the case of accesses to the
prototype and reading the array-length a trampoline is genertaed, and the
branch to the slow-case is relinked to jump to this.
By repatching, we mean rewriting the x86 instruction stream. Instructions are
only modified in a simple fasion - altering immediate operands, memory access
deisplacements, and branch offsets.
For regular get_by_id/put_by_id accesses to an object, a StructureID in an
instruction's immediate operant is updateded, and a memory access operation's
displacement is updated to access the correct field on the object. In the case
of more complex accesses (array length and get_by_id_prototype) the offset on
the branch to slow-case is updated, to now jump to a trampoline.
+2.8% sunspider, +13% v8-tests
* VM/CTI.cpp:
(JSC::CTI::emitCall):
(JSC::CTI::emitJumpSlowCaseIfNotJSCell):
(JSC::CTI::CTI):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
(JSC::CTI::privateCompile):
(JSC::CTI::privateCompileGetByIdSelf):
(JSC::CTI::privateCompileGetByIdProto):
(JSC::CTI::privateCompileGetByIdChain):
(JSC::CTI::privateCompilePutByIdReplace):
(JSC::CTI::privateCompilePutByIdTransition):
(JSC::CTI::privateCompileArrayLengthTrampoline):
(JSC::CTI::privateCompileStringLengthTrampoline):
(JSC::CTI::patchGetByIdSelf):
(JSC::CTI::patchPutByIdReplace):
(JSC::CTI::privateCompilePatchGetArrayLength):
(JSC::CTI::privateCompilePatchGetStringLength):
* VM/CTI.h:
(JSC::CTI::compileGetByIdSelf):
(JSC::CTI::compileGetByIdProto):
(JSC::CTI::compileGetByIdChain):
(JSC::CTI::compilePutByIdReplace):
(JSC::CTI::compilePutByIdTransition):
(JSC::CTI::compileArrayLengthTrampoline):
(JSC::CTI::compileStringLengthTrampoline):
(JSC::CTI::compilePatchGetArrayLength):
(JSC::CTI::compilePatchGetStringLength):
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
(JSC::CodeBlock::~CodeBlock):
* VM/CodeBlock.h:
(JSC::StructureStubInfo::StructureStubInfo):
(JSC::CodeBlock::getStubInfo):
* VM/Machine.cpp:
(JSC::Machine::tryCTICachePutByID):
(JSC::Machine::tryCTICacheGetByID):
(JSC::Machine::cti_op_put_by_val_array):
* VM/Machine.h:
* masm/X86Assembler.h:
(JSC::X86Assembler::):
(JSC::X86Assembler::cmpl_i8m):
(JSC::X86Assembler::emitUnlinkedJa):
(JSC::X86Assembler::getRelocatedAddress):
(JSC::X86Assembler::getDifferenceBetweenLabels):
(JSC::X86Assembler::emitModRm_opmsib):
2008-09-14 Maciej Stachowiak <mjs@apple.com>
Reviewed by Cameron Zwarich.
- split the "prototype" lookup for hasInstance into opcode stream so it can be cached
~5% speedup on v8 earley-boyer test
* API/JSCallbackObject.h: Add a parameter for the pre-looked-up prototype.
* API/JSCallbackObjectFunctions.h:
(JSC::::hasInstance): Ditto.
* API/JSValueRef.cpp:
(JSValueIsInstanceOfConstructor): Look up and pass in prototype.
* JavaScriptCore.exp:
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass): Pass along prototype.
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump): Print third arg.
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitInstanceOf): Implement this, now that there
is a third argument.
* VM/CodeGenerator.h:
* VM/Machine.cpp:
(JSC::Machine::privateExecute): Pass along the prototype.
(JSC::Machine::cti_op_instanceof): ditto
* kjs/JSObject.cpp:
(JSC::JSObject::hasInstance): Expect to get a pre-looked-up prototype.
* kjs/JSObject.h:
* kjs/nodes.cpp:
(JSC::InstanceOfNode::emitCode): Emit a get_by_id of the prototype
property and pass that register to instanceof.
* kjs/nodes.h:
2008-09-14 Gavin Barraclough <barraclough@apple.com>
Reviewed by Sam Weinig.
Remove unnecessary virtual function call from cti_op_call_JSFunction -
~5% on richards, ~2.5% on v8-tests, ~0.5% on sunspider.
* VM/Machine.cpp:
(JSC::Machine::cti_op_call_JSFunction):
2008-09-14 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Bug 20827: the 'typeof' operator is slow
<https://bugs.webkit.org/show_bug.cgi?id=20827>
Optimize the 'typeof' operator when its result is compared to a constant
string.
This is a 5.5% speedup on the V8 Earley-Boyer test.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitEqualityOp):
* VM/CodeGenerator.h:
* VM/Machine.cpp:
(JSC::jsIsObjectType):
(JSC::jsIsFunctionType):
(JSC::Machine::privateExecute):
(JSC::Machine::cti_op_is_undefined):
(JSC::Machine::cti_op_is_boolean):
(JSC::Machine::cti_op_is_number):
(JSC::Machine::cti_op_is_string):
(JSC::Machine::cti_op_is_object):
(JSC::Machine::cti_op_is_function):
* VM/Machine.h:
* VM/Opcode.h:
* kjs/nodes.cpp:
(JSC::BinaryOpNode::emitCode):
(JSC::EqualNode::emitCode):
(JSC::StrictEqualNode::emitCode):
* kjs/nodes.h:
2008-09-14 Sam Weinig <sam@webkit.org>
Reviewed by Cameron Zwarich.
Patch for https://bugs.webkit.org/show_bug.cgi?id=20844
Speed up parseInt for numbers
Sunspider reports this as 1.029x as fast overall and 1.37x as fast on string-unpack-code.
No change on the v8 suite.
* kjs/JSGlobalObjectFunctions.cpp:
(JSC::globalFuncParseInt): Don't convert numbers to strings just to
convert them back to numbers.
2008-09-14 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Oliver Hunt.
Bug 20816: op_lesseq should be optimized
<https://bugs.webkit.org/show_bug.cgi?id=20816>
Add a loop_if_lesseq opcode that is similar to the loop_if_less opcode.
This is a 9.4% speedup on the V8 Crypto benchmark.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitJumpIfTrue):
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
(JSC::Machine::cti_op_loop_if_lesseq):
* VM/Machine.h:
* VM/Opcode.h:
2008-09-14 Sam Weinig <sam@webkit.org>
Reviewed by Cameron Zwarich.
Cleanup Sampling code.
* VM/CTI.cpp:
(JSC::CTI::emitCall):
(JSC::CTI::privateCompileMainPass):
* VM/CTI.h:
(JSC::CTI::execute):
* VM/SamplingTool.cpp:
(JSC::):
(JSC::SamplingTool::run):
(JSC::SamplingTool::dump):
* VM/SamplingTool.h:
(JSC::SamplingTool::callingHostFunction):
2008-09-13 Oliver Hunt <oliver@apple.com>
Reviewed by Cameron Zwarich.
Bug 20821: Cache property transitions to speed up object initialization
https://bugs.webkit.org/show_bug.cgi?id=20821
Implement a transition cache to improve the performance of new properties
being added to objects. This is extremely beneficial in constructors and
shows up as a 34% improvement on access-binary-trees in SunSpider (0.8%
overall)
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
(JSC::):
(JSC::transitionWillNeedStorageRealloc):
(JSC::CTI::privateCompilePutByIdTransition):
* VM/CTI.h:
(JSC::CTI::compilePutByIdTransition):
* VM/CodeBlock.cpp:
(JSC::printPutByIdOp):
(JSC::CodeBlock::printStructureIDs):
(JSC::CodeBlock::dump):
(JSC::CodeBlock::derefStructureIDs):
(JSC::CodeBlock::refStructureIDs):
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::emitPutById):
* VM/Machine.cpp:
(JSC::cachePrototypeChain):
(JSC::Machine::tryCachePutByID):
(JSC::Machine::tryCacheGetByID):
(JSC::Machine::privateExecute):
(JSC::Machine::tryCTICachePutByID):
(JSC::Machine::tryCTICacheGetByID):
* VM/Machine.h:
* VM/Opcode.h:
* kjs/JSObject.h:
(JSC::JSObject::putDirect):
(JSC::JSObject::transitionTo):
* kjs/PutPropertySlot.h:
(JSC::PutPropertySlot::PutPropertySlot):
(JSC::PutPropertySlot::wasTransition):
(JSC::PutPropertySlot::setWasTransition):
* kjs/StructureID.cpp:
(JSC::StructureID::transitionTo):
(JSC::StructureIDChain::StructureIDChain):
* kjs/StructureID.h:
(JSC::StructureID::previousID):
(JSC::StructureID::setCachedPrototypeChain):
(JSC::StructureID::cachedPrototypeChain):
(JSC::StructureID::propertyMap):
* masm/X86Assembler.h:
(JSC::X86Assembler::addl_i8m):
(JSC::X86Assembler::subl_i8m):
2008-09-12 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Bug 20819: JSValue::isObject() is slow
<https://bugs.webkit.org/show_bug.cgi?id=20819>
Optimize JSCell::isObject() and JSCell::isString() by making them
non-virtual calls that rely on the StructureID type information.
This is a 0.7% speedup on SunSpider and a 1.0% speedup on the V8
benchmark suite.
* JavaScriptCore.exp:
* kjs/JSCell.cpp:
* kjs/JSCell.h:
(JSC::JSCell::isObject):
(JSC::JSCell::isString):
* kjs/JSObject.cpp:
* kjs/JSObject.h:
* kjs/JSString.cpp:
* kjs/JSString.h:
(JSC::JSString::JSString):
* kjs/StructureID.h:
(JSC::StructureID::type):
2008-09-11 Stephanie Lewis <slewis@apple.com>
Reviewed by Oliver Hunt.
Turn off PGO Optimization on CTI.cpp -> <rdar://problem/6207709>. Fixes
crash on CNN and on Dromaeo.
Fix Missing close tag in vcproj.
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
2008-09-11 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Not reviewed.
Correct an SVN problem with the last commit and actually add the new
files.
* wrec/CharacterClassConstructor.cpp: Added.
(JSC::):
(JSC::getCharacterClassNewline):
(JSC::getCharacterClassDigits):
(JSC::getCharacterClassSpaces):
(JSC::getCharacterClassWordchar):
(JSC::getCharacterClassNondigits):
(JSC::getCharacterClassNonspaces):
(JSC::getCharacterClassNonwordchar):
(JSC::CharacterClassConstructor::addSorted):
(JSC::CharacterClassConstructor::addSortedRange):
(JSC::CharacterClassConstructor::put):
(JSC::CharacterClassConstructor::flush):
(JSC::CharacterClassConstructor::append):
* wrec/CharacterClassConstructor.h: Added.
(JSC::CharacterClassConstructor::CharacterClassConstructor):
(JSC::CharacterClassConstructor::isUpsideDown):
(JSC::CharacterClassConstructor::charClass):
2008-09-11 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Bug 20788: Split CharacterClassConstructor into its own file
<https://bugs.webkit.org/show_bug.cgi?id=20788>
Split CharacterClassConstructor into its own file and clean up some
style issues.
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* wrec/CharacterClassConstructor.cpp: Added.
(JSC::):
(JSC::getCharacterClassNewline):
(JSC::getCharacterClassDigits):
(JSC::getCharacterClassSpaces):
(JSC::getCharacterClassWordchar):
(JSC::getCharacterClassNondigits):
(JSC::getCharacterClassNonspaces):
(JSC::getCharacterClassNonwordchar):
(JSC::CharacterClassConstructor::addSorted):
(JSC::CharacterClassConstructor::addSortedRange):
(JSC::CharacterClassConstructor::put):
(JSC::CharacterClassConstructor::flush):
(JSC::CharacterClassConstructor::append):
* wrec/CharacterClassConstructor.h: Added.
(JSC::CharacterClassConstructor::CharacterClassConstructor):
(JSC::CharacterClassConstructor::isUpsideDown):
(JSC::CharacterClassConstructor::charClass):
* wrec/WREC.cpp:
(JSC::WRECParser::parseCharacterClass):
2008-09-10 Simon Hausmann <hausmann@webkit.org>
Not reviewed but trivial one-liner for yet unused macro.
Changed PLATFORM(WINCE) to PLATFORM(WIN_CE) as requested by Mark.
(part of https://bugs.webkit.org/show_bug.cgi?id=20746)
* wtf/Platform.h:
2008-09-10 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Rubber-stamped by Oliver Hunt.
Fix a typo by renaming the overloaded orl_rr that takes an immediate to
orl_i32r.
* VM/CTI.cpp:
(JSC::CTI::emitFastArithPotentiallyReTagImmediate):
* masm/X86Assembler.h:
(JSC::X86Assembler::orl_i32r):
* wrec/WREC.cpp:
(JSC::WRECGenerator::generatePatternCharacter):
(JSC::WRECGenerator::generateCharacterClassInverted):
2008-09-10 Sam Weinig <sam@webkit.org>
Reviewed by Geoff Garen.
Add inline property storage for JSObject.
1.2% progression on Sunspider. .5% progression on the v8 test suite.
* JavaScriptCore.exp:
* VM/CTI.cpp:
(JSC::CTI::privateCompileGetByIdProto):
(JSC::CTI::privateCompileGetByIdChain):
* kjs/JSObject.cpp:
(JSC::JSObject::mark): There is no reason to check storageSize now that
we start from 0.
(JSC::JSObject::allocatePropertyStorage): Allocates/reallocates heap storage.
* kjs/JSObject.h:
(JSC::JSObject::offsetForLocation): m_propertyStorage is not an OwnArrayPtr
now so there is no reason to .get()
(JSC::JSObject::usingInlineStorage):
(JSC::JSObject::JSObject): Start with m_propertyStorage pointing to the
inline storage.
(JSC::JSObject::~JSObject): Free the heap storage if not using the inline
storage.
(JSC::JSObject::putDirect): Switch to the heap storage only when we know
we know that we are about to add a property that will overflow the inline
storage.
* kjs/PropertyMap.cpp:
(JSC::PropertyMap::createTable): Don't allocate the propertyStorage, that is
now handled by JSObject.
(JSC::PropertyMap::rehash): PropertyStorage is not a OwnArrayPtr anymore.
* kjs/PropertyMap.h:
(JSC::PropertyMap::storageSize): Rename from markingCount.
* kjs/StructureID.cpp:
(JSC::StructureID::addPropertyTransition): Don't resize the property storage
if we are using inline storage.
* kjs/StructureID.h:
2008-09-10 Oliver Hunt <oliver@apple.com>
Reviewed by Geoff Garen.
Inline immediate number version of op_mul.
Renamed mull_rr to imull_rr as that's what it's
actually doing, and added imull_i32r for the constant
case immediate multiply.
1.1% improvement to SunSpider.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
* masm/X86Assembler.h:
(JSC::X86Assembler::):
(JSC::X86Assembler::imull_rr):
(JSC::X86Assembler::imull_i32r):
2008-09-10 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Not reviewed.
Mac build fix.
* JavaScriptCore.xcodeproj/project.pbxproj:
2008-09-09 Oliver Hunt <oliver@apple.com>
Reviewed by Maciej Stachowiak.
Add optimised access to known properties on the global object.
Improve cross scope access to the global object by emitting
code to access it directly rather than by walking the scope chain.
This is a 0.8% win in SunSpider and a 1.7% win in the v8 benchmarks.
* VM/CTI.cpp:
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::emitGetVariableObjectRegister):
(JSC::CTI::emitPutVariableObjectRegister):
* VM/CTI.h:
* VM/CodeBlock.cpp:
(JSC::CodeBlock::dump):
* VM/CodeGenerator.cpp:
(JSC::CodeGenerator::findScopedProperty):
(JSC::CodeGenerator::emitResolve):
(JSC::CodeGenerator::emitGetScopedVar):
(JSC::CodeGenerator::emitPutScopedVar):
* VM/CodeGenerator.h:
* VM/Machine.cpp:
(JSC::Machine::privateExecute):
* VM/Opcode.h:
* kjs/nodes.cpp:
(JSC::FunctionCallResolveNode::emitCode):
(JSC::PostfixResolveNode::emitCode):
(JSC::PrefixResolveNode::emitCode):
(JSC::ReadModifyResolveNode::emitCode):
(JSC::AssignResolveNode::emitCode):
2008-09-10 Maciej Stachowiak <mjs@apple.com>
Reviewed by Oliver.
- enable polymorphic inline caching of properties of primitives
1.012x speedup on SunSpider.
We create special structure IDs for JSString and
JSNumberCell. Unlike normal structure IDs, these cannot hold the
true prototype. Due to JS autoboxing semantics, the prototype used
when looking up string or number properties depends on the lexical
global object of the call site, not the creation site. Thus we
enable StructureIDs to handle this quirk for primitives.
Everything else should be straightforward.
* VM/CTI.cpp:
(JSC::CTI::privateCompileGetByIdProto):
(JSC::CTI::privateCompileGetByIdChain):
* VM/CTI.h:
(JSC::CTI::compileGetByIdProto):
(JSC::CTI::compileGetByIdChain):
* VM/JSPropertyNameIterator.h:
(JSC::JSPropertyNameIterator::JSPropertyNameIterator):
* VM/Machine.cpp:
(JSC::Machine::Machine):
(JSC::cachePrototypeChain):
(JSC::Machine::tryCachePutByID):
(JSC::Machine::tryCacheGetByID):
(JSC::Machine::privateExecute):
(JSC::Machine::tryCTICachePutByID):
(JSC::Machine::tryCTICacheGetByID):
* kjs/GetterSetter.h:
(JSC::GetterSetter::GetterSetter):
* kjs/JSCell.h:
* kjs/JSGlobalData.cpp:
(JSC::JSGlobalData::JSGlobalData):
* kjs/JSGlobalData.h:
* kjs/JSGlobalObject.h:
(JSC::StructureID::prototypeForLookup):
* kjs/JSNumberCell.h:
(JSC::JSNumberCell::JSNumberCell):
(JSC::jsNumberCell):
* kjs/JSObject.h:
(JSC::JSObject::prototype):
* kjs/JSString.cpp:
(JSC::jsString):
(JSC::jsSubstring):
(JSC::jsOwnedString):
* kjs/JSString.h:
(JSC::JSString::JSString):
(JSC::JSString::):
(JSC::jsSingleCharacterString):
(JSC::jsSingleCharacterSubstring):
(JSC::jsNontrivialString):
* kjs/SmallStrings.cpp:
(JSC::SmallStrings::createEmptyString):
(JSC::SmallStrings::createSingleCharacterString):
* kjs/StructureID.cpp:
(JSC::StructureID::StructureID):
(JSC::StructureID::addPropertyTransition):
(JSC::StructureID::getterSetterTransition):
(JSC::StructureIDChain::StructureIDChain):
* kjs/StructureID.h:
(JSC::StructureID::create):
(JSC::StructureID::storedPrototype):
2008-09-09 Joerg Bornemann <joerg.bornemann@trolltech.com>
Reviewed by Sam Weinig.
https://bugs.webkit.org/show_bug.cgi?id=20746
Added WINCE platform macro.
* wtf/Platform.h:
2008-09-09 Sam Weinig <sam@webkit.org>
Reviewed by Mark Rowe.
Remove unnecessary override of getOffset.
Sunspider reports this as a .6% progression.
* JavaScriptCore.exp:
* kjs/JSObject.h:
(JSC::JSObject::getDirectLocation):
(JSC::JSObject::getOwnPropertySlotForWrite):
(JSC::JSObject::putDirect):
* kjs/PropertyMap.cpp:
* kjs/PropertyMap.h:
2008-09-09 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Bug 20759: Remove MacroAssembler
<https://bugs.webkit.org/show_bug.cgi?id=20759>
Remove MacroAssembler and move its functionality to X86Assembler.
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* VM/CTI.cpp:
(JSC::CTI::emitGetArg):
(JSC::CTI::emitGetPutArg):
(JSC::CTI::emitPutArg):
(JSC::CTI::emitPutCTIParam):
(JSC::CTI::emitGetCTIParam):
(JSC::CTI::emitPutToCallFrameHeader):
(JSC::CTI::emitGetFromCallFrameHeader):
(JSC::CTI::emitPutResult):
(JSC::CTI::emitDebugExceptionCheck):
(JSC::CTI::emitJumpSlowCaseIfNotImm):
(JSC::CTI::emitJumpSlowCaseIfNotImms):
(JSC::CTI::emitFastArithDeTagImmediate):
(JSC::CTI::emitFastArithReTagImmediate):
(JSC::CTI::emitFastArithPotentiallyReTagImmediate):
(JSC::CTI::emitFastArithImmToInt):
(JSC::CTI::emitFastArithIntToImmOrSlowCase):
(JSC::CTI::emitFastArithIntToImmNoCheck):
(JSC::CTI::compileOpCall):
(JSC::CTI::emitSlowScriptCheck):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
(JSC::CTI::privateCompile):
(JSC::CTI::privateCompileGetByIdSelf):
(JSC::CTI::privateCompileGetByIdProto):
(JSC::CTI::privateCompileGetByIdChain):
(JSC::CTI::privateCompilePutByIdReplace):
(JSC::CTI::privateArrayLengthTrampoline):
(JSC::CTI::privateStringLengthTrampoline):
(JSC::CTI::compileRegExp):
* VM/CTI.h:
(JSC::CallRecord::CallRecord):
(JSC::JmpTable::JmpTable):
(JSC::SlowCaseEntry::SlowCaseEntry):
(JSC::CTI::JSRInfo::JSRInfo):
* masm/MacroAssembler.h: Removed.
* masm/MacroAssemblerWin.cpp: Removed.
* masm/X86Assembler.h:
(JSC::X86Assembler::emitConvertToFastCall):
(JSC::X86Assembler::emitRestoreArgumentReference):
* wrec/WREC.h:
(JSC::WRECGenerator::WRECGenerator):
(JSC::WRECParser::WRECParser):
2008-09-09 Sam Weinig <sam@webkit.org>
Reviewed by Cameron Zwarich.
Don't waste the first item in the PropertyStorage.
- Fix typo (makingCount -> markingCount)
- Remove undefined method declaration.
No change on Sunspider.
* kjs/JSObject.cpp:
(JSC::JSObject::mark):
* kjs/PropertyMap.cpp:
(JSC::PropertyMap::put):
(JSC::PropertyMap::remove):
(JSC::PropertyMap::getOffset):
(JSC::PropertyMap::insert):
(JSC::PropertyMap::rehash):
(JSC::PropertyMap::resizePropertyStorage):
(JSC::PropertyMap::checkConsistency):
* kjs/PropertyMap.h:
(JSC::PropertyMap::markingCount): Fix typo.
2008-09-09 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Not reviewed.
Speculative Windows build fix.
* masm/MacroAssemblerWin.cpp:
(JSC::MacroAssembler::emitConvertToFastCall):
(JSC::MacroAssembler::emitRestoreArgumentReference):
2008-09-09 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Bug 20755: Create an X86 namespace for register names and other things
<https://bugs.webkit.org/show_bug.cgi?id=20755>
Create an X86 namespace to put X86 register names. Perhaps I will move
opcode names here later as well.
* VM/CTI.cpp:
(JSC::CTI::emitGetArg):
(JSC::CTI::emitGetPutArg):
(JSC::CTI::emitPutArg):
(JSC::CTI::emitPutArgConstant):
(JSC::CTI::emitPutCTIParam):
(JSC::CTI::emitGetCTIParam):
(JSC::CTI::emitPutToCallFrameHeader):
(JSC::CTI::emitGetFromCallFrameHeader):
(JSC::CTI::emitPutResult):
(JSC::CTI::emitDebugExceptionCheck):
(JSC::CTI::emitJumpSlowCaseIfNotImms):
(JSC::CTI::compileOpCall):
(JSC::CTI::emitSlowScriptCheck):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
(JSC::CTI::privateCompile):
(JSC::CTI::privateCompileGetByIdSelf):
(JSC::CTI::privateCompileGetByIdProto):
(JSC::CTI::privateCompileGetByIdChain):
(JSC::CTI::privateCompilePutByIdReplace):
(JSC::CTI::privateArrayLengthTrampoline):
(JSC::CTI::privateStringLengthTrampoline):
(JSC::CTI::compileRegExp):
* VM/CTI.h:
* masm/X86Assembler.h:
(JSC::X86::):
(JSC::X86Assembler::emitModRm_rm):
(JSC::X86Assembler::emitModRm_rm_Unchecked):
(JSC::X86Assembler::emitModRm_rmsib):
* wrec/WREC.cpp:
(JSC::WRECGenerator::generateNonGreedyQuantifier):
(JSC::WRECGenerator::generateGreedyQuantifier):
(JSC::WRECGenerator::generateParentheses):
(JSC::WRECGenerator::generateBackreference):
(JSC::WRECGenerator::gernerateDisjunction):
* wrec/WREC.h:
2008-09-09 Sam Weinig <sam@webkit.org>
Reviewed by Geoffrey Garen.
Remove unnecessary friend declaration.
* kjs/PropertyMap.h:
2008-09-09 Sam Weinig <sam@webkit.org>
Reviewed by Geoffrey Garen.
Replace uses of PropertyMap::get and PropertyMap::getLocation with
PropertyMap::getOffset.
Sunspider reports this as a .6% improvement.
* JavaScriptCore.exp:
* kjs/JSObject.cpp:
(JSC::JSObject::put):
(JSC::JSObject::deleteProperty):
(JSC::JSObject::getPropertyAttributes):
* kjs/JSObject.h:
(JSC::JSObject::getDirect):
(JSC::JSObject::getDirectLocation):
(JSC::JSObject::locationForOffset):
* kjs/PropertyMap.cpp:
(JSC::PropertyMap::remove):
(JSC::PropertyMap::getOffset):
* kjs/PropertyMap.h:
2008-09-09 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Sam Weinig.
Bug 20754: Remove emit prefix from assembler opcode methods
<https://bugs.webkit.org/show_bug.cgi?id=20754>
* VM/CTI.cpp:
(JSC::CTI::emitGetArg):
(JSC::CTI::emitGetPutArg):
(JSC::CTI::emitPutArg):
(JSC::CTI::emitPutArgConstant):
(JSC::CTI::emitPutCTIParam):
(JSC::CTI::emitGetCTIParam):
(JSC::CTI::emitPutToCallFrameHeader):
(JSC::CTI::emitGetFromCallFrameHeader):
(JSC::CTI::emitPutResult):
(JSC::CTI::emitDebugExceptionCheck):
(JSC::CTI::emitCall):
(JSC::CTI::emitJumpSlowCaseIfNotImm):
(JSC::CTI::emitJumpSlowCaseIfNotImms):
(JSC::CTI::emitFastArithDeTagImmediate):
(JSC::CTI::emitFastArithReTagImmediate):
(JSC::CTI::emitFastArithPotentiallyReTagImmediate):
(JSC::CTI::emitFastArithImmToInt):
(JSC::CTI::emitFastArithIntToImmOrSlowCase):
(JSC::CTI::emitFastArithIntToImmNoCheck):
(JSC::CTI::compileOpCall):
(JSC::CTI::emitSlowScriptCheck):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
(JSC::CTI::privateCompile):
(JSC::CTI::privateCompileGetByIdSelf):
(JSC::CTI::privateCompileGetByIdProto):
(JSC::CTI::privateCompileGetByIdChain):
(JSC::CTI::privateCompilePutByIdReplace):
(JSC::CTI::privateArrayLengthTrampoline):
(JSC::CTI::privateStringLengthTrampoline):
(JSC::CTI::compileRegExp):
* masm/MacroAssemblerWin.cpp:
(JSC::MacroAssembler::emitConvertToFastCall):
(JSC::MacroAssembler::emitRestoreArgumentReference):
* masm/X86Assembler.h:
(JSC::X86Assembler::pushl_r):
(JSC::X86Assembler::pushl_m):
(JSC::X86Assembler::popl_r):
(JSC::X86Assembler::popl_m):
(JSC::X86Assembler::movl_rr):
(JSC::X86Assembler::addl_rr):
(JSC::X86Assembler::addl_i8r):
(JSC::X86Assembler::addl_i32r):
(JSC::X86Assembler::addl_mr):
(JSC::X86Assembler::andl_rr):
(JSC::X86Assembler::andl_i32r):
(JSC::X86Assembler::cmpl_i8r):
(JSC::X86Assembler::cmpl_rr):
(JSC::X86Assembler::cmpl_rm):
(JSC::X86Assembler::cmpl_i32r):
(JSC::X86Assembler::cmpl_i32m):
(JSC::X86Assembler::cmpw_rm):
(JSC::X86Assembler::orl_rr):
(JSC::X86Assembler::subl_rr):
(JSC::X86Assembler::subl_i8r):
(JSC::X86Assembler::subl_i32r):
(JSC::X86Assembler::subl_mr):
(JSC::X86Assembler::testl_i32r):
(JSC::X86Assembler::testl_rr):
(JSC::X86Assembler::xorl_i8r):
(JSC::X86Assembler::xorl_rr):
(JSC::X86Assembler::sarl_i8r):
(JSC::X86Assembler::sarl_CLr):
(JSC::X86Assembler::shl_i8r):
(JSC::X86Assembler::shll_CLr):
(JSC::X86Assembler::mull_rr):
(JSC::X86Assembler::idivl_r):
(JSC::X86Assembler::cdq):
(JSC::X86Assembler::movl_mr):
(JSC::X86Assembler::movzwl_mr):
(JSC::X86Assembler::movl_rm):
(JSC::X86Assembler::movl_i32r):
(JSC::X86Assembler::movl_i32m):
(JSC::X86Assembler::leal_mr):
(JSC::X86Assembler::ret):
(JSC::X86Assembler::jmp_r):
(JSC::X86Assembler::jmp_m):
(JSC::X86Assembler::call_r):
* wrec/WREC.cpp:
(JSC::WRECGenerator::generateBacktrack1):
(JSC::WRECGenerator::generateBacktrackBackreference):
(JSC::WRECGenerator::generateBackreferenceQuantifier):
(JSC::WRECGenerator::generateNonGreedyQuantifier):
(JSC::WRECGenerator::generateGreedyQuantifier):
(JSC::WRECGenerator::generatePatternCharacter):
(JSC::WRECGenerator::generateCharacterClassInvertedRange):
(JSC::WRECGenerator::generateCharacterClassInverted):
(JSC::WRECGenerator::generateCharacterClass):
(JSC::WRECGenerator::generateParentheses):
(JSC::WRECGenerator::gererateParenthesesResetTrampoline):
(JSC::WRECGenerator::generateAssertionBOL):
(JSC::WRECGenerator::generateAssertionEOL):
(JSC::WRECGenerator::generateAssertionWordBoundary):
(JSC::WRECGenerator::generateBackreference):
(JSC::WRECGenerator::gernerateDisjunction):
2008-09-09 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Clean up the WREC code some more.
* VM/CTI.cpp:
(JSC::CTI::compileRegExp):
* wrec/WREC.cpp:
(JSC::getCharacterClassNewline):
(JSC::getCharacterClassDigits):
(JSC::getCharacterClassSpaces):
(JSC::getCharacterClassWordchar):
(JSC::getCharacterClassNondigits):
(JSC::getCharacterClassNonspaces):
(JSC::getCharacterClassNonwordchar):
(JSC::WRECGenerator::generateBacktrack1):
(JSC::WRECGenerator::generateBacktrackBackreference):
(JSC::WRECGenerator::generateBackreferenceQuantifier):
(JSC::WRECGenerator::generateNonGreedyQuantifier):
(JSC::WRECGenerator::generateGreedyQuantifier):
(JSC::WRECGenerator::generatePatternCharacter):
(JSC::WRECGenerator::generateCharacterClassInvertedRange):
(JSC::WRECGenerator::generateCharacterClassInverted):
(JSC::WRECGenerator::generateCharacterClass):
(JSC::WRECGenerator::generateParentheses):
(JSC::WRECGenerator::gererateParenthesesResetTrampoline):
(JSC::WRECGenerator::generateAssertionBOL):
(JSC::WRECGenerator::generateAssertionEOL):
(JSC::WRECGenerator::generateAssertionWordBoundary):
(JSC::WRECGenerator::generateBackreference):
(JSC::WRECGenerator::gernerateDisjunction):
(JSC::WRECParser::parseCharacterClass):
(JSC::WRECParser::parseEscape):
(JSC::WRECParser::parseTerm):
* wrec/WREC.h:
2008-09-09 Mark Rowe <mrowe@apple.com>
Build fix, rubber-stamped by Anders Carlsson.
Silence spurious build warnings about missing format attributes on functions in Assertions.cpp.
* JavaScriptCore.xcodeproj/project.pbxproj:
2008-09-09 Mark Rowe <mrowe@apple.com>
Rubber-stamped by Oliver Hunt.
Fix builds using the "debug" variant.
This reverts r36130 and tweaks Identifier to export the same symbols for Debug
and Release configurations.
* Configurations/JavaScriptCore.xcconfig:
* DerivedSources.make:
* JavaScriptCore.Debug.exp: Removed.
* JavaScriptCore.base.exp: Removed.
* JavaScriptCore.exp: Added.
* JavaScriptCore.xcodeproj/project.pbxproj:
* kjs/identifier.cpp:
(JSC::Identifier::addSlowCase): #ifdef the call to checkSameIdentifierTable so that
there is no overhead in Release builds.
(JSC::Identifier::checkSameIdentifierTable): Add empty functions for Release builds.
* kjs/identifier.h:
(JSC::Identifier::add): #ifdef the calls to checkSameIdentifierTable so that there is
no overhead in Release builds, and remove the inline definitions of checkSameIdentifierTable.
2008-09-09 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Clean up WREC a bit to bring it closer to our coding style guidelines.
* wrec/WREC.cpp:
(JSC::):
(JSC::getCharacterClass_newline):
(JSC::getCharacterClass_d):
(JSC::getCharacterClass_s):
(JSC::getCharacterClass_w):
(JSC::getCharacterClass_D):
(JSC::getCharacterClass_S):
(JSC::getCharacterClass_W):
(JSC::CharacterClassConstructor::append):
(JSC::WRECGenerator::generateNonGreedyQuantifier):
(JSC::WRECGenerator::generateGreedyQuantifier):
(JSC::WRECGenerator::generateCharacterClassInverted):
(JSC::WRECParser::parseQuantifier):
(JSC::WRECParser::parsePatternCharacterQualifier):
(JSC::WRECParser::parseCharacterClassQuantifier):
(JSC::WRECParser::parseBackreferenceQuantifier):
* wrec/WREC.h:
(JSC::Quantifier::):
(JSC::Quantifier::Quantifier):
2008-09-09 Jungshik Shin <jungshik.shin@gmail.com>
Reviewed by Alexey Proskuryakov.
Try MIME charset names before trying IANA names
( https://bugs.webkit.org/show_bug.cgi?id=17537 )
* wtf/StringExtras.h: (strcasecmp): Added.
2008-09-09 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Mark Rowe.
Bug 20719: REGRESSION (r36135-36244): Hangs, then crashes after several seconds
<https://bugs.webkit.org/show_bug.cgi?id=20719>
<rdar://problem/6205787>
Fix a typo in the case-insensitive matching of character patterns.
* wrec/WREC.cpp:
(JSC::WRECGenerator::generatePatternCharacter):
2008-09-09 Maciej Stachowiak <mjs@apple.com>
Reviewed by Sam Weinig.
- allow polymorphic inline cache to handle Math object functions and possibly other similar things
1.012x speedup on SunSpider.
* kjs/MathObject.cpp:
(JSC::MathObject::getOwnPropertySlot):
* kjs/lookup.cpp:
(JSC::setUpStaticFunctionSlot):
* kjs/lookup.h:
(JSC::getStaticPropertySlot):
2008-09-08 Sam Weinig <sam@webkit.org>
Reviewed by Maciej Stachowiak and Oliver Hunt.
Split storage of properties out of the PropertyMap and into the JSObject
to allow sharing PropertyMap on the StructureID. In order to get this
function correctly, the StructureID's transition mappings were changed to
transition based on property name and attribute pairs, instead of just
property name.
- Removes the single property optimization now that the PropertyMap is shared.
This will be replaced by in-lining some values on the JSObject.
This is a wash on Sunspider and a 6.7% win on the v8 test suite.
* JavaScriptCore.base.exp:
* VM/CTI.cpp:
(JSC::CTI::privateCompileGetByIdSelf): Get the storage directly off the JSObject.
(JSC::CTI::privateCompileGetByIdProto): Ditto.
(JSC::CTI::privateCompileGetByIdChain): Ditto.
(JSC::CTI::privateCompilePutByIdReplace): Ditto.
* kjs/JSObject.cpp:
(JSC::JSObject::mark): Mark the PropertyStorage.
(JSC::JSObject::put): Update to get the propertyMap of the StructureID.
(JSC::JSObject::deleteProperty): Ditto.
(JSC::JSObject::defineGetter): Return early if the property is already a getter/setter.
(JSC::JSObject::defineSetter): Ditto.
(JSC::JSObject::getPropertyAttributes): Update to get the propertyMap of the StructureID
(JSC::JSObject::getPropertyNames): Ditto.
(JSC::JSObject::removeDirect): Ditto.
* kjs/JSObject.h: Remove PropertyMap and add PropertyStorage.
(JSC::JSObject::propertyStorage): return the PropertyStorage.
(JSC::JSObject::getDirect): Update to get the propertyMap of the StructureID.
(JSC::JSObject::getDirectLocation): Ditto.
(JSC::JSObject::offsetForLocation): Compute location directly.
(JSC::JSObject::hasCustomProperties): Update to get the propertyMap of the StructureID.
(JSC::JSObject::hasGetterSetterProperties): Ditto.
(JSC::JSObject::getDirectOffset): Get by indexing into PropertyStorage.
(JSC::JSObject::putDirectOffset): Put by indexing into PropertyStorage.
(JSC::JSObject::getOwnPropertySlotForWrite): Update to get the propertyMap of the StructureID.
(JSC::JSObject::getOwnPropertySlot): Ditto.
(JSC::JSObject::putDirect): Move putting into the StructureID unless the property already exists.
* kjs/PropertyMap.cpp: Use the propertyStorage as the storage for the JSValues.
(JSC::PropertyMap::checkConsistency):
(JSC::PropertyMap::operator=):
(JSC::PropertyMap::~PropertyMap):
(JSC::PropertyMap::get):
(JSC::PropertyMap::getLocation):
(JSC::PropertyMap::put):
(JSC::PropertyMap::getOffset):
(JSC::PropertyMap::insert):
(JSC::PropertyMap::expand):
(JSC::PropertyMap::rehash):
(JSC::PropertyMap::createTable):
(JSC::PropertyMap::resizePropertyStorage): Resize the storage to match the size of the map
(JSC::PropertyMap::remove):
(JSC::PropertyMap::getEnumerablePropertyNames):
* kjs/PropertyMap.h:
(JSC::PropertyMapEntry::PropertyMapEntry):
(JSC::PropertyMap::isEmpty):
(JSC::PropertyMap::size):
(JSC::PropertyMap::makingCount):
(JSC::PropertyMap::PropertyMap):
* kjs/StructureID.cpp:
(JSC::StructureID::addPropertyTransition): Transitions now are based off the property name
and attributes.
(JSC::StructureID::toDictionaryTransition): Copy the map.
(JSC::StructureID::changePrototypeTransition): Copy the map.
(JSC::StructureID::getterSetterTransition): Copy the map.
(JSC::StructureID::~StructureID):
* kjs/StructureID.h:
(JSC::TransitionTableHash::hash): Custom hash for transition map.
(JSC::TransitionTableHash::equal): Ditto.
(JSC::TransitionTableHashTraits::emptyValue): Custom traits for transition map
(JSC::TransitionTableHashTraits::constructDeletedValue): Ditto.
(JSC::TransitionTableHashTraits::isDeletedValue): Ditto.
(JSC::StructureID::propertyMap): Added.
2008-09-08 Oliver Hunt <oliver@apple.com>
Reviewed by Mark Rowe.
Bug 20694: Slow Script error pops up when running Dromaeo tests
Correct error in timeout logic where execution tick count would
be reset to incorrect value due to incorrect offset and indirection.
Codegen for the slow script dialog was factored out into a separate
method (emitSlowScriptCheck) rather than having multiple copies of
the same code. Also added calls to generate slow script checks
for loop_if_less and loop_if_true opcodes.
* VM/CTI.cpp:
(JSC::CTI::emitSlowScriptCheck):
(JSC::CTI::privateCompileMainPass):
(JSC::CTI::privateCompileSlowCases):
* VM/CTI.h:
2008-09-08 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Remove references to the removed WRECompiler class.
* VM/Machine.h:
* wrec/WREC.h:
2008-09-08 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Rubber-stamped by Mark Rowe.
Fix the build with CTI enabled but WREC disabled.
* VM/CTI.cpp:
* VM/CTI.h:
2008-09-08 Dan Bernstein <mitz@apple.com>
- build fix
* kjs/nodes.h:
(JSC::StatementNode::):
(JSC::BlockNode::):
2008-09-08 Kevin McCullough <kmccullough@apple.com>
Reviewed by Geoff.
<rdar://problem/6134407> Breakpoints in for loops, while loops or
conditions without curly braces don't break. (19306)
-Statement Lists already emit debug hooks but conditionals without
brackets are not lists.
* kjs/nodes.cpp:
(KJS::IfNode::emitCode):
(KJS::IfElseNode::emitCode):
(KJS::DoWhileNode::emitCode):
(KJS::WhileNode::emitCode):
(KJS::ForNode::emitCode):
(KJS::ForInNode::emitCode):
* kjs/nodes.h:
(KJS::StatementNode::):
(KJS::BlockNode::):
2008-09-08 Maciej Stachowiak <mjs@apple.com>
Reviewed by Anders Carlsson.
- Cache the code generated for eval to speed up SunSpider and web sites
https://bugs.webkit.org/show_bug.cgi?id=20718
1.052x on SunSpider
2.29x on date-format-tofte
Lots of real sites seem to get many hits on this cache as well,
including GMail, Google Spreadsheets, Slate and Digg (the last of
these gets over 100 hits on initial page load).
* VM/CodeBlock.h:
(JSC::EvalCodeCache::get):
* VM/Machine.cpp:
(JSC::Machine::callEval):
(JSC::Machine::privateExecute):
(JSC::Machine::cti_op_call_eval):
* VM/Machine.h:
2008-09-07 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Oliver Hunt.
Bug 20711: Change KJS prefix on preprocessor macros to JSC
<https://bugs.webkit.org/show_bug.cgi?id=20711>
* kjs/CommonIdentifiers.cpp:
(JSC::CommonIdentifiers::CommonIdentifiers):
* kjs/CommonIdentifiers.h:
* kjs/PropertySlot.h:
(JSC::PropertySlot::getValue):
(JSC::PropertySlot::putValue):
(JSC::PropertySlot::setValueSlot):
(JSC::PropertySlot::setValue):
(JSC::PropertySlot::setRegisterSlot):
* kjs/lookup.h:
* kjs/nodes.cpp:
* kjs/nodes.h:
(JSC::Node::):
(JSC::ExpressionNode::):
(JSC::StatementNode::):
(JSC::NullNode::):
(JSC::BooleanNode::):
(JSC::NumberNode::):
(JSC::ImmediateNumberNode::):
(JSC::StringNode::):
(JSC::RegExpNode::):
(JSC::ThisNode::):
(JSC::ResolveNode::):
(JSC::ElementNode::):
(JSC::ArrayNode::):
(JSC::PropertyNode::):
(JSC::PropertyListNode::):
(JSC::ObjectLiteralNode::):
(JSC::BracketAccessorNode::):
(JSC::DotAccessorNode::):
(JSC::ArgumentListNode::):
(JSC::ArgumentsNode::):
(JSC::NewExprNode::):
(JSC::EvalFunctionCallNode::):
(JSC::FunctionCallValueNode::):
(JSC::FunctionCallResolveNode::):
(JSC::FunctionCallBracketNode::):
(JSC::FunctionCallDotNode::):
(JSC::PrePostResolveNode::):
(JSC::PostfixResolveNode::):
(JSC::PostfixBracketNode::):
(JSC::PostfixDotNode::):
(JSC::PostfixErrorNode::):
(JSC::DeleteResolveNode::):
(JSC::DeleteBracketNode::):
(JSC::DeleteDotNode::):
(JSC::DeleteValueNode::):
(JSC::VoidNode::):
(JSC::TypeOfResolveNode::):
(JSC::TypeOfValueNode::):
(JSC::PrefixResolveNode::):
(JSC::PrefixBracketNode::):
(JSC::PrefixDotNode::):
(JSC::PrefixErrorNode::):
(JSC::UnaryPlusNode::):
(JSC::NegateNode::):
(JSC::BitwiseNotNode::):
(JSC::LogicalNotNode::):
(JSC::MultNode::):
(JSC::DivNode::):
(JSC::ModNode::):
(JSC::AddNode::):
(JSC::SubNode::):
(JSC::LeftShiftNode::):
(JSC::RightShiftNode::):
(JSC::UnsignedRightShiftNode::):
(JSC::LessNode::):
(JSC::GreaterNode::):
(JSC::LessEqNode::):
(JSC::GreaterEqNode::):
(JSC::ThrowableBinaryOpNode::):
(JSC::InstanceOfNode::):
(JSC::InNode::):
(JSC::EqualNode::):
(JSC::NotEqualNode::):
(JSC::StrictEqualNode::):
(JSC::NotStrictEqualNode::):
(JSC::BitAndNode::):
(JSC::BitOrNode::):
(JSC::BitXOrNode::):
(JSC::LogicalOpNode::):
(JSC::ConditionalNode::):
(JSC::ReadModifyResolveNode::):
(JSC::AssignResolveNode::):
(JSC::ReadModifyBracketNode::):
(JSC::AssignBracketNode::):
(JSC::AssignDotNode::):
(JSC::ReadModifyDotNode::):
(JSC::AssignErrorNode::):
(JSC::CommaNode::):
(JSC::VarDeclCommaNode::):
(JSC::ConstDeclNode::):
(JSC::ConstStatementNode::):
(JSC::EmptyStatementNode::):
(JSC::DebuggerStatementNode::):
(JSC::ExprStatementNode::):
(JSC::VarStatementNode::):
(JSC::IfNode::):
(JSC::IfElseNode::):
(JSC::DoWhileNode::):
(JSC::WhileNode::):
(JSC::ForNode::):
(JSC::ContinueNode::):
(JSC::BreakNode::):
(JSC::ReturnNode::):
(JSC::WithNode::):
(JSC::LabelNode::):
(JSC::ThrowNode::):
(JSC::TryNode::):
(JSC::ParameterNode::):
(JSC::ScopeNode::):
(JSC::ProgramNode::):
(JSC::EvalNode::):
(JSC::FunctionBodyNode::):
(JSC::FuncExprNode::):
(JSC::FuncDeclNode::):
(JSC::CaseClauseNode::):
(JSC::ClauseListNode::):
(JSC::CaseBlockNode::):
(JSC::SwitchNode::):
2008-09-07 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Bug 20704: Replace the KJS namespace
<https://bugs.webkit.org/show_bug.cgi?id=20704>
Rename the KJS namespace to JSC. There are still some uses of KJS in
preprocessor macros and comments, but these will also be changed some
time in the near future.
* API/APICast.h:
(toJS):
(toRef):
(toGlobalRef):
* API/JSBase.cpp:
* API/JSCallbackConstructor.cpp:
* API/JSCallbackConstructor.h:
* API/JSCallbackFunction.cpp:
* API/JSCallbackFunction.h:
* API/JSCallbackObject.cpp:
* API/JSCallbackObject.h:
* API/JSCallbackObjectFunctions.h:
* API/JSClassRef.cpp:
(OpaqueJSClass::staticValues):
(OpaqueJSClass::staticFunctions):
* API/JSClassRef.h:
* API/JSContextRef.cpp:
* API/JSObjectRef.cpp:
* API/JSProfilerPrivate.cpp:
* API/JSStringRef.cpp:
* API/JSValueRef.cpp:
(JSValueGetType):
* API/OpaqueJSString.cpp:
* API/OpaqueJSString.h:
* JavaScriptCore.Debug.exp:
* JavaScriptCore.base.exp:
* VM/CTI.cpp:
(JSC::):
* VM/CTI.h:
* VM/CodeBlock.cpp:
* VM/CodeBlock.h:
* VM/CodeGenerator.cpp:
* VM/CodeGenerator.h:
* VM/ExceptionHelpers.cpp:
* VM/ExceptionHelpers.h:
* VM/Instruction.h:
* VM/JSPropertyNameIterator.cpp:
* VM/JSPropertyNameIterator.h:
* VM/LabelID.h:
* VM/Machine.cpp:
* VM/Machine.h:
* VM/Opcode.cpp:
* VM/Opcode.h:
* VM/Register.h:
(WTF::):
* VM/RegisterFile.cpp:
* VM/RegisterFile.h:
* VM/RegisterID.h:
(WTF::):
* VM/SamplingTool.cpp:
* VM/SamplingTool.h:
* VM/SegmentedVector.h:
* kjs/ArgList.cpp:
* kjs/ArgList.h:
* kjs/Arguments.cpp:
* kjs/Arguments.h:
* kjs/ArrayConstructor.cpp:
* kjs/ArrayConstructor.h:
* kjs/ArrayPrototype.cpp:
* kjs/ArrayPrototype.h:
* kjs/BatchedTransitionOptimizer.h:
* kjs/BooleanConstructor.cpp:
* kjs/BooleanConstructor.h:
* kjs/BooleanObject.cpp:
* kjs/BooleanObject.h:
* kjs/BooleanPrototype.cpp:
* kjs/BooleanPrototype.h:
* kjs/CallData.cpp:
* kjs/CallData.h:
* kjs/ClassInfo.h:
* kjs/CommonIdentifiers.cpp:
* kjs/CommonIdentifiers.h:
* kjs/ConstructData.cpp:
* kjs/ConstructData.h:
* kjs/DateConstructor.cpp:
* kjs/DateConstructor.h:
* kjs/DateInstance.cpp:
(JSC::DateInstance::msToGregorianDateTime):
* kjs/DateInstance.h:
* kjs/DateMath.cpp:
* kjs/DateMath.h:
* kjs/DatePrototype.cpp:
* kjs/DatePrototype.h:
* kjs/DebuggerCallFrame.cpp:
* kjs/DebuggerCallFrame.h:
* kjs/Error.cpp:
* kjs/Error.h:
* kjs/ErrorConstructor.cpp:
* kjs/ErrorConstructor.h:
* kjs/ErrorInstance.cpp:
* kjs/ErrorInstance.h:
* kjs/ErrorPrototype.cpp:
* kjs/ErrorPrototype.h:
* kjs/ExecState.cpp:
* kjs/ExecState.h:
* kjs/FunctionConstructor.cpp:
* kjs/FunctionConstructor.h:
* kjs/FunctionPrototype.cpp:
* kjs/FunctionPrototype.h:
* kjs/GetterSetter.cpp:
* kjs/GetterSetter.h:
* kjs/GlobalEvalFunction.cpp:
* kjs/GlobalEvalFunction.h:
* kjs/IndexToNameMap.cpp:
* kjs/IndexToNameMap.h:
* kjs/InitializeThreading.cpp:
* kjs/InitializeThreading.h:
* kjs/InternalFunction.cpp:
* kjs/InternalFunction.h:
(JSC::InternalFunction::InternalFunction):
* kjs/JSActivation.cpp:
* kjs/JSActivation.h:
* kjs/JSArray.cpp:
* kjs/JSArray.h:
* kjs/JSCell.cpp:
* kjs/JSCell.h:
* kjs/JSFunction.cpp:
* kjs/JSFunction.h:
(JSC::JSFunction::JSFunction):
* kjs/JSGlobalData.cpp:
(JSC::JSGlobalData::JSGlobalData):
* kjs/JSGlobalData.h:
* kjs/JSGlobalObject.cpp:
* kjs/JSGlobalObject.h:
* kjs/JSGlobalObjectFunctions.cpp:
* kjs/JSGlobalObjectFunctions.h:
* kjs/JSImmediate.cpp:
* kjs/JSImmediate.h:
* kjs/JSLock.cpp:
* kjs/JSLock.h:
* kjs/JSNotAnObject.cpp:
* kjs/JSNotAnObject.h:
* kjs/JSNumberCell.cpp:
* kjs/JSNumberCell.h:
* kjs/JSObject.cpp:
* kjs/JSObject.h:
* kjs/JSStaticScopeObject.cpp:
* kjs/JSStaticScopeObject.h:
* kjs/JSString.cpp:
* kjs/JSString.h:
* kjs/JSType.h:
* kjs/JSValue.cpp:
* kjs/JSValue.h:
* kjs/JSVariableObject.cpp:
* kjs/JSVariableObject.h:
* kjs/JSWrapperObject.cpp:
* kjs/JSWrapperObject.h:
* kjs/LabelStack.cpp:
* kjs/LabelStack.h:
* kjs/MathObject.cpp:
* kjs/MathObject.h:
* kjs/NativeErrorConstructor.cpp:
* kjs/NativeErrorConstructor.h:
* kjs/NativeErrorPrototype.cpp:
* kjs/NativeErrorPrototype.h:
* kjs/NodeInfo.h:
* kjs/NumberConstructor.cpp:
* kjs/NumberConstructor.h:
* kjs/NumberObject.cpp:
* kjs/NumberObject.h:
* kjs/NumberPrototype.cpp:
* kjs/NumberPrototype.h:
* kjs/ObjectConstructor.cpp:
* kjs/ObjectConstructor.h:
* kjs/ObjectPrototype.cpp:
* kjs/ObjectPrototype.h:
* kjs/Parser.cpp:
* kjs/Parser.h:
* kjs/PropertyMap.cpp:
(JSC::PropertyMapStatisticsExitLogger::~PropertyMapStatisticsExitLogger):
* kjs/PropertyMap.h:
* kjs/PropertyNameArray.cpp:
* kjs/PropertyNameArray.h:
* kjs/PropertySlot.cpp:
* kjs/PropertySlot.h:
* kjs/PrototypeFunction.cpp:
* kjs/PrototypeFunction.h:
* kjs/PutPropertySlot.h:
* kjs/RegExpConstructor.cpp:
* kjs/RegExpConstructor.h:
* kjs/RegExpObject.cpp:
* kjs/RegExpObject.h:
* kjs/RegExpPrototype.cpp:
* kjs/RegExpPrototype.h:
* kjs/ScopeChain.cpp:
* kjs/ScopeChain.h:
* kjs/ScopeChainMark.h:
* kjs/Shell.cpp:
(jscmain):
* kjs/SmallStrings.cpp:
* kjs/SmallStrings.h:
* kjs/SourceProvider.h:
* kjs/SourceRange.h:
* kjs/StringConstructor.cpp:
* kjs/StringConstructor.h:
* kjs/StringObject.cpp:
* kjs/StringObject.h:
* kjs/StringObjectThatMasqueradesAsUndefined.h:
* kjs/StringPrototype.cpp:
* kjs/StringPrototype.h:
* kjs/StructureID.cpp:
* kjs/StructureID.h:
* kjs/SymbolTable.h:
* kjs/collector.cpp:
* kjs/collector.h:
* kjs/completion.h:
* kjs/create_hash_table:
* kjs/debugger.cpp:
* kjs/debugger.h:
* kjs/dtoa.cpp:
* kjs/dtoa.h:
* kjs/grammar.y:
* kjs/identifier.cpp:
* kjs/identifier.h:
(JSC::Identifier::equal):
* kjs/interpreter.cpp:
* kjs/interpreter.h:
* kjs/lexer.cpp:
(JSC::Lexer::Lexer):
(JSC::Lexer::clear):
(JSC::Lexer::makeIdentifier):
* kjs/lexer.h:
* kjs/lookup.cpp:
* kjs/lookup.h:
* kjs/nodes.cpp:
* kjs/nodes.h:
* kjs/nodes2string.cpp:
* kjs/operations.cpp:
* kjs/operations.h:
* kjs/protect.h:
* kjs/regexp.cpp:
* kjs/regexp.h:
* kjs/ustring.cpp:
* kjs/ustring.h:
(JSC::operator!=):
(JSC::IdentifierRepHash::hash):
(WTF::):
* masm/MacroAssembler.h:
* masm/MacroAssemblerWin.cpp:
* masm/X86Assembler.h:
* pcre/pcre_exec.cpp:
* profiler/CallIdentifier.h:
(WTF::):
* profiler/HeavyProfile.cpp:
* profiler/HeavyProfile.h:
* profiler/Profile.cpp:
* profiler/Profile.h:
* profiler/ProfileGenerator.cpp:
* profiler/ProfileGenerator.h:
* profiler/ProfileNode.cpp:
* profiler/ProfileNode.h:
* profiler/Profiler.cpp:
* profiler/Profiler.h:
* profiler/TreeProfile.cpp:
* profiler/TreeProfile.h:
* wrec/WREC.cpp:
* wrec/WREC.h:
* wtf/AVLTree.h:
2008-09-07 Maciej Stachowiak <mjs@apple.com>
Reviewed by Dan Bernstein.
- rename IA32MacroAssembler class to X86Assembler
We otherwise call the platform X86, and also, I don't see any macros.
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* masm/IA32MacroAsm.h: Removed.
* masm/MacroAssembler.h:
(KJS::MacroAssembler::MacroAssembler):
* masm/MacroAssemblerWin.cpp:
(KJS::MacroAssembler::emitRestoreArgumentReference):
* masm/X86Assembler.h: Copied from masm/IA32MacroAsm.h.
(KJS::X86Assembler::X86Assembler):
* wrec/WREC.cpp:
(KJS::WRECGenerator::generateNonGreedyQuantifier):
(KJS::WRECGenerator::generateGreedyQuantifier):
(KJS::WRECGenerator::generateParentheses):
(KJS::WRECGenerator::generateBackreference):
(KJS::WRECGenerator::gernerateDisjunction):
* wrec/WREC.h:
2008-09-07 Cameron Zwarich <cwzwarich@webkit.org>
Not reviewed.
Visual C++ seems to have some odd casting rules, so just convert the
offending cast back to a C-style cast for now.
* kjs/collector.cpp:
(KJS::otherThreadStackPointer):
2008-09-07 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Mark Rowe.
Attempt to fix the Windows build by using a const_cast to cast regs.Esp
to a uintptr_t instead of a reinterpret_cast.
* kjs/collector.cpp:
(KJS::otherThreadStackPointer):
2008-09-07 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Sam Weinig.
Remove C-style casts from kjs/collector.cpp.
* kjs/collector.cpp:
(KJS::Heap::heapAllocate):
(KJS::currentThreadStackBase):
(KJS::Heap::markConservatively):
(KJS::otherThreadStackPointer):
(KJS::Heap::markOtherThreadConservatively):
(KJS::Heap::sweep):
2008-09-07 Mark Rowe <mrowe@apple.com>
Build fix for the debug variant.
* DerivedSources.make: Also use the .Debug.exp exports file when building the debug variant.
2008-09-07 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Timothy Hatcher.
Remove C-style casts from the CTI code.
* VM/CTI.cpp:
(KJS::CTI::emitGetArg):
(KJS::CTI::emitGetPutArg):
(KJS::ctiRepatchCallByReturnAddress):
(KJS::CTI::compileOpCall):
(KJS::CTI::privateCompileMainPass):
(KJS::CTI::privateCompileGetByIdSelf):
(KJS::CTI::privateCompileGetByIdProto):
(KJS::CTI::privateCompileGetByIdChain):
(KJS::CTI::privateCompilePutByIdReplace):
(KJS::CTI::privateArrayLengthTrampoline):
(KJS::CTI::privateStringLengthTrampoline):
=== End merge of squirrelfish-extreme ===
2008-09-06 Gavin Barraclough <barraclough@apple.com>
Reviewed by Sam Weinig. Adapted somewhat by Maciej Stachowiak.
- refactor WREC to share more of the JIT infrastructure with CTI
* VM/CTI.cpp:
(KJS::CTI::emitGetArg):
(KJS::CTI::emitGetPutArg):
(KJS::CTI::emitPutArg):
(KJS::CTI::emitPutArgConstant):
(KJS::CTI::emitPutCTIParam):
(KJS::CTI::emitGetCTIParam):
(KJS::CTI::emitPutToCallFrameHeader):
(KJS::CTI::emitGetFromCallFrameHeader):
(KJS::CTI::emitPutResult):
(KJS::CTI::emitDebugExceptionCheck):
(KJS::CTI::emitJumpSlowCaseIfNotImm):
(KJS::CTI::emitJumpSlowCaseIfNotImms):
(KJS::CTI::emitFastArithDeTagImmediate):
(KJS::CTI::emitFastArithReTagImmediate):
(KJS::CTI::emitFastArithPotentiallyReTagImmediate):
(KJS::CTI::emitFastArithImmToInt):
(KJS::CTI::emitFastArithIntToImmOrSlowCase):
(KJS::CTI::emitFastArithIntToImmNoCheck):
(KJS::CTI::CTI):
(KJS::CTI::compileOpCall):
(KJS::CTI::privateCompileMainPass):
(KJS::CTI::privateCompileSlowCases):
(KJS::CTI::privateCompile):
(KJS::CTI::privateCompileGetByIdSelf):
(KJS::CTI::privateCompileGetByIdProto):
(KJS::CTI::privateCompileGetByIdChain):
(KJS::CTI::privateCompilePutByIdReplace):
(KJS::CTI::privateArrayLengthTrampoline):
(KJS::CTI::privateStringLengthTrampoline):
(KJS::CTI::compileRegExp):
* VM/CTI.h:
(KJS::CallRecord::CallRecord):
(KJS::JmpTable::JmpTable):
(KJS::SlowCaseEntry::SlowCaseEntry):
(KJS::CTI::JSRInfo::JSRInfo):
* kjs/regexp.cpp:
(KJS::RegExp::RegExp):
* wrec/WREC.cpp:
(KJS::GenerateParenthesesNonGreedyFunctor::GenerateParenthesesNonGreedyFunctor):
(KJS::GeneratePatternCharacterFunctor::generateAtom):
(KJS::GeneratePatternCharacterFunctor::backtrack):
(KJS::GenerateCharacterClassFunctor::generateAtom):
(KJS::GenerateCharacterClassFunctor::backtrack):
(KJS::GenerateBackreferenceFunctor::generateAtom):
(KJS::GenerateBackreferenceFunctor::backtrack):
(KJS::GenerateParenthesesNonGreedyFunctor::generateAtom):
(KJS::GenerateParenthesesNonGreedyFunctor::backtrack):
(KJS::WRECGenerate::generateBacktrack1):
(KJS::WRECGenerate::generateBacktrackBackreference):
(KJS::WRECGenerate::generateBackreferenceQuantifier):
(KJS::WRECGenerate::generateNonGreedyQuantifier):
(KJS::WRECGenerate::generateGreedyQuantifier):
(KJS::WRECGenerate::generatePatternCharacter):
(KJS::WRECGenerate::generateCharacterClassInvertedRange):
(KJS::WRECGenerate::generateCharacterClassInverted):
(KJS::WRECGenerate::generateCharacterClass):
(KJS::WRECGenerate::generateParentheses):
(KJS::WRECGenerate::generateParenthesesNonGreedy):
(KJS::WRECGenerate::gererateParenthesesResetTrampoline):
(KJS::WRECGenerate::generateAssertionBOL):
(KJS::WRECGenerate::generateAssertionEOL):
(KJS::WRECGenerate::generateAssertionWordBoundary):
(KJS::WRECGenerate::generateBackreference):
(KJS::WRECGenerate::gernerateDisjunction):
(KJS::WRECGenerate::terminateDisjunction):
(KJS::WRECParser::parseGreedyQuantifier):
(KJS::WRECParser::parseQuantifier):
(KJS::WRECParser::parsePatternCharacterQualifier):
(KJS::WRECParser::parseCharacterClassQuantifier):
(KJS::WRECParser::parseBackreferenceQuantifier):
(KJS::WRECParser::parseParentheses):
(KJS::WRECParser::parseCharacterClass):
(KJS::WRECParser::parseOctalEscape):
(KJS::WRECParser::parseEscape):
(KJS::WRECParser::parseTerm):
(KJS::WRECParser::parseDisjunction):
* wrec/WREC.h:
(KJS::WRECGenerate::WRECGenerate):
(KJS::WRECParser::):
(KJS::WRECParser::WRECParser):
(KJS::WRECParser::parseAlternative):
(KJS::WRECParser::isEndOfPattern):
2008-09-06 Oliver Hunt <oliver@apple.com>
Reviewed by NOBODY (Build fix).
Fix the sampler build.
* VM/SamplingTool.h:
2008-09-06 Oliver Hunt <oliver@apple.com>
Reviewed by Maciej Stachowiak.
Jump through the necessary hoops required to make MSVC cooperate with SFX
We now explicitly declare the calling convention on all cti_op_* cfunctions,
and return int instead of bool where appropriate (despite the cdecl calling
convention seems to state MSVC generates code that returns the result value
through ecx). SFX behaves slightly differently under MSVC, specifically it
stores the base argument address for the cti_op_* functions in the first
argument, and then does the required stack manipulation through that pointer.
This is necessary as MSVC's optimisations assume they have complete control
of the stack, and periodically elide our stack manipulations, or move
values in unexpected ways. MSVC also frequently produces tail calls which may
clobber the first argument, so the MSVC path is slightly less efficient due
to the need to restore it.
* JavaScriptCore.xcodeproj/project.pbxproj:
* VM/CTI.cpp:
(KJS::):
(KJS::CTI::compileOpCall):
(KJS::CTI::privateCompileMainPass):
(KJS::CTI::privateCompileSlowCases):
* VM/CTI.h:
* VM/Machine.cpp:
* VM/Machine.h:
* masm/MacroAssembler.h:
(KJS::MacroAssembler::emitConvertToFastCall):
* masm/MacroAssemblerIA32GCC.cpp: Removed.
For performance reasons we need these no-op functions to be inlined.
* masm/MacroAssemblerWin.cpp:
(KJS::MacroAssembler::emitRestoreArgumentReference):
* wtf/Platform.h:
2008-09-05 Geoffrey Garen <ggaren@apple.com>
Reviewed by Maciej Stachowiak, or maybe the other way around.
Added the ability to coalesce JITCode buffer grow operations by first
growing the buffer and then executing unchecked puts to it.
About a 2% speedup on date-format-tofte.
* VM/CTI.cpp:
(KJS::CTI::compileOpCall):
* masm/IA32MacroAsm.h:
(KJS::JITCodeBuffer::ensureSpace):
(KJS::JITCodeBuffer::putByteUnchecked):
(KJS::JITCodeBuffer::putByte):
(KJS::JITCodeBuffer::putShortUnchecked):
(KJS::JITCodeBuffer::putShort):
(KJS::JITCodeBuffer::putIntUnchecked):
(KJS::JITCodeBuffer::putInt):
(KJS::IA32MacroAssembler::emitTestl_i32r):
(KJS::IA32MacroAssembler::emitMovl_mr):
(KJS::IA32MacroAssembler::emitMovl_rm):
(KJS::IA32MacroAssembler::emitMovl_i32m):
(KJS::IA32MacroAssembler::emitUnlinkedJe):
(KJS::IA32MacroAssembler::emitModRm_rr):
(KJS::IA32MacroAssembler::emitModRm_rr_Unchecked):
(KJS::IA32MacroAssembler::emitModRm_rm_Unchecked):
(KJS::IA32MacroAssembler::emitModRm_rm):
(KJS::IA32MacroAssembler::emitModRm_opr):
(KJS::IA32MacroAssembler::emitModRm_opr_Unchecked):
(KJS::IA32MacroAssembler::emitModRm_opm_Unchecked):
2008-09-05 Mark Rowe <mrowe@apple.com>
Reviewed by Sam Weinig.
Disable WREC and CTI on platforms that we have not yet had a chance to test with.
* wtf/Platform.h:
2008-09-05 Geoffrey Garen <ggaren@apple.com>
Reviewed by Sam Weinig.
Use jo instead of a mask compare when fetching array.length and
string.length. 4% speedup on array.length / string.length torture
test.
* VM/CTI.cpp:
(KJS::CTI::privateArrayLengthTrampoline):
(KJS::CTI::privateStringLengthTrampoline):
2008-09-05 Geoffrey Garen <ggaren@apple.com>
Reviewed by Sam Weinig.
Removed a CTI compilation pass by recording labels during bytecode
generation. This is more to reduce complexity than it is to improve
performance.
SunSpider reports no change.
CodeBlock now keeps a "labels" set, which holds the offsets of all the
instructions that can be jumped to.
* VM/CTI.cpp: Nixed a pass.
* VM/CodeBlock.h: Added a "labels" set.
* VM/LabelID.h: No need for a special LableID for holding jump
destinations, since the CodeBlock now knows all jump destinations.
* wtf/HashTraits.h: New hash traits to accomodate putting offset 0 in
the set.
* kjs/nodes.cpp:
(KJS::TryNode::emitCode): Emit a dummy label to record sret targets.
2008-09-05 Mark Rowe <mrowe@apple.com>
Reviewed by Oliver Hunt and Gavin Barraclough.
Move the JITCodeBuffer onto Machine and remove the static variables.
* VM/CTI.cpp: Initialize m_jit with the Machine's code buffer.
* VM/Machine.cpp:
(KJS::Machine::Machine): Allocate a JITCodeBuffer.
* VM/Machine.h:
* kjs/RegExpConstructor.cpp:
(KJS::constructRegExp): Pass the ExecState through.
* kjs/RegExpPrototype.cpp:
(KJS::regExpProtoFuncCompile): Ditto.
* kjs/StringPrototype.cpp:
(KJS::stringProtoFuncMatch): Ditto.
(KJS::stringProtoFuncSearch): Ditto.
* kjs/nodes.cpp:
(KJS::RegExpNode::emitCode): Compile the pattern at code generation time
so that we have access to an ExecState.
* kjs/nodes.h:
(KJS::RegExpNode::):
* kjs/nodes2string.cpp:
* kjs/regexp.cpp:
(KJS::RegExp::RegExp): Pass the ExecState through.
(KJS::RegExp::create): Ditto.
* kjs/regexp.h:
* masm/IA32MacroAsm.h:
(KJS::IA32MacroAssembler::IA32MacroAssembler): Reset the JITCodeBuffer when we are
constructed.
* wrec/WREC.cpp:
(KJS::WRECompiler::compile): Retrieve the JITCodeBuffer from the Machine.
* wrec/WREC.h:
2008-09-05 Mark Rowe <mrowe@apple.com>
Reviewed by Oliver Hunt and Gavin Barraclough.
Fix the build when CTI is disabled.
* VM/CodeBlock.cpp:
(KJS::CodeBlock::~CodeBlock):
* VM/CodeGenerator.cpp:
(KJS::prepareJumpTableForStringSwitch):
* VM/Machine.cpp:
(KJS::Machine::Machine):
(KJS::Machine::~Machine):
2008-09-05 Gavin Barraclough <barraclough@apple.com>
Reviewed by Mark Rowe.
Fix some windows abi issues.
* VM/CTI.cpp:
(KJS::CTI::privateCompileMainPass):
(KJS::CTI::privateCompileSlowCases):
* VM/CTI.h:
(KJS::CallRecord::CallRecord):
(KJS::):
* VM/Machine.cpp:
(KJS::Machine::cti_op_resolve_func):
(KJS::Machine::cti_op_post_inc):
(KJS::Machine::cti_op_resolve_with_base):
(KJS::Machine::cti_op_post_dec):
* VM/Machine.h:
2008-09-05 Mark Rowe <mrowe@apple.com>
Reviewed by Sam Weinig.
Fix ecma/FunctionObjects/15.3.5.3.js after I broke it in r93.
* VM/Machine.cpp:
(KJS::Machine::cti_op_call_NotJSFunction): Restore m_callFrame to the correct value after making the native call.
(KJS::Machine::cti_op_construct_NotJSConstruct): Ditto.
2008-09-04 Mark Rowe <mrowe@apple.com>
Reviewed by Sam Weinig.
Fix fast/dom/Window/console-functions.html.
The call frame on the ExecState was not being updated on calls into native functions. This meant that functions
such as console.log would use the line number of the last JS function on the call stack.
* VM/Machine.cpp:
(KJS::Machine::cti_op_call_NotJSFunction): Update the ExecState's call frame before making a native function call,
and restore it when the function is done.
(KJS::Machine::cti_op_construct_NotJSConstruct): Ditto.
2008-09-05 Oliver Hunt <oliver@apple.com>
Start bringing up SFX on windows.
Reviewed by Mark Rowe and Sam Weinig
Start doing the work to bring up SFX on windows. Initially
just working on WREC, as it does not make any calls so reduces
the amount of code that needs to be corrected.
Start abstracting the CTI JIT codegen engine.
* ChangeLog:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* VM/CTI.cpp:
* masm/IA32MacroAsm.h:
* masm/MacroAssembler.h: Added.
(KJS::MacroAssembler::MacroAssembler):
* masm/MacroAssemblerIA32GCC.cpp: Added.
(KJS::MacroAssembler::emitConvertToFastCall):
* masm/MacroAssemblerWin.cpp: Added.
(KJS::MacroAssembler::emitConvertToFastCall):
* wrec/WREC.cpp:
(KJS::WRECompiler::parseGreedyQuantifier):
(KJS::WRECompiler::parseCharacterClass):
(KJS::WRECompiler::parseEscape):
(KJS::WRECompiler::compilePattern):
* wrec/WREC.h:
2008-09-04 Gavin Barraclough <barraclough@apple.com>
Reviewed by Sam Weinig.
Support for slow scripts (timeout checking).
* VM/CTI.cpp:
(KJS::CTI::privateCompileMainPass):
(KJS::CTI::privateCompile):
* VM/Machine.cpp:
(KJS::slideRegisterWindowForCall):
(KJS::Machine::cti_timeout_check):
(KJS::Machine::cti_vm_throw):
2008-09-04 Sam Weinig <sam@webkit.org>
Reviewed by Mark Rowe.
Third round of style cleanup.
* VM/CTI.cpp:
* VM/CTI.h:
* VM/CodeBlock.h:
* VM/Machine.cpp:
* VM/Machine.h:
* kjs/ExecState.h:
2008-09-04 Sam Weinig <sam@webkit.org>
Reviewed by Jon Honeycutt.
Second round of style cleanup.
* VM/CTI.cpp:
* VM/CTI.h:
* wrec/WREC.h:
2008-09-04 Sam Weinig <sam@webkit.org>
Reviewed by Mark Rowe.
First round of style cleanup.
* VM/CTI.cpp:
* VM/CTI.h:
* masm/IA32MacroAsm.h:
* wrec/WREC.cpp:
* wrec/WREC.h:
2008-09-04 Geoffrey Garen <ggaren@apple.com>
Reviewed by Mark Rowe.
Merged http://trac.webkit.org/changeset/36081 to work with CTI.
* VM/Machine.cpp:
(KJS::Machine::tryCtiCacheGetByID):
2008-09-04 Gavin Barraclough <barraclough@apple.com>
Reviewed by Sam Weinig.
Enable profiling in CTI.
* VM/CTI.h:
(KJS::):
(KJS::CTI::execute):
* VM/Machine.cpp:
(KJS::Machine::cti_op_call_JSFunction):
(KJS::Machine::cti_op_call_NotJSFunction):
(KJS::Machine::cti_op_ret):
(KJS::Machine::cti_op_construct_JSConstruct):
(KJS::Machine::cti_op_construct_NotJSConstruct):
2008-09-04 Victor Hernandez <vhernandez@apple.com>
Reviewed by Geoffrey Garen.
Fixed an #if to support using WREC without CTI.
* kjs/regexp.cpp:
(KJS::RegExp::match):
2008-09-04 Gavin Barraclough <barraclough@apple.com>
Reviewed by Oliver Hunt.
The array/string length trampolines are owned by the Machine, not the codeblock that compiled them.
* VM/CTI.cpp:
(KJS::CTI::privateArrayLengthTrampoline):
(KJS::CTI::privateStringLengthTrampoline):
* VM/Machine.cpp:
(KJS::Machine::~Machine):
* VM/Machine.h:
2008-09-04 Mark Rowe <mrowe@apple.com>
Reviewed by Gavin Barraclough and Sam Weinig.
Fix a crash on launch of jsc when GuardMalloc is enabled.
* kjs/ScopeChain.h:
(KJS::ScopeChain::ScopeChain): Initialize m_node to 0 when we have no valid scope chain.
(KJS::ScopeChain::~ScopeChain): Null-check m_node before calling deref.
2008-09-03 Oliver Hunt <oliver@apple.com>
Reviewed by Gavin Barraclough and Geoff Garen.
Fix inspector and fast array access so that it bounds
checks correctly.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass2_Main):
* masm/IA32MacroAsm.h:
(KJS::IA32MacroAssembler::):
(KJS::IA32MacroAssembler::emitUnlinkedJb):
(KJS::IA32MacroAssembler::emitUnlinkedJbe):
2008-09-03 Mark Rowe <mrowe@apple.com>
Move the assertion after the InitializeAndReturn block, as
that is used even when CTI is enabled.
* VM/Machine.cpp:
(KJS::Machine::privateExecute):
2008-09-03 Mark Rowe <mrowe@apple.com>
Reviewed by Sam Weinig.
Replace calls to exit with ASSERT_WITH_MESSAGE or ASSERT_NOT_REACHED.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::privateCompile_pass2_Main):
(KJS::CTI::privateCompile_pass4_SlowCases):
* VM/Machine.cpp:
(KJS::Machine::privateExecute):
(KJS::Machine::cti_vm_throw):
2008-09-03 Mark Rowe <mrowe@apple.com>
Reviewed by Sam Weinig.
Tweak JavaScriptCore to compile on non-x86 platforms. This is achieved
by wrapping more code with ENABLE(CTI), ENABLE(WREC), and PLATFORM(X86)
#if's.
* VM/CTI.cpp:
* VM/CTI.h:
* VM/CodeBlock.cpp:
(KJS::CodeBlock::printStructureIDs): Use %td as the format specifier for
printing a ptrdiff_t.
* VM/Machine.cpp:
* VM/Machine.h:
* kjs/regexp.cpp:
(KJS::RegExp::RegExp):
(KJS::RegExp::~RegExp):
(KJS::RegExp::match):
* kjs/regexp.h:
* masm/IA32MacroAsm.h:
* wrec/WREC.cpp:
* wrec/WREC.h:
* wtf/Platform.h: Only enable CTI and WREC on x86. Add an extra define to
track whether any MASM-using features are enabled.
2008-09-03 Gavin Barraclough <barraclough@apple.com>
Reviewed by Oliver Hunt.
Copy Geoff's array/string length optimization for CTI.
* VM/CTI.cpp:
(KJS::CTI::privateArrayLengthTrampoline):
(KJS::CTI::privateStringLengthTrampoline):
* VM/CTI.h:
(KJS::CTI::compileArrayLengthTrampoline):
(KJS::CTI::compileStringLengthTrampoline):
* VM/Machine.cpp:
(KJS::Machine::Machine):
(KJS::Machine::getCtiArrayLengthTrampoline):
(KJS::Machine::getCtiStringLengthTrampoline):
(KJS::Machine::tryCtiCacheGetByID):
(KJS::Machine::cti_op_get_by_id_second):
* VM/Machine.h:
* kjs/JSString.h:
* kjs/ustring.h:
2008-09-03 Gavin Barraclough <barraclough@apple.com>
Reviewed by Oliver Hunt.
Implement fast array accesses in CTI - 2-3% progression on sunspider.
* VM/CTI.cpp:
(KJS::CTI::emitFastArithIntToImmNoCheck):
(KJS::CTI::compileOpCall):
(KJS::CTI::privateCompile_pass2_Main):
(KJS::CTI::privateCompile_pass4_SlowCases):
* VM/CTI.h:
* kjs/JSArray.h:
2008-09-02 Gavin Barraclough <barraclough@apple.com>
Reviewed by Oliver Hunt.
Enable fast property access support in CTI.
* VM/CTI.cpp:
(KJS::ctiSetReturnAddress):
(KJS::ctiRepatchCallByReturnAddress):
(KJS::CTI::privateCompile_pass2_Main):
(KJS::CTI::privateCompile):
(KJS::CTI::privateCompileGetByIdSelf):
(KJS::CTI::privateCompileGetByIdProto):
(KJS::CTI::privateCompileGetByIdChain):
(KJS::CTI::privateCompilePutByIdReplace):
* VM/CTI.h:
(KJS::CTI::compileGetByIdSelf):
(KJS::CTI::compileGetByIdProto):
(KJS::CTI::compileGetByIdChain):
(KJS::CTI::compilePutByIdReplace):
* VM/CodeBlock.cpp:
(KJS::CodeBlock::~CodeBlock):
* VM/CodeBlock.h:
* VM/Machine.cpp:
(KJS::doSetReturnAddressVmThrowTrampoline):
(KJS::Machine::tryCtiCachePutByID):
(KJS::Machine::tryCtiCacheGetByID):
(KJS::Machine::cti_op_put_by_id):
(KJS::Machine::cti_op_put_by_id_second):
(KJS::Machine::cti_op_put_by_id_generic):
(KJS::Machine::cti_op_put_by_id_fail):
(KJS::Machine::cti_op_get_by_id):
(KJS::Machine::cti_op_get_by_id_second):
(KJS::Machine::cti_op_get_by_id_generic):
(KJS::Machine::cti_op_get_by_id_fail):
(KJS::Machine::cti_op_throw):
(KJS::Machine::cti_vm_throw):
* VM/Machine.h:
* kjs/JSCell.h:
* kjs/JSObject.h:
* kjs/PropertyMap.h:
* kjs/StructureID.cpp:
(KJS::StructureIDChain::StructureIDChain):
* masm/IA32MacroAsm.h:
(KJS::IA32MacroAssembler::emitCmpl_i32m):
(KJS::IA32MacroAssembler::emitMovl_mr):
(KJS::IA32MacroAssembler::emitMovl_rm):
2008-09-02 Sam Weinig <sam@webkit.org>
Reviewed by Gavin Barraclough and Mark Rowe.
A backslash (\) at the of a RegEx should produce an error.
Fixes fast/regex/test1.html.
* wrec/WREC.cpp:
(KJS::WRECompiler::parseEscape):
2008-09-02 Sam Weinig <sam@webkit.org>
Reviewed by Geoff Garen.
Link jumps for the slow case of op_loop_if_less. Fixes acid3.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass4_SlowCases):
2008-09-01 Sam Weinig <sam@webkit.org>
Rubber-stamped by Maciej Stachowiak.
Switch WREC on by default.
* wtf/Platform.h:
2008-09-01 Sam Weinig <sam@webkit.org>
Reviewed by Mark Rowe.
Fix two failures in fast/regex/test1.html
- \- in a character class should be treated as a literal -
- A missing max quantifier needs to be treated differently than
a null max quantifier.
* wrec/WREC.cpp:
(KJS::WRECompiler::generateNonGreedyQuantifier):
(KJS::WRECompiler::generateGreedyQuantifier):
(KJS::WRECompiler::parseCharacterClass):
* wrec/WREC.h:
(KJS::Quantifier::Quantifier):
2008-09-01 Sam Weinig <sam@webkit.org>
Reviewed by Mark Rowe.
Fix crash in fast/js/kde/evil-n.html
* kjs/regexp.cpp: Always pass a non-null offset vector to the wrec function.
2008-09-01 Sam Weinig <sam@webkit.org>
Reviewed by Gavin Barraclough and Mark Rowe.
Add pattern length limit fixing one test in fast/js.
* wrec/WREC.cpp:
(KJS::WRECompiler::compile):
* wrec/WREC.h:
(KJS::WRECompiler::):
2008-09-01 Sam Weinig <sam@webkit.org>
Reviewed by Gavin Barraclough and Mark Rowe.
Make octal escape parsing/back-reference parsing more closely match
prior behavior fixing one test in fast/js.
* wrec/WREC.cpp:
(KJS::WRECompiler::parseCharacterClass): 8 and 9 should be IdentityEscaped
(KJS::WRECompiler::parseEscape):
* wrec/WREC.h:
(KJS::WRECompiler::peekDigit):
2008-09-01 Sam Weinig <sam@webkit.org>
Reviewed by Gavin Barraclough and Mark Rowe.
Fix one mozilla test.
* wrec/WREC.cpp:
(KJS::WRECompiler::generateCharacterClassInverted): Fix incorrect not
ascii upper check.
2008-09-01 Sam Weinig <sam@webkit.org>
Reviewed by Gavin Barraclough and Mark Rowe.
Parse octal escapes in character classes fixing one mozilla test.
* wrec/WREC.cpp:
(KJS::WRECompiler::parseCharacterClass):
(KJS::WRECompiler::parseOctalEscape):
* wrec/WREC.h:
(KJS::WRECompiler::consumeOctal):
2008-09-01 Sam Weinig <sam@webkit.org>
Reviewed by Oliver Hunt.
Fixes two mozilla tests with WREC enabled.
* wrec/WREC.cpp:
(KJS::CharacterClassConstructor::append): Keep the character class sorted
when appending another character class.
2008-09-01 Sam Weinig <sam@webkit.org>
Reviewed by Gavin Barraclough and Mark Rowe.
Fixes two mozilla tests with WREC enabled.
* wrec/WREC.cpp:
(KJS::CharacterClassConstructor::addSortedRange): Insert the range at the correct position
instead of appending it to the end.
2008-09-01 Gavin Barraclough <barraclough@apple.com>
Reviewed by Oliver Hunt.
Move cross-compilation unit call into NEVER_INLINE function.
* VM/Machine.cpp:
(KJS::doSetReturnAddressVmThrowTrampoline):
2008-09-01 Sam Weinig <sam@webkit.org>
Reviewed by Gavin Barraclough and Geoff Garen.
Fix one test in fast/js.
* VM/Machine.cpp:
(KJS::Machine::cti_op_construct_NotJSConstruct): Throw a createNotAConstructorError,
instead of a createNotAFunctionError.
2008-08-31 Gavin Barraclough <barraclough@apple.com>
Reviewed by Maciej Stachowiak.
Zero-cost exception handling. This patch takes the exception checking
back of the hot path. When an exception occurs in a Machine::cti*
method, the return address to JIT code is recorded, and is then
overwritten with a pointer to a trampoline routine. When the method
returns the trampoline will cause the cti_vm_throw method to be invoked.
cti_vm_throw uses the return address preserved above, to discover the
vPC of the bytecode that raised the exception (using a map build during
translation). From the VPC of the faulting bytecode the vPC of a catch
routine may be discovered (unwinding the stack where necesary), and then
a bytecode address for the catch routine is looked up. Final cti_vm_throw
overwrites its return address to JIT code again, to trampoline directly
to the catch routine.
cti_op_throw is handled in a similar fashion.
* VM/CTI.cpp:
(KJS::CTI::emitPutCTIParam):
(KJS::CTI::emitPutToCallFrameHeader):
(KJS::CTI::emitGetFromCallFrameHeader):
(KJS::ctiSetReturnAddressForArgs):
(KJS::CTI::emitDebugExceptionCheck):
(KJS::CTI::printOpcodeOperandTypes):
(KJS::CTI::emitCall):
(KJS::CTI::compileOpCall):
(KJS::CTI::privateCompile_pass2_Main):
(KJS::CTI::privateCompile):
* VM/CTI.h:
(KJS::CallRecord::CallRecord):
(KJS::):
(KJS::CTI::execute):
* VM/CodeBlock.h:
* VM/Machine.cpp:
(KJS::Machine::privateExecute):
(KJS::Machine::cti_op_instanceof):
(KJS::Machine::cti_op_call_NotJSFunction):
(KJS::Machine::cti_op_resolve):
(KJS::Machine::cti_op_resolve_func):
(KJS::Machine::cti_op_resolve_skip):
(KJS::Machine::cti_op_resolve_with_base):
(KJS::Machine::cti_op_throw):
(KJS::Machine::cti_op_in):
(KJS::Machine::cti_vm_throw):
* VM/RegisterFile.h:
(KJS::RegisterFile::):
* kjs/ExecState.h:
(KJS::ExecState::setCtiReturnAddress):
(KJS::ExecState::ctiReturnAddress):
* masm/IA32MacroAsm.h:
(KJS::IA32MacroAssembler::):
(KJS::IA32MacroAssembler::emitPushl_m):
(KJS::IA32MacroAssembler::emitPopl_m):
(KJS::IA32MacroAssembler::getRelocatedAddress):
2008-08-31 Mark Rowe <mrowe@apple.com>
Reviewed by Oliver Hunt.
Fall back to PCRE for any regexp containing parentheses until we correctly backtrack within them.
* wrec/WREC.cpp:
(KJS::WRECompiler::parseParentheses):
* wrec/WREC.h:
(KJS::WRECompiler::):
2008-08-31 Mark Rowe <mrowe@apple.com>
Reviewed by Oliver Hunt.
Fix several issues within ecma_3/RegExp/perlstress-001.js with WREC enabled.
* wrec/WREC.cpp:
(KJS::WRECompiler::generateNonGreedyQuantifier): Compare with the maximum quantifier count rather than the minimum.
(KJS::WRECompiler::generateAssertionEOL): Do a register-to-register comparison rather than immediate-to-register.
(KJS::WRECompiler::parseCharacterClass): Pass through the correct inversion flag.
2008-08-30 Mark Rowe <mrowe@apple.com>
Reviewed by Oliver Hunt.
Re-fix the six remaining failures in the Mozilla JavaScript tests in a manner that does not kill performance.
This shows up as a 0.6% progression on SunSpider on my machine.
Grow the JITCodeBuffer's underlying buffer when we run out of space rather than just bailing out.
* VM/CodeBlock.h:
(KJS::CodeBlock::~CodeBlock): Switch to using fastFree now that JITCodeBuffer::copy uses fastMalloc.
* kjs/regexp.cpp: Ditto.
* masm/IA32MacroAsm.h:
(KJS::JITCodeBuffer::growBuffer):
(KJS::JITCodeBuffer::JITCodeBuffer):
(KJS::JITCodeBuffer::~JITCodeBuffer):
(KJS::JITCodeBuffer::putByte):
(KJS::JITCodeBuffer::putShort):
(KJS::JITCodeBuffer::putInt):
(KJS::JITCodeBuffer::reset):
(KJS::JITCodeBuffer::copy):
2008-08-29 Oliver Hunt <oliver@apple.com>
RS=Maciej
Roll out previous patch as it causes a 5% performance regression
* JavaScriptCore.xcodeproj/project.pbxproj:
* VM/CTI.cpp:
(KJS::getJCB):
(KJS::CTI::privateCompile):
* VM/CodeBlock.h:
(KJS::CodeBlock::~CodeBlock):
* masm/IA32MacroAsm.h:
(KJS::JITCodeBuffer::JITCodeBuffer):
(KJS::JITCodeBuffer::putByte):
(KJS::JITCodeBuffer::putShort):
(KJS::JITCodeBuffer::putInt):
(KJS::JITCodeBuffer::getEIP):
(KJS::JITCodeBuffer::start):
(KJS::JITCodeBuffer::getOffset):
(KJS::JITCodeBuffer::reset):
(KJS::JITCodeBuffer::copy):
(KJS::IA32MacroAssembler::emitModRm_rr):
(KJS::IA32MacroAssembler::emitModRm_rm):
(KJS::IA32MacroAssembler::emitModRm_rmsib):
(KJS::IA32MacroAssembler::IA32MacroAssembler):
(KJS::IA32MacroAssembler::emitInt3):
(KJS::IA32MacroAssembler::emitPushl_r):
(KJS::IA32MacroAssembler::emitPopl_r):
(KJS::IA32MacroAssembler::emitMovl_rr):
(KJS::IA32MacroAssembler::emitAddl_rr):
(KJS::IA32MacroAssembler::emitAddl_i8r):
(KJS::IA32MacroAssembler::emitAddl_i32r):
(KJS::IA32MacroAssembler::emitAddl_mr):
(KJS::IA32MacroAssembler::emitAndl_rr):
(KJS::IA32MacroAssembler::emitAndl_i32r):
(KJS::IA32MacroAssembler::emitCmpl_i8r):
(KJS::IA32MacroAssembler::emitCmpl_rr):
(KJS::IA32MacroAssembler::emitCmpl_rm):
(KJS::IA32MacroAssembler::emitCmpl_i32r):
(KJS::IA32MacroAssembler::emitCmpl_i32m):
(KJS::IA32MacroAssembler::emitCmpw_rm):
(KJS::IA32MacroAssembler::emitOrl_rr):
(KJS::IA32MacroAssembler::emitOrl_i8r):
(KJS::IA32MacroAssembler::emitSubl_rr):
(KJS::IA32MacroAssembler::emitSubl_i8r):
(KJS::IA32MacroAssembler::emitSubl_i32r):
(KJS::IA32MacroAssembler::emitSubl_mr):
(KJS::IA32MacroAssembler::emitTestl_i32r):
(KJS::IA32MacroAssembler::emitTestl_rr):
(KJS::IA32MacroAssembler::emitXorl_i8r):
(KJS::IA32MacroAssembler::emitXorl_rr):
(KJS::IA32MacroAssembler::emitSarl_i8r):
(KJS::IA32MacroAssembler::emitSarl_CLr):
(KJS::IA32MacroAssembler::emitShl_i8r):
(KJS::IA32MacroAssembler::emitShll_CLr):
(KJS::IA32MacroAssembler::emitMull_rr):
(KJS::IA32MacroAssembler::emitIdivl_r):
(KJS::IA32MacroAssembler::emitCdq):
(KJS::IA32MacroAssembler::emitMovl_mr):
(KJS::IA32MacroAssembler::emitMovzwl_mr):
(KJS::IA32MacroAssembler::emitMovl_rm):
(KJS::IA32MacroAssembler::emitMovl_i32r):
(KJS::IA32MacroAssembler::emitMovl_i32m):
(KJS::IA32MacroAssembler::emitLeal_mr):
(KJS::IA32MacroAssembler::emitRet):
(KJS::IA32MacroAssembler::emitJmpN_r):
(KJS::IA32MacroAssembler::emitJmpN_m):
(KJS::IA32MacroAssembler::emitCall):
(KJS::IA32MacroAssembler::label):
(KJS::IA32MacroAssembler::emitUnlinkedJmp):
(KJS::IA32MacroAssembler::emitUnlinkedJne):
(KJS::IA32MacroAssembler::emitUnlinkedJe):
(KJS::IA32MacroAssembler::emitUnlinkedJl):
(KJS::IA32MacroAssembler::emitUnlinkedJle):
(KJS::IA32MacroAssembler::emitUnlinkedJge):
(KJS::IA32MacroAssembler::emitUnlinkedJae):
(KJS::IA32MacroAssembler::emitUnlinkedJo):
(KJS::IA32MacroAssembler::link):
* wrec/WREC.cpp:
(KJS::WRECompiler::compilePattern):
(KJS::WRECompiler::compile):
* wrec/WREC.h:
2008-08-29 Mark Rowe <mrowe@apple.com>
Reviewed by Oliver Hunt.
Have JITCodeBuffer manage a Vector containing the generated code so that it can grow
as needed when generating code for a large function. This fixes all six remaining failures
in Mozilla tests in both debug and release builds.
* VM/CTI.cpp:
(KJS::CTI::privateCompile):
* VM/CodeBlock.h:
(KJS::CodeBlock::~CodeBlock):
* masm/IA32MacroAsm.h:
(KJS::JITCodeBuffer::putByte):
(KJS::JITCodeBuffer::putShort):
(KJS::JITCodeBuffer::putInt):
(KJS::JITCodeBuffer::getEIP):
(KJS::JITCodeBuffer::start):
(KJS::JITCodeBuffer::getOffset):
(KJS::JITCodeBuffer::getCode):
(KJS::IA32MacroAssembler::emitModRm_rr):
* wrec/WREC.cpp:
(KJS::WRECompiler::compilePattern):
* wrec/WREC.h:
2008-08-29 Mark Rowe <mrowe@apple.com>
Reviewed by Oliver Hunt.
Implement parsing of octal escapes in regular expressions. This fixes three Mozilla tests.
* wrec/WREC.cpp:
(KJS::WRECompiler::parseOctalEscape):
(KJS::WRECompiler::parseEscape): Parse the escape sequence as an octal escape if it has a leading zero.
Add a FIXME about treating invalid backreferences as octal escapes in the future.
* wrec/WREC.h:
(KJS::WRECompiler::consumeNumber): Multiply by 10 rather than 0 so that we handle numbers with more than
one digit.
* wtf/ASCIICType.h:
(WTF::isASCIIOctalDigit):
2008-08-29 Sam Weinig <sam@webkit.org>
Reviewed by Mark Rowe.
Pass vPC to instanceof method. Fixes 2 mozilla tests in debug.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass2_Main):
* VM/Machine.cpp:
(KJS::Machine::cti_op_instanceof):
2008-08-29 Sam Weinig <sam@webkit.org>
Reviewed by Mark Rowe.
Pass vPCs to resolve methods for correct exception creation. Fixes
17 mozilla tests in debug.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass2_Main):
* VM/CTI.h:
* VM/Machine.cpp:
(KJS::Machine::cti_op_resolve):
(KJS::Machine::cti_op_resolve_func):
(KJS::Machine::cti_op_resolve_skip):
(KJS::Machine::cti_op_resolve_with_base):
2008-08-29 Gavin Barraclough <barraclough@apple.com>
Reviewed by Oliver Hunt.
Remembering to actually throw the exception passed to op throw helps.
Regressions 19 -> 6.
* VM/Machine.cpp:
(KJS::Machine::cti_op_throw):
(KJS::Machine::cti_vm_throw):
2008-08-29 Gavin Barraclough <barraclough@apple.com>
Reviewed by Sam Weinig.
Support for exception unwinding the stack.
Once upon a time, Sam asked me for a bettr ChangeLog entry. The return address
is now preserved on entry to a JIT code function (if we preserve lazily we need
restore the native return address during exception stack unwind). This takes
the number of regressions down from ~150 to 19.
* VM/CTI.cpp:
(KJS::getJCB):
(KJS::CTI::emitExceptionCheck):
(KJS::CTI::compileOpCall):
(KJS::CTI::privateCompile_pass2_Main):
(KJS::CTI::privateCompile):
* VM/CTI.h:
(KJS::):
* VM/Machine.cpp:
(KJS::Machine::throwException):
(KJS::Machine::cti_op_call_JSFunction):
(KJS::Machine::cti_op_call_NotJSFunction):
(KJS::Machine::cti_op_construct_JSConstruct):
(KJS::Machine::cti_op_construct_NotJSConstruct):
(KJS::Machine::cti_op_throw):
(KJS::Machine::cti_vm_throw):
2008-08-29 Mark Rowe <mrowe@apple.com>
Reviewed by Oliver Hunt.
Fix js1_2/regexp/word_boundary.js and four other Mozilla tests with WREC enabled.
* wrec/WREC.cpp:
(KJS::WRECompiler::generateCharacterClassInvertedRange): If none of the exact matches
succeeded, jump to failure.
(KJS::WRECompiler::compilePattern): Restore and increment the current position stored
on the stack to ensure that it will be reset to the correct position after a failed
match has consumed input.
2008-08-29 Mark Rowe <mrowe@apple.com>
Reviewed by Oliver Hunt.
Fix a hang in ecma_3/RegExp/15.10.2-1.js with WREC enabled.
A backreference with a quantifier would get stuck in an infinite
loop if the captured range was empty.
* wrec/WREC.cpp:
(KJS::WRECompiler::generateBackreferenceQuantifier): If the captured range
was empty, do not attempt to match the backreference.
(KJS::WRECompiler::parseBackreferenceQuantifier):
* wrec/WREC.h:
(KJS::Quantifier::):
2008-08-28 Sam Weinig <sam@webkit.org>
Reviewed by Oliver Hunt.
Implement op_debug.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::privateCompile_pass2_Main):
* VM/Machine.cpp:
(KJS::Machine::debug):
(KJS::Machine::privateExecute):
(KJS::Machine::cti_op_debug):
* VM/Machine.h:
2008-08-28 Sam Weinig <sam@webkit.org>
Reviewed by Gavin Barraclough and Geoff Garen.
Implement op_switch_string fixing 1 mozilla test and one test in fast/js.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::privateCompile_pass2_Main):
(KJS::CTI::privateCompile):
* VM/CTI.h:
(KJS::SwitchRecord::):
(KJS::SwitchRecord::SwitchRecord):
* VM/CodeBlock.cpp:
(KJS::CodeBlock::dump):
* VM/CodeBlock.h:
(KJS::ExpressionRangeInfo::):
(KJS::StringJumpTable::offsetForValue):
(KJS::StringJumpTable::ctiForValue):
(KJS::SimpleJumpTable::add):
(KJS::SimpleJumpTable::ctiForValue):
* VM/CodeGenerator.cpp:
(KJS::prepareJumpTableForStringSwitch):
* VM/Machine.cpp:
(KJS::Machine::privateExecute):
(KJS::Machine::cti_op_switch_string):
* VM/Machine.h:
2008-08-28 Gavin Barraclough <barraclough@apple.com>
Reviewed by Oliver Hunt.
Do not recurse on the machine stack when executing op_call.
* VM/CTI.cpp:
(KJS::CTI::emitGetPutArg):
(KJS::CTI::emitPutArg):
(KJS::CTI::emitPutArgConstant):
(KJS::CTI::compileOpCall):
(KJS::CTI::privateCompile_pass2_Main):
(KJS::CTI::privateCompile):
* VM/CTI.h:
(KJS::):
(KJS::CTI::compile):
(KJS::CTI::execute):
(KJS::CTI::):
* VM/Machine.cpp:
(KJS::Machine::Machine):
(KJS::Machine::execute):
(KJS::Machine::cti_op_call_JSFunction):
(KJS::Machine::cti_op_call_NotJSFunction):
(KJS::Machine::cti_op_ret):
(KJS::Machine::cti_op_construct_JSConstruct):
(KJS::Machine::cti_op_construct_NotJSConstruct):
(KJS::Machine::cti_op_call_eval):
* VM/Machine.h:
* VM/Register.h:
(KJS::Register::Register):
* VM/RegisterFile.h:
(KJS::RegisterFile::):
* kjs/InternalFunction.h:
(KJS::InternalFunction::InternalFunction):
* kjs/JSFunction.h:
(KJS::JSFunction::JSFunction):
* kjs/ScopeChain.h:
(KJS::ScopeChain::ScopeChain):
* masm/IA32MacroAsm.h:
(KJS::IA32MacroAssembler::):
(KJS::IA32MacroAssembler::emitModRm_opm):
(KJS::IA32MacroAssembler::emitCmpl_i32m):
(KJS::IA32MacroAssembler::emitCallN_r):
2008-08-28 Sam Weinig <sam@webkit.org>
Reviewed by Mark Rowe.
Exit instead of crashing in ctiUnsupported and ctiTimedOut.
* VM/Machine.cpp:
(KJS::ctiUnsupported):
(KJS::ctiTimedOut):
2008-08-28 Oliver Hunt <oliver@apple.com>
Reviewed by Maciej Stachowiak.
Implement codegen for op_jsr and op_sret.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::privateCompile_pass2_Main):
(KJS::CTI::privateCompile):
* VM/CTI.h:
(KJS::CTI::JSRInfo::JSRInfo):
* masm/IA32MacroAsm.h:
(KJS::IA32MacroAssembler::emitJmpN_m):
(KJS::IA32MacroAssembler::linkAbsoluteAddress):
2008-08-28 Gavin Barraclough <barraclough@apple.com>
Reviewed by Oliver Hunt.
Initial support for exceptions (throw / catch must occur in same CodeBlock).
* VM/CTI.cpp:
(KJS::CTI::emitExceptionCheck):
(KJS::CTI::emitCall):
(KJS::CTI::privateCompile_pass2_Main):
(KJS::CTI::privateCompile_pass4_SlowCases):
(KJS::CTI::privateCompile):
* VM/CTI.h:
* VM/CodeBlock.cpp:
(KJS::CodeBlock::nativeExceptionCodeForHandlerVPC):
* VM/CodeBlock.h:
* VM/CodeGenerator.cpp:
(KJS::CodeGenerator::emitCatch):
* VM/Machine.cpp:
(KJS::Machine::throwException):
(KJS::Machine::privateExecute):
(KJS::ctiUnsupported):
(KJS::ctiTimedOut):
(KJS::Machine::cti_op_add):
(KJS::Machine::cti_op_pre_inc):
(KJS::Machine::cti_timeout_check):
(KJS::Machine::cti_op_loop_if_less):
(KJS::Machine::cti_op_put_by_id):
(KJS::Machine::cti_op_get_by_id):
(KJS::Machine::cti_op_instanceof):
(KJS::Machine::cti_op_del_by_id):
(KJS::Machine::cti_op_mul):
(KJS::Machine::cti_op_call):
(KJS::Machine::cti_op_resolve):
(KJS::Machine::cti_op_construct):
(KJS::Machine::cti_op_get_by_val):
(KJS::Machine::cti_op_resolve_func):
(KJS::Machine::cti_op_sub):
(KJS::Machine::cti_op_put_by_val):
(KJS::Machine::cti_op_lesseq):
(KJS::Machine::cti_op_loop_if_true):
(KJS::Machine::cti_op_negate):
(KJS::Machine::cti_op_resolve_skip):
(KJS::Machine::cti_op_div):
(KJS::Machine::cti_op_pre_dec):
(KJS::Machine::cti_op_jless):
(KJS::Machine::cti_op_not):
(KJS::Machine::cti_op_jtrue):
(KJS::Machine::cti_op_post_inc):
(KJS::Machine::cti_op_eq):
(KJS::Machine::cti_op_lshift):
(KJS::Machine::cti_op_bitand):
(KJS::Machine::cti_op_rshift):
(KJS::Machine::cti_op_bitnot):
(KJS::Machine::cti_op_resolve_with_base):
(KJS::Machine::cti_op_mod):
(KJS::Machine::cti_op_less):
(KJS::Machine::cti_op_neq):
(KJS::Machine::cti_op_post_dec):
(KJS::Machine::cti_op_urshift):
(KJS::Machine::cti_op_bitxor):
(KJS::Machine::cti_op_bitor):
(KJS::Machine::cti_op_call_eval):
(KJS::Machine::cti_op_throw):
(KJS::Machine::cti_op_push_scope):
(KJS::Machine::cti_op_stricteq):
(KJS::Machine::cti_op_nstricteq):
(KJS::Machine::cti_op_to_jsnumber):
(KJS::Machine::cti_op_in):
(KJS::Machine::cti_op_del_by_val):
(KJS::Machine::cti_vm_throw):
* VM/Machine.h:
* kjs/ExecState.h:
* masm/IA32MacroAsm.h:
(KJS::IA32MacroAssembler::emitCmpl_i32m):
2008-08-28 Mark Rowe <mrowe@apple.com>
Rubber-stamped by Oliver Hunt.
Print debugging info to stderr so that run-webkit-tests can capture it.
This makes it easy to check whether test failures are due to unimplemented
op codes, missing support for exceptions, etc.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::printOpcodeOperandTypes):
(KJS::CTI::privateCompile_pass2_Main):
(KJS::CTI::privateCompile_pass4_SlowCases):
(KJS::CTI::privateCompile):
* VM/Machine.cpp:
(KJS::Machine::privateExecute):
(KJS::ctiException):
(KJS::ctiUnsupported):
(KJS::Machine::cti_op_call):
(KJS::Machine::cti_op_resolve):
(KJS::Machine::cti_op_construct):
(KJS::Machine::cti_op_get_by_val):
(KJS::Machine::cti_op_resolve_func):
(KJS::Machine::cti_op_resolve_skip):
(KJS::Machine::cti_op_resolve_with_base):
(KJS::Machine::cti_op_call_eval):
2008-08-27 Mark Rowe <mrowe@apple.com>
Reviewed by Gavin Barraclough and Maciej Stachowiak.
Fix fast/js/bitwise-and-on-undefined.html.
A temporary value in the slow path of op_bitand was being stored in edx, but was
being clobbered by emitGetPutArg before we used it. To fix this, emitGetPutArg
now takes a third argument that specifies the scratch register to use when loading
from memory. This allows us to avoid clobbering the temporary in op_bitand.
* VM/CTI.cpp:
(KJS::CTI::emitGetPutArg):
(KJS::CTI::privateCompile_pass2_Main):
(KJS::CTI::privateCompile_pass4_SlowCases):
* VM/CTI.h:
2008-08-27 Mark Rowe <mrowe@apple.com>
Rubber-stamped by Oliver Hunt.
Switch CTI on by default.
* wtf/Platform.h:
2008-08-27 Mark Rowe <mrowe@apple.com>
Reviewed by Oliver Hunt.
Fix the build of the full WebKit stack.
* JavaScriptCore.xcodeproj/project.pbxproj: Mark two new headers as private so they can be pulled in from WebCore.
* VM/CTI.h: Fix build issues that show up when compiled with GCC 4.2 as part of WebCore.
* wrec/WREC.h: Ditto.
2008-08-27 Mark Rowe <mrowe@apple.com>
Reviewed by Sam Weinig.
Implement op_new_error. Does not fix any tests as it is always followed by the unimplemented op_throw.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::privateCompile_pass2_Main):
* VM/Machine.cpp:
(KJS::Machine::cti_op_new_error):
* VM/Machine.h:
2008-08-27 Sam Weinig <sam@webkit.org>
Reviewed by Gavin Barraclough and Geoff Garen.
Implement op_put_getter and op_put_setter.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::privateCompile_pass2_Main):
* VM/Machine.cpp:
(KJS::Machine::cti_op_put_getter):
(KJS::Machine::cti_op_put_setter):
* VM/Machine.h:
2008-08-27 Sam Weinig <sam@webkit.org>
Reviewed by Gavin Barraclough and Geoff Garen.
Implement op_del_by_val fixing 3 mozilla tests.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::privateCompile_pass2_Main):
* VM/Machine.cpp:
(KJS::Machine::cti_op_del_by_val):
* VM/Machine.h:
2008-08-27 Gavin Barraclough <barraclough@apple.com>
Reviewed by Oliver Hunt.
Quick & dirty fix to get SamplingTool sampling op_call.
* VM/SamplingTool.h:
(KJS::SamplingTool::callingHostFunction):
2008-08-27 Sam Weinig <sam@webkit.org>
Reviewed by Gavin Barraclough and Geoff Garen.
Fix op_put_by_index.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass2_Main): Use emitPutArgConstant instead of emitGetPutArg
for the property value.
* VM/Machine.cpp:
(KJS::Machine::cti_op_put_by_index): Get the property value from the correct argument.
2008-08-27 Sam Weinig <sam@webkit.org>
Reviewed by Gavin Barraclough and Geoff Garen.
Implement op_switch_imm in the CTI fixing 13 mozilla tests.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::privateCompile_pass2_Main):
* VM/Machine.cpp:
(KJS::Machine::cti_op_switch_imm):
* VM/Machine.h:
2008-08-27 Gavin Barraclough <barraclough@apple.com>
Reviewed by Oliver Hunt.
Implement op_switch_char in CTI.
* VM/CTI.cpp:
(KJS::CTI::emitCall):
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::privateCompile_pass2_Main):
(KJS::CTI::privateCompile):
* VM/CTI.h:
(KJS::CallRecord::CallRecord):
(KJS::SwitchRecord::SwitchRecord):
* VM/CodeBlock.h:
(KJS::SimpleJumpTable::SimpleJumpTable::ctiForValue):
* VM/Machine.cpp:
(KJS::Machine::cti_op_switch_char):
* VM/Machine.h:
* masm/IA32MacroAsm.h:
(KJS::IA32MacroAssembler::):
(KJS::IA32MacroAssembler::emitJmpN_r):
(KJS::IA32MacroAssembler::getRelocatedAddress):
* wtf/Platform.h:
2008-08-26 Sam Weinig <sam@webkit.org>
Reviewed by Mark Rowe.
Implement op_put_by_index to fix 1 mozilla test.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::privateCompile_pass2_Main):
* VM/Machine.cpp:
(KJS::Machine::cti_op_put_by_index):
* VM/Machine.h:
2008-08-26 Gavin Barraclough <barraclough@apple.com>
Reviewed by Geoff Garen.
More fixes from Geoff's review.
* VM/CTI.cpp:
(KJS::CTI::emitGetArg):
(KJS::CTI::emitGetPutArg):
(KJS::CTI::emitPutArg):
(KJS::CTI::emitPutArgConstant):
(KJS::CTI::getConstantImmediateNumericArg):
(KJS::CTI::emitGetCTIParam):
(KJS::CTI::emitPutResult):
(KJS::CTI::emitCall):
(KJS::CTI::emitJumpSlowCaseIfNotImm):
(KJS::CTI::emitJumpSlowCaseIfNotImms):
(KJS::CTI::getDeTaggedConstantImmediate):
(KJS::CTI::emitFastArithDeTagImmediate):
(KJS::CTI::emitFastArithReTagImmediate):
(KJS::CTI::emitFastArithPotentiallyReTagImmediate):
(KJS::CTI::emitFastArithImmToInt):
(KJS::CTI::emitFastArithIntToImmOrSlowCase):
(KJS::CTI::privateCompile_pass2_Main):
(KJS::CTI::privateCompile_pass4_SlowCases):
(KJS::CTI::privateCompile):
* VM/CTI.h:
2008-08-26 Mark Rowe <mrowe@apple.com>
Reviewed by Gavin Barraclough and Geoff Garen.
Implement op_jmp_scopes to fix 2 Mozilla tests.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::privateCompile_pass2_Main):
* VM/Machine.cpp:
(KJS::Machine::cti_op_push_new_scope): Update ExecState::m_scopeChain after calling ARG_setScopeChain.
(KJS::Machine::cti_op_jmp_scopes):
* VM/Machine.h:
2008-08-26 Gavin Barraclough <barraclough@apple.com>
Reviewed by Oliver Hunt.
WebKit Regular Expression Compiler. (set ENABLE_WREC = 1 in Platform.h).
* JavaScriptCore.xcodeproj/project.pbxproj:
* kjs/regexp.cpp:
* kjs/regexp.h:
* wrec: Added.
* wrec/WREC.cpp: Added.
* wrec/WREC.h: Added.
* wtf/Platform.h:
2008-08-26 Sam Weinig <sam@webkit.org>
Rubber-stamped by Oliver Hunt.
Remove bogus assertion.
* VM/Machine.cpp:
(KJS::Machine::cti_op_del_by_id):
2008-08-26 Mark Rowe <mrowe@apple.com>
Reviewed by Sam Weinig.
Implement op_push_new_scope and stub out op_catch. This fixes 11 Mozilla tests.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::privateCompile_pass2_Main):
* VM/Machine.cpp:
(KJS::Machine::cti_op_push_new_scope):
(KJS::Machine::cti_op_catch):
* VM/Machine.h:
2008-08-26 Mark Rowe <mrowe@apple.com>
Reviewed by Sam Weinig.
Clean up op_resolve_base so that it shares its implementation with the bytecode interpreter.
* VM/Machine.cpp:
(KJS::inlineResolveBase):
(KJS::resolveBase):
2008-08-26 Oliver Hunt <oliver@apple.com>
Reviewed by Sam Weinig.
Add codegen support for op_instanceof, fixing 15 mozilla tests.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::privateCompile_pass2_Main):
* VM/Machine.cpp:
(KJS::Machine::cti_op_instanceof):
(KJS::Machine::cti_op_del_by_id):
* VM/Machine.h:
* wtf/Platform.h:
2008-08-26 Gavin Barraclough <barraclough@apple.com>
Reviewed by Geoff Garen.
Fixes for initial review comments.
* VM/CTI.cpp:
(KJS::CTI::ctiCompileGetArg):
(KJS::CTI::ctiCompileGetPutArg):
(KJS::CTI::ctiCompilePutResult):
(KJS::CTI::ctiCompileCall):
(KJS::CTI::CTI):
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::printOpcodeOperandTypes):
(KJS::CTI::privateCompile_pass2_Main):
(KJS::CTI::privateCompile_pass4_SlowCases):
(KJS::CTI::privateCompile):
* VM/CTI.h:
* VM/Register.h:
* kjs/JSValue.h:
2008-08-26 Sam Weinig <sam@webkit.org>
Reviewed by Gavin Barraclough and Geoff Garen.
Fix up exception checking code.
* VM/Machine.cpp:
(KJS::Machine::cti_op_call):
(KJS::Machine::cti_op_resolve):
(KJS::Machine::cti_op_construct):
(KJS::Machine::cti_op_resolve_func):
(KJS::Machine::cti_op_resolve_skip):
(KJS::Machine::cti_op_resolve_with_base):
(KJS::Machine::cti_op_call_eval):
2008-08-26 Sam Weinig <sam@webkit.org>
Reviewed by Oliver Hunt.
Fix slowcase for op_post_inc and op_post_dec fixing 2 mozilla tests.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass4_SlowCases):
2008-08-26 Mark Rowe <mrowe@apple.com>
Reviewed by Sam Weinig.
Implement op_in, fixing 8 mozilla tests.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::privateCompile_pass2_Main):
* VM/Machine.cpp:
(KJS::Machine::cti_op_in):
* VM/Machine.h:
2008-08-26 Mark Rowe <mrowe@apple.com>
Rubber-stamped by Oliver Hunt.
Don't hardcode the size of a Register for op_new_array. Fixes a crash
seen during the Mozilla tests.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass2_Main):
2008-08-26 Sam Weinig <sam@webkit.org>
Reviewed by Gavin Barraclough and Geoff Garen.
Add support for op_push_scope and op_pop_scope, fixing 20 mozilla tests.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::privateCompile_pass2_Main):
* VM/CTI.h:
* VM/Machine.cpp:
(KJS::Machine::cti_op_push_scope):
(KJS::Machine::cti_op_pop_scope):
* VM/Machine.h:
2008-08-26 Oliver Hunt <oliver@apple.com>
Reviewed by Maciej Stachowiak.
Add codegen support for op_del_by_id, fixing 49 mozilla tests.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::privateCompile_pass2_Main):
* VM/Machine.cpp:
(KJS::Machine::cti_op_del_by_id):
* VM/Machine.h:
2008-08-26 Sam Weinig <sam@webkit.org>
Reviewed by Gavin Barraclough and Geoff Garen.
Don't hardcode the size of a Register for op_get_scoped_var and op_put_scoped_var
fixing 513 mozilla tests in debug build.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass2_Main):
2008-08-26 Oliver Hunt <oliver@apple.com>
Reviewed by Maciej Stachowiak.
Added code generator support for op_loop, fixing around 60 mozilla tests.
* VM/CTI.cpp:
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::privateCompile_pass2_Main):
2008-08-26 Mark Rowe <mrowe@apple.com>
Reviewed by Sam Weinig.
Set -fomit-frame-pointer in the correct location.
* Configurations/JavaScriptCore.xcconfig:
* JavaScriptCore.xcodeproj/project.pbxproj:
2008-08-26 Gavin Barraclough <barraclough@apple.com>
Reviewed by Geoff Garen.
Inital cut of CTI, Geoff's review fixes to follow.
* JavaScriptCore.xcodeproj/project.pbxproj:
* VM/CTI.cpp: Added.
(KJS::getJCB):
(KJS::CTI::ctiCompileGetArg):
(KJS::CTI::ctiCompileGetPutArg):
(KJS::CTI::ctiCompilePutArg):
(KJS::CTI::ctiCompilePutArgImm):
(KJS::CTI::ctiImmediateNumericArg):
(KJS::CTI::ctiCompileGetCTIParam):
(KJS::CTI::ctiCompilePutResult):
(KJS::CTI::ctiCompileCall):
(KJS::CTI::slowCaseIfNotImm):
(KJS::CTI::slowCaseIfNotImms):
(KJS::CTI::ctiFastArithDeTagConstImmediate):
(KJS::CTI::ctiFastArithDeTagImmediate):
(KJS::CTI::ctiFastArithReTagImmediate):
(KJS::CTI::ctiFastArithPotentiallyReTagImmediate):
(KJS::CTI::ctiFastArithImmToInt):
(KJS::CTI::ctiFastArithIntToImmOrSlowCase):
(KJS::CTI::CTI):
(KJS::CTI::privateCompile_pass1_Scan):
(KJS::CTI::ctiCompileAdd):
(KJS::CTI::ctiCompileAddImm):
(KJS::CTI::ctiCompileAddImmNotInt):
(KJS::CTI::TEMP_HACK_PRINT_TYPES):
(KJS::CTI::privateCompile_pass2_Main):
(KJS::CTI::privateCompile_pass3_Link):
(KJS::CTI::privateCompile_pass4_SlowCases):
(KJS::CTI::privateCompile):
* VM/CTI.h: Added.
(KJS::CTI2Result::CTI2Result):
(KJS::CallRecord::CallRecord):
(KJS::JmpTable::JmpTable):
(KJS::SlowCaseEntry::SlowCaseEntry):
(KJS::CTI::compile):
(KJS::CTI::LabelInfo::LabelInfo):
* VM/CodeBlock.h:
(KJS::CodeBlock::CodeBlock):
(KJS::CodeBlock::~CodeBlock):
* VM/Machine.cpp:
(KJS::Machine::execute):
(KJS::Machine::privateExecute):
(KJS::ctiException):
(KJS::ctiUnsupported):
(KJS::ctiTimedOut):
(KJS::Machine::cti_op_end):
(KJS::Machine::cti_op_add):
(KJS::Machine::cti_op_pre_inc):
(KJS::Machine::cti_timeout_check):
(KJS::Machine::cti_op_loop_if_less):
(KJS::Machine::cti_op_new_object):
(KJS::Machine::cti_op_put_by_id):
(KJS::Machine::cti_op_get_by_id):
(KJS::Machine::cti_op_mul):
(KJS::Machine::cti_op_new_func):
(KJS::Machine::cti_op_call):
(KJS::Machine::cti_op_ret):
(KJS::Machine::cti_op_new_array):
(KJS::Machine::cti_op_resolve):
(KJS::Machine::cti_op_construct):
(KJS::Machine::cti_op_get_by_val):
(KJS::Machine::cti_op_resolve_func):
(KJS::Machine::cti_op_sub):
(KJS::Machine::cti_op_put_by_val):
(KJS::Machine::cti_op_lesseq):
(KJS::Machine::cti_op_loop_if_true):
(KJS::Machine::cti_op_negate):
(KJS::Machine::cti_op_resolve_base):
(KJS::Machine::cti_op_resolve_skip):
(KJS::Machine::cti_op_div):
(KJS::Machine::cti_op_pre_dec):
(KJS::Machine::cti_op_jless):
(KJS::Machine::cti_op_not):
(KJS::Machine::cti_op_jtrue):
(KJS::Machine::cti_op_post_inc):
(KJS::Machine::cti_op_eq):
(KJS::Machine::cti_op_lshift):
(KJS::Machine::cti_op_bitand):
(KJS::Machine::cti_op_rshift):
(KJS::Machine::cti_op_bitnot):
(KJS::Machine::cti_op_resolve_with_base):
(KJS::Machine::cti_op_new_func_exp):
(KJS::Machine::cti_op_mod):
(KJS::Machine::cti_op_less):
(KJS::Machine::cti_op_neq):
(KJS::Machine::cti_op_post_dec):
(KJS::Machine::cti_op_urshift):
(KJS::Machine::cti_op_bitxor):
(KJS::Machine::cti_op_new_regexp):
(KJS::Machine::cti_op_bitor):
(KJS::Machine::cti_op_call_eval):
(KJS::Machine::cti_op_throw):
(KJS::Machine::cti_op_get_pnames):
(KJS::Machine::cti_op_next_pname):
(KJS::Machine::cti_op_typeof):
(KJS::Machine::cti_op_stricteq):
(KJS::Machine::cti_op_nstricteq):
(KJS::Machine::cti_op_to_jsnumber):
* VM/Machine.h:
* VM/Register.h:
(KJS::Register::jsValue):
(KJS::Register::getJSValue):
(KJS::Register::codeBlock):
(KJS::Register::scopeChain):
(KJS::Register::i):
(KJS::Register::r):
(KJS::Register::vPC):
(KJS::Register::jsPropertyNameIterator):
* VM/SamplingTool.cpp:
(KJS::):
(KJS::SamplingTool::run):
(KJS::SamplingTool::dump):
* VM/SamplingTool.h:
* kjs/JSImmediate.h:
(KJS::JSImmediate::zeroImmediate):
(KJS::JSImmediate::oneImmediate):
* kjs/JSValue.h:
* kjs/JSVariableObject.h:
(KJS::JSVariableObject::JSVariableObjectData::offsetOf_registers):
(KJS::JSVariableObject::offsetOf_d):
(KJS::JSVariableObject::offsetOf_Data_registers):
* masm: Added.
* masm/IA32MacroAsm.h: Added.
(KJS::JITCodeBuffer::JITCodeBuffer):
(KJS::JITCodeBuffer::putByte):
(KJS::JITCodeBuffer::putShort):
(KJS::JITCodeBuffer::putInt):
(KJS::JITCodeBuffer::getEIP):
(KJS::JITCodeBuffer::start):
(KJS::JITCodeBuffer::getOffset):
(KJS::JITCodeBuffer::reset):
(KJS::JITCodeBuffer::copy):
(KJS::IA32MacroAssembler::):
(KJS::IA32MacroAssembler::emitModRm_rr):
(KJS::IA32MacroAssembler::emitModRm_rm):
(KJS::IA32MacroAssembler::emitModRm_rmsib):
(KJS::IA32MacroAssembler::emitModRm_opr):
(KJS::IA32MacroAssembler::emitModRm_opm):
(KJS::IA32MacroAssembler::IA32MacroAssembler):
(KJS::IA32MacroAssembler::emitInt3):
(KJS::IA32MacroAssembler::emitPushl_r):
(KJS::IA32MacroAssembler::emitPopl_r):
(KJS::IA32MacroAssembler::emitMovl_rr):
(KJS::IA32MacroAssembler::emitAddl_rr):
(KJS::IA32MacroAssembler::emitAddl_i8r):
(KJS::IA32MacroAssembler::emitAddl_i32r):
(KJS::IA32MacroAssembler::emitAddl_mr):
(KJS::IA32MacroAssembler::emitAndl_rr):
(KJS::IA32MacroAssembler::emitAndl_i32r):
(KJS::IA32MacroAssembler::emitCmpl_i8r):
(KJS::IA32MacroAssembler::emitCmpl_rr):
(KJS::IA32MacroAssembler::emitCmpl_rm):
(KJS::IA32MacroAssembler::emitCmpl_i32r):
(KJS::IA32MacroAssembler::emitCmpw_rm):
(KJS::IA32MacroAssembler::emitOrl_rr):
(KJS::IA32MacroAssembler::emitOrl_i8r):
(KJS::IA32MacroAssembler::emitSubl_rr):
(KJS::IA32MacroAssembler::emitSubl_i8r):
(KJS::IA32MacroAssembler::emitSubl_i32r):
(KJS::IA32MacroAssembler::emitSubl_mr):
(KJS::IA32MacroAssembler::emitTestl_i32r):
(KJS::IA32MacroAssembler::emitTestl_rr):
(KJS::IA32MacroAssembler::emitXorl_i8r):
(KJS::IA32MacroAssembler::emitXorl_rr):
(KJS::IA32MacroAssembler::emitSarl_i8r):
(KJS::IA32MacroAssembler::emitSarl_CLr):
(KJS::IA32MacroAssembler::emitShl_i8r):
(KJS::IA32MacroAssembler::emitShll_CLr):
(KJS::IA32MacroAssembler::emitMull_rr):
(KJS::IA32MacroAssembler::emitIdivl_r):
(KJS::IA32MacroAssembler::emitCdq):
(KJS::IA32MacroAssembler::emitMovl_mr):
(KJS::IA32MacroAssembler::emitMovzwl_mr):
(KJS::IA32MacroAssembler::emitMovl_rm):
(KJS::IA32MacroAssembler::emitMovl_i32r):
(KJS::IA32MacroAssembler::emitMovl_i32m):
(KJS::IA32MacroAssembler::emitLeal_mr):
(KJS::IA32MacroAssembler::emitRet):
(KJS::IA32MacroAssembler::JmpSrc::JmpSrc):
(KJS::IA32MacroAssembler::JmpDst::JmpDst):
(KJS::IA32MacroAssembler::emitCall):
(KJS::IA32MacroAssembler::label):
(KJS::IA32MacroAssembler::emitUnlinkedJmp):
(KJS::IA32MacroAssembler::emitUnlinkedJne):
(KJS::IA32MacroAssembler::emitUnlinkedJe):
(KJS::IA32MacroAssembler::emitUnlinkedJl):
(KJS::IA32MacroAssembler::emitUnlinkedJle):
(KJS::IA32MacroAssembler::emitUnlinkedJge):
(KJS::IA32MacroAssembler::emitUnlinkedJae):
(KJS::IA32MacroAssembler::emitUnlinkedJo):
(KJS::IA32MacroAssembler::emitPredictionNotTaken):
(KJS::IA32MacroAssembler::link):
(KJS::IA32MacroAssembler::copy):
* wtf/Platform.h:
2008-08-26 Oliver Hunt <oliver@apple.com>
RS=Maciej.
Enabled -fomit-frame-pointer on Release and Production builds, add additional Profiling build config for shark, etc.
* JavaScriptCore.xcodeproj/project.pbxproj:
=== Start merge of squirrelfish-extreme ===
2008-09-06 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Fix the Mac Debug build by adding symbols that are exported only in a
Debug configuration.
* Configurations/JavaScriptCore.xcconfig:
* DerivedSources.make:
* JavaScriptCore.Debug.exp: Added.
* JavaScriptCore.base.exp: Copied from JavaScriptCore.exp.
* JavaScriptCore.exp: Removed.
* JavaScriptCore.xcodeproj/project.pbxproj:
2008-09-05 Darin Adler <darin@apple.com>
Reviewed by Cameron Zwarich.
- https://bugs.webkit.org/show_bug.cgi?id=20681
JSPropertyNameIterator functions need to be inlined
1.007x as fast on SunSpider overall
1.081x as fast on SunSpider math-cordic
* VM/JSPropertyNameIterator.cpp: Moved functions out of here.
* VM/JSPropertyNameIterator.h:
(KJS::JSPropertyNameIterator::JSPropertyNameIterator): Moved
this into the header and marked it inline.
(KJS::JSPropertyNameIterator::create): Ditto.
(KJS::JSPropertyNameIterator::next): Ditto.
2008-09-05 Darin Adler <darin@apple.com>
Reviewed by Geoffrey Garen.
- fix https://bugs.webkit.org/show_bug.cgi?id=20673
single-character strings are churning in the Identifier table
1.007x as fast on SunSpider overall
1.167x as fast on SunSpider string-fasta
* JavaScriptCore.exp: Updated.
* kjs/SmallStrings.cpp:
(KJS::SmallStrings::singleCharacterStringRep): Added.
* kjs/SmallStrings.h: Added singleCharacterStringRep for clients that
need just a UString, not a JSString.
* kjs/identifier.cpp:
(KJS::Identifier::add): Added special cases for single character strings
so that the UString::Rep that ends up in the identifier table is the one
from the single-character string optimization; otherwise we end up having
to look it up in the identifier table over and over again.
(KJS::Identifier::addSlowCase): Ditto.
(KJS::Identifier::checkSameIdentifierTable): Made this function an empty
inline in release builds so that callers don't have to put #ifndef NDEBUG
at each call site.
* kjs/identifier.h:
(KJS::Identifier::add): Removed #ifndef NDEBUG around the calls to
checkSameIdentifierTable.
(KJS::Identifier::checkSameIdentifierTable): Added. Empty inline version
for NDEBUG builds.
2008-09-05 Mark Rowe <mrowe@apple.com>
Build fix.
* kjs/JSObject.h: Move the inline virtual destructor after a non-inline
virtual function so that the symbol for the vtable is not marked as a
weakly exported symbol.
2008-09-05 Darin Adler <darin@apple.com>
Reviewed by Sam Weinig.
- fix https://bugs.webkit.org/show_bug.cgi?id=20671
JavaScriptCore string manipulation spends too much time in memcpy
1.011x as fast on SunSpider overall
1.028x as fast on SunSpider string tests
For small strings, use a loop rather than calling memcpy. The loop can
be faster because there's no function call overhead, and because it can
assume the pointers are aligned instead of checking that. Currently the
threshold is set at 20 characters, based on some testing on one particular
computer. Later we can tune this for various platforms by setting
USTRING_COPY_CHARS_INLINE_CUTOFF appropriately, but it does no great harm
if not perfectly tuned.
* kjs/ustring.cpp:
(KJS::overflowIndicator): Removed bogus const.
(KJS::maxUChars): Ditto.
(KJS::copyChars): Added.
(KJS::UString::Rep::createCopying): Call copyChars instead of memcpy.
Also eliminated need for const_cast.
(KJS::UString::expandPreCapacity): Ditto.
(KJS::concatenate): Ditto.
(KJS::UString::spliceSubstringsWithSeparators): Ditto.
(KJS::UString::append): Ditto.
2008-09-05 Kevin McCullough <kmccullough@apple.com>
Reviewed by Sam and Alexey.
Make the profiler work with a null exec state. This will allow other
applications start the profiler to get DTrace probes going without
needing a WebView.
* ChangeLog:
* profiler/ProfileGenerator.cpp:
(KJS::ProfileGenerator::ProfileGenerator):
(KJS::ProfileGenerator::willExecute):
(KJS::ProfileGenerator::didExecute):
* profiler/Profiler.cpp:
(KJS::Profiler::startProfiling):
(KJS::Profiler::stopProfiling):
(KJS::dispatchFunctionToProfiles):
2008-09-04 Gavin Barraclough <barraclough@apple.com>
Reviewed by Geoffrey Garen.
Fixed an off-by-one error that would cause the StructureIDChain to
be one object too short.
Can't construct a test case because other factors make this not crash
(yet!).
* kjs/StructureID.cpp:
(KJS::StructureIDChain::StructureIDChain):
2008-09-04 Kevin Ollivier <kevino@theolliviers.com>
wx build fixes.
* JavaScriptCoreSources.bkl:
2008-09-04 Mark Rowe <mrowe@apple.com>
Reviewed by Eric Seidel.
Fix https://bugs.webkit.org/show_bug.cgi?id=20639.
Bug 20639: ENABLE_DASHBOARD_SUPPORT does not need to be a FEATURE_DEFINE
* Configurations/JavaScriptCore.xcconfig: Remove ENABLE_DASHBOARD_SUPPORT from FEATURE_DEFINES.
* wtf/Platform.h: Set ENABLE_DASHBOARD_SUPPORT for PLATFORM(MAC).
2008-09-04 Adele Peterson <adele@apple.com>
Build fix.
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.vcproj/WTF/WTF.vcproj:
* JavaScriptCore.vcproj/jsc/jsc.vcproj:
2008-09-04 Mark Rowe <mrowe@apple.com>
Mac build fix.
* kjs/config.h: Only check the value of HAVE_CONFIG_H if it is defined.
2008-09-04 Marco Barisione <marco.barisione@collabora.co.uk>
Reviewed by Eric Seidel.
http://bugs.webkit.org/show_bug.cgi?id=20380
[GTK][AUTOTOOLS] Include autotoolsconfig.h from config.h
* kjs/config.h: Include the configuration header generated by
autotools if available.
2008-09-04 Tor Arne Vestbø <tavestbo@trolltech.com>
Reviewed by Simon.
Fix the QtWebKit build to match changes in r36016
* JavaScriptCore.pri:
2008-09-04 Mark Rowe <mrowe@apple.com>
Fix the 64-bit build.
* VM/CodeBlock.cpp:
(KJS::CodeBlock::printStructureID): Store the instruction offset into an unsigned local
to avoid a warning related to format specifiers.
(KJS::CodeBlock::printStructureIDs): Ditto.
2008-09-04 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Rubber-stamped by Oliver Hunt.
Correct the spelling of 'entryIndices'.
* kjs/PropertyMap.cpp:
(KJS::PropertyMap::get):
(KJS::PropertyMap::getLocation):
(KJS::PropertyMap::put):
(KJS::PropertyMap::insert):
(KJS::PropertyMap::remove):
(KJS::PropertyMap::checkConsistency):
* kjs/PropertyMap.h:
(KJS::PropertyMapHashTable::entries):
(KJS::PropertyMap::getOffset):
(KJS::PropertyMap::putOffset):
(KJS::PropertyMap::offsetForTableLocation):
2008-09-03 Geoffrey Garen <ggaren@apple.com>
Reviewed by Cameron Zwarich.
Fixed <rdar://problem/6193925> REGRESSION: Crash occurs at
KJS::Machine::privateExecute() when attempting to load my Mobile Gallery
(http://www.me.com/gallery/#home)
also
https://bugs.webkit.org/show_bug.cgi?id=20633 Crash in privateExecute
@ cs.byu.edu
The underlying problem was that we would cache prototype properties
even if the prototype was a dictionary.
The fix is to transition a prototype back from dictionary to normal
status when an opcode caches access to it. (This is better than just
refusing to cache, since a heavily accessed prototype is almost
certainly not a true dictionary.)
* VM/Machine.cpp:
(KJS::Machine::tryCacheGetByID):
* kjs/JSObject.h:
2008-09-03 Eric Seidel <eric@webkit.org>
Reviewed by Sam.
Clean up Platform.h and add PLATFORM(CHROMIUM), PLATFORM(SKIA) and USE(V8_BINDINGS)
* Configurations/JavaScriptCore.xcconfig: add missing ENABLE_*
* wtf/ASCIICType.h: include <wtf/Assertions.h> since it depends on it.
* wtf/Platform.h:
2008-09-03 Kevin McCullough <kmccullough@apple.com>
Reviewed by Tim.
Remove the rest of the "zombie" code from the profiler.
- There is no longer a need for the ProfilerClient callback mechanism.
* API/JSProfilerPrivate.cpp:
(JSStartProfiling):
* JavaScriptCore.exp:
* profiler/HeavyProfile.h:
* profiler/ProfileGenerator.cpp:
(KJS::ProfileGenerator::create):
(KJS::ProfileGenerator::ProfileGenerator):
* profiler/ProfileGenerator.h:
(KJS::ProfileGenerator::profileGroup):
* profiler/Profiler.cpp:
(KJS::Profiler::startProfiling):
(KJS::Profiler::stopProfiling): Immediately return the profile when
stopped instead of using a callback.
* profiler/Profiler.h:
* profiler/TreeProfile.h:
2008-09-03 Adele Peterson <adele@apple.com>
Build fix.
* wtf/win/MainThreadWin.cpp:
2008-09-02 Kevin McCullough <kmccullough@apple.com>
Reviewed by Darin and Tim.
Remove most of the "zombie" mode from the profiler. Next we will need
to remove the client callback mechanism in profiles.
- This simplifies the code, leverages the recent changes I've made in
getting line numbers from SquirrelFish, and is a slight speed
improvement on SunSpider.
- Also the "zombie" mode was a constant source of odd edge cases and
obscure bugs so it's good to remove since all of its issues may not have
been found.
* API/JSProfilerPrivate.cpp: No need to call didFinishAllExecution() any
more.
(JSEndProfiling):
* JavaScriptCore.exp: Export the new signature of retrieveLastCaller()
* VM/Machine.cpp:
(KJS::Machine::execute): No need to call didFinishAllExecution() any
more.
(KJS::Machine::retrieveCaller): Now operates on InternalFunctions now
since the RegisterFile is no longer guaranteeded to store only
JSFunctions
(KJS::Machine::retrieveLastCaller): Now also retrieve the function's
name
(KJS::Machine::callFrame): A result of changing retrieveCaller()
* VM/Machine.h:
* VM/Register.h:
* kjs/JSGlobalObject.cpp:
(KJS::JSGlobalObject::~JSGlobalObject):
* kjs/nodes.h:
* profiler/ProfileGenerator.cpp:
(KJS::ProfileGenerator::create): Now pass the original exec and get the
global exec and client when necessary. We need the original exec so we
can have the stack frame where profiling started.
(KJS::ProfileGenerator::ProfileGenerator): ditto.
(KJS::ProfileGenerator::addParentForConsoleStart): This is where the
parent to star of the profile is added, if there is one.
(KJS::ProfileGenerator::willExecute): Remove uglyness!
(KJS::ProfileGenerator::didExecute): Ditto!
(KJS::ProfileGenerator::stopProfiling):
(KJS::ProfileGenerator::removeProfileStart): Use a better way to find
and remove the function we are looking for.
(KJS::ProfileGenerator::removeProfileEnd): Ditto.
* profiler/ProfileGenerator.h:
(KJS::ProfileGenerator::client):
* profiler/ProfileNode.cpp:
(KJS::ProfileNode::removeChild): Add a better way to remove a child from
a ProfileNode.
(KJS::ProfileNode::stopProfiling):
(KJS::ProfileNode::debugPrintData): Modified a debug-only diagnostic
function to be sane.
* profiler/ProfileNode.h:
* profiler/Profiler.cpp: Change to pass the original exec state.
(KJS::Profiler::startProfiling):
(KJS::Profiler::stopProfiling):
(KJS::Profiler::willExecute):
(KJS::Profiler::didExecute):
(KJS::Profiler::createCallIdentifier):
* profiler/Profiler.h:
2008-09-01 Alexey Proskuryakov <ap@webkit.org>
Reviewed by Darin Adler.
Implement callOnMainThreadAndWait().
This will be useful when a background thread needs to perform UI calls synchronously
(e.g. an openDatabase() call cannot return until the user answers to a confirmation dialog).
* wtf/MainThread.cpp:
(WTF::FunctionWithContext::FunctionWithContext): Added a ThreadCondition member. When
non-zero, the condition is signalled after the function is called.
(WTF::mainThreadFunctionQueueMutex): Renamed from functionQueueMutex, sinc this is no longer
static. Changed to be initialized from initializeThreading() to avoid lock contention.
(WTF::initializeMainThread): On non-Windows platforms, just call mainThreadFunctionQueueMutex.
(WTF::dispatchFunctionsFromMainThread): Signal synchronous calls when done.
(WTF::callOnMainThread): Updated for functionQueueMutex rename.
(WTF::callOnMainThreadAndWait): Added.
* wtf/MainThread.h: Added callOnMainThreadAndWait(); initializeMainThread() now exists on
all platforms.
* wtf/win/MainThreadWin.cpp: (WTF::initializeMainThread): Added a callOnMainThreadAndWait()
call to initialize function queue mutex.
* wtf/ThreadingGtk.cpp: (WTF::initializeThreading):
* wtf/ThreadingPthreads.cpp: (WTF::initializeThreading):
* wtf/ThreadingQt.cpp: (WTF::initializeThreading):
Only initialize mainThreadIdentifier on non-Darwin platforms. It was not guaranteed to be
accurate on Darwin.
2008-09-03 Geoffrey Garen <ggaren@apple.com>
Reviewed by Darin Adler.
Use isUndefinedOrNull() instead of separate checks for each in op_eq_null
and op_neq_null.
* VM/Machine.cpp:
(KJS::Machine::privateExecute):
2008-09-02 Csaba Osztrogonac <oszi@inf.u-szeged.hu>
Reviewed by Darin Adler.
Bug 20296: OpcodeStats doesn't build on platforms which don't have mergesort().
<https://bugs.webkit.org/show_bug.cgi?id=20296>
* VM/Opcode.cpp:
(KJS::OpcodeStats::~OpcodeStats): mergesort() replaced with qsort()
2008-09-02 Geoffrey Garen <ggaren@apple.com>
Reviewed by Oliver Hunt.
Fast path for array.length and string.length.
SunSpider says 0.5% faster.
2008-09-02 Geoffrey Garen <ggaren@apple.com>
Reviewed by Anders Carlsson.
Added optimized paths for comparing to null.
SunSpider says 0.5% faster.
2008-09-02 Geoffrey Garen <ggaren@apple.com>
Reviewed by Sam Weinig.
Changed jsDriver.pl to dump the exact text you would need in order to
reproduce a test result. This enables a fast workflow where you copy
and paste a test failure in the terminal.
* tests/mozilla/jsDriver.pl:
2008-09-02 Geoffrey Garen <ggaren@apple.com>
Reviewed by Sam Weinig.
Implemented the rest of Darin's review comments for the 09-01 inline
caching patch.
SunSpider says 0.5% faster, but that seems like noise.
* JavaScriptCore.xcodeproj/project.pbxproj: Put PutPropertySlot into
its own file, and added BatchedTransitionOptimizer.
* VM/CodeBlock.cpp:
(KJS::CodeBlock::~CodeBlock): Use array indexing instead of a pointer
iterator.
* VM/CodeGenerator.cpp:
(KJS::CodeGenerator::CodeGenerator): Used BatchedTransitionOptimizer to
make batched put and remove for declared variables fast, without forever
pessimizing the global object. Removed the old getDirect/removeDirect hack
that tried to do the same in a more limited way.
* VM/CodeGenerator.h: Moved IdentifierRepHash to the KJS namespace since
it doesn't specialize anything in WTF.
* VM/Machine.cpp:
(KJS::Machine::Machine): Nixed the DummyConstruct tag because it was
confusingly named.
(KJS::Machine::execute): Used BatchedTransitionOptimizer, as above. Fixed
up some comments.
(KJS::cachePrototypeChain): Cast to JSObject*, since it's more specific.
(KJS::Machine::tryCachePutByID): Use isNull() instead of comparing to
jsNull(), since isNull() leaves more options open for the future.
(KJS::Machine::tryCacheGetByID): ditto
(KJS::Machine::privateExecute): ditto
* VM/SamplingTool.cpp:
(KJS::SamplingTool::dump): Use C++-style cast, to match our style
guidelines.
* kjs/BatchedTransitionOptimizer.h: Added. New class that allows host
code to add a batch of properties to an object in an efficient way.
* kjs/JSActivation.cpp: Use isNull(), as above.
* kjs/JSArray.cpp: Get rid of DummyConstruct tag, as above.
* kjs/JSArray.h:
* kjs/JSGlobalData.cpp: Nixed two unused StructureIDs.
* kjs/JSGlobalData.h:
* kjs/JSImmediate.cpp: Use isNull(), as above.
* kjs/JSObject.cpp:
(KJS::JSObject::mark): Moved mark tracing code elsewhere, to make this
function more readable.
(KJS::JSObject::put): Use isNull(), as above.
(KJS::JSObject::createInheritorID): Return a raw pointer, since the
object is owned by a data member, not necessarily the caller.
* kjs/JSObject.h:
* kjs/JSString.cpp: Use isNull(), as above.
* kjs/PropertyMap.h: Updated to use PropertySlot::invalidOffset.
* kjs/PropertySlot.h: Changed KJS_INVALID_OFFSET to WTF::notFound
because C macros are so 80's.
* kjs/PutPropertySlot.h: Added. Split out of PropertySlot.h. Also renamed
PutPropertySlot::SlotType to PutPropertySlot::Type, and slotBase to base,
since "slot" was redundant.
* kjs/StructureID.cpp: Added a new transition *away* from dictionary
status, to support BatchedTransitionOptimizer.
(KJS::StructureIDChain::StructureIDChain): No need to store m_size as
a data member, so keep it in a local, which might be faster.
* kjs/StructureID.h:
* kjs/SymbolTable.h: Moved IdentifierRepHash to KJS namespace, as above.
* kjs/ustring.h:
2008-09-02 Adam Roben <aroben@apple.com>
Windows build fixes
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Add
StructureID.{cpp,h} to the project. Also let VS reorder this file.
* VM/CodeBlock.cpp: Include StringExtras so that snprintf will be
defined on Windows.
2008-09-01 Sam Weinig <sam@webkit.org>
Fix release build.
* JavaScriptCore.exp:
2008-09-01 Jan Michael Alonzo <jmalonzo@webkit.org>
Reviewed by Oliver Hunt.
Gtk buildfix
* GNUmakefile.am:
* kjs/PropertyMap.cpp: rename Identifier.h to identifier.h
* kjs/StructureID.cpp: include JSObject.h
2008-09-01 Geoffrey Garen <ggaren@apple.com>
Reviewed by Darin Adler.
First cut at inline caching for access to vanilla JavaScript properties.
SunSpider says 4% faster. Tests heavy on dictionary-like access have
regressed a bit -- we have a lot of room to improve in this area,
but this patch is over-ripe as-is.
JSCells now have a StructureID that uniquely identifies their layout,
and holds their prototype.
JSValue::put takes a PropertySlot& argument, so it can fill in details
about where it put a value, for the sake of caching.
* VM/CodeGenerator.cpp:
(KJS::CodeGenerator::CodeGenerator): Avoid calling removeDirect if we
can, since it disables inline caching in the global object. This can
probably improve in the future.
* kjs/JSGlobalObject.cpp: Nixed reset(), since it complicates caching, and
wasn't really necessary.
* kjs/JSObject.cpp: Tweaked getter / setter behavior not to rely on the
IsGetterSetter flag, since the flag was buggy. This is necessary in order
to avoid accidentally accessing a getter / setter as a normal property.
Also changed getter / setter creation to honor ReadOnly, matching Mozilla.
* kjs/PropertyMap.cpp: Nixed clear(), since it complicates caching and
isn't necessary.
* kjs/Shell.cpp: Moved SamplingTool dumping outside the loop. This allows
you to aggregate sampling of multiple files (or the same file repeatedly),
which helped me track down regressions.
* kjs/ustring.h: Moved IdentifierRepHash here to share it.
2008-09-01 Geoffrey Garen <ggaren@apple.com>
Reviewed by Sam Weinig.
Eagerly allocate the Math object's numeric constants. This avoids
constantly reallocating them in loops, and also ensures that the Math
object will not use the single property optimization, which makes
properties ineligible for caching.
SunSpider reports a small speedup, in combination with inline caching.
* kjs/MathObject.cpp:
(KJS::MathObject::MathObject):
(KJS::MathObject::getOwnPropertySlot):
* kjs/MathObject.h:
2008-09-01 Jan Michael Alonzo <jmalonzo@webkit.org>
Gtk build fix, not reviewed.
* GNUmakefile.am: Add SmallStrings.cpp in both release and debug builds
2008-08-31 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej Stachowiak.
Bug 20577: REGRESSION (r36006): Gmail is broken
<https://bugs.webkit.org/show_bug.cgi?id=20577>
r36006 changed stringProtoFuncSubstr() so that it is uses the more
efficient jsSubstring(), rather than using UString::substr() and then
calling jsString(). However, the change did not account for the case
where the start and the length of the substring extend beyond the length
of the original string. This patch corrects that.
* kjs/StringPrototype.cpp:
(KJS::stringProtoFuncSubstr):
2008-08-31 Simon Hausmann <hausmann@wekit.org>
Unreviewed build fix (with gcc 4.3)
* kjs/ustring.h: Properly forward declare operator== for UString and
the the concatenate functions inside the KJS namespace.
2008-08-30 Darin Adler <darin@apple.com>
Reviewed by Maciej.
- https://bugs.webkit.org/show_bug.cgi?id=20333
improve JavaScript speed when handling single-character strings
1.035x as fast on SunSpider overall.
1.127x as fast on SunSpider string tests.
1.910x as fast on SunSpider string-base64 test.
* API/JSObjectRef.cpp:
(JSObjectMakeFunction): Removed unneeded explicit construction of UString.
* GNUmakefile.am: Added SmallStrings.h and SmallStrings.cpp.
* JavaScriptCore.pri: Ditto.
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
Ditto.
* JavaScriptCore.xcodeproj/project.pbxproj: Ditto.
* JavaScriptCoreSources.bkl: Ditto.
* JavaScriptCore.exp: Updated.
* VM/Machine.cpp:
(KJS::jsAddSlowCase): Changed to use a code path that doesn't involve
a UString constructor. This avoids an extra jump caused by the "in charge"
vs. "not in charge" constructors.
(KJS::jsAdd): Ditto.
(KJS::jsTypeStringForValue): Adopted jsNontrivialString.
* kjs/ArrayPrototype.cpp:
(KJS::arrayProtoFuncToString): Adopted jsEmptyString.
(KJS::arrayProtoFuncToLocaleString): Ditto.
(KJS::arrayProtoFuncJoin): Ditto.
* kjs/BooleanPrototype.cpp:
(KJS::booleanProtoFuncToString): Adopted jsNontrivialString.
* kjs/DateConstructor.cpp:
(KJS::callDate): Ditto.
* kjs/DatePrototype.cpp:
(KJS::formatLocaleDate): Adopted jsEmptyString and jsNontrivialString.
(KJS::dateProtoFuncToString): Ditto.
(KJS::dateProtoFuncToUTCString): Ditto.
(KJS::dateProtoFuncToDateString): Ditto.
(KJS::dateProtoFuncToTimeString): Ditto.
(KJS::dateProtoFuncToLocaleString): Ditto.
(KJS::dateProtoFuncToLocaleDateString): Ditto.
(KJS::dateProtoFuncToLocaleTimeString): Ditto.
(KJS::dateProtoFuncToGMTString): Ditto.
* kjs/ErrorPrototype.cpp:
(KJS::ErrorPrototype::ErrorPrototype): Ditto.
(KJS::errorProtoFuncToString): Ditto.
* kjs/JSGlobalData.h: Added SmallStrings.
* kjs/JSString.cpp:
(KJS::jsString): Eliminated the overload that takes a const char*.
Added code to use SmallStrings to get strings of small sizes rather
than creating a new JSString every time.
(KJS::jsSubstring): Added. Used when creating a string from a substring
to avoid creating a JSString in cases where the substring will end up
empty or as one character.
(KJS::jsOwnedString): Added the same code as in jsString.
* kjs/JSString.h: Added new functions jsEmptyString, jsSingleCharacterString,
jsSingleCharacterSubstring, jsSubstring, and jsNontrivialString for various
cases where we want to create JSString, and want special handling for small
strings.
(KJS::JSString::JSString): Added an overload that takes a PassRefPtr of
a UString::Rep so you don't have to construct a UString; PassRefPtr can be
more efficient.
(KJS::jsEmptyString): Added.
(KJS::jsSingleCharacterString): Added.
(KJS::jsSingleCharacterSubstring): Added.
(KJS::jsNontrivialString): Added.
(KJS::JSString::getIndex): Adopted jsSingleCharacterSubstring.
(KJS::JSString::getStringPropertySlot): Ditto.
* kjs/NumberPrototype.cpp:
(KJS::numberProtoFuncToFixed): Adopted jsNontrivialString.
(KJS::numberProtoFuncToExponential): Ditto.
(KJS::numberProtoFuncToPrecision): Ditto.
* kjs/ObjectPrototype.cpp:
(KJS::objectProtoFuncToLocaleString): Adopted toThisJSString.
(KJS::objectProtoFuncToString): Adopted jsNontrivialString.
* kjs/RegExpConstructor.cpp: Separated the lastInput value that's used
with the lastOvector to return matches from the input value that can be
changed via JavaScript. They will be equal in many cases, but not all.
(KJS::RegExpConstructor::performMatch): Set input.
(KJS::RegExpMatchesArray::RegExpMatchesArray): Ditto.
(KJS::RegExpMatchesArray::fillArrayInstance): Adopted jsSubstring. Also,
use input rather than lastInput in the appropriate place.
(KJS::RegExpConstructor::getBackref): Adopted jsSubstring and jsEmptyString.
Added code to handle the case where there is no backref -- before this
depended on range checking in UString::substr which is not present in
jsSubstring.
(KJS::RegExpConstructor::getLastParen): Ditto.
(KJS::RegExpConstructor::getLeftContext): Ditto.
(KJS::RegExpConstructor::getRightContext): Ditto.
(KJS::RegExpConstructor::getValueProperty): Use input rather than lastInput.
Also adopt jsEmptyString.
(KJS::RegExpConstructor::putValueProperty): Ditto.
(KJS::RegExpConstructor::input): Ditto.
* kjs/RegExpPrototype.cpp:
(KJS::regExpProtoFuncToString): Adopt jsNonTrivialString. Also changed to
use UString::append to append single characters rather than using += and
a C-style string.
* kjs/SmallStrings.cpp: Added.
(KJS::SmallStringsStorage::SmallStringsStorage): Construct the
buffer and UString::Rep for all 256 single-character strings for
the U+0000 through U+00FF. This covers all the values used in
the base64 test as well as most values seen elsewhere on the web
as well. It's possible that later we might fix this to only work
for U+0000 through U+007F but the others are used quite a bit in
the current version of the base64 test.
(KJS::SmallStringsStorage::~SmallStringsStorage): Free memory.
(KJS::SmallStrings::SmallStrings): Create a set of small strings,
initially not created; created later when they are used.
(KJS::SmallStrings::~SmallStrings): Deallocate. Not left compiler
generated because the SmallStringsStorage class's destructor needs
to be visible.
(KJS::SmallStrings::mark): Mark all the strings.
(KJS::SmallStrings::createEmptyString): Create a cell for the
empty string. Called only the first time.
(KJS::SmallStrings::createSingleCharacterString): Create a cell
for one of the single-character strings. Called only the first time.
* kjs/SmallStrings.h: Added.
* kjs/StringConstructor.cpp:
(KJS::stringFromCharCodeSlowCase): Factored out of strinFromCharCode.
Only used for cases where the caller does not pass exactly one argument.
(KJS::stringFromCharCode): Adopted jsSingleCharacterString.
(KJS::callStringConstructor): Adopted jsEmptyString.
* kjs/StringObject.cpp:
(KJS::StringObject::StringObject): Adopted jsEmptyString.
* kjs/StringPrototype.cpp:
(KJS::stringProtoFuncReplace): Adopted jsSubstring.
(KJS::stringProtoFuncCharAt): Adopted jsEmptyString and
jsSingleCharacterSubstring and also added a special case when the
index is an immediate number to avoid conversion to and from floating
point, since that's the common case.
(KJS::stringProtoFuncCharCodeAt): Ditto.
(KJS::stringProtoFuncMatch): Adopted jsSubstring and jsEmptyString.
(KJS::stringProtoFuncSlice): Adopted jsSubstring and
jsSingleCharacterSubstring. Also got rid of some unneeded locals and
removed unneeded code to set the length property of the array, since it
is automatically updated as values are added to the array.
(KJS::stringProtoFuncSplit): Adopted jsEmptyString.
(KJS::stringProtoFuncSubstr): Adopted jsSubstring.
(KJS::stringProtoFuncSubstring): Ditto.
* kjs/collector.cpp:
(KJS::Heap::collect): Added a call to mark SmallStrings.
* kjs/ustring.cpp:
(KJS::UString::expandedSize): Made this a static member function since
it doesn't need to look at any data members.
(KJS::UString::expandCapacity): Use a non-inline function, makeNull, to
set the rep to null in failure cases. This avoids adding a PIC branch for
the normal case when there is no failure.
(KJS::UString::expandPreCapacity): Ditto.
(KJS::UString::UString): Ditto.
(KJS::concatenate): Refactored the concatenation constructor into this
separate function. Calling the concatenation constructor was leading to
an extra branch because of the in-charge vs. not-in-charge versions not
both being inlined, and this was showing up as nearly 1% on Shark. Also
added a special case for when the second string is a single character,
since it's a common idiom to build up a string that way and we can do
things much more quickly, without involving memcpy for example. Also
adopted the non-inline function, nullRep, for the same reason given for
makeNull above.
(KJS::UString::append): Adopted makeNull for failure cases.
(KJS::UString::operator=): Ditto.
(KJS::UString::toDouble): Added a special case for converting single
character strings to numbers. We're doing this a ton of times while
running the base64 test.
(KJS::operator==): Added special cases so we can compare single-character
strings without calling memcmp. Later we might want to special case other
short lengths similarly.
(KJS::UString::makeNull): Added.
(KJS::UString::nullRep): Added.
* kjs/ustring.h: Added declarations for the nullRep and makeNull. Changed
expandedSize to be a static member function. Added a declaration of the
concatenate function. Removed the concatenation constructor. Rewrote
operator+ to use the concatenate function.
2008-08-29 Anders Carlsson <andersca@apple.com>
Build fix.
* VM/Machine.cpp:
(KJS::getCPUTime):
2008-08-29 Anders Carlsson <andersca@apple.com>
Reviewed by Darin Adler.
<rdar://problem/6174667>
When a machine is under heavy load, the Slow Script dialog often comes up many times and just gets in the way
Instead of using clock time, use the CPU time spent executing the current thread when
determining if the script has been running for too long.
* VM/Machine.cpp:
(KJS::getCPUTime):
(KJS::Machine::checkTimeout):
2008-08-28 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Rubber-stamped by Sam Weinig.
Change 'term' to 'expr' in variable names to standardize terminology.
* kjs/nodes.cpp:
(KJS::BinaryOpNode::emitCode):
(KJS::ReverseBinaryOpNode::emitCode):
(KJS::ThrowableBinaryOpNode::emitCode):
* kjs/nodes.h:
(KJS::BinaryOpNode::BinaryOpNode):
(KJS::ReverseBinaryOpNode::ReverseBinaryOpNode):
(KJS::MultNode::):
(KJS::DivNode::):
(KJS::ModNode::):
(KJS::AddNode::):
(KJS::SubNode::):
(KJS::LeftShiftNode::):
(KJS::RightShiftNode::):
(KJS::UnsignedRightShiftNode::):
(KJS::LessNode::):
(KJS::GreaterNode::):
(KJS::LessEqNode::):
(KJS::GreaterEqNode::):
(KJS::ThrowableBinaryOpNode::):
(KJS::InstanceOfNode::):
(KJS::InNode::):
(KJS::EqualNode::):
(KJS::NotEqualNode::):
(KJS::StrictEqualNode::):
(KJS::NotStrictEqualNode::):
(KJS::BitAndNode::):
(KJS::BitOrNode::):
(KJS::BitXOrNode::):
* kjs/nodes2string.cpp:
(KJS::MultNode::streamTo):
(KJS::DivNode::streamTo):
(KJS::ModNode::streamTo):
(KJS::AddNode::streamTo):
(KJS::SubNode::streamTo):
(KJS::LeftShiftNode::streamTo):
(KJS::RightShiftNode::streamTo):
(KJS::UnsignedRightShiftNode::streamTo):
(KJS::LessNode::streamTo):
(KJS::GreaterNode::streamTo):
(KJS::LessEqNode::streamTo):
(KJS::GreaterEqNode::streamTo):
(KJS::InstanceOfNode::streamTo):
(KJS::InNode::streamTo):
(KJS::EqualNode::streamTo):
(KJS::NotEqualNode::streamTo):
(KJS::StrictEqualNode::streamTo):
(KJS::NotStrictEqualNode::streamTo):
(KJS::BitAndNode::streamTo):
(KJS::BitXOrNode::streamTo):
(KJS::BitOrNode::streamTo):
2008-08-28 Alp Toker <alp@nuanti.com>
GTK+ dist/build fix. List newly added header files.
* GNUmakefile.am:
2008-08-28 Sam Weinig <sam@webkit.org>
Reviewed by Oliver Hunt.
Change to throw a ReferenceError at runtime instead of a ParseError
at parse time, when the left hand side expression of a for-in statement
is not an lvalue.
* kjs/grammar.y:
* kjs/nodes.cpp:
(KJS::ForInNode::emitCode):
2008-08-28 Alexey Proskuryakov <ap@webkit.org>
Not reviewed, build fix (at least for OpenBSD, posssibly more).
https://bugs.webkit.org/show_bug.cgi?id=20545
missing #include <unistd.h> in JavaScriptCore/VM/SamplingTool.cpp
* VM/SamplingTool.cpp: add the missing include.
2008-08-26 Kevin McCullough <kmccullough@apple.com>
Reviewed by Geoff and Cameron.
<rdar://problem/6174603> Hitting assertion in Register::codeBlock when
loading facebook (20516).
- This was a result of my line numbers change. After a host function is
called the stack does not get reset correctly.
- Oddly this also appears to be a slight speedup on SunSpider.
* VM/Machine.cpp:
(KJS::Machine::privateExecute):
2008-08-26 Alexey Proskuryakov <ap@webkit.org>
Reviewed by Geoff and Tim.
Export new API methods.
* JavaScriptCore.exp:
2008-08-25 Kevin McCullough <kmccullough@apple.com>
Reviewed by Geoff, Tim and Mark.
<rdar://problem/6150623> JSProfiler: It would be nice if the profiles
in the console said what file and line number they came from
- Lay the foundation for getting line numbers and other data from the
JavaScript engine. With the cleanup in kjs/ExecState this is actually
a slight performance improvement.
* JavaScriptCore.exp: Export retrieveLastCaller() for WebCore.
* JavaScriptCore.xcodeproj/project.pbxproj:
* VM/Machine.cpp: Now Host and JS functions set a call frame on the
exec state, so this and the profiler code were pulled out of the
branches.
(KJS::Machine::privateExecute):
(KJS::Machine::retrieveLastCaller): This get's the lineNumber, sourceID
and sourceURL for the previously called function.
* VM/Machine.h:
* kjs/ExecState.cpp: Remove references to JSFunction since it's not used
anywhere.
* kjs/ExecState.h:
2008-08-25 Alexey Proskuryakov <ap@webkit.org>
Reviewed by Darin Adler.
Ensure that JSGlobalContextRelease() performs garbage collection, even if there are other
contexts in the current context's group.
This is only really necessary when the last reference is released, but there is no way to
determine that, and no harm in collecting slightly more often.
* API/JSContextRef.cpp: (JSGlobalContextRelease): Explicitly collect the heap if it is not
being destroyed.
2008-08-24 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Oliver Hunt.
Bug 20093: JSC shell does not clear exceptions after it executes toString on an expression
<https://bugs.webkit.org/show_bug.cgi?id=20093>
Clear exceptions after evaluating any code in the JSC shell. We do not
report exceptions that are caused by calling toString on the final
valued, but at least we avoid incorrect behaviour.
Also, print any exceptions that occurred while evaluating code at the
interactive prompt, not just while evaluating code from a file.
* kjs/Shell.cpp:
(runWithScripts):
(runInteractive):
2008-08-24 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Oliver.
Remove an unnecessary RefPtr to a RegisterID.
* kjs/nodes.cpp:
(KJS::DeleteBracketNode::emitCode):
2008-08-24 Mark Rowe <mrowe@apple.com>
Reviewed by Oliver Hunt.
Use the correct version number for when JSGlobalContextCreate was introduced.
* API/JSContextRef.h:
2008-08-23 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Rubber-stamped by Mark Rowe.
Remove modelines.
* API/APICast.h:
* API/JSBase.cpp:
* API/JSCallbackConstructor.cpp:
* API/JSCallbackConstructor.h:
* API/JSCallbackFunction.cpp:
* API/JSCallbackFunction.h:
* API/JSCallbackObject.cpp:
* API/JSCallbackObject.h:
* API/JSCallbackObjectFunctions.h:
* API/JSClassRef.cpp:
* API/JSContextRef.cpp:
* API/JSObjectRef.cpp:
* API/JSProfilerPrivate.cpp:
* API/JSStringRef.cpp:
* API/JSStringRefBSTR.cpp:
* API/JSStringRefCF.cpp:
* API/JSValueRef.cpp:
* API/tests/JSNode.c:
* API/tests/JSNode.h:
* API/tests/JSNodeList.c:
* API/tests/JSNodeList.h:
* API/tests/Node.c:
* API/tests/Node.h:
* API/tests/NodeList.c:
* API/tests/NodeList.h:
* API/tests/minidom.c:
* API/tests/minidom.js:
* API/tests/testapi.c:
* API/tests/testapi.js:
* JavaScriptCore.pro:
* kjs/FunctionConstructor.h:
* kjs/FunctionPrototype.h:
* kjs/JSArray.h:
* kjs/JSString.h:
* kjs/JSWrapperObject.cpp:
* kjs/NumberConstructor.h:
* kjs/NumberObject.h:
* kjs/NumberPrototype.h:
* kjs/lexer.h:
* kjs/lookup.h:
* wtf/Assertions.cpp:
* wtf/Assertions.h:
* wtf/HashCountedSet.h:
* wtf/HashFunctions.h:
* wtf/HashIterators.h:
* wtf/HashMap.h:
* wtf/HashSet.h:
* wtf/HashTable.h:
* wtf/HashTraits.h:
* wtf/ListHashSet.h:
* wtf/ListRefPtr.h:
* wtf/Noncopyable.h:
* wtf/OwnArrayPtr.h:
* wtf/OwnPtr.h:
* wtf/PassRefPtr.h:
* wtf/Platform.h:
* wtf/RefPtr.h:
* wtf/RefPtrHashMap.h:
* wtf/RetainPtr.h:
* wtf/UnusedParam.h:
* wtf/Vector.h:
* wtf/VectorTraits.h:
* wtf/unicode/Unicode.h:
* wtf/unicode/icu/UnicodeIcu.h:
2008-08-22 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Oliver.
Some cleanup to match our coding style.
* VM/CodeGenerator.h:
* VM/Machine.cpp:
(KJS::Machine::privateExecute):
* kjs/ExecState.cpp:
* kjs/ExecState.h:
* kjs/completion.h:
* kjs/identifier.cpp:
(KJS::Identifier::equal):
(KJS::CStringTranslator::hash):
(KJS::CStringTranslator::equal):
(KJS::CStringTranslator::translate):
(KJS::UCharBufferTranslator::equal):
(KJS::UCharBufferTranslator::translate):
(KJS::Identifier::remove):
* kjs/operations.h:
2008-08-20 Alexey Proskuryakov <ap@webkit.org>
Windows build fix.
* API/WebKitAvailability.h: Define DEPRECATED_ATTRIBUTE.
2008-08-19 Alexey Proskuryakov <ap@webkit.org>
Reviewed by Geoff Garen.
Bring back shared JSGlobalData and implicit locking, because too many clients rely on it.
* kjs/JSGlobalData.cpp:
(KJS::JSGlobalData::~JSGlobalData):
(KJS::JSGlobalData::JSGlobalData): Re-add shared instance.
(KJS::JSGlobalData::sharedInstanceExists): Ditto.
(KJS::JSGlobalData::sharedInstance): Ditto.
(KJS::JSGlobalData::sharedInstanceInternal): Ditto.
* API/JSContextRef.h: Deprecated JSGlobalContextCreate(). Added a very conservative
description of its threading model (nothing is allowed).
* API/JSContextRef.cpp:
(JSGlobalContextCreate): Use shared JSGlobalData.
(JSGlobalContextCreateInGroup): Support passing NULL group to request a unique one.
(JSGlobalContextRetain): Added back locking.
(JSGlobalContextRelease): Ditto.
(JSContextGetGlobalObject): Ditto.
* API/tests/minidom.c: (main):
* API/tests/testapi.c: (main):
Switched to JSGlobalContextCreateInGroup() to avoid deprecation warnings.
* JavaScriptCore.exp: Re-added JSLock methods. Added JSGlobalContextCreateInGroup (d'oh!).
* API/JSBase.cpp:
(JSEvaluateScript):
(JSCheckScriptSyntax):
(JSGarbageCollect):
* API/JSCallbackConstructor.cpp:
(KJS::constructJSCallback):
* API/JSCallbackFunction.cpp:
(KJS::JSCallbackFunction::call):
* API/JSCallbackObjectFunctions.h:
(KJS::::init):
(KJS::::getOwnPropertySlot):
(KJS::::put):
(KJS::::deleteProperty):
(KJS::::construct):
(KJS::::hasInstance):
(KJS::::call):
(KJS::::getPropertyNames):
(KJS::::toNumber):
(KJS::::toString):
(KJS::::staticValueGetter):
(KJS::::callbackGetter):
* API/JSObjectRef.cpp:
(JSObjectMake):
(JSObjectMakeFunctionWithCallback):
(JSObjectMakeConstructor):
(JSObjectMakeFunction):
(JSObjectHasProperty):
(JSObjectGetProperty):
(JSObjectSetProperty):
(JSObjectGetPropertyAtIndex):
(JSObjectSetPropertyAtIndex):
(JSObjectDeleteProperty):
(JSObjectCallAsFunction):
(JSObjectCallAsConstructor):
(JSObjectCopyPropertyNames):
(JSPropertyNameArrayRelease):
(JSPropertyNameAccumulatorAddName):
* API/JSValueRef.cpp:
(JSValueIsEqual):
(JSValueIsInstanceOfConstructor):
(JSValueMakeNumber):
(JSValueMakeString):
(JSValueToNumber):
(JSValueToStringCopy):
(JSValueToObject):
(JSValueProtect):
(JSValueUnprotect):
* ForwardingHeaders/JavaScriptCore/JSLock.h: Added.
* GNUmakefile.am:
* JavaScriptCore.pri:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* JavaScriptCoreSources.bkl:
* kjs/AllInOneFile.cpp:
* kjs/JSGlobalData.h:
* kjs/JSGlobalObject.cpp:
(KJS::JSGlobalObject::~JSGlobalObject):
(KJS::JSGlobalObject::init):
* kjs/JSLock.cpp: Added.
(KJS::createJSLockCount):
(KJS::JSLock::lockCount):
(KJS::setLockCount):
(KJS::JSLock::JSLock):
(KJS::JSLock::lock):
(KJS::JSLock::unlock):
(KJS::JSLock::currentThreadIsHoldingLock):
(KJS::JSLock::DropAllLocks::DropAllLocks):
(KJS::JSLock::DropAllLocks::~DropAllLocks):
* kjs/JSLock.h: Added.
(KJS::JSLock::JSLock):
(KJS::JSLock::~JSLock):
* kjs/Shell.cpp:
(functionGC):
(jscmain):
* kjs/collector.cpp:
(KJS::Heap::~Heap):
(KJS::Heap::heapAllocate):
(KJS::Heap::setGCProtectNeedsLocking):
(KJS::Heap::protect):
(KJS::Heap::unprotect):
(KJS::Heap::collect):
* kjs/identifier.cpp:
* kjs/interpreter.cpp:
(KJS::Interpreter::checkSyntax):
(KJS::Interpreter::evaluate):
Re-added implicit locking.
2008-08-19 Kevin McCullough <kmccullough@apple.com>
Reviewed by Tim and Mark.
Implement DTrace hooks for dashcode and instruments.
* API/JSProfilerPrivate.cpp: Added. Expose SPI so that profiling can be
turned on from a client. The DTrace probes were added within the
profiler mechanism for performance reasons so the profiler must be
started to enable tracing.
(JSStartProfiling):
(JSEndProfiling):
* API/JSProfilerPrivate.h: Added. Ditto.
* JavaScriptCore.exp: Exposing the start/stop methods to clients.
* JavaScriptCore.xcodeproj/project.pbxproj:
* kjs/Tracing.d: Define the DTrace probes.
* kjs/Tracing.h: Ditto.
* profiler/ProfileGenerator.cpp: Implement the DTrace probes in the
profiler.
(KJS::ProfileGenerator::willExecute):
(KJS::ProfileGenerator::didExecute):
2008-08-19 Steve Falkenburg <sfalken@apple.com>
Build fix.
* kjs/operations.cpp:
(KJS::equal):
2008-08-18 Timothy Hatcher <timothy@apple.com>
Fix an assertion when generating a heavy profile because the
empty value and deleted value of CallIdentifier where equal.
https://bugs.webkit.org/show_bug.cgi?id=20439
Reviewed by Dan Bernstein.
* profiler/CallIdentifier.h: Make the emptyValue for CallIdentifier
use empty strings for URL and function name.
2008-08-12 Darin Adler <darin@apple.com>
Reviewed by Geoff.
- eliminate JSValue::type()
This will make it slightly easier to change the JSImmediate design without
having to touch so many call sites.
SunSpider says this change is a wash (looked like a slight speedup, but not
statistically significant).
* API/JSStringRef.cpp: Removed include of JSType.h.
* API/JSValueRef.cpp: Removed include of JSType.h.
(JSValueGetType): Replaced use of JSValue::type() with
JSValue::is functions.
* JavaScriptCore.exp: Updated.
* VM/JSPropertyNameIterator.cpp: Removed type() implementation.
(KJS::JSPropertyNameIterator::toPrimitive): Changed to take
PreferredPrimitiveType argument instead of JSType.
* VM/JSPropertyNameIterator.h: Ditto.
* VM/Machine.cpp:
(KJS::fastIsNumber): Updated for name change.
(KJS::fastToInt32): Ditto.
(KJS::fastToUInt32): Ditto.
(KJS::jsAddSlowCase): Updated toPrimitive caller for change from
JSType to PreferredPrimitiveType.
(KJS::jsAdd): Replaced calls to JSValue::type() with calls to
JSValue::isString().
(KJS::jsTypeStringForValue): Replaced calls to JSValue::type()
with multiple calls to JSValue::is -- we could make this a
virtual function instead if we want to have faster performance.
(KJS::Machine::privateExecute): Renamed JSImmediate::toTruncatedUInt32
to JSImmediate::getTruncatedUInt32 for consistency with other functions.
Changed two calls of JSValue::type() to JSValue::isString().
* kjs/GetterSetter.cpp:
(KJS::GetterSetter::toPrimitive): Changed to take
PreferredPrimitiveType argument instead of JSType.
(KJS::GetterSetter::isGetterSetter): Added.
* kjs/GetterSetter.h:
* kjs/JSCell.cpp:
(KJS::JSCell::isString): Added.
(KJS::JSCell::isGetterSetter): Added.
(KJS::JSCell::isObject): Added.
* kjs/JSCell.h: Eliminated type function. Added isGetterSetter.
Made isString and isObject virtual. Changed toPrimitive to take
PreferredPrimitiveType argument instead of JSType.
(KJS::JSCell::isNumber): Use Heap::isNumber for faster performance.
(KJS::JSValue::isGetterSetter): Added.
(KJS::JSValue::toPrimitive): Changed to take
PreferredPrimitiveType argument instead of JSType.
* kjs/JSImmediate.h: Removed JSValue::type() and replaced
JSValue::toTruncatedUInt32 with JSValue::getTruncatedUInt32.
(KJS::JSImmediate::isEitherImmediate): Added.
* kjs/JSNotAnObject.cpp:
(KJS::JSNotAnObject::toPrimitive): Changed to take
PreferredPrimitiveType argument instead of JSType.
* kjs/JSNotAnObject.h: Ditto.
* kjs/JSNumberCell.cpp:
(KJS::JSNumberCell::toPrimitive): Ditto.
* kjs/JSNumberCell.h:
(KJS::JSNumberCell::toInt32): Renamed from fastToInt32. There's no
other "slow" version of this once you have a JSNumberCell, so there's
no need for "fast" in the name. It's a feature that this hides the
base class toInt32, which does the same job less efficiently (and has
an additional ExecState argument).
(KJS::JSNumberCell::toUInt32): Ditto.
* kjs/JSObject.cpp:
(KJS::callDefaultValueFunction): Use isGetterSetter instead of type.
(KJS::JSObject::getPrimitiveNumber): Use PreferredPrimitiveType.
(KJS::JSObject::defaultValue): Ditto.
(KJS::JSObject::defineGetter): Use isGetterSetter.
(KJS::JSObject::defineSetter): Ditto.
(KJS::JSObject::lookupGetter): Ditto.
(KJS::JSObject::lookupSetter): Ditto.
(KJS::JSObject::toNumber): Use PreferredPrimitiveType.
(KJS::JSObject::toString): Ditto.
(KJS::JSObject::isObject): Added.
* kjs/JSObject.h:
(KJS::JSObject::inherits): Call the isObject from JSCell; it's now
hidden by our override of isObject.
(KJS::JSObject::getOwnPropertySlotForWrite): Use isGetterSetter
instead of type.
(KJS::JSObject::getOwnPropertySlot): Ditto.
(KJS::JSObject::toPrimitive): Use PreferredPrimitiveType.
* kjs/JSString.cpp:
(KJS::JSString::toPrimitive): Use PreferredPrimitiveType.
(KJS::JSString::isString): Added.
* kjs/JSString.h: Ditto.
* kjs/JSValue.h: Removed type(), added isGetterSetter(). Added
PreferredPrimitiveType enum and used it as the argument for the
toPrimitive function.
(KJS::JSValue::getBoolean): Simplified a bit an removed a branch.
* kjs/collector.cpp:
(KJS::typeName): Changed to use JSCell::is functions instead of
calling JSCell::type.
* kjs/collector.h:
(KJS::Heap::isNumber): Renamed from fastIsNumber.
* kjs/nodes.h: Added now-needed include of JSType, since the type
is used here to record types of values in the tree.
* kjs/operations.cpp:
(KJS::equal): Rewrote to no longer depend on type().
(KJS::strictEqual): Ditto.
2008-08-18 Kevin McCullough <kmccullough@apple.com>
Reviewed by Tim.
If there are no nodes in a profile all the time should be attributed to
(idle)
* profiler/Profile.cpp: If ther are no nodes make sure we still process
the head.
(KJS::Profile::forEach):
* profiler/ProfileGenerator.cpp: Remove some useless code.
(KJS::ProfileGenerator::stopProfiling):
2008-08-18 Alexey Proskuryakov <ap@webkit.org>
Reviewed by Maciej.
Make JSGlobalContextRetain/Release actually work.
* API/JSContextRef.cpp:
(JSGlobalContextRetain):
(JSGlobalContextRelease):
Ref/deref global data to give checking for globalData.refCount() some sense.
* API/tests/testapi.c: (main): Added a test for this bug.
* kjs/JSGlobalData.cpp:
(KJS::JSGlobalData::~JSGlobalData):
While checking for memory leaks, found that JSGlobalData::emptyList has changed to
a pointer, but it was not destructed, causing a huge leak in run-webkit-tests --threaded.
2008-08-17 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej.
Change the counting of constants so that preincrement and predecrement of
const local variables are considered unexpected loads.
* kjs/nodes.cpp:
(KJS::PrefixResolveNode::emitCode):
* kjs/nodes.h:
(KJS::ScopeNode::neededConstants):
2008-08-17 Oliver Hunt <oliver@apple.com>
Reviewed by Cameron Zwarich.
<rdar://problem/6150322> In Gmail, a crash occurs at KJS::Machine::privateExecute() when applying list styling to text after a quote had been removed
<https://bugs.webkit.org/show_bug.cgi?id=20386>
This crash was caused by "depth()" incorrectly determining the scope depth
of a 0 depth function without a full scope chain. Because such a function
would not have an activation the depth function would return the scope depth
of the parent frame, thus triggering an incorrect unwind. Any subsequent
look up that walked the scope chain would result in incorrect behaviour,
leading to a crash or incorrect variable resolution. This can only actually
happen in try...finally statements as that's the only path that can result in
the need to unwind the scope chain, but not force the function to need a
full scope chain.
The fix is simply to check for this case before attempting to walk the scope chain.
* VM/Machine.cpp:
(KJS::depth):
(KJS::Machine::throwException):
2008-08-17 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Maciej.
Bug 20419: Remove op_jless
<https://bugs.webkit.org/show_bug.cgi?id=20419>
Remove op_jless, which is rarely used now that we have op_loop_if_less.
* VM/CodeBlock.cpp:
(KJS::CodeBlock::dump):
* VM/CodeGenerator.cpp:
(KJS::CodeGenerator::emitJumpIfTrue):
* VM/Machine.cpp:
(KJS::Machine::privateExecute):
* VM/Opcode.h:
2008-08-17 Cameron Zwarich <cwzwarich@uwaterloo.ca>
Reviewed by Dan Bernstein.
Fix a typo in r35807 that is also causing build failures for
non-AllInOne builds.
* kjs/NumberConstructor.cpp:
2008-08-17 Geoffrey Garen <ggaren@apple.com>
Reviewed by Cameron Zwarich.
Made room for a free word in JSCell.
SunSpider says no change.
I changed JSCallbackObjectData, Arguments, JSArray, and RegExpObject to
store auxiliary data in a secondary structure.
I changed InternalFunction to store the function's name in the property
map.
I changed JSGlobalObjectData to use a virtual destructor, so WebCore's
JSDOMWindowBaseData could inherit from it safely. (It's a strange design
for JSDOMWindowBase to allocate an object that JSGlobalObject deletes,
but that's really our only option, given the size constraint.)
I also added a bunch of compile-time ASSERTs, and removed lots of comments
in JSObject.h because they were often out of date, and they got in the
way of reading what was actually going on.
Also renamed JSArray::getLength to JSArray::length, to match our style
guidelines.
2008-08-16 Geoffrey Garen <ggaren@apple.com>
Reviewed by Oliver Hunt.
Sped up property access for array.length and string.length by adding a
mechanism for returning a temporary value directly instead of returning
a pointer to a function that retrieves the value.
Also removed some unused cruft from PropertySlot.
SunSpider says 0.5% - 1.2% faster.
NOTE: This optimization is not a good idea in general, because it's
actually a pessimization in the case of resolve for assignment,
and it may get in the way of other optimizations in the future.
2008-08-16 Dan Bernstein <mitz@apple.com>
Reviewed by Geoffrey Garen.
Disable dead code stripping in debug builds.
* Configurations/Base.xcconfig:
* JavaScriptCore.xcodeproj/project.pbxproj:
2008-08-15 Mark Rowe <mrowe@apple.com>
Reviewed by Oliver Hunt.
<rdar://problem/6143072> FastMallocZone's enumeration code makes assumptions about handling of remote memory regions that overlap
* wtf/FastMalloc.cpp:
(WTF::TCMalloc_Central_FreeList::enumerateFreeObjects): Don't directly compare pointers mapped into the local process with
a pointer that has not been mapped. Instead, calculate a local address for the pointer and compare with that.
(WTF::TCMallocStats::FreeObjectFinder::findFreeObjects): Pass in the remote address of the central free list so that it can
be used when calculating local addresses.
(WTF::TCMallocStats::FastMallocZone::enumerate): Ditto.
2008-08-15 Mark Rowe <mrowe@apple.com>
Rubber-stamped by Geoff Garen.
<rdar://problem/6139914> Please include a _debug version of JavaScriptCore framework
* Configurations/Base.xcconfig: Factor out the debug-only settings so that they can shared
between the Debug configuration and debug Production variant.
* JavaScriptCore.xcodeproj/project.pbxproj: Enable the debug variant.
2008-08-15 Mark Rowe <mrowe@apple.com>
Fix the 64-bit build.
Add extra cast to avoid warnings about loss of precision when casting from
JSValue* to an integer type.
* kjs/JSImmediate.h:
(KJS::JSImmediate::intValue):
(KJS::JSImmediate::uintValue):
2008-08-15 Alexey Proskuryakov <ap@webkit.org>
Still fixing Windows build.
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.make: Added OpaqueJSString
to yet another place.
2008-08-15 Alexey Proskuryakov <ap@webkit.org>
Trying to fix non-Apple builds.
* ForwardingHeaders/JavaScriptCore/OpaqueJSString.h: Added.
2008-08-15 Gavin Barraclough <barraclough@apple.com>
Reviewed by Geoff Garen.
Allow JSImmediate to hold 31 bit signed integer immediate values. The low two bits of a
JSValue* are a tag, with the tag value 00 indicating the JSValue* is a pointer to a
JSCell. Non-zero tag values used to indicate that the JSValue* is not a real pointer,
but instead holds an immediate value encoded within the pointer. This patch changes the
encoding so both the tag values 01 and 11 indicate the value is a signed integer, allowing
a 31 bit value to be stored. All other immediates are tagged with the value 10, and
distinguished by a secondary tag.
Roughly +2% on SunSpider.
* kjs/JSImmediate.h: Encoding of JSImmediates has changed - see comment at head of file for
descption of new layout.
2008-08-15 Alexey Proskuryakov <ap@webkit.org>
More build fixes.
* API/OpaqueJSString.h: Add a namespace to friend declaration to appease MSVC.
* API/JSStringRefCF.h: (JSStringCreateWithCFString) Cast UniChar* to UChar* explicitly.
* JavaScriptCore.exp: Added OpaqueJSString::create(const KJS::UString&) to fix WebCore build.
2008-08-15 Alexey Proskuryakov <ap@webkit.org>
Build fix.
* JavaScriptCore.xcodeproj/project.pbxproj: Marked OpaqueJSString as private
* kjs/identifier.cpp:
(KJS::Identifier::checkSameIdentifierTable):
* kjs/identifier.h:
(KJS::Identifier::add):
Since checkSameIdentifierTable is exported for debug build's sake, gcc wants it to be
non-inline in release builds, too.
* JavaScriptCore.exp: Don't export inline OpaqueJSString destructor.
2008-08-15 Alexey Proskuryakov <ap@webkit.org>
Reviewed by Geoff Garen.
JSStringRef is created context-free, but can get linked to one via an identifier table,
breaking an implicit API contract.
Made JSStringRef point to OpaqueJSString, which is a new string object separate from UString.
* API/APICast.h: Removed toRef/toJS conversions for JSStringRef, as this is no longer a
simple typecast.
* kjs/identifier.cpp:
(KJS::Identifier::checkSameIdentifierTable):
* kjs/identifier.h:
(KJS::Identifier::add):
(KJS::UString::checkSameIdentifierTable):
Added assertions to verify that an identifier is not being added to a different JSGlobalData.
* API/JSObjectRef.cpp:
(OpaqueJSPropertyNameArray::OpaqueJSPropertyNameArray): Changed OpaqueJSPropertyNameArray to
hold JSStringRefs. This is necessary to avoid having to construct (and leak) a new instance
in JSPropertyNameArrayGetNameAtIndex(), now that making a JSStringRef is not just a typecast.
* API/OpaqueJSString.cpp: Added.
(OpaqueJSString::create):
(OpaqueJSString::ustring):
(OpaqueJSString::identifier):
* API/OpaqueJSString.h: Added.
(OpaqueJSString::create):
(OpaqueJSString::characters):
(OpaqueJSString::length):
(OpaqueJSString::OpaqueJSString):
(OpaqueJSString::~OpaqueJSString):
* API/JSBase.cpp:
(JSEvaluateScript):
(JSCheckScriptSyntax):
* API/JSCallbackObjectFunctions.h:
(KJS::::getOwnPropertySlot):
(KJS::::put):
(KJS::::deleteProperty):
(KJS::::staticValueGetter):
(KJS::::callbackGetter):
* API/JSStringRef.cpp:
(JSStringCreateWithCharacters):
(JSStringCreateWithUTF8CString):
(JSStringRetain):
(JSStringRelease):
(JSStringGetLength):
(JSStringGetCharactersPtr):
(JSStringGetMaximumUTF8CStringSize):
(JSStringGetUTF8CString):
(JSStringIsEqual):
* API/JSStringRefCF.cpp:
(JSStringCreateWithCFString):
(JSStringCopyCFString):
* API/JSValueRef.cpp:
(JSValueMakeString):
(JSValueToStringCopy):
Updated to use OpaqueJSString.
* GNUmakefile.am:
* JavaScriptCore.exp:
* JavaScriptCore.pri:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* JavaScriptCoreSources.bkl:
Added OpaqueJSString.
2008-08-14 Kevin McCullough <kmccullough@apple.com>
Reviewed by Tim.
<rdar://problem/6115819> Notify of profile in console
- Profiles now have a unique ID so that they can be linked to the
console message that announces that a profile completed.
* profiler/HeavyProfile.cpp:
(KJS::HeavyProfile::HeavyProfile):
* profiler/Profile.cpp:
(KJS::Profile::create):
(KJS::Profile::Profile):
* profiler/Profile.h:
(KJS::Profile::uid):
* profiler/ProfileGenerator.cpp:
(KJS::ProfileGenerator::create):
(KJS::ProfileGenerator::ProfileGenerator):
* profiler/ProfileGenerator.h:
* profiler/Profiler.cpp:
(KJS::Profiler::startProfiling):
* profiler/TreeProfile.cpp:
(KJS::TreeProfile::create):
(KJS::TreeProfile::TreeProfile):
* profiler/TreeProfile.h:
2008-08-13 Geoffrey Garen <ggaren@apple.com>
Reviewed by Oliver Hunt.
Nixed a PIC branch from JSObject::getOwnPropertySlot, by forcing
fillGetterProperty, which references a global function pointer,
out-of-line.
.2% SunSpider speedup, 4.3% access-nbody speedup, 8.7% speedup on a
custom property access benchmark for objects with one property.
* kjs/JSObject.cpp:
(KJS::JSObject::fillGetterPropertySlot):
2008-08-13 Alp Toker <alp@nuanti.com>
Reviewed by Eric Seidel.
https://bugs.webkit.org/show_bug.cgi?id=20349
WTF::initializeThreading() fails if threading is already initialized
Fix threading initialization logic to support cases where
g_thread_init() has already been called elsewhere.
Resolves database-related crashers reported in several applications.
* wtf/ThreadingGtk.cpp:
(WTF::initializeThreading):
2008-08-13 Brad Hughes <bhughes@trolltech.com>
Reviewed by Simon.
Fix compiling of QtWebKit in release mode with the Intel C++ Compiler for Linux
The latest upgrade of the intel compiler allows us to compile all of
Qt with optimizations enabled (yay!).
* JavaScriptCore.pro:
2008-08-12 Oliver Hunt <oliver@apple.com>
Reviewed by Geoff Garen.
Add peephole optimisation to 'op_not... jfalse...' (eg. if(!...) )
This is a very slight win in sunspider, and a fairly substantial win
in hot code that does if(!...), etc.
* VM/CodeGenerator.cpp:
(KJS::CodeGenerator::retrieveLastUnaryOp):
(KJS::CodeGenerator::rewindBinaryOp):
(KJS::CodeGenerator::rewindUnaryOp):
(KJS::CodeGenerator::emitJumpIfFalse):
* VM/CodeGenerator.h:
2008-08-12 Dan Bernstein <mitz@apple.com>
- JavaScriptCore part of <rdar://problem/6121636>
Make fast*alloc() abort() on failure and add "try" variants that
return NULL on failure.
Reviewed by Darin Adler.
* JavaScriptCore.exp: Exported tryFastCalloc().
* VM/RegisterFile.h:
(KJS::RegisterFile::RegisterFile): Removed an ASSERT().
* kjs/JSArray.cpp:
(KJS::JSArray::putSlowCase): Changed to use tryFastRealloc().
(KJS::JSArray::increaseVectorLength): Ditto.
* kjs/ustring.cpp:
(KJS::allocChars): Changed to use tryFastMalloc().
(KJS::reallocChars): Changed to use tryFastRealloc().
* wtf/FastMalloc.cpp:
(WTF::fastZeroedMalloc): Removed null checking of fastMalloc()'s result
and removed extra call to InvokeNewHook().
(WTF::tryFastZeroedMalloc): Added. Uses tryFastMalloc().
(WTF::tryFastMalloc): Renamed fastMalloc() to this.
(WTF::fastMalloc): Added. This version abort()s if allocation fails.
(WTF::tryFastCalloc): Renamed fastCalloc() to this.
(WTF::fastCalloc): Added. This version abort()s if allocation fails.
(WTF::tryFastRealloc): Renamed fastRealloc() to this.
(WTF::fastRealloc): Added. This version abort()s if allocation fails.
(WTF::do_malloc): Made this a function template. When the abortOnFailure
template parameter is set, the function abort()s on failure to allocate.
Otherwise, it sets errno to ENOMEM and returns zero.
(WTF::TCMallocStats::fastMalloc): Defined to abort() on failure.
(WTF::TCMallocStats::tryFastMalloc): Added. Does not abort() on
failure.
(WTF::TCMallocStats::fastCalloc): Defined to abort() on failure.
(WTF::TCMallocStats::tryFastCalloc): Added. Does not abort() on
failure.
(WTF::TCMallocStats::fastRealloc): Defined to abort() on failure.
(WTF::TCMallocStats::tryFastRealloc): Added. Does not abort() on
failure.
* wtf/FastMalloc.h: Declared the "try" variants.
2008-08-11 Adam Roben <aroben@apple.com>
Move WTF::notFound into its own header so that it can be used
independently of Vector
Rubberstamped by Darin Adler.
* JavaScriptCore.vcproj/WTF/WTF.vcproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
Added NotFound.h to the project.
* wtf/NotFound.h: Added. Moved the notFound constant here...
* wtf/Vector.h: ...from here.
2008-08-11 Alexey Proskuryakov <ap@webkit.org>
Reviewed by Mark Rowe.
<rdar://problem/6130393> REGRESSION: PhotoBooth hangs after launching under TOT Webkit
* API/JSContextRef.cpp: (JSGlobalContextRelease): Corrected a comment.
* kjs/collector.cpp: (KJS::Heap::~Heap): Ensure that JSGlobalData is not deleted while
sweeping the heap.
== Rolled over to ChangeLog-2008-08-10 ==