Context Navigation

← Previous Period
Next Period →

Timeline

Dec 17, 2008:

11:02 PM Changeset in webkit [39371] by barraclough@apple.com

7 edits in trunk/JavaScriptCore

2008-12-17 Gavin Barraclough <barraclough@apple.com>

Reviewed by Cameron Zwarich.

Fixes for Sunspider failures with the JIT enabled on x86-64.

assembler/MacroAssembler.h:

Switch the order of the RegisterID & Address form of je32, to keep it consistent with jne32.

jit/JIT.cpp:
jit/JIT.h:
jit/JITInlineMethods.h:

Port the m_ctiVirtualCall tramopline generation to use the MacroAssembler interface.

jit/JITCall.cpp:

Fix bug in the non-optimizing code path, vptr check should have been to the memory address pointer
to by the register, not to the register itself.

wrec/WRECGenerator.cpp:

See assembler/MacroAssembler.h, above.

7:38 PM Changeset in webkit [39370] by barraclough@apple.com

9 edits in trunk/JavaScriptCore

2008-12-17 Gavin Barraclough <barraclough@apple.com>

Reviewed by Sam Weinig.

print("Hello, 64-bit jitted world!");
Get hello-world working through the JIT, on x86-64.

assembler/X86Assembler.h:

Fix encoding of opcode + RegisterID format instructions for 64-bit.

interpreter/Interpreter.cpp:
interpreter/Interpreter.h:

Make VoidPtrPair actually be a pair of void*s.
(Possibly should make this change for 32-bit Mac platforms, too - but won't change 32-bit behaviour in this patch).

jit/JIT.cpp:
jit/JIT.h:

Provide names for the timeoutCheckRegister & callFrameRegister on x86-64,
force x86-64 ctiTrampoline arguments onto the stack,
implement the asm trampolines for x86-64,
implement the restoreArgumentReference methods for x86-64 calling conventions.

jit/JITCall.cpp:
jit/JITInlineMethods.h:
wtf/Platform.h:

Add switch settings to ENABLE(JIT), on PLATFORM(X86_64) (currently still disabled).

5:56 PM Changeset in webkit [39369] by Nikolas Zimmermann

21 edits in trunk

Reviewed by Maciej Stachowiak.

Fixes: https://bugs.webkit.org/show_bug.cgi?id=22866

wml.css is based on an older copy of html4.css. Synchronize them.
Remove all entries, that are irrelevant for WML.

Fix <do> appearance: should behave like HTMLs <button>.

4:00 PM Changeset in webkit [39368] by weinig@apple.com

2 edits in trunk/JavaScriptCore

2008-12-17 Sam Weinig <sam@webkit.org>

Reviewed by Gavin Barraclough.

Add more CodeBlock statistics.

bytecode/CodeBlock.cpp: (JSC::CodeBlock::dumpStatistics):

2:16 PM Changeset in webkit [39367] by abarth@webkit.org

6 edits in trunk/WebCore

2008-12-17 Yury Semikhatsky <yurys@google.com>

Reviewed by Timothy Hatcher.
Landed by Adam Barth.

Added more checks that WebInspector.panels.{resources,scripts} are
defined where they are accessed from other panels and WebInspector
(some panels are not yet supported in Chrome).

inspector/front-end/Console.js: (WebInspector.Console.prototype.addMessage): (WebInspector.Console.prototype.clearMessages): (WebInspector.Console.prototype.completions):
inspector/front-end/ObjectPropertiesSection.js: (WebInspector.ObjectPropertyTreeElement.prototype.evaluateExpression):
inspector/front-end/ScriptsPanel.js: (WebInspector.ScriptsPanel.prototype.get searchableViews): (WebInspector.ScriptsPanel.prototype._sourceViewForScriptOrResource): (WebInspector.ScriptsPanel.prototype._sourceFrameForScriptOrResource): (WebInspector.ScriptsPanel.prototype._showScriptOrResource):
inspector/front-end/SourceView.js: (WebInspector.SourceView.prototype._addBreakpoint):
inspector/front-end/inspector.js: (WebInspector.addResource): (WebInspector.removeResource): (WebInspector.showResourceForURL):

1:25 PM Changeset in webkit [39366] by weinig@apple.com

5 edits in trunk/JavaScriptCore

2008-12-17 Sam Weinig <sam@webkit.org>

Reviewed by Darin Adler.

Fix for https://bugs.webkit.org/show_bug.cgi?id=22897
<rdar://problem/6428342>
Look into feasibility of discarding bytecode after native codegen

Clear the bytecode Instruction vector at the end JIT generation.

Saves 4.8 MB on Membuster head.

bytecode/CodeBlock.cpp: (JSC::CodeBlock::dump): Add logging for the case that someone tries to dump the instructions of a CodeBlock that has had its bytecode vector cleared. (JSC::CodeBlock::CodeBlock): Initialize the instructionCount (JSC::CodeBlock::handlerForBytecodeOffset): Use instructionCount instead of the size of the instruction vector in the assertion. (JSC::CodeBlock::lineNumberForBytecodeOffset): Ditto. (JSC::CodeBlock::expressionRangeForBytecodeOffset): Ditto. (JSC::CodeBlock::getByIdExceptionInfoForBytecodeOffset): Ditto. (JSC::CodeBlock::functionRegisterForBytecodeOffset): Ditto.
bytecode/CodeBlock.h: (JSC::CodeBlock::setInstructionCount): Store the instruction vector size in debug builds for assertions.
bytecompiler/BytecodeGenerator.cpp: (JSC::BytecodeGenerator::generate):
jit/JIT.cpp: (JSC::JIT::privateCompile): Clear the bytecode vector unless we have compiled with Opcode sampling where we will continue to require it

1:12 PM Changeset in webkit [39365] by abarth@webkit.org

3 edits in trunk/LayoutTests

2008-12-17 Feng Qian <feng@chromium.org>

Reviewed by Alexey Proskuryakov.
Landed by Adam Barth.

https://bugs.webkit.org/show_bug.cgi?id=21948
Change two tests to use the host url in XMLHttpRequest so that
it does not fail because of Chrome's file URL security policy.
It does not affect the expected behaviors of tests.

fast/dom/null-document-xmlhttprequest-open.html:
fast/dom/xmlhttprequest-invalid-values.html:

1:12 PM Changeset in webkit [39364] by abarth@webkit.org

1 edit
2 adds in trunk/LayoutTests

2008-12-17 Pamela Greene <pam@chromium.org>

Reviewed by Darin Adler.
Landed by Adam Barth.

https://bugs.webkit.org/show_bug.cgi?id=22577
Add a test to ensure that checkboxes and radio buttons with duplicate
names map to a collection.

fast/dom/HTMLInputElement/duplicate-element-names-expected.txt: Added.
fast/dom/HTMLInputElement/duplicate-element-names.html: Added.

1:11 PM Changeset in webkit [39363] by abarth@webkit.org

1 edit
4 adds in trunk

2008-12-17 Pamela Greene <pam@chromium.org>

Reviewed by Darin Adler.
Landed by Adam Barth.

https://bugs.webkit.org/show_bug.cgi?id=22612
Add test to verify type enforcement in DOM setters.

fast/dom/setter-type-enforcement-expected.txt: Added.
fast/dom/setter-type-enforcement.html: Added.

1:11 PM Changeset in webkit [39362] by abarth@webkit.org

4 edits in trunk

WebCore:

2008-12-17 Cary Clark < caryclark@google.com>

Reviewed by Darin Adler.
Landed by Adam Barth.

Add ENABLE_TEXT_CARET to permit the ANDROID platform
to invalidate and draw the caret in a separate thread.

page/Frame.cpp: (WebCore::Frame::clearCaretRectIfNeeded): Body of function does nothing if text caret is disabled. (WebCore::Frame::selectionLayoutChanged): Do nothing if text caret is disabled and the caret only blinked. (WebCore::Frame::caretBlinkTimerFired): (WebCore::Frame::paintCaret): (WebCore::Frame::paintDragCaret): Body of functions does nothing if text caret is disabled.

JavaScriptCore:

2008-12-17 Cary Clark <caryclark@google.com>

Reviewed by Darin Adler.
Landed by Adam Barth.

Add ENABLE_TEXT_CARET to permit the ANDROID platform
to invalidate and draw the caret in a separate thread.

wtf/Platform.h: Default ENABLE_TEXT_CARET to 1.

11:01 AM Changeset in webkit [39361] by Nikolas Zimmermann

9 edits in trunk/LayoutTests

Rubber-stamped by Timothy Hatcher.

Fix wml/ tests, by doing s/</\</ and s/>/\>/.
Not sure why this problem wasn't visible before, something has changed.

10:10 AM Changeset in webkit [39360] by ap@webkit.org

2 edits in trunk/JavaScriptCore

Reviewed by Darin Adler.

Don't use unique context group in JSGlobalContextCreate() on Tiger or Leopard, take two.

API/JSContextRef.cpp: The previous patch that claimed to do this was making Tiger and Leopard always use unique context group instead.

5:41 AM Changeset in webkit [39359] by Simon Hausmann

7 edits in trunk

2008-12-16 Simon Hausmann <Simon Hausmann>

Reviewed by Holger Freyther.

Added support for unload event counting to the Qt DRT.

5:41 AM Changeset in webkit [39358] by Simon Hausmann

2 edits in trunk/LayoutTests

2008-12-16 Simon Hausmann <Simon Hausmann>

Reviewed by Holger Freyther.

Added mac API specific loader test to the skip list for the Qt DRT.

4:15 AM Changeset in webkit [39357] by Antti Koivisto

3 edits in trunk/WebKit/mac

2008-12-16 Antti Koivisto <Antti Koivisto>

Reviewed by John Sullivan.

Add version check for shift-reload behavior.

Misc/WebKitVersionChecks.h:
WebView/WebFrame.mm: (-[WebFrame reload]):

Dec 16, 2008:

11:36 PM Changeset in webkit [39356] by mrowe@apple.com

4 edits in trunk

Versioning.

11:35 PM Changeset in webkit [39355] by mrowe@apple.com

1 copy in tags/Safari-6528.7

New tag.

10:30 PM Changeset in webkit [39354] by weinig@apple.com

5 edits in trunk/JavaScriptCore

2008-12-16 Sam Weinig <sam@webkit.org>

Reviewed by Geoffrey Garen.

Fix for https://bugs.webkit.org/show_bug.cgi?id=22838
Remove dependency on the bytecode Instruction buffer in Interpreter::throwException
Part of <rdar://problem/6428342>

bytecode/CodeBlock.cpp: (JSC::CodeBlock::functionRegisterForBytecodeOffset): Added. Function to get a function Register index in a callFrame for a bytecode offset. (JSC::CodeBlock::shrinkToFit): Shrink m_getByIdExceptionInfo and m_functionRegisterInfos.
bytecode/CodeBlock.h: (JSC::FunctionRegisterInfo::FunctionRegisterInfo): Added. (JSC::CodeBlock::addFunctionRegisterInfo):
bytecompiler/BytecodeGenerator.cpp: (JSC::BytecodeGenerator::emitCall):
interpreter/Interpreter.cpp: (JSC::Interpreter::throwException): Use functionRegisterForBytecodeOffset in JIT mode.

10:25 PM Changeset in webkit [39353] by sfalken@apple.com

3 edits in trunk/WebKit/win

Windows build fix.

Reviewed by Stephanie Lewis.

COMPropertyBag.h: (::createInstance): (::adopt): (::QueryInterface): (::AddRef): (::Release): (::Read): (::Write): (::CountProperties): (::GetPropertyInfo): (::LoadObject):
WebURLResponse.cpp: (WebURLResponse::allHeaderFields):

7:29 PM WebKit Team edited by barraclough@apple.com

(diff)

6:35 PM Changeset in webkit [39352] by Stephanie Lewis

2 edits in trunk/WebKit/win

2008-12-16 Stephanie Lewis <Stephanie Lewis>

Another Windows Build Fix.

WebURLResponse.cpp: (WebURLResponse::allHeaderFields):

6:34 PM Changeset in webkit [39351] by weinig@apple.com

6 edits in trunk/JavaScriptCore

2008-12-16 Sam Weinig <sam@webkit.org>

Reviewed by Gavin Barraclough.

Fix for https://bugs.webkit.org/show_bug.cgi?id=22837
Remove dependency on the bytecode Instruction buffer in Interpreter::cti_op_call_NotJSFunction
Part of <rdar://problem/6428342>

interpreter/CallFrame.h: Added comment regarding returnPC storing a void*.
interpreter/Interpreter.cpp: (JSC::bytecodeOffsetForPC): We no longer have any cases of the PC being in the instruction stream for JIT, so we can remove the check. (JSC::Interpreter::cti_op_call_NotJSFunction): Use the CTI_RETURN_ADDRESS as the call frame returnPC as it is only necessary for looking up when throwing an exception.
interpreter/RegisterFile.h: (JSC::RegisterFile::): Added comment regarding returnPC storing a void*.
jit/JIT.h: Remove ARG_instr4.
jit/JITCall.cpp: (JSC::JIT::compileOpCallSetupArgs): Don't pass the instruction pointer.

6:23 PM Changeset in webkit [39350] by cwzwarich@webkit.org

4 edits in trunk

2008-12-16 Darin Adler <Darin Adler>

Reviewed and landed by Cameron Zwarich.

Preparatory work for fixing

Bug 22887: Make UString::Rep use RefCounted rather than implementing its own ref counting
<https://bugs.webkit.org/show_bug.cgi?id=22887>

Change the various string translators used by Identifier:add() so that
they never zero the ref count of a newly created UString::Rep. Also,
change similar code for AtomicString to use the same idiomatic style.

JavaScriptCore:

runtime/Identifier.cpp: (JSC::CStringTranslator::translate): (JSC::Identifier::add): (JSC::UCharBufferTranslator::translate):

WebCore:

platform/text/AtomicString.cpp: (WebCore::AtomicString::add):

5:54 PM Changeset in webkit [39349] by adele@apple.com

2 edits in trunk/WebCore

2008-12-16 Adele Peterson <adele@apple.com>

Reviewed by Darin Adler.

Fix for https://bugs.webkit.org/show_bug.cgi?id=22827
<rdar://problem/6247724> RenderThemeWin buttons have weird heavy text

rendering/RenderThemeWin.cpp: (WebCore::fillFontDescription): Add version that takes a font size. (WebCore::RenderThemeWin::systemFont): Use the default GUI font for control fonts, but specify the size. This will match Firefox.

5:27 PM Changeset in webkit [39348] by barraclough@apple.com

2 edits in trunk/JavaScriptCore

2008-12-16 Gavin Barraclough <barraclough@apple.com>

Build fix for 'doze.

assembler/AssemblerBuffer.h:

5:11 PM Changeset in webkit [39347] by Stephanie Lewis

2 edits in trunk/WebCore

2008-12-16 Stephanie Lewis <Stephanie Lewis>

Another build fix.

platform/text/AtomicString.h: (WebCore::AtomicString::AtomicString):

5:10 PM Changeset in webkit [39346] by andersca@apple.com

5 edits in trunk/WebKit/mac

2008-12-16 Anders Carlsson <andersca@apple.com>

Reviewed by Dan Bernstein.

Start sending keyboard events to the plug-in host.

Plugins/Hosted/NetscapePluginInstanceProxy.h:
Plugins/Hosted/NetscapePluginInstanceProxy.mm: (WebKit::NetscapePluginInstanceProxy::keyEvent):
Plugins/Hosted/WebHostedNetscapePluginView.mm: (-[WebHostedNetscapePluginView keyDown:]): (-[WebHostedNetscapePluginView keyUp:]):
Plugins/Hosted/WebKitPluginHost.defs:

4:57 PM Changeset in webkit [39345] by Stephanie Lewis

2 edits in trunk/WebCore

2008-12-16 Stephanie Lewis <Stephanie Lewis>

Windows Build Fix.

platform/text/AtomicString.h: (WebCore::AtomicString::AtomicString): (WebCore::AtomicString::createCFString):

4:11 PM Changeset in webkit [39344] by andersca@apple.com

2 edits in trunk/WebKit/mac

2008-12-16 Anders Carlsson <andersca@apple.com>

Reviewed by Kevin Decker.

<rdar://problem/6450538>

Fix flag enumeration.

Plugins/Hosted/WebKitPluginHostTypes.h:

4:03 PM Changeset in webkit [39343] by andersca@apple.com

6 edits in trunk/WebKit/mac

2008-12-16 Anders Carlsson <andersca@apple.com>

Reviewed by Kevin Decker.

Instead of passing a gazillion booleans to WKPCLoadURL, pass a single set of flags.

Plugins/Hosted/NetscapePluginHostProxy.mm: (WKPCLoadURL):
Plugins/Hosted/NetscapePluginInstanceProxy.h:
Plugins/Hosted/NetscapePluginInstanceProxy.mm: (WebKit::NetscapePluginInstanceProxy::loadURL):
Plugins/Hosted/WebKitPluginClient.defs:
Plugins/Hosted/WebKitPluginHostTypes.h:

4:03 PM Changeset in webkit [39342] by barraclough@apple.com

9 edits in trunk/JavaScriptCore

2008-12-16 Gavin Barraclough <barraclough@apple.com>

Reviewed by Cameron Zwarich.

Make the JIT compile on x86-64.
This largely involves populting the missing calls in MacroAssembler.h.
In addition some reinterpret_casts need removing from the JIT, and the
repatching property access code will need to be fully compiled out for
now. The changes in interpret.cpp are to reorder the functions so that
the _generic forms come before all other property access methods, and
then to place all property access methods other than the generic forms
under control of the ENABLE_JIT_OPTIMIZE_PROPERTY_ACCESS macro.

No performance impact.

assembler/AssemblerBuffer.h: (JSC::AssemblerBuffer::putInt64Unchecked):
assembler/MacroAssembler.h: (JSC::MacroAssembler::loadPtr): (JSC::MacroAssembler::load32): (JSC::MacroAssembler::storePtr): (JSC::MacroAssembler::storePtrWithRepatch): (JSC::MacroAssembler::store32): (JSC::MacroAssembler::poke): (JSC::MacroAssembler::move): (JSC::MacroAssembler::testImm64): (JSC::MacroAssembler::jePtr): (JSC::MacroAssembler::jnePtr): (JSC::MacroAssembler::jnzPtr): (JSC::MacroAssembler::jzPtr):
assembler/X86Assembler.h: (JSC::X86Assembler::): (JSC::X86Assembler::cmpq_rr): (JSC::X86Assembler::cmpq_rm): (JSC::X86Assembler::cmpq_im): (JSC::X86Assembler::testq_i32m): (JSC::X86Assembler::movl_mEAX): (JSC::X86Assembler::movl_i32r): (JSC::X86Assembler::movl_EAXm): (JSC::X86Assembler::movq_rm): (JSC::X86Assembler::movq_mEAX): (JSC::X86Assembler::movq_mr): (JSC::X86Assembler::movq_i64r): (JSC::X86Assembler::movl_mr): (JSC::X86Assembler::X86InstructionFormatter::oneByteOp64): (JSC::X86Assembler::X86InstructionFormatter::immediate64):
interpreter/Interpreter.cpp: (JSC::Interpreter::cti_op_put_by_id_generic): (JSC::Interpreter::cti_op_get_by_id_generic): (JSC::Interpreter::cti_op_put_by_id): (JSC::Interpreter::cti_op_put_by_id_second):
jit/JIT.cpp: (JSC::JIT::privateCompileMainPass): (JSC::JIT::privateCompile): (JSC::JIT::privateCompileCTIMachineTrampolines):
jit/JITCall.cpp: (JSC::JIT::compileOpCallSetupArgs): (JSC::JIT::compileOpCall):
jit/JITPropertyAccess.cpp: (JSC::JIT::compileGetByIdHotPath): (JSC::JIT::compilePutByIdHotPath):
runtime/JSImmediate.h: (JSC::JSImmediate::makeInt):

2:36 PM Changeset in webkit [39341] by cwzwarich@webkit.org

3 edits in trunk/JavaScriptCore

2008-12-16 Cameron Zwarich <zwarich@apple.com>

Reviewed by Darin Adler.

Bug 22869: REGRESSION (r38407): http://news.cnet.com/8301-13579_3-9953533-37.html crashes
<https://bugs.webkit.org/show_bug.cgi?id=22869>
<rdar://problem/6402499>

Before r38407, Structure::m_nameInPrevious was ref'd due to it being
stored in a PropertyMap. However, PropertyMaps are created lazily after
r38407, so Structure::m_nameInPrevious is not necessarily ref'd while
it is being used. Making it a RefPtr instead of a raw pointer fixes
the problem.

Unfortunately, the crash in the bug is rather intermittent, and it is
impossible to add an assertion in UString::Ref::ref() to catch this bug
because some users of UString::Rep deliberately zero out the reference
count. Therefore, there is no layout test accompanying this bug fix.

runtime/Structure.cpp: (JSC::Structure::~Structure): Use get(). (JSC::Structure::materializePropertyMap): Use get(). (JSC::Structure::addPropertyTransitionToExistingStructure): Use get(). (JSC::Structure::addPropertyTransition): Use get().
runtime/Structure.h: Make Structure::m_nameInPrevious a RefPtr instead of a raw pointer.

1:44 PM Changeset in webkit [39340] by pkasting@chromium.org

2 edits in trunk/WebCore

2008-12-16 Peter Kasting <pkasting@google.com>

Reviewed by David Hyatt.

https://bugs.webkit.org/show_bug.cgi?id=22885
Fix memory corruption in GIFImageDecoder.cpp with certain GIFs.

platform/image-decoders/gif/GIFImageDecoder.cpp: (WebCore::GIFImageDecoder::frameComplete):

1:33 PM Changeset in webkit [39339] by Nikolas Zimmermann

2 edits in trunk/JavaScriptCore

Not reviewed. Attempt to fix win build. No 'using namespace WTF' in this file, needs manual WTF:: prefix.
Not sure why the build works as is here.

1:20 PM Changeset in webkit [39338] by Stephanie Lewis

12 edits in trunk/WebCore

2008-12-16 Stephanie Lewis <Stephanie Lewis>

Reviewed by Geoff Garen.

Change HTTPHeaderMap to use an AtomicString as its key.
Shaves ~1MB off of the Mozilla Memory Test
No functionality difference

WebCore.xcodeproj/project.pbxproj:
inspector/InspectorController.cpp: (WebCore::addHeaders):
loader/appcache/ApplicationCacheStorage.cpp: (WebCore::parseHeader):
platform/network/HTTPHeaderMap.h:
platform/network/ResourceRequestBase.cpp: (WebCore::ResourceRequestBase::httpHeaderField): (WebCore::ResourceRequestBase::setHTTPHeaderField): (WebCore::ResourceRequestBase::addHTTPHeaderField):
platform/network/ResourceRequestBase.h:
platform/network/ResourceResponseBase.cpp: (WebCore::ResourceResponseBase::httpHeaderField): (WebCore::ResourceResponseBase::setHTTPHeaderField):
platform/network/ResourceResponseBase.h:
platform/text/StringHash.h: (WebCore::CaseFoldingHash::hash): (WebCore::CaseFoldingHash::equal):
xml/XMLHttpRequest.cpp: (WebCore::isSetCookieHeader): (WebCore::XMLHttpRequest::setRequestHeader): (WebCore::XMLHttpRequest::setRequestHeaderInternal): (WebCore::XMLHttpRequest::getRequestHeader): (WebCore::XMLHttpRequest::getResponseHeader):
xml/XMLHttpRequest.h:

1:15 PM Changeset in webkit [39337] by Nikolas Zimmermann

16 edits
4 adds in trunk

Reviewed by Darin Adler.

Fixes: https://bugs.webkit.org/show_bug.cgi?id=22876

Unify random number generation in JavaScriptCore & WebCore, by introducing
wtf/RandomNumber.h and moving wtf_random/wtf_random_init out of MathExtras.h.

wtf_random_init() has been renamed to initializeRandomNumberGenerator() and
lives in it's own private header: wtf/RandomNumberSeed.h, only intended to
be used from within JavaScriptCore.

wtf_random() has been renamed to randomNumber() and lives in a public header
wtf/RandomNumber.h, usable from within JavaScriptCore & WebCore. It encapsulates
the code taking care of initializing the random number generator (only when
building without ENABLE(JSC_MULTIPLE_THREADS), otherwhise initializeThreading()
already took care of that).

Functional change on darwin: Use random() instead of rand(), as it got a larger
period (more randomness). HTMLFormElement already contains this implementation
and I just moved it in randomNumber(), as special case for PLATFORM(DARWIN).

11:49 AM Changeset in webkit [39336] by andersca@apple.com