Timeline


and

06/05/09:

21:49 Changeset [44479] by jmalonzo@webkit.org

2009-06-05 WANG Lu <coolwanglu@gmail.com>

Reviewed by Jan Alonzo.

[Gtk] WebKit GTK with libsoup won't recognize proxies
https://bugs.webkit.org/show_bug.cgi?id=25263

Added --with-libsoup-gnome (default=no) to enable GNOME-specific
features of libsoup, an important one of which is the proxy-resolver
feature.

  • configure.ac:

2009-06-05 WANG Lu <coolwanglu@gmail.com>

Reviewed by Jan Alonzo.

[Gtk] WebKit GTK with libsoup won't recognize proxies
https://bugs.webkit.org/show_bug.cgi?id=25263

If libsoup-gnome is enabled, proper parameters are added when creating
a new soup session, in order to enable GNOME-specific features of
libsoup.

  • GNUmakefile.am:
  • platform/network/soup/ResourceHandleSoup.cpp: (WebCore::createSoupSession):
19:54 Changeset [44478] by barraclough@apple.com

2009-06-05 Gavin Barraclough <barraclough@apple.com>

Rudder stamped by Sam Weinig.

Add missing ASSERT.

  • assembler/X86Assembler.h: (JSC::X86Assembler::getRelocatedAddress):
19:39 Changeset [44477] by barraclough@apple.com

2009-06-05 Gavin Barraclough <barraclough@apple.com>

Reviewed by Sam Weinig.

Switch storePtrWithPatch to take the initial immediate value as an argument.

  • assembler/MacroAssemblerX86.h: (JSC::MacroAssemblerX86::storePtrWithPatch):
  • assembler/MacroAssemblerX86_64.h: (JSC::MacroAssemblerX86_64::storePtrWithPatch):
  • jit/JITOpcodes.cpp: (JSC::JIT::emit_op_jsr):
17:08 Changeset [44476] by barraclough@apple.com

2009-06-05 Gavin Barraclough <barraclough@apple.com>

Reviewed by Sam Weinig.

Remove patchLength..tByIdExternalLoadPrefix magic numbers from JIT.h.

These aren't really suitable values to be tracking within common code
of the JIT, since they are not (and realistically cannot) be checked
by ASSERTs, as the other repatch offsets are. Move this functionality
(skipping the REX prefix when patching load instructions to LEAs on
x86-64) into the X86Assembler.

  • assembler/AbstractMacroAssembler.h: (JSC::AbstractMacroAssembler::CodeLocationInstruction::repatchLoadPtrToLEA):
  • assembler/X86Assembler.h: (JSC::X86Assembler::repatchLoadPtrToLEA):
  • jit/JIT.h:
  • jit/JITPropertyAccess.cpp: (JSC::JIT::patchGetByIdSelf): (JSC::JIT::patchPutByIdReplace):
16:24 Changeset [44475] by weinig@apple.com

2009-06-05 Sam Weinig <sam@webkit.org>

Reviewed by Anders Carlsson.

Add ononline and onoffline attributes for the <body> element.

  • html/HTMLAttributeNames.in: Added ononlineAttr and onofflineAttr.
  • html/HTMLBodyElement.cpp: (WebCore::HTMLBodyElement::parseMappedAttribute): Map ononlineAttr and onofflineAttr to window event listeners.
13:22 Changeset [44474] by hyatt@apple.com

2009-06-05 David Hyatt <hyatt@apple.com>

Reviewed by Mark Rowe.

https://bugs.webkit.org/show_bug.cgi?id=26220

Update a WebView's active state when it moves to a different window.

  • WebView/WebView.mm: (-[WebView viewDidMoveToWindow]):
10:43 Changeset [44473] by ddkilzer@apple.com

2009-06-05 Kenneth Rohde Christiansen <kenneth.christiansen@openbossa.com>

Bug 26188: Clarify some of the cases in the coding style guide

<https://bugs.webkit.org/show_bug.cgi?id=26188>

Reviewed and tweaked by David Kilzer.

Changed the coding style document to clarify some cases of ambiguity.

  • coding/coding-style.html:
09:02 Changeset [44472] by vestbo@webkit.org

2009-06-05 Tor Arne Vestbø <tor.arne.vestbo@nokia.com>

Reviewed by Simon Hausmann.

[Qt] Add missing includes of config.h

07:53 Changeset [44471] by dglazkov@chromium.org

2009-06-05 Fumitoshi Ukai <ukai@google.com>

Reviewed by Dimitri Glazkov.

https://bugs.webkit.org/show_bug.cgi?id=26215
Try to fix the Chromium build.

  • platform/graphics/chromium/GlyphPageTreeNodeChromiumWin.cpp: (WebCore::fillBMPGlyphs): (WebCore::fillNonBMPGlyphs):
07:53 QtWebKitTodo edited by ariya@webkit.org
(diff)
07:43 Changeset [44470] by ddkilzer@apple.com

JavaScriptCore:

2009-06-05 Shinichiro Hamaji <hamaji@chromium.org>

Bug 26160: Compile fails in MacOSX when GNU fileutils are installed

<https://bugs.webkit.org/show_bug.cgi?id=26160>

Reviewed by Alexey Proskuryakov.

Use /bin/ln instead of ln for cases where this command is used with -h option.
As this option is not supported by GNU fileutils, this change helps users
who have GNU fileutils in their PATH.

WebCore:

2009-06-05 Shinichiro Hamaji <hamaji@chromium.org>

Bug 26160: Compile fails in MacOSX when GNU fileutils are installed

<https://bugs.webkit.org/show_bug.cgi?id=26160>

Reviewed by Alexey Proskuryakov.

Use /bin/ln instead of ln for cases where this command is used with -h option.
As this option is not supported by GNU fileutils, this change helps users
who have GNU fileutils in their PATH.

  • WebCore.xcodeproj/project.pbxproj:
07:10 Changeset [44469] by vestbo@webkit.org

Fix the Qt build after r44413

07:00 Changeset [44468] by oliver@apple.com

2009-06-03 Ben Murdoch <benm@google.com>

<https://bugs.webkit.org/show_bug.cgi?id=25710> HTML5 Database stops executing transactions if the URL hash changes while a transaction is open and an XHR is in progress.

Reviewed by Alexey Proskuryakov.

Fix a bug that causes database transactions to fail if a history navigation to a hash fragment of the same document is made while resources (e.g. an XHR) are loading
https://bugs.webkit.org/show_bug.cgi?id=25710

Test: storage/hash-change-with-xhr.html

  • loader/DocumentLoader.cpp: (WebCore::DocumentLoader::stopLoading):
  • loader/DocumentLoader.h:
  • loader/FrameLoader.cpp: (WebCore::FrameLoader::stopLoading): (WebCore::FrameLoader::stopAllLoaders):
  • loader/FrameLoader.h:
  • loader/FrameLoaderTypes.h: (WebCore::):
  • page/Page.cpp: (WebCore::Page::goToItem):
  • WebCore.base.exp:
05:12 Changeset [44467] by vestbo@webkit.org

2009-06-03 Tor Arne Vestbø <tor.arne.vestbo@nokia.com>

Reviewed by Simon Hausmann.

[Qt] Make sure the correct config.h is included when shadow-building

05:05 Changeset [44466] by vestbo@webkit.org

2009-06-03 Tor Arne Vestbø <tor.arne.vestbo@nokia.com>

Rubber-stamped by Simon Hausmann.

[Qt] Fix Perl warning when calling chdir with an empty/undefined path

This happened when the included makespec was in the same directory as
the parent makespec, so the relative path between the two was empty.

05:03 Changeset [44465] by vestbo@webkit.org

2009-06-05 Tor Arne Vestbø <tor.arne.vestbo@nokia.com>

Reviewed by Simon Hausmann.

Fix Qt build after r44452

04:27 Changeset [44464] by vestbo@webkit.org

Add missing changelog for r44462

04:24 Changeset [44463] by vestbo@webkit.org

Fix the Qt build (missing symbols)

04:24 Changeset [44462] by vestbo@webkit.org

[Qt] Remove some dead code (MenuEventProxy)

03:58 Changeset [44461] by barraclough@apple.com

2009-06-05 Gavin Barraclough <barraclough@apple.com>

Reviewed by Oliver Hunt.

Remove DoubleNotEqual floating point comparison condition for now -
it is not used, and it is unclear the semantics are correct (I think
this comparison would actually give you not-equal-or-unordered, which
might be what is wanted... we can revisit this interface & get it
right when required).

Also, fix asserts in branchArith32 ops. All adds & subs can check
for Signed, multiply only sets OF so can only check for overflow.

  • assembler/MacroAssemblerX86Common.h: (JSC::MacroAssemblerX86Common::): (JSC::MacroAssemblerX86Common::branchAdd32): (JSC::MacroAssemblerX86Common::branchMul32): (JSC::MacroAssemblerX86Common::branchSub32):
03:46 Changeset [44460] by barraclough@apple.com

2009-06-05 Gavin Barraclough <barraclough@apple.com>

Reviewed by Oliver Hunt.

Minor tidy up in JITStubs.

  • jit/JITStubs.cpp: (JSC::StackHack::StackHack):
  • jit/JITStubs.h:
03:15 Changeset [44459] by xan@webkit.org

2009-06-05 Koen Kooi <koen@dominion.thruhere.net>

Reviewed by Xan Lopez.

Build fix for glib unicode backend.

  • wtf/unicode/glib/UnicodeMacrosFromICU.h:
02:31 Changeset [44458] by xan@webkit.org

2009-06-05 Xan Lopez <xlopez@igalia.com>

More build fixes.

  • platform/graphics/gtk/GlyphPageTreeNodePango.cpp: (WebCore::GlyphPage::fill):
02:19 Changeset [44457] by barraclough@apple.com

2009-06-05 Gavin Barraclough <barraclough@apple.com>

Reviewed by Oliver Hunt.

3 tiny cleanups:

  • assembler/MacroAssemblerX86.h:
  • assembler/MacroAssemblerX86_64.h: (JSC::MacroAssemblerX86_64::storePtrWithPatch):

store*() methods should take an ImplicitAddress, rather than an Address.

  • assembler/X86Assembler.h:

Make patchPointer private.

  • jit/JITOpcodes.cpp: (JSC::JIT::emit_op_ret):

Remove empty line at end of function.

02:16 Changeset [44456] by xan@webkit.org

2009-06-05 Xan Lopez <xlopez@igalia.com>

Fix the GTK+ build.

  • GNUmakefile.am:
  • platform/graphics/gtk/FontCacheGtk.cpp: (WebCore::FontCache::getFontDataForCharacters):
  • platform/graphics/gtk/GlyphPageTreeNodeGtk.cpp: (WebCore::GlyphPage::fill):
00:55 Changeset [44455] by barraclough@apple.com

2009-06-05 Gavin Barraclough <barraclough@apple.com>

Reviewed by Oliver Hunt.

Encapsulate many uses of void* in the assembler & jit with types that provide
more semantic information. The new types are:


  • MacroAssemblerCodePtr - this wraps a pointer into JIT generated code.
  • FunctionPtr - this wraps a pointer to a C/C++ function in JSC.
  • ReturnAddressPtr - this wraps a return address resulting from a 'call' instruction.

Wrapping these types allows for stronger type-checking than is possible with everything
represented a void*. For example, it is now enforced by the type system that near
calls can only be linked to JIT code and not to C functions in JSC (this was previously
required, but could not be enforced on the interface).

  • assembler/AbstractMacroAssembler.h: (JSC::AbstractMacroAssembler::CodeLocationCommon::CodeLocationCommon): (JSC::AbstractMacroAssembler::CodeLocationCommon::dataLocation): (JSC::AbstractMacroAssembler::CodeLocationCommon::executableAddress): (JSC::AbstractMacroAssembler::CodeLocationCommon::reset): (JSC::AbstractMacroAssembler::CodeLocationInstruction::repatchLoadToLEA): (JSC::AbstractMacroAssembler::CodeLocationInstruction::CodeLocationInstruction): (JSC::AbstractMacroAssembler::CodeLocationLabel::addressForSwitch): (JSC::AbstractMacroAssembler::CodeLocationLabel::addressForExceptionHandler): (JSC::AbstractMacroAssembler::CodeLocationLabel::addressForJSR): (JSC::AbstractMacroAssembler::CodeLocationLabel::operator!): (JSC::AbstractMacroAssembler::CodeLocationLabel::reset): (JSC::AbstractMacroAssembler::CodeLocationLabel::CodeLocationLabel): (JSC::AbstractMacroAssembler::CodeLocationLabel::getJumpDestination): (JSC::AbstractMacroAssembler::CodeLocationJump::relink): (JSC::AbstractMacroAssembler::CodeLocationJump::CodeLocationJump): (JSC::AbstractMacroAssembler::CodeLocationCall::relink): (JSC::AbstractMacroAssembler::CodeLocationCall::calleeReturnAddressValue): (JSC::AbstractMacroAssembler::CodeLocationCall::CodeLocationCall): (JSC::AbstractMacroAssembler::CodeLocationNearCall::relink): (JSC::AbstractMacroAssembler::CodeLocationNearCall::calleeReturnAddressValue): (JSC::AbstractMacroAssembler::CodeLocationNearCall::CodeLocationNearCall): (JSC::AbstractMacroAssembler::CodeLocationDataLabel32::repatch): (JSC::AbstractMacroAssembler::CodeLocationDataLabel32::CodeLocationDataLabel32): (JSC::AbstractMacroAssembler::CodeLocationDataLabelPtr::repatch): (JSC::AbstractMacroAssembler::CodeLocationDataLabelPtr::CodeLocationDataLabelPtr): (JSC::AbstractMacroAssembler::ProcessorReturnAddress::relinkCallerToTrampoline): (JSC::AbstractMacroAssembler::ProcessorReturnAddress::relinkCallerToFunction): (JSC::AbstractMacroAssembler::ProcessorReturnAddress::relinkNearCallerToTrampoline): (JSC::AbstractMacroAssembler::ProcessorReturnAddress::addressForLookup): (JSC::AbstractMacroAssembler::trampolineAt): (JSC::AbstractMacroAssembler::PatchBuffer::link): (JSC::AbstractMacroAssembler::PatchBuffer::performFinalization): (JSC::::CodeLocationCommon::instructionAtOffset): (JSC::::CodeLocationCommon::labelAtOffset): (JSC::::CodeLocationCommon::jumpAtOffset): (JSC::::CodeLocationCommon::callAtOffset): (JSC::::CodeLocationCommon::nearCallAtOffset): (JSC::::CodeLocationCommon::dataLabelPtrAtOffset): (JSC::::CodeLocationCommon::dataLabel32AtOffset):
  • assembler/MacroAssemblerCodeRef.h: (JSC::FunctionPtr::FunctionPtr): (JSC::FunctionPtr::value): (JSC::FunctionPtr::executableAddress): (JSC::ReturnAddressPtr::ReturnAddressPtr): (JSC::ReturnAddressPtr::value): (JSC::MacroAssemblerCodePtr::MacroAssemblerCodePtr): (JSC::MacroAssemblerCodePtr::executableAddress): (JSC::MacroAssemblerCodePtr::dataLocation): (JSC::MacroAssemblerCodeRef::MacroAssemblerCodeRef):
  • assembler/X86Assembler.h: (JSC::X86Assembler::patchPointerForCall):
  • jit/JIT.cpp: (JSC::ctiPatchNearCallByReturnAddress): (JSC::ctiPatchCallByReturnAddress): (JSC::JIT::privateCompile): (JSC::JIT::privateCompileCTIMachineTrampolines):
  • jit/JIT.h: (JSC::JIT::compileCTIMachineTrampolines):
  • jit/JITCall.cpp: (JSC::JIT::compileOpCall):
  • jit/JITCode.h: (JSC::JITCode::operator !): (JSC::JITCode::addressForCall): (JSC::JITCode::offsetOf): (JSC::JITCode::execute): (JSC::JITCode::size): (JSC::JITCode::HostFunction):
  • jit/JITInlineMethods.h: (JSC::JIT::emitNakedCall):
  • jit/JITPropertyAccess.cpp: (JSC::JIT::privateCompilePutByIdTransition): (JSC::JIT::patchGetByIdSelf): (JSC::JIT::patchPutByIdReplace): (JSC::JIT::privateCompilePatchGetArrayLength): (JSC::JIT::privateCompileGetByIdProto): (JSC::JIT::privateCompileGetByIdChain):
  • jit/JITStubs.cpp: (JSC::JITThunks::JITThunks): (JSC::JITThunks::tryCachePutByID): (JSC::JITThunks::tryCacheGetByID): (JSC::JITStubs::DEFINE_STUB_FUNCTION):
  • jit/JITStubs.h: (JSC::JITThunks::ctiArrayLengthTrampoline): (JSC::JITThunks::ctiStringLengthTrampoline): (JSC::JITThunks::ctiVirtualCallPreLink): (JSC::JITThunks::ctiVirtualCallLink): (JSC::JITThunks::ctiVirtualCall): (JSC::JITThunks::ctiNativeCallThunk):
  • yarr/RegexJIT.h: (JSC::Yarr::RegexCodeBlock::operator!): (JSC::Yarr::RegexCodeBlock::execute):
00:51 Changeset [44454] by antti@apple.com

2009-06-05 Antti Koivisto <antti@apple.com>

Try to unbreak Windows build.

00:36 Changeset [44453] by antti@apple.com

2009-06-05 Antti Koivisto <antti@apple.com>

Try to fix Windows (and possibly other platforms) build.

Restore ResourceResponseBase::lastModifiedDate() and setLastModifiedDate() removed in previous commit.
for now since PluginStream used on some platforms expects them and calculations differ from plain
Last-modified header value.


Also include <wtf/MathExtras.h> to get isfinite().

  • platform/network/ResourceResponseBase.cpp: (WebCore::ResourceResponseBase::adopt): (WebCore::ResourceResponseBase::copyData): (WebCore::ResourceResponseBase::setLastModifiedDate): (WebCore::ResourceResponseBase::lastModifiedDate):
  • platform/network/ResourceResponseBase.h:
  • platform/network/cf/ResourceResponseCFNet.cpp: (WebCore::ResourceResponse::platformLazyInit):

06/04/09:

23:24 Changeset [44452] by antti@apple.com

JavaScriptCore:

2009-06-03 Antti Koivisto <antti@apple.com>

Reviewed by Dave Kilzer.


https://bugs.webkit.org/show_bug.cgi?id=13128
Safari not obeying cache header


Export JSC::parseDate()

WebCore:

2009-06-03 Antti Koivisto <antti@apple.com>

Reviewed by Dave Kilzer.

https://bugs.webkit.org/show_bug.cgi?id=13128
Safari not obeying cache header


Implement RFC 2616 cache expiration calculations in WebKit instead of
relying on the networking layer.

  • ForwardingHeaders/runtime/DateMath.h: Added.
  • WebCore.base.exp:
  • loader/Cache.cpp: (WebCore::Cache::revalidationSucceeded):
  • loader/CachedResource.cpp: (WebCore::CachedResource::CachedResource): (WebCore::CachedResource::isExpired): (WebCore::CachedResource::currentAge): (WebCore::CachedResource::freshnessLifetime): (WebCore::CachedResource::setResponse): (WebCore::CachedResource::updateResponseAfterRevalidation): (WebCore::CachedResource::mustRevalidate):
  • loader/CachedResource.h:
  • platform/network/ResourceResponseBase.cpp: (WebCore::ResourceResponseBase::ResourceResponseBase): (WebCore::ResourceResponseBase::adopt): (WebCore::ResourceResponseBase::copyData): (WebCore::ResourceResponseBase::setHTTPHeaderField): (WebCore::ResourceResponseBase::parseCacheControlDirectives): (WebCore::ResourceResponseBase::cacheControlContainsNoCache): (WebCore::ResourceResponseBase::cacheControlContainsMustRevalidate): (WebCore::ResourceResponseBase::cacheControlMaxAge): (WebCore::parseDateValueInHeader): (WebCore::ResourceResponseBase::date): (WebCore::ResourceResponseBase::age): (WebCore::ResourceResponseBase::expires): (WebCore::ResourceResponseBase::lastModified): (WebCore::ResourceResponseBase::isAttachment): (WebCore::ResourceResponseBase::compare):
  • platform/network/ResourceResponseBase.h:
  • platform/network/cf/ResourceResponseCFNet.cpp: (WebCore::ResourceResponse::platformLazyInit):
  • platform/network/mac/ResourceResponseMac.mm: (WebCore::ResourceResponse::platformLazyInit):

LayoutTests:

2009-06-03 Antti Koivisto <antti@apple.com>

Reviewed by Dave Kilzer.


https://bugs.webkit.org/show_bug.cgi?id=13128
Safari not obeying cache header


Test subresource caching and expiration.

  • http/tests/cache: Added.
  • http/tests/cache/resources: Added.
  • http/tests/cache/resources/cache-simulator.cgi: Added.
  • http/tests/cache/resources/subresource-test.js: Added. (frame1.onload): (frame2.onload): (loadTestFrame.scr.onload): (nextTest): (testComplete): (runTests):
  • http/tests/cache/resources/test-frame.html: Added.
  • http/tests/cache/subresource-expiration-expected.txt: Added.
  • http/tests/cache/subresource-expiration.html: Added.
22:29 Changeset [44451] by weinig@apple.com

WebKit:

2009-06-04 Sam Weinig <sam@webkit.org>

Reviewed by Alice Liu.

Move WebView internal data into WebViewData.h/mm.

  • WebKit.xcodeproj/project.pbxproj:

WebKit/mac:

2009-06-04 Sam Weinig <sam@webkit.org>

Reviewed by Alice Liu.

Move WebView internal data into WebViewData.h/mm.

  • WebView/WebDelegateImplementationCaching.mm:
  • WebView/WebView.mm:
  • WebView/WebViewData.h: Copied from mac/WebView/WebViewInternal.h.
  • WebView/WebViewData.mm: Copied from mac/WebView/WebViewInternal.mm.
  • WebView/WebViewInternal.h:
  • WebView/WebViewInternal.mm: Removed.
20:45 Changeset [44450] by eric@webkit.org

2009-06-04 Roland Steiner <rolandsteiner@google.com>

Reviewed by Eric Seidel.

Bug 26201: Remove superfluous 'if' statements in RenderTable::addChild
https://bugs.webkit.org/show_bug.cgi?id=26201

  • rendering/RenderTable.cpp: (WebCore::RenderTable::addChild): remove superfluous 'if' statements
20:31 Changeset [44449] by eric@webkit.org

2009-06-04 Roland Steiner <rolandsteiner@google.com>

Reviewed by Eric Seidel.

Bug 26202: add macros for primitive values to simplify CSSStyleSelector::applyProperty
https://bugs.webkit.org/show_bug.cgi?id=26202

  • css/CSSStyleSelector.cpp: add HANDLE_INHERIT_AND_INITIAL_AND_PRIMITIVE[_WITH_VALUE] macros (WebCore::CSSStyleSelector::applyProperty): use new macros
20:30 Changeset [44448] by eric@webkit.org

2009-06-04 Roland Steiner <rolandsteiner@google.com>

Reviewed by Eric Seidel

Bug 26203: Move parsing of 'attr(X)' values to own method
https://bugs.webkit.org/show_bug.cgi?id=26203

  • css/CSSParser.cpp: (WebCore::CSSParser::parseAttr): new method (WebCore::CSSParser::parseContent): use new parseAttr method
  • css/CSSParser.h: (WebCore::CSSParser::parseAttr): new method
20:25 Changeset [44447] by eric@webkit.org

2009-06-04 Roland Steiner <rolandsteiner@google.com>

Reviewed by Eric Seidel.

Bug 26205: RenderTableSection::addChild : correct comment
https://bugs.webkit.org/show_bug.cgi?id=26205

  • rendering/RenderTableSection.cpp: (WebCore::RenderTableSection::addChild): correct comment
20:22 Changeset [44446] by eric@webkit.org

2009-06-04 Roland Steiner <rolandsteiner@google.com>

Reviewed by Eric Seidel.

Bug 26204: RenderBlock : simplify handleSpecialChild, comment correction
https://bugs.webkit.org/show_bug.cgi?id=26204

  • rendering/RenderBlock.cpp: (WebCore::RenderBlock::handleSpecialChild): simplify usage (WebCore::RenderBlock::handlePositionedChild): simplify usage (WebCore::RenderBlock::handleFloatingChild): simplify usage (WebCore::RenderBlock::handleRunInChild): simplify usage (WebCore::RenderBlock::layoutBlock): correct comment (WebCore::RenderBlock::layoutBlockChildren): simplify loop, change call to handleSpecialChild
  • rendering/RenderBlock.h: (WebCore::RenderBlock::handleSpecialChild): change signature (WebCore::RenderBlock::handlePositionedChild): change signature (WebCore::RenderBlock::handleFloatingChild): change signature (WebCore::RenderBlock::handleRunInChild): change signature
18:27 Changeset [44445] by oliver@apple.com

Bug in property caching of getters and setters.

Reviewed by Gavin Barraclough

Make sure that the transition logic accounts for getters and setters.
If we don't we end up screwing up the transition tables so that some
transitions will start incorrectly believing that they need to check
for getters and setters.

17:58 Changeset [44444] by mitz@apple.com
  • retry to fix the Tiger build
  • platform/graphics/mac/SimpleFontDataMac.mm: (WebCore::initFontData):
17:36 Changeset [44443] by barraclough@apple.com

2009-06-04 Gavin Barraclough <barraclough@apple.com>

Reviewed by Sam Weinig.

Minor tweak to PatchBuffer, change it so it no longer holds a CodeRef, and instead
holds a separate code pointer and executable pool. Since it now always holds its
own copy of the code size, and to simplify the construction sequence, it's neater
this way.

  • assembler/AbstractMacroAssembler.h: (JSC::AbstractMacroAssembler::PatchBuffer::PatchBuffer): (JSC::AbstractMacroAssembler::PatchBuffer::finalizeCode): (JSC::AbstractMacroAssembler::PatchBuffer::code): (JSC::AbstractMacroAssembler::PatchBuffer::performFinalization):
17:36 Changeset [44442] by mitz@apple.com
  • try to fix the Tiger build
  • platform/graphics/SimpleFontData.h:
17:26 Changeset [44441] by mitz@apple.com
  • try to fix the Leopard and Tiger builds
  • platform/graphics/SimpleFontData.h:
17:24 Changeset [44440] by mitz@apple.com
  • try to fix the Windows build
  • platform/graphics/win/UniscribeController.cpp: (WebCore::UniscribeController::shapeAndPlaceItem):
17:11 Changeset [44439] by mitz@apple.com

Reviewed by Sam Weinig.

  • make SimpleFontData's data members private
  • rename SimpleFontData's m_font member to m_platformData
  • platform/graphics/Font.h: (WebCore::Font::spaceWidth):
  • platform/graphics/SimpleFontData.cpp: (WebCore::SimpleFontData::SimpleFontData):
  • platform/graphics/SimpleFontData.h: (WebCore::SimpleFontData::platformData): (WebCore::SimpleFontData::spaceWidth): (WebCore::SimpleFontData::adjustedSpaceWidth): (WebCore::SimpleFontData::syntheticBoldOffset): (WebCore::SimpleFontData::spaceGlyph): (WebCore::SimpleFontData::getNSFont): (WebCore::SimpleFontData::getQtFont): (WebCore::SimpleFontData::getWxFont):
  • platform/graphics/WidthIterator.cpp: (WebCore::WidthIterator::advance):
  • platform/graphics/cairo/FontCairo.cpp: (WebCore::Font::drawGlyphs):
  • platform/graphics/chromium/SimpleFontDataChromiumWin.cpp: (WebCore::SimpleFontData::platformInit): (WebCore::SimpleFontData::smallCapsFontData): (WebCore::SimpleFontData::determinePitch): (WebCore::SimpleFontData::platformWidthForGlyph):
  • platform/graphics/chromium/SimpleFontDataLinux.cpp: (WebCore::SimpleFontData::platformInit): (WebCore::SimpleFontData::smallCapsFontData): (WebCore::SimpleFontData::containsCharacters): (WebCore::SimpleFontData::platformWidthForGlyph):
  • platform/graphics/gtk/FontGtk.cpp: (WebCore::setPangoAttributes):
  • platform/graphics/gtk/SimpleFontDataGtk.cpp: (WebCore::SimpleFontData::platformInit): (WebCore::SimpleFontData::containsCharacters): (WebCore::SimpleFontData::determinePitch): (WebCore::SimpleFontData::platformWidthForGlyph): (WebCore::SimpleFontData::setFont):
  • platform/graphics/gtk/SimpleFontDataPango.cpp: (WebCore::SimpleFontData::platformInit): (WebCore::SimpleFontData::containsCharacters): (WebCore::SimpleFontData::determinePitch): (WebCore::SimpleFontData::platformWidthForGlyph): (WebCore::SimpleFontData::setFont):
  • platform/graphics/mac/CoreTextController.cpp: (WebCore::CoreTextController::adjustGlyphsAndAdvances):
  • platform/graphics/mac/FontMac.mm: (WebCore::Font::drawGlyphs):
  • platform/graphics/mac/FontMacATSUI.mm: (WebCore::initializeATSUStyle): (WebCore::overrideLayoutOperation): (WebCore::ATSULayoutParameters::initialize):
  • platform/graphics/mac/SimpleFontDataMac.mm: (WebCore::initFontData): (WebCore::SimpleFontData::platformInit): (WebCore::SimpleFontData::platformCharWidthInit): (WebCore::SimpleFontData::smallCapsFontData): (WebCore::SimpleFontData::containsCharacters): (WebCore::SimpleFontData::determinePitch): (WebCore::SimpleFontData::platformWidthForGlyph): (WebCore::SimpleFontData::checkShapesArabic): (WebCore::SimpleFontData::getCTFont):
  • platform/graphics/qt/SimpleFontDataQt.cpp: (WebCore::SimpleFontData::determinePitch): (WebCore::SimpleFontData::platformInit): (WebCore::SimpleFontData::platformCharWidthInit):
  • platform/graphics/win/FontCGWin.cpp: (WebCore::drawGDIGlyphs): (WebCore::Font::drawGlyphs):
  • platform/graphics/win/SimpleFontDataCGWin.cpp: (WebCore::SimpleFontData::platformInit): (WebCore::SimpleFontData::platformCharWidthInit): (WebCore::SimpleFontData::platformWidthForGlyph):
  • platform/graphics/win/SimpleFontDataCairoWin.cpp: (WebCore::SimpleFontData::platformInit): (WebCore::SimpleFontData::platformDestroy): (WebCore::SimpleFontData::platformWidthForGlyph): (WebCore::SimpleFontData::setFont):
  • platform/graphics/win/SimpleFontDataWin.cpp: (WebCore::SimpleFontData::initGDIFont): (WebCore::SimpleFontData::smallCapsFontData): (WebCore::SimpleFontData::containsCharacters): (WebCore::SimpleFontData::determinePitch): (WebCore::SimpleFontData::widthForGDIGlyph): (WebCore::SimpleFontData::scriptFontProperties):
  • platform/graphics/wx/SimpleFontDataWx.cpp: (WebCore::SimpleFontData::platformInit): (WebCore::SimpleFontData::determinePitch): (WebCore::SimpleFontData::platformWidthForGlyph):
16:01 Changeset [44438] by eric@webkit.org

2009-06-04 Paul Godavari <paul@chromium.org>

Reviewed by Eric Seidel.

Initialize the width of PopupMenuListBox properly for Mac Chromium.

Added a test that works only with this patch applied. The test is
a manual one, since the hit testing infrastructure in the layout
tests sends keyboard and mouse events to the main window and not
the cocoa control that implements the popup up, which means we can't
select items from the popup up.

https://bugs.webkit.org/show_bug.cgi?id=25904

  • manual-tests/select-narrow-width.html: Added.
  • platform/chromium/PopupMenuChromium.cpp: (WebCore::PopupListBox::PopupListBox): (WebCore::PopupContainer::showExternal):
15:52 Changeset [44437] by barraclough@apple.com

2009-06-04 Gavin Barraclough <barraclough@apple.com>

Reviewed by Oliver Hunt.

Remove 'JIT_STUB_ARGUMENT_STACK' this is unused and untested.

This just leaves JIT_STUB_ARGUMENT_REGISTER and JIT_STUB_ARGUMENT_VA_LIST.
Since JIT_STUB_ARGUMENT_REGISTER is the sensible configuration on most platforms,
remove this define and make this the default behaviour.
Platforms must now define JIT_STUB_ARGUMENT_VA_LIST to get crazy va_list voodoo,
if they so desire.

(Refactoring of #ifdefs only, no functional change, no performance impact.)

  • jit/JIT.h:
  • jit/JITInlineMethods.h: (JSC::JIT::restoreArgumentReference): (JSC::JIT::restoreArgumentReferenceForTrampoline):
  • jit/JITStubs.cpp: (JSC::):
  • jit/JITStubs.h:
  • wtf/Platform.h:
15:34 Changeset [44436] by bfulgham@webkit.org

2009-06-04 Brent Fulgham <bfulgham@webkit.org>

Unreviewed Windows build correction.

  • WebCore.vcproj/WebCore.vcproj: Add missing 'ReplaceNodeWithSpanCommand.cpp' and 'ReplaceNodeWithSpanCommand.h'
15:17 Changeset [44435] by eric@webkit.org

2009-02-03 Eric Seidel <eric@webkit.org>

Reviewed by Justin Garcia.

Make sure execCommand("bold") on <b style="text-decoration: underline">test</b>
only removes the bold and not the underline.
https://bugs.webkit.org/show_bug.cgi?id=23496

Test: editing/execCommand/convert-style-elements-to-spans.html

  • WebCore.xcodeproj/project.pbxproj:
  • css/CSSStyleDeclaration.h: (WebCore::CSSStyleDeclaration::isEmpty):
  • dom/NamedAttrMap.h: (WebCore::NamedAttrMap::isEmpty):
  • editing/ApplyStyleCommand.cpp: (WebCore::isUnstyledStyleSpan): (WebCore::isSpanWithoutAttributesOrUnstyleStyleSpan): (WebCore::ApplyStyleCommand::applyBlockStyle): (WebCore::ApplyStyleCommand::applyRelativeFontStyleChange): (WebCore::ApplyStyleCommand::implicitlyStyledElementShouldBeRemovedWhenApplyingStyle): (WebCore::ApplyStyleCommand::replaceWithSpanOrRemoveIfWithoutAttributes): (WebCore::ApplyStyleCommand::removeCSSStyle): (WebCore::ApplyStyleCommand::applyTextDecorationStyle): (WebCore::ApplyStyleCommand::removeInlineStyle): (WebCore::ApplyStyleCommand::addInlineStyleIfNeeded):
  • editing/ApplyStyleCommand.h:
  • editing/CompositeEditCommand.cpp: (WebCore::CompositeEditCommand::replaceNodeWithSpanPreservingChildrenAndAttributes):
  • editing/CompositeEditCommand.h:
  • editing/RemoveNodePreservingChildrenCommand.cpp: (WebCore::RemoveNodePreservingChildrenCommand::RemoveNodePreservingChildrenCommand):
  • editing/ReplaceNodeWithSpanCommand.cpp: Added. (WebCore::ReplaceNodeWithSpanCommand::ReplaceNodeWithSpanCommand): (WebCore::swapInNodePreservingAttributesAndChildren): (WebCore::ReplaceNodeWithSpanCommand::doApply): (WebCore::ReplaceNodeWithSpanCommand::doUnapply):
  • editing/ReplaceNodeWithSpanCommand.h: Added. (WebCore::ReplaceNodeWithSpanCommand::create):
14:51 Changeset [44434] by bfulgham@webkit.org

2009-06-04 Brent Fulgham <bfulgham@webkit.org>

Unreviewed build fix for Windows Cairo target.

Add missing post-build command to copy history/cf contents
to output directory.

  • WebCore.vcproj/WebCore.vcproj: Update Debug_Cairo and Release_Cairo target post-build steps with copy commands.
14:51 Changeset [44433] by barraclough@apple.com

2009-06-04 Gavin Barraclough <barraclough@apple.com>

Rubber stamped by Sam Weinig.

  • jit/JITArithmetic.cpp:

Remove some redundant typedefs, unused since arithmetic was added to the MacroAssembler interface.

14:31 Changeset [44432] by eric.carlson@apple.com

2009-06-04 Eric Carlson <eric.carlson@apple.com>

Reviewed by Simon Fraser.

<rdar://problem/6854695> Movie controller thumb fails to scale with full page zoom


Account for zoom level when drawing media controller thumb on Windows.

  • rendering/RenderMediaControls.cpp: (WebCore::RenderMediaControls::adjustMediaSliderThumbSize):
14:00 Changeset [44431] by mrowe@apple.com

Versioning.

13:59 Changeset [44430] by mrowe@apple.com

New tag.

13:55 Changeset [44429] by hyatt@apple.com

WebCore:

2009-06-04 David Hyatt <hyatt@apple.com>

Reviewed by Sam Weinig.

Move DOM window focus/blur out of SelectionController and into FocusController. Make sure it
fires on the focused frame when the page activation state changes also. This is covered by an existing
layout test (albeit badly). I have modified the test to be correct.

  • editing/SelectionController.cpp: (WebCore::SelectionController::setFocused):
  • page/FocusController.cpp: (WebCore::FocusController::setFocusedFrame): (WebCore::FocusController::setActive):

WebKit/mac:

2009-06-04 David Hyatt <hyatt@apple.com>

Reviewed by Sam Weinig.

Remove _updateFocusedStateForFrame, since it's actually not even necessary now that I made
setFocusedFrame get called explicitly from become/ResignFirstResponder. setFocusedFrame does
the work of focusing the selection already.

  • WebCoreSupport/WebFrameLoaderClient.mm: (WebFrameLoaderClient::transitionToCommittedForNewPage):
  • WebView/WebHTMLView.mm: (-[WebHTMLView becomeFirstResponder]): (-[WebHTMLView resignFirstResponder]):
  • WebView/WebView.mm:
  • WebView/WebViewInternal.h:

LayoutTests:

2009-06-04 David Hyatt <hyatt@apple.com>

Reviewed by Sam Weinig.

Fix the Window focus test to not resign first responder, since that actually made the test
incorrect. An unfocused WebHTMLView's DOM window should not receive focus/blur events when the activation
changes.

Now the test just checks that a first responder WebHTMLView will actually fire focus/blur on the DOM window
as the window gains/loses key.

  • fast/dom/Window/window-onFocus.html:
13:47 Changeset [44428] by mrowe@apple.com

Versioning.

13:47 Changeset [44427] by mrowe@apple.com

Merge r44286.

13:44 Changeset [44426] by eric@webkit.org

2009-06-04 Albert J. Wong <ajwong@chromium.org>

Reviewed by Eric Seidel.

https://bugs.webkit.org/show_bug.cgi?id=26148
Adding in empty files to stage the extract of RenderThemeChromiumSkia
from RenderThemeChromiumLinux and RenderThemeChromiumWindows.

  • rendering/RenderThemeChromiumSkia.cpp: Added.
  • rendering/RenderThemeChromiumSkia.h: Added.
13:39 Changeset [44425] by darin@apple.com

2009-06-04 Darin Adler <darin@apple.com>

Reviewed by Tim Hatcher.

Fix crashes due to unbalanced start/stop observer calls.

  • DumpRenderTree/mac/DumpRenderTreeWindow.h: Added observingWebView.
  • DumpRenderTree/mac/DumpRenderTreeWindow.mm: (-[DumpRenderTreeWindow startObservingWebView]): Call stopObservingWebView in case we're already observing some other WebView. Set observingWebView. (-[DumpRenderTreeWindow stopObservingWebView]): Do nothing if observingWebView is already NO. Clear observingWebView.
13:07 Changeset [44424] by bfulgham@webkit.org

2009-06-04 Andrei Popescu <andreip@google.com>

Reviewed by Alexey Proskuryakov.

https://bugs.webkit.org/show_bug.cgi?id=25562
Potential crash after ApplicationCacheStorage::storeNewestCache() fails

Fix the crash by checking the return value of cacheStorage().storeNewestCache(this)
in WebCore::ApplicationCacheGroup::checkIfLoadIsComplete. If storeNewestCache failed,
we run the cache failure steps:

  1. Fire the error events to all pending master entries, as well any other cache hosts currently associated with a cache in this group.
  2. Disassociate the pending master entries from the failed new cache.
  3. Reinstate the old "newest cache", if there was one.

We also introduce two other changes:

  1. a mechanism to rollback storageID changes to the in-memory resource objects when the storing of an ApplicationCache object fails.
  1. defer removing the pending master entries from the list of pending master entries until the entire load is complete. This matches the HTML 5 spec better. To track if the load is complete we now introduce a counter for those pending master entries that haven't yet finshed downloading.
  • loader/appcache/ApplicationCacheGroup.cpp: (WebCore::ApplicationCacheGroup::ApplicationCacheGroup): initializes the new counter to 0 (WebCore::ApplicationCacheGroup::selectCache): increments the counter when a new pending master entry is added. (WebCore::ApplicationCacheGroup::finishedLoadingMainResource): decrements the counter instead of removing the pending master entry. (WebCore::ApplicationCacheGroup::failedLoadingMainResource): decrements the counter instead of removing the pending master entry. (WebCore::ApplicationCacheGroup::setNewestCache): removes an assertion that no longer holds true. In particular, the newest cache is not necessarily new anymore. We can set an old cache as the new cache. This can happen if we failed to store a newly downloaded cache to the database and we are now reinstating the former newest cache. (WebCore::ApplicationCacheGroup::manifestNotFound): resets the counter to 0. (WebCore::ApplicationCacheGroup::checkIfLoadIsComplete): check if the loading is complete based on the counter instead of the list of pending master entries. Empty the list of master entries if the load is complete.
  • loader/appcache/ApplicationCacheGroup.h: add the new counter.
  • loader/appcache/ApplicationCacheStorage.cpp: introduce the journaling mechanism for in-memory resource objects. (WebCore::ResourceStorageIDJournal::~ResourceStorageIDJournal): (WebCore::ResourceStorageIDJournal::add): (WebCore::ResourceStorageIDJournal::commit): (WebCore::ResourceStorageIDJournal::Record::Record): (WebCore::ResourceStorageIDJournal::Record::restore): (WebCore::ApplicationCacheStorage::store): log the changes to the in-memory resource objects. (WebCore::ApplicationCacheStorage::storeNewestCache): create the journal object.
  • loader/appcache/ApplicationCacheStorage.h: modify the signature of bool store(ApplicationCache*) to add a pointer to the logger object used to trace the changes to the storageID of the resource objects.
12:29 Changeset [44423] by bfulgham@webkit.org

2009-06-04 Jeremy Orlow <jorlow@chromium.org>

Reviewed by Darin Adler.

https://bugs.webkit.org/show_bug.cgi?id=26154
Allow underscores in the hostnames we parse out of databaseIdentifiers.
This code is used for HTML 5 database support.

  • page/SecurityOrigin.cpp: (WebCore::SecurityOrigin::createFromDatabaseIdentifier):
11:44 Changeset [44422] by bfulgham@webkit.org

2009-06-04 Mihnea Ovidenie <mihnea@adobe.com>

Reviewed by Darin Adler.

Bug 26084: Multiple missing images in webkit-mask-image prevent rendering
https://bugs.webkit.org/show_bug.cgi?id=26084

When painting multiple images, make sure that at least one image is valid before pushing a transparency layer.

Added a manual test.

  • manual-tests/mask-composite-missing-images.html: Added.
  • rendering/RenderBox.cpp: (WebCore::RenderBox::paintMaskImages):
11:40 Changeset [44421] by mitz@apple.com
  • build fix
  • WebView/WebViewInternal.mm:
11:34 Changeset [44420] by bfulgham@webkit.org

2009-06-04 Jeremy Orlow <jorlow@chromium.org>

Reviewed by Darin Adler.

https://bugs.webkit.org/show_bug.cgi?id=26180
Add a fast path for SecurityOrigin::equal. If "other == this" (where
other is the other security origin), then we really don't need to do
all the other (expensive) comparisons. We know it's equal.

  • page/SecurityOrigin.cpp: (WebCore::SecurityOrigin::equal):
11:03 Changeset [44419] by dimich@chromium.org

2009-06-05 Dmitry Titov <dimich@chromium.org>

Reviewed by Eric Seidel.

https://bugs.webkit.org/show_bug.cgi?id=26192
Fix the expected results for editing/execCommand/list-wrapping-image-crash.html

  • editing/execCommand/list-wrapping-image-crash-expected.txt:
  • editing/execCommand/list-wrapping-image-crash.html:
10:44 Changeset [44418] by bfulgham@webkit.org

2009-06-04 Brent Fulgham <bfulgham@webkit.org>

Build fix due to header include problem.

  • interpreter/Interpreter.h: Remove wtf from includes so that compile can find the headers in expected places.
10:30 Changeset [44417] by bfulgham@webkit.org

2009-06-04 Zoltan Horvath <hzoltan@inf.u-szeged.hu>

Reviewed by Darin Adler.

HashTable class (JavaScriptCore/wtf/HashTable.h) doesn't instantiated by 'new', so
inheritance was removed. HashTable struct has been instantiated by operator new in
JSGlobalData.cpp:106.
HashTable couldn't inherited from FastAllocBase since struct with inheritance is
no longer POD, so HashTable struct has been instantiated by fastNew, destroyed by
fastDelete.

  • interpreter/Interpreter.h:
  • runtime/JSGlobalData.cpp: (JSC::JSGlobalData::JSGlobalData): (JSC::JSGlobalData::~JSGlobalData):
  • wtf/HashTable.h:
09:16 Changeset [44416] by hyatt@apple.com

WebCore:

2009-06-03 David Hyatt <hyatt@apple.com>

Reviewed by Sam Weinig.

Improvements in how selection behaves with focus/activation and a reversion back to using isActive
in the scrollbar theme code to remove a Chromium ifdef.

  • editing/SelectionController.cpp: (WebCore::SelectionController::SelectionController): Make the controller set its focused state correctly upon initial creation.

(WebCore::SelectionController::setSelection):
Make selection willing to shift the focus node if the selection is focused even if the
selection is not active. Whether or not the Page is active is irrelevant to focus changes.

(WebCore::SelectionController::setFocused):

  • editing/SelectionController.h: (WebCore::SelectionController::isFocused): Add a new isFocused() method so that code can check if the Selection is focused without caring about the active state.
  • page/Frame.cpp: (WebCore::Frame::setFocusedNodeIfNeeded): Allow focus shifts even when the selection is not active.
  • platform/mac/ScrollbarThemeMac.mm: (WebCore::ScrollbarThemeMac::paint): Revert Dan's change to directly talk to AppKit for checking active state. Now that the WebCore isActive method works, ditch the Chromium-specific #ifdef and go back to the original code.

WebKit/mac:

2009-06-03 David Hyatt <hyatt@apple.com>

Reviewed by Sam Weinig.

Reworking of focus and active state updating for WebHTMLViews so that it actually works instead of reporting wrong answers.

  • WebCoreSupport/WebFrameLoaderClient.mm: (WebFrameLoaderClient::transitionToCommittedForNewPage): WebHTMLViews become first responders before they are able to reach their Frames/Pages. This happens because the NSClipView becomes first responder when the WebHTMLView gets destroyed, and then we transfer the responder state back to the new WebHTMLView when it is first connected. Once we have transitioned to a new page and have the Frame/Page available, go ahead and explicitly focus the frame in WebCore and update our focused state. This change allows us to remove the updateFocusedActive timer and the code from viewDidMoveToWindow.


  • WebView/WebHTMLView.mm: (-[WebHTMLView close]): (-[WebHTMLView addWindowObservers]): (-[WebHTMLView viewWillMoveToWindow:]): (-[WebHTMLView viewDidMoveToWindow]): (-[WebHTMLView _removeWindowObservers]): Remove all the updateFocusedActiveState timer code, since it no longer exists.

(-[WebHTMLView windowDidBecomeKey:]):
(-[WebHTMLView windowDidResignKey:]):
Active state updating is no longer done by WebHTMLViews. It is handled by the WebView instead.


(-[WebHTMLView becomeFirstResponder]):
(-[WebHTMLView resignFirstResponder]):
Reworking of WebHTMLView's code for gaining/losing responder status. No longer update active state here, since the active state
can never change just because of responder changes. Make sure that the focused frame gets updated properly (and most importantly
actually cleared when a WebHTMLView resigns responder status).

  • WebView/WebHTMLViewInternal.h:
  • WebView/WebHTMLViewPrivate.h: State updating for focused and active states has been made independent.
  • WebView/WebView.mm: (-[WebView _updateActiveState]): WebView now handles updating of active state in _updateActiveState. It is now purely based off whether the window is key and nothing else.

(-[WebView addWindowObserversForWindow:]):
(-[WebView removeWindowObservers]):
Start listening for the window becoming/losing key even in the usesDocumentViews case.

(-[WebView _updateFocusedStateForFrame:]):
Focused state updating is split into its own method now and called when WebHTMLViews gain and lose responder status.

(-[WebView _windowDidBecomeKey:]):
(-[WebView _windowDidResignKey:]):
Make sure to call _updateActiveState as the window gains/loses key.

(-[WebView _windowWillOrderOnScreen:]):
Run this code now that WebHTMLView no longer does it.

  • WebView/WebViewInternal.h:
  • WebView/WebViewPrivate.h: Updating for the split of focused and active state updating into separate methods.

WebKitTools:

2009-06-03 David Hyatt <hyatt@apple.com>

Reviewed by Sam Weinig.

  • DumpRenderTree/mac/FrameLoadDelegate.mm: (-[FrameLoadDelegate webView:didCommitLoadForFrame:]):
  • DumpRenderTree/mac/LayoutTestControllerMac.mm: (LayoutTestController::setMainFrameIsFirstResponder): Remove explicit calls to _updateFocusedActiveState when the view is made first responder. Focused state already properly updates when views become first responders, so these calls are unnecessary.

(LayoutTestController::setWindowIsKey):
Explicitly call _updateActiveState when our window becomes key, since we're just faking it
and the notification won't be sent.

09:04 Changeset [44415] by eric.carlson@apple.com

2009-06-04 Pierre d'Herbemont <pdherbemont@apple.com>

Reviewed by Darin Adler.

Test: media/before-load-member-access.html


https://bugs.webkit.org/show_bug.cgi?id=26081

  • html/HTMLMediaElement.cpp: (WebCore::HTMLMediaElement::played): Ensure that if m_playedTimeRanges, is not initialized we return a valid range, and don't attempt to use it.

2009-06-04 Pierre d'Herbemont <pdherbemont@apple.com>

Reviewed by Darin Adler.

https://bugs.webkit.org/show_bug.cgi?id=26081

  • media/before-load-member-access-expected.txt: Added.
  • media/before-load-member-access.html: Test access for property accessed before loading occurs. That tests for 26081 crasher specifically.
08:11 Changeset [44414] by ariya@webkit.org

2009-06-04 Ariya Hidayat <ariya.hidayat@nokia.com>

Rubber-stamped by Tor Arne Vestbø.

Added Qt-specific expected result for Canvas pointInPath's test.

Qt's QPainterPath::contains(point) returns false if the point is
exactly along one of the edges (except the origin). Until we
find a workaround for this, compensate it in the expected result.

  • platform/qt/fast/canvas/pointInPath-expected.txt: Added.
05:44 Changeset [44413] by hausmann@webkit.org

2009-06-03 Kenneth Rohde Christiansen <kenneth.christiansen@openbossa.org>

Reviewed by Simon Hausmann.

Add API to make it possible to clear all memory caches used by
QtWebKit.

  • Api/qwebsettings.cpp: (QWebSettings::clearMemoryCaches):
  • Api/qwebsettings.h:

2009-06-03 Kenneth Rohde Christiansen <kenneth.christiansen@openbossa.org>

Reviewed by Simon Hausmann.

Implement a way to invalidate the FontCache used by the Qt port.

  • platform/graphics/qt/FontCacheQt.cpp: (WebCore::FontCache::invalidate):
04:00 Changeset [44412] by barraclough@apple.com

2009-06-04 Gavin Barraclough <barraclough@apple.com>

Reviewed by Oliver Hunt.

Wrap the code that plants pushes/pops planted by JIT in explanatorily named
methods; move property storage reallocation into a standard stub function.

~No performance impact (possible <1% progression on x86-64, likely just noise).

  • jit/JIT.cpp: (JSC::JIT::privateCompile): (JSC::JIT::privateCompileCTIMachineTrampolines):

Wrap calls to push/pop.

  • jit/JIT.h:

Declare the new wrapper methods.

  • jit/JITInlineMethods.h: (JSC::JIT::preverveReturnAddressAfterCall): (JSC::JIT::restoreReturnAddressBeforeReturn):

Define the new wrapper methods.

  • jit/JITOpcodes.cpp: (JSC::JIT::emit_op_end): (JSC::JIT::emit_op_ret):

Wrap calls to push/pop.

  • jit/JITPropertyAccess.cpp: (JSC::JIT::privateCompilePutByIdTransition):

Move property storage reallocation into a standard stub function.

  • jit/JITStubs.cpp: (JSC::JITStubs::DEFINE_STUB_FUNCTION):
  • jit/JITStubs.h: (JSC::JITStubs::):
01:58 Changeset [44411] by ariya@webkit.org

2009-06-04 Laszlo Gombos <laszlo.1.gombos@nokia.com>

Reviewed by Ariya Hidayat.

https://bugs.webkit.org/show_bug.cgi?id=26015

[Qt] Single-threaded QtWebKit configuration

Turn off Database, DOM storage, icon database and Web Workers support
when ENABLE_SINGLE_THREADED is turned on.

Set SQLITE_THREADSAFE to false to turn off SQLite mutexes
when ENABLE_SINGLE_THREADED is turned on.

  • WebCore.pro:
01:22 Changeset [44410] by mrowe@apple.com

Speculative Tiger build fix.

06/03/09:

20:32 Changeset [44409] by weinig@apple.com

WebKit:

2009-06-03 Sam Weinig <sam@webkit.org>

Reviewed by Mark Rowe.

Added WebDelegateImplementationCaching.h/mm

  • WebKit.xcodeproj/project.pbxproj:

WebKit/mac:

2009-06-03 Sam Weinig <sam@webkit.org>

Reviewed by Mark Rowe.

Move delegate implementation caching logic into its own files.

  • WebView/WebDelegateImplementationCaching.h: Copied from mac/WebView/WebViewInternal.h.
  • WebView/WebDelegateImplementationCaching.mm: Copied from mac/WebView/WebView.mm. (WebViewGetResourceLoadDelegateImplementations): (WebViewGetFrameLoadDelegateImplementations): (WebViewGetScriptDebugDelegateImplementations):
  • WebView/WebView.mm:
  • WebView/WebViewInternal.h:
20:30 Changeset [44408] by mitz@apple.com

Reviewed by Sam Weinig.

  • add some assertions that Font methods are used on the main thread
  • platform/graphics/Font.cpp: (WebCore::Font::setShouldUseSmoothing):
  • platform/graphics/Font.h: (WebCore::Font::primaryFont):
  • platform/graphics/FontFastPath.cpp: (WebCore::Font::glyphDataForCharacter):
19:39 Changeset [44407] by weinig@apple.com

WebKit:

2009-06-03 Sam Weinig <sam@webkit.org>

Reviewed by Mark Rowe.

Add WebViewInternal.mm.

  • WebKit.xcodeproj/project.pbxproj:

WebKit/mac:

2009-06-03 Sam Weinig <sam@webkit.org>

Reviewed by Mark Rowe.

Move WebViewPrivate structure to WebViewInternal.h/mm.

  • WebView/WebView.mm:
  • WebView/WebViewInternal.h:
  • WebView/WebViewInternal.mm: Added. (+[WebViewPrivate initialize]): (-[WebViewPrivate init]): (-[WebViewPrivate dealloc]): (-[WebViewPrivate finalize]):
18:13 Changeset [44406] by weinig@apple.com

2009-06-03 Sam Weinig <sam@webkit.org>

Reviewed by Darin Adler.

Small cleanup in WebView.

  • Move global data initialization to WebView's initialization method.
  • Move _clearViewUpdateRunLoopObserver from WebViewPrivate to WebView (FileInternal)
  • WebView/WebView.mm: (-[WebViewPrivate init]): Remove global initializers. (-[WebViewPrivate dealloc]): Cleanup whitespace. (-[WebViewPrivate finalize]): Ditto. (-[WebView _close]): Call [self _clearViewUpdateRunLoopObserver] instead of [_private _clearViewUpdateRunLoopObserver] (+[WebView initialize]): Move global initializers here. (-[WebView _clearViewUpdateRunLoopObserver]): Moved from WebViewPrivate. (viewUpdateRunLoopObserverCallBack): Call [self _clearViewUpdateRunLoopObserver] instead of [_private _clearViewUpdateRunLoopObserver]
17:11 Changeset [44405] by mitz@apple.com
  • Windows build fix

Rolled out apparently-accidental changes to config.h from r44398. These
were not part of the patch as reviewed.

  • config.h:
17:07 Changeset [44404] by dimich@chromium.org

2009-06-03 Dmitry Titov <dimich@chromium.org>

Not reviewed, Chromium build fix.

https://bugs.webkit.org/show_bug.cgi?id=26177
Reverting 'private' to 'protected' on 2 classes.
Chromium glue layer (not yet in Webkit tree) relies on ability to derive
Chromium-specific platform classes and access the data members.
See bug for more details.

  • platform/PlatformMouseEvent.h:
  • platform/PlatformWheelEvent.h:
16:37 Changeset [44403] by cmarrin@apple.com

Reviewed by Simon Fraser <simonfr@apple.com>.

Fixed https://bugs.webkit.org/show_bug.cgi?id=26162

This corrects an error when destroying an animation
or transition where endAnimation was never getting
called and therefore the hardware animation was never
getting removed.

This includes a manual-test since it's really impossible
to make a meaningful automatic test for an animation
bug like this.

This has no effect unless accelerated compositing is
enabled.

16:12 Changeset [44402] by kevino@webkit.org

wx build fix, remove no longer needed function call.

16:08 Changeset [44401] by agl@chromium.org

2009-06-03 Adam Langley <agl@google.com>

Reviewed by Eric Seidel.

Make the scrollbar thumb size twice the width for Chromium Linux. This
matches Firefox on Linux.

This will need layout test pixel results to be rebaselined in the
Chromium tree.

http://code.google.com/p/chromium/issues/detail?id=12602
https://bugs.webkit.org/show_bug.cgi?id=26176

  • platform/chromium/ScrollbarThemeChromiumLinux.cpp: (WebCore::ScrollbarThemeChromiumLinux::minimumThumbLength):
  • platform/chromium/ScrollbarThemeChromiumLinux.h:
15:46 Changeset [44400] by agl@chromium.org

2009-06-03 Adam Langley <agl@google.com>

Reviewed by Eric Seidel.

Change Chromium scrollbar theme code to use different classes on
Windows and Linux rather than suppling symbols. The ScrollbarTheme
class is already using virtual dispatch, so there's no reason not to.

This should not affect any layout tests.

https://bugs.webkit.org/show_bug.cgi?id=26174

  • platform/chromium/ScrollbarThemeChromium.cpp:
  • platform/chromium/ScrollbarThemeChromium.h:
  • platform/chromium/ScrollbarThemeChromiumLinux.cpp: (WebCore::ScrollbarTheme::nativeTheme): (WebCore::ScrollbarThemeChromiumLinux::scrollbarThickness): (WebCore::ScrollbarThemeChromiumLinux::paintTrackPiece): (WebCore::ScrollbarThemeChromiumLinux::paintButton): (WebCore::ScrollbarThemeChromiumLinux::paintThumb): (WebCore::ScrollbarThemeChromiumLinux::buttonSize):
  • platform/chromium/ScrollbarThemeChromiumLinux.h: Added.
  • platform/chromium/ScrollbarThemeChromiumWin.cpp: (WebCore::ScrollbarTheme::nativeTheme): (WebCore::ScrollbarThemeChromiumWin::scrollbarThickness): (WebCore::ScrollbarThemeChromiumWin::invalidateOnMouseEnterExit): (WebCore::ScrollbarThemeChromiumWin::shouldSnapBackToDragOrigin): (WebCore::ScrollbarThemeChromiumWin::paintTrackPiece): (WebCore::ScrollbarThemeChromiumWin::paintButton): (WebCore::ScrollbarThemeChromiumWin::paintThumb): (WebCore::ScrollbarThemeChromiumWin::getThemeState): (WebCore::ScrollbarThemeChromiumWin::getThemeArrowState): (WebCore::ScrollbarThemeChromiumWin::getClassicThemeState): (WebCore::ScrollbarThemeChromiumWin::buttonSize):
  • platform/chromium/ScrollbarThemeChromiumWin.h: Added.
15:45 Changeset [44399] by bfulgham@webkit.org

2009-06-03 Brent Fulgham <bfulgham@webkit.org>

Build fix for Windows Cairo.

  • WebKit.vcproj/WebKit.vcproj: Remove unneeded giflib.lib link command.
15:32 Changeset [44398] by bfulgham@webkit.org

2009-06-03 Pavel Feldman <pfeldman@chromium.org>

Reviewed by Timothy Hatcher.

Enabling debugger requires that Scripts panel is already attached to the
render tree. The reason is that recompile events result in script sources
being added into the source frames. Prior to the global options introduced,
debugger was enabled from the Scripts panel, so that it was guaranteed to
exist. The InspectorController::enableDebugger API calls with no inspector
frontend showing were failing though.

https://bugs.webkit.org/show_bug.cgi?id=26145

  • WebCore.base.exp:
  • inspector/InspectorController.cpp: (WebCore::InspectorController::setWindowVisible): (WebCore::InspectorController::scriptObjectReady): (WebCore::InspectorController::enableDebuggerFromFrontend): (WebCore::InspectorController::enableDebugger):
  • inspector/InspectorController.h:
  • inspector/InspectorController.idl:
  • inspector/InspectorFrontend.cpp: (WebCore::InspectorFrontend::attachDebuggerWhenShown):
  • inspector/InspectorFrontend.h:
  • inspector/front-end/ScriptsPanel.js: (WebInspector.ScriptsPanel.prototype.show): (WebInspector.ScriptsPanel.prototype.attachDebuggerWhenShown):
  • inspector/front-end/inspector.js: (WebInspector.attachDebuggerWhenShown):
15:31 Changeset [44397] by weinig@apple.com

2009-06-03 Sam Weinig <sam@webkit.org>

Reviewed by Geoffrey Garen.

Add back known this value optimization by abstracting
slow case if not JSCell jumps.

  • jit/JIT.h:
  • jit/JITCall.cpp: (JSC::JIT::compileOpCallVarargs): (JSC::JIT::compileOpCallVarargsSlowCase): (JSC::JIT::compileOpCall): (JSC::JIT::compileOpCallSlowCase):
  • jit/JITInlineMethods.h: (JSC::JIT::emitJumpSlowCaseIfNotJSCell): (JSC::JIT::linkSlowCaseIfNotJSCell):
  • jit/JITOpcodes.cpp: (JSC::JIT::emit_op_instanceof): (JSC::JIT::emitSlow_op_instanceof):
  • jit/JITPropertyAccess.cpp: (JSC::JIT::emit_op_get_by_val): (JSC::JIT::emitSlow_op_get_by_val): (JSC::JIT::emit_op_put_by_val): (JSC::JIT::emitSlow_op_put_by_val): (JSC::JIT::emit_op_get_by_id): (JSC::JIT::emitSlow_op_get_by_id): (JSC::JIT::emit_op_put_by_id): (JSC::JIT::emitSlow_op_put_by_id):
15:10 Changeset [44396] by mitz@apple.com

Reviewed by Anders Carlsson.

  • fix a regression from the previous patch
  • platform/graphics/Font.cpp: Initialize shouldUseFontSmoothing to true.
15:03 Changeset [44395] by mitz@apple.com

WebCore:

Reviewed by Anders Carlsson.

  • eliminate WebCoreTextRenderer
  • WebCore.base.exp: Updated.
  • WebCore.xcodeproj/project.pbxproj: Removed WebCoreTextRenderer.{h,mm}

and promoted WebFontCache.h to private.

  • platform/graphics/Font.cpp: (WebCore::Font::setShouldUseSmoothing): Added this static setter for

a new file-static boolean.

(WebCore::Font::shouldUseSmoothing): Added this static getter.

  • platform/graphics/Font.h: Decleared setShouldUseSmoothing() and

shouldUseSmoothing().

  • platform/graphics/mac/FontMac.mm: (WebCore::Font::drawGlyphs): Use Font::shouldUseSmoothing() instead of WebCoreShouldUseFontSmoothing().
  • platform/graphics/mac/WebLayer.mm: Removed unneeded #import.
  • platform/mac/WebCoreTextRenderer.h: Removed.
  • platform/mac/WebCoreTextRenderer.mm: Removed.

WebKit:

Reviewed by Anders Carlsson.

  • WebKit project part of eliminating WebCoreTextRenderer
  • WebKit.xcodeproj/project.pbxproj: Changed WebKitNSStringExtras.m to

compile as Objective-C++.

WebKit/mac:

Reviewed by Anders Carlsson.

  • WebKit part of eliminating WebCoreTextRenderer
  • Misc/WebKitNSStringExtras.m: (webkit_CGCeiling): Added. (-[NSString _web_drawAtPoint:font:textColor:]): Replaced the malloc()ed

string buffer with a Vector. Moved code from
WebCoreDrawTextAtPoint() to here.

(-[NSString _web_drawDoubledAtPoint:withTopColor:bottomColor:font:]):
(-[NSString _web_widthWithFont:]): Replaced the malloc()ed string buffer

with a Vector. Moved code from WebCoreTextFloatWidth() to here.

  • WebView/WebHTMLView.mm: (-[WebHTMLView _addToStyle:fontA:fontB:]): Replaced call to

WebCoreFindFont() with use of
+[WebFontCache fontWithFamily:traits:weight:size:].

  • WebView/WebView.mm: (+[WebView _setAlwaysUsesComplexTextCodePath:]): Changed to call

Font::setCodePath() directly.

(+[WebView _setShouldUseFontSmoothing:]): Changed to call

Font::setShouldUseSmoothing().

(+[WebView _shouldUseFontSmoothing]): Changed to call

Font::shouldUseSmoothing().

14:11 Changeset [44394] by mrowe@apple.com

Versioning.

14:09 Changeset [44393] by mrowe@apple.com

New tag.

14:05 Changeset [44392] by mrowe@apple.com

Merge r44356.

13:32 Changeset [44391] by levin@chromium.org

2009-06-03 David Levin <levin@chromium.org>

Reviewed by Dimitri Glazkov.

v8's ScriptController::evaluate should protect the Frame like the jsc version.
https://bugs.webkit.org/show_bug.cgi?id=26172

This change is simply copying protections done for Frame in the method
WebCore::ScriptController::evaluate in the file js/ScriptController.cpp.

  • bindings/v8/ScriptController.cpp: (WebCore::ScriptController::evaluate):
13:17 Changeset [44390] by mitz@apple.com

Rubber-stamped by Mark Rowe.

  • remove a private method that was used only by Safari 3.0.
  • WebView/WebView.mm: Removed +_minimumRequiredSafariBuildNumber.
  • WebView/WebViewPrivate.h: Ditto.
11:24 Changeset [44389] by mitz@apple.com

Reviewed by John Sullivan.

  • fix <rdar://problem/6841120> Use CTFontManager notifications instead of ATS notifications
  • platform/graphics/FontCache.h: Made it an error to destroy a FontCache instance.
  • platform/graphics/mac/FontCacheMac.mm: (WebCore::fontCacheRegisteredFontsChangedNotificationCallback): Added this notification callback for the kCTFontManagerRegisteredFontsChangedNotification, which calls invalidate(). (WebCore::FontCache::platformInit): Register for kCTFontManagerRegisteredFontsChangedNotification.
11:10 Changeset [44388] by mitz@apple.com
  • build fix
  • WebView/WebHTMLView.mm:
10:07 Changeset [44387] by kevino@webkit.org

Reviewed by Kevin Ollivier.

Use CGContextShowGlyphsWithAdvances to get more accurate text rendering on Mac.

https://bugs.webkit.org/show_bug.cgi?id=26161

09:43 Changeset [44386] by dglazkov@chromium.org

2009-06-03 Pavel Feldman <pfeldman@chromium.org>

Reviewed by Timothy Hatcher.

Reorder ResourcesPanel components initialization to unfreeze resource list scroller.

https://bugs.webkit.org/show_bug.cgi?id=26159

  • inspector/front-end/ResourcesPanel.js: (WebInspector.ResourcesPanel):
09:41 Changeset [44385] by dglazkov@chromium.org

2009-06-03 Pavel Feldman <pfeldman@chromium.org>

Reviewed by Timothy Hatcher.

  • Fix for crash (preceded by assertion) in InspectorController::didCommitLoad when reloading or navigating with the Inspector open.
  • Fix for Inspector's Elements panel being empty when Inspector first appears.

https://bugs.webkit.org/show_bug.cgi?id=26134
https://bugs.webkit.org/show_bug.cgi?id=26135

  • inspector/InspectorController.cpp: (WebCore::InspectorController::scriptObjectReady): (WebCore::InspectorController::didLoadResourceFromMemoryCache): (WebCore::InspectorController::identifierForInitialRequest): (WebCore::InspectorController::ensureResourceTrackingSettingsLoaded):
  • inspector/InspectorController.h:
09:35 Changeset [44384] by mitz@apple.com

Reviewed by Anders Carlsson.

  • build fix
  • WebView/WebHTMLView.mm:
  • WebView/WebView.mm:
08:29 Changeset [44383] by aroben@apple.com

Windows build fix after r44379

  • svg/graphics/SVGImage.cpp: Move EmptyClients.h back down below the other #includes to fix a compiler warning on Windows.
07:22 Changeset [44382] by hausmann@webkit.org

2009-06-02 Kenneth Rohde Christiansen <kenneth.christiansen@openbossa.org>

Reviewed by Simon Hausmann.

Add workaround for crash in Linux Flash Player when hosted by
another toolkit than GTK+. Bug fixed at the Flash Player bugzilla,
issue (FP-2140).

07:21 Changeset [44381] by hausmann@webkit.org

2009-06-01 Kenneth Rohde Christiansen <kenneth.christiansen@openbossa.org>

Reviewed by Simon Hausmann.

Refactor the Qt plugin code to use NPP_SetWindow correctly,
to make resizing of plugins work.

Attention was paid to make sure that the windowed plugins scroll
synchronized with the page view. A manual test has been added.

00:33 Changeset [44380] by ap@webkit.org

Rubber-stamped by Mark Rowe.

Fix appcache/fallback.html test broken by network simulator changes made for a preloading test.

  • http/tests/resources/network-simulator.php: Tell PHP that a variable is global.

06/02/09:

22:37 Changeset [44379] by darin@apple.com

WebCore:

2009-06-02 Darin Adler <darin@apple.com>

Reviewed by David Hyatt.

Bug 26112: viewless WebKit -- make events work
https://bugs.webkit.org/show_bug.cgi?id=26112

The main fix here is to make mouse and wheel event coordinates in the coordinate
system of the top level NSView rather than the NSWindow when in the viewless mode.
This is the design Hyatt chose, but the event part of it wasn't done yet.

Also fix FrameView to do normal reference counting instead of a strange model with
an explicit deref near creation time.

  • WebCore.base.exp: Updated.
  • page/EventHandler.cpp: (WebCore::EventHandler::eventLoopHandleMouseUp): Moved this function into the file to reduce conditionals in the header. (WebCore::EventHandler::eventLoopHandleMouseDragged): Ditto.
  • page/EventHandler.h: Reduced includes. Fixed formatting of Objective-C types. Made currentNSEvent a static member function. Added sendContextMenuEvent and eventMayStartDrag functions that takes NSEvent * so the conversion to PlatformMouseEvent can be done here rather than in WebKit. Reduced #if by making eventLoopHandleMouseUp and eventLoopHandleMouseDragged unconditional.
  • page/Frame.cpp: (WebCore::Frame::setView): Made this take a PassRefPtr since it takes ownership. (WebCore::Frame::createView): Changed to use RefPtr and FrameView::create and remove the explicit deref.
  • page/Frame.h: Changed setView to take a PassRefPtr.
  • page/FrameTree.cpp: Added newly-needed include.
  • page/FrameView.cpp: (WebCore::FrameView::FrameView): Got rid of one of the two constructors, and removed the initialization of m_refCount and call to show from the reamining one. (WebCore::FrameView::create): Added two create functions that do what the two constructors did before, except that they return a PassRefPtr to make sure the reference counting is handled correctly. (WebCore::FrameView::~FrameView): Removed assertion from when FrameView implemented its own reference counting.
  • page/FrameView.h: Inherit from RefCounted for reference counting. Made the constructor private and added create functions. Got rid of the hand-implemented reference counting in this class.
  • page/mac/EventHandlerMac.mm: (WebCore::currentNSEventSlot): Renamed currentEvent to currentNSEventSlot to make it more clear how it relates to currentNSEvent. (WebCore::EventHandler::currentNSEvent): Updated. (WebCore::CurrentEventScope::CurrentEventScope): Added. Use to set/reset the current event in a foolproof way. (WebCore::CurrentEventScope::~CurrentEventScope): Ditto. (WebCore::EventHandler::wheelEvent): Use CurrentEventScope. Pass the platform window in when constructing the PlatformWheelEvent. (WebCore::EventHandler::keyEvent): Use CurrentEventScope. (WebCore::lastEventIsMouseUp): Use currentNSEvent. (WebCore::EventHandler::passMouseDownEventToWidget): Ditto. (WebCore::EventHandler::eventLoopHandleMouseDragged): Ditto. (WebCore::EventHandler::eventLoopHandleMouseUp): Ditto. (WebCore::EventHandler::passSubframeEventToSubframe): Use currentPlatformMouseEvent to get a mouse event that has the appropriate platform window passed to create it. (WebCore::EventHandler::passWheelEventToWidget): Ditto. (WebCore::EventHandler::mouseDown): Ditto. (WebCore::EventHandler::mouseDragged): Ditto. (WebCore::EventHandler::mouseUp): Ditto. (WebCore::EventHandler::mouseMoved): Ditto. (WebCore::EventHandler::currentPlatformMouseEvent): Added. Passes the platform window that's now needed to create a PlatformMouseEvent. (WebCore::EventHandler::sendContextMenuEvent): Added. (WebCore::EventHandler::eventMayStartDrag): Added.
  • platform/HostWindow.h: Removed unneeded includes and constructor definition.
  • platform/PlatformMouseEvent.h: Sorted things in alphabetical order. Changed Mac constructor to take a windowView as well as the event. This is needed in viewless mode, since the "window" is actually an NSView, so the event has to know which view to compute the coordinates with. Made the same change to pointForEvent.
  • platform/PlatformWheelEvent.h: Ditto.
  • platform/mac/PlatformMouseEventMac.mm: (WebCore::pointForEvent): Convert point from window coordinates to view coordinates if a windowView is passed in. This is used in viewless mode. (WebCore::PlatformMouseEvent::PlatformMouseEvent): Ditto.
  • platform/mac/WheelEventMac.mm: (WebCore::PlatformWheelEvent::PlatformWheelEvent): Ditto.
  • platform/mac/WidgetMac.mm: (WebCore::Widget::convertFromContainingWindow): Fixed case where there is no platform widget and no parent. Before it would yield unpredictable results because of dispatching to a nil object and returning a structure. Now it returns the point without changing coordinates at all, which is what we need for this case in viewless mode.
  • rendering/RenderApplet.cpp: Removed unneeded includes. (WebCore::RenderApplet::intrinsicSize): Use widget function. (WebCore::RenderApplet::createWidgetIfNecessary): Ditto.
  • rendering/RenderApplet.h: Make more things private. Get rid of unneeded explicit destructor.
  • rendering/RenderFrame.cpp: Removed unneeded includes. (WebCore::RenderFrame::edgeInfo): Updated to use node function so header doesn't have to define element function. (WebCore::RenderFrame::viewCleared): Ditto. Also changed to use widget function.
  • rendering/RenderFrame.h: Removed unneeded includes. Made some things private. Got rid of element function.
  • rendering/RenderPart.cpp: Removed unneeded includes. (WebCore::RenderPart::~RenderPart): Changed to use clearWidget function. (WebCore::RenderPart::setWidget): Changed to use widget function. (WebCore::RenderPart::deleteWidget): Changed to use passed-in widget. This is now only called by the clearWidget function.
  • rendering/RenderPart.h: Removed unneeded forward declarations. Made more functions private. Updated deleteWidget to take widget argument.
  • rendering/RenderPartObject.cpp: (WebCore::RenderPartObject::~RenderPartObject): use frameView function instead of getting at m_view directly. (WebCore::RenderPartObject::updateWidget): Ditto. (WebCore::RenderPartObject::layout): Ditto. Same for widget and m_widget. (WebCore::RenderPartObject::viewCleared): Ditto.
  • rendering/RenderPartObject.h: Made some functions private.
  • rendering/RenderWidget.cpp: (WebCore::RenderWidget::RenderWidget): Initialize m_frameView with construction syntax instead of assignment. (WebCore::RenderWidget::destroy): Updated for name change of m_view to m_frameView. (WebCore::RenderWidget::~RenderWidget): Use clearWidget to delete the widget. (WebCore::RenderWidget::setWidget): Use clearWidget. (WebCore::RenderWidget::paint): Updated for name change of m_view to m_frameView. (WebCore::RenderWidget::deleteWidget): Changed to use passed-in widget.
  • rendering/RenderWidget.h: Made many functions protected, others private and made all data members private.
  • svg/animation/SMILTime.h: Removed unhelpful max and min functions. These just do what std::max and std::min will already do automatically for this type, so they are not helpful.
  • svg/graphics/SVGImage.cpp: (WebCore::SVGImage::SVGImage): Removed unneeded initial values for types that initialize to zero without anything explicit. (WebCore::SVGImage::~SVGImage): Update since m_frame and m_frameView are no longer stored. (WebCore::SVGImage::setContainerSize): Ditto. (WebCore::SVGImage::usesContainerSize): Ditto. (WebCore::SVGImage::size): Ditto. (WebCore::SVGImage::hasRelativeWidth): Ditto. (WebCore::SVGImage::hasRelativeHeight): Ditto. (WebCore::SVGImage::draw): Ditto. (WebCore::SVGImage::nativeImageForCurrentFrame): Ditto. (WebCore::SVGImage::dataChanged): Ditto.
  • svg/graphics/SVGImage.h: Removed unneeded includes and unneeded data members m_document, m_frame, m_frameView, and m_minSize.

WebKit/mac:

2009-06-02 Darin Adler <darin@apple.com>

Reviewed by David Hyatt.

Bug 26112: viewless WebKit -- make events work
https://bugs.webkit.org/show_bug.cgi?id=26112

One change here is to make the -[WebFrame frameView] function assert if ever called
in viewless mode, and fix many callers that would trip that assertion.

A major change is to put some methods and data in WebView that are currently in
WebHTMLView, used only in viewless mode. A next step will be to use the WebView
methods whenever possible, even when not in the viewless mode.

Also fix FrameView to do normal reference counting instead of a strange model with
an explicit deref near creation time.

  • WebCoreSupport/WebChromeClient.mm: (WebChromeClient::scrollRectIntoView): Add code to start in the appropriate place when dealing with viewless mode. This gets triggered when visiting the Google home page. (WebChromeClient::setToolTip): Changed to call WebView instead of WebHTMLView. (WebChromeClient::print): Changed to use a new selector that doesn't require a WebFrameView if present. Also don't even try to use the old selector in viewless mode.
  • WebCoreSupport/WebEditorClient.mm: (WebEditorClient::respondToChangedSelection): Changed to call WebView instead of WebHTMLView.
  • WebCoreSupport/WebFrameLoaderClient.mm: (WebFrameLoaderClient::transitionToCommittedForNewPage): Added more code paths for viewless mode to skip things we can't do in that mode, with appropriate FIXME. Use Frame::create and RefPtr and eliminate the strange reference counting of FrameView.
  • WebView/WebDataSource.mm: (-[WebDataSource _receivedData:]): Added a _usesDocumentViews guard around code that's specific to document views.
  • WebView/WebFrame.mm: (-[WebFrame _updateBackgroundAndUpdatesWhileOffscreen]): Added a _usesDocumentViews guard around code that's specific to document views. (-[WebFrame _hasSelection]): Added an implementation for viewless mode. (-[WebFrame _clearSelection]): Assert we're not in viewless mode; it doesn't really make sense to clear the selection in only one frame in viewless mode. Later we can decide what to do. (-[WebFrame _dragSourceMovedTo:]): Assert we're not in viewless mode; the dragging code for viewless mode shouldn't have to involve the WebFrame object at all. (-[WebFrame _dragSourceEndedAt:operation:]): Ditto. (-[WebFrame frameView]): Assert we're not in viewless mode. This assertion fires often, but it's a great pointer to code that needs to be changed.
  • WebView/WebHTMLView.mm: (-[WebHTMLView hitTest:]): Tweaked a comment. (-[WebHTMLView _updateMouseoverWithEvent:]): Fixed a bug where the fake event for moving the mouse out of the old view ended up overwriting the event for moving the mouse within the new view. (-[WebHTMLView mouseDown:]): Got rid of explicit conversion of event to PlatformMouseEvent in call to sendContextMenuEvent; that's no longer possible without passing another argument, and it's now handled in EventHandler.
  • WebView/WebTextCompletionController.h: Copied from WebKit/mac/WebView/WebHTMLView.mm. Removed everything except for the WebTextCompletionController class.
  • WebView/WebTextCompletionController.mm: Copied from WebKit/mac/WebView/WebHTMLView.mm. Ditto.
  • WebView/WebUIDelegatePrivate.h: Added webView:printFrame: method.
  • WebView/WebView.mm: Moved includes and fields in from WebHTMLView. (-[WebView _usesDocumentViews]): Updated for name change from useDocumentViews to usesDocumentViews. (-[WebView _commonInitializationWithFrameName:groupName:usesDocumentViews:]): Ditto. (-[WebView drawSingleRect:]): Ditto. (-[WebView isFlipped]): Ditto. (-[WebView setFrameSize:]): Ditto. Also changed to use _mainCoreFrame method. (-[WebView _viewWillDrawInternal]): Ditto. (-[WebView viewWillDraw]): Ditto. (-[WebView drawRect:]): Ditto. (-[WebView _close]): Added code to nil out the lastMouseoverView global. (-[WebView _dashboardRegions]): Use _mainCoreFrame. (-[WebView setProhibitsMainFrameScrolling:]): Ditto. (-[WebView _setInViewSourceMode:]): Ditto. (-[WebView _inViewSourceMode]): Ditto. (-[WebView _attachScriptDebuggerToAllFrames]): Ditto. (-[WebView _detachScriptDebuggerFromAllFrames]): Ditto. (-[WebView textIteratorForRect:]): Ditto. (-[WebView _executeCoreCommandByName:value:]): Ditto. (-[WebView addWindowObserversForWindow:]): Ditto. (-[WebView removeWindowObservers]): Ditto. (-[WebView _updateFocusedAndActiveState]): Ditto. (-[WebView _updateFocusedAndActiveStateForFrame:]): Turned into a class method. Added code to handle the viewless case without calling frameView. (-[WebView _windowDidBecomeKey:]): Updated for name change from useDocumentViews to usesDocumentViews. (-[WebView _windowDidResignKey:]): Ditto. (-[WebView _windowWillOrderOnScreen:]): Ditto. (-[WebView mainFrame]): Tweaked. (-[WebView selectedFrame]): Added a conditional to avoid trying to get at the frame view in viewless case. (-[WebView _setZoomMultiplier:isTextOnly:]): Use _mainCoreFrame. (-[WebView setCustomTextEncodingName:]): Ditto. (-[WebView windowScriptObject]): Ditto. (-[WebView setHostWindow:]): Ditto. Also put some code that's needed only for document views inside _private->usesDocumentViews. (-[WebView _hitTest:dragTypes:]): Tweaked. (-[WebView acceptsFirstResponder]): Added case for viewless mode along with a FIXME, since it's not complete. (-[WebView becomeFirstResponder]): Ditto. (-[WebView _webcore_effectiveFirstResponder]): Put the body of this inside a usesDocumentView check, because we don't need the first responder forwarding in viewless mode. (-[WebView setNextKeyView:]): Ditto. (-[WebView mouseDown:]): Added. Copied from WebHTMLView. FIXME in here suggests that we make WebHTMLView share this one soon, which I think is practical. (-[WebView mouseUp:]): Ditto. (-[WebView setHoverFeedbackSuspended:]): Added a code path for viewless mode. (-[WebView shouldClose]): Use _mainCoreFrame. (-[WebView aeDescByEvaluatingJavaScriptFromString:]): Ditto. (-[WebView setEditable:]): Ditto. (-[WebView _frameViewAtWindowPoint:]): Added an assertion that we're not in viewless mode since this method makes no sense in that mode. (-[WebView _setMouseDownEvent:]): Added. Copied from WebHTMLView. I plan to eliminate the one in WebHTMLView soon. (-[WebView _cancelUpdateMouseoverTimer]): Ditto. (-[WebView _stopAutoscrollTimer]): Ditto. (+[WebView _updateMouseoverWithEvent:]): Ditto. (-[WebView _updateMouseoverWithFakeEvent]): Ditto. (-[WebView _setToolTip:]): Added. Calls through to the WebHTMLView version. (-[WebView _selectionChanged]): Ditto. (-[WebView _mainCoreFrame]): Added. (-[WebView _needsOneShotDrawingSynchronization]): Moved into the WebInternal category. (-[WebView _setNeedsOneShotDrawingSynchronization:]): Ditto. (-[WebView _startedAcceleratedCompositingForFrame:]): Ditto. (-[WebView _stoppedAcceleratedCompositingForFrame:]): Ditto. (viewUpdateRunLoopObserverCallBack): Ditto. (-[WebView _scheduleViewUpdate]): Ditto.
  • WebView/WebViewInternal.h: Made most of the file not compile at all when included from non-C++ source files, elminating some excess declarations and typedefs. Moved more methods into the real internal category. Added new methods _setMouseDownEvent, _cancelUpdateMouseoverTimer, _stopAutoscrollTimer, _updateMouseoverWithFakeEvent, _selectionChanged, and _setToolTip:.

WebKit/wx:

2009-06-02 Darin Adler <darin@apple.com>

Reviewed by David Hyatt.

  • WebKitSupport/FrameLoaderClientWx.cpp: (WebCore::FrameLoaderClientWx::transitionToCommittedForNewPage): Use FrameView::create and RefPtr instead of the old "crazy" reference counting.
20:15 Changeset [44378] by agl@chromium.org

2009-06-02 Adam Langley <agl@google.com>

Reviewed by Eric Seidel.

The previous code was assuming that we'll be painting buttons on the scrollbar
which isn't true on Linux. To reproduce, resize a page with scrollbars until
they are less than two widths high.

This will need pixel test baselines to be updated in the Chromium tree.

  • platform/chromium/ScrollbarThemeChromium.cpp: move this function...
  • platform/chromium/ScrollbarThemeChromiumWin.cpp: (WebCore::ScrollbarThemeChromium::trackRect): ... to here
  • platform/chromium/ScrollbarThemeChromiumLinux.cpp: (WebCore::ScrollbarThemeChromium::trackRect): add an alternative which doesn't remove the track when the scrollbar is less than two widths high.
19:57 Changeset [44377] by jberlin@apple.com

2009-06-02 Jessie Berlin <jberlin@apple.com>

Reviewed by Mark Rowe.


Add "replace" behavior so that it shows the addition part of the
replacement of the file first, and then shows the changes.


Future: Add in showing the removal part of the replacement.

  • Scripts/svn-create-patch:
19:01 Changeset [44376] by mrowe@apple.com

Remove workaround that was added to address <rdar://problem/5488678> as it no longer affects our Tiger builds.

Reviewed by Anders Carlsson.

  • Configurations/Base.xcconfig:
17:19 Changeset [44375] by eric@webkit.org

2009-06-02 Eric Seidel <eric@webkit.org>

Reviewed by Maciej Stachowiak.

Insert*List on an stand-alone image in a content editable region ASSERTS
ASSERTION FAILED: isStartOfParagraph(startOfParagraphToMove)
https://bugs.webkit.org/show_bug.cgi?id=19066

Attempt to fix this by noticing that we inserted the list inside
the selection which includes the image, and re-adjust the selection
to not include the list before trying to move the image into
the list item.

Test: editing/execCommand/list-wrapping-image-crash.html

  • editing/InsertListCommand.cpp: (WebCore::InsertListCommand::doApply):
16:09 Changeset [44374] by eric@webkit.org

2009-06-02 Eric Seidel <eric@webkit.org>

Reviewed by Darin Adler.

Rename PositionIterator members in hope of futher clarity
https://bugs.webkit.org/show_bug.cgi?id=24854

Rename m_parent to m_anchorNode (since although it's always the parent
of the previous m_child member, it is not always the parent of the effective position)
Rename m_child to m_nodeAfterPositionInAnchor to make clear that it's the node
directly following the position. This member is often NULL, but is always
a child of m_parent, now m_anchorNode if set.
Rename m_offset to m_offsetInAnchor (since it's interpreted relative to m_anchorNode)

  • dom/PositionIterator.cpp: (WebCore::PositionIterator::operator Position): (WebCore::PositionIterator::increment): (WebCore::PositionIterator::decrement): (WebCore::PositionIterator::atStart): (WebCore::PositionIterator::atEnd): (WebCore::PositionIterator::atStartOfNode): (WebCore::PositionIterator::atEndOfNode): (WebCore::PositionIterator::isCandidate):
  • dom/PositionIterator.h: (WebCore::PositionIterator::PositionIterator): (WebCore::PositionIterator::node): (WebCore::PositionIterator::offsetInLeafNode):
15:08 Changeset [44373] by jianli@chromium.org

Fix the incorrect bug link in WebCore/ChangeLog for r44366.

14:46 Changeset [44372] by ap@webkit.org

Land correct results for a test I just added (forgot to regenerate them after adding new
test cases).

  • fast/xpath/py-dom-xpath/expressions-expected.txt:
14:43 Changeset [44371] by pkasting@chromium.org

Revert r44276.

14:42 Changeset [44370] by cmarrin@apple.com

Moved CSSMatrix tests into LayoutTests/transforms so both 2D and 3D versions get tested in all cases, as per discussion with Maciej

14:33 Changeset [44369] by eric@webkit.org

2009-06-02 Julien Chaffraix <jchaffraix@webkit.org>

Reviewed by Eric Seidel.

Bug 17167: Failures in fast/dom/Node/initial-values.html

This partly solve a compatibility issue with other browsers. It will also
make us more consistent while handling XHTML element.

The issue is that when we create an XHTML element inside an HTML document
(as it is the case when calling createElementNS), we default to the HTML
behaviour in nodeName. As we cannot test if an HTMLElement is an XHTML
element, our fix is to check whether it has a prefix and then default
to XML behaviour for nodeName.

  • html/HTMLElement.cpp: (WebCore::HTMLElement::nodeName): Add a prefix check before returning the uppercase tagName (HTML behaviour).
14:31 Changeset [44368] by eric@webkit.org

2009-06-02 Eric Seidel <eric@webkit.org>

Reviewed by Maciej Stachowiak.

Add a compareBoundaryPoints which takes RangeBoundaryPoints
https://bugs.webkit.org/show_bug.cgi?id=25500

I noticed the need for this function when removing compareBoundaryPoints(Position, Position)
This patch is almost entirely minus lines.

No functional changes, thus no tests.

  • dom/Range.cpp: (WebCore::Range::setStart): (WebCore::Range::setEnd): (WebCore::Range::compareBoundaryPoints): (WebCore::Range::boundaryPointsValid):
  • dom/Range.h:
14:31 Changeset [44367] by eric@webkit.org

2009-06-02 Eric Seidel <eric@webkit.org>

Reviewed by Maciej Stachowiak.

Remove Range::compareBoundaryPoints(Position, Position) per Darin's suggestion in bug 25056
https://bugs.webkit.org/show_bug.cgi?id=25500

Darin indicated the Range should deal only with primitive DOM node/offset
pairs, and that Position (which is a more robust editing construct) should have
its own comparison functions and that Range.h should not mention Position at all.

Turns out that Position already has a comparePositions() function (which knows
how to additionally handled positions in shadow trees). So I've just changed
all callers of compareBoundaryPoints(Position, Position) to use the existing
comparePositions() function. I've also added a comparePositions which takes
VisiblePositions for convenience.

  • dom/Range.cpp:
  • dom/Range.h:
  • editing/ApplyStyleCommand.cpp: (WebCore::ApplyStyleCommand::updateStartEnd): (WebCore::ApplyStyleCommand::applyBlockStyle): (WebCore::ApplyStyleCommand::applyRelativeFontStyleChange): (WebCore::ApplyStyleCommand::applyInlineStyle): (WebCore::ApplyStyleCommand::applyInlineStyleToRange): (WebCore::ApplyStyleCommand::removeInlineStyle): (WebCore::ApplyStyleCommand::nodeFullySelected): (WebCore::ApplyStyleCommand::nodeFullyUnselected):
  • editing/CompositeEditCommand.cpp: (WebCore::CompositeEditCommand::deleteInsignificantText): (WebCore::CompositeEditCommand::moveParagraphContentsToNewBlockIfNecessary): (WebCore::CompositeEditCommand::moveParagraphs):
  • editing/DeleteSelectionCommand.cpp: (WebCore::DeleteSelectionCommand::initializeStartEnd): (WebCore::DeleteSelectionCommand::handleGeneralDelete): (WebCore::DeleteSelectionCommand::mergeParagraphs):
  • editing/SelectionController.cpp: (WebCore::SelectionController::nodeWillBeRemoved):
  • editing/VisibleSelection.cpp: (WebCore::VisibleSelection::toNormalizedRange):
  • editing/htmlediting.cpp: (WebCore::comparePositions):
  • editing/htmlediting.h:
  • page/EventHandler.cpp: (WebCore::EventHandler::handleMousePressEventSingleClick):
14:09 Changeset [44366] by jianli@chromium.org

2009-06-02 Albert J. Wong <ajwong@chromium.org>

Reviewed by Dimitri Glazkov.

https://bugs.webkit.org/show_bug.cgi?id=25595
Upstream v8_utility.h functions into V8Utilities.h. This patch has
some transitional code to make upstreaming easier. This code will
be deleted in a few days.

  • bindings/v8/ScriptFunctionCall.cpp: (WebCore::ScriptFunctionCall::construct): NewInstance -> newInstance.
  • bindings/v8/V8Utilities.h: (WebCore::AllowAllocation::AllowAllocation): Function added. (WebCore::AllowAllocation::~AllowAllocation): Function added. (WebCore::SafeAllocation::NewInstance): Function added.
  • bindings/v8/WorkerContextExecutionProxy.cpp: (WebCore::WorkerContextExecutionProxy::initContextIfNeeded): NewInstance -> newInstance. (WebCore::WorkerContextExecutionProxy::toV8): NewInstance -> newInstance.
13:59 Changeset [44365] by andersca@apple.com

2009-06-02 Anders Carlsson <andersca@apple.com>

Rubber-stamped by Mark Rowe.

Update WebKitSystemInterface.

  • libWebKitSystemInterfaceLeopard.a:
  • libWebKitSystemInterfaceSnowLeopard.a:
  • libWebKitSystemInterfaceTiger.a:
13:25 Changeset [44364] by mitz@apple.com

Reviewed by John Sullivan.

  • fix <rdar://problem/6940747> Search field’s focus ring is outset
  • rendering/RenderThemeWin.cpp: (WebCore::RenderThemeWin::adjustSearchFieldStyle): Set the focused search field outline offset to -2.
12:01 Changeset [44363] by bfulgham@webkit.org

2009-06-02 Brent Fulgham <bfulgham@webkit.org>

Reviewed by Adam Roben.

Update WinLauncher for WebKitCreateInstance.
https://bugs.webkit.org/show_bug.cgi?id=26136

  • WinLauncher/WinLauncher.cpp: Revise implementation to use WebKitCreateInstance (rather than CoCreateInstance).
  • WinLauncher/WinLauncher.h: Modify include path for WebKit.h to match DumpRenderTree style.
  • WinLauncher/WinLauncher.vcproj: Remove unnecessary WebKit.tlb and EmbeddedIDL flags from build.
11:10 Changeset [44362] by weinig@apple.com

2009-06-02 Sam Weinig <sam@webkit.org>

Reviewed by Alexey Proskuryakov.

Add list of unimplemented event handlers to DOMWindow.

  • page/DOMWindow.idl:
10:59 Changeset [44361] by ap@webkit.org

Reviewed by Sam Weinig.

https://bugs.webkit.org/show_bug.cgi?id=26133
Adapt and import py-dom-xpath tests

Tests: fast/xpath/py-dom-xpath/abbreviations.html

fast/xpath/py-dom-xpath/axes.html
fast/xpath/py-dom-xpath/data.html
fast/xpath/py-dom-xpath/expressions.html
fast/xpath/py-dom-xpath/functions.html
fast/xpath/py-dom-xpath/nodetests.html
fast/xpath/py-dom-xpath/paths.html
fast/xpath/py-dom-xpath/predicates.html

Fix bugs found with this test suite:

  • name and local-name were incorrect for processing instructions (XPath expanded-name doesn't match DOM exactly);
  • name, local-name and namespace functions should crash on attribute nodes;
  • attemps to make node sets from other types were not detected as errors.

No performance impact.

  • xml/XPathExpressionNode.h: Track type conversion errors that happen during evaluation. An error won't stop evaluation, but an exception will be raised afterwards. We could also detect conversion errors at compile time, but not if we're going to support XPath variables (which is unnecessary for XPathEvaluator, but will be necessary if we decide to make our own XSLT one day).
  • xml/XPathExpression.cpp: (WebCore::XPathExpression::evaluate): Check whether a type conversion exception occurred during evaluation, and raise an excpetion if it did.
  • xml/XPathFunctions.cpp: (WebCore::XPath::expandedNameLocalPart): (WebCore::XPath::expandedName): XPath name(), local-name() and namespace-uri() functions are defined in terms of expanded-name, which doesn't match anything available via DOM exactly. Calculate the expanded name properly. (WebCore::XPath::FunNamespaceURI::evaluate): This function could crash if used with an attribute node, because it released what was possibly the only reference to attribute node before using it. Changed the function to avoid such situation. (WebCore::XPath::FunLocalName::evaluate): Ditto. Also, used the new expandedNameLocalPart() to work properly with processing instruction nodes. (WebCore::XPath::FunName::evaluate): Ditto (using expandedName()). (WebCore::XPath::FunCount::evaluate): Signal an error if the argument is not a node-set (by using toNodeSet unconditionally, which will raise an error, and return an empty set).
  • xml/XPathPath.cpp: (WebCore::XPath::Filter::evaluate): Signal an error if the expression evaluation result is not a node-set.
  • xml/XPathPath.h: (WebCore::XPath::Filter::resultType): A Filter's result is actually always a node-set (this is not so for FilterExpr production in the spec, but is for us, because we don't naively map BNF productions to classes).
  • xml/XPathPredicate.cpp: (WebCore::XPath::Union::evaluate): Signal an error if either side is not a node-set.
  • xml/XPathStep.cpp: Removed an unnecesary include.
  • xml/XPathValue.cpp: (WebCore::XPath::Value::toNodeSet): Signal an error if conversion fails. (WebCore::XPath::Value::modifiableNodeSet): Ditto. (WebCore::XPath::Value::toNumber): Don't allow inputs that don't match XPath Number production (in particular, those using exponential notation).
10:33 Changeset [44360] by weinig@apple.com

WebCore:

2009-06-01 Sam Weinig <sam@webkit.org>

Reviewed by Brady Eidson.

Part of https://bugs.webkit.org/show_bug.cgi?id=26100
Add missing event handler properties to the DOMWindow

Added oncontextmenu, oninput, and onmessage event handlers to
the DOMWindow. Aditionally, the onloadstart, onprogress, onstalled,
and onsuspend event handlers were implemented but not added to
DOMWindow.idl.

  • page/DOMWindow.cpp: (WebCore::DOMWindow::oninput): (WebCore::DOMWindow::setOninput): (WebCore::DOMWindow::onmessage): (WebCore::DOMWindow::setOnmessage): (WebCore::DOMWindow::oncontextmenu): (WebCore::DOMWindow::setOncontextmenu):
  • page/DOMWindow.h:
  • page/DOMWindow.idl:

LayoutTests:

2009-06-01 Sam Weinig <sam@webkit.org>

Reviewed by Brady Eidson.

Part of https://bugs.webkit.org/show_bug.cgi?id=26100
Add missing event handler properties to the DOMWindow

Added oncontextmenu, oninput, and onmessage event handlers to
the DOMWindow. Aditionally, the onloadstart, onprogress, onstalled,
and onsuspend event handlers were implemented but not added to
DOMWindow.idl.

  • fast/dom/Window/window-properties-expected.txt:
09:57 Changeset [44359] by xan@webkit.org

2009-06-02 Xan Lopez <xlopez@igalia.com>

Reviewed by Sam Weinig.

Use C-style comments in Platform.h so it can be included from C
files.

  • wtf/Platform.h:
07:54 Changeset [44358] by vestbo@webkit.org

2009-06-02 Tor Arne Vestbø <tor.arne.vestbo@nokia.com>

Rubber-stamped by Adam Roben.

[Qt] Fix Perl warning about uninitialized value (passedConfiguration())

06:54 Changeset [44357] by vestbo@webkit.org

2009-06-02 Tor Arne Vestbø <tor.arne.vestbo@nokia.com>

Rubber-stamped by Simon Hausmann.

Use File::Spec->tmpdir instead of hardcoded paths for tempfile() dir

This fixes the Windows-build if the user does not have a /tmp directory.

02:34 Changeset [44356] by barraclough@apple.com

2009-06-02 Gavin Barraclough <barraclough@apple.com>

Reviewed by Oliver ">>" Hunt.

emitSlow_op_rshift is linking the wrong number of slow cases, if !supportsFloatingPoint().
Fixerate, and refactor/comment the code a little to make it clearer what is going on.

  • jit/JITArithmetic.cpp: (JSC::JIT::emit_op_rshift): (JSC::JIT::emitSlow_op_rshift):
02:17 Changeset [44355] by abarth@webkit.org

Revert 44354. Test does not pass.

01:37 Changeset [44354] by abarth@webkit.org

WebCore:

2009-06-02 Grace Kloba <klobag@gmail.com>

Reviewed by Darin Adler. Landed by Adam Barth.

https://bugs.webkit.org/show_bug.cgi?id=26016

As HTMLKeygenElement is derived from HTMLSelectElement, HTMLOptionElement::ownerSelectElement() needs to check both selectTag and keygenTag.
Test: fast/dom/HTMLKeygenElement/keygen-option-select.html

  • html/HTMLOptionElement.cpp: (WebCore::HTMLOptionElement::ownerSelectElement):

LayoutTests:

2009-06-02 Grace Kloba <klobag@gmail.com>

Reviewed by Darin Adler. Landed by Adam Barth.

https://bugs.webkit.org/show_bug.cgi?id=26016

As HTMLKeygenElement is derived from HTMLSelectElement, HTMLOptionElement::ownerSelectElement() needs to check both selectTag and keygenTag. Added layout test to make sure set selected on option will affect keygen element.

  • fast/dom/HTMLKeygenElement: Added.
  • fast/dom/HTMLKeygenElement/keygen-option-select-expected.txt: Added.
  • fast/dom/HTMLKeygenElement/keygen-option-select.html: Added.
01:32 Changeset [44353] by abarth@webkit.org

Add link to bug.

01:30 Changeset [44352] by abarth@webkit.org

Revert 44351.

00:56 Changeset [44351] by abarth@webkit.org

WebCore:

2009-06-01 Ben Murdoch <benm@google.com>

Reviewed by Darin Adler. Landed by Adam Barth.

https://bugs.webkit.org/show_bug.cgi?id=25710

Fix a bug that causes database transactions to fail if a history navigation to a hash fragment of the same document is made while resources (e.g. an XHR) are loading.

Test: storage/hash-change-with-xhr.html

  • loader/DocumentLoader.cpp: (WebCore::DocumentLoader::stopLoading):
  • loader/DocumentLoader.h:
  • loader/FrameLoader.cpp: (WebCore::FrameLoader::stopLoading): (WebCore::FrameLoader::stopAllLoaders):
  • loader/FrameLoader.h:
  • loader/FrameLoaderTypes.h: (WebCore::):
  • page/Page.cpp: (WebCore::Page::goToItem):

LayoutTests:

2009-06-02 Ben Murdoch <benm@google.com>

Reviewed by Darin Alder. Landed by Adam Barth.

https://bugs.webkit.org/show_bug.cgi?id=25710

Fix a bug that causes database transactions to fail if a history navigation to a hash fragment of the same document is made while resources (e.g. an XHR) are loading.

  • storage/hash-change-with-xhr-expected.txt: Added.
  • storage/hash-change-with-xhr.html: Added.
00:26 Changeset [44350] by abarth@webkit.org

2009-06-02 Johnny Ding <johnnyding.webkit@gmail.com>

Reviewed by Alexey Proskuryakov. Landed by Adam Barth.

https://bugs.webkit.org/show_bug.cgi?id=24747

Add a test case for multiple requests for same sub-resource due to preload.

  • http/tests/loading/preload-img-test-expected.txt: Added.
  • http/tests/loading/preload-img-test.html: Added.
  • http/tests/loading/resources/preload-test.jpg: Added.
  • http/tests/resources/network-simulator.php:
00:00 Changeset [44349] by abarth@webkit.org

2009-06-01 Yongjun Zhang <yongjun.zhang@nokia.com>>

Reviewed by Holger Freyther. Landed by Adam Barth.

<https://bugs.webkit.org/show_bug.cgi?id=26070>

[Qt] EventSender.keyDown doesn't support Arrow Up and Down keys.
Add Down and Up key support.


  • DumpRenderTree/qt/jsobjects.cpp: (EventSender::keyDown):

06/01/09:

23:40 Changeset [44348] by abarth@webkit.org

2009-06-01 Jeremy Orlow <jorlow@chromium.org>

Reviewed by Darin Adler. Landed by Adam Barth.

Remove a redundant checkEncodedString call when constructing a
KURL object from a string.

  • platform/KURL.cpp: (WebCore::KURL::KURL):
22:59 Changeset [44347] by barraclough@apple.com

2009-06-01 Gavin Barraclough <barraclough@apple.com>

Reviewed by NOBODY - speculative windows build fix (errm, for the other patch!).

  • jit/JITStubs.cpp: (JSC::):
22:54 Changeset [44346] by barraclough@apple.com

2009-06-01 Gavin Barraclough <barraclough@apple.com>

Reviewed by NOBODY - speculative windows build fix.

  • assembler/AbstractMacroAssembler.h: (JSC::::CodeLocationCall::CodeLocationCall): (JSC::::CodeLocationNearCall::CodeLocationNearCall):
22:46 Changeset [44345] by ggaren@apple.com

JavaScriptCore:

2009-06-01 Geoffrey Garen <ggaren@apple.com>

Reviewed by Sam Weinig.


Fixed some of the regression in crypto-aes.js. (8.5% speedup in
crypto-aes.js.)


SunSpider reports no change overall.


Division was producing double results, which took the slow path through
array access code.


Strangely, all my attempts at versions of this patch that modified array
access code to accept ints encoded as doubles along the fast or slow paths
were regressions. So I did this instead.

  • jit/JITArithmetic.cpp: (JSC::JIT::emit_op_div): When dividing an int by an int, go ahead and try to turn the result into an int. Don't just do int division, though, because testing shows it to be slower than SSE double division, and the corner cases are pretty complicated / lengthy on top of that. Also, don't try to canonicalize division of known tiny numerators into ints, since that's a waste of time.

LayoutTests:

2009-06-01 Geoffrey Garen <ggaren@apple.com>

Reviewed by Sam Weinig.


Added test cases for some division edge cases I broke.

  • fast/js/arithmetic-expected.txt:
  • fast/js/resources/arithmetic.js: Also modified this test to avoid constant folding.
22:36 Changeset [44344] by barraclough@apple.com

2009-06-01 Gavin Barraclough <barraclough@apple.com>

Reviewed by Olliej Hunt.

Change JITStub functions from being static members on the JITStub class to be
global extern "C" functions, and switch their the function signature declaration
in the definition of the functions to be C-macro generated. This makes it easier
to work with the stub functions from assembler code (since the names no longer
require mangling), and by delaring the functions with a macro we can look at
also auto-generating asm thunks to wrap the JITStub functions to perform the
work currently in 'restoreArgumentReference' (as a memory saving).

Making this change also forces us to be a bit more realistic about what is private
on the Register and CallFrame objects. Presently most everything on these classes
is private, and the classes have plenty of friends. We could befriend all the
global functions to perpetuate the delusion of encapsulation, but using friends is
a bit of a sledgehammer solution here - since friends can poke around with all of
the class's privates, and since all the major classes taht operate on Regsiters are
currently friends, right there is currently in practice very little protection at
all. Better to start removing friend delclarations, and exposing just the parts
that need to be exposed.

  • interpreter/CallFrame.h: (JSC::ExecState::returnPC): (JSC::ExecState::setCallerFrame): (JSC::ExecState::returnValueRegister): (JSC::ExecState::setArgumentCount): (JSC::ExecState::setCallee): (JSC::ExecState::setCodeBlock):
  • interpreter/Interpreter.h:
  • interpreter/Register.h: (JSC::Register::Register): (JSC::Register::i):
  • jit/JITStubs.cpp: (JSC::): (JSC::JITThunks::JITThunks): (JSC::JITThunks::tryCachePutByID): (JSC::JITThunks::tryCacheGetByID): (JSC::JITStubs::DEFINE_STUB_FUNCTION):
  • jit/JITStubs.h: (JSC::JITStubs::):
  • runtime/JSFunction.h: (JSC::JSFunction::nativeFunction): (JSC::JSFunction::classInfo):
  • runtime/JSGlobalData.h:
21:40 Changeset [44343] by oliver@apple.com

Tidy up the literal parser.

Reviewed by Gavin Barraclogh.

Make the number lexing in the LiteralParser exactly match the JSON spec, which
makes us cover more cases, but also more strict. Also made string lexing only
allow double-quoted strings.

20:03 Changeset [44342] by dglazkov@chromium.org

2009-06-01 Dimitri Glazkov <dglazkov@chromium.org>

Unreviewed, build fix.

Reposition platform guard, improperly placed by http://trac.webkit.org/changeset/44340
Bad Dimitri.

  • platform/KeyboardCodes.h: Repositioned the guard to avoid nested

WebCore namespace declarations.

18:20 Changeset [44341] by barraclough@apple.com

2009-06-01 Gavin Barraclough <barraclough@apple.com>

Reviewed by Sam "WX" Weinig.

Allow the JIT to operate without relying on use of RWX memory, on platforms where this is supported.

This patch adds a switch to Platform.h (ENABLE_ASSEMBLER_WX_EXCLUSIVE) which enables this mode of operation.
When this flag is set, all executable memory will be allocated RX, and switched to RW only whilst being
modified. Upon completion of code generation the protection is switched back to RX to allow execution.

Further optimization will be required before it is desirable to enable this mode of operation by default;
enabling this presently incurs a 5%-10% regression.

(Submitting disabled - no performance impact).

  • assembler/AbstractMacroAssembler.h: (JSC::AbstractMacroAssembler::CodeLocationInstruction::repatchLoadToLEA): (JSC::AbstractMacroAssembler::CodeLocationLabel::fromFunctionPointer): (JSC::AbstractMacroAssembler::CodeLocationJump::relink): (JSC::AbstractMacroAssembler::CodeLocationCall::relink): (JSC::AbstractMacroAssembler::CodeLocationNearCall::relink): (JSC::AbstractMacroAssembler::CodeLocationDataLabel32::repatch): (JSC::AbstractMacroAssembler::CodeLocationDataLabelPtr::repatch): (JSC::AbstractMacroAssembler::ProcessorReturnAddress::relinkCallerToTrampoline): (JSC::AbstractMacroAssembler::ProcessorReturnAddress::relinkCallerToFunction): (JSC::AbstractMacroAssembler::ProcessorReturnAddress::relinkNearCallerToTrampoline): (JSC::AbstractMacroAssembler::ProcessorReturnAddress::relinkNearCallerToFunction): (JSC::AbstractMacroAssembler::PatchBuffer::PatchBuffer): (JSC::AbstractMacroAssembler::PatchBuffer::~PatchBuffer): (JSC::AbstractMacroAssembler::PatchBuffer::link): (JSC::AbstractMacroAssembler::PatchBuffer::patch): (JSC::AbstractMacroAssembler::PatchBuffer::performFinalization): (JSC::::CodeLocationCommon::nearCallAtOffset): (JSC::::CodeLocationCall::CodeLocationCall): (JSC::::CodeLocationNearCall::CodeLocationNearCall):
  • assembler/AssemblerBuffer.h: (JSC::AssemblerBuffer::executableCopy):
  • assembler/X86Assembler.h: (JSC::CAN_SIGN_EXTEND_U32_64): (JSC::X86Assembler::linkJump): (JSC::X86Assembler::linkCall): (JSC::X86Assembler::patchPointer): (JSC::X86Assembler::relinkJump): (JSC::X86Assembler::relinkCall): (JSC::X86Assembler::repatchInt32): (JSC::X86Assembler::repatchPointer): (JSC::X86Assembler::repatchLoadToLEA): (JSC::X86Assembler::patchInt32): (JSC::X86Assembler::patchRel32):
  • jit/ExecutableAllocator.h: (JSC::ExecutableAllocator::): (JSC::ExecutableAllocator::makeWritable): (JSC::ExecutableAllocator::makeExecutable):
  • jit/ExecutableAllocatorFixedVMPool.cpp: (JSC::FixedVMPoolAllocator::FixedVMPoolAllocator):
  • jit/ExecutableAllocatorPosix.cpp: (JSC::ExecutablePool::systemAlloc): (JSC::ExecutablePool::systemRelease): (JSC::ExecutableAllocator::reprotectRegion):
  • jit/ExecutableAllocatorWin.cpp:
  • jit/JITPropertyAccess.cpp: (JSC::JIT::patchGetByIdSelf): (JSC::JIT::patchPutByIdReplace):
  • wtf/Platform.h:
16:48 Changeset [44340] by dglazkov@chromium.org

2009-06-01 Dimitri Glazkov <dglazkov@chromium.org>

Unreviewed, build fix.

  • platform/KeyboardCodes.h: Integrated contents of platform/chromium/KeyboardCodes.h

with a stern FIXME.

  • platform/chromium/KeyboardCodes.h: Removed.
16:38 Changeset [44339] by mrowe@apple.com

svn-create-patch displays confusing "Broken pipe" errors from Subversion
when files have been renamed in local working copy. This was happening
because we invoke 'svn info' to determine the URL that the file was copied
from, but we were closing the pipe to the subprocess before SVN had finished
writing its output.

Reviewed by Sam Weinig.

  • Scripts/prepare-ChangeLog: Consume all 'svn info' output before closing the file handle.
  • Scripts/svn-create-patch: Ditto.
16:09 Changeset [44338] by zimmermann@webkit.org

Reviewed by Eric Seidel.

Fix assertion error in --filters enabled debug builds.
Instead of using RefPtr<AtomicStringImpl> as keys for the hash maps in SVGFilterBuilder, just use AtomicString objects.

15:20 Changeset [44337] by zimmermann@webkit.org

Reviewed by Eric Seidel.

Fix --filters enabled build on Mac.
Remove unnecessary 'boundingBox' parameter from finishRenderSVGContent() method.
Kill several warnings, to make build pass.

15:19 Changeset [44336] by bfulgham@webkit.org

2009-05-29 Zoltan Horvath <hzoltan@inf.u-szeged.hu>

Reviewed by Darin Adler.

Inherits Interpreter class from FastAllocBase because it has been
instantiated by 'new' in JavaScriptCore/runtime/JSGlobalData.cpp.

  • interpreter/Interpreter.h:
14:52 Changeset [44335] by dglazkov@chromium.org

ChangeLog spelling, meaning fixes.

14:50 Changeset [44334] by dglazkov@chromium.org

2009-06-01 Dimitri Glazkov <dglazkov@chromium.org>

Reviewed by Dave Hyatt.

Fix Chromium build by adding an #ifdef.

  • platform/mac/ScrollbarThemeMac.mm: (WebCore::ScrollbarThemeMac::paint): Added an #ifdef.
14:37 Changeset [44333] by kov@webkit.org

2009-06-01 Benjamin C Meyer <benjamin.meyer@torchmobile.com>

Rubber-stamped by Gustavo Noronha.

Before guessing the url trim the string to remove whitespace added to the start/end by the user.

  • Api/qwebview.cpp: (QWebView::guessUrlFromString):
  • tests/qwebview/tst_qwebview.cpp: (tst_QWebView::guessUrlFromString_data):
14:28 Changeset [44332] by krit@webkit.org

2009-06-01 Dirk Schulze <krit@webkit.org>

Reviewed by Nikolas Zimmermann.

Remove last pieces of the old SVG filte system. They are not
usable with our current filter system. The new filter effects
will replace the functionality step by step.

  • GNUmakefile.am:
  • WebCore.pro:
  • WebCore.vcproj/WebCore.vcproj:
  • WebCore.xcodeproj/project.pbxproj:
  • svg/graphics/cairo: Removed.
  • svg/graphics/cairo/SVGResourceFilterCairo.cpp: Removed.
  • svg/graphics/cg: Removed.
  • svg/graphics/cg/SVGResourceFilterCg.cpp: Removed.
  • svg/graphics/cg/SVGResourceFilterCg.mm: Removed.
  • svg/graphics/filters/SVGFilterEffect.cpp: Removed.
  • svg/graphics/filters/SVGFilterEffect.h: Removed.
  • svg/graphics/filters/cg: Removed.
  • svg/graphics/filters/cg/SVGFEHelpersCg.h: Removed.
  • svg/graphics/filters/cg/SVGFEHelpersCg.mm: Removed.
  • svg/graphics/filters/cg/SVGFilterEffectCg.mm: Removed.
  • svg/graphics/filters/cg/WKArithmeticFilter.cikernel: Removed.
  • svg/graphics/filters/cg/WKArithmeticFilter.h: Removed.
  • svg/graphics/filters/cg/WKArithmeticFilter.m: Removed.
  • svg/graphics/filters/cg/WKComponentMergeFilter.cikernel: Removed.
  • svg/graphics/filters/cg/WKComponentMergeFilter.h: Removed.
  • svg/graphics/filters/cg/WKComponentMergeFilter.m: Removed.
  • svg/graphics/filters/cg/WKDiffuseLightingFilter.cikernel: Removed.
  • svg/graphics/filters/cg/WKDiffuseLightingFilter.h: Removed.
  • svg/graphics/filters/cg/WKDiffuseLightingFilter.m: Removed.
  • svg/graphics/filters/cg/WKDiscreteTransferFilter.cikernel: Removed.
  • svg/graphics/filters/cg/WKDiscreteTransferFilter.h: Removed.
  • svg/graphics/filters/cg/WKDiscreteTransferFilter.m: Removed.
  • svg/graphics/filters/cg/WKDisplacementMapFilter.cikernel: Removed.
  • svg/graphics/filters/cg/WKDisplacementMapFilter.h: Removed.
  • svg/graphics/filters/cg/WKDisplacementMapFilter.m: Removed.
  • svg/graphics/filters/cg/WKDistantLightFilter.cikernel: Removed.
  • svg/graphics/filters/cg/WKDistantLightFilter.h: Removed.
  • svg/graphics/filters/cg/WKDistantLightFilter.m: Removed.
  • svg/graphics/filters/cg/WKGammaTransferFilter.cikernel: Removed.
  • svg/graphics/filters/cg/WKGammaTransferFilter.h: Removed.
  • svg/graphics/filters/cg/WKGammaTransferFilter.m: Removed.
  • svg/graphics/filters/cg/WKIdentityTransferFilter.h: Removed.
  • svg/graphics/filters/cg/WKIdentityTransferFilter.m: Removed.
  • svg/graphics/filters/cg/WKLinearTransferFilter.cikernel: Removed.
  • svg/graphics/filters/cg/WKLinearTransferFilter.h: Removed.
  • svg/graphics/filters/cg/WKLinearTransferFilter.m: Removed.
  • svg/graphics/filters/cg/WKNormalMapFilter.cikernel: Removed.
  • svg/graphics/filters/cg/WKNormalMapFilter.h: Removed.
  • svg/graphics/filters/cg/WKNormalMapFilter.m: Removed.
  • svg/graphics/filters/cg/WKPointLightFilter.cikernel: Removed.
  • svg/graphics/filters/cg/WKPointLightFilter.h: Removed.
  • svg/graphics/filters/cg/WKPointLightFilter.m: Removed.
  • svg/graphics/filters/cg/WKSpecularLightingFilter.cikernel: Removed.
  • svg/graphics/filters/cg/WKSpecularLightingFilter.h: Removed.
  • svg/graphics/filters/cg/WKSpecularLightingFilter.m: Removed.
  • svg/graphics/filters/cg/WKSpotLightFilter.cikernel: Removed.
  • svg/graphics/filters/cg/WKSpotLightFilter.h: Removed.
  • svg/graphics/filters/cg/WKSpotLightFilter.m: Removed.
  • svg/graphics/filters/cg/WKTableTransferFilter.cikernel: Removed.
  • svg/graphics/filters/cg/WKTableTransferFilter.h: Removed.
  • svg/graphics/filters/cg/WKTableTransferFilter.m: Removed.
  • svg/graphics/mac: Removed.
  • svg/graphics/mac/SVGResourceFilterPlatformDataMac.h: Removed.
  • svg/graphics/mac/SVGResourceFilterPlatformDataMac.mm: Removed.
  • svg/graphics/qt: Removed.
  • svg/graphics/qt/SVGResourceFilterQt.cpp: Removed.
  • svg/graphics/skia: Removed.
  • svg/graphics/skia/SVGResourceFilterSkia.cpp: Removed.
14:05 Changeset [44331] by kov@webkit.org

2009-06-01 Benjamin C Meyer <benjamin.meyer@torchmobile.com>

Rubber-stamped by Gustavo Noronha.

Use QLatin1Char() to compile when QT_NO_CAST_FROM_ASCII and QT_NO_CAST_TO_ASCII is defined

  • Api/qwebview.cpp: (QWebView::guessUrlFromString):
13:06 Changeset [44330] by kov@webkit.org

2009-06-01 Benjamin C Meyer <benjamin.meyer@torchmobile.com>

Rubber-stamped by Gustavo Noronha.

Cleanup autotest file

  • use proper includes
  • remove useless constructor and destructor
  • add missing newline at the end of the file
  • tests/qwebview/tst_qwebview.cpp:
13:01 Changeset [44329] by kov@webkit.org

2009-06-01 Benjamin C Meyer <benjamin.meyer@torchmobile.com>

Rubber-stamped by Gustavo Noronha.

Add a missing space between two words in the documentation.

  • Api/qwebview.cpp:
12:56 Changeset [44328] by dimich@chromium.org

2009-06-01 Dmitry Titov <dimich@chromium.org>

Fix the previous checkin (ttp://trac.webkit.org/changeset/44327).
This adds a comment suggested during review.

  • platform/ThreadGlobalData.cpp: (WebCore::ThreadGlobalData::~ThreadGlobalData): Add comment clarifying the change.
12:44 Changeset [44327] by dimich@chromium.org

2009-06-01 Dmitry Titov <dimich@chromium.org>

Reviewed by Darin Adler.

https://bugs.webkit.org/show_bug.cgi?id=25973
Avoid calling CurrentThread() in thread-specific destructors in OSX Chromium.
Pthreads invoke thread-specific destructors after WTF::detachThread() is called and ThreadIdentifier
for the thread removed from the WTF thread map. Calling CurrentThread() in such destructor causes
the ThreadIdentifier to be re-created and inserted into the map again. Since Pthreads on OSX reuse
the pthread_t between threads, the next created thread will have the same pthread_t and cause an assert
in establishIdentifierForPthreadHandle() since the id is already in the map.

The behavior is covered by existing test LayoutTests/fast/workers/worker-terminate.html, which currently fails
on OSX Chromium and will stop failing after this change.

  • platform/ThreadGlobalData.h:
  • platform/ThreadGlobalData.cpp: (WebCore::ThreadGlobalData::~ThreadGlobalData): Store the result of "isMainThread()" in a member variable during construction of thread-specific data to avoid calling IsMainThread() in destructor, since the latter calls CurrentThread() in OSX Chromium.
12:18 Changeset [44326] by levin@chromium.org

2009-06-01 David Levin <levin@chromium.org>

Reviewed by NOBODY (windows build fix).

Add exports for windows (corresponding to the JavaScriptCore.exp modification
in the previous change).

11:41 Changeset [44325] by levin@chromium.org

JavaScriptCore:

2009-06-01 David Levin <levin@chromium.org>

Reviewed by Darin Alder and Maciej Stachowiak.

Bug 26057: StringImpl should share buffers with UString.
https://bugs.webkit.org/show_bug.cgi?id=26057

  • JavaScriptCore.exp:
  • runtime/UString.cpp: (JSC::UString::Rep::create): (JSC::UString::BaseString::sharedBuffer): Only do the sharing when the buffer exceeds a certain size. The size was tuned by running various dom benchmarks with numbers ranging from 20 to 800 and finding a place that seemed to do the best overall.
  • runtime/UString.h:

WebCore:

2009-06-01 David Levin <levin@chromium.org>

Reviewed by Darin Alder and Maciej Stachowiak.

Bug 26057: StringImpl should share buffers with UString.
https://bugs.webkit.org/show_bug.cgi?id=26057

This change results in the following performance improvements:
On http://www.hixie.ch/tests/adhoc/perf/dom/artificial/core/001.html
the time went from 78ms to 40ms for append (other times remained constant).

On http://www.hixie.ch/tests/adhoc/perf/dom/artificial/core/002.html,
the time went from 3900ms to 2600ms.

For http://dromaeo.com/?dom, the time for DomModification improved by ~6%.
Other tests in dom seemed to be faster across several runs but within the
margin of error (except DOM Attributes which was slightly ~1.5% worse).

Existing tests cover this code and there is no new functionality
that is exposed to test.

  • platform/text/AtomicString.cpp: (WebCore::AtomicString::add):
  • platform/text/String.cpp: (WebCore::String::String): (WebCore::String::operator UString):
  • platform/text/StringImpl.cpp: (WebCore::StringImpl::StringImpl): (WebCore::StringImpl::~StringImpl): (WebCore::StringImpl::create): Consumes a shared buffer. (WebCore::StringImpl::ustring): Shares the StringImpl's buffer with the UString. (WebCore::StringImpl::sharedBuffer): Exposes the buffer that may be shared.
  • platform/text/StringImpl.h: (WebCore::StringImpl::hasTerminatingNullCharacter): (WebCore::StringImpl::inTable): (WebCore::StringImpl::setInTable): Converted the bools to be inside of PtrAndFlags to avoid growing StringImpl in size.
10:35 WebKit Team edited by kov@webkit.org
(diff)
10:16 Changeset [44324] by bfulgham@webkit.org

2009-06-01 Brent Fulgham <bfulgham@webkit.org>

Build fix for Cairo Windows target. Windows link definition
files were out of sync with the Apple Windows link files.

  • WebKit.vcproj/WebKit_Cairo.def:
  • WebKit.vcproj/WebKit_Cairo_debug.def:
09:53 Changeset [44323] by dglazkov@chromium.org

2009-06-01 Dimitri Glazkov <dglazkov@chromium.org>

Unreviewed, build fix.

V8 bindings follow-up to to DOMWindow-related cleanup:
http://trac.webkit.org/changeset/44215

  • bindings/v8/ScriptController.cpp: Removed disconnectFrame(), relocated

its body to destructor.

  • bindings/v8/ScriptController.h: Removed disconnectFrame() decl.
09:16 Changeset [44322] by dglazkov@chromium.org

2009-05-28 Pavel Feldman <pfeldman@chromium.org>

Reviewed by Timothy Hatcher.

  • Add panel enabler to the resources panel.
  • Add session / always options into the panel enabler.
  • Make enabled status for three panels sticky (globally).
  • Persist enabled status using InspectorController::Settings
  • Make InspectorController produce no network-related overhead when resources panel is not enabled.

https://bugs.webkit.org/show_bug.cgi?id=26046

  • inspector/InspectorController.cpp: (WebCore::InspectorController::InspectorController): (WebCore::InspectorController::setWindowVisible): (WebCore::InspectorController::populateScriptObjects): (WebCore::InspectorController::identifierForInitialRequest): (WebCore::InspectorController::willSendRequest): (WebCore::InspectorController::didReceiveResponse): (WebCore::InspectorController::didReceiveContentLength): (WebCore::InspectorController::didFinishLoading): (WebCore::InspectorController::didFailLoading): (WebCore::InspectorController::resourceRetrievedByXMLHttpRequest): (WebCore::InspectorController::scriptImported): (WebCore::InspectorController::enableResourceTracking): (WebCore::InspectorController::disableResourceTracking): (WebCore::InspectorController::startUserInitiatedProfiling): (WebCore::InspectorController::enableProfiler): (WebCore::InspectorController::disableProfiler): (WebCore::InspectorController::enableDebugger): (WebCore::InspectorController::disableDebugger):
  • inspector/InspectorController.h: (WebCore::InspectorController::Setting::Setting): (WebCore::InspectorController::resourceTrackingEnabled):
  • inspector/InspectorController.idl:
  • inspector/InspectorFrontend.cpp: (WebCore::InspectorFrontend::resourceTrackingWasEnabled): (WebCore::InspectorFrontend::resourceTrackingWasDisabled):
  • inspector/InspectorFrontend.h:
  • inspector/front-end/PanelEnablerView.js: (WebInspector.PanelEnablerView.enableOption): (WebInspector.PanelEnablerView): (WebInspector.PanelEnablerView.prototype._windowResized): (WebInspector.PanelEnablerView.prototype.alwaysWasChosen):
  • inspector/front-end/ProfilesPanel.js: (WebInspector.ProfilesPanel.prototype._enableProfiling): (WebInspector.ProfilesPanel.prototype._toggleProfiling):
  • inspector/front-end/ResourcesPanel.js: (WebInspector.ResourcesPanel): (WebInspector.ResourcesPanel.prototype.get statusBarItems): (WebInspector.ResourcesPanel.prototype.resourceTrackingWasEnabled): (WebInspector.ResourcesPanel.prototype.resourceTrackingWasDisabled): (WebInspector.ResourcesPanel.prototype.reset): (WebInspector.ResourcesPanel.prototype._updateSidebarWidth): (WebInspector.ResourcesPanel.prototype._enableResourceTracking): (WebInspector.ResourcesPanel.prototype._toggleResourceTracking):
  • inspector/front-end/ScriptsPanel.js: (WebInspector.ScriptsPanel.prototype._enableDebugging): (WebInspector.ScriptsPanel.prototype._toggleDebugging):
  • inspector/front-end/inspector.css:
  • inspector/front-end/inspector.js: (WebInspector.resourceTrackingWasEnabled): (WebInspector.resourceTrackingWasDisabled):
03:14 Changeset [44321] by abarth@webkit.org

Revert 44320. The test didn't pass on Windows.

02:05 Changeset [44320] by abarth@webkit.org

2009-06-01 Jungshik Shin <jshin@chromium.org>

Reviewed by Maciej Stachowiak. Landed by Adam Barth.

Add an additional test for bug 23786 (mistreatment of CR/LF
as zero-width characters in Chromium port). Tests using Bidi text were
added when the fix was checked in. Without the patch, Chromium used
to fail the same way with LTR complex scripts as well. (that is,
CR and LF were treated as zero-width space). To prevent a potential
regression in the future, a new test with Hindi (a LTR complex script)
is added.

https://bugs.webkit.org/show_bug.cgi?id=25698

  • fast/text/international/hindi-whitespace.html: Added.
  • platform/mac/fast/text/international/hindi-whitespace-expected.checksum: Added.
  • platform/mac/fast/text/international/hindi-whitespace-expected.png: Added.
  • platform/mac/fast/text/international/hindi-whitespace-expected.txt: Added.
02:00 Changeset [44319] by abarth@webkit.org

WebCore:

2009-06-01 Drew Wilson <atwilson@google.com>

Reviewed by Darin Adler. Landed (and tweaked) by Adam Barth.

https://bugs.webkit.org/show_bug.cgi?id=25902

Added WorkerContext.close()

Test: fast/workers/worker-close.html

  • workers/WorkerContext.cpp: (WebCore::WorkerContext::close):
  • workers/WorkerContext.h:
  • workers/WorkerContext.idl:
  • workers/WorkerMessagingProxy.cpp: (WebCore::WorkerMessagingProxy::workerContextDestroyedInternal):

LayoutTests:

2009-06-01 Drew Wilson <atwilson@google.com>

Reviewed by Darin Adler. Landed by Adam Barth.

https://bugs.webkit.org/show_bug.cgi?id=25902

Added WorkerContext.close()

  • fast/workers/worker-close-expected.txt: Added.
  • fast/workers/worker-close.html: Added.
  • fast/workers/resources/worker-close.js: Added.
  • http/tests/xmlhttprequest/workers/close.html: Added.
01:30 Changeset [44318] by abarth@webkit.org

Add missing expected result.

01:27 Changeset [44317] by ap@webkit.org

Reviewed by Darin Adler.

https://bugs.webkit.org/show_bug.cgi?id=12471
XPathResult not invalidated for documents retrieved via XMLHttpRequest

Test: fast/xpath/detached-subtree-invalidate-iterator.html and existing tests in dom/svg/level3/xpath.

Use DOM tree version instead of DOMSubtreeModified events to invalidate, which is more
reliable and much faster.

  • xml/XPathExpression.cpp: (WebCore::XPathExpression::evaluate):
  • xml/XPathResult.cpp: (WebCore::XPathResult::XPathResult): (WebCore::XPathResult::~XPathResult): (WebCore::XPathResult::invalidIteratorState): (WebCore::XPathResult::iterateNext):
  • xml/XPathResult.h: (WebCore::XPathResult::create):
01:18 Changeset [44316] by darin@apple.com

2009-06-01 Darin Adler <darin@apple.com>

  • WebView/WebTextCompletionController.mm: Fix Tiger build by adding import of WebTypesInternal.h.
01:11 Changeset [44315] by abarth@webkit.org

WebCore:

2009-06-01 Brett Wilson <brettw@chromium.org>

Reviewed by Darin Adler. Landed by Adam Barth.

https://bugs.webkit.org/show_bug.cgi?id=25750

Test: fast/transforms/bounding-rect-zoom.html

Make getClientRects and getBoundingClientRect account for ther zoom
factor.

  • dom/Element.cpp: (WebCore::adjustFloatPointForAbsoluteZoom): (WebCore::adjustFloatQuadForAbsoluteZoom): (WebCore::adjustIntRectForAbsoluteZoom): (WebCore::Element::getClientRects): (WebCore::Element::getBoundingClientRect):

LayoutTests:

2009-06-01 Brett Wilson <brettw@chromium.org>

Reviewed by Darin Adler. Landed by Adam Barth.

https://bugs.webkit.org/show_bug.cgi?id=25750

Test: fast/transforms/bounding-rect-zoom.html

Make getClientRects and getBoundingClientRect account for ther zoom
factor.

  • fast/transforms/bounding-rect-zoom.html: Added.
  • platform/mac/fast/transforms/bounding-rect-zoom-expected.txt: Added.
01:07 Changeset [44314] by abarth@webkit.org

Add bogus expected result to trick the bot into showing me the real expected results (as suggested by bdash).

00:46 Changeset [44313] by darin@apple.com

WebKit:

2009-06-01 Darin Adler <darin@apple.com>

Reviewed by Maciej Stachowiak.

Bug 26113: break WebTextCompletionController out into its own source file
https://bugs.webkit.org/show_bug.cgi?id=26113

  • WebKit.xcodeproj/project.pbxproj: Add WebTextCompletionController.h/mm.

WebKit/mac:

2009-06-01 Darin Adler <darin@apple.com>

Reviewed by Maciej Stachowiak.

Bug 26113: break WebTextCompletionController out into its own source file
https://bugs.webkit.org/show_bug.cgi?id=26113

  • WebView/WebHTMLView.mm: (-[WebHTMLViewPrivate dealloc]): Updated for new name. (-[WebHTMLViewPrivate clear]): Ditto. (-[WebHTMLView _frameOrBoundsChanged]): Ditto. (-[WebHTMLView menuForEvent:]): Ditto. (-[WebHTMLView windowDidResignKey:]): Ditto. (-[WebHTMLView windowWillClose:]): Ditto. (-[WebHTMLView mouseDown:]): Ditto. (-[WebHTMLView resignFirstResponder]): Ditto. (-[WebHTMLView keyDown:]): Ditto. (-[WebHTMLView complete:]): Ditto. Also pass WebView to init method. (-[WebHTMLView _interceptEditingKeyEvent:shouldSaveCommand:]): Ditto.
  • WebView/WebTextCompletionController.h: Copied from mac/WebView/WebHTMLView.mm.
  • WebView/WebTextCompletionController.mm: Copied from mac/WebView/WebHTMLView.mm. Changed initializer to pass in a spearate WebView and WebHTMLView, to smooth the way for handling viewless mode properly in the future.
00:43 Changeset [44312] by abarth@webkit.org

2009-06-01 Tony Chang <tony@chromium.org>

Reviewed by Dimitri Glazkov. Landed by Adam Barth.

https://bugs.webkit.org/show_bug.cgi?id=26026

Fix an infinite loop when using the keyboard in Chromium select
popups.

Not testable since it involves sending a keyboard event to
the popup, which is not possible (eventSender sends the key
events through webview, we want to go through the webwidget).

  • platform/chromium/PopupMenuChromium.cpp:
00:34 Changeset [44311] by abarth@webkit.org

WebCore:

2009-06-01 Nate Chapin <japhet@google.com>

Reviewed by Dimitri Glazkov. Landed by Adam Barth.


If a url with an anchor is being loaded, ensure that the anchor remains locked in view until the page
has finished loading compeltely or the user has manually scrolled. Refreshing an anchor url after
scrolling to a new location on the page will still result in jumping to the new location.


https://bugs.webkit.org/show_bug.cgi?id=26034

  • loader/FrameLoader.cpp: (WebCore::FrameLoader::gotoAnchor): Set anchor lock after navigating to anchor. (WebCore::FrameLoader::completed): Release anchor lock.
  • page/FrameView.cpp: (WebCore::FrameView::FrameView): Ensure anchor lock starts unset. (WebCore::FrameView::reset): Ensure anchor lock starts unset. (WebCore::FrameView::layout): If anchor lock is set, force a gotoAnchor() after layout. (WebCore::FrameView::scrollRectIntoViewRecursively): Release anchor lock if a programmatic scroll begins. (WebCore::FrameView::setWasScrolledByUser): Release anchor lock if user manually scrolls. (WebCore::FrameView::setScrollPosition): Release anchor lock if a programmatic scroll begins.
  • page/FrameView.h: (WebCore::FrameView::lockedToAnchor): Added. (WebCore::FrameView::setLockedToAnchor): Added.

LayoutTests:

2009-06-01 Nate Chapin <japhet@google.com>

Reviewed by Dimitri Glazkov. Landed by Adam Barth.

Ensure an anchor remains locked in view as images (without height already specified) load around it.


https://bugs.webkit.org/show_bug.cgi?id=26034

  • fast/dynamic/anchor-lock.html: Added.
  • fast/dynamic/resources/largeblank.png: Added.
Note: See TracTimeline for information about the timeline view.