Timeline
02/17/09:
- 21:48 Changeset [41054] by
-
Windows build fix
- loader/FrameLoader.cpp: (WebCore::toPlugInElement): Don't rely on #if being processed before ASSERT.
- 20:17 Changeset [41053] by
-
2009-02-17 Geoffrey Garen <ggaren@apple.com>
Reviewed by Sam Weinig.
Fixed failure in js1_5/Regress/regress-168347.js, as seen on the Oliver
bot.
Technically, both behaviors are OK, but we might as well keep this test
passing.
- runtime/FunctionPrototype.cpp: (JSC::insertSemicolonIfNeeded): No need to add a trailing semicolon after a trailing '}', since '}' ends a block, indicating the end of a statement.
- 20:03 Changeset [41052] by
-
2009-02-17 Geoffrey Garen <ggaren@apple.com>
Build fix.
- runtime/FunctionPrototype.cpp:
- 19:36 Changeset [41051] by
-
Add assertion to guard against oversized pc relative calls.
Reviewed by Geoff Garen
- 18:15 Changeset [41050] by
-
Versioning.
- 18:14 Changeset [41049] by
-
Disable the JIT for x86_64 until <rdar://problem/6594577> is resolved.
Rubber-stamped by Oliver Hunt.
- wtf/Platform.h:
- 17:20 Changeset [41048] by
-
New tag.
- 16:50 Changeset [41047] by
-
2009-02-17 David Levin <levin@chromium.org>
Reviewed by Alexey Proskuryakov.
Bug 23977: Unnecessary timer related headers in files.
<https://bugs.webkit.org/show_bug.cgi?id=23977>
No observable change in behavior, so no test.
- dom/Document.cpp:
- dom/WorkerRunLoop.cpp:
- dom/WorkerRunLoop.h:
- 16:21 Changeset [41046] by
-
WebCore:
2009-02-17 Peter Abrahamsen <rainhead@gmail.com>
Reviewed by Sam Weinig.
https://bugs.webkit.org/show_bug.cgi?id=23958
<rdar://problem/6587815>
Updated XMLHttpRequest with new header names from the latest Access
Control draft: http://www.w3.org/TR/access-control/
- Access-Control-Origin becomes Access-Control-Allow-Origin
- Access-Control-Credentials becomes Access-Control-Allow-Credentials
- xml/XMLHttpRequest.cpp: (WebCore::XMLHttpRequest::accessControlCheck):
LayoutTests:
2009-02-17 Peter Abrahamsen <rainhead@gmail.com>
Reviewed by Sam Weinig.
https://bugs.webkit.org/show_bug.cgi?id=23958
<rdar://problem/6587815>
Updated XMLHttpRequests tests to match the newest Access Control draft.
- http/tests/security/originHeader/resources/print-origin.cgi:
- http/tests/xmlhttprequest/resources/access-control-basic-allow-access-control-origin-header.cgi:
- http/tests/xmlhttprequest/resources/access-control-basic-allow-print-headers.cgi:
- http/tests/xmlhttprequest/resources/access-control-basic-allow-star.cgi:
- http/tests/xmlhttprequest/resources/access-control-basic-allow.cgi:
- http/tests/xmlhttprequest/resources/access-control-basic-get-fail-non-simple.cgi:
- http/tests/xmlhttprequest/resources/access-control-basic-non-get-allow.cgi:
- http/tests/xmlhttprequest/resources/access-control-basic-preflight-cache-invalidation.php:
- http/tests/xmlhttprequest/resources/access-control-basic-preflight-cache-timeout.php:
- http/tests/xmlhttprequest/resources/access-control-basic-preflight-cache.php:
- http/tests/xmlhttprequest/resources/access-control-basic-whitelist-response-headers.cgi:
- 16:14 Changeset [41045] by
-
2009-02-17 Geoffrey Garen <ggaren@apple.com>
Reviewed by Sam Weinig.
Fixed <rdar://problem/6595040> REGRESSION: http://www.amnestyusa.org/
fails to load.
amnestyusa.org uses the Optimist JavaScript library, which adds event
listeners by concatenating string-ified functions. This is only sure to
be syntactically valid if the string-ified functions end in semicolons.
- parser/Lexer.cpp: (JSC::Lexer::isWhiteSpace):
- parser/Lexer.h: (JSC::Lexer::isWhiteSpace): (JSC::Lexer::isLineTerminator): Added some helper functions for examining whitespace.
- runtime/FunctionPrototype.cpp: (JSC::appendSemicolonIfNeeded): (JSC::functionProtoFuncToString): When string-ifying a function, insert a semicolon in the last non-whitespace position, if one doesn't already exist.
LayoutTests:
2009-02-17 Geoffrey Garen <ggaren@apple.com>
Reviewed by Sam Weinig.
Test for <rdar://problem/6595040> REGRESSION: http://www.amnestyusa.org/
fails to load.
- fast/js/function-toString-semicolon-insertion-expected.txt: Added.
- fast/js/function-toString-semicolon-insertion.html: Added.
- fast/js/resources/function-toString-semicolon-insertion.js: Added. (compileAndSerialize):
- 14:59 Changeset [41044] by
-
wx build fix. Add missing constructor used for empty values.
- 14:16 Changeset [41043] by
-
2009-02-17 Antti Koivisto <antti@apple.com>
Reviewed by Dave Kilzer.
<rdar://problem/6592446> dynamically updating page doesn't seem to draw when updated
Need update after callback.
- bindings/js/JSCustomPositionCallback.cpp: (WebCore::JSCustomPositionCallback::handleEvent):
- bindings/js/JSCustomPositionErrorCallback.cpp: (WebCore::JSCustomPositionErrorCallback::handleEvent):
- 12:43 Changeset [41042] by
-
2009-02-17 Eric Carlson <eric.carlson@apple.com>
Reviewed by Antti Koivisto.
https://bugs.webkit.org/show_bug.cgi?id=23917
Allow a WebKit plug-in to act as a proxy for the <audio> and <video>
element.
- DerivedSources.make: add media element proxy exports to .exp file when feature is defined.
- WebCore.VideoProxy.exp: New, define the informal protocol exported by a media element proxy.
- WebCore.xcodeproj/project.pbxproj: Add MediaPlayerProxy.h.
- html/HTMLMediaElement.cpp: (WebCore::HTMLMediaElement::HTMLMediaElement): Initialize m_needWidgetUpdate. (WebCore::HTMLMediaElement::attributeChanged): Don't detach+attach when PLUGIN_PROXY_FOR_VIDEO, the proxy plug-in handles the poster frame. (WebCore::HTMLMediaElement::rendererIsNeeded): New logic for PLUGIN_PROXY_FOR_VIDEO. (WebCore::HTMLMediaElement::createRenderer): Create RenderPartObject when PLUGIN_PROXY_FOR_VIDEO. (WebCore::HTMLMediaElement::attach): Set m_needWidgetUpdate when PLUGIN_PROXY_FOR_VIDEO (WebCore::HTMLMediaElement::load): Don't reallocate MediaPlayer when PLUGIN_PROXY_FOR_VIDEO, we keep the same plug-in for the life of the element. (WebCore::HTMLMediaElement::mediaPlayerNetworkStateChanged): update m_networkState when media player network state changes to EMPTY, otherwise we can get out of sync with engine. (WebCore::HTMLMediaElement::defaultEventHandler): pass event to widget when PLUGIN_PROXY_FOR_VIDEO (WebCore::HTMLMediaElement::deliverNotification): New, deliver notification from proxy plug-in to media player. (WebCore::HTMLMediaElement::setMediaPlayerProxy): New, pass proxy object to media player. (WebCore::HTMLMediaElement::initialURL): New, return the url from the "src" attr or the appropriate <source> element to be used as the initial url for the proxy. (WebCore::HTMLMediaElement::finishParsingChildren): New, allocate MediaPlayer and update widget.
- html/HTMLMediaElement.h: Declare new methods for proxy, add m_needWidgetUpdate. (WebCore::HTMLMediaElement::setNeedWidgetUpdate):
- html/HTMLVideoElement.cpp: (WebCore::HTMLVideoElement::attach): Poster image is handled by proxy when PLUGIN_PROXY_FOR_VIDEO. (WebCore::HTMLVideoElement::parseMappedAttribute): Ditto.
- html/HTMLVideoElement.h:
- loader/FrameLoader.cpp: (WebCore::toPlugInElement): Allow cast if element is <video> or <audio>
- platform/graphics/MediaPlayer.cpp: (WebCore::MediaPlayer::MediaPlayer): Remove white space. (WebCore::MediaPlayer::setPoster): New, forward call to private player. (WebCore::MediaPlayer::deliverNotification): Ditto. (WebCore::MediaPlayer::setMediaPlayerProxy): Ditto.
- platform/graphics/MediaPlayer.h: (WebCore::MediaPlayer::mediaPlayerClient):
- platform/graphics/mac/MediaPlayerProxy.h: New, defines media player proxy interface.
- rendering/RenderPart.cpp: (WebCore::RenderPart::RenderPart): Change constructor to take Element* instead of Node* as a non-element node doesn't need a renderer
- rendering/RenderPart.h: Ditto.
- rendering/RenderPartObject.cpp: (WebCore::RenderPartObject::RenderPartObject): Ditto. (WebCore::RenderPartObject::updateWidget): Package params for proxy plug-in when element is <video> or <audio>.
- rendering/RenderPartObject.h:
2009-02-17 Eric Carlson <eric.carlson@apple.com>
Reviewed by Antti Koivisto.
https://bugs.webkit.org/show_bug.cgi?id=23917
Allow a WebKit plug-in to act as a proxy for the <audio> and <video>
element.
- Plugins/WebPluginContainerPrivate.h:
- Plugins/WebPluginController.mm: (mediaProxyClient): New, cast to HTMLMediaElement if it is a video or audio element (-[WebPluginController _setMediaPlayerProxy:forElement:]): New, pass proxy to HTMLMediaElement (-[WebPluginController _postMediaPlayerNotification:forElement:]): New, deliver event to HTMLMediaElement
- WebCoreSupport/WebFrameLoaderClient.mm: (WebFrameLoaderClient::createPlugin): Don't allow a media player proxy plug-in to be chosen by file extension, only want a match for the new MIME type proxy plug-ins should have.
- 11:33 Changeset [41041] by
-
WebCore:
2009-02-17 David Hyatt <hyatt@apple.com>
Reviewed by Eric Seidel
Don't allow legends to be anything but display:block.
Added fast/forms/inline-ignored-on-legend.html
- GNUmakefile.am:
- WebCore.pro:
- WebCore.vcproj/WebCore.vcproj:
- WebCore.xcodeproj/project.pbxproj:
- WebCoreSources.bkl:
- css/html4.css:
- html/HTMLLegendElement.cpp:
- html/HTMLLegendElement.h:
- rendering/RenderLegend.cpp: Removed.
- rendering/RenderLegend.h: Removed.
- wml/WMLInsertedLegendElement.cpp:
- wml/WMLInsertedLegendElement.h:
LayoutTests:
2009-02-17 David Hyatt <hyatt@apple.com>
Reviewed by Eric Seidel
Don't allow legends to be anything but display:block.
- fast/forms/inline-ignored-on-legend.html: Added.
- platform/mac/fast/block/basic/fieldset-stretch-to-legend-expected.txt:
- platform/mac/fast/borders/fieldsetBorderRadius-expected.txt:
- platform/mac/fast/forms/006-expected.txt:
- platform/mac/fast/forms/007-expected.txt:
- platform/mac/fast/forms/fieldset-align-expected.txt:
- platform/mac/fast/forms/float-before-fieldset-expected.txt:
- platform/mac/fast/forms/inline-ignored-on-legend-expected.checksum: Added.
- platform/mac/fast/forms/inline-ignored-on-legend-expected.png: Added.
- platform/mac/fast/forms/inline-ignored-on-legend-expected.txt: Added.
- platform/mac/fast/invalid/residual-style-expected.txt:
- platform/mac/svg/custom/inline-svg-in-xhtml-expected.txt:
- 10:14 Changeset [41040] by
-
Reviewed by Simon Fraser.
https://bugs.webkit.org/attachment.cgi?id=27666
Fix Chromium build build: forgotten include in RenderObject, sync
RenderTheme to the recent changes in RenderObject.
- rendering/RenderObject.cpp:
- rendering/RenderThemeChromiumMac.mm: (WebCore::RenderThemeChromiumMac::updatePressedState):
- 08:52 Changeset [41039] by
-
2009-02-17 Xan Lopez <xan@gnome.org>
Rubber-stamped by Alexey Proskuryakov.
Restoring change landed in r40715, which was accidentally undone
by r40918.
- webkit/webkitwebframe.cpp:
- 04:43 Changeset [41038] by
-
2009-02-17 Gustavo Noronha Silva <gns@gnome.org>
Reviewed by Alexey Proskuryakov.
https://bugs.webkit.org/show_bug.cgi?id=23939
Release build being built with debugging symbols
- configure.ac: Revert change done in revision 40790, since we already have a AC_PROG_CXX macro call as part of WEBKIT_INIT
- 04:32 Changeset [41037] by
-
2009-02-17 Gustavo Noronha Silva <gns@gnome.org>
Reviewed by Darin Adler.
https://bugs.webkit.org/show_bug.cgi?id=22966
crash when destroying a webview that opened a page containing <script>
tags
Fix m_group being set to 0 instead of to m_singlePageGroup when
GroupName is set to empty
- page/Page.cpp: (WebCore::Page::setGroupName):
02/16/09:
- 19:38 Changeset [41036] by
-
Roll out r41022 as it breaks qt and gtk builds
- 19:38 Changeset [41035] by
-
2009-02-16 Beth Dakin <bdakin@apple.com>
Reviewed by Sam Weinig.
Fix for <rdar://problem/6386623>
I made m_numParsedProperties and m_maxParsedProperties unsigned
instead of int, and then added an early return from addPropery() if
m_maxParsedProperties exceeds the max value.
- css/CSSParser.cpp: (WebCore::CSSParser::addProperty): (WebCore::CSSParser::rollbackLastProperties): (WebCore::CSSParser::clearProperties): (WebCore::CSSParser::createFontFaceRule): (WebCore::CSSParser::deleteFontFaceOnlyValues):
- css/CSSParser.h:
- 19:36 Changeset [41034] by
-
WebCore:
2009-02-16 Justin Garcia <justin.garcia@apple.com>
Reviewed by Mark Rowe.
https://bugs.webkit.org/show_bug.cgi?id=16309
HTML5: The third execCommand argument for insert{un}orderedlist should be ignored
When we originally implemented this command, MSDN documentation said that IE set the
id attribute of inserted lists to the third argument to execCommand, but IE doesn't
do this nor do any other browsers.
- editing/EditorCommand.cpp: (WebCore::executeInsertOrderedList): (WebCore::executeInsertUnorderedList):
- editing/IndentOutdentCommand.cpp: (WebCore::IndentOutdentCommand::outdentParagraph):
- editing/InsertListCommand.cpp: (WebCore::InsertListCommand::insertList): (WebCore::InsertListCommand::InsertListCommand): (WebCore::InsertListCommand::doApply):
- editing/InsertListCommand.h: (WebCore::InsertListCommand::create):
LayoutTests:
2009-02-16 Justin Garcia <justin.garcia@apple.com>
Reviewed by Mark Rowe.
https://bugs.webkit.org/show_bug.cgi?id=16309
HTML5: The third execCommand argument for insert{un}orderedlist should be ignored
- editing/execCommand/insert-list-with-id-expected.txt: Added.
- editing/execCommand/insert-list-with-id.html:
- platform/mac/editing/execCommand/insert-list-with-id-expected.checksum: Removed.
- platform/mac/editing/execCommand/insert-list-with-id-expected.png: Removed.
- platform/mac/editing/execCommand/insert-list-with-id-expected.txt: Removed.
- 18:59 Changeset [41033] by
-
2009-02-16 Gustavo Noronha Silva <gns@gnome.org>
Reviewed by Adam Roben
Actually use the installation prefix defined in WebKitInstallationPrefix,
in autotools builds.
- Scripts/webkitdirs.pm:
- 18:00 Changeset [41032] by
-
Versioning.
- 17:59 Changeset [41031] by
-
New tag.
- 17:56 Changeset [41030] by
-
Versioning.
- 17:56 Changeset [41029] by
-
Merge r41027.
- 16:40 Changeset [41028] by
-
2009-02-16 Sam Weinig <sam@webkit.org>
Reviewed by Geoffrey Garen.
Fix for <rdar://problem/6468156>
REGRESSION (r36779): Adding link, images, flash in TinyMCE blocks entire page (21382)
No performance regression.
- runtime/Arguments.cpp: (JSC::Arguments::fillArgList): Add codepath for when the "length" property has been overridden.
LayoutTests:
2009-02-16 Sam Weinig <sam@webkit.org>
Reviewed by Geoffrey Garen.
Add tests for <rdar://problem/6468156>
REGRESSION (r36779): Adding link, images, flash in TinyMCE blocks entire page (21382)
- fast/js/function-apply-expected.txt:
- fast/js/resources/function-apply.js: Add cases covering setting arugments.length and Array.length explicitly or implicitly using Array.prototype.unshift.
- 15:56 Changeset [41027] by
-
WebCore:
Reviewed by Geoffrey Garen.
- fix <rdar://problem/6561077> REGRESSION (r39634): Unable to select an item in stocks widget with the mouse
Test: platform/mac/fast/events/pointer-events-dashboard.html
- css/CSSStyleSelector.cpp: (WebCore::CSSStyleSelector::applyProperty): Do not apply the pointer-events property in Dashboard backwards compatibility mode, in order to work around misuse of that property by the Stocks widget.
LayoutTests:
Reviewed by Geoffrey Garen.
- test for <rdar://problem/6561077> REGRESSION (r39634): Unable to select an item in stocks widget with the mouse
- platform/mac/fast/events/pointer-events-dashboard-expected.checksum: Added.
- platform/mac/fast/events/pointer-events-dashboard-expected.png: Added.
- platform/mac/fast/events/pointer-events-dashboard-expected.txt: Added.
- platform/mac/fast/events/pointer-events-dashboard.html: Copied from LayoutTests/fast/events/pointer-events.html.
- 15:33 Changeset [41026] by
-
WebCore:
2009-02-16 Justin Garcia <justin.garcia@apple.com>
Reviewed by Eric Seidel.
https://bugs.webkit.org/show_bug.cgi?id=15381
execCommand justify modifies the contentEditable node
- editing/CompositeEditCommand.cpp: (WebCore::CompositeEditCommand::moveParagraphContentsToNewBlockIfNecessary): Avoid modifying not just the body element, but any root editable element.
LayoutTests:
2009-02-16 Justin Garcia <justin.garcia@apple.com>
Reviewed by Eric Seidel.
https://bugs.webkit.org/show_bug.cgi?id=15381
execCommand justify modifies the contentEditable node
- editing/execCommand/15381-expected.txt: Added.
- editing/execCommand/15381.html: Added.
Made this test cross platform and changed it to reflect the fix:
- editing/execCommand/5062376-expected.txt: Added.
- editing/execCommand/5062376.html:
- platform/mac/editing/execCommand/5062376-expected.checksum: Removed.
- platform/mac/editing/execCommand/5062376-expected.png: Removed.
- platform/mac/editing/execCommand/5062376-expected.txt: Removed.
- 15:19 Changeset [41025] by
-
Bug 23979: AX: alt tag not returned for <input type="image">
https://bugs.webkit.org/show_bug.cgi?id=23979
- 14:36 Changeset [41024] by
-
Merge r41023.
- 14:23 Changeset [41023] by
-
Build fix.
- 14:15 Changeset [41022] by
-
Bug 23787: Allow JIT to generate SSE2 code if using GCC
<https://bugs.webkit.org/show_bug.cgi?id=23787>
Provided by Csaba Osztrogonac
Reviewed by Oliver Hunt
GCC version of the cpuid check.
- 14:06 Changeset [41021] by
-
Reviewed by Kevin Ollivier.
The ref data is not properly set unless we use the CreateMatrix function.
https://bugs.webkit.org/show_bug.cgi?id=23978
- 13:57 Changeset [41020] by
-
Reviewed by John Sullivan.
- fix <rdar://problem/6516829> FontPlatformData constructor passes NULL to CTFontCopyGraphicsFont and CTFontGetPlatformFont
- platform/graphics/FontCache.cpp: (WebCore::FontDataCacheKeyTraits::emptyValue): Changed to use the FontPlatformData(float, bool, bool) constructor.
- platform/graphics/mac/FontCacheMac.mm: (WebCore::FontCache::createFontPlatformData): Changed to pass a valid NSFont to the FontPlatformData constructor, instead of implicitly passing 0 and then calling setFont().
- platform/graphics/mac/FontPlatformData.h: Made the NSFont parameter of the constructor mandatory.
- 02:57 Changeset [41019] by
-
2009-02-16 Simon Hausmann <simon.hausmann@nokia.com>
Reviewed by Tor Arne Vestbø.
Avoid loading plugins when they're disabled in the settings.
02/15/09:
- 13:39 Changeset [41018] by
-
Move duplicate caretWidth constants to RenderObject.h
Reviewed by Dan Bernstein.
No test since no change in behavior.
- rendering/RenderBlock.cpp: (WebCore::RenderBlock::localCaretRect): Removed caretWidth.
- rendering/RenderBox.cpp: (WebCore::RenderBox::localCaretRect): Ditto.
- rendering/RenderObject.h: Added caretWidth definition.
- rendering/RenderText.cpp: (WebCore::RenderText::localCaretRect): Removed caretWidth.
02/14/09:
- 08:39 Changeset [41017] by
-
2009-02-14 Dimitri Glazkov <dglazkov@chromium.org>
Reviewed by Eric Seidel.
https://bugs.webkit.org/show_bug.cgi?id=23955
V8HTMLPlugInCustom returns undefined, which fools interceptors.
- bindings/v8/custom/V8HTMLPlugInElementCustom.cpp: (WebCore::NAMED_PROPERTY_GETTER): Use deferToInterceptor() return value. (WebCore::NAMED_PROPERTY_SETTER): Ditto. (WebCore::INDEXED_PROPERTY_GETTER): Ditto. (WebCore::INDEXED_PROPERTY_SETTER): Ditto.
02/13/09:
- 18:17 Changeset [41016] by
-
2009-02-13 Dimitri Glazkov <dglazkov@chromium.org>
Reviewed by Eric Seidel.
https://bugs.webkit.org/show_bug.cgi?id=23954
Add V8 custom bindings for CanvasRenderingContext2D.
- bindings/v8/custom/V8CanvasRenderingContext2DCustom.cpp: Added. (WebCore::toV8): (WebCore::toCanvasStyle): (WebCore::ACCESSOR_GETTER): (WebCore::ACCESSOR_SETTER):
- 18:05 Changeset [41015] by
-
Versioning.
- 18:05 Changeset [41014] by
-
New tag.
- 17:48 Changeset [41013] by
-
2009-02-13 Adam Treat <adam.treat@torchmobile.com>
Reviewed by George Staikos.
Don't depend on 'initializeThreading()' to come before a call to 'isMainThread()'
as QtWebKit only calls 'initializeThreading()' during QWebPage construction.
A client app may well make a call to QWebSettings::iconForUrl() for instance
before creating a QWebPage and that call to QWebSettings triggers an
ASSERT(isMainThread()) deep within WebCore.
- wtf/ThreadingQt.cpp: (WTF::isMainThread):
- 17:19 Changeset [41012] by
-
Merge r40975.
- 17:19 Changeset [41011] by
-
Merge r40945.
- 17:08 Changeset [41010] by
-
Merge r41000.
- 17:08 Changeset [41009] by
-
Merge r40995.
- 17:08 Changeset [41008] by
-
Merge r40993.
- 17:08 Changeset [41007] by
-
Merge r40968.
- 17:08 Changeset [41006] by
-
Merge r40967.
- 17:08 Changeset [41005] by
-
Merge r40937.
- 17:08 Changeset [41004] by
-
Merge r40935.
- 17:08 Changeset [41003] by
-
Merge r40932.
- 17:08 Changeset [41002] by
-
Merge r40940.
- 17:08 Changeset [41001] by
-
Merge r40920.
- 17:00 Changeset [41000] by
-
2009-02-13 Gavin Barraclough <barraclough@apple.com>
Reviewed by Darin Adler.
Some data in the instruction stream is potentially uninitialized - fix this.
Change the OperandTypes constructor so that uninitialized memory in the int
is zeroed, and modify the Instruction constructor taking an Opcode so that
if !HAVE(COMPUTED_GOTO) (i.e. when Opcode is an enum, and is potentially only
a byte) it zeros the Instruction first before writing the opcode.
- bytecode/Instruction.h: (JSC::Instruction::Instruction):
- parser/ResultType.h: (JSC::OperandTypes::OperandTypes):
- 16:46 Changeset [40999] by
-
Fix failing layout pixel test so bdash doesn't come to my house in the middle of the night and beat me up.
- 16:41 Changeset [40998] by
-
Fix ChangeLog.
- 16:35 Changeset [40997] by
-
2009-02-13 Anders Carlsson <andersca@apple.com>
Reviewed by Darin Adler.
<rdar://problem/6583187>
- page/NavigatorBase.cpp: Handle PPC64 and X86_64 as well.
- 16:04 Changeset [40996] by
-
2009-02-13 Anders Carlsson <andersca@apple.com>
Reviewed by Kevin Decker.
<rdar://problem/6584834> ESPN radio live stream link hangs Safari
When a plug-in invokes JavaScript code that will destroy the plug-in, we need to
defer destruction until we're done executing the script.
- Plugins/Hosted/NetscapePluginHostProxy.mm: (WebKit::PluginDestroyDeferrer::PluginDestroyDeferrer): (WebKit::PluginDestroyDeferrer::~PluginDestroyDeferrer): Add a simple RAII object for deferring destruction of the plug-in instance.
(WKPCEvaluate):
(WKPCInvoke):
(WKPCInvokeDefault):
(WKPCConstruct):
(WKPCGetProperty):
(WKPCSetProperty):
(WKPCRemoveProperty):
(WKPCHasProperty):
(WKPCHasMethod):
Use the PluginDestroyDeferrer.
- Plugins/Hosted/NetscapePluginInstanceProxy.h: (WebKit::NetscapePluginInstanceProxy::pluginID): Assert that the plug-in ID is not 0 here.
- Plugins/Hosted/NetscapePluginInstanceProxy.mm: (WebKit::NetscapePluginInstanceProxy::NetscapePluginInstanceProxy): Initialize the call depth.
(WebKit::NetscapePluginInstanceProxy::~NetscapePluginInstanceProxy):
Set the plug-in ID to 0 to aid debugging.
(WebKit::NetscapePluginInstanceProxy::willCallPluginFunction):
Increment the call depth.
(WebKit::NetscapePluginInstanceProxy::didCallPluginFunction):
Decrement the call depth, if it's 0 and we should stop the plug-in, do so.
(WebKit::NetscapePluginInstanceProxy::shouldStop):
If we're called this with a non-zero call depth, set shouldStopSoon to true.
- Plugins/Hosted/WebHostedNetscapePluginView.mm: (-[WebHostedNetscapePluginView shouldStop]): Call the proxy.
- 16:01 Changeset [40995] by
-
2009-02-13 Geoffrey Garen <ggaren@apple.com>
Build fix for non_JIT platforms.
- bytecode/CodeBlock.h: (JSC::CodeBlock::setIsNumericCompareFunction): (JSC::CodeBlock::isNumericCompareFunction):
- 15:56 Changeset [40994] by
-
2009-02-13 Darin Adler <darin@apple.com>
Reviewed by Mark Rowe.
Fix broken tokenizer regression test that reflected a lifetime bug
in the document.write optimization. The test failure was trivial to
reproduce in COLLECT_ON_EVERY_ALLOCATION mode.
- bindings/js/JSHTMLDocumentCustom.cpp: (WebCore::documentWrite): Convert strings to String rather than passing the pointer and length to SegmentedString. The optimization is thus mostly gone. However, there are two ways to bring it back: 1) Apply the patch that makes UString and String share the same buffers. 2) Add a UString feature to SegmentedString; simple to do but might risk slowing down normal document parsing.
- 15:28 Changeset [40993] by
-
2009-02-13 Geoffrey Garen <ggaren@apple.com>
Reviewed by Darin Adler.
Fixed <rdar://problem/6584057> Optimize sort by JS numeric comparison
function not to run the comparison function
- bytecode/CodeBlock.cpp: (JSC::CodeBlock::CodeBlock):
- bytecode/CodeBlock.h: (JSC::CodeBlock::setIsNumericCompareFunction): (JSC::CodeBlock::isNumericCompareFunction): Added the ability to track whether a CodeBlock performs a sort-like numeric comparison.
- bytecompiler/BytecodeGenerator.cpp: (JSC::BytecodeGenerator::generate): Set the isNumericCompareFunction bit after compiling.
- parser/Nodes.cpp: (JSC::FunctionBodyNode::emitBytecode): Fixed a bug that caused us to codegen an extra return at the end of all functions (eek!), since this made it harder / weirder to detect the numeric comparison pattern in bytecode.
- runtime/ArrayPrototype.cpp: (JSC::arrayProtoFuncSort): Use the isNumericCompareFunction bit to do a faster sort if we can.
- runtime/FunctionConstructor.cpp: (JSC::extractFunctionBody): (JSC::constructFunction):
- runtime/FunctionConstructor.h: Renamed and exported extractFunctionBody for use in initializing lazyNumericCompareFunction.
- runtime/JSArray.cpp: (JSC::compareNumbersForQSort): (JSC::compareByStringPairForQSort): (JSC::JSArray::sortNumeric): (JSC::JSArray::sort):
- runtime/JSArray.h: Added a fast numeric sort. Renamed ArrayQSortPair to be more specific since we do different kinds of qsort now.
- runtime/JSGlobalData.cpp: (JSC::JSGlobalData::JSGlobalData): (JSC::JSGlobalData::numericCompareFunction): (JSC::JSGlobalData::ClientData::~ClientData):
- runtime/JSGlobalData.h: Added helper data for computing the isNumericCompareFunction bit.
LayoutTests:
2009-02-13 Geoffrey Garen <ggaren@apple.com>
Reviewed by Sam Weinig.
Added a test for an edge case in <rdar://problem/6584057>.
- fast/js/resources/sort-non-numbers.js: Added.
- fast/js/sort-non-numbers.html: Added.
- fast/js/sort-non-numbers-expected.txt: Added.
- 14:49 Changeset [40992] by
-
2009-02-13 Adam Treat <adam.treat@torchmobile.com>
Caught by Darin Adler.
No reason to check m_tokenizer twice for 0 in a row.
- dom/Document.cpp: (WebCore::Document::write):
- 13:49 Changeset [40991] by
-
2009-02-13 Lyon Chen <lyon.chen@torchmobile.com>
Reviewed by Dave Hyatt.
Take into account y() + height() when calculating getLowerRightCorner.
- dom/ContainerNode.cpp: (WebCore::ContainerNode::getLowerRightCorner):
- 13:17 Changeset [40990] by
-
2009-02-13 Antti Koivisto <antti@apple.com>
Reviewed by Dave Hyatt.
Most of this code was in RenderBox, restore the copyright lines.
- rendering/RenderBoxModelObject.cpp:
- rendering/RenderBoxModelObject.h:
- 13:07 Changeset [40989] by
-
2009-02-13 David Hyatt <hyatt@apple.com>
Get rid of printBoxDecorations, since it is dead code.
Reviewed by Dan Bernstein
- rendering/RenderObject.h:
- 13:01 Changeset [40988] by
-
2009-02-13 Adam Treat <adam.treat@torchmobile.com>
Reviewed by Darin Adler.
https://bugs.webkit.org/show_bug.cgi?id=23952
Add Document::write overload which takes a SegmentedString since this
is what the tokenizer expects anyway. Modified JSHTMLDocument so it once
again calls Document::write instead of injecting the string directly into
the tokenizer. This ensures that all document.write's are funneled through
one method again, but should still be just as fast for JSHTMLDocumentCustom.
- bindings/js/JSHTMLDocumentCustom.cpp: (WebCore::documentWrite):
- dom/Document.cpp: (WebCore::Document::write):
- dom/Document.h:
- 12:38 Changeset [40987] by
-
2009-02-13 David Hyatt <hyatt@apple.com>
Reviewed by Simon Fraser
Remove updateWidgetPosition on RenderObject. Change RenderView's set of widgets
to have a tighter type: RenderWidget instead of RenderObject. Devirtualize
updateWidgetPosition, since all RenderWidgets (except for applets) are now
RenderParts also.
- rendering/RenderObject.cpp:
- rendering/RenderObject.h:
- rendering/RenderPart.cpp:
- rendering/RenderPart.h:
- rendering/RenderView.cpp: (WebCore::RenderView::updateWidgetPositions): (WebCore::RenderView::addWidget): (WebCore::RenderView::removeWidget):
- rendering/RenderView.h:
- rendering/RenderWidget.cpp: (WebCore::RenderWidget::updateWidgetPosition):
- rendering/RenderWidget.h:
- 12:11 Changeset [40986] by
-
2009-02-13 Darin Adler <darin@apple.com>
- StringsNotToBeLocalized.txt: Updated for recent changes.
- 12:04 Changeset [40985] by
-
2009-02-13 David Hyatt <hyatt@apple.com>
Reviewed by Eric Seidel
Remove isEditable() on RenderObject. It is dead code.
- rendering/RenderObject.cpp:
- rendering/RenderObject.h:
- 11:49 Changeset [40984] by
-
2009-02-13 David Hyatt <hyatt@apple.com>
Reviewed by Eric Seidel
Add isTextControl() to RenderObject. Patch call sites thats asked isTextField() isTextArea() to just use isTextControl() instead.
Add a toRenderTextControl converter for doing checked casting to RenderTextControls.
- dom/Document.cpp: (WebCore::Document::setFocusedNode):
- dom/InputElement.cpp: (WebCore::InputElement::updateSelectionRange):
- editing/ReplaceSelectionCommand.cpp: (WebCore::ReplacementFragment::ReplacementFragment):
- editing/TextIterator.cpp: (WebCore::TextIterator::handleReplacedElement):
- html/HTMLInputElement.cpp: (WebCore::HTMLInputElement::selectionStart): (WebCore::HTMLInputElement::selectionEnd): (WebCore::HTMLInputElement::setSelectionStart): (WebCore::HTMLInputElement::setSelectionEnd): (WebCore::HTMLInputElement::select): (WebCore::HTMLInputElement::defaultEventHandler): (WebCore::HTMLInputElement::selection):
- html/HTMLTextAreaElement.cpp: (WebCore::HTMLTextAreaElement::selectionStart): (WebCore::HTMLTextAreaElement::selectionEnd): (WebCore::HTMLTextAreaElement::setSelectionStart): (WebCore::HTMLTextAreaElement::setSelectionEnd): (WebCore::HTMLTextAreaElement::select): (WebCore::HTMLTextAreaElement::setSelectionRange): (WebCore::HTMLTextAreaElement::appendFormData): (WebCore::HTMLTextAreaElement::updateValue): (WebCore::HTMLTextAreaElement::selection):
- page/AccessibilityRenderObject.cpp: (WebCore::AccessibilityRenderObject::isNativeTextControl): (WebCore::AccessibilityRenderObject::text): (WebCore::AccessibilityRenderObject::selectedText): (WebCore::AccessibilityRenderObject::selectedTextRange): (WebCore::AccessibilityRenderObject::setSelectedTextRange): (WebCore::AccessibilityRenderObject::visiblePositionForIndex): (WebCore::AccessibilityRenderObject::indexForVisiblePosition): (WebCore::AccessibilityRenderObject::observableObject):
- page/Frame.cpp: (WebCore::Frame::notifyRendererOfSelectionChange):
- rendering/RenderObject.h: (WebCore::RenderObject::isTextControl): (WebCore::objectIsRelayoutBoundary):
- rendering/RenderTextControl.h: (WebCore::RenderTextControl::isTextControl): (WebCore::toRenderTextControl):
- rendering/TextControlInnerElements.cpp: (WebCore::TextControlInnerTextElement::defaultEventHandler):
- wml/WMLInputElement.cpp: (WebCore::WMLInputElement::select): (WebCore::WMLInputElement::defaultEventHandler): (WebCore::WMLInputElement::isConformedToInputMask):
- 11:18 Changeset [40983] by
-
2009-02-13 David Hyatt <hyatt@apple.com>
Get rid of isEdited/setEdited on RenderObject. Devirtualize isEdited/setEdited on RenderTextControl.
Callers were already asking if the renderer was a text field (or a text area), so it was especially
pointless to waste a second virtual function call after making one just to determine the object type.
Reviewed by Simon Fraser
- dom/Document.cpp: (WebCore::Document::setFocusedNode):
- html/HTMLInputElement.cpp: (WebCore::HTMLInputElement::defaultEventHandler):
- rendering/RenderObject.h:
- rendering/RenderTextControl.h: (WebCore::RenderTextControl::isEdited): (WebCore::RenderTextControl::setEdited):
- 11:04 Changeset [40982] by
-
2009-02-13 Scott Violet <sky@google.com>
Reviewed by Eric Seidel.
https://bugs.webkit.org/show_bug.cgi?id=23934
Skia platform doesn't render text decoration shadows.
Makes Skia render text decoration shadows correctly. We weren't
preparing the SkPaint correctly and didn't have a couple of checks
CG has. Additionally makes the fillColor/strokeColor methods
consistent.
This behavior is covered by existing layout tests (see bug for list).
- platform/graphics/chromium/FontChromiumWin.cpp: (WebCore::Font::drawGlyphs): (WebCore::Font::drawComplexText):
- platform/graphics/skia/GraphicsContextSkia.cpp: (WebCore::GraphicsContext::drawLineForText):
- platform/graphics/skia/PlatformContextSkia.cpp: (PlatformContextSkia::effectiveFillColor): (PlatformContextSkia::effectiveStrokeColor):
- platform/graphics/skia/PlatformContextSkia.h:
- 11:01 Changeset [40981] by
-
2009-02-12 Darin Fisher <darin@chromium.org>
Reviewed by Eric Seidel.
https://bugs.webkit.org/show_bug.cgi?id=23944
KURLGoogle's pre-parsed canonicalSpec constructor should take a CString
- platform/KURL.h:
- platform/KURLGoogle.cpp: (WebCore::KURLGooglePrivate::setUtf8): (WebCore::KURLGooglePrivate::setAscii): (WebCore::KURLGooglePrivate::init): (WebCore::KURLGooglePrivate::replaceComponents): (WebCore::KURL::KURL):
- platform/KURLGooglePrivate.h:
- 10:17 Changeset [40980] by
-
2009-02-13 David Hyatt <hyatt@apple.com>
Remove leftmost/rightmost/lowestPosition from RenderObject. They only need to be called on boxes.
Reviewed by Dan Bernstein
- rendering/RenderBlock.cpp: (WebCore::RenderBlock::lowestPosition): (WebCore::RenderBlock::rightmostPosition): (WebCore::RenderBlock::leftmostPosition):
- rendering/RenderLayer.cpp: (WebCore::RenderLayer::computeScrollDimensions):
- rendering/RenderMedia.cpp: (WebCore::RenderMedia::lowestPosition): (WebCore::RenderMedia::rightmostPosition): (WebCore::RenderMedia::leftmostPosition):
- rendering/RenderObject.h:
- rendering/RenderTableSection.cpp: (WebCore::RenderTableSection::lowestPosition): (WebCore::RenderTableSection::rightmostPosition): (WebCore::RenderTableSection::leftmostPosition):
- 09:49 Changeset [40979] by
-
2009-02-12 Simon Fraser <simon.fraser@apple.com>
Fix build with ACCELERATED_COMPOSITING turned on.
toRenderBox()->x() asserts for a layer on an inline; use
m_owningLayer->renderBoxX() instead.
- rendering/RenderLayerBacking.cpp: (WebCore::RenderLayerBacking::paintIntoLayer):
- 08:24 Changeset [40978] by
-
2009-02-13 Darin Adler <darin@apple.com>
- Configurations/JavaScriptCore.xcconfig: Undo accidental commit of this file.
- 08:21 Changeset [40977] by
-
Reviewed by Adam Roben.
Fixed 3 more test cases
- platform/win/fast/css/computed-style-expected.txt:
- platform/win/fast/css/computed-style-without-renderer-expected.txt:
- svg/css/getComputedStyle-basic-expected.txt:
- 08:20 Changeset [40976] by
-
2009-02-12 Darin Adler <darin@apple.com>
Reviewed by Oliver Hunt.
Speed up document.write a bit.
- bindings/js/JSHTMLDocumentCustom.cpp: (WebCore::documentWrite): Added. Uses SegmentedString to avoid conversion from UString to String and to avoid appending strings. Also added code to handle newlines efficiently. (WebCore::JSHTMLDocument::write): Changed to use documentWrite. (WebCore::JSHTMLDocument::writeln): Ditto.
- dom/Document.cpp: (WebCore::Document::prepareToWrite): Added. Refactored the initialization part of write into a separate function for use in the JavaScript binding. (WebCore::Document::write): Changed to call prepareToWrite.
- dom/Document.h: Added declaration for prepareToWrite.
- 08:16 Changeset [40975] by
-
2009-02-12 Darin Adler <darin@apple.com>
Reviewed by Oliver Hunt and Alexey Proskuryakov.
Speed up a couple string functions.
- runtime/StringPrototype.cpp: (JSC::stringProtoFuncIndexOf): Added a fast path for cases where the second argument is either missing or an integer. (JSC::stringProtoFuncBig): Use jsNontrivialString since the string is guaranteed to be 2 or more characters long. (JSC::stringProtoFuncSmall): Ditto. (JSC::stringProtoFuncBlink): Ditto. (JSC::stringProtoFuncBold): Ditto. (JSC::stringProtoFuncItalics): Ditto. (JSC::stringProtoFuncStrike): Ditto. (JSC::stringProtoFuncSub): Ditto. (JSC::stringProtoFuncSup): Ditto. (JSC::stringProtoFuncFontcolor): Ditto. (JSC::stringProtoFuncFontsize): Make the fast path Sam recently added even faster by avoiding all but the minimum memory allocation. (JSC::stringProtoFuncAnchor): Use jsNontrivialString. (JSC::stringProtoFuncLink): Added a fast path.
- runtime/UString.cpp: (JSC::UString::find): Added a fast path for single-character search strings.
- 07:53 Changeset [40974] by
-
Reviewed by Simon Fraser.
Fixed 2 more test cases (mac specific)
- 06:45 Changeset [40973] by
-
2009-02-13 Benjamin C Meyer <benjamin.meyer@torchmobile.com>
Reviewed by Nikolas Zimmermann.
https://bugs.webkit.org/show_bug.cgi?id=23738
Expose the url elements target frame string. This is the sister function
to the existing linkTargetFrame which returns the QWebFrame*. When the
linkTargetFrame is 0 it is useful to know what the target was to be.
- Api/qwebframe.cpp: (QWebHitTestResultPrivate::QWebHitTestResultPrivate): (QWebHitTestResult::linkTarget):
- Api/qwebframe.h:
- Api/qwebframe_p.h:
- tests/qwebframe/tst_qwebframe.cpp:
- 05:19 ApplicationsGtk edited by
- Update the correct URL for "Evolution RSS Reader Plugin" (diff)
- 04:02 Changeset [40972] by
-
2009-02-13 David Boddie <dboddie@trolltech.com>
Reviewed by Simon Hausmann.
Removed documentation for internal functions.
- 04:02 Changeset [40971] by
-
2009-02-13 Martin Smith <msmith@trolltech.com>
Reviewed by Simon Hausmann.
Corrected some minor qdoc errors.
- 04:01 Changeset [40970] by
-
2009-02-13 Prasanth Ullattil <pullatti@trolltech.com>
Reviewed by Simon Hausmann.
Fixes crash in the Qt port on Windows with comboboxes
The focusWidget in the input context is not reset properly.
The QApplication::setFocusWidget() wont reset the IC if the reason is
PopupFocusReason, this is not ideal. For the time being we are going
to do this from webkit itself.
- 01:29 Changeset [40969] by
-
2009-02-13 David Levin <levin@chromium.org>
Reviewed by Darin Adler.
Bug 23926: Race condition in callOnMainThreadAndWait
<https://bugs.webkit.org/show_bug.cgi?id=23926>
- wtf/MainThread.cpp: Removed callOnMainThreadAndWait since it isn't used.
- 01:18 Changeset [40968] by
-
Math.random is really slow on windows.
Reviewed by Jon Honeycutt.
Math.random calls WTF::randomNumber which is implemented as
the secure rand_s on windows. Unfortunately rand_s is an order
of magnitude slower than arc4random. For this reason I've
added "weakRandomNumber" for use by JavaScript's Math Object.
In the long term we should look at using our own secure PRNG
in place of the system, but this will do for now.
30% win on SunSpider on Windows, resolving most of the remaining
disparity vs. Mac.
