Timeline


and

11/22/07:

22:43 Changeset [27983] by ap@webkit.org

Reviewed by Maciej.

http://bugs.webkit.org/show_bug.cgi?id=14977
Hixie's DOM Core performance test shows insert >10x slower than append

Each invocation of Element.childNodes[n] was creating and registering a new ChildNodeList,
which persisted and listened to notifications until GC.

A fix is to avoid registering child node lists for notifications - they don't need them, as
they share a common cache in Node.

  • dom/Node.cpp: (WebCore::Node::registerNodeList): (WebCore::Node::unregisterNodeList): (WebCore::Node::notifyLocalNodeListsAttributeChanged): (WebCore::Node::notifyLocalNodeListsChildrenChanged):
  • dom/NodeList.h: (WebCore::NodeList::needsNotifications):
22:04 Changeset [27982] by mitz@apple.com

WebCore:

Reviewed by Antti Koivisto.

Defer plug-in loading until after attach and recalcStyle using the
post-attach callback mechanism. Netscape plug-ins are still loaded only
after layout.

  • dom/ContainerNode.cpp: Made NodeCallbackQueue elements retain the Node because callbacks might delete nodes that are in the callback queue. (WebCore::ContainerNode::suspendPostAttachCallbacks): Added. Needed to prevent post-attach callbacks from being dispatched under recalcStyle(). (WebCore::ContainerNode::resumePostAttachCallbacks): Ditto. (WebCore::ContainerNode::dispatchPostAttachCallbacks): Factored out from attach(). (WebCore::ContainerNode::attach):
  • dom/ContainerNode.h:
  • dom/Document.cpp: (WebCore::Document::recalcStyle): Added calls to suspendPostAttachCallbacks() and resumePostAttachCallbacks().
  • html/HTMLEmbedElement.cpp: (WebCore::HTMLEmbedElement::HTMLEmbedElement): (WebCore::HTMLEmbedElement::attach): Changed to queue the widget update for post-attach. (WebCore::HTMLEmbedElement::updateWidget): Added. Called by the post-attach callback.
  • html/HTMLEmbedElement.h: Added an m_needWidgetUpdate member needed to prevent a double update if another plug-in's post-attach updateWidget() triggers a layout which updates the widget before this plug-in's post-attach callback is invoked. (WebCore::HTMLEmbedElement::setNeedWidgetUpdate): Added a setter for m_needWidgetUpdate.
  • html/HTMLObjectElement.cpp: (WebCore::HTMLObjectElement::attach): Changed to queue the widget update for post-attach. (WebCore::HTMLObjectElement::updateWidget): Added. Called by the post-attach callback.
  • html/HTMLObjectElement.h: (WebCore::HTMLObjectElement::setNeedWidgetUpdate): Added a setter for m_needWidgetUpdate.
  • html/HTMLPlugInElement.cpp: (WebCore::HTMLPlugInElement::updateWidgetCallback): Added.
  • html/HTMLPlugInElement.h: (WebCore::HTMLPlugInElement::updateWidget):
  • rendering/RenderPartObject.cpp: (WebCore::RenderPartObject::updateWidget): Added calls to setNeedWidgetUpdate(false) so that if this method is called from FrameView::layout() during post-attach dispatch of another plug-in, it will not be called again when this plug-in's post-attach callback is dispatched.
  • rendering/RenderPartObject.h: (WebCore::RenderPartObject::updateWidget) Renamed argument to match the method definition.

WebKit/mac:

Reviewed by Antti Koivisto.

20:45 Changeset [27981] by timothy@apple.com

Reviewed by Dan Bernstein.

Fix the Element.hasStyleClass and Element.removeStyleClass helpers
to not find and replace substrings, but whole class names at the
beginning or end of the string or surrounded by whitespace.

  • page/inspector/utilities.js:
20:23 Changeset [27980] by timothy@apple.com

Fix typo.

20:20 Changeset [27979] by timothy@apple.com

Revert part of my r27935 change that made the Tip balloons
animate with CSS transitions. The balloon was only opacity 0,
so it would cause it's row to highlight when hovering over
another row.

  • page/inspector/NetworkPanel.js:
  • page/inspector/inspector.css:
20:11 Changeset [27978] by mitz@apple.com

WebCore:

Reviewed by Maciej Stachowiak.

Test: fast/transforms/matrix-01.html

Parse all six matrix entries as numbers.

  • css/CSSParser.cpp: (WebCore::TransformOperationInfo::TransformOperationInfo): (WebCore::CSSParser::parseTransform):
  • css/CSSStyleSelector.cpp: (WebCore::CSSStyleSelector::applyProperty):
  • rendering/RenderStyle.cpp: (WebCore::MatrixTransformOperation::blend):
  • rendering/RenderStyle.h: (WebCore::MatrixTransformOperation::MatrixTransformOperation): (WebCore::MatrixTransformOperation::apply):

LayoutTests:

Reviewed by Maciej Stachowiak.

  • fast/transforms/matrix-01.html: Added.
  • platform/mac/fast/transforms/matrix-01-expected.checksum: Added.
  • platform/mac/fast/transforms/matrix-01-expected.png: Added.
  • platform/mac/fast/transforms/matrix-01-expected.txt: Added.
18:53 Changeset [27977] by alp@webkit.org

2007-11-22 Alp Toker <alp@atoker.com>

Whitespace fixes only.

  • Api/webkitgtkframe.cpp:
  • Api/webkitgtkpage.cpp:
  • Api/webkitgtkpage.h:
  • Api/webkitgtkprivate.h:
  • WebCoreSupport/ChromeClientGtk.cpp:
  • WebCoreSupport/ContextMenuClientGtk.cpp:
  • WebCoreSupport/ContextMenuClientGtk.h:
  • WebCoreSupport/DragClientGtk.h:
  • WebCoreSupport/EditorClientGtk.h:
  • WebCoreSupport/FrameLoaderClientGtk.cpp: (WebKit::FrameLoaderClient::dispatchWillSendRequest): (WebKit::FrameLoaderClient::assignIdentifierToInitialRequest): (WebKit::FrameLoaderClient::postProgressEstimateChangedNotification): (WebKit::FrameLoaderClient::createFrame): (WebKit::FrameLoaderClient::createJavaAppletWidget): (WebKit::FrameLoaderClient::registerForIconNotification): (WebKit::FrameLoaderClient::setMainFrameDocumentReady): (WebKit::FrameLoaderClient::dispatchDidFinishLoad): (WebKit::FrameLoaderClient::frameLoadCompleted): (WebKit::FrameLoaderClient::saveViewStateToItem): (WebKit::FrameLoaderClient::restoreViewState): (WebKit::FrameLoaderClient::shouldGoToHistoryItem): (WebKit::FrameLoaderClient::setTitle):
  • WebCoreSupport/FrameLoaderClientGtk.h:
17:52 Changeset [27976] by alp@webkit.org

2007-11-22 Michael Natterer <mitch@imendio.com>

Reviewed by Alp Toker.

http://bugs.webkit.org/show_bug.cgi?id=15984
Implement "navigation-requested" signal for WebKit Gtk

  • Api/webkitgtknetworkrequest.h: fix parent class member.
  • Api/webkitgtknetworkrequest.cpp
  • Api/webkitgtkprivate.h: basic implementation featuring an "url" member and API.
  • Api/webkitgtkdefines.h: added network request typedefs.
  • Api/webkitgtkpage.h: fix enum name: s/WEBKIT_NAVIGATION_REQUEST_RESPONSE/WebKitNavigationRequestResponse/ and sanitized enum values.
  • Api/webkitgtkpage.cpp: made "navigation-requested" a signal.
  • Api/webkitgtk-marshal.list: added INT:OBJECT,OBJECT
  • WebCoreSupport/FrameLoaderClientGtk.cpp: emit the new signal in dispatchDecidePolicyForNavigationAction().
17:48 Changeset [27975] by mrowe@apple.com

2007-11-22 Mark Rowe <mrowe@apple.com>

Reviewed by Alp Toker.

Fix build-webkit to propagate make's exit status if it fails.

  • Scripts/webkitdirs.pm:
17:03 Changeset [27974] by alp@webkit.org

2007-11-22 Alp Toker <alp@atoker.com>

GTK+ build fix. Pass a correct LIBDIR value.

  • WebKit.pri:
16:04 Changeset [27973] by alp@webkit.org

2007-11-22 Laszlo Gombos <laszlo.gombos@gmail.com>

Reviewed by Alp Toker.

Fix some compilation warnings. Do not reference undefined (platform) specific type (WebCore::CachedFont::m_fontData)

  • loader/CachedFont.cpp: (WebCore::CachedFont::~CachedFont): (WebCore::CachedFont::allReferencesRemoved):
13:02 Changeset [27972] by alp@webkit.org

2007-11-22 Alp Toker <alp@atoker.com>

Reviewed by Mark Rowe.

GTK+ drawing fix.

Add a check to avoid crashing when the GraphicsContext is not
associated with a GdkEventExpose. This was noticed when adding
printing support but might be triggered in other situations too.

  • platform/gtk/WidgetGtk.cpp: (WebCore::Widget::paint):
09:36 Changeset [27971] by mitz@apple.com

Reviewed by Mark Rowe.

  • fix crash when running pixel tests
  • DumpRenderTree/mac/DumpRenderTree.mm: (dumpRenderTree): Parse the command line options before setting up the environment so that we know if we need to set up the pixel dump machinery.
09:02 QtWebKitContrib edited by treat@kde.org
(diff)
05:04 QtWebKitTodo edited by treat@kde.org
(diff)
05:03 QtWebKit edited by treat@kde.org
(diff)
03:35 WikiStart edited by treat@kde.org
Reorganize a bit and add page for Git (diff)
03:30 UsingGitWithWebKit edited by treat@kde.org
Tips and tricks (diff)
03:03 Changeset [27970] by ap@webkit.org

2007-11-22 Julien Chaffraix <julien.chaffraix@gmail.com>

Reviewed by Alexey.

Bug 15530: XMLHttpRequest should not support certain methods

Test: http/tests/xmlhttprequest/xmlhttprequest-forbidden-methods-exception.html

  • xml/XMLHttpRequest.cpp: (WebCore::XMLHttpRequest::open):
02:55 Changeset [27969] by mrowe@apple.com

Remove WARNING notes from change logs that committers neglected to remove when landing.

02:42 WebKit Team edited by mjs@apple.com
(diff)
02:37 WebKit Team edited by mjs@apple.com
(diff)
02:24 WebKit Team edited by mjs@apple.com
(diff)
02:20 WebKit Team edited by mrowe@apple.com
(diff)
02:18 WebKit Team edited by alp@atoker.com
Update (diff)
02:16 Changeset [27968] by hausmann@webkit.org

Fix changelog

02:13 Changeset [27967] by hausmann@webkit.org

Fix compilation (don't define notImplemented twice)

02:12 Changeset [27966] by hausmann@webkit.org

Use Q_SIGNALS/Q_SLOTS in the public API

02:12 Changeset [27965] by hausmann@webkit.org

Fixed the signature of the QWebPage constructor to follow Qt guidelines.

02:12 Changeset [27964] by hausmann@webkit.org

Use <QtModule/headerfile.h> instead of <Classname> in public HEADER files, to not require the include paths for the other modules to be present

02:11 Changeset [27963] by hausmann@webkit.org

Removed inline copy of QExplicitlySharedDataPointer. This is not needed anymore since we require Qt >= 4.3.0.

02:11 Changeset [27962] by hausmann@webkit.org

For the build inside Qt include qbase.pri and don't duplicate what qbase.pri otherwise does (install targets, etc.)

02:11 Changeset [27961] by hausmann@webkit.org

Don't set OBJECTS_DIR when building inside Qt

02:10 Changeset [27960] by hausmann@webkit.org

Use ../generated as path for the generated sources for builds inside Qt

02:10 Changeset [27959] by hausmann@webkit.org

Make the directory of where to put the generated sources configurable through the GENERATED_SOURCE_DIR variable

02:10 Changeset [27958] by hausmann@webkit.org

Centralize the setup for all the extra compilers in a addExtraCompiler function.

This allows adding a "generated_files" target that builds all generated files
using "make generated_files". For the build inside Qt we do not generate
actual rules for the extra compilers but instead do the variable substitution
of compiler.output manually and add the generated sources to SOURCES.

02:06 UsingGitWithWebKit created by treat@kde.org
Start new page for collecting info for using git with webkit
02:03 QtWebKitContrib edited by treat@kde.org
Add a link for new page describing tips and tricks for using git with … (diff)
01:58 WikiStart edited by treat@kde.org
(diff)
01:40 WebKit Team edited by alp@atoker.com
Fix own entry (diff)
01:21 QtWebKit edited by treat@kde.org
(diff)
01:19 QtWebKitContrib edited by treat@kde.org
More information for new contributors (diff)
01:12 Open Source Contributors edited by treat@kde.org
(diff)
01:11 WikiStart edited by treat@kde.org
Don't need this anymore as we can add people to the Team page. (diff)
01:10 WebKit Team edited by treat@kde.org
Add lots more contributors and add review/commit info. (diff)
01:08 Changeset [27957] by hausmann@webkit.org

Fix license headers

Signed-off-by: Simon Hausmann <hausmann@kde.org>

00:50 Changeset [27956] by hausmann@webkit.org

Fix shadowing of "page" variable that is passed as argument as well as a member variable.

This also fixes compilation with gcc 4.3.

Signed-off-by: George Staikos <staikos@kde.org>

11/21/07:

23:06 Changeset [27955] by eric@webkit.org

2007-11-21 Eric Seidel <eric@webkit.org>

Speculative build fix for Tiger.

  • DumpRenderTree/mac/PixelDumpSupport.mm: include unistd.h
22:48 Changeset [27954] by mrowe@apple.com

Update changelog entry to match reality.

21:54 Changeset [27953] by timothy@apple.com

Reviewed by Eric Seidel.

Show Fonts as Yellow in the network timeline.

  • page/inspector/NetworkPanel.js:
  • page/inspector/inspector.css:
21:23 Changeset [27952] by mitz@apple.com

WebCore:

Reviewed by Eric Seidel.

  • fix <rdar://problem/5607037> REGRESSION (r27351): Departure date does not repaint when changed on Google Maps public transit planner (16034)

Test: fast/repaint/subtree-root-skipped.html

  • page/FrameView.cpp: (WebCore::FrameViewPrivate::FrameViewPrivate): Initialize the layout root to 0. (WebCore::FrameView::layoutRoot): Changed to return a RenderObject instead of a Node. (WebCore::FrameView::layout): Changed for layout root being a renderer rather than a DOM node. Also replaced clearing the repaint rects set with asserting that it is empty if this is the top-level call to layout(). If it is not, the set may contain rects from enclosing layout() and those should not be removed. (WebCore::FrameView::scheduleRelayout): Changed for layout root being a renderer rather than a DOM node. (WebCore::isObjectAncestorContainerOf): Added this helper function that tests whether one object will be marked by calling markContainingBlocksForLayout() on the other. (WebCore::FrameView::scheduleRelayoutOfSubtree): Changed for layout root being a renderer rather than a DOM node. Changed the check if new and current layout roots are on the same path from the root to use the subgraph of the render tree defined by container()hood instead of the DOM tree and parenthood.
  • page/FrameView.h:
  • rendering/RenderBox.cpp: (WebCore::RenderBox::calcWidth): Changed for layout root being a renderer rather than a DOM node.
  • rendering/RenderObject.cpp: (WebCore::RenderObject::~RenderObject): Added an assertion that the object being deleted is not currently the layout root. (WebCore::RenderObject::scheduleRelayout): Changed for layout root being a renderer rather than a DOM node.

LayoutTests:

Reviewed by Eric Seidel.

  • repaint test for <rdar://problem/5607037> REGRESSION (r27351): Departure date does not repaint when changed on Google Maps public transit planner (16034)
  • fast/repaint/subtree-root-skipped.html: Added.
  • platform/mac-leopard/fast/repaint/subtree-root-skipped-expected.checksum: Added.
  • platform/mac-leopard/fast/repaint/subtree-root-skipped-expected.png: Added.
  • platform/mac/fast/repaint/subtree-root-skipped-expected.txt: Added.
20:32 Changeset [27951] by mrowe@apple.com

Fix JavaScriptGlue, WebCore and WebKit to build without warnings using GCC 4.2.

Reviewed by Eric.

20:13 Changeset [27950] by mrowe@apple.com

2007-11-21 Mark Rowe <mrowe@apple.com>

Reviewed by Eric.

Update coding style to cover control statements with no body.

  • coding/coding-style.html:
20:03 Changeset [27949] by eric@webkit.org

2007-11-21 Eric Seidel <eric@webkit.org>

Reviewed by Tim Hatcher.

Break out more of DumpRenderTree.mm into individual files

  • DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
  • DumpRenderTree/mac/CheckedMalloc.cpp: Added. (checkedMalloc): (checkedRealloc): (makeLargeMallocFailSilently):
  • DumpRenderTree/mac/CheckedMalloc.h: Added.
  • DumpRenderTree/mac/DumpRenderTree.mm: (crashHandler): (dump):
  • DumpRenderTree/mac/DumpRenderTreeMac.h:
  • DumpRenderTree/mac/DumpRenderTreePasteboard.h:
  • DumpRenderTree/mac/DumpRenderTreePasteboard.m:
  • DumpRenderTree/mac/DumpRenderTreeWindow.h:
  • DumpRenderTree/mac/DumpRenderTreeWindow.mm:
  • DumpRenderTree/mac/JavaScriptThreading.cpp: Added. (javaScriptThreads): (runJavaScriptThread): (startJavaScriptThreads): (stopJavaScriptThreads):
  • DumpRenderTree/mac/JavaScriptThreading.h: Added.
  • DumpRenderTree/mac/PixelDumpSupport.h: Added.
  • DumpRenderTree/mac/PixelDumpSupport.mm: Added. (restoreColorSpace): (setDefaultColorProfileToRGB): (initializeColorSpaceAndScreeBufferForPixelTests): (md5HashStringForBitmap): (dumpWebViewAsPixelsAndCompareWithExpected):
19:28 Changeset [27948] by alp@webkit.org

2007-11-21 Alp Toker <alp@atoker.com>

Reviewed by Mark Rowe.

http://bugs.webkit.org/show_bug.cgi?id=16071
Curl backend handles EINTR incorrectly

Defer timers during select() to avoid interruption by timer signals.

  • platform/network/curl/ResourceHandleManager.cpp: (WebCore::ResourceHandleManager::downloadTimerCallback):
19:08 Changeset [27947] by mrowe@apple.com

Fix <rdar://problem/5602936> Need to resolve new GCC 4.2 warnings.

Reviewed by Tim Hatcher.

Fix all warnings emitted by GCC 4.2 when building JavaScriptCore. This allows builds with
-Werror to succeed. At present they will crash when executed due to code that is not safe
under strict aliasing (<rdar://problem/5536806>).

This required some format strings to be modified in WebCore and WebKit as their format
specifiers did not match the argument type.

19:01 Changeset [27946] by kevino@webkit.org

Restore the wx buildbot for Mac by automatically installing dependencies not provided by the OS into WebKitLibraries.

Reviewed by Mark Rowe

18:19 Changeset [27945] by eric@webkit.org

2007-11-21 Eric Seidel <eric@webkit.org>

Reviewed by Adam.

More refactoring for greater code readability

  • DumpRenderTree/mac/DumpRenderTree.mm: (initializeGlobalsFromCommandLineOptions): (initializeColorSpaceAndScreeBufferForPixelTests): (addTestPluginsToPluginSearchPath): (useLongRunningServerMode): (runTestingServerLoop): (prepareConsistentTestingEnvironment): (dumpRenderTree): (main): (dumpFramesAsText): (dumpBackForwardListForWebView): (sizeWebViewForCurrentTest): (methodNameStringForFailedTest): (dumpBackForwardListForAllWindows): (dumpWebViewAsPixelsAndCompareWithExpected): (invalidateAnyPreviousWaitToDumpWatchdog): (dump):
  • DumpRenderTree/mac/DumpRenderTreeMac.h:
  • DumpRenderTree/mac/UIDelegate.mm: (-[UIDelegate webView:createWebViewWithRequest:]):
16:24 Changeset [27944] by alp@webkit.org

2007-11-21 Xan Lopez <xan@gnome.org>

Reviewed by Alp Toker.

Clarify scroll event processing with a comment.

  • platform/gtk/PlatformScrollBarGtk.cpp: (gtkScrollEventCallback):
15:40 Changeset [27943] by timothy@apple.com

Reviewed by Adam Roben.

Bug 16085: Web Inspector's Network Timeline graph code clean up and misc. fixes
http://bugs.webkit.org/show_bug.cgi?id=16085


Changes include:

  • Refactor the drawing code as nested functions instead of global functions.
  • Compute the segment percentages only once per call to drawSummaryGraph.
  • Account for percentages that rounded down to total less-than 100%.
  • Draw the pill shadow better using the canvas shadow drawing properties.
  • Removes a couple canvas context saves and restores.
  • page/inspector/NetworkPanel.js:
11:29 Changeset [27942] by eric@webkit.org

2007-11-21 Eric Seidel <eric@webkit.org>

Reviewed by Tim Hatcher.

Abstract more of DRT into static methods

  • DumpRenderTree/mac/DumpRenderTree.mm: (setDefaultsToConsistentValuesForTesting): (setupSignalHandlers): (allocateGlobalControllers): (releaseAndZero): (releaseGlobalControllers): (dumpRenderTree): (shouldLogFrameLoadDelegates): (createCFURLFromPathOrURL): (resetWebViewToConsistentStateBeforeTesting): (runTest):
11:28 Changeset [27941] by eric@webkit.org

2007-11-21 Eric Seidel <eric@webkit.org>

Reviewed by Tim Hatcher.

Pull DumpRenderTreeWindow and DumpRenderTreePasteboard out into their own files

  • DumpRenderTree/DumpRenderTree.h:
  • DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
  • DumpRenderTree/mac/DumpRenderTree.mm: (dumpRenderTree): (dump): (runTest):
  • DumpRenderTree/mac/DumpRenderTreePasteboard.h: Added.
  • DumpRenderTree/mac/DumpRenderTreePasteboard.m: Added. (+[DumpRenderTreePasteboard _pasteboardWithName:]): (+[DumpRenderTreePasteboard releaseLocalPasteboards]): (-[DumpRenderTreePasteboard declareType:owner:]): (+[LocalPasteboard alloc]): (-[LocalPasteboard init]): (-[LocalPasteboard dealloc]): (-[LocalPasteboard name]): (-[LocalPasteboard releaseGlobally]): (-[LocalPasteboard declareTypes:owner:]): (-[LocalPasteboard addTypes:owner:]): (-[LocalPasteboard changeCount]): (-[LocalPasteboard types]): (-[LocalPasteboard availableTypeFromArray:]): (-[LocalPasteboard setData:forType:]): (-[LocalPasteboard dataForType:]): (-[LocalPasteboard setPropertyList:forType:]): (-[LocalPasteboard setString:forType:]):
  • DumpRenderTree/mac/DumpRenderTreeWindow.h: Added.
  • DumpRenderTree/mac/DumpRenderTreeWindow.mm: Added. (+[DumpRenderTreeWindow allWindows]): (-[DumpRenderTreeWindow initWithContentRect:styleMask:backing:defer:]): (-[DumpRenderTreeWindow dealloc]): (-[DumpRenderTreeWindow isKeyWindow]): (-[DumpRenderTreeWindow keyDown:]):
10:48 BuildingGtk edited by alp@atoker.com
Add public API and GtkLauncher to the list of code modules (diff)
08:18 QtWebKitContrib edited by treat@kde.org
(diff)
08:15 BuildingQtOnWindows edited by treat@kde.org
(diff)
08:14 QtWebKitContrib edited by treat@kde.org
(diff)
08:13 BuildingQtOnLinux edited by treat@kde.org
(diff)
08:11 BuildingQtOnLinux edited by treat@kde.org
(diff)
08:08 QtWebKitContrib edited by treat@kde.org
(diff)
08:07 QtWebKitTodo edited by treat@kde.org
(diff)
08:07 QtWebKit edited by treat@kde.org
(diff)
08:06 QtWebKitContrib edited by treat@kde.org
(diff)
08:01 QtWebKitContrib created by treat@kde.org
07:16 QtWebKit edited by treat@kde.org
(diff)
05:02 Changeset [27940] by treat@webkit.org
  • Connect up the signal for all subframes too.
04:20 Changeset [27939] by ap@webkit.org

Reviewed by Maciej.

<rdar://problem/5230233> dom/xhtml/level2/html/frame.xhtml is failing on Windows

This isn't actually a test, but an unused resource - we should re-work the W3C test rig to
support subresource loading from (X)HTML.
Anyway, it doesn't fail for me, I suppose this problem was fixed since the bug was filed.

  • platform/win/Skipped:
04:07 QtWebKitTodo edited by hausmann@webkit.org
(diff)
03:54 Changeset [27938] by ap@webkit.org

Reviewed by Maciej.

Fix a copy/paste error in the test (introduced in bug 12194).

  • http/tests/xmlhttprequest/zero-length-response-sync-expected.txt:
  • http/tests/xmlhttprequest/zero-length-response-sync.html:
02:01 WikiStart edited by aroben@apple.com
(diff)
00:10 Changeset [27937] by ap@webkit.org

Reviewed by Adam Roben.

http://bugs.webkit.org/show_bug.cgi?id=16056
Unicode not being interpreted correctly in Web Inspector source view

Cannot be tested automatically.

  • loader/CachedCSSStyleSheet.cpp: (WebCore::CachedCSSStyleSheet::encoding):
  • loader/CachedCSSStyleSheet.h:
  • loader/CachedResource.h: (WebCore::CachedResource::encoding):
  • loader/CachedScript.cpp: (WebCore::CachedScript::encoding):
  • loader/CachedScript.h:
  • loader/CachedXBLDocument.cpp: (WebCore::CachedXBLDocument::encoding):
  • loader/CachedXBLDocument.h:
  • loader/CachedXSLStyleSheet.cpp: (WebCore::CachedXSLStyleSheet::encoding):
  • loader/CachedXSLStyleSheet.h: Teach textual CachedResources to report their encodings.
  • page/InspectorController.cpp: (WebCore::addSourceToFrame): (WebCore::updateResourceResponse): Use the actual encoding - the network layer has little idea about it.

11/20/07:

23:29 Changeset [27936] by mrowe@apple.com

Fix <rdar://problem/5609579> (DOMParser().parseFromString() freezes Safari when parsing large nodes with XML entities)
http://bugs.webkit.org/show_bug.cgi?id=16076

Reviewed by Maciej Stachowiak.

XMLTokenizer was calling CharacterData::appendData twice per entity in the fragment of XML being
parsed (once for text before the entity, once for the entity itself). This triggered O(n2) copying
of the CharacterData's string due to resizing. We now prevent this happening by buffering all the
content for a given Text node in the XMLTokenizer before sending it out to the node in a single go.

22:55 Changeset [27935] by timothy@apple.com

Reviewed by Mark Rowe.

Animate the status area and tip balloons in the Web Inspector
with CSS animations.

  • page/inspector/NetworkPanel.js:
  • page/inspector/inspector.css:
  • page/inspector/inspector.js:
17:14 Changeset [27934] by kevino@webkit.org

wx build fix for Windows. Don't use WebCore/move-js-headers.sh as it indiscriminately copies any headers inside JavaScriptCore, which includes Tiger ICU headers.

14:11 Changeset [27933] by alp@webkit.org

2007-11-20 Alp Toker <alp@atoker.com>

Reviewed by Mark Rowe.

Avoid pointlessly mallocing and freeing this transform matrix.

  • platform/gtk/FontPlatformDataGtk.cpp: (WebCore::FontPlatformData::FontPlatformData):
13:46 Changeset [27932] by alp@webkit.org

2007-11-20 Naiem Shaik <naiem.shaik@gmail.com>

Reviewed by Alp Toker.

http://bugs.webkit.org/show_bug.cgi?id=15763
[GTK] Enter key does not take to the link highlighted.

Add missing keycode cases.

  • platform/gtk/KeyEventGtk.cpp: (WebCore::keyIdentifierForGdkKeyCode): (WebCore::windowsKeyCodeForKeyEvent): (WebCore::PlatformKeyboardEvent::PlatformKeyboardEvent):
13:36 Changeset [27931] by mrowe@apple.com

Fix http://bugs.webkit.org/show_bug.cgi?id=16074
Bug 16074: execCommand("InsertHorizontalRule", false, "") results in id="" being inserted

For consistency with InsertOrderedList and InsertUnorderedList a value parameter of "" should
not result in an id being set.

Reviewed by Sam Weinig.

10:01 Changeset [27930] by treat@webkit.org
  • Prepend git branch name to $baseProductDir
09:38 Changeset [27929] by mitz@apple.com

Reviewed by Adam Roben.

  • fix <rdar://problem/5090708> Textareas render broken-looking scrollbars when too short to show full scrollbar
  • platform/win/PlatformScrollBar.h:
  • platform/win/PlatformScrollBarSafari.cpp: Added the "hit inset" constants (representing how far the track eats into the button). (WebCore::PlatformScrollbar::paint): Changed to paint the buttons and the thumb only if they should be showing. (WebCore::PlatformScrollbar::hasButtons): Added. Returns whether the arrows should be drawn. (WebCore::PlatformScrollbar::hasThumb): Added. Returns whether the thumb should be drawn. (WebCore::PlatformScrollbar::forwardButtonRect): (WebCore::PlatformScrollbar::trackRect): Changed to return the entire bounds of the scrollbar if the scrollbar has no buttons. (WebCore::PlatformScrollbar::paintTrack): Changed to paint a disabled track along the entire scrollbar if it has not buttons. (WebCore::PlatformScrollbar::hitTest): Changed to hit test only the parts that the scrollbar has in its current dimensions.
08:06 Changeset [27928] by mrowe@apple.com

2007-11-20 Mark Rowe <mrowe@apple.com>

Reviewed by Alp Toker.

  • WebKit.pri: Add in -Wno-unused-parameter to silence warnings in WebCore.

2007-11-20 Mark Rowe <mrowe@apple.com>

Reviewed by Alp Toker.

  • config.h: Change #if to #ifdef to silence warnings on non-Apple platforms.
07:53 Changeset [27927] by mrowe@apple.com

2007-11-20 Mark Rowe <mrowe@apple.com>

Reviewed by Simon Hausmann.

  • Scripts/build-webkit: Pass "clean" flag down into buildQMakeProject.
  • Scripts/webkitdirs.pm: Respect the "clean" flag passed down from build-webkit. Have it trigger a "make distclean" rather than "make clean" to ensure that the built product and generated Makefile's are removed.
07:35 Changeset [27926] by alp@webkit.org

2007-11-20 Alp Toker <alp@atoker.com>

Reviewed by Simon Hausmann.

Clean up GTK+ port configuration.

Use similar compiler warning flags to the Mac build.

Add the qmake changes needed for DirectFB support.

  • WebKit.pri:
07:27 Changeset [27925] by hausmann@webkit.org

Remove static linkage of QtWebKit against the ICO image format plugin.

Instead build the support for the ICO image format as a proper standalone qt image format plugin and install it.

Signed-off-by: Adam Treat <treat@kde.org>

06:52 Changeset [27924] by ddkilzer@apple.com

Fix typo.

06:48 Changeset [27923] by ddkilzer@apple.com

no-svg build broken after r27278
<http://bugs.webkit.org/show_bug.cgi?id=16061>

Reviewed by Eric.

  • bindings/scripts/CodeGeneratorObjC.pm: Always generate DOMHTMLEmbedElementPrivate.h and DOMHTMLObjectElementPrivate.h for no-svg build.
06:23 Changeset [27922] by hausmann@webkit.org

Added a default argument to textContent for the hoveringOverLink signal to keep existing two-argument connections working

Signed-off-by: George Staikos <staikos@kde.org>

05:26 Changeset [27921] by treat@webkit.org
  • Don't use so much heap memory.
04:37 Changeset [27920] by treat@webkit.org
  • Make gif animations work for instance.
01:06 Changeset [27919] by hausmann@webkit.org

Add an argument to pass the link text in the hovering signal

Signed-off-by: Simon Hausmann <hausmann@kde.org>

01:03 Changeset [27918] by hausmann@webkit.org

Adjust some test result after fixing a bug related to computation
of QFont::xHeight() in Qt.

01:02 Changeset [27917] by hausmann@webkit.org

Fix text break interators.

The break iterators where trying to be smart about the input
string and caching their results. Unfortunately a pointer/string length
comparison is not good enough in all cases (since some input is stack based
temp strings). Removed the caching but at the same time started to use a more
efficient constructor of QTextBoundaryFinder that doesn't need to malloc for
most strings.

Fixes two test cases that test "text-transform: capitalize".

00:50 Changeset [27916] by treat@webkit.org
  • Build in release mode
00:29 Changeset [27915] by treat@webkit.org
  • Be quiet and allow suppression of NotImplemented calls at runtime.

11/19/07:

18:39 Changeset [27914] by alp@webkit.org

2007-11-19 Doug Turner <dougt@meer.net>

Reviewed by Alp Toker.

http://bugs.webkit.org/show_bug.cgi?id=16054
Crash when GlyphPage::fill is called with more than 256 bytes of data

http://bugs.webkit.org/show_bug.cgi?id=14446
[GDK] Crash on http://www.wikipedia.org/

setGlyphDataForIndex() uses a fixed array of size 256 which we can't
exceed. We need to return failure if the buffer has Unicode
supplementary characters for now.

This strategy matches the Win port, which also doesn't support this
case yet.

Add an assertion so nobody makes this mistake again.

  • platform/GlyphPageTreeNode.h: (WebCore::GlyphPage::setGlyphDataForIndex):
  • platform/gtk/GlyphPageTreeNodeGtk.cpp: (WebCore::GlyphPage::fill):
16:24 Changeset [27913] by kevino@webkit.org

wx port build fix (wx headers include ctype functions).

14:53 Changeset [27912] by antti@apple.com

Rubber-stamped by bdash.


Improve coverage of these tests.

  • media/video-muted-expected.txt:
  • media/video-muted.html:
  • media/video-volume-expected.txt:
  • media/video-volume.html:
11:39 BuildingGtk edited by alp@atoker.com
Note the targets supported by the GTK+ port (diff)
11:00 Changeset [27911] by kevino@webkit.org

Remove outdated and unused Windows port files.

Reviewed by Adam Roben.

10:01 Changeset [27910] by alp@webkit.org

2007-11-19 Doug Turner <dougt@meer.net>

Reviewed by Timothy Hatcher.

http://bugs.webkit.org/show_bug.cgi?id=16050
sqlite3_prepare16_v2 build bustage.

Using the correct SQLITE_VERSION_NUMBER around sqlite_prepare16_v2
to fix build bustage.

  • platform/sql/SQLiteStatement.cpp: (WebCore::SQLiteStatement::prepare):
09:56 Changeset [27909] by mrowe@apple.com

2007-11-19 Mark Rowe <mrowe@apple.com>

Build fix. Don't over-qualify the constructor name.

  • storage/DatabaseDetails.h:
09:51 Changeset [27908] by beidson@apple.com

WebCore:

Reviewed by Maciej

Stub out the WebCore parts of the WebKit API

  • WebCore.base.exp:
  • WebCore.xcodeproj/project.pbxproj:
  • storage/DatabaseDetails.h: Added. Simple container for vitals on a specific database (WebCore::DatabaseDetails::DatabaseDetails::DatabaseDetails): (WebCore::DatabaseDetails::DatabaseDetails): (WebCore::DatabaseDetails::isValid): (WebCore::DatabaseDetails::name): (WebCore::DatabaseDetails::version): (WebCore::DatabaseDetails::displayName): (WebCore::DatabaseDetails::expectedUsage): (WebCore::DatabaseDetails::currentUsage):
  • storage/DatabaseTracker.cpp: Added various methods for API usage (WebCore::DatabaseTracker::detailsForNameAndOrigin): (WebCore::DatabaseTracker::usageForOrigin): (WebCore::DatabaseTracker::quotaForOrigin): (WebCore::DatabaseTracker::setQuota):
  • storage/DatabaseTracker.h:

WebKit:

Reviewed by Maciej

  • WebKit.xcodeproj/project.pbxproj: Sorted!

WebKit/mac:

Reviewed by Maciej

Finished hooking up the WebKit API for database management.
Most of the API is actually implemented in WebCore and some of those methods might
only be stubs for now.

  • Storage/WebDatabaseManager.mm: (-[WebDatabaseManager origins]): Call through to the WebCore tracker and construct an API result (-[WebDatabaseManager databasesWithOrigin:]): Ditto (-[WebDatabaseManager detailsForDatabase:withOrigin:]): Ditto
  • Storage/WebSecurityOrigin.mm: (-[WebSecurityOrigin usage]): Call through to WebCore (-[WebSecurityOrigin quota]): Ditto (-[WebSecurityOrigin setQuota:]): Ditto (-[WebSecurityOrigin _core]): Get WebCore version of this object
  • Storage/WebSecurityOriginInternal.h:
09:28 Changeset [27907] by kevino@webkit.org

wx port build fix for !USE(WXGC).

07:59 Changeset [27906] by alp@webkit.org

2007-11-19 Alp Toker <alp@atoker.com>

Reviewed by Mark Rowe.

http://bugs.webkit.org/show_bug.cgi?id=16040
[GTK] GtkLauncher should be written in C

Port GtkLauncher to plain C.

Use a more conventional GTK+ coding style.

Use a toolbar instead of menus.

Various signature fixes and cleanups.

Add a license header. Assume all previous modifications were copyright
assigned to Apple Inc. by default.

  • GtkLauncher/GtkLauncher.pro:
  • GtkLauncher/main.c: Added. (activate_url_entry_cb): (update_title): (link_hover_cb): (title_change_cb): (progress_change_cb): (destroy_cb): (go_back_cb): (go_forward_cb): (create_browser): (create_statusbar): (create_toolbar): (create_window): (main):
  • GtkLauncher/main.cpp: Removed.
07:53 Changeset [27905] by alp@webkit.org

2007-11-19 Alp Toker <alp@atoker.com>

Reviewed by Mark Rowe.

GTK+ drawing fixes.

Add a check to avoid crashing when the GraphicsContext is not
associated with a GdkDrawable. This was noticed when adding printing
support but might be triggered in other situations too.

Do not render themes when painting is disabled. This is an
optimisation for cases where GraphicsContext is used to calculate page
dimensions etc. without actually rendering.

  • platform/graphics/cairo/GraphicsContextCairo.cpp: (WebCore::GraphicsContext::gdkDrawable):
  • platform/gtk/RenderThemeGtk.cpp: (WebCore::paintMozWidget):
07:33 Changeset [27904] by treat@webkit.org
  • Don't segfault when event pos is outside of the widget.
05:43 Changeset [27903] by mrowe@apple.com

2007-11-13 Rahul Abrol <ra5ul@comcast.net>

Reviewed by Tim Hatcher.

http://bugs.webkit.org/show_bug.cgi?id=15977
Resizing images preference now toggles default image state.

  • loader/ImageDocument.cpp: (WebCore::ImageDocument::ImageDocument): (WebCore::ImageDocument::createDocumentStructure): (WebCore::ImageDocument::imageChanged):
05:02 Changeset [27902] by mrowe@apple.com

Roll out r27896 as it caused lots of build breakages.

04:51 Changeset [27901] by mrowe@apple.com

2007-11-19 Mark Rowe <mrowe@apple.com>

Build fix.

  • platform/image-decoders/gif/GIFImageDecoder.cpp: (WebCore::GIFImageDecoder::initFrameBuffer):
04:47 Changeset [27900] by mrowe@apple.com

2007-11-19 Mark Rowe <mrowe@apple.com>

Build fix.

  • platform/image-decoders/gif/GIFImageDecoder.cpp: (WebCore::GIFImageDecoder::initFrameBuffer): Remove reference to SkBitmap type that doesn't belong.
04:37 Changeset [27899] by mrowe@apple.com

2007-11-19 Kevin Ollivier <kevino@theolliviers.com>

Build script fixes to ensure they do the right thing for the
wx port, and update build-wxwebkit to reflect the way the
build scripts now work.

Reviewed by Adam.

  • Scripts/build-testkjs:
  • Scripts/build-webkit:
  • Scripts/run-javascriptcore-tests:
  • Scripts/webkitdirs.pm:
  • wx/build-wxwebkit:

2007-11-19 Kevin Ollivier <kevino@theolliviers.com>

Add pcre directory to JSCore includes, and update the wx port
to reflect the Shared -> RefCounted name change. Also, fix
WebFrame.cpp to re-enable code that should never have been
committed disabled.

Reviewed by Adam.

2007-11-19 Kevin Ollivier <kevino@theolliviers.com>

Update wx port build sources with recent changes, update
wx FontPlatformData to be a class (MSVC7 gets confused
otherwise...) and implement its hash() method.

Reviewed by Adam.

  • WebCoreSources.bkl:
  • platform/wx/FontPlatformData.h: (WebCore::FontPlatformData::hash):
04:30 Changeset [27898] by mrowe@apple.com

2007-11-19 Julien Chaffraix <julien.chaffraix@gmail.com>

Reviewed by Darin.

http://bugs.webkit.org/show_bug.cgi?id=12194
Bug 12194: Trying to access XMLHttpRequest.responseText or responseXML when they
are not available should raise an exception

Tests: http/tests/xmlhttprequest/xmlhttprequest-responseText-exception.html

http/tests/xmlhttprequest/xmlhttprequest-responseXML-exception.html

  • bindings/js/JSXMLHttpRequest.cpp: (KJS::JSXMLHttpRequest::getValueProperty):
  • xml/XMLHttpRequest.cpp: (WebCore::XMLHttpRequest::getResponseText): (WebCore::XMLHttpRequest::getResponseXML):
  • xml/XMLHttpRequest.h:

2007-11-19 Julien Chaffraix <julien.chaffraix@gmail.com>

Reviewed by Darin.

http://bugs.webkit.org/show_bug.cgi?id=12194
Bug 12194: Trying to access XMLHttpRequest.responseText or responseXML when they are not
available should raise an exception

  • http/tests/xmlhttprequest/xmlhttprequest-responseText-exception-expected.txt: Added.
  • http/tests/xmlhttprequest/xmlhttprequest-responseText-exception.html: Added.
  • http/tests/xmlhttprequest/xmlhttprequest-responseXML-exception-expected.txt: Added.
  • http/tests/xmlhttprequest/xmlhttprequest-responseXML-exception.html: Added.
  • http/tests/xmlhttprequest/zero-length-response-expected.txt: Updated result
  • http/tests/xmlhttprequest/zero-length-response-sync-expected.txt: Updated result
  • http/tests/xmlhttprequest/zero-length-response-sync.html: Corrected URL & added exception handling to test case
04:22 Changeset [27897] by mrowe@apple.com

2007-11-19 Peter Kasting <zerodpx@gmail.com>

Reviewed by Darin Adler.

http://bugs.webkit.org/show_bug.cgi?id=15971
The GIF decoder should not fail decoding if the caller asks it to
decode again when no new data has arrived since the last call.

  • platform/image-decoders/gif/GIFImageReader.cpp: (GIFImageReader::read):
04:22 Changeset [27896] by mrowe@apple.com

2007-11-19 Peter Kasting <zerodpx@gmail.com>

Reviewed by Darin Adler.

http://bugs.webkit.org/show_bug.cgi?id=15974
GIF decoding should respect frames' specified disposal methods.

  • platform/image-decoders/ImageDecoder.h: (WebCore::RGBA32Buffer::): (WebCore::RGBA32Buffer::RGBA32Buffer): (WebCore::RGBA32Buffer::disposalMethod): (WebCore::RGBA32Buffer::setDisposalMethod):
  • platform/image-decoders/gif/GIFImageDecoder.cpp: (WebCore::GIFImageDecoder::frameBufferAtIndex): (WebCore::GIFImageDecoder::initFrameBuffer): (WebCore::GIFImageDecoder::prepEmptyFrameBuffer): (WebCore::GIFImageDecoder::haveDecodedRow): (WebCore::GIFImageDecoder::frameComplete):
  • platform/image-decoders/gif/GIFImageDecoder.h:
  • platform/image-decoders/gif/GIFImageReader.cpp: (GIFImageReader::read):
  • platform/image-decoders/gif/GIFImageReader.h: (GIFFrameReader::GIFFrameReader):

11/18/07:

23:14 Changeset [27895] by beidson@apple.com

Reviewed by Eric

Fix a crash hashing a default SecurityOriginData object

  • storage/DatabaseTracker.cpp: (WebCore::SecurityOriginDataHash::hash): The default object has some default strings with NULL StringImpls - don't try to hash those!
16:40 Changeset [27894] by eric@webkit.org

2007-11-18 Eric Seidel <eric@webkit.org>

Reviewed by Oliver.

  • tests/mozilla/jsDriver.pl: exit non-0 when user aborts test run
15:04 Changeset [27893] by timothy@apple.com

Reviewed by Sam Weinig.

Bug 16043: Remove InspectorController.log now that console.log works
http://bugs.webkit.org/show_bug.cgi?id=16043

Remove the last uses of InspectorController.log. Once use was no longer needed,
checking an erro case that can't happen anymore. The other case now uses console.error.

  • page/InspectorController.cpp: Remove the log function. (WebCore::InspectorController::windowScriptObjectAvailable): Remove the log function from InspectorController script class.
  • page/inspector/DocumentPanel.js: Remove a use of InspectorController.log that isn't needed.
  • page/inspector/Panel.js: Change InspectorController.log to console.log and early return.
15:01 Changeset [27892] by timothy@apple.com

Reviewed by Brady.

Bug 16041: REGRESSION: the Database panel in the Inspector stopped showing errors
http://bugs.webkit.org/show_bug.cgi?id=16041

  • page/inspector/DatabasePanel.js: Implement error callbacks to pass to executeSql() and transaction(). Refactored some code for the success and error callbacks to use.
14:18 Changeset [27891] by eric@webkit.org

2007-11-18 Eric Seidel <eric@webkit.org>

Reviewed by Anders.

Make run-javascriptcore-tests report failures on exit (to support git bisect)

  • Scripts/run-javascriptcore-tests:
09:54 Changeset [27890] by ap@webkit.org

Reviewed by Adam Roben.

Make run-webkit-tests work with Windows debug build.

  • DumpRenderTree/win/DumpRenderTree.cpp: (main): Only use memory checks with debug CRT.
  • DumpRenderTree/win/DumpRenderTree.vcproj: Switched Debug configuration to release CRT, as it is supposed to run with release Apple libraries. Removed _DEBUG preprocessor symbol, as it goes with debug CRT (AFAIK, it is supposed to be added automatically, and shouldn't be needed in Debug_internal configuration, but I didn't dare to change that).
09:15 Changeset [27889] by kevino@webkit.org

Add wxWebKit sample and build script, and integrate it with build-webkit. Also make build-webkit --clean work for all ports.

Reviewed by Darin Adler.

07:44 QtWebKitTodo edited by treat@kde.org
Updates (diff)
03:31 Changeset [27888] by alp@webkit.org

2007-11-18 Christian Dywan <christian@twotoasts.de>

Reviewed by Alp.

http://bugs.webkit.org/show_bug.cgi?id=15891
[GTK] Javascript console and dialogs are not implemented

Correct script-prompt to return NULL when cancelled.

Small changes to avoid compiler warnings.

  • Api/webkitgtkpage.cpp:
01:49 Changeset [27887] by ap@webkit.org

Restoring a binary svn:mime-type on the test.

01:46 Changeset [27886] by ap@webkit.org

Reviewed by Maciej.

<rdar://problem/5546393> Whitespace handling doesn't match HTML5.

HTML5 definition is the same as MSIE's, with the exception that the latter strips null
characters. Firefox also treats U+0008 as whitespace, but not U+000B or U+000C.

Test: fast/parser/html-whitespace.html

  • html/HTMLTokenizer.cpp: (WebCore::HTMLTokenizer::parseSpecial): (WebCore::HTMLTokenizer::parseTag): Use isASCIISpace, which matches HTML5 definition of whitespace, and also what we use to check for whitespace almost everywhere.
01:09 Changeset [27885] by mrowe@apple.com

Fix: <rdar://problem/5607032> (REGRESSION: testapi exits with assertion failure in debug build) and <rdar://problem/5440659> (JSGlobalContextCreate throws away globalObjectClass's prototype)

Split Interpreter's initialization into two distinct steps: the creation of the global prototypes
and constructors, and storing them on the global object. This allows JSClassRef's passed to
JSGlobalContextCreate to be instantiated with the correct prototype.

Reviewed by Darin Adler.

Note: See TracTimeline for information about the timeline view.