Timeline


and

09/02/09:

19:31 Changeset [48004] by eric@webkit.org
  • 2 edits in trunk/WebKitTools

2009-09-02 Laurent Cerveau <lcerveau@me.com>

Reviewed by David Kilzer.

<http://webkit.org/b/25517> build-webkit script should print build time at end

  • Scripts/build-webkit: Added startTime and endTime variable so that the build time is computed and printed as part of the build message; display formatting has been separated in a dedicated subroutine.
18:44 Changeset [48003] by ddkilzer@apple.com
  • 2 edits in trunk/WebKitTools

<http://webkit.org/b/28881> svn-create-patch should check if the repo path is the same when trying to find the root

Reviewed by Eric Seidel.

  • Scripts/VCSUtils.pm: (determineSvnRoot): Added back check for repository root that was removed in r46134 when this code lived in svn-create-patch. It's necessary to check both the repository root and the repository UUID in case two different working directories are checked out from the same repository.
18:10 Changeset [48002] by timothy@apple.com
  • 2 edits in trunk/WebKitTools

Use new 512x512 icons for nightly builds.

Rubber-stamped by Mark Rowe.

  • WebKitLauncher/webkit.icns:
17:55 Changeset [48001] by beidson@apple.com
  • 5 edits in trunk/WebCore

More partial work towards "Page Cache should support pages with Frames"
https://bugs.webkit.org/show_bug.cgi?id=13631

Reviewed by Darin Adler.

No new tests. (No change in behavior, current tests pass).

  • loader/DocumentLoader.cpp:

(WebCore::DocumentLoader::stopLoading): Adopt the new enum form for FrameLoader::stopLoading().

  • loader/FrameLoader.cpp:

(WebCore::FrameLoader::stopLoading): Take an enum to decide whether to dispatch no unload events,

only unload, or unload and pagehide.

(WebCore::FrameLoader::closeURL):
(WebCore::FrameLoader::scheduleRedirection):
(WebCore::FrameLoader::cachePageForHistoryItem): Call pageHidden().
(WebCore::FrameLoader::pageHidden): Add to dispatch the pagehide event to all frames in the case

where a page is added to the PageCache.

  • loader/FrameLoader.h:
  • loader/FrameLoaderTypes.h: Add an UnloadEventPolicy enum.
17:18 Changeset [48000] by barraclough@apple.com
  • 3 edits in trunk/JavaScriptCore

Should crash if JIT code buffer allocation fails.

Patch by Gavin Barraclough <barraclough@apple.com> on 2009-09-02
https://bugs.webkit.org/show_bug.cgi?id=28926
<rdar://problem/7031922>

  • jit/ExecutableAllocatorPosix.cpp:

(JSC::ExecutablePool::systemAlloc):

  • jit/ExecutableAllocatorWin.cpp:

(JSC::ExecutablePool::systemAlloc):

17:07 Changeset [47999] by beidson@apple.com
  • 8 edits in trunk

WebCore: More partial work towards "Page Cache should support pages with Frames"
https://bugs.webkit.org/show_bug.cgi?id=13631

Reviewed by Darin Adler.

  • More CachedFrame destruction work.
  • Including related Frame and FrameLoader changes and cleanup.

No new tests. (No change in behavior, current tests pass).

  • history/CachedFrame.cpp:

(WebCore::CachedFrameBase::restore): Rebuild up the whole FrameTree when restoring.
(WebCore::CachedFrame::CachedFrame): Tear down the entire FrameTree when caching.
(WebCore::CachedFrame::destroy): For child frames that are being destroyed while in

the PageCache, do lightweight cleanup to:
A - Prevent referring to a stale Page object.
B - Prevent World Leaks of WebKit objects.

  • loader/FrameLoader.cpp:

(WebCore::FrameLoader::stopAllLoaders): stopAllLoaders() should never be called on Frames that are

in the PageCache. ASSERT that fact.

(WebCore::FrameLoader::open): Remove a bogus release-build workaround, as the comment right above it

explains it is not effective, and it'd be better to see the crash locally instead of downstream.

(WebCore::FrameLoader::closeAndRemoveChild): Make sure the owner element has a page pointer before

referencing it.

(WebCore::FrameLoader::detachFromParent): Use the new detachViewsAndDocumentLoader() call
(WebCore::FrameLoader::detachViewsAndDocumentLoader): Does lightweight cleanup of client (WebKit) objects.

  • loader/FrameLoader.h:
  • page/Frame.cpp:

(WebCore::Frame::detachFromPage): Simply clear the Page pointer without doing any other work.
(WebCore::Frame::pageDestroyed): Use the new detachFromPage().

  • page/Frame.h:

WebKit/mac: More partial work towards "Page Cache should support pages with Frames"
https://bugs.webkit.org/show_bug.cgi?id=13631

Reviewed by Darin Adler.

  • WebView/WebHTMLView.mm:

(-[WebHTMLView _topHTMLView]): Rework the ASSERT in this method to reflect the

reality of calling this method for pages currently in the PageCache.

16:41 Changeset [47998] by jshin@chromium.org
  • 2 edits in trunk/WebCore

2009-09-02 Jungshik Shin <jshin@chromium.org>

Reviewed by Dimitri Glazkov

[Chromium] Add 'icu::' qualifier in preparation for ICU upgrade to 4.2
in Chromium. This is to fix a new file added in r28742 since the
last patch for this issue.

https://bugs.webkit.org/show_bug.cgi?id=28410

No change in the test result.

  • platform/graphics/chromium/FontLinux.cpp: (WebCore::TextRunWalker::getNormalizedTextRun):
16:01 Changeset [47997] by kevino@webkit.org
  • 2 edits in trunk/WebKitTools

waf build fix. Remove local variable shadowing global.

15:21 Changeset [47996] by dglazkov@chromium.org
  • 2 edits in trunk/WebCore

2009-09-02 Dimitri Glazkov <dglazkov@chromium.org>

Unreviewed, build fix.

[Chromium] Correct a typo to fix bustage, caused by r47991.
https://bugs.webkit.org/show_bug.cgi?id=27933

  • platform/graphics/skia/ImageBufferSkia.cpp: Changed "Unpremultipled" to "Unmultiplied".
14:31 Changeset [47995] by dglazkov@chromium.org
  • 2 edits in trunk/WebCore

2009-09-02 Dimitri Glazkov <dglazkov@chromium.org>

Unreviewed, build fix.

[Chromium] Add a missing include to fix bustage, caused by r47991.
https://bugs.webkit.org/show_bug.cgi?id=27933

  • platform/graphics/skia/ImageBufferSkia.cpp: Added SkColorPriv include.
14:30 Changeset [47994] by jianli@chromium.org
  • 2 edits in trunk/WebCore

[Chromium] Implement file support for DragData.
https://bugs.webkit.org/show_bug.cgi?id=28896

Reviewed by Dimitri Glazkov.

  • platform/chromium/DragDataChromium.cpp:

(WebCore::DragData::containsURL):
(WebCore::DragData::asURL):
(WebCore::DragData::containsCompatibleContent):

14:07 Changeset [47993] by kevino@webkit.org
  • 2 edits in trunk/WebCore

Reviewed by Kevin Ollivier.

Protect libcurl shared data with Mutex objects via curl_set_share_opt.
https://bugs.webkit.org/show_bug.cgi?id=28920

13:45 Changeset [47992] by eric@webkit.org
  • 3 edits in trunk/WebCore

2009-09-02 Evan Stade <estade@chromium.org>

Reviewed by Eric Seidel.

Expose functions to change the focus ring color for Linux Chromium

https://bugs.webkit.org/show_bug.cgi?id=28887

  • rendering/RenderThemeChromiumLinux.cpp: (WebCore::RenderThemeChromiumLinux::RenderThemeChromiumLinux): (WebCore::RenderThemeChromiumLinux::setFocusRingColor): (WebCore::RenderThemeChromiumLinux::platformFocusRingColor):
  • rendering/RenderThemeChromiumLinux.h:
13:38 Changeset [47991] by eric@webkit.org
  • 2 edits in trunk/WebCore

2009-09-02 Dirk Schulze <krit@webkit.org>

Reviewed by Eric Seidel.

SVG Filter premultiplied color support for getImageDate/putImageData
https://bugs.webkit.org/show_bug.cgi?id=27933

Patch to get premultiplied color support for Skia on getImageDate/putImageData.

  • platform/graphics/skia/ImageBufferSkia.cpp: (WebCore::getImageData): (WebCore::putImageData):
13:23 Changeset [47990] by beidson@apple.com
  • 4 edits in trunk/WebCore

Forget to rename openInFrameloader() to open(), which was requested in Darin's review of my last patch.

  • history/CachedFrame.cpp:

(WebCore::CachedFrameBase::restore):
(WebCore::CachedFrame::open):

  • history/CachedFrame.h:
  • history/CachedPage.cpp:

(WebCore::CachedPage::restore):

13:16 Changeset [47989] by beidson@apple.com
  • 7 edits in trunk/WebCore

In http://trac.webkit.org/changeset/47943 I added new header dependencies to FrameLoader.h.
Removing them is a little trouble, but worth it going forward.
The pattern of creating a "CachedFrameBase" as the interface CachedFrame provides to FrameLoader
was suggested by Darin Adler.

Reviewed by Darin Adler.

No new tests. (No change in behavior)

Make CachedFrame inherit privately from CachedFrameBase, which becomes the interface FrameLoader uses:

  • history/CachedFrame.cpp:

(WebCore::CachedFrameBase::CachedFrameBase):
(WebCore::CachedFrameBase::~CachedFrameBase):
(WebCore::CachedFrameBase::restore):
(WebCore::CachedFrame::CachedFrame):
(WebCore::CachedFrame::openInFrameLoader):

  • history/CachedFrame.h:

(WebCore::CachedFrameBase::document):
(WebCore::CachedFrameBase::view):
(WebCore::CachedFrameBase::domWindow):
(WebCore::CachedFrame::create):
(WebCore::CachedFrame::documentLoader):
(WebCore::CachedFrame::mousePressNode):

Call the new ::openInFrameLoader() method on the CachedFrame, as CachedFrameBase is now the only
thing capable of interacting with FrameLoader::open(CachedFrameBase):

  • history/CachedPage.cpp:

(WebCore::CachedPage::restore):

  • history/CachedPage.h:

(WebCore::CachedPage::documentLoader):

Make ::open(CachedFrame) public, and change it to ::open(CachedFrameBase):

  • loader/FrameLoader.cpp:

(WebCore::FrameLoader::open):

  • loader/FrameLoader.h:
13:13 Changeset [47988] by ddkilzer@apple.com
  • 2 edits in trunk/WebKitTools

2009-09-02 Zan Dobersek <zandobersek@gmail.com>

Reviewed by David Kilzer.

Calls exitStatus function from the main package where it is also defined.

  • Scripts/VCSUtils.pm:
13:13 Changeset [47987] by rniwa@webkit.org
  • 3 edits
    3 adds in trunk

WebCore: Cannot pres Enter to escape from bulleted list when <ul> is a child of <li>.
https://bugs.webkit.org/show_bug.cgi?id=24866

Patch by Ryosuke Niwa <rniwa@webkit.org> on 2009-09-02
Reviewed by Eric Seidel.

This patch modifies breakOutOfEmptyListItem to break out of a properly nested list
(a list to break out of is inside another list item).

When the empty list item appears at the end of another list item, WebKit breaks out of
the outer list item, and adds new item below the outer list item.
Otherwise, WebKit breaks out of the the inner list item and inserts new paragraph.

Test: editing/execCommand/break-out-of-empty-list-item.html

  • editing/CompositeEditCommand.cpp:

(WebCore::CompositeEditCommand::breakOutOfEmptyListItem):

LayoutTests: Cannot press Enter to escape from bulleted list when <ul> is a child of <li>.
https://bugs.webkit.org/show_bug.cgi?id=24866

Patch by Ryosuke Niwa <rniwa@webkit.org> on 2009-09-02
Reviewed by Eric Seidel.

This patch adds a test to make sure WebKit break out of an empty list item.

  • editing/execCommand/break-out-of-empty-list-item-expected.txt: Added.
  • editing/execCommand/break-out-of-empty-list-item.html: Added.
  • editing/execCommand/resources/break-out-of-empty-list-item.js: Added.

(pressKey):
(enterAtTarget):
(testBreakOutOfEmptyListItem):

12:55 Changeset [47986] by ap@apple.com
  • 3 edits in trunk/LayoutTests

Reviewed by Brady Eidson.

Fix the test to pass on Tiger.

Tiger's ICU decodes the invalid text differently. Moved it inside the comment - this also
matches the original from the Web better.

  • fast/encoding/japanese-encoding-mix-expected.txt:
  • fast/encoding/japanese-encoding-mix.html:
12:32 Changeset [47985] by beidson@apple.com
  • 5 edits in trunk/WebCore

More partial work towards "Page Cache should support pages with Frames"
https://bugs.webkit.org/show_bug.cgi?id=13631

Reviewed by Sam Weinig.

This patch is primarily about the "CachedFrame tree" and making slightly different decisions
based on whether a CachedFrame is the main frame or a subframe.

  • Store whether a CachedFrame represents the main frame or a subframe.
  • "Chop off the frame tree" for the main frame when creating its CachedFrame, as main frames are reused and need to start new page loads with an empty frame tree.
  • Reattach the frame tree to the main frame when restoring it.
  • open() the cached subframes in their respective loaders when restoring.
  • Properly clear() and destroy() cached subframes.
  • When committing to a new page load, and after caching the previous page, only clear the FrameView for the main frame.

Note that all of above will eventually be necessary as I continue to land this work in pieces,
but doesn't actually change any behavior right now because we still refuse to even *try* to
cache pages with frames.

No new tests. No change in behavior, and current tests continue to pass.

  • history/CachedFrame.cpp:

(WebCore::CachedFrame::CachedFrame):
(WebCore::CachedFrame::restore):
(WebCore::CachedFrame::clear):
(WebCore::CachedFrame::destroy):

  • history/CachedFrame.h:

(WebCore::CachedFrame::isMainFrame):

  • loader/FrameLoader.cpp:

(WebCore::FrameLoader::clear):
(WebCore::FrameLoader::open):

  • loader/FrameLoader.h:
11:32 Changeset [47984] by beidson@apple.com
  • 2 edits in trunk/WebCore

No review, build fix.

http://trac.webkit.org/changeset/47976 broke the build.
Let's try compiling before we land things, shall we?

  • platform/graphics/SimpleFontData.cpp:

(WebCore::SimpleFontData::SimpleFontData): Order the initialization list correctly.

11:28 Changeset [47983] by kevino@webkit.org
  • 11 edits in trunk

waf build fixes for Windows/MSVC and Mac/Snow Leopard.

11:14 Changeset [47982] by benm@google.com
  • 2 edits in trunk/WebCore

Unreviewed build fix for Chromium.

Missed #including V8Proxy.h in WebCore/bindings/v8/V8Binding.cpp.

  • bindings/v8/V8Binding.cpp: add V8Proxy.h as an include
11:03 Changeset [47981] by kevino@webkit.org
  • 2 edits in trunk/WebKit/wx

Reviewed by Kevin Ollivier.

wx port: Call WTF::initializeThreading().
https://bugs.webkit.org/show_bug.cgi?id=28912

10:41 Changeset [47980] by kevino@webkit.org
  • 2 edits in trunk/WebCore

waf build fix, don't define symbols we now grab from other sources.

10:39 Changeset [47979] by kevino@webkit.org
  • 2 edits in trunk/JavaScriptCore

Build fix for building on Windows.

10:34 Changeset [47978] by kevino@webkit.org
  • 3 edits in trunk/WebCore

Build fixes for PluginViewNone.cpp and a wx build fix for PluginView.cpp.

10:32 Changeset [47977] by kevino@webkit.org
  • 2 edits in trunk/WebKit/wx

wx build fix. Use the 2 param KURL constructor.

10:10 Changeset [47976] by yong.li@torchmobile.com
  • 3 edits in trunk/WebCore

2009-09-02 Charles Wei <charles.wei@torchmobile.com.cn>

Reviewed by Eric Seidel.

Fix the build failure of WebKit for Linux/Qt when WML is enabled

No test cases needed since this just fixes the build problem.

  • wml/WMLElement.cpp: (WebCore::WMLElement::create):
  • wml/WMLTableElement.cpp: (WebCore::WMLTableElement::joinSuperflousColumns): (WebCore::WMLTableElement::padWithEmptyColumns):
09:37 Changeset [47975] by yong.li@torchmobile.com
  • 5 edits in trunk/WebCore

2009-09-02 Yong Li <yong.li@torchmobile.com>

Reviewed by Adam Barth.

WINCE PORT: use shared buffer for custom font data
https://bugs.webkit.org/show_bug.cgi?id=27734

Refactored by Joe Mason <joe.mason@torchmobile.com>

  • loader/CachedFont.cpp: add WINCE to platforms using cached custom data
  • platform/graphics/opentype/OpenTypeUtilities.cpp: (WebCore::renameFont): implement for WinCE
  • platform/graphics/opentype/OpenTypeUtilities.h: build fixes
  • platform/graphics/wince/FontCustomPlatformData.cpp: (WebCore::setCustomFontCache): add accessor (WebCore::createFontCustomPlatformData): change param to SharedBuffer
  • platform/graphics/wince/FontCustomPlatformData.h: update function signatures
09:32 Changeset [47974] by yong.li@torchmobile.com
  • 5 edits in trunk/WebCore

2009-09-02 Yong Li <yong.li@torchmobile.com>

Reviewed by Adam Barth.

WINCE PORT: font-related build fixes and minor bugfixes
https://bugs.webkit.org/show_bug.cgi?id=27734

  • platform/graphics/FontCache.h: declare functions added in https://bugs.webkit.org/show_bug.cgi?id=27509
  • platform/graphics/GlyphPageTreeNode.cpp: (WebCore::GlyphPageTreeNode::initializePage): fix ambiguous overload error in min/max on Windows
  • platform/graphics/SimpleFontData.cpp: (WebCore::SimpleFontData::SimpleFontData): initialize member vars
  • platform/graphics/SimpleFontData.h: disable unused members on Wince to save space
08:54 Changeset [47973] by abarth@webkit.org
  • 6 edits in trunk/WebCore

2009-09-02 Jeremy Orlow <jorlow@chromium.org>

Reviewed by Adam Barth.

Another pass at releasing the storage mutex.
https://bugs.webkit.org/show_bug.cgi?id=28904

Another pass at releasing the storage mutex (when applicable) when we finish
executing JavaScript. http://dev.w3.org/html5/spec/Overview.html#storage-mutex
describes the required behavior.

As it turns out, https://bugs.webkit.org/show_bug.cgi?id=28789 didn't solve
the problem correctly. First of all, I missed that events and timeouts go
through a different call path. Second of all, I didn't consider recursion
correctly. Third of all, my check to see if LocalStorage is alive actually
instantiates it if it isn't yet. Fourth, I forgot DOM_STORAGE guards around
it. So I'm reverting that change completely and doing it more cleanly.
Unfortunately, the solution isn't as portable (and thus I've left out the JSC
implementation for now).

NOTE: This change now tracks recursion on function callbacks in V8. I ran this
by Mads Ager and Mike Belshe and they seemed to think adding the checks were
fine. Most callbacks are asynchronous and thus wouldn't be nested. The few
scenareos where you can have nested callbacks probably should have always been
protected by a recursion guard.

Unfortunately, this can't be tested without a hook added to the layout test
controllers since the lock is (by design) not supposed to be observable by
JavaScript. https://bugs.webkit.org/show_bug.cgi?id=28906

  • bindings/js/ScriptController.cpp: (WebCore::ScriptController::evaluate): revert
  • bindings/v8/ScriptController.cpp: (WebCore::ScriptController::evaluate): revert
  • bindings/v8/V8Proxy.cpp: (WebCore::V8Proxy::runScript):

Add releaseStorageMutex call. A small amount of clean up.

(WebCore::V8Proxy::callFunction):

Add releaseStorageMutex call. A small amount of clean up.

(WebCore::V8Proxy::releaseStorageMutex):

Check to see if localStorage has been instantiated. If so, tell it to
unlock all storage areas.

  • bindings/v8/V8Proxy.h:
  • page/PageGroup.h: (WebCore::PageGroup::hasLocalStorage): Make hasLocalStorage public.
08:48 Changeset [47972] by kevino@webkit.org
  • 4 edits in trunk/WebKitTools

Reviewed by Eric Seidel.

Changes needed for build-webkit to support the waf build system for the wx port.

https://bugs.webkit.org/show_bug.cgi?id=27619

08:24 Changeset [47971] by ddkilzer@apple.com
  • 2 edits in trunk/WebKitTools

<http://webkit.org/b/27168> With Subversion 1.6, update-webkit prompts on conflicts

Reviewed by Eric Seidel.

  • Scripts/update-webkit: Added "--accept postpone" to @svnOptions when running with svn-1.6 or newer.
08:24 Changeset [47970] by ddkilzer@apple.com
  • 5 edits in trunk/WebKitTools

Moved svn 1.6 version check into VCSUtils::isSVNVersion16OrNewer()

Reviewed by Eric Seidel.

  • Scripts/VCSUtils.pm: (@EXPORT): Added &isSVNVersion16OrNewer. (svnVersion): Added. Internal method that gets the SVN version and caches it. (isSVNVersion16OrNewer): Added. Method that does the SVN 1.6 version check.
  • Scripts/prepare-ChangeLog: Switched to use new isSVNVersion16OrNewer() method.
  • Scripts/resolve-ChangeLogs: Ditto.
  • Scripts/svn-create-patch: Ditto.
08:24 Changeset [47969] by ddkilzer@apple.com
  • 2 edits in trunk/WebKitTools

Clean up VCSUtils.pm

Reviewed by Eric Seidel.

  • Scripts/VCSUtils.pm: Added proper package statement. Fixed indentation of BEGIN block. Listed each exported method on a line by itself. Added methods to the export list after adding the package statement. Sorted module variables. Moved definiton of $gitRoot next to other module variables.
08:19 Changeset [47968] by abarth@webkit.org
  • 1 edit
    2 adds in trunk/WebCore

2009-09-02 Yong Li <yong.li@torchmobile.com>

Reviewed by Adam Barth.

WINCE PORT: Implement ImageBuffer for WINCE
Implement Image interface with a light class BufferedImage
for rendering ImageBuffer
https://bugs.webkit.org/show_bug.cgi?id=28167

  • platform/graphics/wince/ImageBufferData.h: Added.
  • platform/graphics/wince/ImageBufferWince.cpp: Added.
07:24 Changeset [47967] by vestbo@webkit.org
  • 8 edits in trunk/WebCore

Join the various PluginView destructors into one shared implementation

Reviewed by Eric Seidel.

https://bugs.webkit.org/show_bug.cgi?id=27706

The Qt port used to delete the platform plugin widget as the last step.
Now this is done before cleaning up the script objects and unloading the
plugin package, similar to how the Win port does it.

  • plugins/PluginView.cpp:
  • plugins/PluginViewNone.cpp:
  • plugins/gtk/PluginViewGtk.cpp:
  • plugins/mac/PluginViewMac.cpp:
  • plugins/qt/PluginViewQt.cpp:
  • plugins/win/PluginViewWin.cpp:
06:50 Changeset [47966] by vestbo@webkit.org
  • 8 edits in trunk/WebCore

Share PluginView::init() between ports

Reviewed by Eric Seidel.

https://bugs.webkit.org/show_bug.cgi?id=27706

The port-spesific bits are moved to platformStart(), which now returns
a bool based on the success of the platformStart.

m_hasPendingGeometryChange is now initialized in the initializer list
instead of in PluginViewQt's init()

PluginViewMac (currently used only by the Qt port) used to manually call
stop() in the case where an unsupported drawing or event model was detected.
This was wrong, as the m_status and m_isStarted fields should be used to
guard against doing operations on a plugin in this intermediate state.

  • plugins/PluginView.cpp:
  • plugins/PluginView.h:
  • plugins/PluginViewNone.cpp:
  • plugins/gtk/PluginViewGtk.cpp:
  • plugins/mac/PluginViewMac.cpp:
  • plugins/qt/PluginViewQt.cpp:
  • plugins/win/PluginViewWin.cpp:
06:06 Changeset [47965] by jmalonzo@webkit.org
  • 5 edits
    1 add in trunk

2009-09-02 Jan Michael Alonzo <jmalonzo@webkit.org>

Reviewed by Xan Lopez and Gustavo Noronha.

[Gtk] Implement a WebDataSource for the gtk port
https://bugs.webkit.org/show_bug.cgi?id=24758

Add data source API getters to WebKitWebFrame.
Also add WebKitWebDataSource unit test.

  • tests/testwebdatasource.c: Added.
  • webkit/webkitwebframe.cpp:
  • webkit/webkitwebframe.h:

Add WebKitWebDataSource unit test to the build script

  • GNUmakefile.am:
05:58 Changeset [47964] by jmalonzo@webkit.org
  • 8 edits
    4 adds in trunk

2009-09-02 Jan Michael Alonzo <jmalonzo@webkit.org>

Reviewed by Xan Lopez and Gustavo Noronha.

[Gtk] Implement a WebDataSource for the gtk port
https://bugs.webkit.org/show_bug.cgi?id=24758

Subclass DocumentLoader for Gtk and implement it under the WebKit
namespace. Add WebKitWebDataSource implementation that closely
follows the mac implementation of the API.

  • WebCoreSupport/DocumentLoaderGtk.cpp: Added.
  • WebCoreSupport/DocumentLoaderGtk.h: Added.
  • webkit/webkitwebdatasource.cpp: Added.
  • webkit/webkitwebdatasource.h: Added.

Add WebKitWebDataSource and DocumentLoaderGtk to the build script.

  • GNUmakefile.am:
05:40 Changeset [47963] by vestbo@webkit.org
  • 3 edits in trunk/WebKit/qt

[Qt] Make sure we relayout the page after evaluating JS

Patch by Richard Moore <rich@kde.org> on 2009-09-02
Reviewed by Tor Arne Vestbø.

https://bugs.webkit.org/show_bug.cgi?id=28235

QtWebKit does not always seem to reflow the page when evaluating
javascript. This patch changes the way evaluateJavaScript works to
use the frameloader which ensures that this is done properly.

  • Api/qwebframe.cpp:

(QWebFrame::evaluateJavaScript):

  • tests/qwebframe/tst_qwebframe.cpp:
04:38 Changeset [47962] by benm@google.com
  • 7 edits in trunk

WebCore: https://bugs.webkit.org/show_bug.cgi?id=28737

Patch by Ben Murdoch <benm@google.com> on 2009-09-02
Update JavaScript bindings in JSC and V8 for SQLTransaction to throw exceptions in the right places.

  • bindings/js/JSSQLTransactionCustom.cpp:

(WebCore::JSSQLTransaction::executeSql): Throw an exception if no parameters are passed to executeSql().

  • bindings/v8/V8Binding.cpp:

(WebCore::v8ValueToWebCoreString): Catch an exception if thrown from toString().

  • bindings/v8/custom/V8SQLTransactionCustom.cpp:

(WebCore::CALLBACK_FUNC_DECL): Allow "arrayish objects" to be passed to executeSql.

LayoutTests: https://bugs.webkit.org/show_bug.cgi?id=28737

Patch by Ben Murdoch <benm@google.com> on 2009-09-02
Update the execute-sql-args.html to be agnostic to exception message formatting between JSC and V8 and throw an exception when no parameters are passed to the executeSql() function.

  • storage/execute-sql-args-expected.txt:
  • storage/execute-sql-args.html:
03:16 Changeset [47961] by vestbo@webkit.org
  • 4 edits in trunk

[Qt] Fix layout-test plugins/plugin-javascript-access.html

Rubber-stamped by Simon Hausmann.

  • DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
  • platform/qt/Skipped:
02:59 Changeset [47960] by pfeldman@chromium.org
  • 6 edits in trunk/WebCore

2009-09-01 Pavel Feldman <pfeldman@chromium.org>

Reviewed by Timothy Hatcher.

WebInspector: Migrate Databases tab to InjectedScript /
serialized interaction.

https://bugs.webkit.org/show_bug.cgi?id=28873

  • inspector/InspectorDOMAgent.cpp: (WebCore::InspectorDOMAgent::handleEvent):
  • inspector/front-end/Database.js: (WebInspector.Database.prototype.isDB): (WebInspector.Database.prototype.runWithTableNames.sortingCallback): (WebInspector.Database.prototype.runWithTableNames): (WebInspector.Database.prototype.executeSql): (WebInspector.Database.prototype.executeSql.errorCallback): (WebInspector.Database.prototype.executeSql.queryTransaction):
  • inspector/front-end/DatabaseQueryView.js: (WebInspector.DatabaseQueryView.prototype._enterKeyPressed): (WebInspector.DatabaseQueryView.prototype._queryFinished):
  • inspector/front-end/DatabaseTableView.js: (WebInspector.DatabaseTableView):
  • inspector/front-end/StoragePanel.js: (WebInspector.StoragePanel.prototype.selectDatabase): (WebInspector.StoragePanel.prototype.updateDatabaseTables.withTableNames): (WebInspector.StoragePanel.prototype.updateDatabaseTables): (WebInspector.DatabaseSidebarTreeElement.prototype.onpopulate.withTableNames): (WebInspector.DatabaseSidebarTreeElement.prototype.onpopulate):
02:49 Changeset [47959] by eric@webkit.org
  • 3 edits in trunk/JavaScriptCore

2009-09-02 Norbert Leser <norbert.leser@nokia.com>

Reviewed by Eric Seidel.

Use fastMalloc when neither MMAP nor VIRTUALALLOC are enabled

RegisterFile constructor currently throws #error when both
MMAP and VIRTUALALLOC conditions fail.
On any platform that does not provide these features
(for instance, Symbian),
the fallback should be regular malloc (or fastMalloc).
It is functionally equivalent in this case, even though it may
have certain drawbacks such as lack of dynamic pre-allocation.

  • interpreter/RegisterFile.cpp: (JSC::RegisterFile::~RegisterFile):
  • interpreter/RegisterFile.h: (JSC::RegisterFile::RegisterFile):
02:46 Changeset [47958] by pfeldman@chromium.org
  • 2 edits in trunk/WebCore

2009-09-02 Pavel Feldman <pfeldman@chromium.org>

Not reviewed (trivial followup fix).


Web Inspector: Trivial follow up fix to r47944.

https://bugs.webkit.org/show_bug.cgi?id=28800

  • inspector/InspectorBackend.cpp: (WebCore::InspectorBackend::nodeForId):
02:44 Changeset [47957] by zoltan@webkit.org
  • 24 edits in trunk/LayoutTests

2009-09-02 Renata Hodovan <hodovan.renata@stud.u-szeged.hu>

Reviewed by Ariya Hidayat.

[Qt] Update css2.1/t1202-counter-* and css2.1/t1205* tests after new metrics.
Remove these files from Skipped (and fixed Skipped according to Changeset 47815).

  • platform/qt/Skipped:
  • platform/qt/css2.1/t1202-counter-00-b-expected.txt:
  • platform/qt/css2.1/t1202-counter-01-b-expected.txt:
  • platform/qt/css2.1/t1202-counter-02-b-expected.txt:
  • platform/qt/css2.1/t1202-counter-03-b-expected.txt:
  • platform/qt/css2.1/t1202-counter-04-b-expected.txt:
  • platform/qt/css2.1/t1202-counter-05-b-expected.txt:
  • platform/qt/css2.1/t1202-counter-06-b-expected.txt:
  • platform/qt/css2.1/t1202-counter-07-b-expected.txt:
  • platform/qt/css2.1/t1202-counter-08-b-expected.txt:
  • platform/qt/css2.1/t1202-counter-09-b-expected.txt:
  • platform/qt/css2.1/t1202-counter-11-b-expected.txt:
  • platform/qt/css2.1/t1202-counter-12-b-expected.txt:
  • platform/qt/css2.1/t1202-counter-13-b-expected.txt:
  • platform/qt/css2.1/t1202-counter-14-b-expected.txt:
  • platform/qt/css2.1/t1202-counter-15-b-expected.txt:
  • platform/qt/css2.1/t1202-counter-16-f-expected.txt:
  • platform/qt/css2.1/t1205-c561-list-displ-00-b-expected.txt:
  • platform/qt/css2.1/t1205-c563-list-type-00-b-expected.txt:
  • platform/qt/css2.1/t1205-c563-list-type-01-b-expected.txt:
  • platform/qt/css2.1/t1205-c564-list-img-00-b-g-expected.txt:
  • platform/qt/css2.1/t1205-c565-list-pos-00-b-expected.txt:
  • platform/qt/css2.1/t1205-c566-list-stl-01-c-g-expected.txt:
01:08 Changeset [47956] by eric@webkit.org
  • 5 edits
    2 moves
    1 add in trunk

2009-09-02 Shinichiro Hamaji <hamaji@chromium.org>

Reviewed by Eric Seidel.

js tests should move into jstests subdirectory instead of resources/
https://bugs.webkit.org/show_bug.cgi?id=25880

make-script-wrappers supports both resources and script-tests directories.
run-webkit-tests ignores files in script-tests.
Move script tests of animations to check the new script is working.

  • animations/animation-css-rule-types.html:
  • animations/resources/TEMPLATE.html: Removed.
  • animations/resources/animation-css-rule-types.js: Removed.
  • animations/script-tests: Added.
  • animations/script-tests/TEMPLATE.html: Copied from LayoutTests/animations/resources/TEMPLATE.html.
  • animations/script-tests/animation-css-rule-types.js: Copied from LayoutTests/animations/resources/animation-css-rule-types.js.

2009-09-02 Shinichiro Hamaji <hamaji@chromium.org>

Reviewed by Eric Seidel.

js tests should move into jstests subdirectory instead of resources/
https://bugs.webkit.org/show_bug.cgi?id=25880

make-script-wrappers supports both resources and script-tests directories.
run-webkit-tests ignores files in script-tests.
Move script tests of animations to check the new script is working.

  • Scripts/make-script-test-wrappers:
  • Scripts/run-webkit-tests:
01:01 Changeset [47955] by eric@webkit.org
  • 5 edits in trunk/WebKitTools

2009-09-02 Szabo Carol <carol.szabo@nokia.com>

Reviewed by David Levin.

check-webkit-style uses python from /usr/bin instead of the PATH
https://bugs.webkit.org/show_bug.cgi?id=28225

  • Scripts/bugzilla-tool:
  • Scripts/check-webkit-style:
  • Scripts/run-webkit-unittests:
  • Scripts/update-sources-list.py: Changed the first line from #!/usr/bin/python to #!/usr/bin/env python which causes python to be invoked from the path location returned by "which python" when any of these scripts are launched. these are currently all the python scripts in WebKitTools/Scripts.

09/01/09:

22:03 Changeset [47954] by bweinstein@apple.com
  • 5 edits in trunk

2009-09-01 Brian Weinstein <bweinstein@apple.com>

Reviewed by John Sullivan.

Added a new case to aria-readonly.html to account for textareas instead of just input fields.

  • accessibility/aria-readonly-expected.txt:
  • accessibility/aria-readonly.html:

2009-09-01 Brian Weinstein <bweinstein@apple.com>

Reviewed by John Sullivan.

Fixed a bug in handling of isReadOnly, where textareas were casted to HTMLInputElements.

  • accessibility/AccessibilityRenderObject.cpp: (WebCore::AccessibilityRenderObject::isReadOnly):
19:26 Changeset [47953] by ddkilzer@apple.com
  • 3 edits in trunk/WebKitTools

<http://webkit.org/b/28601> bugzilla-tool post-commits posts commits backwards

Reviewed by Adam Barth.

  • Scripts/modules/scm.py: (Git.commit_ids_from_commitish_arguments): Reverse the list of commits returned from git-rev-list since we always want to post the oldest patches first to bugs.webkit.org.
  • Scripts/modules/scm_unittest.py: (run): Added return statement to return the output now that we want it sometimes. (SVNTestRepository._setup_test_commits): Added a fourth commit so the GitTest.test_commitish_order() test has more commits to work with. (GitTest.test_commitish_order): Added unit test for change to Git.commit_ids_from_commitish_arguments() in scm.py.
17:38 Changeset [47952] by eric@webkit.org
  • 6 edits
    1 add in trunk/WebCore

2009-09-01 Patrick Mueller <Patrick_Mueller@us.ibm.com>

Reviewed by Timothy Hatcher.

https://bugs.webkit.org/show_bug.cgi?id=28047
add utility method to enable logging messages for inspector development

manual test case added (see below)

  • inspector/front-end/ConsoleView.js: (WebInspector.ConsoleMessage.prototype.toMessageElement): (WebInspector.ConsoleMessage.prototype.toString):
  • inspector/front-end/inspector.css:
  • inspector/front-end/inspector.js: (WebInspector.log):
  • manual-tests/inspector/webinspector-log.html: Added.
  • page/Console.cpp: (WebCore::printMessageSourceAndLevelPrefix):
  • page/Console.h: (WebCore::):
17:15 Changeset [47951] by eric@webkit.org
  • 1 edit
    1 move in trunk/LayoutTests

2009-09-01 Eric Seidel <eric@webkit.org>

Rubber stamped by Simon Fraser.

REGRESSION: media/video-source-add-src.html (and other media tests?) crashing/timing-out intermittently
https://bugs.webkit.org/show_bug.cgi?id=28845

Disabling this test on all platforms until
a media expert can have a chance to look at it.
See bug 28845 for further discussion.

  • media/video-source-add-src.html-disabled: Renamed from LayoutTests/media/video-source-add-src.html.
17:06 Changeset [47950] by ap@apple.com
  • 3 edits
    2 adds in trunk

Reviewed by Darin Adler.

https://bugs.webkit.org/show_bug.cgi?id=21990
When a rare EUC-JP character is present, explicitly (and correctly) labelled EUC-JP document
is mistreated as Shift_JIS

<rdar://problem/5828506> Safari should not try auto-detecting if charset information exists

WebKit used to unconditionally auto-detect document encoding if it was one of Japanese ones.
This didn't really match any browser, and did cause problems on some sites.

Our detector is very quick to label an EUC-JP page as Shift_JIS if it notices a byte that
cannot be part of EUC-JP source. However, such bytes do appear in Web pages sometimes,
especially in comments. Also, some valid EUC-JP sequences are not known to it.

Some pages that are fixed by this change:

<rdar://problem/5934750> Encoding in incorrect at http://cocodeparis.blog45.fc2.com/. This
page has incorrect encoding in title, so no browser displays it 100% fine. We now match
Firefox and IE.

<rdar://problem/6007713> The web page http://www.sf.us.emb-japan.go.jp/top.htm is garbled
in Safari. There is a stray byte in comments that was makes charset detector think it can't
be EUC-JP.

<rdar://problem/6965711> A Japanese web page www.nippontv.studiolegend.tv is garbled in
Safari. Many comments on this page are encoded as Shift_JIS - IE magically displays that
fine in View Source mode (it doesn't seem to have an ability to switch between EUC-JP and
Shift_JIS in normal HTML view though).

Test: fast/encoding/japanese-encoding-mix.html

  • loader/TextResourceDecoder.cpp: (WebCore::TextResourceDecoder::decode): Only invoke auto- detection if there is no solid charset information yet.
16:07 Changeset [47949] by ddkilzer@apple.com
  • 4 edits in trunk/WebKitTools

<http://webkit.org/b/28877> Implement bugzilla-tool mark-fixed

Reviewed by David Levin.

The mark-fixed subcommand is for those times when you don't use
bugzilla-tool to commit a patch, but you want to use it to close
the bug with a committed-revision message.

  • Scripts/bugzilla-tool: (bug_comment_from_svn_revision): Added. Extracted from bug_comment_from_commit_text(). (bug_comment_from_commit_text): Extracted bug_comment_from_svn_revision() from this method. (MarkBugFixed.init): Added. (MarkBugFixed._fetch_commit_log): Added. Retrieves the commit log from the last commit if no svn revision is specified, else the commit log for the specified svn revision. (MarkBugFixed._determine_bug_id_and_svn_revision): Added. Attempts to determine the bug id and svn revision if one or both were not defined on the command line. (MarkBugFixed.execute): Added. Adds a comment about the revision that fixed the bug and closes the bug. (BugzillaTool.init): Added mark-fixed subcommand.
  • Scripts/modules/bugzilla.py: (Bugzilla.fetch_title_from_bug): Added. Returns the title of a bug given a bug id.
  • Scripts/modules/scm.py: (SCM.strip_r_from_svn_revision): Added. Utility method to strip the leading 'r' from an svn revision. (SCM.svn_commit_log): Added. Subclasses must override. (SCM.last_svn_commit_log): Added. Subclasses must override. (SVN.svn_commit_log): Added. Returns svn log for a given revision. (SVN.last_svn_commit_log): Added. Uses svnversion to find the last commit in an svn working directory and then runs svn log. (Git.svn_commit_log): Added. Returns svn log for a given revision. (Git.last_svn_commit_log): Added. Runs git-svn-log with a limit of one log message.
13:18 Changeset [47948] by dglazkov@chromium.org
  • 3 edits in trunk/LayoutTests

2009-09-01 Victor Wang <victorw@chromium.org>

Reviewed by Darin Adler.

https://bugs.webkit.org/show_bug.cgi?id=28879

Fix fast/dom/StyleSheet/ownerNode-lifetime-2.html
The problem is WebKit <link> nodes will only have a
style sheet created if they have href attribute.
The src attribute to non-existant css doesn't work.
Change src to href to make this test works in WebKit.

Also update the expectation result.

  • fast/dom/StyleSheet/ownerNode-lifetime-2-expected.txt:
  • fast/dom/StyleSheet/ownerNode-lifetime-2.html:
13:16 Changeset [47947] by jianli@chromium.org
  • 4 edits
    1 add in trunk/WebCore

[V8] FileList cannot be accessed via index in Chromium.
https://bugs.webkit.org/show_bug.cgi?id=28883

Reviewed by Dimitri Glazkov.

Tested by clipboard-file-access.html.

  • WebCore.gypi:
  • bindings/v8/V8DOMWrapper.cpp:

(WebCore::V8DOMWrapper::getTemplate):

  • bindings/v8/custom/V8CustomBinding.h:
  • bindings/v8/custom/V8FileListCustom.cpp: Added.
12:52 Changeset [47946] by jhoneycutt@apple.com
  • 9 edits in trunk/WebCore

Remove the workaround added in r47316.

Reviewed by Sam Weinig.

  • accessibility/AXObjectCache.cpp:

(WebCore::AXObjectCache::AXObjectCache):

  • accessibility/AXObjectCache.h:

Remove the Document argument to the AXObjectCache constructor and the
m_document member variable, and replace the no-parameter
handleFocusedUIElementChanged() with the two-parameter GTK function.
(WebCore::AXObjectCache::handleFocusedUIElementChanged):

  • accessibility/chromium/AXObjectCacheChromium.cpp:

(WebCore::AXObjectCache::handleFocusedUIElementChanged):

  • accessibility/gtk/AXObjectCacheAtk.cpp:

(WebCore::AXObjectCache::handleFocusedUIElementChanged):

  • accessibility/gtk/AccessibilityObjectWrapperAtk.cpp:

(fallbackObject):
Don't pass a null Document when constructing the AXObjectCache.

  • accessibility/mac/AXObjectCacheMac.mm:

(WebCore::AXObjectCache::handleFocusedUIElementChanged):

  • accessibility/win/AXObjectCacheWin.cpp:

(WebCore::AXObjectCache::handleFocusedUIElementChanged):
Get the document from the RenderObject; m_document was removed.

  • dom/Document.cpp:

(WebCore::Document::axObjectCache):
Don't pass the Document when constructing the AXObjectCache.
(WebCore::Document::setFocusedNode):
Pass the old and new focused RenderObjects when calling
handleFocusedUIElementChanged().

11:42 Changeset [47945] by ddkilzer@apple.com
  • 2 edits in trunk/WebKitTools

<http://webkit.org/b/28880> svn-apply --force doesn't actually work

Reviewed by Brady Eidson.

  • Scripts/svn-apply: (applyPatch): Add "--force" to $options arrayref if $force is set.
11:17 Changeset [47944] by pfeldman@chromium.org
  • 3 edits in trunk/WebCore

2009-09-01 Yury Semikhatsky <yurys@chromium.org>

Reviewed by Timothy Hatcher.

Check that InspectorDOMAgent and InspectorFrontend are not null
before accessing them in InspectorBackend. They may be null if
InspectorController is closed before an asynchronous operation is
started.

https://bugs.webkit.org/show_bug.cgi?id=28800

  • inspector/InspectorBackend.cpp: (WebCore::InspectorBackend::getCookies): (WebCore::InspectorBackend::getChildNodes): (WebCore::InspectorBackend::setAttribute): (WebCore::InspectorBackend::removeAttribute): (WebCore::InspectorBackend::setTextNodeValue): (WebCore::InspectorBackend::highlight): (WebCore::InspectorBackend::nodeForId): (WebCore::InspectorBackend::pushNodePathToFrontend): (WebCore::InspectorBackend::addNodesToSearchResult): (WebCore::InspectorBackend::selectDatabase): (WebCore::InspectorBackend::selectDOMStorage): (WebCore::InspectorBackend::inspectorDOMAgent): (WebCore::InspectorBackend::inspectorFrontend):
  • inspector/InspectorBackend.h:
11:13 Changeset [47943] by beidson@apple.com
  • 8 edits in trunk/WebCore

Page Cache should support pages with frames
https://bugs.webkit.org/show_bug.cgi?id=13631

Reviewed by Sam Weinig.

Things learned in the full task that this patch addresses:

  • The notions of clearing a CachedFrame and destroying a CachedFrame were two concepts that needed to be managed separately.
  • Once we start restoring pages with multiple CachedFrames, we'll have to be calling FrameLoader::open(CachedFrame&) multiple times with different CachedFrames, and this process will be guided by the CachedFrames themselves.

No new tests. (No change in behavior, current layout tests pass)

  • history/CachedFrame.cpp:

(WebCore::CachedFrame::CachedFrame):
(WebCore::CachedFrame::~CachedFrame):
(WebCore::CachedFrame::clear): Only clear pointers out.
(WebCore::CachedFrame::destroy): Perform the destructive cleanup work here - Things that shouldn't

occur simply by navigating back to a page with a CachedFrame.

  • history/CachedFrame.h:
  • history/CachedPage.cpp:

(WebCore::CachedPage::~CachedPage):
(WebCore::CachedPage::restore): Call clear() after restoring.
(WebCore::CachedPage::clear):
(WebCore::CachedPage::destroy):

  • history/CachedPage.h:
  • history/PageCache.cpp:

(WebCore::PageCache::releaseAutoreleasedPagesNow): destroy() CachedPages that are being pruned,

instead of clearing them.

  • loader/FrameLoader.cpp:

(WebCore::FrameLoader::commitProvisionalLoad): Let the CachedPage clear() itself after restoring.
(WebCore::FrameLoader::open): Don't ::open() the CachedFrame, as the CachedPage does that now.

  • loader/FrameLoader.h: Add the friending so CachedPage and CachedFrame can better guide the process. Even though CachedFrame::restore() isn't used for now, it will be soon.
11:06 Coding Style Guidelines edited by berkeley@ptc.com
(diff)
11:00 Changeset [47942] by benm@google.com
  • 3 edits in trunk/WebCore

Add a "lowMemoryNotification" method to the V8 ScriptController class.
This is needed in order to allow the host application (Android Web browser in this case)
to call the v8::V8::LowMemoryNotification() method added to V8 in
http://code.google.com/p/v8/source/detail?spec=svn2777&r=2725

Patch by Andrei Popescu <andreip@google.com> on 2009-09-01
Reviewed by David Levin.

https://bugs.webkit.org/show_bug.cgi?id=28776

  • bindings/v8/ScriptController.cpp:

(WebCore::ScriptController::lowMemoryNotification):

  • bindings/v8/ScriptController.h:
09:39 Changeset [47941] by abarth@webkit.org
  • 9 edits
    1 add in trunk/WebCore

2009-09-01 Fumitoshi Ukai <ukai@chromium.org>

Reviewed by Adam Barth.

V8 Bindings for WebSocket API.
https://bugs.webkit.org/show_bug.cgi?id=28844

  • bindings/v8/DOMObjectsInclude.h:
  • bindings/v8/DerivedSourcesAllInOne.cpp:
  • bindings/v8/V8DOMWrapper.cpp: (WebCore::V8DOMWrapper::getTemplate): (WebCore::V8DOMWrapper::convertToV8Object): (WebCore::V8DOMWrapper::convertEventTargetToV8Object):
  • bindings/v8/V8DOMWrapper.h:
  • bindings/v8/V8Index.cpp:
  • bindings/v8/V8Index.h:
  • bindings/v8/WorkerContextExecutionProxy.cpp:
  • bindings/v8/custom/V8CustomBinding.h:
  • bindings/v8/custom/V8WebSocketCustom.cpp: Added.
09:35 Changeset [47940] by aroben@apple.com
  • 4 edits in trunk/WebKit/win

Post a notification when a WebView is closing

This is useful for situations where some code wants to react to the
WebView closing but isn't in charge of the WebView and can't otherwise
be notified. This notification could replace
IWebUIDelegatePrivate::webViewClosing if we decide that the
notification system isn't too burdensome.

<http://webkit.org/b/28875>

Reviewed by John Sullivan.

  • Interfaces/IWebViewPrivate.idl: Added the

WebViewWillCloseNotification string constant.

  • WebKit.vcproj/WebKit.vcproj: Link against comsuppw.lib in all

configurations so that we can use _bstr_t.

  • WebView.cpp:

(WebView::close): Post the WebViewWillCloseNotification.

09:17 Changeset [47939] by pkasting@chromium.org
  • 2 edits in trunk/WebCore

Unreviewed (build fix).

Missed a "." -> "->" conversion.

  • platform/graphics/qt/ImageDecoderQt.cpp:

(WebCore::ImageDecoderQt::ReadContext::ReadContext):

09:12 Changeset [47938] by pkasting@chromium.org
  • 3 edits in trunk/WebCore

Unreviewed (build fix).

Try to fix Qt build, take 2. The Qt code had what looks like a
broken virtual function override.

  • platform/graphics/qt/ImageDecoderQt.cpp:

(WebCore::ImageDecoderQt::ReadContext::ReadContext):
(WebCore::ImageDecoderQt::setData):

  • platform/graphics/qt/ImageDecoderQt.h:
09:09 Changeset [47937] by abarth@webkit.org
  • 2 edits in trunk/WebCore

2009-09-01 Adam Barth <abarth@webkit.org>

Reviewed by Sam Weinig.

Uninitialized variable in XSSAuditor
https://bugs.webkit.org/show_bug.cgi?id=28874

One ought to initialize variables before use even when it doesn't
matter. No behavior change.

  • page/XSSAuditor.h: (WebCore::XSSAuditor::CachingURLCanonicalizer::CachingURLCanonicalizer):
08:57 Changeset [47936] by pkasting@chromium.org
  • 2 edits in trunk/WebCore

Unreviewed (build fix).

Try to fix Qt build.

  • platform/graphics/ImageSource.cpp:
08:50 Changeset [47935] by pkasting@chromium.org
  • 4 edits in trunk/WebCore

https://bugs.webkit.org/show_bug.cgi?id=27965
Eliminate most of ImageSourceQt.cpp in favor of ImageSource.cpp.

Reviewed by Eric Seidel.

  • WebCore.pro:
  • platform/graphics/ImageSource.cpp:

(WebCore::ImageSource::frameHasAlphaAtIndex):

  • platform/graphics/qt/ImageSourceQt.cpp:

(WebCore::ImageSource::createFrameAtIndex):
(WebCore::ImageSource::frameHasAlphaAtIndex):
(WebCore::ImageSource::frameIsCompleteAtIndex):

07:58 Changeset [47934] by pfeldman@chromium.org
  • 4 edits in trunk/WebCore

2009-09-01 Pavel Feldman <pfeldman@chromium.org>

Reviewed by Timothy Hatcher.

WebInspector: Make InjectedScript self-contained (move necessary
methods from utilities.js into InjectedScript).

https://bugs.webkit.org/show_bug.cgi?id=28871

  • inspector/front-end/ElementsPanel.js: (WebInspector.ElementsPanel.prototype.generateStylesheet):
  • inspector/front-end/InjectedScript.js: (InjectedScript._ensureCommandLineAPIInstalled.inspectObject): (InjectedScript._ensureCommandLineAPIInstalled): (InjectedScript.pushNodeToFrontend): (InjectedScript.createProxyObject): (Object.type): (Object.hasProperties): (Object.className): (String.prototype.escapeCharacters):
  • inspector/front-end/utilities.js:
07:41 Changeset [47933] by pfeldman@chromium.org
  • 2 edits in trunk/WebCore

2009-09-01 Pavel Feldman <pfeldman@chromium.org>

Reviewed by Timothy Hatcher.

WebInspector: console log message repeat count double.

https://bugs.webkit.org/show_bug.cgi?id=28856

  • inspector/front-end/InjectedScript.js: (InjectedScript._evaluateOn):
07:11 Changeset [47932] by ddkilzer@apple.com
  • 2 edits in trunk/WebKitTools

<http://webkit.org/b/28725> resolve-ChangeLogs: determineVCSRoot() returns incorrect repository root during git filter-branch

Reviewed by Adam Roben.

When git-filter-branch has been invoked to rewrite ChangeLog
files on series of git commits, it changes directories into
.git-rewrite/t before re-running resolve-ChangeLogs. This
causes determineVCSRoot() in VCSUtils.pm to return
".git-rewrite/t", which causes that path to be prepended to all
ChangeLog paths, which results in an error like this:

error: pathspec '.git-rewrite/t/ChangeLog' did not match any file(s) known to git.
Died at WebKitTools/Scripts/resolve-ChangeLogs line 376.

The correct way to fix this is not to try to find the repository
root when invoked by git-filter-branch.

  • Scripts/resolve-ChangeLogs: If isInGitFilterBranch() is true, set $relativePath to '.' instead of calling chdirReturningRelativePath(determineVCSRoot()). (isInGitFilterBranch): Added. Checks for the existence of the MAPPED_PREVIOUS_COMMIT environment variable.
06:06 Changeset [47931] by kov@webkit.org
  • 4 edits in trunk/WebKit/gtk

2009-09-01 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk>

Reviewed by Xan Lopez.

Avoid causing unexpected issues with window sizing when sites
disable scrollbars completely.

  • WebCoreSupport/ChromeClientGtk.cpp: (WebKit::ChromeClient::scrollbarsModeDidChange):
  • tests/testwindow.c: (test_webkit_window_scrollbar_policy):
  • webkit/webkitwebframe.cpp: (webkit_web_frame_class_init):
05:50 Changeset [47930] by xan@webkit.org
  • 5 edits
    1 add in trunk

LayoutTests:

2009-09-01 Xan Lopez <xlopez@igalia.com>

Reviewed by Gustavo Noronha.

Add support for Fedora distros in the http tests
https://bugs.webkit.org/show_bug.cgi?id=28263

Add httpd conf file for Fedora.

  • http/conf/fedora-httpd.conf: Added.

WebKitTools:

2009-09-01 Xan Lopez <xlopez@igalia.com>

Reviewed by Gustavo Noronha.

Add support for Fedora distros in the http tests
https://bugs.webkit.org/show_bug.cgi?id=28263

Add detection code for Fedora distribution, and use the proper
httpd conf file when needed.

  • Scripts/run-webkit-httpd:
  • Scripts/run-webkit-tests:
  • Scripts/webkitdirs.pm:
05:19 Changeset [47929] by jmalonzo@webkit.org
  • 6 edits
    3 adds in trunk

2009-09-01 Jan Michael Alonzo <jmalonzo@webkit.org>

Reviewed by Xan Lopez and Gustavo Noronha.

[Gtk] Implement a WebDataSource for the gtk port
https://bugs.webkit.org/show_bug.cgi?id=24758

Implement WebKitWebResource for the resource-related API for
WebKitWebDataSource.

  • GNUmakefile.am:

Add WebKitWebResource to the build script.

  • GNUmakefile.am:
05:05 Changeset [47928] by jmalonzo@webkit.org
  • 5 edits in trunk

2009-09-01 Jan Michael Alonzo <jmalonzo@webkit.org>

Reviewed by Eric Seidel.

[Gtk] DRT needs implementation of overridePreference
https://bugs.webkit.org/show_bug.cgi?id=28830

Unskip tests that are now working.

  • platform/gtk/Skipped:

Implement overridePreference.

  • DumpRenderTree/gtk/DumpRenderTree.cpp: (resetDefaultsToConsistentValues): (setDefaultsToConsistentStateValuesForTesting): (runTest):
  • DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: (toWebSettingKey): (LayoutTestController::overridePreference):
04:55 Changeset [47927] by jmalonzo@webkit.org
  • 4 edits in trunk

2009-09-01 Jan Michael Alonzo <jmalonzo@webkit.org>

Reviewed by Mark Rowe.

[Gtk] Fix DATALIST build
https://bugs.webkit.org/show_bug.cgi?id=28826

Unskip now working fast/forms/input-list.html.

  • platform/gtk/Skipped:

Add ENABLE_DATALIST to FEATURE_DEFINES.

  • GNUmakefile.am:
04:11 Changeset [47926] by eric@webkit.org
  • 3 edits in trunk/WebCore

2009-09-01 Cameron McCormack <cam@mcc.id.au>

Reviewed by Eric Seidel.

SVGSVGElement suspend methods argument/return types incorrect
https://bugs.webkit.org/show_bug.cgi?id=28860

Also fix the style of the argument names.

  • svg/SVGSVGElement.cpp: (WebCore::SVGSVGElement::suspendRedraw): (WebCore::SVGSVGElement::unsuspendRedraw):
  • svg/SVGSVGElement.h:
04:01 Changeset [47925] by eric@webkit.org
  • 3 edits in trunk/WebCore

2009-09-01 Roland Steiner <rolandsteiner@google.com>

Reviewed by Eric Seidel.

Fix bug 28808: [Skia] Fix flaky layout test svg/dynamic-updates/SVGClipPathElement-dom-clipPathUnits-attr.html [DEBUG]
(https://bugs.webkit.org/show_bug.cgi?id=28808)

The bug fix adds an explicit check for the validity of the path when it is
transformed into local coordinates.
This was the most suitable place I could find that didn't cause an inordinate
amount of refactoring. Also, it seems pertinent and may also catch other cases
that are not limited to clipping paths.

TEST: existing svg/dynamic-updates/SVGClipPathElement-dom-clipPathUnits-attr.html

  • platform/graphics/skia/GraphicsContextSkia.cpp: make isPathSkiaSafe accessible (WebCore::):
  • platform/graphics/skia/PlatformContextSkia.cpp: (PlatformContextSkia::currentPathInLocalCoordinates): check Skia safety of path
03:54 Changeset [47924] by xan@webkit.org
  • 4 edits in trunk/WebKit/gtk

2009-09-01 Xan Lopez <xlopez@igalia.com>

Reviewed by Jan Alonzo.

https://bugs.webkit.org/show_bug.cgi?id=28842
[GTK] Do not emit extra FINISHED load-status signals

Do not emit FINISHED load-status signals on our own, just rely on
WebCore for that. Avoids emitting FINISHED several times when we
fail to load a URI.

  • WebCoreSupport/FrameLoaderClientGtk.cpp: (WebKit::FrameLoaderClient::dispatchDidFailProvisionalLoad): (WebKit::FrameLoaderClient::dispatchDidFailLoad):
  • tests/testloading.c: (web_loading_fixture_setup): (web_loading_fixture_teardown): (load_finished_cb): (status_changed_cb): (test_loading_status): (load_error_status_changed_cb): (load_error_cb): (test_loading_error): (main):
  • webkit/webkitwebframe.h:
03:24 Changeset [47923] by eric@webkit.org
  • 3 edits in trunk/WebKitTools

2009-09-01 Joseph Pecoraro <joepeck@webkit.org>

<http://webkit.org/b/28623> svn-[un]apply should change directories to the repository root before [un]applying

Reviewed by Eric Seidel.

Jump back and forth between the repository root directory (to apply) and the
directory the script was run from (to find the patch).

  • Scripts/svn-apply:
  • Scripts/svn-unapply:
03:14 Changeset [47922] by eric@webkit.org
  • 2 edits in trunk/WebCore

2009-09-01 Yusuke Sato <yusukes@chromium.org>

Reviewed by Eric Seidel.

[Chromium] Combining Diacritical Marks (U+0300..) are not handled correctly.
https://bugs.webkit.org/show_bug.cgi?id=28742

Normalize (NFC) a TextRun when the run contains combining diacritical marks
so that Skia can pick a correct glyph without relying on GSUB table in a font.
This change is for Chromium Linux.

  • platform/graphics/chromium/FontLinux.cpp: (WebCore::TextRunWalker::getTextRun): Added. (WebCore::TextRunWalker::getNormalizedTextRun): Added.

This function is called when the run contains U+0300..U+036F and converts characters in
the run to the combined form (NFC) using ICU.

03:05 Changeset [47921] by eric@webkit.org
  • 5 edits in trunk/WebCore

2009-09-01 Vitaly Repeshko <vitalyr@chromium.org>

Reviewed by Dimitri Glazkov.

[V8] Inlined fast paths of V8DOMWrapper::convertNodeToV8Object and
V8Proxy::createWrapperFromCache.
https://bugs.webkit.org/show_bug.cgi?id=28848

  • bindings/v8/V8DOMWrapper.cpp: (WebCore::V8DOMWrapper::convertDocumentToV8Object): Special function for document wrapping. (WebCore::V8DOMWrapper::convertNewNodeToV8Object): Slow case of wrapping a new node.
  • bindings/v8/V8DOMWrapper.h: (WebCore::V8DOMWrapper::convertNodeToV8Object): Extracted fast path from V8DOMWrapper.cpp.
  • bindings/v8/V8Proxy.cpp: (WebCore::V8Proxy::createWrapperFromCacheSlowCase):
  • bindings/v8/V8Proxy.h: (WebCore::V8Proxy::createWrapperFromCache): Extracted fast path from V8Proxy.cpp.
02:34 Changeset [47920] by eric@webkit.org
  • 2 edits in trunk/WebCore

2009-09-01 Eric Seidel <eric@webkit.org>

No review, windows build fix only. Clearly I should have read the patch closer. :(

Allow excluding certain plugins from loading
https://bugs.webkit.org/show_bug.cgi?id=28677

  • plugins/PluginDatabase.cpp: (WebCore::PluginDatabase::refresh):
02:31 Changeset [47919] by pfeldman@chromium.org
  • 9 edits in trunk/WebCore

2009-09-01 Alexander Pavlov <apavlov@chromium.org>

Reviewed by Timothy Hatcher.

Implement conditional breakpoints in the Web Inspector backend
and add frontend JavaScript stubs.
https://bugs.webkit.org/show_bug.cgi?id=28846

  • inspector/InspectorBackend.cpp: (WebCore::InspectorBackend::profiles): (WebCore::InspectorBackend::addBreakpoint): (WebCore::InspectorBackend::updateBreakpoint):
  • inspector/InspectorBackend.h:
  • inspector/InspectorBackend.idl:
  • inspector/JavaScriptDebugServer.cpp: (WebCore::JavaScriptDebugServer::BreakpointInfo::condition): (WebCore::JavaScriptDebugServer::BreakpointInfo::setCondition): (WebCore::JavaScriptDebugServer::addBreakpoint): (WebCore::JavaScriptDebugServer::breakpointInfo): (WebCore::JavaScriptDebugServer::updateBreakpoint): (WebCore::JavaScriptDebugServer::updateBreakpointInfo): (WebCore::JavaScriptDebugServer::removeBreakpoint): (WebCore::JavaScriptDebugServer::hasBreakpoint): (WebCore::JavaScriptDebugServer::clearBreakpoints):
  • inspector/JavaScriptDebugServer.h: (WebCore::JavaScriptDebugServer::BreakpointInfo::BreakpointInfo):
  • inspector/front-end/Breakpoint.js: (WebInspector.Breakpoint): (WebInspector.Breakpoint.prototype.get id): (WebInspector.Breakpoint.prototype.get condition): (WebInspector.Breakpoint.prototype.set condition):
  • inspector/front-end/BreakpointsSidebarPane.js: (WebInspector.BreakpointsSidebarPane.prototype.addBreakpoint): (WebInspector.BreakpointsSidebarPane.prototype._breakpointEnableChanged):
  • inspector/front-end/ScriptsPanel.js: (WebInspector.ScriptsPanel.prototype.addScript):
02:13 Changeset [47918] by eric@webkit.org
  • 3 edits
    1 add in trunk/WebCore

2009-09-01 Marius Renn <damarvy@gmail.com>

Reviewed by Eric Seidel.

Added delegate to PluginDatabase to disable loading certain plugins.
This is useful when you want to load from the standard plugin
directory, but want to exclude certain plugins. Plugins may be
excluded early on by name, so that no plugin code is executed (which
may have caused a crash).

  • plugins/PluginDatabase.cpp: (WebCore::PluginDatabase::PluginDatabase): (WebCore::PluginDatabase::refresh):
  • plugins/PluginDatabase.h: (WebCore::PluginDatabase::setClient):
  • plugins/PluginDatabaseClient.h: Added. (WebCore::PluginDatabaseClient::~PluginDatabaseClient):

08/31/09:

20:45 Changeset [47917] by beidson@apple.com
  • 2 edits in trunk/WebCore

Rubberstamped by Sam Weinig

  • WebCore.base.exp: Removed an unneeded symbol export.
18:34 Changeset [47916] by kov@webkit.org
  • 7 edits in trunk

WebCore

2009-08-31 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk>

Reviewed by Alexey Proskuryakov.

https://bugs.webkit.org/show_bug.cgi?id=28466
When downloading a file, two GET requests are sent to the HTTP server

Can't think of a way to test this.

  • loader/ResourceLoader.cpp: (WebCore::ResourceLoader::releaseResources): only nullify the handle's client when it is the ResourceLoader, to support the case in which a download API uses a different client

WebKit/gtk

2009-08-31 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk>

Reviewed by Alexey Proskuryakov.

https://bugs.webkit.org/show_bug.cgi?id=28466
When downloading a file, two GET requests are sent to the HTTP server

Support reusing a connection that is already ongoing, instead of
cancelling the connection and creating another.

  • WebCoreSupport/FrameLoaderClientGtk.cpp: (WebKit::FrameLoaderClient::download):
  • webkit/webkitdownload.cpp: (webkit_download_new_with_handle): (webkit_download_start):
  • webkit/webkitprivate.h:
  • webkit/webkitwebview.cpp: (webkit_web_view_request_download):
16:44 Changeset [47915] by bweinstein@apple.com
  • 5 edits in trunk

2009-08-31 Brian Weinstein <bweinstein@apple.com>

Reviewed by Darin Adler.

Text Fields and Text Areas are reported as read-only by inspect32.exe.
https://bugs.webkit.org/show_bug.cgi?id=28854

Added another case in AcccessibiltyRenderObject::isReadOnly to test if text fields
and texts areas are read-only.

  • accessibility/AccessibilityRenderObject.cpp: (WebCore::AccessibilityRenderObject::isReadOnly): (WebCore::AccessibilityRenderObject::canSetValueAttribute):

2009-08-31 Brian Weinstein <bweinstein@apple.com>

Reviewed by Darin Adler.

Added a case to check for readonly and non-readonly set through the readonly attribute
instead of through aria-readonly. This tests our new case in AccessibilityRenderObject::isReadOnly.

  • accessibility/aria-readonly-expected.txt:
  • accessibility/aria-readonly.html:
16:40 Changeset [47914] by atwilson@chromium.org
  • 12 edits in trunk

WebCore: fast/workers/worker-lifecycle.html crashes intermittently on snowleopard
https://bugs.webkit.org/show_bug.cgi?id=28795

Reviewed by Alexey Proskuryakov.

Changed WorkerContext::close() to no longer stop the worker thread.
It is now the parent's responsibility to shut down the worker thread when it is notified that the context is closed.

  • workers/AbstractWorker.cpp:

Updated dispatchScriptErrorEvent to pass-through the "handled" value from dispatchEvent(), to allow event handlers added with addEventListener() to mark events as handled.

  • workers/DefaultSharedWorkerRepository.cpp:

(WebCore::SharedWorkerProxy::workerContextClosed):
Now shuts down the worker thread when notified that the context is closed.
(WebCore::SharedWorkerProxy::close):
Now handles being invoked when the context is already in the process of shutting down.

  • workers/WorkerContext.cpp:

(WebCore::WorkerContext::close):
No longer calls WorkerThread::stop() to avoid race conditions with shutting down the thread while the parent is still interacting with it.

  • workers/WorkerContext.h:

(WebCore::WorkerContext::isClosing):
isClosing() is now public so WorkerRunLoop can call it to determine whether to process tasks.

  • workers/WorkerMessagingProxy.cpp:

(WebCore::WorkerExceptionTask::performTask):
Error events should still be delivered even if the worker thread is closing.
Also fixed problem where error events were not dispatched if error handlers were added via addEventListener().
(WebCore::WorkerTerminateTask::create):
New task that invokes terminateWorkerContext() on the parent thread when the context is closed.
(WebCore::WorkerTerminateTask::WorkerTerminateTask):
(WebCore::WorkerTerminateTask::performTask):
(WebCore::WorkerMessagingProxy::workerContextClosed):
Now fires off a WorkerTerminateTask to shut down the thread when the context is closed.

  • workers/WorkerMessagingProxy.h:

WorkerMessagingProxy now overrides workerContextClosed().

  • workers/WorkerRunLoop.cpp:

(WebCore::WorkerRunLoop::runInMode):
Now drops tasks on the floor if the WorkerContext is closing.

LayoutTests: fast/workers/worker-lifecycle.html crashes intermittently on snowleopard
https://bugs.webkit.org/show_bug.cgi?id=28795

Reviewed by Alexey Proskuryakov.

Removed worker-lifecycle.html from the list of Skipped tests since it no longer crashes.
Added test for case where we capture script errors via an event listener added via addEventListener().

  • fast/workers/worker-script-error.html
  • fast/workers/worker-script-error-expected.txt

Added testScriptErrorAddEventListener to test catching script errors via addEventListener().

  • platform/mac-snowleopard/Skipped:
16:36 Changeset [47913] by bdakin@apple.com
  • 2 edits in trunk/LayoutTests

Rubber-stamped by Mark Rowe.

Meant to update these results with my earlier check-in, but I
forgot!

  • platform/mac/fast/inspector/style-expected.txt:
16:30 Changeset [47912] by ojan@chromium.org
  • 4 edits in trunk/WebCore

Not reviewed (build fix)

Build fix for Chromium to match r49707.

  • bindings/v8/ScriptCallFrame.cpp:

(WebCore::ScriptCallFrame::ScriptCallFrame):

  • bindings/v8/ScriptSourceCode.h:

(WebCore::ScriptSourceCode::ScriptSourceCode):

  • platform/KURLGoogle.cpp:

(WebCore::KURL::KURL):
(WebCore::blankURL):

16:01 Changeset [47911] by ap@apple.com
  • 2 edits in trunk/WebKit/gtk

Gtk build fix.

This is also a behavior fix, at least on debug builds - a caller of
webkit_web_history_item_new_with_data() could pass an URL that could cause an assertion
failure in KURL.

  • webkit/webkitwebhistoryitem.cpp: (webkit_web_history_item_new_with_data): Pass a null KURL as base to parse properly.
15:49 Changeset [47910] by ap@apple.com
  • 3 edits
    2 adds in trunk

Reviewed by Darin Adler.

https://bugs.webkit.org/show_bug.cgi?id=28858
Element.baseURI parses xml:base attribute incorrectly

Test: fast/dom/base-attribute-parsing.xhtml

  • dom/Element.cpp: (WebCore::Element::baseURI): Avoid assertion failure (no change in release mode).
15:27 Changeset [47909] by dglazkov@chromium.org
  • 4 edits in trunk

WebCore:

2009-08-31 Dimitri Glazkov <dglazkov@chromium.org>

Reverting http://trac.webkit.org/changeset/47904, because it caused
layout test failure.

Test: fast/dom/Window/new-window-opener.html

WebKit/mac:

2009-08-31 Dimitri Glazkov <dglazkov@chromium.org>

Reverting http://trac.webkit.org/changeset/47904, because it caused
layout test failure.

15:09 Changeset [47908] by mrowe@apple.com
  • 2 edits
    2 deletes in trunk/WebKit

Remove WebViewEditingContextMenu.nib and WebViewEditingContextMenuOld.nib.

Rubber-stamped by Beth Dakin.

We stopped using these back in r18592 when we switched over to context menus
driven by WebCore.

  • English.lproj/WebViewEditingContextMenu.nib/classes.nib: Removed.
  • English.lproj/WebViewEditingContextMenu.nib/info.nib: Removed.
  • English.lproj/WebViewEditingContextMenu.nib/objects.nib: Removed.
  • English.lproj/WebViewEditingContextMenuOld.nib/classes.nib: Removed.
  • English.lproj/WebViewEditingContextMenuOld.nib/info.nib: Removed.
  • English.lproj/WebViewEditingContextMenuOld.nib/objects.nib: Removed.
  • WebKit.xcodeproj/project.pbxproj:
15:07 Changeset [47907] by ap@apple.com
  • 50 edits in trunk

Reviewed by Darin Adler.

https://bugs.webkit.org/show_bug.cgi?id=28852
Rename KURL single argument constructor to avoid confusion

  • platform/KURL.h: The constructor that used to be single argument should now be invoked as KURL(ParsedURLString, myString).
15:05 Changeset [47906] by bdakin@apple.com
  • 22 edits
    2 deletes in trunk

WebCore: Fix for https://bugs.webkit.org/show_bug.cgi?id=28635 [CSS3
Backgrounds and Borders] Add support for 2-keyword values for
background-repeat

Reviewed by Darin Adler.

This patch allows background-repeat to take two values by making
background-repeat just like background-position internally. There
is a little extra legwork for background-repeat because the spec
indicates that its computed value should be equivalent to how it
was specified. I keep track of the specified thing by setting the
implicit flag whenever background-repeat is defined with only one
value (since internally, we now store this as 2 values.)

Here we can't access the implicit flag, so for backwards-
compatibility's sake, we always return one value when that makes
sense.

  • css/CSSComputedStyleDeclaration.cpp:

(WebCore::fillRepeatToCSSValue):
(WebCore::CSSComputedStyleDeclaration::getPropertyCSSValue):

Here we check for the implicit flag and return one value when it is
set.

  • css/CSSMutableStyleDeclaration.cpp:

(WebCore::CSSMutableStyleDeclaration::getPropertyValue):
(WebCore::CSSMutableStyleDeclaration::getLayeredShorthandValue):
(WebCore::CSSMutableStyleDeclaration::cssText):

Add support for CSSPropertyBackgroundRepeatX and
CSSPropertyBackgroundRepeatY

  • css/CSSParser.cpp:

(WebCore::CSSParser::parseValue):
(WebCore::CSSParser::parseFillShorthand):
(WebCore::CSSParser::parseFillRepeat):
(WebCore::CSSParser::parseFillProperty):

  • css/CSSParser.h:

Get rid of mappings to RepeatXFill and RepeatYFill since we don't
need those parts of the EFillRepeat enum anymore.

  • css/CSSPrimitiveValueMappings.h:

(WebCore::CSSPrimitiveValue::CSSPrimitiveValue):
(WebCore::CSSPrimitiveValue::operator EFillRepeat):

Add CSSPropertyBackgroundRepeatX and CSSPropertyBackgroundRepeatY

  • css/CSSPropertyLonghand.cpp:

(WebCore::initShorthandMap):

Add background-repeat-x and -y.

  • css/CSSPropertyNames.in:

Break repeat into x and y.

  • css/CSSStyleSelector.cpp:

(WebCore::CSSStyleSelector::applyProperty):
(WebCore::CSSStyleSelector::mapFillRepeatX):
(WebCore::CSSStyleSelector::mapFillRepeatY):

  • css/CSSStyleSelector.h:

There is no more RepeatXFill. Instead, look for
fillLayer->repeatX() == RepeatFill

  • rendering/RenderBoxModelObject.cpp:

(WebCore::RenderBoxModelObject::calculateBackgroundImageGeometry):

Break m_repeat into m_repeatX and m_repeatY

  • rendering/style/FillLayer.cpp:

(WebCore::FillLayer::FillLayer):
(WebCore::FillLayer::operator=):
(WebCore::FillLayer::operator==):
(WebCore::FillLayer::fillUnsetProperties):
(WebCore::FillLayer::cullEmptyLayers):

  • rendering/style/FillLayer.h:

(WebCore::FillLayer::repeatX):
(WebCore::FillLayer::repeatY):
(WebCore::FillLayer::isRepeatXSet):
(WebCore::FillLayer::isRepeatYSet):
(WebCore::FillLayer::setRepeatX):
(WebCore::FillLayer::setRepeatY):
(WebCore::FillLayer::clearRepeatX):
(WebCore::FillLayer::clearRepeatY):
(WebCore::FillLayer::initialFillRepeatX):
(WebCore::FillLayer::initialFillRepeatY):

Again, break backgroundRepeat into backgroundRepeatX and
backgroundRepeatY

  • rendering/style/RenderStyle.h:

(WebCore::InheritedFlags::backgroundRepeatX):
(WebCore::InheritedFlags::backgroundRepeatY):
(WebCore::InheritedFlags::maskRepeatX):
(WebCore::InheritedFlags::maskRepeatY):

We don't need RepeatXFill or RepeatYFill. Now that we store two
values, they are represented by Repeat-NoRepeat and NoRepeat-
Repeat, respectively.

  • rendering/style/RenderStyleConstants.h:

(WebCore::):

LayoutTests: Tests for https://bugs.webkit.org/show_bug.cgi?id=28635 [CSS3
Backgrounds and Borders] Add support for 2-keyword values for
background-repeat

Reviewed by Darin Adler.

This test is updated to test for two values.

  • fast/backgrounds/repeat/parsing-background-repeat-expected.txt:
  • fast/backgrounds/repeat/resources/parsing-background-repeat.js:

These tests are updated to reflect new order.

  • fast/css/background-position-serialize-expected.txt:
  • fast/css/remove-shorthand-expected.txt:

Reflect new order in results, and make the test dump as text.

  • fast/inspector/style.html:
  • platform/mac/fast/inspector/style-expected.checksum: Removed.
  • platform/mac/fast/inspector/style-expected.png: Removed.
  • platform/mac/fast/inspector/style-expected.txt:
14:00 Changeset [47905] by simon.fraser@apple.com
  • 3 edits
    4 adds in trunk

2009-08-31 Simon Fraser <simon.fraser@apple.com>

Reviewed by Dan Bernstein.

Compositing layer size needs to take into account software reflections on children.
https://bugs.webkit.org/show_bug.cgi?id=28837

When computing the bounds of a composited layer, take software-rendered
reflections into account.

Test: compositing/reflections/reflection-in-composited.html

  • rendering/RenderLayerCompositor.cpp: (WebCore::RenderLayerCompositor::calculateCompositedBounds):
13:37 Changeset [47904] by dglazkov@chromium.org
  • 4 edits in trunk

WebCore:

2009-08-31 Mark Mentovai <mark@chromium.org>

Reviewed by Dave Hyatt.

https://bugs.webkit.org/show_bug.cgi?id=28614

Perform a layout prior to checking whether the scrollbar modes are
off, on, or automatic. The modes may change during layout.

  • platform/ScrollView.cpp: (WebCore::ScrollView::updateScrollbars):

WebKit/mac:

2009-08-31 Mark Mentovai <mark@chromium.org>

Reviewed by Dave Hyatt.

https://bugs.webkit.org/show_bug.cgi?id=28614

Perform a layout prior to checking whether the scrollbar modes are
off, on, or automatic. The modes may change during layout.

  • WebView/WebDynamicScrollBarsView.mm: (-[WebDynamicScrollBarsView updateScrollers]):
13:15 Changeset [47903] by eric@webkit.org
  • 8 edits
    1 add in trunk/WebCore

2009-08-31 Patrick Mueller <Patrick_Mueller@us.ibm.com>

Reviewed by Timothy Hatcher.

Inspector Request Headers Should Show Data Sent With Request
https://bugs.webkit.org/show_bug.cgi?id=22920

Manual test added - see below.

  • English.lproj/localizedStrings.js:
  • inspector/InspectorResource.cpp: (WebCore::InspectorResource::updateRequest): (WebCore::InspectorResource::createScriptObject): (WebCore::InspectorResource::updateScriptObject):
  • inspector/InspectorResource.h:
  • inspector/front-end/Resource.js: (WebInspector.Resource):
  • inspector/front-end/ResourceView.js: (WebInspector.ResourceView): (WebInspector.ResourceView.prototype._refreshURL): (WebInspector.ResourceView.prototype._refreshQueryString): (WebInspector.ResourceView.prototype._refreshFormData): (WebInspector.ResourceView.prototype._refreshRequestPayload): (WebInspector.ResourceView.prototype._refreshParms): (WebInspector.ResourceView.prototype._toggleURLdecoding): (WebInspector.ResourceView.prototype._getHeaderValue): (WebInspector.ResourceView.prototype._refreshRequestHeaders):
  • inspector/front-end/inspector.css:
  • inspector/front-end/inspector.js: (WebInspector.addResource): (WebInspector.updateResource):
  • manual-tests/inspector/display-form-data.html: Added.
12:31 Changeset [47902] by barraclough@apple.com
  • 2 edits in trunk/JavaScriptCore

Fixed typo.
https://bugs.webkit.org/show_bug.cgi?id=28691

Patch by Robert Agoston <Agoston.Robert@stud.u-szeged.hu> on 2009-08-31
Reviewed by Gavin Barraclough.

  • parser/Parser.h:

(JSC::Parser::parse):

12:01 Changeset [47901] by simon.fraser@apple.com
  • 3 edits in trunk/LayoutTests

2009-08-31 Simon Fraser <simon.fraser@apple.com>

Followup from r47877; add test for the Obj-C wrapper class for WebkitCSSTransformValue
https://bugs.webkit.org/show_bug.cgi?id=27727

  • platform/mac/fast/dom/wrapper-classes-objc-expected.txt:
  • platform/mac/fast/dom/wrapper-classes-objc.html:
11:37 Changeset [47900] by simon.fraser@apple.com
  • 3 edits
    8 adds in trunk

2009-08-28 Simon Fraser <simon.fraser@apple.com>

Reviewed by Dan Bernstein.

Fix positioning error when a compositing, absolutely-positioned element has an ancestor with opacity.
https://bugs.webkit.org/show_bug.cgi?id=28754

For absolutely-positioned layers, convertToLayerCoords() jumps to the positioned
ancestor, since the layer x and y are relative to that ancestor. However, this could
skip over the ancestorLayer, thus giving the wrong result.

Fix by duplicating the enclosingPositionedAncestor() logic in convertToLayerCoords(),
and checking for ancestorLayer along the way. If found, compute offset of both to
the enclosingPositionedAncestor() and subtract.

This also fixes a positioning bug with abs. positioned elements in reflections,
so there is a new reflection test with a pixel result.

Tests: compositing/geometry/abs-position-inside-opacity.html

fast/reflections/abs-position-in-reflection.html

  • rendering/RenderLayer.cpp: (WebCore::isPositionedContainer): (WebCore::RenderLayer::enclosingPositionedAncestor): (WebCore::RenderLayer::enclosingTransformedAncestor): (WebCore::RenderLayer::convertToLayerCoords):
11:12 Changeset [47899] by aroben@apple.com
  • 2 edits in trunk/WebKitTools

Fall back to a Release version of Safari if a Debug one doesn't exist

<http://webkit.org/b/28849>

Reviewed by Sam Weinig.

  • Scripts/webkitdirs.pm:

(safariPath): If the user is working with a Debug build, but there's
no Debug version of Safari present, fall back to using a Release
version of Safari.

10:54 Changeset [47898] by aroben@apple.com
  • 2 edits in trunk/WebKitTools

Make safariPath() work for Debug builds of Safari on Windows

<http://webkit.org/b/28849>

Reviewed by Sam Weinig.

  • Scripts/webkitdirs.pm:

(safariPath): If the user is working with a Debug build, add the
_debug suffix to Safari.exe.

10:20 Changeset [47897] by pfeldman@chromium.org
  • 7 edits in trunk/WebCore

2009-08-31 Pavel Feldman <pfeldman@chromium.org>

Reviewed by Timothy Hatcher.

WebInspector: pass call frames into the frontend as a part of
pausedScript event.

https://bugs.webkit.org/show_bug.cgi?id=28847

  • inspector/InspectorController.cpp: (WebCore::InspectorController::didPause):
  • inspector/InspectorFrontend.cpp: (WebCore::InspectorFrontend::pausedScript):
  • inspector/InspectorFrontend.h:
  • inspector/front-end/ScriptsPanel.js: (WebInspector.ScriptsPanel.prototype.debuggerPaused):
  • inspector/front-end/inspector.js: (WebInspector.pausedScript):
08:06 Changeset [47896] by aroben@apple.com
  • 4 edits in trunk/WebKit/win

Give WebKit clients access to the window features specified in window.open

Reviewed by Steve Falkenburg.

  • COMVariantSetter.h: Added COMVariantSetters for bool and float.
  • Interfaces/IWebUIDelegatePrivate.idl: Added IWebUIDelegatePrivate2,

which contains a new version of createWebViewWithRequest that also
passes the window features as an IPropertyBag. This is analagous to
the -createWebViewWithRequest:windowFeatures: method added to
WebUIDelegate on Mac in r27452. Also added key strings for use with
the window features property bag.

  • WebCoreSupport/WebChromeClient.cpp:

(createWindowFeaturesPropertyBag): Added this helper function to turn
a WebCore::WindowFeatures into an IPropertyBag for passing to WebKit
clients.
(WebChromeClient::createWindow): If the UI delegate implements
IWebUIDelegatePrivate2, call the new createWebViewWithRequest function
that takes a window features property bag. If the delegate doesn't
implement the new function, fall back to the old functions.

08:05 Changeset [47895] by aroben@apple.com
  • 2 edits in trunk/WebKit/win

A little clean-up in WebChromeClient::createWindow

Reviewed by Steve Falkenburg.

  • WebCoreSupport/WebChromeClient.cpp:

(WebChromeClient::createWindow): Refactored to share more code between
the dialog and non-dialog cases.

05:50 Changeset [47894] by zoltan@webkit.org
  • 14 edits in trunk/LayoutTests

2009-08-31 Andras Becsi <becsi.andras@stud.u-szeged.hu>

Reviewed by Ariya Hidayat.

[Qt] Update outdated font metrics in the remaining css1 tests. All tests pass.

  • platform/qt/Skipped:
  • platform/qt/css1/text_properties/letter_spacing-expected.txt:
  • platform/qt/css1/text_properties/line_height-expected.txt:
  • platform/qt/css1/text_properties/text_align-expected.txt:
  • platform/qt/css1/text_properties/text_decoration-expected.txt:
  • platform/qt/css1/text_properties/text_indent-expected.txt:
  • platform/qt/css1/text_properties/text_transform-expected.txt:
  • platform/qt/css1/text_properties/vertical_align-expected.txt:
  • platform/qt/css1/text_properties/word_spacing-expected.txt:
  • platform/qt/css1/units/color_units-expected.txt:
  • platform/qt/css1/units/length_units-expected.txt:
  • platform/qt/css1/units/percentage_units-expected.txt:
  • platform/qt/css1/units/urls-expected.txt:
05:15 Changeset [47893] by zoltan@webkit.org
  • 14 edits in trunk/LayoutTests

2009-08-31 Andras Becsi <becsi.andras@stud.u-szeged.hu>

Reviewed by Ariya Hidayat.

[Qt] Update font metrics for these css1 tests.

  • platform/qt/Skipped:
  • platform/qt/css1/formatting_model/canvas-expected.txt:
  • platform/qt/css1/formatting_model/floating_elements-expected.txt:
  • platform/qt/css1/formatting_model/height_of_lines-expected.txt:
  • platform/qt/css1/formatting_model/horizontal_formatting-expected.txt:
  • platform/qt/css1/formatting_model/inline_elements-expected.txt:
  • platform/qt/css1/formatting_model/replaced_elements-expected.txt:
  • platform/qt/css1/formatting_model/vertical_formatting-expected.txt:
  • platform/qt/css1/pseudo/anchor-expected.txt:
  • platform/qt/css1/pseudo/firstletter-expected.txt:
  • platform/qt/css1/pseudo/firstline-expected.txt:
  • platform/qt/css1/pseudo/multiple_pseudo_elements-expected.txt:
  • platform/qt/css1/pseudo/pseudo_elements_in_selectors-expected.txt:
05:09 Changeset [47892] by vestbo@webkit.org
  • 2 edits in trunk/WebCore

[Qt] Un-blacklist Silverlight on Mac

Reviewed by Ariya Hidayat.

Does not seem to crash anymore, probably due to how we now handle
drawing and event model handshake better.

  • plugins/mac/PluginPackageMac.cpp:

(WebCore::PluginPackage::isPluginBlacklisted):

04:17 Changeset [47891] by zoltan@webkit.org
  • 10 edits in trunk/LayoutTests

2009-08-31 Andras Becsi <becsi.andras@stud.u-szeged.hu>

Reviewed by Ariya Hidayat.

[Qt] Update outdated font metrics for these css1 tests.

  • platform/qt/Skipped:
  • platform/qt/css1/color_and_background/background-expected.txt:
  • platform/qt/css1/color_and_background/background_attachment-expected.txt:
  • platform/qt/css1/color_and_background/background_color-expected.txt:
  • platform/qt/css1/color_and_background/background_image-expected.txt:
  • platform/qt/css1/color_and_background/background_position-expected.txt:
  • platform/qt/css1/color_and_background/background_repeat-expected.txt:
  • platform/qt/css1/color_and_background/color-expected.txt:
  • platform/qt/css1/conformance/forward_compatible_parsing-expected.txt:
04:06 Changeset [47890] by eric@webkit.org
  • 4 edits in trunk/WebCore

2009-08-31 Cameron McCormack <cam@mcc.id.au>

Reviewed by Eric Seidel.

https://bugs.webkit.org/show_bug.cgi?id=28827
SVGSVGElement.unsuspendRedraw() shouldn't throw

  • svg/SVGSVGElement.cpp: (WebCore::SVGSVGElement::unsuspendRedraw):
  • svg/SVGSVGElement.h:
  • svg/SVGSVGElement.idl:
03:55 Changeset [47889] by eric@webkit.org
  • 4 edits in trunk/WebCore

2009-08-31 Kent Tamura <tkent@chromium.org>

Reviewed by Eric Seidel.

  • html/HTMLInputElement.cpp: (WebCore::HTMLInputElement::list): (WebCore::HTMLInputElement::dataList): (WebCore::HTMLInputElement::selectedOption):
  • html/HTMLInputElement.h:
  • html/HTMLInputElement.idl:
03:45 Changeset [47888] by eric@webkit.org
  • 2 edits in trunk/WebCore

2009-08-31 Kwang Yul Seo <skyul@company100.net>

Reviewed by Eric Seidel.

Remove ASSERT(frame) in ResourceHandle::start (ResourceHandleCurl.cpp)
https://bugs.webkit.org/show_bug.cgi?id=28802

Remove ASSERT because the frame could be null if the
ResourceHandle is not associated with any frame, e.g. if we are
downloading a file.
If the frame is not null but the page is null this must be an
attempted load from an onUnload handler, so let's just block it.

  • platform/network/curl/ResourceHandleCurl.cpp: (WebCore::ResourceHandle::start):
03:18 Changeset [47887] by eric@webkit.org
  • 3 edits in trunk/WebCore

2009-08-31 Maxime Simon <simon.maxime@gmail.com>

Reviewed by Eric Seidel.

Build fix for platforms which don't enable DOM_STORAGE.
https://bugs.webkit.org/show_bug.cgi?id=28834

  • bindings/js/ScriptController.cpp: (WebCore::ScriptController::evaluate):
  • bindings/v8/ScriptController.cpp: (WebCore::ScriptController::evaluate):
02:59 Changeset [47886] by vestbo@webkit.org
  • 1 edit in trunk/WebKitTools/DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro

Fix the Qt/Mac build of TestNetscapePlugin after Qt change 3940f25

02:53 Changeset [47885] by vestbo@webkit.org
  • 1 edit in trunk/WebCore/WebCore.pro

Fix the Qt/Mac build after Qt change 3940f25 (LIBS_PRIVATE)

We no longer pick up the libraries/frameworks that Qt was built
against, so we have to add them manually for our targets.

00:49 Changeset [47884] by zoltan@webkit.org
  • 23 edits in trunk/LayoutTests

2009-08-31 Renata Hodovan <hodovan.renata@stud.u-szeged.hu>

Reviewed by Ariya Hidayat.

[Qt] Update outdated metrics in css2.1/t040*. (19 files)
Fix relative path in .css files.

  • css2.1/support/css1test64a.css:
  • css2.1/support/css1test64b.css:
  • platform/qt/Skipped:
  • platform/qt/css2.1/t0402-c71-fwd-parsing-00-f-expected.txt:
  • platform/qt/css2.1/t0402-c71-fwd-parsing-01-f-expected.txt:
  • platform/qt/css2.1/t0402-c71-fwd-parsing-02-f-expected.txt:
  • platform/qt/css2.1/t0402-c71-fwd-parsing-03-f-expected.txt:
  • platform/qt/css2.1/t0402-c71-fwd-parsing-04-f-expected.txt:
  • platform/qt/css2.1/t0402-syntax-01-f-expected.txt:
  • platform/qt/css2.1/t0402-syntax-02-f-expected.txt:
  • platform/qt/css2.1/t0402-syntax-03-f-expected.txt:
  • platform/qt/css2.1/t0402-syntax-04-f-expected.txt:
  • platform/qt/css2.1/t0402-syntax-05-f-expected.txt:
  • platform/qt/css2.1/t0402-syntax-06-f-expected.txt:
  • platform/qt/css2.1/t040302-c61-ex-len-00-b-a-expected.txt:
  • platform/qt/css2.1/t040302-c61-phys-len-00-b-expected.txt:
  • platform/qt/css2.1/t040302-c61-rel-len-00-b-ag-expected.txt:
  • platform/qt/css2.1/t040303-c62-percent-00-b-ag-expected.txt:
  • platform/qt/css2.1/t040304-c64-uri-00-a-g-expected.txt:
  • platform/qt/css2.1/t040306-c63-color-00-b-ag-expected.txt:
  • platform/qt/css2.1/t040306-syntax-01-f-expected.txt:
  • platform/qt/css2.1/t040307-syntax-01-b-expected.txt:

08/29/09:

16:59 Changeset [47883] by adele@apple.com
  • 3 edits
    2 adds in trunk

WebCore: Fix for https://bugs.webkit.org/show_bug.cgi?id=28829
Crash in AccessibilityRenderObject::activeDescendant when trying to set aria-activedescendant to something without a renderer

Reviewed by Dan Bernstein.

Test: accessibility/aria-activedescendant-crash.html

  • accessibility/AccessibilityRenderObject.cpp: (WebCore::AccessibilityRenderObject::activeDescendant): Nil check.

LayoutTests: Test for https://bugs.webkit.org/show_bug.cgi?id=28829
Crash in AccessibilityRenderObject::activeDescendant when trying to set aria-activedescendant to something without a renderer

Reviewed by Dan Bernstein.

  • accessibility/aria-activedescendant-crash-expected.txt: Added.
  • accessibility/aria-activedescendant-crash.html: Added.
Note: See TracTimeline for information about the timeline view.