⚠ Archived content — this site is no longer maintained.   Current WebKit documentation is at docs.webkit.org.

Timeline



May 15, 2008:

10:53 PM Changeset in webkit [33512] by kevino@webkit.org
  • 4 edits in trunk

wx build fix. Add rendering/style to includes dir.

10:10 PM Changeset in webkit [33511] by kmccullough@apple.com
  • 2 edits in trunk/JavaScriptCore

2008-05-15 Kevin McCullough <​kmccullough@apple.com>

  • Build fix.
8:27 PM Changeset in webkit [33510] by alice.liu@apple.com
  • 15 edits in trunk/WebCore

2008-05-15 Alice Liu <​alice.liu@apple.com>

Reviewed by Beth Dakin.

Fixed <rdar://problem/5882283> Implement ARIA aria-activedescendant
Also added support for aria roles listbox and option, and aria-selected,
and aria-multiselectable for listboxes

  • dom/Element.cpp: (WebCore::Element::attributeChanged):
  • dom/Element.h:
  • dom/StyledElement.cpp: (WebCore::StyledElement::attributeChanged):
  • html/HTMLAttributeNames.in:
  • page/AXObjectCache.cpp: (WebCore::AXObjectCache::handleActiveDescendantChanged):
  • page/AXObjectCache.h: (WebCore::AXObjectCache::accessibilityEnabled): (WebCore::AXObjectCache::handleActiveDescendantChanged):
  • page/AccessibilityListBox.cpp: (WebCore::AccessibilityListBox::selectedChildren): (WebCore::AccessibilityListBox::visibleChildren):
  • page/AccessibilityListBox.h:
  • page/AccessibilityListBoxOption.h: (WebCore::AccessibilityListBoxOption::selectedChildren): (WebCore::AccessibilityListBoxOption::visibleChildren):
  • page/AccessibilityObject.cpp:
  • page/AccessibilityObject.h: (WebCore::AccessibilityObject::isListBox): (WebCore::AccessibilityObject::shouldFocusActiveDescendant): (WebCore::AccessibilityObject::activeDescendant): (WebCore::AccessibilityObject::handleActiveDescendantChanged):
  • page/AccessibilityRenderObject.cpp: (WebCore::AccessibilityRenderObject::title): (WebCore::AccessibilityRenderObject::accessibilityShouldUseUniqueId): (WebCore::AccessibilityRenderObject::focusedUIElement): (WebCore::AccessibilityRenderObject::shouldFocusActiveDescendant): (WebCore::AccessibilityRenderObject::activeDescendant): (WebCore::AccessibilityRenderObject::handleActiveDescendantChanged): (WebCore::RoleEntry::): (WebCore::AccessibilityRenderObject::ariaRoleAttribute): (WebCore::AccessibilityRenderObject::ariaListboxSelectedChildren): (WebCore::AccessibilityRenderObject::selectedChildren): (WebCore::AccessibilityRenderObject::ariaListboxVisibleChildren): (WebCore::AccessibilityRenderObject::visibleChildren):
  • page/AccessibilityRenderObject.h: (WebCore::AccessibilityRenderObject::isDetached):
  • page/mac/AccessibilityObjectWrapper.mm: (-[AccessibilityObjectWrapper accessibilityAttributeValue:]):
8:20 PM Changeset in webkit [33509] by timothy@apple.com
  • 3 edits
    3 adds in trunk

Fixes a bug in the view source parsing of text within <script>,
<style>, <xmp>, <textarea> and other special tags. The text was
being parsed as HTML and would be syntax highlighted if a "<"
was encountered.

<rdar://problem/5790805> Inspector's source view missing some words,
doubles others in JS source in <script> tag (17752)

Reviewed by Dave Hyatt.

Added test: fast/frames/viewsource-plain-text-tags.html

  • html/HTMLTokenizer.cpp:

(WebCore::HTMLTokenizer::scriptHandler): Set scriptCodeSize and
scriptCodeResync to zero before an early return in view source mode.
This was causing text from a previous script tag to show up in
then next style tag.
(WebCore::HTMLTokenizer::parseTag): Enter an if block in view source
mode even if n is null. The n is property guarded in other places
and not used by view source mode. Removes a redundant null check on n.
Add a new null check of n before setting scriptSrcCharset. Other code
in HTMLTokenizer is already guarded for this code path.

8:20 PM Changeset in webkit [33508] by timothy@apple.com
  • 3 edits in trunk/WebCore

Changes the Web Inspector node highlight to be less aggressive
and only highlight when hovering over the selected node in the DOM
tree or breadcrumbs. You can highlight on hover for any node if
you hold down the Alt/Option key while hovering.

Reviewed by Kevin McCullough.

  • page/inspector/ElementsPanel.js:
  • page/inspector/inspector.js:
7:51 PM Changeset in webkit [33507] by kmccullough@apple.com
  • 5 edits in trunk/JavaScriptCore

2008-05-15 Kevin McCullough <​kmccullough@apple.com>

Reviewed by Tim.

<rdar://problem/5770054> JavaScript profiler (10928)

  • Cache some values to save on computing them repetitively. This will be a big savings when we sort since we won't have to walk the tree for every comparison!
  • We cache these values when we end profiling because otherwise we won't know which profile to get the totalTime for the whole profile from without retaining a reference to the head profile or looking up the profile from the list of all profiles.
  • Also it's safe to assume we won't be asked for these values while we are still profiling since the WebInspector only get's profileNodes from profiles that are in the allProfiles() list and a profile is only added to that list after it has finished and these values will no longer change.
  • JavaScriptCore.exp:
  • profiler/ProfileNode.cpp: (KJS::ProfileNode::ProfileNode): (KJS::ProfileNode::stopProfiling): (KJS::ProfileNode::printDataInspectorStyle): (KJS::ProfileNode::printDataSampleStyle): (KJS::ProfileNode::endAndRecordCall):
  • profiler/ProfileNode.h: (KJS::ProfileNode::totalTime): (KJS::ProfileNode::selfTime): (KJS::ProfileNode::totalPercent): (KJS::ProfileNode::selfPercent):
  • profiler/Profiler.cpp: (KJS::Profiler::stopProfiling):
7:37 PM Changeset in webkit [33506] by Stephanie Lewis
  • 2 edits in trunk/WebKit/mac

2008-05-15 Stephanie Lewis <Stephanie Lewis>

fix mac build

  • WebView/WebView.mm: (-[WebView closeWithFastTeardown]):
7:31 PM Changeset in webkit [33505] by Stephanie Lewis
  • 2 edits in trunk/WebCore

2008-05-15 Stephanie Lewis <Stephanie Lewis>

Fix windows build

  • loader/FrameLoader.cpp: (WebCore::FrameLoader::stopLoading):
6:07 PM Changeset in webkit [33504] by Stephanie Lewis
  • 8 edits in trunk

2008-05-15 Stephanie Lewis <Stephanie Lewis>

Reviewed by Anders.

Turn on fast teardown. I added a preference for using full teardown because
the LEAKS output will be useless without a full teardown.

preference for fullteardown

  • WebView/WebPreferenceKeysPrivate.h:
  • WebView/WebPreferences.m: (-[WebPreferences setFullDocumentTeardownEnabled:]): (-[WebPreferences fullDocumentTeardownEnabled]):
  • WebView/WebPreferencesPrivate.h:


on application quit dispatch unload events and destroy plugins then exit

  • WebView/WebView.mm: (-[WebView closeWithFastTeardown]): (-[WebView _close]):
  • WebView/WebViewPrivate.h:

export a symbol needed for fast teardown

  • WebCore.base.exp:
5:51 PM Changeset in webkit [33503] by Stephanie Lewis
  • 25 edits in trunk

2008-05-15 Stephanie Lewis <Stephanie Lewis>

Reviewed by Anders.

Keep track on pending unload event counts

No change in behavior so no new tests, but I am dumping the
unload event count to track the new variable.

  • WebCore.base.exp:

increment/decrement the count when an event is added/removed

  • dom/Document.cpp: (WebCore::Document::removeHTMLWindowEventListener): (WebCore::Document::addWindowEventListener): (WebCore::Document::removeWindowEventListener): (WebCore::Document::setPendingFrameUnloadEventCount):
  • dom/Document.h:

set the count to 0 for each frame after the unload event has been dispatched

  • loader/FrameLoader.cpp: (WebCore::FrameLoader::stopLoading):

keep track of the number of pending unload events for the eventHandler's frame

  • page/EventHandler.cpp: (WebCore::EventHandler::EventHandler):

(WebCore::EventHandler::pendingFrameUnloadEventCount):

(WebCore::EventHandler::setPendingFrameUnloadEventCount):

  • page/EventHandler.h:

keep track of the total number of pending unload events for all of the frames

  • page/Page.cpp: (WebCore::Page::Page): (WebCore::Page::pendingUnloadEventCount): (WebCore::Page::setPendingUnloadEventCount):

age/Page.h:

get the pending frame unload count from WebCore

  • WebView/WebFrame.mm: (-[WebFrame _pendingFrameUnloadEventCount]):
  • WebView/WebFramePrivate.h:

Dump the unload count for a frame after parsing is finished.

  • DumpRenderTree/mac/FrameLoadDelegate.mm: (-[FrameLoadDelegate webView:didFinishDocumentLoadForFrame:]):

Dump unload event counts for frames.

  • fast/dom/Window/get-set-properties-expected.txt:
  • fast/events/onunload-expected.txt:
  • fast/events/onunload-not-on-body-expected.txt:
  • fast/events/onunload-window-property-expected.txt:
  • fast/forms/button-state-restore-expected.txt:
  • fast/history/history_reload-expected.txt:
  • fast/loader/onunload-form-submit-crash-2-expected.txt:
  • fast/loader/onunload-form-submit-crash-expected.txt:
  • http/tests/xmlhttprequest/xhr-onunload-expected.txt:
  • platform/mac/fast/loader/start-load-in-unload-expected.txt:
4:40 PM Changeset in webkit [33502] by sfalken@apple.com
  • 2 edits in trunk/WebCore

2008-05-15 Steve Falkenburg <​sfalken@apple.com>

Fix build.

  • WebCore.vcproj/WebCore.vcproj:
4:39 PM Changeset in webkit [33501] by sullivan@apple.com
  • 2 edits in trunk/WebKit/mac

2008-05-15 John Sullivan <​sullivan@apple.com>

Reviewed by Kevin Decker


  • fixed <rdar://problem/5940275> Inspector highlighting moves to bottom-left corner of screen when new tab appears


The highlight should go away entirely, but this simple patch just makes it not jump away.
The issue with it not going away entirely is harder to fix and covered by <rdar://problem/5322306>

  • WebInspector/WebNodeHighlight.m: (-[WebNodeHighlight _repositionHighlightWindow]): Bail out if target view isn't in a window
4:37 PM Changeset in webkit [33500] by Stephanie Lewis
  • 10 edits in trunk/WebKit/mac

2008-05-15 Stephanie Lewis <Stephanie Lewis>

Reviewed by Anders.

Track views that contain plugin instances so that they can be destroyed at application
quit without walking the entire document tree.

Add/Remove Netscape plugin views from instance list. Start/stop are when netscape
plugins are created and destroyed

  • Plugins/WebBaseNetscapePluginView.mm: (-[WebBaseNetscapePluginView start]): (-[WebBaseNetscapePluginView stop]):


Add/remove WebKit plugin views from instance list

  • Plugins/WebPluginController.mm: (-[WebPluginController addPlugin:]): (-[WebPluginController destroyPlugin:]): (-[WebPluginController destroyAllPlugins]):

Add a set of views with plugin instances to the WebPluginDatabase

  • Plugins/WebPluginDatabase.h:
  • Plugins/WebPluginDatabase.mm: (-[WebPluginDatabase init]): (-[WebPluginDatabase dealloc]): (-[WebPluginDatabase addPluginInstanceView:]): (-[WebPluginDatabase removePluginInstanceView:]): (-[WebPluginDatabase removePluginInstanceViewsFor:]): (-[WebPluginDatabase destroyAllPluginInstanceViews]):


Handle cases where plugin views are detached before the plugin is destroyed.

  • WebCoreSupport/WebFrameLoaderClient.mm: (WebFrameLoaderClient::detachedFromParent2): (WebFrameLoaderClient::transitionToCommittedFromCachedPage): (WebFrameLoaderClient::transitionToCommittedForNewPage):

Add plugin instances to the set in the WebPluginDatabase by way of the WebView

  • WebView/WebHTMLView.mm:


  • WebView/WebHTMLViewInternal.h:
  • WebView/WebView.mm: (-[WebView addPluginInstanceView:]): (-[WebView removePluginInstanceView:]): (-[WebView removePluginInstanceViewsFor:]):
  • WebView/WebViewInternal.h:
4:10 PM Changeset in webkit [33499] by beidson@apple.com
  • 4 edits in trunk/WebKit

WebKit/mac:

2008-05-15 Brady Eidson <​beidson@apple.com>

Reviewed by Tim Hatcher

Until the settings/preferences equation can be reworked, we'll need to manually set the local storage path
before setting the page group of the new page.


  • WebView/WebView.mm: (-[WebView _commonInitializationWithFrameName:groupName:]): Set the LocalStorage path immediately after creating the page so it is in place for initializing the LocalStorageThread

WebKit/win:

2008-05-06 Brady Eidson <​beidson@apple.com>

Reviewed by Tim Hatcher

Until the settings/preferences equation can be reworked, we'll need to manually set the local storage path
before setting the page group of the new page.

Since I still need to add the local storage path to WebPreferences on Windows, mark this with a FIXME for now.

  • WebView.cpp: (WebView::initWithFrame): Mark an important FIXME for setting the LocalStorage path as appropriate. Until this FIXME is handled, Windows will have no persistent LocalStorage
4:03 PM Changeset in webkit [33498] by sfalken@apple.com
  • 2 edits in trunk/WebCore

2008-05-15 Steve Falkenburg <​sfalken@apple.com>

Fix build.

  • WebCore.vcproj/WebCore.vcproj:
3:25 PM Changeset in webkit [33497] by sfalken@apple.com
  • 2 edits in trunk/WebCore

2008-05-15 Steve Falkenburg <​sfalken@apple.com>

Fix build.

  • WebCore.vcproj/WebCore.vcproj:
2:08 PM Changeset in webkit [33496] by ddkilzer@apple.com
  • 3 edits
    2 adds in trunk

WebCore:

Bug 19085: font-family: initial sets font-family differently than font: initial

<​https://bugs.webkit.org/show_bug.cgi?id=19085>
<rdar://problem/5939391>

Reviewed by David Hyatt.

Test: fast/css/font-family-initial.html

  • css/CSSStyleSelector.cpp: (WebCore::CSSStyleSelector::applyProperty): Don't set family if first family of initial font description is empty.

LayoutTests:

Bug 19085: font-family: initial sets font-family differently than font: initial

<​https://bugs.webkit.org/show_bug.cgi?id=19085>
<rdar://problem/5939391>

Reviewed by David Hyatt.

  • fast/css/font-family-initial-expected.txt: Added.
  • fast/css/font-family-initial.html: Added.
2:00 PM Changeset in webkit [33495] by hyatt@apple.com
  • 6 edits
    2 moves
    1 add in trunk/WebCore

2008-05-15 David Hyatt <​hyatt@apple.com>

Move RenderStyle into a new subdirectory, rendering/style/, in preparation for breaking it up into
a bunch of smaller files.

Reviewed by Dan Bernstein

  • GNUmakefile.am:
  • WebCore.pro:
  • WebCore.vcproj/WebCore.vcproj:
  • WebCore.xcodeproj/project.pbxproj:
  • WebCoreSources.bkl:
  • rendering/RenderStyle.cpp: Removed.
  • rendering/RenderStyle.h: Removed.
  • rendering/style: Added.
  • rendering/style/RenderStyle.cpp: Copied from rendering/RenderStyle.cpp.
  • rendering/style/RenderStyle.h: Copied from rendering/RenderStyle.h.
1:42 PM Changeset in webkit [33494] by ddkilzer@apple.com
  • 2 edits in trunk/WebCore

Fix leaks when stopping a page from loading

Patch suggested by Antti. Reviewed by Antti.

  • loader/DocLoader.cpp: (WebCore::DocLoader::clearPreloads): If a resource is in m_preloads but never made it to the cache, delete it to prevent it from leaking.
12:58 PM Changeset in webkit [33493] by timothy@apple.com
  • 2 edits in trunk/WebKit/mac

Fixes the bug where the Web Inspector would flash white while resizing.
This was cause by deferring the window creation.

<rdar://problem/5873549> REGRESSION: Inspector flickers horribly while resizing (17979)

Reviewed by Darin Adler.

  • WebCoreSupport/WebInspectorClient.mm:

(-[WebInspectorWindowController window]): Don't defer the window creation.

12:49 PM Changeset in webkit [33492] by jchaffraix@webkit.org
  • 3 edits
    2 adds in trunk

WebCore:

2008-05-15 Julien Chaffraix <​jchaffraix@webkit.org>

Reviewed by Darin.

​https://bugs.webkit.org/show_bug.cgi?id=13942
ASSERTION FAILED: !attrName.contains('/') in HTMLTokenizer.cpp:132 when loading ​http://bamanzi.blogeden.cn/

In HTML, when an attribute was null (for example when we parse '="somevalue"'
(attribute forgotten or there is a space between the attribute and the '=')),
the fallback was to assign the value to the attribute. However if the value was
a url or did contain a '/', we would trigger the assertion.

To avoid that, we check the value before assigning it now and do not assign it
if it means adding a '/'.

Test: fast/parser/assertion-empty-attribute.html

  • html/HTMLTokenizer.cpp: (WebCore::HTMLTokenizer::parseTag): Add check for '/' in value before assigning it to an attribute when the attribute is null.

LayoutTests:

2008-05-15 Julien Chaffraix <​jchaffraix@webkit.org>

Reviewed by Darin.

Test case for ​https://bugs.webkit.org/show_bug.cgi?id=13942
ASSERTION FAILED: !attrName.contains('/') in HTMLTokenizer.cpp:132 when loading ​http://bamanzi.blogeden.cn/

  • fast/parser/assertion-empty-attribute-expected.txt: Added.
  • fast/parser/assertion-empty-attribute.html: Added.
10:22 AM Changeset in webkit [33491] by ap@webkit.org
  • 9 edits in trunk

Reviewed by Dan Bernstein.

​https://bugs.webkit.org/show_bug.cgi?id=10707
DumpRenderTree should not be able to access non-local resources

  • DumpRenderTree/mac/ResourceLoadDelegate.mm: (-[ResourceLoadDelegate webView:resource:willSendRequest:redirectResponse:fromDataSource:]): Block them, and complain.
10:02 AM Changeset in webkit [33490] by ap@webkit.org
  • 2 edits in trunk/WebKit/mac

Tiger build fix.

  • Misc/WebNSAttributedStringExtras.mm: Import WebTypesInternal.h for NSUInteger.
9:18 AM Changeset in webkit [33489] by kevino@webkit.org
  • 2 edits in trunk/WebKitTools

wx build fix. Update the version of libpng to download and instsall.

8:52 AM Changeset in webkit [33488] by kevino@webkit.org
  • 4 edits
    6 adds in trunk/WebCore

wx build fix for recent breakages. Add EventLoopWx.cpp, update the Bakefiles, and add stubs for new Plugins functions. Also move plugin functions outside of TemporaryLinkStubs.cpp so that they are more noticeable.

7:51 AM Changeset in webkit [33487] by Simon Hausmann
  • 2 edits in trunk/WebCore

2008-05-15 Adam Treat <​treat@kde.org>

Reviewed by Simon.

Don't try to setCompositionMode for devices that do not support PorterDuff

6:45 AM Changeset in webkit [33486] by ap@webkit.org
  • 7 edits in trunk

2008-05-15 Adele Peterson <​adele@apple.com>

Reviewed and landed by Alexey.

Use TextIterator in +[NSAttributedString _web_attributedStringFromRange:].

6:25 AM Changeset in webkit [33485] by Simon Hausmann
  • 2 edits in trunk/JavaScriptCore

Simon Hausmann <​shausman@trolltech.com>

Fix compilation when compiling with MSVC and wchar_t support.

5:45 AM Changeset in webkit [33484] by oliver@apple.com
  • 11 edits in branches/squirrelfish/JavaScriptCore

Revert "Bug 18626: SQUIRRELFISH: support the "slow script" dialog"

This seems to cause a significant perf regression on some systems

2:31 AM Changeset in webkit [33483] by oliver@apple.com
  • 11 edits in branches/squirrelfish/JavaScriptCore

Bug 18626: SQUIRRELFISH: support the "slow script" dialog
<​https://bugs.webkit.org/show_bug.cgi?id=18626>

Reviewed by Maciej

Support the slow script dialog for for(;;), while, and do..while
loops.

12:26 AM Changeset in webkit [33482] by Simon Hausmann
  • 2 edits in trunk/WebCore

2008-05-15 Simon Hausmann <Simon Hausmann>

Fix the Qt build by adding NetworkStateNotifier.cpp to the build.

12:14 AM Changeset in webkit [33481] by ggaren@apple.com
  • 2 edits in branches/squirrelfish/JavaScriptCore

2008-05-15 Geoffrey Garen <​ggaren@apple.com>

Not reviewed.


We regret to inform you that your program is crashing because you were
stupid.

  • VM/Machine.cpp: (KJS::Machine::privateExecute): Math is hard.

May 14, 2008:

11:50 PM Changeset in webkit [33480] by ggaren@apple.com
  • 10 edits in branches/squirrelfish/JavaScriptCore

2008-05-14 Geoffrey Garen <​ggaren@apple.com>

Reviewed by Oliver Hunt.

A little more debugger action: filled in op_debug. All debugger control
flow works now, but variable inspection and backtraces still don't.


SunSpider reports no change.

  • VM/CodeGenerator.cpp: Changed op_debug to accept line number parameters.
  • VM/Machine.cpp: (KJS::Machine::getFunctionAndArguments): Moved op_debug into a NEVER_INLINE function to avoid a stunning 10% performance regression. Also factored out a common function for retrieving the function and arguments from a call frame.
  • kjs/JSActivation.cpp: (KJS::JSActivation::createArgumentsObject): Use the new factored out function mentioned above.
  • kjs/Parser.cpp: (KJS::Parser::parse): Increment m_sourceId before assigning it, so the sourceId we send to the debugger matches the sourceId recorded in the node.
  • kjs/nodes.cpp: Emit debugging hooks.
11:06 PM Changeset in webkit [33479] by mrowe@apple.com
  • 9 edits in tags/Safari-6526.5

Merge r33017.

10:52 PM Changeset in webkit [33478] by mrowe@apple.com
  • 4 edits in trunk

Versioning.

10:49 PM Changeset in webkit [33477] by mrowe@apple.com
  • 1 copy in tags/Safari-6526.5

New tag.

10:49 PM Changeset in webkit [33476] by mitz@apple.com
  • 3 edits
    2 adds in trunk

WebCore:

Reviewed by Mark Rowe.

  • fix <rdar://problem/5907633> Memory corruption after closing a document that uses box-reflect

Test: fast/reflections/teardown-crash.html

  • rendering/RenderLayer.cpp: (WebCore::RenderLayer::~RenderLayer): Changed to not call removeLayers() on the reflection if the document is being destroyed, because in that case the layer tree is not being maintained.

LayoutTests:

Reviewed by Mark Rowe.

  • test for <rdar://problem/5907633> Memory corruption after closing a document that uses box-reflect
  • fast/reflections/teardown-crash-expected.txt: Added.
  • fast/reflections/teardown-crash.html: Added.
7:36 PM Changeset in webkit [33475] by kmccullough@apple.com
  • 2 edits in trunk/JavaScriptCore

2008-05-14 Kevin McCullough <​kmccullough@apple.com>

Reviewed by Tim.

<rdar://problem/5770054> JavaScript profiler (10928)

  • Turn on the profiler.
  • kjs/config.h:
7:06 PM Changeset in webkit [33474] by kmccullough@apple.com
  • 4 edits in trunk

JavaScriptCore:

2008-05-14 Kevin McCullough <​kmccullough@apple.com>

Reviewed by Tim.

<rdar://problem/5770054> JavaScript profiler (10928)

WebCore:

2008-05-14 Kevin McCullough <​kmccullough@apple.com>

Reviewed by Tim.

<rdar://problem/5770054> JavaScript profiler (10928)

  • page/JavaScriptProfileNode.cpp: (WebCore::getTotalPercent): (WebCore::getSelfPercent): (WebCore::ProfileNodeClass):
6:59 PM Changeset in webkit [33473] by timothy@apple.com
  • 2 edits
    1 add in trunk/WebCore

Fixes a crash in JavaScriptDebugServer::returnEvent when debugging
code that contains an eval. This change makes stepping into eval
code work as expected.

​http://bugs.webkit.org/show_bug.cgi?id=19038

Reviewed by Kevin McCullough.

Tested with: manual-tests/inspector/returnEvent-crash.html

  • manual-tests/inspector/returnEvent-crash.html: Added.
  • page/JavaScriptDebugServer.cpp:

(WebCore::JavaScriptDebugServer::sourceParsed): Adds #ifdefed
debugging code to prevent the sourceID and URL.
(WebCore::updateCurrentCallFrame): Added. A helper function that
is called from all 4 of the debugger hooks below. This function will
update and/or create JavaScriptCallFrames to match the exec state,
sourceID and lineNumber passed into it. Contains #ifdefed debugging
code that was helpful while fixing this bug.
(WebCore::JavaScriptDebugServer::callEvent): Call updateCurrentCallFrame
before pauseIfNeeded.
(WebCore::JavaScriptDebugServer::atStatement): Ditto.
(WebCore::JavaScriptDebugServer::returnEvent): Ditto.
(WebCore::JavaScriptDebugServer::exception): Ditto.

6:31 PM Changeset in webkit [33472] by alp@webkit.org
  • 2 edits in trunk/WebCore

2008-05-14 Alp Toker <​alp@nuanti.com>

GTK+ build fix for r33457. Add NetworkStateNotifier.cpp to the build.

  • GNUmakefile.am:
6:10 PM Changeset in webkit [33471] by kmccullough@apple.com
  • 1 edit in trunk/JavaScriptCore/ChangeLog

2008-05-14 Kevin McCullough <​kmccullough@apple.com>

Giving credit where credit is due.

  • ChangeLog:
6:09 PM Changeset in webkit [33470] by kmccullough@apple.com
  • 5 edits in trunk/JavaScriptCore

2008-05-14 Kevin McCullough <​kmccullough@apple.com>

Reviewed by Geoff and Sam.

<rdar://problem/5770054> JavaScript profiler (10928)
Add the ability to get percentages of total and self time for displaying
in the WebInspector.

  • profiler/Profile.h: (KJS::Profile::totalProfileTime):
  • profiler/ProfileNode.cpp: (KJS::ProfileNode::totalPercent): (KJS::ProfileNode::selfPercent):
  • profiler/ProfileNode.h:
  • profiler/Profiler.h: (KJS::Profiler::currentProfile):
6:06 PM Changeset in webkit [33469] by andersca@apple.com
  • 2 edits
    1 add in trunk/WebKitSite

2008-05-14 Anders Carlsson <​andersca@apple.com>

Reviewed by Mitz.

Make our Sticky notes example an offline web app.

  • misc/DatabaseExample.html:
  • misc/DatabaseExample.manifest: Added.
6:02 PM Changeset in webkit [33468] by alp@webkit.org
  • 3 edits in trunk/WebCore

2008-05-14 Alp Toker <​alp@nuanti.com>

GTK+ build fix. Include value.h from kjs, not JavaScriptCore.

  • page/JavaScriptProfile.cpp:
  • page/JavaScriptProfileNode.cpp:
5:43 PM Changeset in webkit [33467] by Adam Roben
  • 2 edits in trunk/WebCore

Windows release build fix

  • WebCore.vcproj/WebCore.vcproj: Make sure the warnings we disable on XPathGrammar.cpp are disabled for all configurations, not just Debug_Internal.
5:25 PM Changeset in webkit [33466] by kmccullough@apple.com
  • 17 edits
    5 moves in trunk

JavaScriptCore:

2008-05-14 Kevin McCullough <​kmccullough@apple.com>

Reviewed by Sam.

<rdar://problem/5770054> JavaScript profiler (10928)

  • Rename FunctionCallProfile to ProfileNode.
  • GNUmakefile.am:
  • JavaScriptCore.exp:
  • JavaScriptCore.pri:
  • JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
  • JavaScriptCore.xcodeproj/project.pbxproj:
  • JavaScriptCoreSources.bkl:
  • profiler/FunctionCallProfile.cpp: Removed.
  • profiler/FunctionCallProfile.h: Removed.
  • profiler/Profile.cpp: (KJS::Profile::Profile): (KJS::Profile::willExecute):
  • profiler/Profile.h: (KJS::Profile::callTree):
  • profiler/ProfileNode.cpp: Copied from profiler/FunctionCallProfile.cpp. (KJS::ProfileNode::ProfileNode): (KJS::ProfileNode::willExecute): (KJS::ProfileNode::didExecute): (KJS::ProfileNode::addChild): (KJS::ProfileNode::findChild): (KJS::ProfileNode::stopProfiling): (KJS::ProfileNode::selfTime): (KJS::ProfileNode::printDataInspectorStyle): (KJS::ProfileNode::printDataSampleStyle): (KJS::ProfileNode::endAndRecordCall):
  • profiler/ProfileNode.h: Copied from profiler/FunctionCallProfile.h. (KJS::ProfileNode::create): (KJS::ProfileNode::children):
  • profiler/Profiler.cpp:

WebCore:

2008-05-14 Kevin McCullough <​kmccullough@apple.com>

Reviewed by Sam.

<rdar://problem/5770054> JavaScript profiler (10928)

  • Rename FunctionCallProfile to ProfileNode
  • ForwardingHeaders/profiler/FunctionCallProfile.h: Removed.
  • ForwardingHeaders/profiler/ProfileNode.h: Copied from ForwardingHeaders/profiler/FunctionCallProfile.h.
  • GNUmakefile.am:
  • WebCore.pro:
  • WebCore.vcproj/WebCore.vcproj:
  • WebCore.xcodeproj/project.pbxproj:
  • WebCoreSources.bkl:
  • page/JavaScriptFunctionCallProfile.cpp: Removed.
  • page/JavaScriptFunctionCallProfile.h: Removed.
  • page/JavaScriptProfile.cpp:
  • page/JavaScriptProfileNode.cpp: Copied from page/JavaScriptFunctionCallProfile.cpp. (WebCore::ProfileNodeCache): (WebCore::getFunctionName): (WebCore::getTotalTime): (WebCore::getSelfTime): (WebCore::getNumberOfCalls): (WebCore::getChildren): (WebCore::finalize): (WebCore::ProfileNodeClass): (WebCore::toJS):
  • page/JavaScriptProfileNode.h: Copied from page/JavaScriptFunctionCallProfile.h.
5:02 PM Changeset in webkit [33465] by Beth Dakin
  • 6 edits in trunk/WebCore

2008-05-13 Beth Dakin <Beth Dakin>

Reviewed by Brady.

This patch adds support for the ARIA progressbar role and the
following ARIA states (ie HTML attributes): aria-checked, aria-
level, aria-pressed, aria-valuenow, aria-valuemin, and aria-
valuemax.

  • html/HTMLAttributeNames.in:
  • page/AccessibilityObject.h: (WebCore::AccessibilityObject::isProgressIndicator): (WebCore::AccessibilityObject::valueForRange): (WebCore::AccessibilityObject::maxValueForRange): (WebCore::AccessibilityObject::minValueForRange):
  • page/AccessibilityRenderObject.cpp: (WebCore::AccessibilityRenderObject::isProgressIndicator): (WebCore::AccessibilityRenderObject::isPressed): Check the aria- pressed attribute if this is an ARIA button. (WebCore::AccessibilityRenderObject::headingLevel): Check the aria- level attribute if this is an ARIA heading. (WebCore::AccessibilityRenderObject::intValue): Check the aria- checked attribute if this is an ARIA radio button or checkbox. (WebCore::AccessibilityRenderObject::valueForRange): (WebCore::AccessibilityRenderObject::maxValueForRange): (WebCore::AccessibilityRenderObject::minValueForRange): (WebCore::RoleEntry::): (WebCore::AccessibilityRenderObject::canSetValueAttribute):
  • page/AccessibilityRenderObject.h:
  • page/mac/AccessibilityObjectWrapper.mm: (-[AccessibilityObjectWrapper accessibilityAttributeNames]): (-[AccessibilityObjectWrapper accessibilityAttributeValue:]):
4:02 PM Changeset in webkit [33464] by kmccullough@apple.com
  • 6 edits in trunk

JavaScriptCore:

2008-05-14 Kevin McCullough <​kmccullough@apple.com>

Reviewed by John.

<rdar://problem/5770054> JavaScript profiler (10928)

  • Have each FunctionCallProfile be able to return it's total and self time.
  • JavaScriptCore.exp:
  • profiler/FunctionCallProfile.cpp: (KJS::FunctionCallProfile::selfTime):
  • profiler/FunctionCallProfile.h: (KJS::FunctionCallProfile::totalTime):

WebCore:

2008-05-14 Kevin McCullough <​kmccullough@apple.com>

Reviewed by John.

-<rdar://problem/5770054> JavaScript profiler (10928)
Use the FunctionCallProfile's new total and self time functions.

  • page/JavaScriptFunctionCallProfile.cpp: (WebCore::getTotalTime): (WebCore::getSelfTime):
3:50 PM Changeset in webkit [33463] by andersca@apple.com
  • 1 edit
    1 add in trunk/LayoutTests

2008-05-14 Anders Carlsson <​andersca@apple.com>

Check in the test result.


  • http/tests/appcache/manifest-with-empty-file-expected.txt: Added.
3:44 PM Changeset in webkit [33462] by andersca@apple.com
  • 3 edits
    3 adds in trunk

WebCore:

2008-05-14 Anders Carlsson <​andersca@apple.com>

Reviewed by Adam.

Don't try to store the resource data if it's empty.


  • loader/appcache/ApplicationCacheStorage.cpp: (WebCore::ApplicationCacheStorage::store):

LayoutTests:

2008-05-14 Anders Carlsson <​andersca@apple.com>

Reviewed by Adam.

Add test where a manifest references an empty file.


  • http/tests/appcache/manifest-with-empty-file.html: Added.
  • http/tests/appcache/resources/empty.txt: Added.
  • http/tests/appcache/resources/manifest-with-empty-file.manifest: Added.
3:35 PM Changeset in webkit [33461] by Adam Roben
  • 3 edits in trunk/WebCore

Windows build fix

  • platform/network/NetworkStateNotifier.cpp: Added a missing #include.
  • platform/network/NetworkStateNotifier.h: (WebCore::NetworkStateNotifier::NetworkStateNotifier): Removed 'void' return type from the constructor, and fixed the spelling of a member variable.
3:26 PM Changeset in webkit [33460] by kmccullough@apple.com
  • 7 edits
    6 adds in trunk/WebCore

2008-05-14 Kevin McCullough <​kmccullough@apple.com>

Reviewed by Adam.

<rdar://problem/5770054> JavaScript profiler (10928)

  • ForwardingHeaders/profiler/FunctionCallProfile.h: Added.
  • ForwardingHeaders/profiler/Profile.h: Added.
  • GNUmakefile.am: Add the new files to all the projects.
  • WebCore.pro:
  • WebCore.vcproj/WebCore.vcproj:
  • WebCore.xcodeproj/project.pbxproj:
  • WebCoreSources.bkl:
  • page/InspectorController.cpp: Add the ability to retrieve all of the profiles in JavaScript from the InspectorController. (WebCore::allProfiles): (WebCore::InspectorController::windowScriptObjectAvailable):
  • page/JavaScriptFunctionCallProfile.cpp: Added. Wrapper around the JSC FunctionCallProfile. (WebCore::functionCallProfileCache): This cache makes sure we don't needlessly create duplicate wrappers around the same FunctionCallProfile*. (WebCore::getFunctionName): Callback to return the name of the function. (WebCore::getTotalTime): Callback to return the total time spent in the function. (WebCore::getSelfTime): Callback to return the time spent in self (total
  • time in children) in the function. (WebCore::getNumberOfCalls): Callback to return the number of times this function was called. (WebCore::getChildren): Callback to return the children of this function where the children are the functions this function called. (WebCore::finalize): Cleanup. (WebCore::functionCallProfileClass): Define the JSClassRef for this object. (WebCore::toJS): Create a conversion function to make a usable JSValue to wrap the FunctionCallProfile in JSC.
  • page/JavaScriptFunctionCallProfile.h: Added.
  • page/JavaScriptProfile.cpp: Added. Wrapper around the JSC Profile. (WebCore::profileCache): This cache makes sure we don't needlessly create duplicate wrappers aroudn the same Profile*. (WebCore::getHeadCallback): Callback to return the head of the graph of the functions profiled durring this profile's run. (WebCore::finalize): (WebCore::profileClass): Define the JSClassRef for this object. (WebCore::toJS): Create a conversion function to make a usable JSValue to wrap the Profile from JSC.
  • page/JavaScriptProfile.h: Added.
3:20 PM Changeset in webkit [33459] by Adam Roben
  • 2 edits in trunk/WebKit/win

Fix Bug 18767: Inspector is currently pretty much useless on Windows

<​https://bugs.webkit.org/show_bug.cgi?id=18767>

Reviewed by Tim Hatcher.

  • WebCoreSupport/WebInspectorClient.cpp: (WebInspectorClient::highlight): Only move the highlight behind the Inspector window when first creating the highlight, not every time we update it.
3:07 PM Changeset in webkit [33458] by hyatt@apple.com
  • 2 edits
    2 adds in trunk

2008-05-14 David Hyatt <​hyatt@apple.com>

Fix for bug ​https://bugs.webkit.org/show_bug.cgi?id=16130.

Opacity, transforms, reflections and masks should not honor z-index unless the element is also
positioned. Always reset to auto if the element is not positioned (and then the following code will
adjust auto to 0 as needed).

Reviewed by Dan Bernstein

  • css/CSSStyleSelector.cpp: (WebCore::CSSStyleSelector::adjustRenderStyle):
2:30 PM Changeset in webkit [33457] by andersca@apple.com
  • 8 edits
    3 adds in trunk/WebCore

2008-05-14 Anders Carlsson <​andersca@apple.com>

Reviewed by Brady.

<rdar://problem/5931742> Support navigator.onLine from HTML5


  • WebCore.vcproj/WebCore.vcproj:
  • WebCore.xcodeproj/project.pbxproj: Add NetworkStateNotifier files.


  • dom/EventNames.h: Add offline and online events.


  • page/Navigator.cpp: (WebCore::Navigator::onLine):
  • page/Navigator.h:
  • page/Navigator.idl: Add navigator.onLine.


  • page/Page.cpp: (WebCore::networkStateChanged): Dispatch offline/online events to all frames.


(WebCore::Page::Page):
Set state change callback.


  • platform/network/NetworkStateNotifier.cpp: Added.
  • platform/network/NetworkStateNotifier.h: Added.
  • platform/network/mac/NetworkStateNotifierMac.cpp: Added. Mac network state notifier.
2:26 PM Changeset in webkit [33456] by eric@webkit.org
  • 2 edits in trunk/WebKit/mac

Reviewed by Oliver.

Add missing NULL check to match rest of file, this was found by the editing fuzzer.

  • WebView/WebResource.mm: (-[WebResource data]):
2:17 PM Changeset in webkit [33455] by Adam Roben
  • 1 edit in trunk/WebCore/ChangeLog

Add some Bugzilla URLs

2:14 PM Changeset in webkit [33454] by Adam Roben
  • 2 edits in trunk/WebCore

Fix Bug 18888: REGRESSION (r32731): console.log always prints "undefined"

Reviewed by Tim Hatcher.

Tested by pre-existing
manual-tests/inspector/console-log-formatting.html.

  • page/inspector/Console.js: (WebInspector.ConsoleMessage.prototype._format): Changed to explicitly wrap Console._format instead of using bind to do it. Using bind was causing us to pass plainText in as the first parameter instead of as the second.
1:59 PM Changeset in webkit [33453] by timothy@apple.com
  • 2 edits in trunk/WebCore

Fixes the bug where stepping out of a function would show the wrong
call stack (would look like it stepped out twice.)

Reviewed by Adam Roben.

  • page/JavaScriptDebugServer.cpp:

(WebCore::JavaScriptDebugServer::returnEvent):
Call pauseIfNeeded before setting m_currentCallFrame to the caller.

1:40 PM Changeset in webkit [33452] by Stephanie Lewis
  • 2 edits in trunk/WebCore

2008-05-14 Stephanie Lewis <Stephanie Lewis>

Fix 64bit build

  • plugins/PluginView.h:
1:36 PM Changeset in webkit [33451] by ddkilzer@apple.com
  • 6 edits in trunk

WebCore:

REGRESSION (r26474): WebKit fails jQuery test 64 core module: text(String) subtest 1 Check escaped text (createTextNode)

<​https://bugs.webkit.org/show_bug.cgi?id=18976>
<rdar://problem/5924793>

Reviewed by Darin.

Tests: fast/dom/dom-parse-serialize-expected.txt:

fast/xsl/xslt-processor-expected.txt:
fast/xsl/xslt-text-expected.txt:

Commit r26474 removed escaping of '>' characters to '&gt;' in serialized
HTML content, which broke one jQuery test. This change restores that
serialization behavior to both attributes and content for consistency.

  • editing/markup.cpp: (WebCore::appendAttributeValue): Escape '>' as '&gt;'. (WebCore::escapeContentText): Ditto. (WebCore::appendEscapedContent): Ditto. This is the bug fix for the jQuery test.

LayoutTests:

REGRESSION (r26474): WebKit fails jQuery test 64 core module: text(String) subtest 1 Check escaped text (createTextNode)

<​https://bugs.webkit.org/show_bug.cgi?id=18976>
<rdar://problem/5924793>

Reviewed by Darin.

Updated test results now that '>' is escaped to '&gt;' in both
attributes and content.

  • fast/xsl/xslt-processor-expected.txt:
  • fast/xsl/xslt-text-expected.txt:
  • platform/mac/fast/dom/dom-parse-serialize-expected.txt:
1:04 PM Changeset in webkit [33450] by Adam Roben
  • 3 edits in trunk/WebCore

Fix Bug 19051: REGRESSION (r33447): Assertion in JSDebugServer::removeListener closing tab with Inspector open

m_inspectedPage is now cleared when the Page is destroyed, which never
used to happen. Some of InspectorController's code was not prepared
for this.

Reviewed by Dave Hyatt.

  • page/InspectorController.cpp: (WebCore::InspectorController::~InspectorController): Removed a call to stopDebugging and added an assertion. Debugging should have already been stopped by this point. (WebCore::InspectorController::inspectedPageDestroyed): Moved here from the header file. Added a call to stopDebugging. (WebCore::InspectorController::enabled): Return false if our inspected Page has already been destroyed, since we can't really do much at this point anyway. (WebCore::InspectorController::didCommitLoad): (WebCore::InspectorController::didLoadResourceFromMemoryCache): (WebCore::InspectorController::identifierForInitialRequest): (WebCore::InspectorController::startDebuggingAndReloadInspectedPage): (WebCore::InspectorController::stopDebugging): (WebCore::InspectorController::drawNodeHighlight): Assert that m_inspectedPage hasn't been cleared.
  • page/InspectorController.h:
12:30 PM Changeset in webkit [33449] by justin.garcia@apple.com
  • 10 edits in trunk

WebCore:

2008-05-14 Justin Garcia <​justin.garcia@apple.com>

Reviewed by Darin.

<rdar://problem/5914803> Improve performance of WebCore::Editor::setComposition

  • editing/Editor.cpp: (WebCore::Editor::confirmComposition): Remove the previous composition when we insert the new one, not with a separate, slower, delete operation. (WebCore::Editor::setComposition): Ditto.
  • editing/InsertTextCommand.cpp: (WebCore::InsertTextCommand::performTrivialReplace): Remove the selected text with a low level operation that doesn't perform a layout and insert the new text in a way that won't trigger a layout from the removal. (WebCore::InsertTextCommand::input): Call the optimized replace.
  • editing/InsertTextCommand.h:
  • editing/htmlediting.cpp: (WebCore::isTabSpanNode): Check to see if the node is a span, to avoid the expense of getAttribute in the common case.
  • page/Frame.cpp: (WebCore::Frame::selectionLayoutChanged): Selection::start() and end() will already be at VisiblePosition deepEquivalents. Selection::validate() ensures this.

LayoutTests:

2008-05-14 Justin Garcia <​justin.garcia@apple.com>

Reviewed by Darin.

<rdar://problem/5914803> Improve performance of WebCore::Editor::setComposition


Replacing all the text in a node is now much less likely to remove it
and cause selection changes.

  • platform/mac/editing/input/text-input-controller-expected.txt:
  • platform/mac/editing/inserting/4959067-expected.txt:
  • platform/mac/editing/style/style-3681552-fix-001-expected.txt:
12:01 PM Changeset in webkit [33448] by ap@webkit.org
  • 2 edits in trunk/WebKit/mac

Reviewed by Dan Bernstein.

NPP_ValidAttributesForMarkedText should return NSArray*, not NSArray.

  • Plugins/nptextinput.h:
11:33 AM Changeset in webkit [33447] by Adam Roben
  • 4 edits in trunk/WebCore

Make the Inspector able to handle being reloaded

When Reload was selected from the context menu in the Inspector, the
Inspector window would close and never be openable again. Now it
closes, then reopens when the page finishes loading.

Reviewed by Tim Hatcher.

  • page/InspectorController.cpp: (WebCore::InspectorController::close): Removed code that cleared out m_page. This is now handled by pageDestroyed.
  • page/InspectorController.h: (WebCore::InspectorController::inspectedPageDestroyed): Added.
  • page/Page.cpp: (WebCore::Page::~Page): Fixed a bug and added a call to the new inspectedPageDestroyed. If we have a parent InspectorController, that means that we are the Page in the Inspector's window (i.e., we are the Inspector's UI), so when we are destroyed we should call InspectorController::pageDestroyed. We also need to call inspectedPageDestroyed on our own InspectorController so that it won't try to call into us.
11:21 AM Changeset in webkit [33446] by jchaffraix@webkit.org
  • 3 edits in trunk/WebKitTools

2008-05-14 Julien Chaffraix <​jchaffraix@webkit.org>

Reviewed by Eric.


  • isDarwin() and isCygwin() returned an empty string if the platform equality check fails. We now force the return value to be numeric.
  • Removed platform checks as it was a work around the previous issue.
  • Replaced isDarwin() by isOSX() as they have now the same behaviour.
  • Scripts/build-webkit:
  • Scripts/webkitdirs.pm:
10:55 AM Changeset in webkit [33445] by timothy@apple.com
  • 8 edits
    1 add in trunk/WebCore

Add pause on exception support to the Inspector debugger
with a toggle buttons to enable or disable it.

Reviewed by Adam Roben.

  • page/InspectorController.cpp:

(WebCore::pauseOnExceptions): Call the InspectorController.
(WebCore::setPauseOnExceptions): Ditto.
(WebCore::InspectorController::windowScriptObjectAvailable): Add the new
setPauseOnExceptions and pauseOnExceptions functions to the JS class.
(WebCore::InspectorController::pauseOnExceptions): Call JavaScriptDebugServer.
(WebCore::InspectorController::setPauseOnExceptions): Ditto.

  • page/InspectorController.h:
  • page/JavaScriptDebugServer.cpp:

(WebCore::JavaScriptDebugServer::setPauseOnExceptions): Set m_pauseOnExceptions.
(WebCore::JavaScriptDebugServer::exception): Check m_pauseOnExceptions
and set m_pauseOnNextStatement to true before calling pauseIfNeeded.

  • page/JavaScriptDebugServer.h:
  • page/inspector/Images/pauseOnExceptionButtons.png: Added.
  • page/inspector/ScriptsPanel.js: Adds a status bar button that calls

the new InspectorController functions.

  • page/inspector/inspector.css: New styles.
10:55 AM Changeset in webkit [33444] by timothy@apple.com
  • 4 edits in trunk/WebCore

Rename the Scope Chain pane to Scope Variables. Name the global
object Global. Name objects before the locals "With Block".
Name colsure scopes "Closure". Name any element or document
after Local, "Event Target" and "Event Document", respectfully.

Reviewed by Adam Roben.

  • English.lproj/localizedStrings.js: Add new strings.
  • page/inspector/ObjectPropertiesSection.js:

(WebInspector.ObjectPropertiesSection): Add a new argument that is used
for the empty placeholder text.
(WebInspector.ObjectPropertiesSection.prototype.onpopulate): Use the
empty placeholder text.

  • page/inspector/ScopeChainSidebarPane.js:

(WebInspector.ScopeChainSidebarPane): Rename Scope Chain to Scope Variables.
(WebInspector.ScopeChainSidebarPane.prototype.update): Use the object name as
the subtitle. Use "Global" for the global object title. Use "With Block" for
the title of objects before the locals. Use "No Variables" for the empty
placeholder text for locals.

9:34 AM Changeset in webkit [33443] by ap@webkit.org
  • 3 edits
    3 adds in trunk

Reviewed by Darin.

<rdar://problem/5934376> REGRESSION: A script fails because of a straw BOM character in it.

<​https://bugs.webkit.org/show_bug.cgi?id=4931>
Unicode format characters (Cf) should be removed from JavaScript source

Of all Cf characters, we are only removing BOM, because this is what Firefox trunk has
settled upon, after extensive discussion and investigation.

Based on Darin's work on this bug.

Test: fast/js/removing-Cf-characters.html

  • kjs/lexer.cpp: (KJS::Lexer::setCode): Tweak formatting. Use a call to shift(4) to read in the first characters, instead of having special case code here. (KJS::Lexer::shift): Add a loop when reading a character to skip BOM characters.
9:18 AM Changeset in webkit [33442] by ap@webkit.org
  • 3 edits
    2 adds in trunk

2008-05-14 Michael A. Puls II <​shadow2531@gmail.com>

Reviewed by Alexey.

Fix for ​https://bugs.webkit.org/show_bug.cgi?id=16923
Attr nodes with a value of "" should not have any childNodes


In Opera, Firefox and IE, when an attribute node has a value
of "", the attribute node doesn't have any childNodes. This
is true in WebKit also, but not when you assign "" to the
Attr's value when the existing value is non-empty.


Test: fast/dom/attribute-empty-value-no-children.html

  • dom/Attr.cpp: (WebCore::Attr::setValue): Use createTextChild(), which only appends a child text node if the value being assigned is not empty.
8:50 AM Changeset in webkit [33441] by jchaffraix@webkit.org
  • 10 edits in trunk/WebCore

2008-05-14 Julien Chaffraix <​jchaffraix@webkit.org>

Reviewed by Eric.

GCC 4.3 build fix.
Removes some compiler warnings.

  • dom/Document.h: (WebCore::FormElementKey::hashTableDeletedValue): GCC complained about the const keyword so remove it as it is only used inside FormElementKey.
  • html/HTMLSelectElement.cpp: (WebCore::HTMLSelectElement::menuListDefaultEventHandler): Correct coding style (removes a warning too).
  • loader/appcache/ApplicationCacheResource.cpp: Include stdio.h for gcc4.3 build.
  • plugins/PluginPackage.cpp: (WebCore::PluginPackage::PluginPackage): Changed initialization order in the constructor.
  • plugins/PluginView.cpp: (WebCore::PluginView::PluginView): Ditto.
  • plugins/PluginView.h:
  • xml/AccessControlList.cpp: Include stdio.h
  • xml/AccessItem.cpp: Ditto.
  • xml/AccessItemRule.cpp: Ditto.
7:31 AM Changeset in webkit [33440] by Simon Hausmann
  • 2 edits
    1 add in trunk/WebCore

2008-05-14 Ariya Hidayat <​ariya.hidayat@trolltech.com>

Reviewed by Simon.

Qt build fix. Add JSJavaScriptCallFrameCustom.cpp to the build
and implemented EventLoopQt.

5:53 AM Changeset in webkit [33439] by Simon Hausmann
  • 2 edits in trunk/WebCore

2008-05-14 Ariya Hidayat <​ariya.hidayat@trolltech.com>

Reviewed by Simon.

Qt build fix. Add JavaScriptCallFrame.{idl,cpp} to the build.

5:14 AM Changeset in webkit [33438] by oliver@apple.com
  • 6 edits
    2 adds in branches/squirrelfish

Bug 19024: SQUIRRELFISH: ASSERTION FAILED: activation->isActivationObject() in Machine::unwindCallFrame
<​https://bugs.webkit.org/show_bug.cgi?id=19024>

Reviewed by Maciej

This fixes a number of issues. The most important is that we now check every register
file for tainting rather than just looking for function register files as that was
insufficient. Additionally guarded against implicit re-entry into Eval code.

Also added a few additional assertions to reduce the amout of time between something
going wrong and us seeing the error.

1:45 AM Changeset in webkit [33437] by ggaren@apple.com
  • 8 edits in branches/squirrelfish/JavaScriptCore

2008-05-14 Geoffrey Garen <​ggaren@apple.com>

Reviewed by Oliver Hunt.


A little more debugger action: emit opcodes for debugger hooks. Right
now, the opcode implementation is just a stub.


SunSpider reports no change.


Some example codegen for "function f() { 1; }":

[ 0] dbg DidEnterCallFrame
[ 2] dbg WillExecuteStatement
[ 4] load tr0, 1(@k0)
[ 7] load tr0, undefined(@k1)
[ 10] dbg WillLeaveCallFrame
[ 12] ret tr0

12:25 AM Changeset in webkit [33436] by oliver@apple.com
  • 3 edits
    2 adds in branches/squirrelfish

Bug 19025: SQUIRRELFISH: malformed syntax in onload handler causes crash
<​https://bugs.webkit.org/show_bug.cgi?id=19025>

Reviewed by Geoff

Simple fix -- move the use of functionBodyNode to after the null check.

12:24 AM Changeset in webkit [33435] by oliver@apple.com
  • 3 edits
    1 copy
    2 adds in branches/squirrelfish/LayoutTests

Split the binops tests up a bit more

RS=Geoff

Note: See TracTimeline for information about the timeline view.