⚠ Archived content — this site is no longer maintained.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

Timeline



Jul 23, 2007:

11:54 PM Changeset in webkit [24564] by oliver
  • 2 edits in trunk/WebCore

Reviewed by Steve.

Windows build fix, remove superfluous !

  • page/EventHandler.cpp: (WebCore::EventHandler::keyEvent):
11:31 PM Changeset in webkit [24563] by sfalken
  • 2 edits in trunk/WebKit/win

Fixes to versioning script.


Reviewed by Ada.

  • WebKit.vcproj/auto-version.sh:
9:11 PM Changeset in webkit [24562] by aroben
  • 7 edits in trunk

Implement IWebHistoryItem::[set]AlternateTitle

WebCore:

Add an assignment operator for BString that takes a BSTR

Reviewed by Geoff and Oliver.

  • platform/win/BString.cpp: (WebCore::BString::operator=):
  • platform/win/BString.h:

WebKit/win:

Implement IWebHistoryItem::[set]AlternateTitle

The argument types were reversed for these two methods, so I fixed that as
well.

Reviewed by Geoff and Oliver.

  • Interfaces/IWebHistoryItem.idl:
  • WebHistoryItem.cpp: (WebHistoryItem::setAlternateTitle): Implemented. (WebHistoryItem::alternateTitle): Implemented.
  • WebHistoryItem.h:
8:59 PM Changeset in webkit [24561] by justing
  • 3 edits
    8 adds in trunk

LayoutTests:

Reviewed by Geoff.


<rdar://problem/5354455> Word selection when right-clicking can be confusing (14728)

  • editing/selection/5354455-1-expected.checksum: Added.
  • editing/selection/5354455-1-expected.png: Added.
  • editing/selection/5354455-1-expected.txt: Added.
  • editing/selection/5354455-1.html: Added.
  • editing/selection/5354455-2-expected.checksum: Added.
  • editing/selection/5354455-2-expected.png: Added.
  • editing/selection/5354455-2-expected.txt: Added.
  • editing/selection/5354455-2.html: Added.

WebCore:

Reviewed by Geoff.


<rdar://problem/5354455> Word selection when right-clicking can be confusing (14728)

  • page/EventHandler.cpp: (WebCore::EventHandler::sendContextMenuEvent): In non-editable content, only do word selection over text. In editable content, we will continue to select images, line breaks and other elements on right-clicks, to match TextEdit. For now, in editable content, we'll live with the cases where positionForPoint creates selections that aren't underneath the mouse. These aren't regressions because we've always done word selection on right clicks in editable content.
8:09 PM Changeset in webkit [24560] by andersca
  • 3 edits
    2 adds in trunk

LayoutTests:

Reviewed by Oliver.

<rdar://problem/5335354>
REGRESSION(r21359-r21368): Can't edit inside RTEF editable region


  • fast/dom/Document/open-with-pending-load-expected.txt: Added.
  • fast/dom/Document/open-with-pending-load.html: Added.

WebCore:

Reviewed by Oliver.

<rdar://problem/5335354>
REGRESSION(r21359-r21368): Can't edit inside RTEF editable region


If Document::open is called and there's a pending load that has not yet started,
that load should be cancelled.


  • dom/Document.cpp: (WebCore::Document::open):
8:01 PM Changeset in webkit [24559] by andersca
  • 2 edits in trunk/JavaScriptCore

Reviewed by Geoff.

<rdar://problem/5121461> REGRESSION: Unable to load JigZone puzzle


  • bindings/jni/jni_jsobject.cpp: (JavaJSObject::createNative):


Call RootObject::gcProtect on the global object, thereby putting it in the
"protect count" set which is used for checking if a native handle is valid.

7:36 PM RunWebKitTestsDesign edited by pewtermoose@gmail.com
(diff)
6:22 PM Changeset in webkit [24558] by andersca
  • 2 edits in trunk/LayoutTests

Reviewed by Mark.

Remove http/tests/multipart/invalid-image-data.html which no longer fails on Leopard.


  • mac/leopard/Skipped:
6:00 PM Changeset in webkit [24557] by justing
  • 6 edits in trunk/LayoutTests

<rdar://problem/5228167> REGRESSION: editing/execCommand/5119244.html failing on Leopard


Removed another use of the &#x23B7 entity. See below.

  • editing/execCommand/5119244-expected.checksum:
  • editing/execCommand/5119244-expected.png:
  • editing/execCommand/5119244-expected.txt:
  • editing/execCommand/5119244.html:
  • mac/leopard/Skipped:
5:46 PM Changeset in webkit [24556] by justing
  • 6 edits in trunk/LayoutTests

Reviewed by Oliver.


<rdar://problem/5228164> REGRESSION: editing/deleting/5144139-2.html fails on Leopard


Removed the use of the &#x23B7 entity, which is available on
Leopard but not Tiger. Perhaps because it is not available in Ahem
but is available in some fallback font that ships with Leopard but not Tiger.
It's OK to remove because its use wasn't necessary to test the fix for 5144139.

  • editing/deleting/5144139-2-expected.checksum:
  • editing/deleting/5144139-2-expected.png:
  • editing/deleting/5144139-2-expected.txt:
  • editing/deleting/5144139-2.html:
  • mac/leopard/Skipped:
5:44 PM Changeset in webkit [24555] by bdash
  • 4 edits in tags/Safari-5522.12.1

Versioning.

5:44 PM Changeset in webkit [24554] by bdash
  • 4 edits in tags/Safari-5522.12.1

Merge r24545.

Build fix, newer Leopard builds expect GCC_ENABLE_OBJC_GC to have a value of 'supported'.

5:39 PM Changeset in webkit [24553] by adachan
  • 5 edits in trunk/WebKit/win

2007-07-23 Ada Chan <​adachan@apple.com>

Reviewed by Steve.


Added a new text drawing method that allows caller to override the font smoothing level.

  • WebKit.vcproj/WebKit.def:
  • WebKit.vcproj/WebKit_debug.def:
  • WebKitGraphics.cpp: (WebDrawText):
  • WebKitGraphics.h:


5:37 PM Changeset in webkit [24552] by adachan
  • 4 edits in trunk/WebKitLibraries

2007-07-23 Ada Chan <​adachan@apple.com>

Reviewed by Steve.


Update WebKitSystemInterface.{h,lib}.

  • win/include/WebKitSystemInterface/WebKitSystemInterface.h:
  • win/lib/WebKitSystemInterface.lib:
  • win/lib/WebKitSystemInterface_debug.lib:
5:33 PM Changeset in webkit [24551] by bdash
  • 1 copy in tags/Safari-5522.12.1

New tag.

5:04 PM Changeset in webkit [24550] by thatcher
  • 5 edits in trunk/WebCore

Reviewed by Hyatt.

<rdar://problem/5242145> REGRESSION: Clicking on symbol in documentation often doesn't scroll to symbol

A renderer for the anchor wasn't always available at the time parsing finished. So we need
to bail out of gotoAnchor if stylesheets are pending and remember to call gotoAnchor later
once all of the pending stylesheets load.

  • dom/Document.cpp: (WebCore::Document::Document): Initialize m_gotoAnchorNeededAfterStylesheetsLoad to false. (WebCore::Document::stylesheetLoaded): If we have no more pending stylesheets, call gotoAnchor if needed.
  • dom/Document.h: (WebCore::Document::gotoAnchorNeededAfterStylesheetsLoad): New method. (WebCore::Document::setGotoAnchorNeededAfterStylesheetsLoad): Ditto.
  • loader/FrameLoader.cpp: (WebCore::FrameLoader::gotoAnchor): Bail early if the document still has pending stylesheets.
  • loader/FrameLoader.h: Make gotoAnchor() public.
4:16 PM Changeset in webkit [24549] by sullivan
  • 2 edits in trunk/WebCore

Reviewed by Darin


  • fixed <rdar://problem/5327887> Printing Mail note (or Safari page) with misspelled word prints red underline
  • rendering/InlineTextBox.cpp: (WebCore::InlineTextBox::paintSpellingOrGrammarMarker): When printing, bail out without doing any work. Also moved a couple of lines down below another early bailout since they were pointless above it.
3:17 PM Changeset in webkit [24548] by sfalken
  • 1 copy in tags/Safari-522.15.3b

New tag.

2:58 PM Changeset in webkit [24547] by treat
  • 2 edits in trunk/WebKitTools

Fix qt DRT to suppress js popup alerts and log instead.

2:35 PM Changeset in webkit [24546] by darin
  • 2 edits in trunk/JavaScriptCore
  • pcre/pcre_compile.c: Roll back a tiny accidental change in the unused !JAVASCRIPT side of an #ifdef. This has no effect when using PCRE in JAVASCRIPT mode as we do, but seems worth rolling back.
2:29 PM Changeset in webkit [24545] by thatcher
  • 4 edits in trunk

Build fix, newer Leopard builds expect GCC_ENABLE_OBJC_GC to have a value of 'supported'.

2:19 PM Changeset in webkit [24544] by sfalken
  • 6 edits in branches/Safari-522-2/JavaScriptCore

Merged fix from r24453.

2:15 PM Changeset in webkit [24543] by sfalken
  • 3 edits
    2 copies in branches/Safari-522-2

Merged fix from r24430.

2:06 PM TestFormUpload created by andersca@apple.com
2:06 PM WebKit Team edited by andersca@apple.com
(diff)
2:03 PM Changeset in webkit [24542] by oliver
  • 8 edits
    3 adds in trunk

LayoutTests:

Reviewed by Darin.

Added initial pseudo-hangul IM, and tests for behaviour of hangul (<rdar://problem/5352152>)
Updated keydown-keypress-preventDefault to be correct, and to have an demonstrate that keys have been inserted.
Updated kotoeri test to perform more validation and check against updated behaviour

  • editing/input/mac/hangul-enter-confirms-and-sends-keypress-expected.txt: Added.
  • editing/input/mac/hangul-enter-confirms-and-sends-keypress.html: Added.
  • editing/input/mac/hangul.js: Added.
  • editing/input/mac/kotoeri-enter-to-confirm-and-newline.html:
  • fast/events/keydown-keypress-preventDefault-expected.txt:
  • fast/events/keydown-keypress-preventDefault.html:

WebCore:

Reviewed by Darin.


<rdar://problem/5353577> WebKit key event behaviour needs to better match other browsers keyevent behaviour
<rdar://problem/5352152> REGRESSION (24399-24479): Pressing Enter in a search field at apple.com no longer starts a search with Korean IM
<rdar://problem/5352649> WebKit should send keyDown event on autorepeat keyDowns
​http://bugs.webkit.org/show_bug.cgi?id=14690

Perform Input Method call first so that our behaviour is determined by whether or not
the input method has actually handle the event, rather than by making guesses based
on the existences of marked text. This fixes issues with IMs that have side effects
even on events they do not handle.

Refactored the code in order to improve clarity given the hoisting required by performing
the IM call earlier.

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

WebKit:

Reviewed by Darin.

We have to be able to support insertText: followed by doCommandBySelector: in
order to support the 2- and 3-Set Korean and RuSwitcher IMs at least.

  • WebView/WebHTMLView.mm: (-[WebHTMLView insertText:]):
1:27 PM Changeset in webkit [24541] by andersca
  • 9 edits
    3 adds in trunk

LayoutTests:

Reviewed by Darin.

<rdar://problem/5045711>
​http://bugs.webkit.org/show_bug.cgi?id=12938
Google calendar settings page crashes

  • fast/dom/destroy-selected-radio-button-crash-expected.txt: Added.
  • fast/dom/destroy-selected-radio-button-crash.html: Added.

WebCore:

Reviewed by Darin.

<rdar://problem/5045711>
​http://bugs.webkit.org/show_bug.cgi?id=12938
Google calendar settings page crashes


Move handling of checked radio buttons to a new class, HTMLFormElement::CheckedRadioButtons.
Each <form> element has an instance of this class. For radio buttons without a containing
form, the document has an instance of HTMLFormElement::CheckedRadioButtons where the
state of those radio buttons will be stored.
be stored.


This also fixes another bug where removing a checked radio button that does not have a
containing form would not remove the radio button from the checked radio buttons map,
which could lead to a crash due to a dangling pointer.


  • dom/Document.cpp: (WebCore::Document::~Document):
  • dom/Document.h: (WebCore::Document::checkedRadioButtons):
  • html/HTMLFormElement.cpp: (WebCore::HTMLFormElement::registerFormElement): (WebCore::HTMLFormElement::removeFormElement): (WebCore::HTMLFormElement::CheckedRadioButtons::didCheckButton): (WebCore::HTMLFormElement::CheckedRadioButtons::checkedButtonForGroup): (WebCore::HTMLFormElement::CheckedRadioButtons::removeButtonIfChecked):
  • html/HTMLFormElement.h: (WebCore::HTMLFormElement::checkedRadioButtons):
  • html/HTMLGenericFormElement.cpp: (WebCore::HTMLGenericFormElement::insertedIntoTree):
  • html/HTMLInputElement.cpp: (WebCore::HTMLInputElement::~HTMLInputElement): (WebCore::checkedRadioButtons): (WebCore::HTMLInputElement::isKeyboardFocusable): (WebCore::HTMLInputElement::setInputType): (WebCore::HTMLInputElement::parseMappedAttribute): (WebCore::HTMLInputElement::setChecked): (WebCore::HTMLInputElement::preDispatchEventHandler): (WebCore::HTMLInputElement::willMoveToNewOwnerDocument):
  • manual-tests/remove-form-node-with-radio-buttons-crash.html: Added.
1:23 PM Changeset in webkit [24540] by aliceli1
  • 3 edits in trunk/WebKit

Reverting change 24535 now that a solution has been found that doesn't
involve exposing an interface unnecessarily.

  • WebView/WebHTMLView.mm:
  • WebView/WebHTMLViewPrivate.h:
11:51 AM Changeset in webkit [24539] by hyatt
  • 4 adds in trunk/LayoutTests/fast/inline

Add layout test for continuation hang.

11:50 AM Changeset in webkit [24538] by hyatt
  • 3 edits in trunk/WebCore

Fix for 5350587, hang in iExploder text 8101. The code that painted outlines for continuations was buggy
in that it assumed the chain was always inline-block-inline, but this is only true for the innermost chained
continuation. To make this more robust, we always just use the inline's containing block's containing block,
which should be guaranteed to enclose all renderers in the continuation chain. In addition, there is now a
null check (to guarantee no more hangs) and an assert (to check if we hit this situation again where the block
used to paint does not properly enclose the continuation chain).

Reviewed by beth

fast/inline/outline-continuation.html

  • rendering/RenderBlock.cpp: (WebCore::RenderBlock::paintContinuationOutlines):
  • rendering/InlineFlowBox.cpp: (WebCore::RenderInline::paint):
11:16 AM Changeset in webkit [24537] by zecke
  • 16 edits
    1 move
    21 adds in trunk

2007-07-22 Holger Hans Peter Freyther <​zecke@selfish.org>

Reviewed by Adam.

Add WebKit/gtk/Api and WebCoreSupport API to the INCLUDEPATH to allow GdkLauncher to be
build against it.

  • WebKit.pri:

WebCore:
2007-07-22 Holger Hans Peter Freyther <​zecke@selfish.org>

Reviewed by Adam.

WARNING: NO TEST CASES ADDED OR CHANGED

Add the WebKit/Gtk API to the buildsystem, ask qmake to create a
pkg-config file and install headers and libraries.

Start to emit signals from FrameLoaderClientGdk, hide the kit Frame
inside the the FrameLoaderClientGdk.

Move ChromeClientGdk to WebKit/gtk/WebCoreSupport

  • WebCore.pro:
  • loader/gdk/FrameLoaderClientGdk.cpp:
  • loader/gdk/FrameLoaderClientGdk.h:
  • platform/gdk/ChromeClientGdk.h: Removed.
  • platform/gdk/FrameGdk.cpp: Remove code not belonging here
  • platform/gdk/FrameGdk.h:
  • platform/gdk/ScrollViewGdk.cpp: The gdkDrawable won't exist at that time and it is fine
  • platform/gdk/TemporaryLinkStubs.cpp: Removed ChromeClientGdk stubs

WebKit:
2007-07-22 Holger Hans Peter Freyther <​zecke@selfish.org>

Reviewed by Adam.

Add the first API and partial implementation of the WebKit/Gtk port as
of ​http://bugs.webkit.org/show_bug.cgi?id=14678.

The delegates of WebView will be modeled as signals. Delegates like the
WebPolicyDelegate will be implemented as signals with default handlers.

Start moving files to WebKit/gtk/WebCoreSupport.

  • gtk: Added.
  • gtk/Api: Added.
  • gtk/Api/headers.pri: Added.
  • gtk/Api/webkitgtk-marshal.list: Added.
  • gtk/Api/webkitgtkdefines.h: Added.
  • gtk/Api/webkitgtkframe.cpp: Added.
  • gtk/Api/webkitgtkframe.h: Added.
  • gtk/Api/webkitgtkframedata.cpp: Added.
  • gtk/Api/webkitgtkframedata.h: Added.
  • gtk/Api/webkitgtkglobal.cpp: Added.
  • gtk/Api/webkitgtkglobal.h: Added.
  • gtk/Api/webkitgtknetworkrequest.cpp: Added.
  • gtk/Api/webkitgtknetworkrequest.h: Added.
  • gtk/Api/webkitgtkpage.cpp: Added.
  • gtk/Api/webkitgtkpage.h: Added.
  • gtk/Api/webkitgtkprivate.cpp: Added.
  • gtk/Api/webkitgtkprivate.h: Added.
  • gtk/Api/webkitgtksettings.cpp: Added.
  • gtk/Api/webkitgtksettings.h: Added.
  • gtk/WebCoreSupport: Added.
  • gtk/WebCoreSupport/ChromeClientGdk.cpp: Added. (WebCore::ChromeClientGdk::ChromeClientGdk): (WebCore::ChromeClientGdk::chromeDestroyed): (WebCore::ChromeClientGdk::windowRect): (WebCore::ChromeClientGdk::setWindowRect): (WebCore::ChromeClientGdk::pageRect): (WebCore::ChromeClientGdk::scaleFactor): (WebCore::ChromeClientGdk::focus): (WebCore::ChromeClientGdk::unfocus): (WebCore::ChromeClientGdk::createWindow): (WebCore::ChromeClientGdk::createModalDialog): (WebCore::ChromeClientGdk::show): (WebCore::ChromeClientGdk::canRunModal): (WebCore::ChromeClientGdk::runModal): (WebCore::ChromeClientGdk::setToolbarsVisible): (WebCore::ChromeClientGdk::toolbarsVisible): (WebCore::ChromeClientGdk::setStatusbarVisible): (WebCore::ChromeClientGdk::statusbarVisible): (WebCore::ChromeClientGdk::setScrollbarsVisible): (WebCore::ChromeClientGdk::scrollbarsVisible): (WebCore::ChromeClientGdk::setMenubarVisible): (WebCore::ChromeClientGdk::menubarVisible): (WebCore::ChromeClientGdk::setResizable): (WebCore::ChromeClientGdk::closeWindowSoon): (WebCore::ChromeClientGdk::canTakeFocus): (WebCore::ChromeClientGdk::takeFocus): (WebCore::ChromeClientGdk::canRunBeforeUnloadConfirmPanel): (WebCore::ChromeClientGdk::runBeforeUnloadConfirmPanel): (WebCore::ChromeClientGdk::addMessageToConsole): (WebCore::ChromeClientGdk::runJavaScriptAlert): (WebCore::ChromeClientGdk::runJavaScriptConfirm): (WebCore::ChromeClientGdk::runJavaScriptPrompt): (WebCore::ChromeClientGdk::setStatusbarText): (WebCore::ChromeClientGdk::shouldInterruptJavaScript): (WebCore::ChromeClientGdk::tabsToLinks): (WebCore::ChromeClientGdk::windowResizerRect): (WebCore::ChromeClientGdk::addToDirtyRegion): (WebCore::ChromeClientGdk::scrollBackingStore): (WebCore::ChromeClientGdk::updateBackingStore): (WebCore::ChromeClientGdk::mouseDidMoveOverElement): (WebCore::ChromeClientGdk::setToolTip): (WebCore::ChromeClientGdk::print):
  • gtk/WebCoreSupport/ChromeClientGdk.h: Added.

WebKitTools:
2007-07-22 Holger Hans Peter Freyther <​zecke@selfish.org>

Reviewed by Adam.

Make the GdkLauncher use the new WebKit/Gtk API. Change webkitdirs.pm to
honor --qmakearg for the Gdk/Gtk build as well.

  • GdkLauncher/main.cpp: Switch to the new API
  • Scripts/webkitdirs.pm: Allow to specify --qmakearg, e.g. to control the WEBKIT_{INC,LIB}_DIR
10:16 AM Changeset in webkit [24536] by aroben
  • 5 edits in trunk/WebKit/win

Implement IDOMElement::setAttribute

This method was mistakenly called "setResult" in DOMCore.idl, so I
fixed that as well.

Needed for <rdar://problem/5314906>.

Reviewed by Ada.

  • DOMCoreClasses.cpp: (DOMElement::setAttribute): Implemented/renamed.
  • DOMCoreClasses.h: Renamed setResult => setAttribute.
  • DOMHTMLClasses.h: Ditto.
  • Interfaces/DOMCore.idl: Ditto.
9:37 AM Changeset in webkit [24535] by aliceli1
  • 3 edits in trunk/WebKit

Reviewed by Oliver Hunt.

Expose [WebHTMLView hasMarkedText] to
fix <rdar://problem/4830074> autocomplete breaks Japanese typing

  • WebView/WebHTMLView.mm:
  • WebView/WebHTMLViewPrivate.h:
3:26 AM Changeset in webkit [24534] by mjs
  • 2 edits in trunk/JavaScriptCore

Reviewed by Oliver.


  • fix remaining problems with Window shadowing
  • kjs/nodes.cpp: (VarDeclNode::evaluate): Tweak the special case a little.
1:48 AM Changeset in webkit [24533] by mjs
  • 2 edits in trunk/JavaScriptCore

Reviewed by Oliver.


  • fix Window shadowing regressions caused by the previous commit.
  • kjs/nodes.cpp: (VarDeclNode::evaluate): Handle the case of global scope specially.
12:10 AM Changeset in webkit [24532] by mjs
  • 3 edits in trunk/JavaScriptCore

Reviewed by Darin.

-fixed <rdar://problem/5353293> REGRESSION (r24287): 1% i-Bench JS slowdown from JavaScript compatibility fix (14719)
​http://bugs.webkit.org/show_bug.cgi?id=14719


My fix for this actually resulted in JS iBench being 1% faster than before the regression
and the Celtic Kane benchmark being 5% faster than before the regression.


  • kjs/nodes.cpp: (VarDeclNode::handleSlowCase): factored out the slow code path to be out of line. (VarDeclNode::evaluate): I did a couple of things: (1) Don't check if the variable is already declared by looking for the property in the variable object, that code path was dead code. (2) Special-case the common case where the top of the scope and the variable object are the same; in that case the variable must always be in the variable object. (3) Don't return a jsString() of the variable name, nothing uses the return value from this node types evaluate method.
  • kjs/nodes.h:

Jul 22, 2007:

11:53 PM RunWebKitTestsDesign edited by Adam Roben
(diff)
10:45 PM Changeset in webkit [24531] by mjs
  • 5 edits in trunk/WebCore

Reviewed by Adam.

<rdar://problem/5353174> REGRESSION: 1% JavaScript performance regression fro Window refactoring (14717)
​http://bugs.webkit.org/show_bug.cgi?id=14717


  • bindings/js/kjs_window.cpp:
  • bindings/js/kjs_window.h: (KJS::Window::impl): inlined
  • page/DOMWindow.cpp:
  • page/DOMWindow.h: (WebCore::DOMWindow::frame): inlined
10:42 PM Changeset in webkit [24530] by weinig
  • 2 edits in trunk/LayoutTests

Fix typo.

  • fast/dom/Window/window-function-name-getter-precedence.html:
10:30 PM Changeset in webkit [24529] by darin
  • 2 edits in trunk/WebCore
  • platform/ContextMenu.cpp: Added missing include.
10:28 PM Changeset in webkit [24528] by weinig
  • 3 edits
    2 adds in trunk

LayoutTests:

Reviewed by Darin.

Test for ​http://bugs.webkit.org/show_bug.cgi?id=14682
REGRESSION(24285-24399): alerts do not appear at W3C RTL test page

  • fast/dom/Window/window-function-name-getter-precedence-expected.txt: Added.
  • fast/dom/Window/window-function-name-getter-precedence.html: Added.

WebCore:

Reviewed by Darin.

Fix for ​http://bugs.webkit.org/show_bug.cgi?id=14682
REGRESSION(24285-24399): alerts do not appear at W3C RTL test page

Test: fast/dom/Window/window-function-name-getter-precedence.html

  • bindings/js/JSDOMWindowCustom.cpp: (WebCore::JSDOMWindow::customGetOwnPropertySlot): Make sure to return functions before the custom name getter.
10:25 PM Changeset in webkit [24527] by darin
  • 2 edits in trunk/WebCore

Reviewed by Adam.

  • fix <rdar://problem/5353263> ContextMenuItem and ContextMenu objects leaking (seen on buildbot)
  • platform/ContextMenu.cpp: (WebCore::separatorItem): Return an auto_ptr instead of a raw pointer, since this function creates an object that's the caller's responsibility to delete. (WebCore::createAndAppendFontSubMenu): Create the context menu on the stack, not the heap, since setSubMenu does not take ownership, and hence the objects were leaking. (WebCore::createAndAppendSpellingAndGrammarSubMenu): Ditto. (WebCore::createAndAppendSpellingSubMenu): Ditto. (WebCore::createAndAppendSpeechSubMenu): Ditto. (WebCore::createAndAppendWritingDirectionSubMenu): Ditto.
10:00 PM Changeset in webkit [24526] by pewtermoose
  • 2 edits
    6 adds in trunk/WebKitSite

Reviewed by Adam Roben.

Add information on getting a useful crash log on Mac and Windows.

  • nav.inc:
  • quality/crashlogs.html: Added.
  • quality/mac_reportbox.jpg: Added.
  • quality/mac_viewtrace.jpg: Added.
  • quality/win_installwatson.jpg: Added.
  • quality/win_watsongui.jpg: Added.
  • quality/win_watsoninstalled.jpg: Added.
9:38 PM Changeset in webkit [24525] by darin
  • 2 edits in trunk/WebKitTools
  • DumpRenderTree/TextInputController.m: (-[TextInputController interpretKeyEvents:withSender:]): Fix a leak by releasing the array used here.
7:59 PM RunWebKitTestsDesign edited by Adam Roben
(diff)
7:59 PM RunWebKitTestsDesign edited by Adam Roben
(diff)
7:53 PM RunWebKitTestsDesign edited by pewtermoose@gmail.com
(diff)
7:53 PM RunWebKitTestsDesign edited by Adam Roben
(diff)
7:51 PM RunWebKitTestsDesign edited by Adam Roben
(diff)
3:47 PM Changeset in webkit [24524] by darin
  • 4 edits in trunk

JavaScriptCore:

Reviewed by Kevin Decker.

  • bindings/objc/objc_utility.mm: (KJS::Bindings::convertObjcValueToValue): If the object returns 0 for _imp, convert that to "undefined", since callers can't cope with a JSValue of 0.

WebCore:

Reviewed by Kevin Decker.

  • bindings/objc/WebScriptObject.mm: (_didExecute): Removed unnecessary check for isValid() since the _rootObject method already takes care of that check. (-[WebScriptObject _imp]): Made this return 0 if there is not a valid rootObject. (-[WebScriptObject _isSafeScript]): Made this always return false if there is not a valid rootObject, eliminating the need to check _rootObject for nil if the code is checking _isSafeScript. (-[WebScriptObject callWebScriptMethod:withArguments:]): Removed check of _rootObject, since it's immediately followed by a check of _isSafeScript. Removed awkward conversion of name from NSString * to KJS::Identifier and did it using WebCore::String instead -- makes it clear that there's no arbitrary JavaScript execution after _isSafeScript and before _imp, so we don't need to null-check the _imp pointer. Added a second check of _isSafeScript after calling get to get the function object, since arbitrary changes could occur in there, including navigation to a new frame. This also takes care of null checking the second call to _imp. (-[WebScriptObject evaluateWebScript:]): Removed check of _rootObject, since it's immediately followed by a check of _isSafeScript. Removed awkward conversion of script from NSString * to KJS::Identifier and did it using WebCore::String instead -- makes it clear that there's no arbitrary JavaScript execution after _isSafeScript and before _imp, so we don't need to null-check the _imp pointer. (-[WebScriptObject setValue:forKey:]): Removed check of _rootObject, since it's immediately followed by a check of _isSafeScript. Removed awkward conversion of key from NSString * to KJS::Identifier and did it using WebCore::String instead -- makes it clear that there's no arbitrary JavaScript execution after _isSafeScript and before _imp, so we don't need to null-check the _imp pointer. (-[WebScriptObject valueForKey:]): Ditto. (-[WebScriptObject removeWebScriptKey:]): Ditto. (-[WebScriptObject stringRepresentation]): Removed check of _rootObject, since it's immediately followed by a check of _isSafeScript. Removed unnecessary local variable and const_cast from the call to the _imp method. (-[WebScriptObject webScriptValueAtIndex:]): Removed check of _rootObject, since it's immediately followed by a check of _isSafeScript. (-[WebScriptObject setWebScriptValueAtIndex:value:]): Ditto. (-[WebScriptObject JSObject]): Ditto.
3:24 PM Changeset in webkit [24523] by ddkilzer
  • 2 edits
    1 add in trunk/WebKitTools

WebKitTools:

Reviewed by Darin.

Added script to update WebKitTools/iExploder/htdocs/cssproperties.in based on the contents
of WebCore/css/CSSPropertyNames.in. Also updated cssproperties.in.

  • Scripts/update-iexploder-cssproperties: Added.
  • iExploder/htdocs/cssproperties.in: Updated by running update-iexploder-cssproperties script. Added new CSS3 property section and Moved box-sizing property to it.
2:10 PM Changeset in webkit [24522] by oliver
  • 4 edits
    1 move in trunk/LayoutTests

LayoutTests:

Reviewed by Sam.

Moving input method tests to editing/input/mac

  • editing/input/mac: Copied from fast/events/inputmethods.
  • fast/events/inputmethods: Removed.
  • fast/events/inputmethods/kotoeri-enter-to-confirm-and-newline-expected.txt: Removed.
  • fast/events/inputmethods/kotoeri-enter-to-confirm-and-newline.html: Removed.
  • fast/events/inputmethods/kotoeri.js: Removed.
  • fast/events/inputmethods/logger.js: Removed.
  • win/Skipped:

qt:

Reviewed by Sam.


Add Mac input methods to qt skip list

  • Skipped:
1:32 PM Changeset in webkit [24521] by darin
  • 1 edit in trunk/WebCore/ChangeLog

Updated reviewers.

1:31 PM Changeset in webkit [24520] by darin
  • 3 edits in trunk/WebCore

Reviewed by Anders.

  • speculative fix for <rdar://problem/5337500> CrashTracer: [USER] 4 crashes in Mail at WebCore::HTMLImageElement::isURLAttribute(WebCore::Attribute*) const
  • editing/markup.cpp: Added AttributeChange class. (WebCore::completeURLs): Changed function so that all the URL completion is done in a separate pass after finding all the URL attributes. This is safer, since actually applying an attribute change could have any arbitrary effect on the document; it's tricky to iterate a document while it's being modified and we don't have the checks here that would be needed to make that work in pathological cases.
1:02 PM Changeset in webkit [24519] by oliver
  • 2 edits in trunk/LayoutTests

Reviewed by Adam.

Add the input methods tests to the windows skip list
until we work to how to make them platform agnostic

  • win/Skipped:
12:51 PM Changeset in webkit [24518] by darin
  • 1 edit in trunk/WebCore/ChangeLog

Update Bugzilla bug number in Radar bug title based on feedback from Mitz.

12:46 PM Changeset in webkit [24517] by oliver
  • 5 edits
    5 adds in trunk

LayoutTests:

Reviewed by Darin.

Initial input method test case, this tests the behaviour of the
enter key in our "kotoeri" engine.

  • fast/events/inputmethods: Added.
  • fast/events/inputmethods/kotoeri-enter-to-confirm-and-newline-expected.txt: Added.
  • fast/events/inputmethods/kotoeri-enter-to-confirm-and-newline.html: Added.
  • fast/events/inputmethods/kotoeri.js: Added.
  • fast/events/inputmethods/logger.js: Added.

WebKitTools:

Reviewed by Darin.

​http://bugs.webkit.org/show_bug.cgi?id=14710

Add preliminary support for testing Input Method/WebKit behaviour and interaction
in DRT. This provides the NSTextInput API which is most of what should be necessary
to mimic the event sequences Input Methods trigger.

  • DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
  • DumpRenderTree/TextInputController.h:
  • DumpRenderTree/TextInputController.m: (-[WebHTMLView interpretKeyEvents:]): (+[TextInputController isSelectorExcludedFromWebScript:]): (+[TextInputController webScriptNameForSelector:]): (-[TextInputController initWithWebView:]): (-[TextInputController dealloc]): (-[TextInputController textInput]): (-[TextInputController setInputMethodHandler:]): (-[TextInputController interpretKeyEvents:withSender:]):
12:43 PM Changeset in webkit [24516] by darin
  • 2 edits in trunk/WebCore

Reviewed by Adam.

  • fix <rdar://problem/5350604> REGRESSION: Crash in inspector controller code after clicking back button in test application (13250)
  • page/InspectorController.cpp: (WebCore::InspectorController::addScriptResource): Replace an assertion that was firing with some code that at least doesn't immediately crash.
12:08 PM Changeset in webkit [24515] by darin
  • 2 edits in trunk/WebKit
  • StringsNotToBeLocalized.txt: Updated for recent changes.
11:43 AM Changeset in webkit [24514] by darin
  • 2 edits in trunk/WebCore

Reviewed by Anders.

  • improve the fix for <rdar://problem/5298816> Crash redirecting message, in -[SharedBufferData initWithSharedBuffer:]; fixes a crash seen on the buildbot
  • loader/DocLoader.cpp: (WebCore::DocLoader::~DocLoader): Call setDocLoader(0) on all the resources so they don't try to use a stale DocLoader pointer later.
Note: See TracTimeline for information about the timeline view.