Timeline



Jan 23, 2008:

11:44 PM Changeset in webkit [29755] by mrowe@apple.com
  • 4 edits in trunk

Versioning.

11:42 PM Changeset in webkit [29754] by mrowe@apple.com
  • 1 copy in tags/Safari-5525.6

New tag.

4:47 PM Changeset in webkit [29753] by ddkilzer@apple.com
  • 2 edits in trunk/WebCore

<rdar://problem/5702947> WebCore: CGContextDrawPDFDocument will be deprecated

Reviewed by Darin.

No test cases added since there is no change in behavior.

  • platform/graphics/cg/PDFDocumentImage.cpp: (WebCore::PDFDocumentImage::draw): Replace CGContextDrawPDFDocument() with calls to CG methods that perform the equivalent work.
3:08 PM Changeset in webkit [29752] by Antti Koivisto
  • 4 edits in trunk/WebCore

Reviewed by Hyatt, Adele.

Timers for media controls keep running when the page is in the page cache.


Stop timers in RenderMedia and get rid of the controls tree when the page goes to the cache.

  • html/HTMLMediaElement.cpp: (WebCore::HTMLMediaElement::HTMLMediaElement): (WebCore::HTMLMediaElement::willSaveToCache): (WebCore::HTMLMediaElement::didRestoreFromCache):
  • html/HTMLMediaElement.h: (WebCore::HTMLMediaElement::inPageCache):
  • rendering/RenderMedia.cpp: (WebCore::RenderMedia::updateControls):
2:26 PM Changeset in webkit [29751] by Antti Koivisto
  • 2 edits in trunk/WebCore

Reviewed by Adam.

Video that had already played to the end would start playing (from beginning) on back/forward navigation.


Use setPausedInternal() instead of pause(). This avoid generating pause/play events and matches specification text.

  • html/HTMLMediaElement.cpp: (WebCore::HTMLMediaElement::willSaveToCache):
2:20 PM Changeset in webkit [29750] by ap@webkit.org
  • 5 edits
    6 adds in trunk

Reviewed by Darin.

<rdar://problem/4200075> Missing support for accented chars in mailto forms

With some combinations of form parameters, this was resulting in regressed behavior.

Tests: fast/forms/mailto/get-non-ascii-text-plain-latin-1.html

fast/forms/mailto/post-text-plain-with-accept-charset.html
fast/forms/mailto/post-text-plain.html

  • html/HTMLFormElement.cpp: (WebCore::HTMLFormElement::formData): Ignore accept-charset for mailto forms, which are always encoded as UTF-8. (WebCore::HTMLFormElement::isMailtoForm): Added. (WebCore::HTMLFormElement::dataEncoding): Added (factored out from formData). (WebCore::HTMLFormElement::submit): When round-tripping in text/plain case, use utf-8 encoding.
  • html/HTMLFormElement.h: Added a isMailtoForm() function. It may be a bit unfortunate that its result is calculated repeatedly, but this doesn't seem to be anywhere near hot code.
2:04 PM Changeset in webkit [29749] by sfalken@apple.com
  • 7 edits in trunk

WebCore:

Reviewed by Alice Liu.

<rdar://problem/5658727> Undo/redo for pasted or typed text in yahoo mail becomes disabled after one round of being selected

  • rendering/RenderTextControl.cpp: (WebCore::RenderTextControl::updateFromElement): If value == text() and m_innerText is empty, then value and text() are "". In that case both DOM mutations are no-ops and there is no reason to clear the Undo stack.

WebKit/win:

<rdar://problem/5698732> Copyright strings should include 2008


Reviewed by Sam.

  • WebKit.vcproj/WebKit.rc:

WebKitTools:

<rdar://problem/5698732> Copyright strings should include 2008

Reviewed by Sam.

  • Drosera/win/Drosera.vcproj/Drosera.rc:
  • DumpRenderTree/win/TestNetscapePlugin/TestNetscapePlugin.rc:
1:51 PM Changeset in webkit [29748] by justin.garcia@apple.com
  • 3 edits
    2 adds in trunk

WebCore:

Reviewed by Alice Liu.

<rdar://problem/5658727> Undo/redo for pasted or typed text in yahoo mail becomes disabled after one round of being selected

  • rendering/RenderTextControl.cpp: (WebCore::RenderTextControl::updateFromElement): If value == text() and m_innerText is empty, then value and text() are "". In that case both DOM mutations are no-ops and there is no reason to clear the Undo stack.

LayoutTests:

Reviewed by Alice Liu.


<rdar://problem/5658727> Undo/redo for pasted or typed text in yahoo mail becomes disabled after one round of being selected

  • editing/undo/5658727-expected.txt: Added.
  • editing/undo/5658727.html: Added.
1:08 PM Changeset in webkit [29747] by Adam Roben
  • 2 edits in trunk/WebKit/win

Fix <rdar://5698502> Inspector window reverts to default position/size unexpectedly

Reviewed by Hyatt.

  • WebInspectorClient.cpp: (defaultWindowRect): Added. (WebInspectorClient::createPage):
    • Use defaultWindowRect() to set the initial size/position of the Inspector window.
    • Pass the Inspector window's client rect to IWebView::initWithFrame so that the WebView will be initially sized to fill the window.

(WebInspectorClient::showWindow): Don't move/resize the window back to
the defaults every time this method is called.

1:07 PM Changeset in webkit [29746] by Adam Roben
  • 3 edits
    2 adds in trunk

Fix <rdar://5700620> REGRESSION (r28190): "Open frame in new tab" instead of "Open link in new tab" @ digg.com

WebKit/win:

Fix <rdar://5700620> REGRESSION (r28190): "Open frame in new tab" instead of "Open link in new tab" @ digg.com

Test: fast/events/contextmenu-scrolled-page-with-frame.html

Reviewed by Dan.

  • WebView.cpp: (WebView::handleContextMenuEvent): Translate the mouse coordinates into document coordinates before performing the hit-test. This is what EventHandler does internally.

LayoutTests:

Add a test for <rdar://5700620> REGRESSION (r28190): "Open frame in
new tab" instead of "Open link in new tab" @ digg.com

Reviewed by Dan.

  • fast/events/contextmenu-scrolled-page-with-frame-expected.txt: Added.
  • fast/events/contextmenu-scrolled-page-with-frame.html: Added.
11:43 AM Changeset in webkit [29745] by sfalken@apple.com
  • 9 edits in trunk

WebCore:

<rdar://problem/5699509> Allow file upload dialog to be localized.

Reviewed by Darin.

  • platform/LocalizedStrings.h:
  • platform/win/FileChooserWin.cpp: (WebCore::FileChooser::openFileChooser):

WebKit:

Update localization exceptions.


Reviewed by Darin.

  • English.lproj/Localizable.strings:
  • StringsNotToBeLocalized.txt:

WebKit/win:

<rdar://problem/5699509> Allow file upload dialog to be localized.


Reviewed by Darin.

  • English.lproj/Localizable.strings: Updated.
  • WebCoreLocalizedStrings.cpp: (WebCore::uploadFileText): Added. (WebCore::allFilesText): Added.
9:49 AM Changeset in webkit [29744] by dsmith@webkit.org
  • 1 edit
    19 adds in trunk/WebKitSite

Reviewed by Darin.


Added a modified version of the MooTools SlickSpeed Selector Test/Benchmark to show off querySelectorAll. MIT licensed.

  • perf/slickspeed: Added.
  • perf/slickspeed/config.ini: Added.
  • perf/slickspeed/footer.html: Added.
  • perf/slickspeed/frameworks: Added.
  • perf/slickspeed/frameworks/DomQuery.js: Added.
  • perf/slickspeed/frameworks/dummy.js: Added.
  • perf/slickspeed/frameworks/jquery-1.2.1.js: Added.
  • perf/slickspeed/frameworks/prototype.js: Added.
  • perf/slickspeed/header.html: Added.
  • perf/slickspeed/index.php: Added.
  • perf/slickspeed/instructions.txt: Added.
  • perf/slickspeed/logo.png: Added.
  • perf/slickspeed/selectors.list: Added.
  • perf/slickspeed/style.css: Added.
  • perf/slickspeed/system: Added.
  • perf/slickspeed/system/index.php: Added.
  • perf/slickspeed/system/slickspeed.js: Added.
  • perf/slickspeed/system/template.php: Added.
  • perf/slickspeed/template.html: Added.
9:38 AM Changeset in webkit [29743] by Adam Roben
  • 2 edits in trunk/WebCore

Fix behavior of type-to-select in <select> menus

In r27825 we started posting both a WM_KEYDOWN and WM_CHAR message
whenever a key was pressed in the dropdown menu. However, since posted
messages go through TranslateMessage, the WM_KEYDOWN was generating
another WM_CHAR message, so two WM_CHAR messages in a row would reach
the WebView. This caused the searching to happen twice for each key,
so typing "dog" would search for "ddoogg".

Reviewed by Darin.

  • platform/win/PopupMenuWin.cpp: (WebCore::PopupWndProc): Repost the WM_KEYDOWN message to the WebView. This is will generate a WM_CHAR message which will trigger the type-to-select behavior. This is very similar to our behavior prior to r27825, except that we're now calling PostMessage instead of SendMessage.
8:18 AM Changeset in webkit [29742] by mitz@apple.com
  • 4 edits in trunk/LayoutTests
  • platform/mac/fast/frames/viewsource-attribute-expected.checksum:
  • platform/mac/fast/frames/viewsource-attribute-expected.png:
  • platform/mac/fast/frames/viewsource-attribute-expected.txt:
7:55 AM Changeset in webkit [29741] by alp@webkit.org
  • 8 edits in trunk

2008-01-23 Alp Toker <alp@atoker.com>

Rubber-stamped by Mark Rowe.

Remove whitespace after -I in automake include lists.

3:16 AM Changeset in webkit [29740] by Simon Hausmann
  • 2 edits in trunk/WebKit/qt

Holger Hans Peter Freyther <holger.freyther@trolltech.com>

  • Fix leaking of sub frames (WebCore::Frame). We keep one reference too many. This was found while working on the page-cache, other ports are not affected.
  • Rubber stamped by Simon
3:14 AM Changeset in webkit [29739] by Simon Hausmann
  • 2 edits in trunk/WebCore

Holger Hans Peter Freyther <holger.freyther@trolltech.com>

  • Load images incrementally. This will show warnings on the console and we want to fix them for Qt4.4. If that happens we have to remove the comment from the enum inside the ImageDecoderQt.cpp
  • Rubber stamped by Simon
3:11 AM Changeset in webkit [29738] by Simon Hausmann
  • 2 edits in trunk/WebKit/qt

Rohan McGovern <rohan.mcgovern@trolltech.com>

Fix Qtopia compilation with QT_NO_TOOLTIP

3:11 AM Changeset in webkit [29737] by Simon Hausmann
  • 2 edits in trunk/WebKit/qt

Geir Vattekar <gvatteka@trolltech.com>

Doc: Replaced \code with snippets in the docs

3:10 AM Changeset in webkit [29736] by Simon Hausmann
  • 2 edits in trunk/WebCore

Lars Knoll <lars@trolltech.com>

Fix rendering of the Scrollbar as well as mouse handling for some styles.

The QStyle expects that that painter is set up to clip to the scrollbar bounds
and some of the styles seem to paint somewhat outside of these bounds. Clipping to
the scrollbar bounds before drawing removes some artifacts.

Also set m_opt.rect.topLeft to (0/0) in the mouse handlers. Fixes some issues with
the oxygen style.

1:55 AM Changeset in webkit [29735] by Simon Hausmann
  • 3 edits in trunk/WebCore

Michael Brasser <michael.brasser@trolltech.com>

Rendering fixes for Qtopia.

Make the focus ring work with line breaks (until we have QPainterPath::simplify) and
call QStyle with better arguments when painting themed elements.

1:55 AM Changeset in webkit [29734] by Simon Hausmann
  • 3 edits in trunk/WebCore

Holger Hans Peter Freyther <holger.freyther@trolltech.com>

  • FormData::flatten does not include to be uploaded files. Create our own QIODevice for the QNetworkAccessManager and feed the complete content of the FormData* to the (HTTP) backend.
  • Try to optimize the ::readData implementation to fill the buffer as best as possible
1:55 AM Changeset in webkit [29733] by Simon Hausmann
  • 2 edits in trunk/WebCore

Holger Hans Peter Freyther <holger.freyther@trolltech.com>

  • Coding-Style fixes
1:48 AM Changeset in webkit [29732] by Simon Hausmann
  • 2 edits in trunk/WebCore

Michael Brasser <michael.brasser@trolltech.com>

use PopupClient's font for popup menu. Otherwise, in Qtopia, the popup's font is much bigger (QApplication::font()?) and the text doesn't fit.

1:44 AM Changeset in webkit [29731] by Simon Hausmann
  • 2 edits in trunk/WebKit/qt

Benjamin Meyer <bmeyer@trolltech.com>

Tweak key presses even to match commonly expected behavior

  • space key == page down
  • page down moved down not a page, but slightly less then a page so you don't loose your spot when reading.
  • Use font height rather then a hard coded "10" for left, right, up, down
  • Ctrl-Up moves to the top of the page
  • Ctrl-Down move to the bottom of the page
  • Backspace == GoBack
  • Shift-Backspace == GoForward
1:43 AM Changeset in webkit [29730] by Simon Hausmann
  • 7 edits in trunk

Implemented FrameLoaderClient::startDownload() and FrameLoaderClient::download().

Added two signals to QWebPage to handle downloading of links and handling of
unsupported content.

1:39 AM Changeset in webkit [29729] by Simon Hausmann
  • 12 edits in trunk

Michael Goddard <michael.goddard@trolltech.com>

Reworked the JavaScriptCore Qt bindings:

  • Add initial support for string and variant arrays, as well as sub QObjects in the JS bindings.
  • Don't expose fields marked as not scriptable by moc.
  • Add support for dynamic properties and accessing named QObject children of an object (like QtScript and older IE DOM style JS).
  • Add support for custom toString methods.
  • Fine tune some bindings to be closer to QtScript. Make void functions return undefined, and empty/ null QStrings return a zero length string.
  • Create framework for allowing more direct method calls. Since RuntimeMethod doesn't allow us to add additional methods/properties to a function, add these classes. Start prototyping object.signal.connect(...).
  • Add signal support to the Qt bindings. Allow connecting to signals (object.signal.connect(slot)), disconnecting, and emitting signals. Currently chooses the first signal that matches the name, so this will need improvement.
  • Add property names, and resolve signals closer to use. Enumerating properties now returns some of the Qt properties and signals. Slots and methods aren't quite present. Also, resolve signal connections etc. closer to the time of use, so we can do more dynamic resolution based on argument type etc. Still picks the first one with the same name, at the moment.
  • Make signature comparison code consistent. Use the same code for checking meta signatures in the method and fallback getters, and avoid a QByteArray construction when we can.
  • Fix minor memory leak, and handle pointers better. Delete the private object in the dtors, and use RefPtrs for holding Instances etc.
  • Handle method lookup better. Allow invocation time method lookup based on the arguments, which is closer to QtScript behaviour. Also, cache the method lists and delete them in the QtClass dtor (stops a memory leak).
  • Improve JS to Qt data type conversions. Add some support for Date & RegExp JS objects, and provide some metrics on the quality of the conversion.
  • A couple of fixes for autotest failures. Better support for converting lists, read/write only QMetaProperty support, modified slot search order...)
  • Update JS DRT controller for Qt JS binding changes. There were two functions that needed some changes so that the layout tests would work, so this makes a few tests pass again.
  • Bump the timeout for layout tests up to 11s. At least some tests have an internal timeout of 10 seconds, so make the waitUntilDone approach wait at least 11s. fast/dom/open-and-close-by-DOM.html is one of these - now the failure message is more accurate.

Jan 22, 2008:

8:01 PM Changeset in webkit [29728] by mitz@apple.com
  • 2 edits in trunk/WebCore

Reviewed by Adele Peterson.

The regression test for this is fast/css/css2-system-color.html which
is currently disabled.

  • rendering/RenderThemeMac.mm: (WebCore::menuBackgroundColor): Added. Uses HIThemeDrawMenuBackground to draw the menu item background into a bitmap graphics context and returns the color of the pixel at (0, 0). (WebCore::RenderThemeMac::systemColor): Changed to call menuBackgroundColor for the CSS2 menu color.
4:29 PM Changeset in webkit [29727] by Darin Adler
  • 2 edits in trunk/WebCore

Reviewed by Oliver Hunt.

  • turn full repainting back on for <canvas> until we get incremental repaint working properly

Ollie and I came up with a plan for testing this, but for now we should
just fall back on the "repaint everything" behavior we have always had.

  • html/HTMLCanvasElement.cpp: (WebCore::HTMLCanvasElement::willDraw): Just call repaint.
3:16 PM Changeset in webkit [29726] by andersca@apple.com
  • 5 edits
    2 adds in trunk

JavaScriptCore:

Reviewed by Darin and Adam.

<rdar://problem/5688975>
div element on microsoft site has wrong left offset.


Return true even if NPN_GetProperty returns null or undefined. This matches Firefox
(and is what the Silverlight plug-in expects).


  • bindings/NP_jsobject.cpp: (_NPN_GetProperty):

WebKitTools:

Reviewed by Darin and Adam.

<rdar://problem/5688975>
div element on microsoft site has wrong left offset.


Add new method for testing the return vale of NPN_GetProperty.


  • DumpRenderTree/TestNetscapePlugIn.subproj/PluginObject.cpp: (pluginInvoke):

LayoutTests:

Reviewed by Darin and Adam.

<rdar://problem/5688975>
div element on microsoft site has wrong left offset.


Add test.


  • plugins/netscape-get-property-return-value-expected.txt: Added.
  • plugins/netscape-get-property-return-value.html: Added.
1:23 PM Changeset in webkit [29725] by alp@webkit.org
  • 2 edits in trunk/WebCore

2008-01-22 Alp Toker <alp@atoker.com>

GTK+/qmake build fix.

  • WebCore.pro:
12:46 PM Changeset in webkit [29724] by ap@webkit.org
  • 4 edits in trunk

Reviewed by Adam Roben.

<rdar://problem/5670257> editing/selection/4895428-3.html makes editing/selection/5131716-2.html fail

  • DumpRenderTree/win/EventSender.cpp: (makeEventSender): Reset static variables for a new test.
12:45 PM Changeset in webkit [29723] by alp@webkit.org
  • 9 edits
    2 adds
    2 deletes in trunk

2008-01-22 Christian Dywan <christian@imendio.com>

Reviewed by Alp Toker.

[GTK] API: WebKitWebSettings is not usable
http://bugs.webkit.org/show_bug.cgi?id=16219

Implement WebKitWebSettings.

11:33 AM Changeset in webkit [29722] by Adam Roben
  • 6 edits
    9 adds in trunk

Fix <rdar://5693599> REGRESSION (r29369): scroll bars are drawing when they shouldn't be on an ad on cnn.com

WebCore:

Add HTMLFrameOwnerElement::scrollingMode

This new virtual method is overridden by HTMLFrameElementBase to
provide the scrolling mode set on the frame or iframe element. Object
elements always have scrolling=auto, so there's no need to override
this method for HTMLPlugInElement.

Reviewed by Hyatt.

  • html/HTMLFrameElementBase.h: Declare scrollingMode as virtual.
  • html/HTMLFrameOwnerElement.h: (WebCore::HTMLFrameOwnerElement::scrollingMode): Added.

WebKit/win:

Fix <rdar://5693599> REGRESSION (r29369): scroll bars are drawing when they shouldn't be on an ad on cnn.com

Now that we have multiple FrameViews per Frame on Windows, we need to
make sure each FrameView gets its scrolling mode set correctly. On Mac
this works because WebFrameView holds the scrolling mode and a single
WebFrameView is shared by many FrameViews. On Windows, we have no such
shared object, so we just ask the Frame's owner element for the
scrolling mode.

Tests: fast/frames/frame-scrolling-attribute.html

fast/frames/iframe-scrolling-attribute.html

Reviewed by Hyatt.

  • WebFrame.cpp: (WebFrame::transitionToCommittedForNewPage): Set the scrolling mode on the FrameView based on the scrolling mode of the Frame's owner element.

LayoutTests:

Tests for <rdar://5693599> REGRESSION (r29369): scroll bars are drawing when they shouldn't be on an ad on cnn.com

Reviewed by Hyatt.

  • fast/frames/frame-scrolling-attribute.html: Added.
  • fast/frames/iframe-scrolling-attribute.html: Added.
  • platform/mac/fast/frames/frame-scrolling-attribute-expected.checksum: Added.
  • platform/mac/fast/frames/frame-scrolling-attribute-expected.png: Added.
  • platform/mac/fast/frames/frame-scrolling-attribute-expected.txt: Added.
  • platform/mac/fast/frames/iframe-scrolling-attribute-expected.checksum: Added.
  • platform/mac/fast/frames/iframe-scrolling-attribute-expected.png: Added.
  • platform/mac/fast/frames/iframe-scrolling-attribute-expected.txt: Added.
9:36 AM Changeset in webkit [29721] by hyatt@apple.com
  • 2 edits in trunk/WebCore

Fix for <rdar://problem/5698481> REGRESSION: Web Inspector source view gutter double divider and no padding

Now that cellpadding cascades properly on <td>s, we need to make the view source padding rules in the user
agent sheet have !important in order to override author settings.

Reviewed by Tim

  • css/view-source.css:
8:09 AM Changeset in webkit [29720] by Nikolas Zimmermann
  • 5 edits in trunk/WebCore

Reviewed by Alp.
Allow to use SVGPaintServerSolid without RenderObject/RenderStyle. Used later to draw SVG Fonts in HTML.

6:57 AM Changeset in webkit [29719] by Nikolas Zimmermann
  • 2 edits in trunk/WebCore

Reviewed by Alp.

Initialize missing variables, m_lineGap / m_lineSpacing.
Only noticeable when trying to render HTML text using SVG Fonts (which is not possible with trunk, but with my local version).

5:53 AM Changeset in webkit [29718] by alp@webkit.org
  • 4 edits in trunk/WebCore

2008-01-22 Alp Toker <alp@atoker.com>

Reviewed by Nikolas.

Fix Cairo SVG fonts following changes in r29700.

Set the font description's details as needed in FontPlatformData and
eliminate GraphicsContext::setPlatformFont() since it was redundant
and assumed all fonts were Cairo fonts.

  • platform/graphics/GraphicsContext.cpp:
  • platform/graphics/cairo/GraphicsContextCairo.cpp:
  • platform/graphics/gtk/FontPlatformDataGtk.cpp: (WebCore::FontPlatformData::FontPlatformData):
3:10 AM Changeset in webkit [29717] by Nikolas Zimmermann
  • 3 edits
    4 adds in trunk

Reviewed by Eric.

<rdar://problem/5699344> Reproducible crash when navigating back to a page using SVG fonts

Test: svg/custom/font-platformDestroy-crash.svg

2:03 AM Changeset in webkit [29716] by Simon Hausmann
  • 2 edits in trunk/WebKit/qt

Lars Knoll <lars@trolltech.com>

setup a family for cursive and fantasy fonts as well.

2:03 AM Changeset in webkit [29715] by Simon Hausmann
  • 2 edits in trunk/WebCore

Lars Knoll <lars@trolltech.com>

cleanup the CookieJarQt implementation and implement cookiesEnabled.

2:03 AM Changeset in webkit [29714] by Simon Hausmann
  • 2 edits in trunk

Simon Hausmann <Simon Hausmann>

Don't compile the ICO plugin when building against Qt >= 4.4

2:03 AM Changeset in webkit [29713] by Simon Hausmann
  • 5 edits in trunk/WebKit/qt

Benjamin Meyer <bmeyer@trolltech.com>

Implement ChromeClientQt::setToolTip
Implement ChromeClientQt::mouseDidMoveOverElement

12:38 AM Changeset in webkit [29712] by Simon Hausmann
  • 19 edits
    1 add in trunk

Lars Knoll <lars@trolltech.com>

Ported the font implementation to use the new QFont features of Qt 4.4.

Jan 21, 2008:

10:48 PM Changeset in webkit [29711] by eric@webkit.org
  • 4 edits in trunk

Reviewed by Nikolas & Oliver.

Tested by svg/custom/svg-features.html

  • dom/DOMImplementation.cpp: (WebCore::isSVG10Feature): use ENABLE(SVG_*) conditionals (WebCore::isSVG11Feature): use ENABLE(SVG_*) conditionals
10:18 PM Changeset in webkit [29710] by ggaren@apple.com
  • 12 edits
    2 adds in trunk

JavaScriptCore:

Reviewed by Maciej Stachowiak.


Fixed http://bugs.webkit.org/show_bug.cgi?id=16909
REGRESSION: Amazon.com crash (ActivationImp)


(and a bunch of other crashes)


Plus, a .7% SunSpider speedup to boot.


Replaced the buggy currentExec and savedExec mechanisms with an
explicit ExecState stack.

  • kjs/collector.cpp: (KJS::Collector::collect): Explicitly mark the ExecState stack.

(KJS::Collector::reportOutOfMemoryToAllExecStates): Slight change in
behavior: We no longer throw an exception in any global ExecStates,
since global ExecStates are more like pseudo-ExecStates, and aren't
used for script execution. (It's unclear what would happen if you left
an exception waiting around in a global ExecState, but it probably
wouldn't be good.)

WebCore:

Reviewed by Maciej Stachowiak.

Adapted WebCore to the fix for http://bugs.webkit.org/show_bug.cgi?id=16909
REGRESSION: Amazon.com crash (ActivationImp)

  • bindings/js/kjs_proxy.cpp: (WebCore::KJSProxy::~KJSProxy): No convenient way to make this assertion anymore. (It wasn't firing for anyone, anyway, so it's no big loss.)
  • bindings/objc/WebScriptObject.mm: (+[WebScriptObject throwException:]): Use the ExecState stack, instead of currentExec. (-[WebScriptObject setException:]): ditto. Also, a slight change in behavior: If no ExecStates are active, we no longer throw an exception in the global ExecState. The JavaScriptCore ChangeLog explains why. This also matches the behavior of +throwException.

LayoutTests:

Layout test for http://bugs.webkit.org/show_bug.cgi?id=16909
REGRESSION: Amazon.com crash (ActivationImp)

  • fast/js/exec-state-marking-expected.txt: Added.
  • fast/js/exec-state-marking.html: Added.
4:52 PM Changeset in webkit [29709] by Nikolas Zimmermann
  • 2 edits in trunk/WebCore

Not reviewed. Try to fix Qt build, after the rmdir() fixes.

4:49 PM Changeset in webkit [29708] by Nikolas Zimmermann
  • 2 edits in trunk/WebCore

Not reviewed. Another Wx build fix.

4:30 PM Changeset in webkit [29707] by Darin Adler
  • 2 edits in trunk/WebCore
  • try to fix Qt build
  • platform/qt/FileSystemQt.cpp: (WebCore::deleteEmptyDirectory): QDir::root().rmdir() instead of QDir::rmdir().
4:05 PM Changeset in webkit [29706] by Nikolas Zimmermann
  • 3 edits in trunk/WebCore

Not reviewed. Yet another win/gtk build fix.

3:43 PM Changeset in webkit [29705] by Nikolas Zimmermann
  • 2 edits in trunk/WebCore

Not reviewed. Another build fix for Gtk/Linux & Wx/Mac.

3:26 PM Changeset in webkit [29704] by Nikolas Zimmermann
  • 4 edits in trunk/WebCore

Not reviewed. Build fix for Qt/Gtk & Wx.

3:16 PM Changeset in webkit [29703] by Nikolas Zimmermann
  • 3 edits in trunk/WebCore

Not reviewed. Build fix for ports that don't build SVG: wx.

3:12 PM Changeset in webkit [29702] by Nikolas Zimmermann
  • 2 edits in trunk/WebCore

Not reviewed. (First) Build fix for Qt/Gtk.

3:06 PM Changeset in webkit [29701] by alp@webkit.org
  • 7 edits in trunk

2008-01-21 Jan Michael Alonzo <jmalonzo@unpluggable.com>

Reviewed by Alp Toker.

http://bugs.webkit.org/show_bug.cgi?id=16955
Get errors when cross-compile webkit-gtk

2:56 PM Changeset in webkit [29700] by Nikolas Zimmermann
  • 86 edits
    6 adds
    2 deletes in trunk

Reviewed by Eric. Older parts reviewed by Dan.

Fixes: http://bugs.webkit.org/show_bug.cgi?id=16880 (SVGCSSFontFace should die, instead integrate within the FontCache.)
Fixes: http://bugs.webkit.org/show_bug.cgi?id=16784 (In-document fonts work only as the first child to font-face-src)

Proper integration of SVG Fonts within the existing custom font concept, removing the need for a custom SVGCSSFontFace.
Integrate within the CSSFontFaceSource logic for caching support, as well as the framework for external SVG Fonts.

Proper handling of missing glyphs, using <missing-glyph> element. If that is not specified in a font, fallback to non-SVG font rendering.

2:43 PM Changeset in webkit [29699] by alp@webkit.org
  • 2 edits in trunk/WebKit/gtk

2008-01-21 Alp Toker <alp@atoker.com>

GTK+ build fix for breakage introduced in r29698.

  • WebCoreSupport/ChromeClientGtk.h:
2:23 PM Changeset in webkit [29698] by Darin Adler
  • 53 edits in trunk

WebCore:

Reviewed by John Sullivan.

  • fix <rdar://problem/5644324> Delegate-less WebKit clients should have no databases
  • remove the default database quota setting
  • simplify the quota-related client calls by merging the one for a new database with the one for an existing database, adding a feature where you can get details about a database being created using the same functions that work on the other databases, and removing the parameters and return values, instead having clients use a function to set the quota
  • fix unsafe multi-thread access to the database tracker's quota map
  • fix bug in deleteAllDatabases where it would iterate a HashMap while modifying it

The tracker database is now only created when we set the quota for a database origin.
Thus asking for info about databases won't cause anything to be written to disk.

  • WebCore.base.exp: Updated.
  • bindings/js/GCController.cpp: Added an #if to get rid of an unused function warning.
  • page/Chrome.cpp: Removed database-related functions. There's no problem having the code deal directly with the client.
  • page/Chrome.h: Ditto. Also made Chrome inherit from Noncopyable.
  • page/ChromeClient.h: Replaced the two quota-related functions with a single one. The details about the state of databases in the origin are now available by asking for database details. There's also no need to pass the security origin, since it's easy to get that from the frame's document.
  • page/Settings.cpp: Removed the default quota setting.
  • page/Settings.h: Ditto.
  • platform/FileSystem.h: Added deleteEmptyDirectory.
  • platform/gtk/FileSystemGtk.cpp: (WebCore::deleteEmptyDirectory): Added.
  • platform/posix/FileSystemPOSIX.cpp: (WebCore::deleteEmptyDirectory): Added.
  • platform/qt/FileSystemQt.cpp: (WebCore::deleteEmptyDirectory): Added.
  • platform/win/FileSystemWin.cpp: (WebCore::deleteEmptyDirectory): Added.
  • platform/wx/FileSystemWx.cpp: (WebCore::deleteEmptyDirectory): Added placeholder.
  • storage/DatabaseDetails.h: Removed the isValid() function since its name is confusing -- we removed our other isValid() functions. For the few callers that need this, it's fine to just check name().isEmpty(). Made the member functions all const.
  • storage/DatabaseTracker.cpp: (WebCore::DatabaseTracker::DatabaseTracker): Removed m_defaultQuota initialization. Added m_proposedDatabase and m_thread. (WebCore::DatabaseTracker::setDatabaseDirectoryPath): Got rid of code that would open the tracker database here. This might slightly speed up launch time, since we won't do the I/O until we have to, and before we were actually creating an SQL database in this code that's run when we go to the first webpage. (WebCore::DatabaseTracker::trackerDatabasePath): Added. (WebCore::DatabaseTracker::openTrackerDatabase): Added a boolean parameter telling this function whether to create the database. Made this function safe to call if the database is already open. Also made the function fail quietly if the path has not been set yet. (WebCore::DatabaseTracker::canEstablishDatabase): Got rid of the call to the establishEntryForOrigin function. The origin is now put in the tracker database when the quota is set to a non-zero value. When judging if there's enough space for the new database, require at least one byte even if estimatedSize is 0, and check for overflow. Also added code here to populate the origins map, which guarantees it will be ready when the database uses it on another thread later. Also changed this to call the new ChromeClient function. (WebCore::DatabaseTracker::hasEntryForDatabase): Added code to open the tracker database, since that's no longer done by setDatabaseDirectoryPath. (WebCore::DatabaseTracker::originPath): Added. (WebCore::DatabaseTracker::fullPathForDatabase): Added code so that this will return a null string for the proposed database if called from within the ChromeClient function. Also switched from empty string to null string for the error cases. (WebCore::DatabaseTracker::populateOrigins): Added code to open the tracker database. (WebCore::DatabaseTracker::databaseNamesForOrigin): Ditto. (WebCore::DatabaseTracker::detailsForNameAndOrigin): Ditto. Also added code that will return the details of the proposed database if called from within the ChromeClient function. This is how the client can learn of the display name and the estimated size of the new database. (WebCore::DatabaseTracker::setDatabaseDetails): Added code to open the tracker database. (WebCore::DatabaseTracker::quotaForOrigin): Made this code OK to call on a non-main thread by using m_quotaMapGuard to guard access to the map. Other code runs on the main thread only, and only functions that write to the map use the lock. (WebCore::DatabaseTracker::setQuota): Changed this function so it can insert the initial quota as well as updating an existing quota. Added locking since this function modifies the quota map. Added code to open the tracker database. Added an early exit if the quota is already correct, which is guarantees that if you set a quota to 0 it won't trigger creation of a tracker database. (WebCore::DatabaseTracker::addDatabase): Added code to open the tracker database. (WebCore::DatabaseTracker::deleteAllDatabases): Made a copy of the quota map before iterating it to find all the origins. This fixes a problem with the old code where it would modify the map while iterating it, which gives assertions in debug builds and unpredictable results. (WebCore::DatabaseTracker::deleteOrigin): Replaced deleteDatabasesWithOrigin with this function. Added code to open the tracker database if needed. Added code to delete the origin from the tracker database, and to close the tracker database and delete files and directories as needed if we are deleting the final origin. (WebCore::DatabaseTracker::deleteDatabase): Added code to open the tracker database if needed.
  • storage/DatabaseTracker.h: Renamed databasePath to databaseDirectoryPath for clarity, including the data member, and the getter and setter functions. Replaced deleteDatabasesWithOrigin with deleteOrigin. Removed the functions dealing with default origin quota. There is no default any more; origins start with no quota and the client must set a quota. Added trackerDatabasePath and originPath helper functions. Added a boolean parameter to openTrackerDatabase to tell it whether to create the database or not. Removed the establishEntryForOrigin function. Renamed m_originQuotaMap to just m_quotaMap, and added m_quotaMapGuard. Added a QuotaMap typedef. Added m_proposedDatabase, which holds the origin and details for the current proposed database during the client callback function that must decide whether to grant quota. Added a m_thread data member for debugging use to assert if functions that can only be called on a single thread are misused.
  • storage/SQLTransaction.cpp: (WebCore::SQLTransaction::deliverQuotaIncreaseCallback): Changed to call the new exceededDatabaseQuota function instead of the old one.
  • svg/graphics/SVGImageEmptyClients.h: Updated for the change to ChromeClient.

WebKit/gtk:

Reviewed by John Sullivan.

  • updated for changes to database functions
  • WebCoreSupport/ChromeClientGtk.cpp: (WebKit::ChromeClient::exceededDatabaseQuota):
  • WebCoreSupport/ChromeClientGtk.h:
  • WebView/webkitprivate.cpp: (webkit_init):

WebKit/mac:

Reviewed by John Sullivan.

  • fix <rdar://problem/5644324> Delegate-less WebKit clients should have no databases
  • add a missing export of WebDatabaseExpectedSizeKey
  • implement deleteOrigin: and remove deleteDatabasesWithOrigin:
  • Storage/WebDatabaseManager.mm: (-[WebDatabaseManager detailsForDatabase:withOrigin:]): Updated to check for a null name instead of calling isValid(). (-[WebDatabaseManager deleteOrigin:]): Implemented. (WebKitInitializeDatabasesIfNecessary): Updated for name change.
  • Storage/WebDatabaseManagerPrivate.h: Removed deleteDatabasesWithOrigin:.
  • WebCoreSupport/WebChromeClient.h: Updated for changes to ChromeClient.
  • WebCoreSupport/WebChromeClient.mm: (WebChromeClient::exceededDatabaseQuota): Replaced the two different client functions we had before with a single one.
  • WebKit.exp: Added missing export for WebDatabaseExpectedSizeKey.
  • WebView/WebPreferenceKeysPrivate.h: Removed WebKitDefaultDatabaseQuotaKey.
  • WebView/WebPreferences.m: (+[WebPreferences initialize]): Removed the default for WebKitDefaultDatabaseQuotaKey.
  • WebView/WebPreferencesPrivate.h: Removed defaultDatabaseQuota and setDefaultDatabaseQuota:.
  • WebView/WebUIDelegatePrivate.h: Replaced the two different database quota delegate methods we had before with a single one.
  • WebView/WebView.mm: (-[WebView _preferencesChangedNotification:]): Removed the code to set the default database origin quota in WebCore::Settings based on WebPreferences.
  • WebView/WebViewInternal.h: Removed delegate method dispatch functions for unusual types of parameters that the database UI delegate methods had before.

WebKit/qt:

Reviewed by John Sullivan.

  • updated for changes to ChromeClient database functions
  • WebCoreSupport/ChromeClientQt.cpp: (WebCore::ChromeClientQt::exceededDatabaseQuota):
  • WebCoreSupport/ChromeClientQt.h:

WebKit/win:

Reviewed by John Sullivan.

  • fix <rdar://problem/5644324> Delegate-less WebKit clients should have no databases
  • fix incorrect value for WebDatabaseDisplayNameKey in header


  • Interfaces/IWebDatabaseManager.idl: Renamed detailsForDatabaseWithOrigin to detailsForDatabase. Replaced deleteDatabasesWithOrigin with deleteOrigin. Renamed deleteDatabaseWithOrigin to deleteDatabase. Fixed incorrect value for WebDatabaseDisplayNameKey.
  • Interfaces/IWebPreferencesPrivate.idl: Removed defaultDatabaseQuota and setDefaultDatabaseQuota.
  • Interfaces/IWebUIDelegatePrivate.idl: Replaced the two database quota functions with a single one.
  • WebChromeClient.cpp: (WebChromeClient::exceededDatabaseQuota): Replaced the two database quota functions with a single one.
  • WebChromeClient.h: Ditto.
  • WebDatabaseManager.cpp: (WebDatabaseManager::detailsForDatabase): Changed to check for a name of null to detect a nonexistent database rather than using isValid. (WebDatabaseManager::deleteOrigin): Updated for name change (actually a semantic change too, but both have the same name). (WebDatabaseManager::deleteDatabase): Updated for name change. (WebKitSetWebDatabasesPathIfNecessary): Ditto.
  • WebDatabaseManager.h: Updated for name changes.
  • WebPreferenceKeysPrivate.h: Removed WebKitDefaultDatabaseQuotaKey.
  • WebPreferences.cpp: (WebPreferences::initializeDefaultSettings): Removed the default for WebKitDefaultDatabaseQuotaKey.
  • WebPreferences.h: Removed defaultDatabaseQuota and setDefaultDatabaseQuota.
  • WebView.cpp: (WebView::notifyPreferencesChanged): Removed the code to set the default database origin quota in WebCore::Settings based on IWebPreferencesPrivate.

WebKit/wx:

Reviewed by John Sullivan.

  • updated for changes to ChromeClient database functions
  • WebKitSupport/ChromeClientWx.cpp: (WebCore::ChromeClientWx::exceededDatabaseQuota):
  • WebKitSupport/ChromeClientWx.h:
1:52 PM Changeset in webkit [29697] by Nikolas Zimmermann
  • 1 edit
    4 adds in trunk/LayoutTests

Rubber stamped by Oliver.
Add missing layout test results.

1:19 PM Changeset in webkit [29696] by hyatt@apple.com
  • 3 edits
    4 adds in trunk

WebCore:

Fix for http://bugs.webkit.org/show_bug.cgi?id=16935

Fix a bug in layers where the positioned ancestor was being computed incorrectly (the root should not
automatically be included).

Reviewed by olliej

Added fast/layers/positioned-inside-root-with-margins.html

  • rendering/RenderLayer.cpp: (WebCore::RenderLayer::stackingContext): (WebCore::RenderLayer::enclosingPositionedAncestor):

LayoutTests:

http://bugs.webkit.org/show_bug.cgi?id=16935

Reviewed by olliej

  • fast/layers/positioned-inside-root-with-margins.html: Added.
  • platform/mac/fast/layers/positioned-inside-root-with-margins-expected.checksum: Added.
  • platform/mac/fast/layers/positioned-inside-root-with-margins-expected.png: Added.
  • platform/mac/fast/layers/positioned-inside-root-with-margins-expected.txt: Added.
11:00 AM Changeset in webkit [29695] by adele@apple.com
  • 2 edits in trunk/LayoutTests

Adding media/{audio,video}-controls-rendering.html back to the skipped list, since they're still failing on some machines.

  • platform/win/Skipped:
8:59 AM Changeset in webkit [29694] by Darin Adler
  • 2 edits in trunk/WebKitTools

Reviewed by Mitz and Adam.

  • Scripts/svn-create-patch: Sort ChangeLog files first. Also slightly improved the sorting speed by doing all sort criteria in a single pass instead of three sorts.
3:37 AM QtWebKitContrib edited by Simon Hausmann
(diff)
3:06 AM Changeset in webkit [29693] by Simon Hausmann
  • 2 edits in trunk/WebKit/qt

Benjamin Meyer <bmeyer@trolltech.com>

Change QWebHistoryInterface::addHistoryEntry() from const to non-const

3:06 AM Changeset in webkit [29692] by Simon Hausmann
  • 2 edits in trunk/WebKit/qt

Adam Treat <treat@kde.org>

  • Don't leak the d-pointer in QWebSettings.
2:47 AM Changeset in webkit [29691] by Simon Hausmann
  • 2 edits in trunk/WebKit/qt

Simon Hausmann <Simon Hausmann>

Update the WebKit version number from WebKit/win/WebKit.vcproj/VERSION.

This has the fortunate side-effect that gmail sends us sensible HTML/JS again :)

2:41 AM Changeset in webkit [29690] by Simon Hausmann
  • 2 edits in trunk/WebCore

Fix KURL to QUrl conversion.

Stick to the encoded version of the URL and in addition to KURL we encode the
characters mentioned in section 2.4.3 of RFC 2396 as QUrl requires these, too.

This fixes fast/css/import-rule-regression-11590.html,
fast/css/import-style-update.html, svg/hixie/processing-model/003.xml and
svg/hixie/processing-model/004.xml.

2:40 AM Changeset in webkit [29689] by Simon Hausmann
  • 10 edits in trunk

Fix focus chain handling and cycling through focusable objects (links) using tab/backtab.

  • Fix GraphicsContext::drawFocusRing to also draw single focus rects.
  • Implemented QWebPage::focusNextPrevChild by sending fake tab/shift-tab events and make the return value depend on whether we successfully determined a focusable node or not.
  • Changed QWebView::focusNextPrevChild() to call the base QWidget implementation correctly if we could not handle the focus chain ourselves.
  • Changed the focus policy of QWebView to correctly use WheelFocus instead of ClickFocus.
  • Made ChromeClientQt::canTakeFocus() and takeFocus() dummy method since they are only used to control the situation of stepping out of the focus chain inside the page.
  • Made inclusion of links in the focus chain configurable through QWebSettings::LinksIncludedInFocusChain. The layout tests expect this to be disabled but for the user it seems sensible to have it on by default, hence the default in qwebsettings.cpp
2:40 AM Changeset in webkit [29688] by Simon Hausmann
  • 2 edits in trunk/WebCore

Fix access key support and fast/forms/legend-access-key.html

SVN revision 26664 changed the default access key for the non-mac build to Alt
while this test relies on Ctrl.

2:39 AM Changeset in webkit [29687] by Simon Hausmann
  • 2 edits in trunk/WebCore

Fix error reporting when parsing X(HT)ML fragments.

We use the regular XMLTokenizer write()/.../end() sequence to parse those
fragments, but we should not report any parsing errors inline inside the
content. Instead we should just return the error and the DOM layer will take
care of throwing an exception.

This fixes fast/innerHTML/innerHTML-changing-document-properties.xhtml

1:10 AM Changeset in webkit [29686] by Simon Hausmann
  • 2 edits in trunk/WebCore

Holger Freyther <holger.freyther@trolltech.com>

Fix some issues with redirections.

  • Ensure that we deliver the response to the ResourceHandleClient only once by setting m_responseSent back to false only before calling start() for the redirection
  • Added 307 as another HTTP status code that causes a redirection after POST to become get, just like in MainResourceLoader::isPostOrRedirectAfterPost
  • Also set the HTTP method on the ResourceRequest to GET before passing it to willSendRequest.
  • willSendRequest's newRequest argument is actually an in/out argument and could be modified in theory, so set m_request accordingly after the call.
1:09 AM Changeset in webkit [29685] by Simon Hausmann
  • 2 edits in trunk/WebCore

Holger Freyther <holger.freyther@trolltech.com>

Fixed WebCore::cookies() to return all cookies, not only the first one.

Fixes login into mail.yahoo.com.

1:01 AM Changeset in webkit [29684] by Antti Koivisto
  • 3 edits in trunk/LayoutTests

Try to make the test less timing dependent to fix occasional test bot failures.

  • media/video-seek-past-end-paused-expected.txt:
  • media/video-seek-past-end-paused.html:
12:36 AM Mobile edited by mrowe@apple.com
(diff)

Jan 20, 2008:

11:07 PM Changeset in webkit [29683] by alp@webkit.org
  • 2 edits in trunk/WebKit/gtk

2008-01-20 Alp Toker <alp@atoker.com>

Reviewed by Mark Rowe.

Attach the widget's style to its window and set the background to the
base background to avoid black flicker when repainting. This is
similar to what GtkTextView does.

  • WebView/webkitwebview.cpp:
9:34 PM Changeset in webkit [29682] by mrowe@apple.com
  • 4 edits in trunk/WebKit

Remove code bracketed by REMOVE_SAFARI_DOM_TREE_DEBUG_ITEM as we are no longer interested in supporting Safari 2 with TOT WebKit.

Reviewed by Dan Bernstein.

  • WebKit.xcodeproj/project.pbxproj: Don't define REMOVE_SAFARI_DOM_TREE_DEBUG_ITEM as it is no longer used.
  • mac/WebView/WebView.mm:

(+[WebView initialize]):

8:30 PM Changeset in webkit [29681] by oliver@apple.com
  • 2 edits in trunk/WebCore

Fix http://bugs.webkit.org/show_bug.cgi?id=16816 , rdar://problem/5682985

Reviewed by Mitz

Correctly trigger willPerformDragDestinationAction when a drop causes a
load to occur. The logic that originally did this was lost during the
great drag migration of '07.

8:04 PM Changeset in webkit [29680] by mitz@apple.com
  • 2 edits in trunk/LayoutTests
  • update Tiger-specific results for r29667
  • platform/mac/editing/pasteboard/paste-RTFD-expected.txt:
6:01 PM Changeset in webkit [29679] by mitz@apple.com
  • 2 edits in trunk/WebCore

Reviewed by Darin Adler.

Assertion failure in FrameView::scheduleRelayout (!m_frame->document()
!m_frame->document()->inPageCache()) when going back from a page with a focused popup
  • dom/Document.cpp: (WebCore::Document::setFocusedNode): Bail out if the document is in the page cache. Documents in the back/forward cache are "frozen" and should not change state.
3:18 PM Changeset in webkit [29678] by weinig@apple.com
  • 3 edits
    10 adds in trunk

WebCore:

Reviewed by Sam Weinig.

Fix for http://bugs.webkit.org/show_bug.cgi?id=16775

We now use frame()->loader()->url() for postMessage, preventing a
malicious sender from overwriting the uri property (using a <base> tag,
for example). Also, use frame->loader()->url().host() instead of
instead of document()->SecurityOrigin()->domain() to reflect a recent
clarification in the HTML5 spec.

Tests: http/tests/security/postMessage/domain-affected-by-document-domain.html

http/tests/security/postMessage/domain-and-uri-unaffected-by-base-tag.html
http/tests/security/postMessage/javascript-page-still-sends-domain.html

  • bindings/js/JSDOMWindowCustom.cpp: (WebCore::JSDOMWindow::postMessage):

LayoutTests:

Reviewed by Sam Weinig.

Tests for http://bugs.webkit.org/show_bug.cgi?id=16775

  • http/tests/security/postMessage: Added.
  • http/tests/security/postMessage/domain-and-uri-unaffected-by-base-tag-expected.txt: Added.
  • http/tests/security/postMessage/domain-and-uri-unaffected-by-base-tag.html: Added.
  • http/tests/security/postMessage/domain-unaffected-by-document-domain-expected.txt: Added.
  • http/tests/security/postMessage/domain-unaffected-by-document-domain.html: Added.
  • http/tests/security/postMessage/javascript-page-still-sends-domain-expected.txt: Added.
  • http/tests/security/postMessage/javascript-page-still-sends-domain.html: Added.
  • http/tests/security/postMessage/resources: Added.
  • http/tests/security/postMessage/resources/javascript-post-message-sender.html: Added.
  • http/tests/security/postMessage/resources/post-message-listener.html: Added.
1:22 PM Changeset in webkit [29677] by Antti Koivisto
  • 5 edits in trunk

WebCore:

Reviewed by Adele.

Fix <rdar://problem/5695451>
Middle part of the media timeline missing on Windows if load failed

Add some null checking to avoid getting garbage results from the functions.

  • platform/graphics/MediaPlayer.cpp:
  • platform/graphics/win/QTMovieWin.cpp: (QTMovieWin::duration): (QTMovieWin::currentTime): (QTMovieWin::setCurrentTime): (QTMovieWin::maxTimeLoaded):

LayoutTests:

Reviewed by Adele.

Update test so it checks the state of time related properties after the failure.

  • media/video-error-does-not-exist-expected.txt:
  • media/video-error-does-not-exist.html:

Jan 19, 2008:

8:53 PM Changeset in webkit [29676] by alp@webkit.org
  • 1 edit in trunk/WebCore/ChangeLog

Fix 'Reviewed by' typo in ChangeLog

8:49 PM Changeset in webkit [29675] by alp@webkit.org
  • 2 edits in trunk/WebCore

2008-01-19 Alp Toker <alp@atoker.com>

Reviewed by Alp Toker.

http://bugs.webkit.org/show_bug.cgi?id=16944
Use of GST_PLUGIN_DEFINE_STATIC results in a module-level constructor

http://bugzilla.gnome.org/show_bug.cgi?id=510547
Epiphany/Webkit fails to start due to initializing threads twice

Don't call GST_PLUGIN_DEFINE_STATIC() as it performs g_thread_init()
in the global initializer, breaking subsequent calls to
g_thread_init() in applications.

  • platform/graphics/gtk/VideoSinkGStreamer.cpp:
8:25 PM WikiStart edited by webkit@mattlilek.com
Remove spam (diff)
10:51 AM Changeset in webkit [29674] by mrowe@apple.com
  • 2 edits in trunk/WebKitTools

Allow the --http flag to run-webkit-tests to override
the default behaviour of disabling HTTP tests for Qt,
Gtk and Wx.

Reviewed by Alp Toker.

  • Scripts/run-webkit-tests:
10:34 AM Changeset in webkit [29673] by alp@webkit.org
  • 2 edits in trunk/WebCore

2008-01-19 Alp Toker <alp@atoker.com>

Reviewed by Mark Rowe.

Remove any fragment part from the URL to be requested just before
passing it to curl, otherwise curl sends it as part of the HTTP/local
file request, causing page loads to fail.

  • platform/network/curl/ResourceHandleManager.cpp: (WebCore::ResourceHandleManager::startJob):
10:27 AM Changeset in webkit [29672] by mrowe@apple.com
  • 272 edits in trunk

Set svn:eol-style=native on .idl and .c files that were missed in r29663.

10:10 AM Changeset in webkit [29671] by mrowe@apple.com
  • 9 edits in trunk/WebKit/win

2008-01-19 Mark Rowe <mrowe@apple.com>

Fix inconsistent line endings and set svn:eol-style=native
on IDL files that did not have it set.

  • Interfaces/IGEN_DOMObject.idl:
7:46 AM Changeset in webkit [29670] by alp@webkit.org
  • 4 edits in trunk

2008-01-19 Jan Michael Alonzo <jmalonzo@unpluggable.com>

Reviewed by Alp Toker.

Add svg foreign object and svg experimental to the GTK+/autotools
build system.

5:58 AM Changeset in webkit [29669] by alp@webkit.org
  • 2 edits in trunk/WebKit/gtk

2008-01-19 Christian Dywan <christian@imendio.com>

Reviewed by Alp Toker.

http://bugs.webkit.org/show_bug.cgi?id=16882
[GTK] ChromeClientGtk is incompete

Implement these functions.

  • WebCoreSupport/ChromeClientGtk.cpp: (WebKit::ChromeClient::windowRect): (WebKit::ChromeClient::pageRect): (WebKit::ChromeClient::scaleFactor): (WebKit::ChromeClient::focus): (WebKit::ChromeClient::unfocus): (WebKit::ChromeClient::canTakeFocus): (WebKit::ChromeClient::takeFocus): (WebKit::ChromeClient::canRunBeforeUnloadConfirmPanel):
2:43 AM Changeset in webkit [29668] by ddkilzer@apple.com
  • 2 edits in trunk/WebKitTools

<rdar://problem/5695344> check-for-global-initializers script never checks any object files

Reviewed by Darin.

We now touch a check-for-global-initializers.timestamp file in
the TARGET_TEMP_DIR directory to determine when new object files
have been compiled and thus need to be checked. If the timestamp
file doesn't exist, all object files will be checked.

Previously the modification time of the "executable" (the
framework binary, e.g., WebKit.framework/WebKit) was used, but
since this was the last file modified at the end of the compile
phase, no object files would ever get checked!

Also added JSCustomSQLTransactionCallback.o to the list of files
since it has static initializers in Debug builds of WebCore.

  • Scripts/check-for-global-initializers:
2:12 AM Changeset in webkit [29667] by mitz@apple.com
  • 21 edits
    5 adds in trunk

WebCore:

Reviewed by Maciej Stachowiak.

  • fix <rdar://problem/5645813> CrashTracer: [USER] 6 crashes in Safari at com.apple.WebCore: WebCore::RenderBox::destroy + 116

Test: editing/selection/inconsistent-in-removeChildNode.html

  • editing/SelectionController.cpp: (WebCore::SelectionController::nodeWillBeRemoved): If the selection base or extent are not visible any more, adjust the selection.

LayoutTests:

Reviewed by Maciej Stachowiak.

  • test and updated results for <rdar://problem/5645813> CrashTracer: [USER] 6 crashes in Safari at com.apple.WebCore: WebCore::RenderBox::destroy + 116
  • editing/selection/inconsistent-in-removeChildNode.html: Added.
  • platform/mac-leopard/editing/pasteboard/paste-RTFD-expected.txt:
  • platform/mac-leopard/editing/selection: Added.
  • platform/mac-leopard/editing/selection/inconsistent-in-removeChildNode-expected.checksum: Added.
  • platform/mac-leopard/editing/selection/inconsistent-in-removeChildNode-expected.png: Added.
  • platform/mac/editing/deleting/collapse-whitespace-3587601-fix-expected.txt:
  • platform/mac/editing/deleting/delete-3608462-fix-expected.txt:
  • platform/mac/editing/deleting/delete-4083333-fix-expected.txt:
  • platform/mac/editing/deleting/delete-leading-ws-001-expected.txt:
  • platform/mac/editing/deleting/delete-line-011-expected.txt:
  • platform/mac/editing/execCommand/find-after-replace-expected.txt:
  • platform/mac/editing/execCommand/paste-1-expected.txt:
  • platform/mac/editing/execCommand/paste-2-expected.txt:
  • platform/mac/editing/pasteboard/emacs-ctrl-k-y-001-expected.txt:
  • platform/mac/editing/selection/inconsistent-in-removeChildNode-expected.txt: Added.
  • platform/mac/editing/selection/move-between-blocks-no-001-expected.txt:
  • platform/mac/editing/selection/replace-selection-1-expected.txt:
  • platform/mac/editing/style/remove-underline-across-paragraph-expected.txt:
  • platform/mac/editing/style/remove-underline-across-paragraph-in-bold-expected.txt:
  • platform/mac/editing/style/remove-underline-after-paragraph-in-bold-expected.txt:
  • platform/mac/editing/style/remove-underline-expected.txt:
  • platform/mac/editing/style/remove-underline-in-bold-expected.txt:
  • platform/mac/editing/style/unbold-in-bold-expected.txt:
1:43 AM Changeset in webkit [29666] by oliver@apple.com
  • 2 edits in trunk/WebCore

Build fix for --svg-foreign-object path

Reviewed by Eric Seidel

Note: See TracTimeline for information about the timeline view.