Timeline


and

03/24/08:

23:57 squirrelfish edited by oliver@apple.com
(diff)
23:56 Changeset [31278] by eric@webkit.org

Reviewed by Mark.

Clean up SVG features message to be less confusing.

  • Scripts/build-webkit:
23:21 squirrelfish edited by cwzwarich@uwaterloo.ca
(diff)
23:20 squirrelfish edited by cwzwarich@uwaterloo.ca
(diff)
22:48 Changeset [31277] by ggaren@apple.com

2008-03-24 Geoffrey Garen <ggaren@apple.com>

Reviewed by Oliver Hunt.


Fixed recent 25% regression on simple for loop test. GCC seems to be
very finicky about the code that gets inlined into
Machine::privateExecute.

Everything in this patch is simply the result of experiment.


The resolve and resolve_base opcodes do not seem to have gotten slower
from this change.

  • VM/Machine.cpp: (KJS::resolve): (KJS::resolveBase): (KJS::Machine::privateExecute):
  • kjs/nodes.h:
22:05 Changeset [31276] by oliver@apple.com

Bug 18059: squirrelfish needs to compile on platforms without computed goto
<http://bugs.webkit.org/show_bug.cgi?id=18059>

Reviewed by Geoff Garen

"Standard" macro style support for conditionalising the use of computed goto.

22:02 Changeset [31275] by darin@apple.com

2008-03-24 Darin Adler <darin@apple.com>

  • try to fix broken Wx build
  • platform/wx/TemporaryLinkStubs.cpp: (GraphicsContext::setPlatformShadow): (GraphicsContext::clearPlatformShadow):
21:35 Changeset [31274] by oliver@apple.com

Bug 18030: REGRESSION(r31236): Space bar fails to scroll down page
<http://bugs.webkit.org/show_bug.cgi?id=18030>

Reviewed by Mark Rowe

Rollout keyDown changes from r31236 -- fix for keyDown behaviour is
tracked by Bug 18057: keyDown incorrectly propagates up the frame tree
<http://bugs.webkit.org/show_bug.cgi?id=18057>

21:30 Changeset [31273] by darin@apple.com

2008-03-24 Darin Adler <darin@apple.com>

Reviewed by Sam.

  • use AtomicString more consistently for attribute values

1% speedup on Acid3 test 26

  • dom/Element.cpp: (WebCore::Element::setAttribute): Changed value argument to AtomicString. Also fixed call to lower() to only be done if the name isn't already lowercase. Also took out one unneeded condition in a cascade of if statements. (WebCore::Element::createAttribute): Ditto. (WebCore::Element::setAttributeNS): Ditto.
  • dom/Element.h: Ditto.
  • dom/StyledElement.cpp: (WebCore::StyledElement::createAttribute): Ditto.
  • dom/StyledElement.h: Ditto.
21:28 Changeset [31272] by darin@apple.com

2008-03-24 Darin Adler <darin@apple.com>

Reviewed by Sam.

  • get rid of extra level of function call in toJS functions for Node

0.6% speedup on Acid3 test 26

  • bindings/js/JSNodeCustom.cpp: (WebCore::toJS): Take a raw pointer, not a PassRefPtr, like other toJS functions.
  • bindings/scripts/CodeGeneratorJS.pm: Add an inlined toJS for EventTargetNode* to avoid ambiguity that otherwise exists between the toJS for EventTarget* and for Node*.
20:43 Changeset [31271] by weinig@apple.com

2008-03-24 Sam Weinig <sam@webkit.org>

Reviewed by Dave Hyatt.

Add EventSender.zoomPageIn/zoomPageOut support to DRT.

  • DumpRenderTree/mac/DumpRenderTree.mm: (resetWebViewToConsistentStateBeforeTesting):
  • DumpRenderTree/mac/EventSendingController.mm: (+[EventSendingController isSelectorExcludedFromWebScript:]): (-[EventSendingController zoomPageIn]): (-[EventSendingController zoomPageOut]):
  • DumpRenderTree/win/DumpRenderTree.cpp: (resetWebViewToConsistentStateBeforeTesting):
  • DumpRenderTree/win/EventSender.cpp: (textZoomInCallback): (textZoomOutCallback): (zoomPageInCallback): (zoomPageOutCallback):
20:39 Changeset [31270] by darin@apple.com

2008-03-24 Darin Adler <darin@apple.com>

Reviewed by Sam.

  • convert a JavaScript immediate number to a string more efficiently

2% speedup of Acid3 test 26

  • kjs/JSImmediate.cpp: (KJS::JSImmediate::toString): Take advantage of the fact that all immediate numbers are integers, and use the faster UString function for formatting integers instead of the slower one that works for floating point. I think this is a leftover from when immediate numbers were floating point.
20:26 Changeset [31269] by darin@apple.com

2008-03-24 Darin Adler <darin@apple.com>

Reviewed by Hyatt.

  • add fast path to isValidName that does not use advanced Unicode functions

2% speedup on Acid3 test 26

  • dom/Document.cpp: (WebCore::isValidNameNonASCII): Added. Contains the old isValidName function. (WebCore::isValidNameASCII): Added. Contains a simpler function that returns true for most valid names that are also all ASCII. (WebCore::Document::isValidName): Changed to call both of the above functions. The common case is that isValidNameASCII returns true so isValidNameNonASCII is never called.
19:39 Changeset [31268] by hyatt@apple.com

2008-03-24 David Hyatt <hyatt@apple.com>

Fix for bug 18052, window properties need to respect zoom factor.

Reviewed by Beth

  • page/DOMWindow.cpp: (WebCore::DOMWindow::innerHeight): (WebCore::DOMWindow::innerWidth): (WebCore::DOMWindow::scrollX): (WebCore::DOMWindow::scrollY): (WebCore::DOMWindow::scrollTo):
  • page/FrameView.cpp: (WebCore::FrameViewPrivate::reset): (WebCore::FrameView::layout): (WebCore::FrameView::performPostLayoutTasks):
19:26 Changeset [31267] by weinig@apple.com

JavaScriptCore:

2008-03-23 Sam Weinig <sam@webkit.org>

Reviewed by Darin Adler.

Fix http://bugs.webkit.org/show_bug.cgi?id=18048
The "thisObject" parameter to JSEvaluateScript is not used properly

Making passing a thisObject to JSEvaluateScript actually set the thisObject of the created
ExecState.

  • API/testapi.c: (main): Add tests for setting the thisObject when calling JSEvaluateScript.
  • kjs/ExecState.cpp: (KJS::ExecState::ExecState): Assign the thisObject to m_thisValue and remove the comment.

WebCore:

2008-03-24 Sam Weinig <sam@webkit.org>

Reviewed by Darin Adler.

Fix http://bugs.webkit.org/show_bug.cgi?id=18048
The "thisObject" parameter to JSEvaluateScript is not used properly

  • bindings/js/kjs_proxy.cpp: (WebCore::KJSProxy::evaluate): No need to pass a thisObject since we want the global object to be used.
  • bridge/jni/jni_jsobject.mm: (JavaJSObject::eval): To avoid any change to this function, don't pass a thisObject to keep the same behavior.
18:19 Changeset [31266] by beidson@apple.com

2008-03-24 Brady Eidson <beidson@apple.com>

Made new layout test files executable, as needed for the tests to not fail!

  • http/tests/navigation/resources/reload-subframe-content.pl: Added.
  • http/tests/navigation/resources/reload-subframe.pl: Added.
18:10 squirrelfish edited by cwzwarich@uwaterloo.ca
(diff)
17:49 Changeset [31265] by ggaren@apple.com

2008-03-24 Geoffrey Garen <ggaren@apple.com>

Moved my notes from nodes.h to the wiki.

  • kjs/nodes.h:
17:48 squirrelfish edited by ggaren@apple.com
(diff)
17:44 squirrelfish edited by oliver@apple.com
(diff)
17:44 WikiStart edited by ggaren@apple.com
(diff)
17:43 squirrelfish edited by ggaren@apple.com
(diff)
17:41 squirrelfish created by ggaren@apple.com
initial rough cut of tasks and ideas
17:37 Changeset [31264] by beidson@apple.com

WebKit/mac:

2008-03-24 Cameron Zwarich <cwzwarich@uwaterloo.ca>

Reviewed by Maciej, landed by Brady

Bug 3580: iFrames Appear to be Cached
<http://bugs.webkit.org/show_bug.cgi?id=3580>

Bug 15486: REGRESSION: Reload causes WebKit to *forget* fragment URLs
<http://bugs.webkit.org/show_bug.cgi?id=15486>

Bug 15554: Reload causes <object> to use old data
<http://bugs.webkit.org/show_bug.cgi?id=15554>

If a page is reloaded, a child frame's URL can not be taken from a history item.

  • WebView/WebFrame.mm: (-[WebFrame _loadURL:referrer:intoChild:]):

LayoutTests:

2008-03-24 Cameron Zwarich <cwzwarich@uwaterloo.ca>

Reviewed by Maciej, landed by Brady

Tests for the following bugs:

Bug 3580: iFrames Appear to be Cached
<http://bugs.webkit.org/show_bug.cgi?id=3580>

Bug 15486: REGRESSION: Reload causes WebKit to *forget* fragment URLs
<http://bugs.webkit.org/show_bug.cgi?id=15486>

Bug 15554: Reload causes <object> to use old data
<http://bugs.webkit.org/show_bug.cgi?id=15554>

  • http/tests/navigation/reload-subframe-frame-expected.txt: Added.
  • http/tests/navigation/reload-subframe-frame.html: Added.
  • http/tests/navigation/reload-subframe-iframe-expected.txt: Added.
  • http/tests/navigation/reload-subframe-iframe.html: Added.
  • http/tests/navigation/reload-subframe-object-expected.txt: Added.
  • http/tests/navigation/reload-subframe-object.html: Added.
  • http/tests/navigation/resources/reload-subframe-content.pl: Added.
  • http/tests/navigation/resources/reload-subframe.pl: Added.
17:16 Changeset [31263] by ggaren@apple.com

2008-03-24 Geoffrey Garen <ggaren@apple.com>

Reviewed by NOBODY.


SquirrelFish lives.


Initial check-in of the code I've been carrying around. Lots of stuff
doesn't work. Plus a bunch of empty files.

17:10 Changeset [31262] by mrowe@apple.com

New branch.

17:09 Changeset [31261] by mrowe@apple.com
16:19 Changeset [31260] by darin@apple.com

2008-03-24 Darin Adler <darin@apple.com>

Reviewed by Beth.

  • fix <rdar://problem/5817067> -[WebDataSource unreachableURL] invokes KURL's copy constructor
  • History/WebHistoryItem.mm: (-[WebHistoryItem URL]): Use a reference to avoid making a copy.
  • WebView/WebDataSource.mm: (-[WebDataSource _URL]): Ditto. (-[WebDataSource unreachableURL]): Ditto.
  • WebView/WebHTMLView.mm: (-[WebHTMLView namesOfPromisedFilesDroppedAtDestination:]): Ditto.
  • DefaultDelegates/WebDefaultContextMenuDelegate.mm:
  • History/WebHistory.mm:
  • Misc/WebElementDictionary.mm:
  • Misc/WebNSAttributedStringExtras.mm: Remove unneeded imports of KURL.h.
16:15 Changeset [31259] by ggaren@apple.com

No, I am not kidding.

15:59 Changeset [31258] by beidson@apple.com

WebCore:

2008-03-24 Brady Eidson <beidson@apple.com>

Reviewed by Darin's rubberstamp

Land a load of empty files for upcoming work to make sure I'm not breaking any platform's build

  • GNUmakefile.am:
  • WebCore.pro:
  • WebCore.vcproj/WebCore.vcproj:
  • WebCore.xcodeproj/project.pbxproj:
  • WebCoreSources.bkl:
  • loader/archive: Added.
  • loader/archive/Archive.h: Added.
  • loader/archive/ArchiveFactory.cpp: Added.
  • loader/archive/ArchiveFactory.h: Added.
  • loader/archive/ArchiveResource.cpp: Added.
  • loader/archive/ArchiveResource.h: Added.
  • loader/archive/ArchiveResourceCollection.cpp: Added.
  • loader/archive/ArchiveResourceCollection.h: Added.
  • loader/archive/cf: Added.
  • loader/archive/cf/LegacyWebArchive.cpp: Added.
  • loader/archive/cf/LegacyWebArchive.h: Added.
  • loader/archive/cf/LegacyWebArchiveMac.mm: Added.

WebKit:

2008-03-24 Brady Eidson <beidson@apple.com>

Reviewed by Darin's rubberstamp

Rename a .m to .mm

  • WebKit.xcodeproj/project.pbxproj:

WebKit/mac:

2008-03-24 Brady Eidson <beidson@apple.com>

Reviewed by Darin's rubberstamp

Rename this file for upcoming work.

  • WebView/WebArchive.m: Removed.
  • WebView/WebArchive.mm: Copied from WebKit/mac/WebView/WebArchive.m.
15:51 Changeset [31257] by hyatt@apple.com

2008-03-24 David Hyatt <hyatt@apple.com>

Fix for bug 18009. Make sure the IE offset/client/scroll* extensions respect full page zoom.

Reviewed by Beth

  • dom/Element.cpp: (WebCore::localZoomForRenderer): (WebCore::adjustForLocalZoom): (WebCore::adjustForAbsoluteZoom): (WebCore::Element::offsetLeft): (WebCore::Element::offsetTop): (WebCore::Element::offsetWidth): (WebCore::Element::offsetHeight): (WebCore::Element::clientLeft): (WebCore::Element::clientTop): (WebCore::Element::clientWidth): (WebCore::Element::clientHeight): (WebCore::Element::scrollLeft): (WebCore::Element::scrollTop): (WebCore::Element::setScrollLeft): (WebCore::Element::setScrollTop): (WebCore::Element::scrollWidth): (WebCore::Element::scrollHeight):
  • rendering/RenderBox.cpp: (WebCore::RenderBox::setStyle):
  • rendering/RenderObject.cpp: (WebCore::RenderObject::offsetParent):
  • rendering/RenderObject.h:
14:59 Changeset [31256] by mrowe@apple.com

Versioning.

14:57 Changeset [31255] by mrowe@apple.com

New tag.

14:53 Changeset [31254] by mitz@apple.com

Reviewed by Mark Rowe.

  • update bisect-builds for Safari 3.1
  • Scripts/bisect-builds: Added Safari 3.1 and the corresponding minimal revision, r29711.
14:51 Changeset [31253] by mrowe@apple.com

Merge r31236.

14:51 Changeset [31252] by mrowe@apple.com

Merge r31144.

14:19 Google Summer of Code 2008 edited by eric@webkit.org
(diff)
14:17 Google Summer of Code 2008 edited by eric@webkit.org
(diff)
13:43 Changeset [31251] by beidson@apple.com

2008-03-24 Brady Eidson <beidson@apple.com>

Reviewed by Darin

Some style cleanup and add a logging channel for upcoming work

  • platform/Logging.cpp:
  • platform/Logging.h:
  • platform/MIMETypeRegistry.cpp: (WebCore::initializeSupportedImageMIMETypes): (WebCore::initializeSupportedNonImageMimeTypes):
13:31 Changeset [31250] by mitz@apple.com

WebCore:

Reviewed by Sam Weinig.

  • fix <rdar://problem/5812292> CrashTracer: [REGRESSION] 60 crashes in Safari at com.apple.WebCore: WebCore::RenderObject::repaintRectangle + 47

Test: fast/repaint/renderer-destruction-by-invalidateSelection-crash.html

  • page/FrameView.cpp: (WebCore::FrameViewPrivate::reset): Removed repaintRects code. (WebCore::FrameView::layout): Ditto. (WebCore::FrameView::addRepaintInfo): Removed.
  • page/FrameView.h: Removed addRepaintInfo().
  • rendering/RenderBlock.cpp: (WebCore::RenderBlock::layoutBlock): Changed to call repaintRectangle instead of FrameView::addRepaintInfo().
  • rendering/RenderTable.cpp: (WebCore::RenderTable::layout): Ditto.

LayoutTests:

Reviewed by Sam Weinig.

  • test for <rdar://problem/5812292> CrashTracer: [REGRESSION] 60 crashes in Safari at com.apple.WebCore: WebCore::RenderObject::repaintRectangle + 47
  • fast/repaint/renderer-destruction-by-invalidateSelection-crash.html: Added.
  • platform/mac/fast/repaint/renderer-destruction-by-invalidateSelection-crash-expected.checksum: Added.
  • platform/mac/fast/repaint/renderer-destruction-by-invalidateSelection-crash-expected.png: Added.
  • platform/mac/fast/repaint/renderer-destruction-by-invalidateSelection-crash-expected.txt: Added.
12:26 Changeset [31249] by alp@webkit.org

2008-03-24 Alp Toker <alp@atoker.com>

GTK+/soup http backend build fix for breakage introduced in r31141.
Fix suggested by Xan.

  • platform/network/ResourceHandle.h:
12:15 Changeset [31248] by ggaren@apple.com

Imported r31247 of JavaScriptCore from trunk.

12:07 Changeset [31247] by ggaren@apple.com

Created registercode branch.

10:55 Changeset [31246] by mitz@apple.com

Reviewed by Sam Weinig.

  • fix <rdar://problem/5792582> CSS opacity does not work with GDI text
  • fix <rdar://problem/5792619> non-opaque CSS colors do not work with GDI text (RGBA, HSLA)
  • fix <rdar://problem/5792624> GDI text can't be stroked/filled independently
  • make text-shadow work with GDI text
  • make -webkit-background-clip: text work with GDI text
  • platform/graphics/GraphicsContext.cpp: (WebCore::GraphicsContext::setShadow): Made this a shared method that updates the state and calls the platform-specific method. (WebCore::GraphicsContext::clearShadow): Ditto. (WebCore::GraphicsContext::getShadow): Added.
  • platform/graphics/GraphicsContext.h: (WebCore::GraphicsContext::WindowsBitmap): Added this class that represents a bitmap with a Windows device context that GDI can draw into.
  • platform/graphics/GraphicsContextPrivate.h: (WebCore::GraphicsContextState::GraphicsContextState): Added the shadow parameters to the graphics state.
  • platform/graphics/cairo/GraphicsContextCairo.cpp: (WebCore::GraphicsContext::setPlatformShadow): Renamed the platform- specific implementation to this. (WebCore::GraphicsContext::clearPlatformShadow): Renamed the platform- specific implementation to this.
  • platform/graphics/cg/GraphicsContextCG.cpp: (WebCore::GraphicsContext::setPlatformShadow): Renamed the platform- specific implementation to this. (WebCore::GraphicsContext::clearPlatformShadow): Renamed the platform- specific implementation to this.
  • platform/graphics/qt/GraphicsContextQt.cpp: (WebCore::GraphicsContext::setPlatformShadow): Renamed the platform- specific implementation to this. (WebCore::GraphicsContext::clearPlatformShadow): Renamed the platform- specific implementation to this.
  • platform/graphics/win/FontCGWin.cpp: (WebCore::toCGFloat): Added. Converts a Windows FIXED. (WebCore::createPathForGlyph): Added. Retrieves a glyph's hinted outline from GDI and creates a CGPath containing it. (WebCore::Font::drawGlyphs): Added two code paths in the GDI case. The existing code path, which uses GDI to draw text directly to the graphics context, is used for opaque-colored, non-transformed, non-stroked, non-shadowed text in a GDI-backed graphics context that is not in a transparency layer. An additional code path is used for non-stroked text in all other cases. It uses GDI to draw black text into an all-white Windows bitmap, then uses the green channel as a mask to create a bitmap with the desired fill color and alpha. The bitmap is then drawn into the graphics context, at which time transparency layers, transforms and shadows are handled properly. A third code path is used for stroked text. It constructs a path from hinted glyph outlines retrieved from GDI and strokes (and optionally fills) it.
  • platform/graphics/win/GraphicsContextCGWin.cpp: (WebCore::GraphicsContext::getWindowsContext): Added a mayCreateBitmap parameter. When false, prevents this method from creating a Windows device context if the graphics context does not already have one. (WebCore::GraphicsContext::releaseWindowsContext): Added a mayCreateBitmap parameter to match getWindowsContext(). (WebCore::GraphicsContext::WindowsBitmap::WindowsBitmap): (WebCore::GraphicsContext::WindowsBitmap::~WindowsBitmap): (WebCore::GraphicsContext::createWindowsBitmap): Added. (WebCore::GraphicsContext::drawWindowsBitmap): Added. Draws the bitmap as a non-alpha-premultiplied image.
10:45 Changeset [31245] by kmccullough@apple.com

2008-03-24 Kevin McCullough <kmccullough@apple.com>

  • Fixed Changelog
10:16 Changeset [31244] by alp@webkit.org

2008-03-24 Marco Barisione <marco.barisione@collabora.co.uk>

Reviewed by Alp Toker.

http://bugs.webkit.org/show_bug.cgi?id=17845
[GTK] combo box menu displayed in the wrong position

  • platform/gtk/PopupMenuGtk.cpp: (WebCore::PopupMenu::show): Change the popup menu vertical position so that the active item is over the combo box. (WebCore::PopupMenu::menuPositionFunction): use the "push in" style like native GTK+ combo boxes to handle menus that don't fit in the screen.
05:00 Changeset [31243] by aroben@apple.com
  • WebCore.vcproj/WebCore.vcproj: Add files from WebCore/page/inspector to the project.
01:15 Changeset [31242] by ap@webkit.org

Build fix.

  • WebCore/WebCore.xcodeproj/project.pbxproj: Added DOMSVGAltGlyphElement.h and DOMSVGAltGlyphElementInternal.h to Copy Generated Headers phase.
  • WebKit/MigrateHeaders.make: Added DOMSVGAltGlyphElement.h and DOMSVGAltGlyphElementInternal.h.
00:24 Changeset [31241] by eric@webkit.org

Build fix, no review.

Fix project paths to be group relative instead of project relative

  • WebCore.xcodeproj/project.pbxproj:

03/23/08:

22:36 Changeset [31240] by mjs@apple.com

WebCore:

2008-03-23 Maciej Stachowiak <mjs@apple.com>

Reviewed by Eric.

  • svg/SVGAltGlyphElement.cpp: Added. Implementation for new element. (WebCore::SVGAltGlyphElement::SVGAltGlyphElement): (WebCore::SVGAltGlyphElement::~SVGAltGlyphElement): (WebCore::SVGAltGlyphElement::setGlyphRef): (WebCore::SVGAltGlyphElement::glyphRef): (WebCore::SVGAltGlyphElement::setFormat): (WebCore::SVGAltGlyphElement::format): (WebCore::SVGAltGlyphElement::childShouldCreateRenderer): (WebCore::SVGAltGlyphElement::createRenderer): For now, this just renders as a tspan, which is right for everything besides the alternate glyph substitution itself.
  • svg/SVGAltGlyphElement.h: Added. Header for new element. (WebCore::SVGAltGlyphElement::contextElement):
  • svg/SVGAltGlyphElement.idl: Added. Interface for new element.
  • svg/SVGTSpanElement.cpp: (WebCore::SVGTSpanElement::childShouldCreateRenderer): Allow altGlyph to render here.
  • svg/SVGTextElement.cpp: (WebCore::SVGTextElement::childShouldCreateRenderer): ditto
  • svg/SVGTextPathElement.cpp: (WebCore::SVGTextPathElement::childShouldCreateRenderer): ditto
  • svg/svgtags.in: Add altGlyph
  • DerivedSources.make: Add new files
  • GNUmakefile.am: ditto
  • WebCore.SVG.exp: Add new ObjC DOM class
  • WebCore.pro: Add new files
  • WebCore.vcproj/WebCore.vcproj: ditto
  • WebCore.xcodeproj/project.pbxproj: ditto
  • bindings/js/JSSVGElementWrapperFactory.cpp: Handle altGlyph
  • bindings/objc/DOM.mm: (WebCore::createElementClassMap): ditto
  • bindings/objc/DOMInternal.h: Include altGlyph header
  • bindings/objc/DOMSVG.h: Include altGlyph header

LayoutTests:

2008-03-23 Maciej Stachowiak <mjs@apple.com>

Reviewed by Eric.

  • platform/mac/svg/W3C-SVG-1.1/text-altglyph-01-b-expected.txt:
  • platform/mac/svg/text/text-altglyph-01-b-expected.txt:
  • svg/dom/altGlyph-dom-expected.txt: Added.
  • svg/dom/altGlyph-dom.xhtml: Added.
  • svg/dom/resources/altGlyph-dom.js: Added.
18:21 Changeset [31239] by antti@apple.com

2008-03-23 Antti Koivisto <antti@apple.com>

  • blog-files/wsj-vs.png: Added.
14:08 Changeset [31238] by eric@webkit.org

Reviewed by eseidel. Landed by eseidel.

Bug 17685: [~=] attribute selector failing to match empty string
<http://bugs.webkit.org/show_bug.cgi?id=17685>

Test: fast/css/attribute-selector-empty-value.html

  • css/CSSStyleSelector.cpp: (WebCore::CSSStyleSelector::checkOneSelector):
14:03 Changeset [31237] by eric@webkit.org

Reviewed by Sam.

Make our SVG status document slightly less inaccurate.

  • projects/svg/status.xml:
03:48 Changeset [31236] by oliver@apple.com

Bug 17670: Key events may improperly propagate from iframe to parent frame
Bug 16381: REGRESSION: Shift, command, option, ctrl keys in Gmail Rich Text changes focus

Reviewed by Maciej

Prevent the Cocoa event system from propagating key events to the parent WebHTMLView,
as that results in us dispatching the key events for each frame going up the frame
tree.

03/22/08:

15:21 Changeset [31235] by mrowe@apple.com

2008-03-22 Marco Barisione <marco.barisione@collabora.co.uk>

Reviewed by Darin Adler.

http://bugs.webkit.org/show_bug.cgi?id=16339
Frame scrollbar positions are not updated on page scroll

Widget::geometryChanged() was called instead of
ScrollViewScrollbar::geometryChanged() because of a missing const.

  • platform/gtk/ScrollViewGtk.cpp: (WebCore::ScrollViewScrollbar::geometryChanged): add the missing const.
15:01 Changeset [31234] by mrowe@apple.com

2008-03-22 Rodney Dawes <dobey@wayofthemonkey.com>

Reviewed by Mark Rowe.

Bug 17974: REGRESSION: npfunctions.h should not import npruntime_internal.h
<http://bugs.webkit.org/show_bug.cgi?id=17974>
<rdar://problem/5782310>

Include npapi.h and npruntime.h directly in npfunctions.h.
Add npfunctions.h and npapi.h to npruntime_internal.h.
Replace includes of npfunctions.h with npruntime_internal.h.

  • bridge/npruntime_internal.h:
  • plugins/PluginPackage.h:
  • plugins/PluginStream.h:
  • plugins/npapi.cpp:
  • plugins/npfunctions.h:
14:04 Changeset [31233] by mrowe@apple.com

Fix the Qt build.

03:31 Changeset [31232] by eric@webkit.org

Update a (passing) result I missed in my last checkin. No review.

  • dom/xhtml/level3/core/documentgetdoctype01-expected.txt:
02:50 Changeset [31231] by eric@webkit.org

Reviewed by mjs.

Unify handling of NAMESPACE_ERR and fix Acid3 test 25
http://bugs.webkit.org/show_bug.cgi?id=16693

Test: fast/dom/DOMImplementation/createDocumentType-err.html

  • dom/DOMImplementation.cpp: (WebCore::DOMImplementation::createDocumentType): (WebCore::DOMImplementation::createDocument):
  • dom/DOMImplementation.idl:
  • dom/Document.cpp: (WebCore::Document::hasPrefixNamespaceMismatch): (WebCore::Document::createElementNS): (WebCore::Document::parseQualifiedName): (WebCore::Document::createAttributeNS):
  • dom/Document.h:
  • dom/Element.cpp: (WebCore::Element::setAttributeNS):
  • editing/FormatBlockCommand.cpp: (WebCore::FormatBlockCommand::doApply):
02:49 Changeset [31230] by eric@webkit.org

Reviewed by mjs.

Fix createElementNS to throw exceptions for invalid qualified names
Fixes Acid3 sub-test 23
http://bugs.webkit.org/show_bug.cgi?id=16833

Tests: fast/dom/Document/createAttributeNS-namespace-err.html

fast/dom/Document/createElementNS-namespace-err.html

  • dom/Document.cpp: (WebCore::Document::createElement): (WebCore::hasNamespaceError): (WebCore::Document::createElementNS): (WebCore::Document::createAttributeNS):
  • dom/Document.idl:
02:14 Changeset [31229] by eric@webkit.org

Reviewed by Sam Weinig. Landed by eseidel.


Make testkjs flush stdout after printing.

  • JavaScriptCore.xcodeproj/project.pbxproj:
  • kjs/testkjs.cpp: (functionPrint):
00:14 Changeset [31228] by oliver@apple.com

2008-03-21 Cameron Zwarich <cwzwarich@uwaterloo.ca>

Bug 13693: REGRESSION (r13615): Acid2 Test Eyes render improperly after a page refresh
<http://bugs.webkit.org/show_bug.cgi?id=13693>

Reviewed by Oliver Hunt

Do not create bogus history items for unloaded <object> elements, because they
cause fallback content to be ignored on reload.

03/21/08:

23:57 Changeset [31227] by mitz@apple.com

WebCore:

Reviewed by Oliver Hunt.

Test: fast/css-generated-content/empty-content-with-float-crash.html

  • rendering/bidi.cpp: (WebCore::RenderBlock::layoutInlineChildren): Added a null check for the case where the block contains no in-flow elements but still lays out as a line flow, which can happen as a result of empty generated content.

LayoutTests:

Reviewed by Oliver Hunt.

  • fast/css-generated-content/empty-content-with-float-crash-expected.txt: Added.
  • fast/css-generated-content/empty-content-with-float-crash.html: Added.
21:59 Changeset [31226] by oliver@apple.com

Optimise lookup of Math, undefined, NaN and Infinity

Reviewed by Maciej

Added a method to JSVariableObject to allow us to inject DontDelete properties
into the symbol table and localStorage. This results in a 0.4% progression in
SunSpider, with a 8% gain in math-partial-sums.

19:36 Changeset [31225] by oliver@apple.com

Global properties that use LocalStorage are not correctly listed as enumerable.

Reviewed by Geoff Garen

The problem was caused by JSObject::getPropertyAttributes not being aware
of the JSVariableObject SymbolTable. The fix is to make getPropertyAttributes
virtual and override in JSVariableObject. This does not produce any performance
regression.

17:39 Changeset [31224] by weinig@apple.com

WebCore:

2008-03-21 Sam Weinig <sam@webkit.org>

Reviewed by Oliver Hunt.

Fix <rdar://problem/5788451>
toDataURL not implemented for Windows (need mapping of MIME type to UTI)

Hard code support for just PNG's on windows, the minimum the spec requires.

  • platform/MIMETypeRegistry.cpp: (WebCore::initializeSupportedImageMIMETypesForEncoding):
  • platform/graphics/cg/ImageBufferCG.cpp: (WebCore::utiFromMIMEType): (WebCore::ImageBuffer::toDataURL):

LayoutTests:

2008-03-21 Sam Weinig <sam@webkit.org>

Reviewed by Oliver Hunt.

Enable tests for <rdar://problem/5788451>
toDataURL not implemented for Windows (need mapping of MIME type to UTI)

  • platform/win/Skipped:
  • platform/win/fast/canvas: Added.
  • platform/win/fast/canvas/toDataURL-supportedTypes-expected.txt: Added.
16:58 Changeset [31223] by pewtermoose@webkit.org

2008-03-21 Matt Lilek <webkit@mattlilek.com>

Not reviewed, Mac build fix.

  • bindings/objc/WebScriptObject.mm: (+[WebScriptObject _convertValueToObjcValue:originRootObject:rootObject:]):
  • bridge/jni/jni_utility.cpp: (KJS::Bindings::convertValueToJValue):
  • bridge/objc/objc_runtime.mm: (ObjcFallbackObjectImp::callAsFunction):
16:31 Changeset [31222] by antti@apple.com

2008-03-21 Antti Koivisto <antti@apple.com>

Add some images for a blog post.

  • blog-files/wsj-nopreload.png: Added.
  • blog-files/wsj-preload.png: Added.
16:30 Changeset [31221] by aroben@apple.com

Attempt to fix the Mac build

  • bindings/objc/DOMUtility.mm: (KJS::createDOMWrapper):
15:09 Changeset [31220] by aroben@apple.com

Rename static info members to s_info

This avoids conflicts with custom "info" methods, such as the one that
will soon be added to Console.

Reviewed by Ada Chan and Sam Weinig.

15:00 Changeset [31219] by hyatt@apple.com

2008-03-21 David Hyatt <hyatt@apple.com>

<rdar://problem/5808863> Zooming should not let non-zero border width become zero

Make sure zoomed values that begin >= 1 are not allowed to fall below 1.

Reviewed by adele

  • css/CSSPrimitiveValue.cpp: (WebCore::CSSPrimitiveValue::computeLengthDouble):
  • loader/CachedImage.cpp: (WebCore::CachedImage::imageSize): (WebCore::CachedImage::imageRect):
14:05 Changeset [31218] by kmccullough@apple.com

2008-03-21 Kevin McCullough <kmccullough@apple.com>

Reviewed by Sam.

-<rdar://problem/5712804> XMLHttpRequests do not show response contents, preview images (14313)
-<rdar://problem/5712931> XMLHttpRequests should be shown separately and grouped (14315)
-<rdar://problem/5732836> XMLHttpRequest: Inspector should show network activity/XHR in Console (17233)

  • Instrument the InspectorController to cache XHR resrouces since they are not cached by WebCore. The resources will be kept in the same map as all other resources so they will have the exact same lifecycle.
  • ChangeLog:
  • page/InspectorController.cpp: (WebCore::InspectorCachedXMLHttpRequestResource::type): (WebCore::InspectorCachedXMLHttpRequestResource::setType): (WebCore::InspectorCachedXMLHttpRequestResource::data): (WebCore::InspectorCachedXMLHttpRequestResource::setData): (WebCore::InspectorCachedXMLHttpRequestResource::encoding): (WebCore::InspectorCachedXMLHttpRequestResource::setEncoding): (WebCore::InspectorCachedXMLHttpRequestResource::create): (WebCore::InspectorCachedXMLHttpRequestResource::InspectorCachedXMLHttpRequestResource):
  • xml/XMLHttpRequest.cpp: (WebCore::XMLHttpRequest::send):
13:40 Changeset [31217] by zecke@webkit.org

Add missing ChangeLog entry for r31214.

2008-03-21 Rodney Dawes <dobey@wayofthemonkey.com>

Reviewed by Holger.

Remove the JSCore include path options from global_cppflags.
Place JSCore include path options in javascriptcore_cppflags.
Add javascriptcore_cppflags to libWebCore_la_CPPFLAGS.

  • GNUmakefile.am:
12:26 Changeset [31216] by ddkilzer@apple.com

2008-03-21 Arkadiusz Miskiewicz <arekm@maven.pl>

Webkit does not build on linux powerpc

<http://bugs.webkit.org/show_bug.cgi?id=17019>

Reviewed by David Kilzer.

  • wtf/TCSpinLock.h: (TCMalloc_SpinLock::Unlock):
11:48 Changeset [31215] by ap@webkit.org

Reviewed by Geoff.

  • sunspider: Pause update daemon on Darwin to make results reliable again (workaround for <rdar://problem/5811127>).
11:21 Changeset [31214] by zecke@webkit.org

JavaScriptCore/ChangeLog:
2008-03-21 Rodney Dawes <dobey@wayofthemonkey.com>

Reviewed by Holger.

http://bugs.webkit.org/show_bug.cgi?id=17981

Add javascriptcore_cppflags to Programs_minidom_CPPFLAGS.

  • GNUmakefile.am:

WebCore/ChaneLog:
2008-03-21 Rodney Dawes <dobey@wayofthemonkey.com>

Reviewed by Holger.

http://bugs.webkit.org/show_bug.cgi?id=17981

Remove WebCore/ForwardingHeaders from cppflags as it is not needed,
and causes build problems with GTK+.

  • GNUmakefile.am:

WebKitTools/ChangeLog:
2008-03-21 Rodney Dawes <dobey@wayofthemonkey.com>

Reviewed by Holger.

http://bugs.webkit.org/show_bug.cgi?id=17981

Add webcore and javascriptcore cppflags to programs' _CPPFLAGS.

  • GNUmakefile.am:
10:17 Changeset [31213] by mitz@apple.com

WebCore:

Reviewed by Adele Peterson.

Test: fast/block/float/br-with-clear.html

  • rendering/bidi.cpp: (WebCore::RenderBlock::layoutInlineChildren): Fixed the logic for the very special case of an all-lines-clean block that ends with a float- clearing <br> by looking for the <br> at the end of the line rather than at its beginning.

LayoutTests:

Reviewed by Adele Peterson.

  • fast/block/float/br-with-clear.html: Added.
  • platform/mac/fast/block/float/br-with-clear-expected.checksum: Added.
  • platform/mac/fast/block/float/br-with-clear-expected.png: Added.
  • platform/mac/fast/block/float/br-with-clear-expected.txt: Added.
09:36 Changeset [31212] by adachan@apple.com

<rdar://problem/5810324> visitCount not updated in History.plist
Call WebCore::HistoryItem::mergeAutoCompleteHints() in WebHistoryItem::mergeAutoCompleteHints().

Reviewed by Steve.

  • WebHistoryItem.cpp: (WebHistoryItem::mergeAutoCompleteHints): (WebHistoryItem::QueryInterface):
09:29 Changeset [31211] by aroben@apple.com

Remove a non-working mirror from cygwin-downloader and add the ruby package

  • CygwinDownloader/cygwin-downloader.py:
  • CygwinDownloader/cygwin-downloader.zip: Updated.
09:27 Changeset [31210] by aroben@apple.com

Don't allow any newly-scheduled plugin requests to be serviced when JS is paused

This is a followup to r31199.

Reviewed by Tim Hatcher.

  • plugins/PluginView.cpp: (WebCore::PluginView::requestTimerFired): Added an assertion that JS is not paused. (WebCore::PluginView::scheduleRequest): Don't start the request timer if JS is paused. This is the bug fix. (WebCore::PluginView::setJavaScriptPaused): Replaced the use of m_requestTimerWasActive with a check to see if we have any pending requests. m_requestTimerWasActive would not be accurate if a new request had been scheduled while JS was paused. (WebCore::PluginView::PluginView): Removed initialization of m_requestTimerWasActive.
  • plugins/PluginView.h: Removed m_requestTimerWasActive.
09:19 UpdatingCygwinDownloader created by aroben@apple.com
Initial version
09:14 WikiStart edited by aroben@apple.com
Add a link to a page to describe how to update cygwin-downloader (diff)
09:00 Changeset [31209] by timothy@apple.com

Bug 17980: Regression: Inspector highlighting of webpage not cleared when going to new URL
http://bugs.webkit.org/show_bug.cgi?id=17980

Reviewed by Adam.

The new highlight drawing was not honoring the fade value, so it was
always drawing at full opacity. The animation code didn't match Windows
and the new highlight anyway, so it has been removed. The highlight
how just detaches when it is hidden.

  • WebCoreSupport/WebInspectorClient.mm:

(-[WebInspectorWindowController windowShouldClose:]): Call hideHighlight.
(-[WebInspectorWindowController close]): Ditto.
(-[WebInspectorWindowController highlightNode:]): Call attach.
(-[WebInspectorWindowController hideHighlight]): Call detach and release _currentHighlight.

  • WebInspector/WebNodeHighlight.h:
  • WebInspector/WebNodeHighlight.m:

(-[WebNodeHighlight initWithTargetView:inspectorController:]):
(-[WebNodeHighlight dealloc]): Assert we have no _highlightView.
(-[WebNodeHighlight attach]): Renamed from attachHighlight.
(-[WebNodeHighlight detach]): Renamed from detachHighlight.
(-[WebNodeHighlight setNeedsUpdateInTargetViewRect:]): Renamed from setHolesNeedUpdateInTargetViewRect:.

  • WebInspector/WebNodeHighlightView.h:
  • WebInspector/WebNodeHighlightView.m:

(-[WebNodeHighlightView setNeedsDisplayInRect:]): Renamed from setHolesNeedUpdateInRect:.

04:14 Changeset [31208] by ap@webkit.org

Reviewed by Oliver Hunt.

Consolidate static identifier initializers within CommonIdentifiers.

No reliably measurable change on SunSpider; maybe a tiny improvement (within 0.2%).

  • kjs/CommonIdentifiers.h: Added static identifiers that were lazily initialized throughout the code.
  • kjs/date_object.cpp: (KJS::DateObjectImp::DateObjectImp):
  • kjs/function_object.cpp: (KJS::FunctionPrototype::FunctionPrototype):
  • kjs/object_object.cpp: (KJS::ObjectPrototype::ObjectPrototype):
  • kjs/regexp_object.cpp: (KJS::RegExpPrototype::RegExpPrototype): Use the values from CommonIdentifiers.
  • kjs/lookup.h: Caching the identifier in a static wasn't a win on SunSpider, removed it.
  • kjs/value.h: (KJS::jsNaN): We already have a shared NaN value, no need for a duplicate here.
  • wtf/MathExtras.h: (wtf_atan2): Having local variables for numeric_limits constants is good for readability, but there is no reason to keep them static.
00:07 Changeset [31207] by mitz@apple.com
  • fast/css/computed-style-expected.txt:
  • fast/css/computed-style-without-renderer-expected.txt:
  • svg/css/getComputedStyle-basic-expected.txt:

03/20/08:

22:30 Changeset [31206] by mitz@apple.com

WebCore:

2008-03-20 Dan Bernstein <mitz@apple.com>

Reviewed by Anders Carlsson.

  • fix an assertion failure due to a line being marked dirty during computeVerticalPositionsForLine() due to temporary box destruction

Test: fast/inline/clean-after-removing-temp-boxes.html

  • rendering/RenderText.cpp: (WebCore::RenderText::setTextWithOffset): Added code to dirty lines when this method is called on an empty text node, which has no line boxes.
  • rendering/bidi.cpp: (WebCore::RenderBlock::computeVerticalPositionsForLine): Mark the line clean in the end.

LayoutTests:

2008-03-20 Dan Bernstein <mitz@apple.com>

Reviewed by Anders Carlsson.

  • test for an assertion failure due to a line being marked dirty during computeVerticalPositionsForLine() due to temporary box destruction
  • fast/inline/clean-after-removing-temp-boxes-expected.txt: Added.
  • fast/inline/clean-after-removing-temp-boxes.html: Added.
22:00 Changeset [31205] by oliver@apple.com

Fix for leak introduced by inline ScopeChainNode use

Reviewed by Maciej

To avoid any extra branches when managing an inline ScopeChainNode
in the ScopeChain the inline node gets inserted with a refcount of

  1. This meant than when the ScopeChain was destroyed the ScopeChainNodes

above the inline node would be leaked.

We resolve this by manually popping the inline node in the
FunctionExecState destructor.

18:32 Changeset [31204] by andersca@apple.com

2008-03-20 Anders Carlsson <andersca@apple.com>

Reviewed by Maciej.

Make some member variables private.


  • loader/ResourceLoader.cpp: (WebCore::ResourceLoader::ResourceLoader):
  • loader/ResourceLoader.h:
18:18 Changeset [31203] by andersca@apple.com

2008-03-20 Anders Carlsson <andersca@apple.com>

Reviewed by Maciej.

Get rid of m_originalURL in ResourceLoader.

  • loader/ResourceLoader.cpp: (WebCore::ResourceLoader::load): Don't store m_originalURL.


  • loader/ResourceLoader.h:
  • loader/SubresourceLoader.cpp: (WebCore::SubresourceLoader::willSendRequest): Use the url of the current request instead of the original URL.
18:07 Changeset [31202] by andersca@apple.com

2008-03-20 Anders Carlsson <andersca@apple.com>

Reviewed by Maciej.

Small loader cleanups.


  • loader/MainResourceLoader.cpp: (WebCore::MainResourceLoader::willSendRequest): No need to call setRequest here, ResourceLoader::willSendRequest takes care of that. Also, there is no need to check that newRequest is null here, because we assert earlier that it is not.


  • loader/ResourceLoader.h: Get rid of setRequest, it is no longer needed.
18:00 Changeset [31201] by andersca@apple.com

2008-03-20 Anders Carlsson <andersca@apple.com>

Get rid of a couple of unnecessary class declarations.


  • loader/SubresourceLoader.h:
17:30 Changeset [31200] by hyatt@apple.com

2008-03-20 David Hyatt <hyatt@apple.com>

Add a new value to background-clip, "text", that allows backgrounds to be clipped to foreground text (and
associated text decorations like shadows and underlines).

Reviewed by Beth

  • css/CSSParser.cpp: (WebCore::CSSParser::parseBackgroundProperty):
  • css/CSSPrimitiveValueMappings.h: (WebCore::CSSPrimitiveValue::CSSPrimitiveValue): (WebCore::CSSPrimitiveValue::operator EBackgroundBox):
  • platform/graphics/GraphicsContext.cpp: (WebCore::GraphicsContext::clipToImageBuffer):
  • platform/graphics/GraphicsContext.h:
  • platform/graphics/cg/GraphicsContextCG.cpp: (WebCore::GraphicsContext::clipToImageBuffer):
  • rendering/InlineBox.cpp: (WebCore::InlineBox::paint):
  • rendering/InlineFlowBox.cpp: (WebCore::InlineFlowBox::paintBackgrounds): (WebCore::InlineFlowBox::paintBackground): (WebCore::InlineFlowBox::paintBoxDecorations):
  • rendering/InlineFlowBox.h:
  • rendering/InlineTextBox.cpp: (WebCore::InlineTextBox::paint):
  • rendering/RenderBlock.cpp: (WebCore::RenderBlock::paintColumns): (WebCore::RenderBlock::paintObject): (WebCore::RenderBlock::paintFloats):
  • rendering/RenderBlock.h:
  • rendering/RenderBox.cpp: (WebCore::RenderBox::paintRootBoxDecorations): (WebCore::RenderBox::paintBoxDecorations): (WebCore::RenderBox::paintBackgrounds): (WebCore::RenderBox::paintBackground): (WebCore::RenderBox::paintBackgroundExtended):
  • rendering/RenderBox.h:
  • rendering/RenderFieldset.cpp: (WebCore::RenderFieldset::paintBoxDecorations):
  • rendering/RenderFlow.cpp: (WebCore::RenderFlow::paintLines):
  • rendering/RenderObject.h: (WebCore::): (WebCore::RenderObject::paintBackgroundExtended):
  • rendering/RenderStyle.h: (WebCore::):
  • rendering/RenderTable.cpp: (WebCore::RenderTable::paintBoxDecorations):
  • rendering/RenderTableCell.cpp: (WebCore::RenderTableCell::paintBackgroundsBehindCell):
17:14 Changeset [31199] by aroben@apple.com

Allow pausing/blocking of JS execution by plugins

Part of Bug 17133: Should support pausing JavaScript execution without
hanging the process

<http://bugs.webkit.org/show_bug.cgi?id=17133>
<rdar://problem/5719551>

This patch doesn't affect Mac, which doesn't use the shared PluginView
code.

Note that this patch doesn't prevent plugins from executing JS via an
NPObject they've already gotten hold of. It just blocks obtaining new
NPObjects and pauses any requests to evaluate javascript: URIs. This
is probably good enough for now because most plugins seem to always
obtain the Window object each time they want to execute JS.

Reviewed by Tim Hatcher.

  • plugins/PluginView.cpp: (WebCore::PluginView::getValue): Return an error if JS is paused. (WebCore::PluginView::setJavaScriptPaused): Stop the request timer if we're pausing, and resume it if we're unpausing. (WebCore::PluginView::PluginView): Initialize new members.
  • plugins/PluginView.h:
17:14 Changeset [31198] by aroben@apple.com

Allow pausing of callOnMainThread callbacks

Part of Bug 17133: Should support pausing JavaScript execution without
hanging the process

<http://bugs.webkit.org/show_bug.cgi?id=17133>
<rdar://problem/5719551>

Reviewed by Tim Hatcher.

  • platform/MainThread.cpp: (WebCore::dispatchFunctionsFromMainThread): If callbacks are paused, don't dispatch the functions. (WebCore::setMainThreadCallbacksPaused): Added. If we're being unpaused, call scheduleDispatchFunctionsOnMainThread so that any queued callbacks will get dispatched in the near future.
  • platform/MainThread.h:
17:13 Changeset [31197] by aroben@apple.com

Allow blocking of JS event handlers/javascript: URIs per-Frame

Part of Bug 17133: Should support pausing JavaScript execution without
hanging the process

<http://bugs.webkit.org/show_bug.cgi?id=17133>
<rdar://problem/5719551>

Two new methods are added to KJSProxy: setPaused and isPaused. When
setPaused(true) is called, JS event handlers are blocked and
javascript: URIs will not be evaluated.

Reviewed by Tim Hatcher.

  • bindings/js/kjs_events.cpp: (WebCore::JSAbstractEventListener::handleEvent):
    • Removed some old KJS_DEBUGGER code
    • Don't run the handler if the KJSProxy is paused.
  • bindings/js/kjs_proxy.cpp: (WebCore::KJSProxy::KJSProxy): Initialize new member.
  • bindings/js/kjs_proxy.h: Added new methods.
  • loader/FrameLoader.cpp: (WebCore::FrameLoader::executeScript): Don't execute the script if the KJSProxy is paused.
16:50 Changeset [31196] by jhoneycutt@apple.com

2008-03-20 Jon Honeycutt <jhoneycutt@apple.com>

Reviewed by Anders.

Fix Windows warning / leak: warning C4150: deletion of pointer to
incomplete type 'WebCore::PluginRequest'; no destructor called

  • plugins/PluginView.cpp: Move PluginRequest class to PluginView.h so Windows PluginView destructor can use WTF::deleteAllValues to clean up m_requests.
  • plugins/PluginView.h: (WebCore::PluginRequest::PluginRequest): (WebCore::PluginRequest::frameLoadRequest): (WebCore::PluginRequest::notifyData): (WebCore::PluginRequest::sendNotification): (WebCore::PluginRequest::shouldAllowPopups):
16:18 Changeset [31195] by mrowe@apple.com

Ensure that the defines in FEATURE_DEFINES are sorted so that they will match the default settings of build-webkit.
This will prevent the world from being rebuilt if you happen to switch between building in Xcode and with build-webkit on the
command-line.

Reviewed by Sam Weinig.

16:17 Changeset [31194] by aroben@apple.com

Add some more bug references to ChangeLog from r31079

16:16 Developer Tools Hitlist edited by aroben@apple.com
Removed another fixed bug (diff)
16:11 Developer Tools Hitlist edited by aroben@apple.com
Removed a fixed bug (diff)
16:00 Changeset [31193] by jhoneycutt@apple.com

2008-03-20 Jon Honeycutt <jhoneycutt@apple.com>

Reviewed by Anders.

<rdar://problem/5809600> REGRESSION:
http/tests/plugins/cross-frame-object-access.html hangs Windows Layout
Tests

r30897 changed the way we conditionalize this feature.

  • plugins/PluginView.cpp: (WebCore::PluginView::getValue): Test ENABLE(NETSCAPE_PLUGIN_API) instead of USE(NPOBJECT). (WebCore::PluginView::bindingInstance): Same.
15:59 Changeset [31192] by aroben@apple.com

Highlight nodes when you hover over them in the Console

Part of Bug 16532: Inspector should highlight nodes in page when
hovering over nodes in Inspector's interface

<http://bugs.webkit.org/show_bug.cgi?id=16532>
<rdar://problem/5712896>

Reviewed by Tim Hatcher.

  • page/inspector/ConsolePanel.js: (WebInspector.ConsolePanel._formatnode): Add mouseover/mouseout event listeners to highlight the node and clear the highlight.
15:59 Changeset [31191] by aroben@apple.com

Highlight nodes in the page when you hover over their DOM breadcrumbs

Part of Bug 16532: Inspector should highlight nodes in page when
hovering over nodes in Inspector's interface

<http://bugs.webkit.org/show_bug.cgi?id=16532>
<rdar://problem/5712896>

Reviewed by Tim Hatcher.

  • page/inspector/DocumentPanel.js: (WebInspector.DocumentPanel): Add an mouseout event listener to the breadcrumbs element to clear the highlighted node. (WebInspector.DocumentPanel.updateBreadcrumbs): Change the mouseover event listener to highlight the node represented by the hovered breadcrumb.
15:58 Changeset [31190] by aroben@apple.com

Highlight nodes in the page when you hover over them in the DOM tree

Part of Bug 16532: Inspector should highlight nodes in page when
hovering over nodes in Inspector's interface

<http://bugs.webkit.org/show_bug.cgi?id=16532>
<rdar://problem/5712896>

The inspected node is no longer highlighted (unless, of course, you
hover over it).

Reviewed by Tim Hatcher.

  • page/inspector/DocumentPanel.js: (WebInspector.DocumentPanel):
    • Don't highlight the focused node when the DOM tree is shown
    • Added mousemove/mouseout event listeners to set/clear the highlighted node. These are added to the root of the DOM tree instead of to each individual list item to avoid flashing as the mouse moves between nodes.

(WebInspector.DocumentPanel.set focusedDOMNode): Don't highlight the
focused node.
(WebInspector.DocumentPanel._onmousemove): Highlight the node under
the mouse.

15:58 Changeset [31189] by aroben@apple.com

Show nodes' content/padding/border/margin boxes in the node highlight

Part of Bug 17221: Node highlight should show node metrics

<http://bugs.webkit.org/show_bug.cgi?id=17221>
<rdar://problem/5732822>

Outstanding issues:

1) We don't show padding/border/margins for inlines
2) We don't show any numeric metrics, we just draw the boxes
3) We'll probably want to tweak the look of the highlight some, at

least to make the boxes better distinguishable

Reviewed by Tim Hatcher.

  • page/InspectorController.cpp: (WebCore::drawOutlinedRect): Added. Just draws a single rect. (WebCore::drawHighlightForBoxes): Added. Takes the rects we calculated for the node and draws the highlight. (WebCore::InspectorController::drawNodeHighlight): Calculates the content/padding/border/margin boxes for blocks and passes them off to drawHighlightForBoxes. The behavior for inlines is for now unchanged.
15:57 Changeset [31188] by aroben@apple.com

Fix RenderContainer's override of addLineBoxRects to match RenderObject's

RenderContainer::addLineBoxRects was never getting called because its
parameters didn't match those of RenderObject::addLineBoxRects.

Reviewed by Mitz Pettel.

No test possible.

  • rendering/RenderContainer.cpp: Added an optional bool useSelectionHeight parameter to match RenderObject's method.
  • rendering/RenderContainer.h: Ditto.
15:38 Changeset [31187] by mrowe@apple.com

Tweak NodeList tests to ensure that they force a GC outside of DRT.

Rubber-stamped by Sam Weinig.

  • fast/dom/NodeList/5725058-crash-scenario-1.html:
  • fast/dom/NodeList/5725058-crash-scenario-2.html:
15:37 Changeset [31186] by mitz@apple.com

Reviewed by Adam Roben.

  • make makeTextLarger() and zoomPageIn() do what they are supposed to do
  • WebView.cpp: (WebView::makeTextLarger): Changed call to canZoomIn() to zoomIn(). (WebView::zoomPageIn): Ditto.
14:00 Changeset [31185] by alp@webkit.org

2008-03-20 Marco Barisione <marco.barisione@collabora.co.uk>

Reviewed by Alp Toker.

http://bugs.webkit.org/show_bug.cgi?id=17946
[GTK] Widgets are not clipped

  • platform/gtk/RenderThemeGtk.cpp: (WebCore::paintMozWidget): pass the clipping rectangle to moz_gtk_widget_paint() instead of just the widget rectangle.
13:44 Changeset [31184] by alp@webkit.org

2008-03-20 Alp Toker <alp@atoker.com>

GTK+ build fix. Back out r31183. This patch also introduced API style
issues.

  • WebCoreSupport/FrameLoaderClientGtk.cpp: (WebKit::FrameLoaderClient::dispatchDecidePolicyForNavigationAction):
  • webkit/headers.pri:
  • webkit/webkit-marshal.list:
  • webkit/webkitdefines.h:
  • webkit/webkitnavigationaction.cpp: Removed.
  • webkit/webkitnavigationaction.h: Removed.
  • webkit/webkitprivate.cpp:
  • webkit/webkitprivate.h:
  • webkit/webkitwebview.cpp:
  • webkit/webkitwebview.h:
11:19 Changeset [31183] by andersca@apple.com

.:

2008-03-20 Jasper Bryant-Greene <jasper@unix.geek.nz>

Reviewed by Anders.

Resolves http://bugs.webkit.org/show_bug.cgi?id=16092
"[GTK] Middle-mouse click should allow opening a URL in a new tab"

Added WebKitNavigationAction object to GTK API

  • GNUmakefile.am:

WebKit/gtk:

2008-03-20 Jasper Bryant-Greene <jasper@unix.geek.nz>

Reviewed and tweaked by Anders.

Resolves http://bugs.webkit.org/show_bug.cgi?id=16092
"[GTK] Middle-mouse click should allow opening a URL in a new tab"

Created WebKitNavigationAction object exported through the API. The
navigation-requested signal provides this object as context for the
requested navigation, allowing the application to decide what to do
with the navigation based on which mouse button was used, which
modifier keys were held down, etc. This allows, for example, the
application to open links in a new tab when either middle-click or
control-click are used to initiate the navigation.

  • WebCoreSupport/FrameLoaderClientGtk.cpp: (WebKit::FrameLoaderClient::dispatchDecidePolicyForNavigationAction):
  • webkit/headers.pri:
  • webkit/webkit-marshal.list:
  • webkit/webkitdefines.h:
  • webkit/webkitnavigationaction.cpp: Added. (webkit_navigation_action_finalize): (webkit_navigation_action_class_init): (webkit_navigation_action_init): (webkit_navigation_action_get_button): (webkit_navigation_action_get_modifier_flags): (webkit_navigation_action_get_navigation_type): (webkit_navigation_action_get_original_url):
  • webkit/webkitnavigationaction.h: Added.
  • webkit/webkitprivate.cpp: (WebKit::kit):
  • webkit/webkitprivate.h:
  • webkit/webkitwebview.cpp:
  • webkit/webkitwebview.h:
10:59 Changeset [31182] by hyatt@apple.com

Added layout test for dynamic sibling selector patch.

10:59 Changeset [31181] by hyatt@apple.com

2008-03-20 David Hyatt <hyatt@apple.com>

Fix for http://bugs.webkit.org/show_bug.cgi?id=9279

Make :hover work with the adjacent sibling selector. This fix makes all forms of dynamic changes
(class name changes, :hover, :focus, etc.) work properly when used with the + selector.

Reviewed by weinig

Added fast/css/dynamic-sibling-selector.html

  • dom/Element.cpp: (WebCore::Element::recalcStyle):
10:22 Changeset [31180] by hyatt@apple.com

Add full page zoom API for Windows WebKit.

09:18 Changeset [31179] by sullivan@apple.com

WebKit:

2008-03-20 John Sullivan <sullivan@apple.com>

  • StringsNotToBeLocalized.txt: Brought this file up to date

WebKit/win:

2008-03-20 John Sullivan <sullivan@apple.com>

  • English.lproj/Localizable.strings: Brought this file up to date
09:09 Changeset [31178] by ddkilzer@apple.com

WebCore:

2008-03-20 Aaron Golden <aegolden@gmail.com>

Reviewed by Darin and David Kilzer.

Addresses <http://bugs.webkit.org/show_bug.cgi?id=15263>, which caused certain marquees to not display.

Test: fast/html/marquee-scroll.html

  • html/HTMLMarqueeElement.cpp: (WebCore::HTMLMarqueeElement::parseMappedAttribute):

LayoutTests:

2008-03-20 Aaron Golden <aegolden@gmail.com>

Reviewed by Darin and David Kilzer.

  • fast/html/marquee-scroll.html: Added.
  • platform/mac/fast/html/marquee-scroll-expected.checksum: Added.
  • platform/mac/fast/html/marquee-scroll-expected.png: Added.
  • platform/mac/fast/html/marquee-scroll-expected.txt: Added.
09:01 Changeset [31177] by aroben@apple.com

Make WebNodeHighlightView use InspectorController to do its painting

WebCore:

Export InspectorController::drawNodeHighlight

Reviewed by Tim Hatcher.

  • WebCore.base.exp: Also sorted this file.

WebKit:

Mark WebNodeHighlight.m and WebNodeHighlightView.m Obj-C++

Reviewed by Tim Hatcher.

  • WebKit.xcodeproj/project.pbxproj:

WebKit/mac:

Make WebNodeHighlightView use InspectorController to do its painting

Reviewed by Tim Hatcher.

  • WebCoreSupport/WebInspectorClient.mm: (-[WebInspectorWindowController highlightNode:]): Pass the InspectorController to the WebNodeHighlight, and don't call setHighlightedNode: (which has been removed). (-[WebInspectorWindowController hideHighlight]): Removed call to setHighlightedNode:.
  • WebInspector/WebNodeHighlight.h:
    • Replaced _highlightNode with _inspectorController
    • Removed _highlightedNode accessors
    • Added -inspectorController method
  • WebInspector/WebNodeHighlight.m: (-[WebNodeHighlight initWithTargetView:inspectorController:]): Now takes an InspectorController* and stores it in _inspectorController. (-[WebNodeHighlight dealloc]): Removed code dealing with _highlightedNode. (-[WebNodeHighlight inspectorController]): Added.
  • WebInspector/WebNodeHighlightView.m: Removed FileInternal category. (-[WebNodeHighlightView isFlipped]): Added. WebCore expects all GraphicsContexts to be based on a flipped CGContext, so we have to specify that this view is flipped. (-[WebNodeHighlightView drawRect:]): Changed to create a GraphicsContext and pass it to InspectorController::drawNodeHighlight.
07:33 Changeset [31176] by ddkilzer@apple.com

2008-03-20 David Krause <david.krause@gmail.com>

Reviewed by David Kilzer.

Fix http://bugs.webkit.org/show_bug.cgi?id=17923
Bug 17923: ARM platform endian defines inaccurate

  • wtf/Platform.h: Replaced !defined(ARMEL) check with !defined(VFP_FP) for PLATFORM(MIDDLE_ENDIAN)
03:20 Changeset [31175] by mjs@apple.com

2008-03-20 Maciej Stachowiak <mjs@apple.com>

  • fix build
  • ForwardingHeaders/wtf/ListRefPtr.h: Added.
03:20 Changeset [31174] by mjs@apple.com

2008-03-20 Maciej Stachowiak <mjs@apple.com>

  • fix build
  • JavaScriptCore.xcodeproj/project.pbxproj: install Activation.h as private
02:34 Changeset [31173] by mjs@apple.com

2008-03-20 Maciej Stachowiak <mjs@apple.com>

Reviewed by Oliver.


  • reduce function call overhead for 1.014x speedup on SunSpider

I moved some functions from ExecState.cpp to ExecStateInline.h and
from JSGlobalObject.cpp to JSGlobalObject.h, and declared them
inline; machine function call overhead for these was hurting JS
funcion call overhead.


  • kjs/ExecState.cpp:
  • kjs/ExecStateInlines.h: Added. (KJS::ExecState::ExecState): (KJS::ExecState::~ExecState): (KJS::FunctionExecState::FunctionExecState): (KJS::FunctionExecState::~FunctionExecState):
  • kjs/JSGlobalObject.cpp:
  • kjs/JSGlobalObject.h: (KJS::JSGlobalObject::pushActivation): (KJS::JSGlobalObject::checkActivationCount): (KJS::JSGlobalObject::popActivation):
  • kjs/function.cpp:
Note: See TracTimeline for information about the timeline view.