Timeline
Jan 27, 2006:
- 7:26 PM Changeset in webkit [12424] by
-
- 3 edits in trunk/WebCore
Reviewed by Geoff.
- GC after destroying the very last frame to avoid node leak false positives http://bugzilla.opendarwin.org/show_bug.cgi?id=6884
- ForwardingHeaders/JSLock.h: added
- page/Page.cpp: (Page::~Page): In debug mode, when destroying the last page, 0 out m_mainFrame to destroy the very last frame and then garbage collect, so a GC happens at a time when the stack won't have references into the JS heap for the conservative GC to catch.
- 7:07 PM Changeset in webkit [12423] by
-
- 9 edits1 add in trunk/WebCore
Reviewed by Eric.
- remove nearly all the substance of ObjectContents and put it on Frame http://bugzilla.opendarwin.org/show_bug.cgi?id=6885
- ForwardingHeaders/kjs/JSLock.h: Added.
- bridge/mac/MacFrame.mm: (Plugin::Plugin): renamed from KWQPluginPart, make it able to store a view. (Plugin::view): And return the QWidget* view. (MacFrame::createPart): Use Plugin instead of KWQPluginPart, avoid setWidget. (MacFrame::setView): Don't use m_url directly so much, and use through d-> if needed. (MacFrame::userAgent): ditto (MacFrame::openURLFromPageCache): ditto
- kwq/KWQObject.cpp: (QObject::inherits): Removed case for ObjectContents, which is not long for this world.
- kwq/KWQObject.h:
- page/Frame.cpp: (Frame::init): Don't pass parent() to FramePrivate, we can't possibly have one yet. (Frame::restoreURL): d->m_url, not m_url (Frame::didOpenURL): ditto (Frame::executeScript): ditto (Frame::receivedFirstData): ditto (Frame::begin): ditto (Frame::gotoAnchor): ditto (Frame::requestObject): ditto, plus only use Frame methods when we really have a Frame. (Frame::processObjectRequest): ditto, also get view() not widget(). (Frame::childFrame): remove excess spaces and pointless assert. (Frame::parentFrame): just call parent() (could probably remove this). (Frame::reparseConfiguration): d->m_url not m_url (Frame::canCachePage): ditto (Frame::updatePolicyBaseURL): ditto (Frame::scrollToAnchor): ditto (Frame::setName): set in FramePrivate (Frame::name): get from FramePrivate (Frame::setParent): similarly (Frame::parent): similarly (Frame::url): similarly
- page/Frame.h:
- page/FramePrivate.h: (FramePrivate::FramePrivate): Added m_parent, m_name and m_url.
- page/FrameView.cpp: (FrameView::viewportMousePressEvent): Call frame directly instead of sending through QApplication. (FrameView::viewportMouseDoubleClickEvent): ditto (FrameView::viewportMouseMoveEvent): ditto (FrameView::viewportMouseReleaseEvent): ditto
- page/ObjectContents.h: (ObjectContents::ObjectContents): Removed url, setParent, parent, name, setName, openURL, closeURL, widget, setWidget, event and customEvent methods, and corresponding fields. Add pure virtual view() method.
- 5:19 PM Changeset in webkit [12422] by
-
- 1 edit2 adds in trunk/LayoutTests
Reviewed by Adele.
- test for http://bugzilla.opendarwin.org/show_bug.cgi?id=6075 REGRESSION: event.relatedTarget not working, breaks menus at www.ati.com
- fast/events/related-target-expected.txt: Added.
- fast/events/related-target.html: Added.
- 5:18 PM Changeset in webkit [12421] by
-
- 4 edits in trunk/WebCore
Reviewed by Adele.
- fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6075 REGRESSION: event.relatedTarget not working, breaks menus at www.ati.com
Test: fast/events/related-target.html
- khtml/xml/NodeImpl.h:
- khtml/xml/NodeImpl.cpp: (WebCore::NodeImpl::dispatchMouseEvent): Added a relatedTarget parameter to a couple overloads.
- page/FrameView.cpp: (FrameView::viewportMouseMoveEvent): Moved code to set prevMouseX/Y out of here and into dispatchMouseEvent, because it now handles mouseout and mouseover events, and that's what these globals are used for. This fixes DumpRenderTree to work properly with mouseout and mouseover (because it delivers only mouse down and up, no mouse move events). (FrameView::dispatchMouseEvent): Passed in the "other" node as the related target for mouseout and mouseover events. Also set prevMouseX/Y.
- 5:05 PM Changeset in webkit [12420] by
-
- 2 edits in trunk/WebKitTools
2006-01-27 Eric Seidel <eseidel@apple.com>
Reviewed by adele.
- Scripts/run-webkit-tests: make new tests use absolute urls
- 3:41 PM Changeset in webkit [12419] by
-
- 2 edits4 adds in trunk
LayoutTests:
Adding test for
<rdar://problem/4422624> REGRESSION: crash loading webmail.mac.com
- fast/forms/input-no-renderer-expected.checksum: Added.
- fast/forms/input-no-renderer-expected.png: Added.
- fast/forms/input-no-renderer-expected.txt: Added.
- fast/forms/input-no-renderer.html: Added.
WebCore:
Adding comment for added layout test.
- 3:06 PM Changeset in webkit [12418] by
-
- 1 edit in trunk/WebCore/ChangeLog
Updated ChangeLog to mention layout tests landed with previous commits.
- 2:39 PM Changeset in webkit [12417] by
-
- 1 edit in trunk/WebCore/ChangeLog
Updating ChangeLog with explanation of missing test case.
- 1:18 PM Changeset in webkit [12416] by
-
- 1 edit in trunk/WebCore/ChangeLog
Updated ChangeLog to explain why my last commit did not include a test case.
- 2:52 AM Changeset in webkit [12415] by
-
- 2 edits in trunk/WebKitTools
2006-01-27 Eric Seidel <eseidel@apple.com>
Reviewed by mjs.
make run-webkit-tests output total leaks count
- Scripts/run-webkit-tests:
- 2:22 AM Changeset in webkit [12414] by
-
- 76 edits2 moves7 deletes in trunk/WebCore
Fix for 6867. Rename QPixmap to Image and move it to platform/.
Eliminate QPaintDevice and QPaintDeviceMetrics completely, since
they aren't necessary.
Reviewed by anders
- ForwardingHeaders/q3paintdevicemetrics.h: Removed.
- ForwardingHeaders/qpaintdevice.h: Removed.
- ForwardingHeaders/qpaintdevicemetrics.h: Removed.
- ForwardingHeaders/qpixmap.h: Removed.
- WebCore.xcodeproj/project.pbxproj:
- bridge/mac/MacFrame.mm: (MacFrame::fileWrapperForElement):
- bridge/mac/WebCoreFrameBridge.mm: (-[WebCoreFrameBridge reapplyStylesForDeviceType:]): (nowPrinting): (-[WebCoreFrameBridge elementAtPoint:]):
- css/css_valueimpl.cpp: (WebCore::CSSPrimitiveValueImpl::computeLength): (WebCore::CSSPrimitiveValueImpl::computeLengthFloat):
- css/css_valueimpl.h:
- css/csshelper.cpp:
- css/cssstyleselector.cpp: (WebCore::CSSStyleSelector::CSSStyleSelector): (WebCore::CSSStyleSelector::init): (WebCore::CSSStyleSelector::initForStyleResolve): (WebCore::CSSStyleSelector::styleForElement): (WebCore::CSSStyleSelector::pseudoStyleForElement): (WebCore::convertToLength): (WebCore::CSSStyleSelector::applyProperty): (WebCore::CSSStyleSelector::mapBackgroundXPosition): (WebCore::CSSStyleSelector::mapBackgroundYPosition):
- css/cssstyleselector.h:
- kcanvas/KCanvasFilters.h: (KCanvasFEImage::image): (KCanvasFEImage::setImage):
- kcanvas/KCanvasImage.h:
- kcanvas/device/quartz/KCanvasFilterQuartz.mm: (KCanvasFEImageQuartz::getCIFilter):
- kcanvas/device/quartz/KCanvasResourcesQuartz.h: (KCanvasImageQuartz::init):
- khtml/ecma/kjs_events.cpp: (KJS::ClipboardProtoFunc::callAsFunction):
- khtml/ecma/kjs_html.cpp: (KJS::KJS::Context2DFunction::callAsFunction): (KJS::drawPattern): (KJS::_rh): (KJS::ImagePattern::createPattern):
- khtml/ecma/kjs_html.h: (KJS::ImagePattern::image):
- khtml/ecma/kjs_window.cpp: (KJS::Screen::getValueProperty):
- khtml/editing/SelectionController.h:
- khtml/html/html_imageimpl.h: (WebCore::HTMLImageElementImpl::image):
- khtml/misc/helper.h:
- khtml/xml/DocumentImpl.cpp: (WebCore::DocumentImpl::DocumentImpl): (WebCore::DocumentImpl::~DocumentImpl): (WebCore::DocumentImpl::recalcStyle): (WebCore::DocumentImpl::attach): (WebCore::DocumentImpl::detach):
- khtml/xml/DocumentImpl.h: (WebCore::DocumentImpl::printing): (WebCore::DocumentImpl::setPrinting):
- khtml/xml/dom2_eventsimpl.h:
- ksvg2/css/SVGCSSStyleSelector.cpp:
- ksvg2/misc/KCanvasRenderingStyle.cpp: (WebCore::KSVGPainterFactory::cssPrimitiveToLength): (WebCore::KSVGPainterFactory::strokePainter):
- ksvg2/svg/SVGCursorElementImpl.cpp: (SVGCursorElementImpl::notifyFinished):
- ksvg2/svg/SVGCursorElementImpl.h: (KSVG::SVGCursorElementImpl::image):
- ksvg2/svg/SVGDocumentImpl.cpp: (SVGDocumentImpl::SVGDocumentImpl):
- ksvg2/svg/SVGFEImageElementImpl.cpp: (SVGFEImageElementImpl::notifyFinished):
- ksvg2/svg/SVGMaskElementImpl.cpp: (KSVG::SVGMaskElementImpl::drawMaskerContent):
- ksvg2/svg/SVGSVGElementImpl.cpp:
- kwq/KWQApplication.h:
- kwq/KWQApplication.mm: (QDesktopWidget::screenDepth):
- kwq/KWQClipboard.h:
- kwq/KWQClipboard.mm: (WebCore::KWQClipboard::dragImage): (WebCore::KWQClipboard::setDragImage): (WebCore::KWQClipboard::setDragImageElement):
- kwq/KWQComboBox.mm:
- kwq/KWQCursor.h:
- kwq/KWQCursor.mm: (WebCore::createCustomCursor): (WebCore::QCursor::QCursor):
- kwq/KWQKPartsEvent.h:
- kwq/KWQObject.h:
- kwq/KWQPaintDevice.h: Removed.
- kwq/KWQPaintDeviceMetrics.h: Removed.
- kwq/KWQPaintDeviceMetrics.mm: Removed.
- kwq/KWQPainter.h: (WebCore::QPainter::printing):
- kwq/KWQPainter.mm: (WebCore::QPainter::drawImage): (WebCore::QPainter::drawFloatImage): (WebCore::QPainter::drawTiledImage): (WebCore::QPainter::drawScaledAndTiledImage):
- kwq/KWQPixmap.h: Removed.
- kwq/KWQPixmap.mm: Removed.
- kwq/KWQPrinter.h: (QPrinter::QPrinter):
- kwq/KWQWidget.h:
- loader/Cache.cpp: (khtml::Cache::init): (khtml::Cache::clear): (khtml::Cache::requestImage): (khtml::Cache::getStatistics):
- loader/Cache.h:
- loader/CachedImage.cpp: (WebCore::CachedImage::CachedImage): (WebCore::CachedImage::ref): (WebCore::CachedImage::tiled_image): (WebCore::CachedImage::image): (WebCore::CachedImage::image_size): (WebCore::CachedImage::do_notify): (WebCore::CachedImage::data): (WebCore::CachedImage::error):
- loader/CachedImage.h:
- loader/CachedImageCallback.cpp: (WebCore::CachedImageCallback::notifyUpdate): (WebCore::CachedImageCallback::notifyFinished): (WebCore::CachedImageCallback::handleError):
- loader/CachedObject.h: (WebCore::CachedObject::):
- loader/CachedObjectClient.h: (WebCore::CachedObjectClient::setImage):
- loader/DocLoader.cpp: (khtml::DocLoader::setAutoloadImages): (khtml::DocLoader::setShowAnimations):
- loader/loader.cpp: (WebCore::Loader::servePendingRequests): (WebCore::Loader::slotFinished):
- page/Frame.cpp: (Frame::paint):
- page/Frame.h:
- page/FrameView.cpp: (selectCursor):
- page/FrameView.h:
- platform/Image.h: Added.
- platform/mac/Image.mm: Added. (WebCore::Image::loadResource): (WebCore::Image::Image): (WebCore::Image::~Image): (WebCore::Image::imageRef): (WebCore::Image::resetAnimation): (WebCore::Image::setAnimationRect): (-[WebImageCallback initWithCallback:WebCore::]): (WebCore::Image::shouldUseThreadedDecoding): (WebCore::Image::receivedData): (WebCore::Image::mask): (WebCore::Image::isNull): (WebCore::Image::size): (WebCore::Image::rect): (WebCore::Image::width): (WebCore::Image::height): (WebCore::Image::resize): (WebCore::Image::operator=): (WebCore::Image::increaseUseCount): (WebCore::Image::decreaseUseCount): (WebCore::Image::stopAnimations): (WebCore::Image::flushRasterCache):
- rendering/InlineTextBox.cpp: (khtml::InlineTextBox::paint):
- rendering/InlineTextBox.h:
- rendering/RenderBlock.cpp: (WebCore::RenderBlock::paintChildren): (WebCore::RenderBlock::paintObject):
- rendering/RenderText.h:
- rendering/font.cpp: (khtml::Font::update):
- rendering/font.h:
- rendering/render_box.cpp: (WebCore::RenderBox::paintBackgroundExtended):
- rendering/render_canvasimage.cpp: (WebCore::RenderCanvasImage::paint):
- rendering/render_flow.cpp: (RenderFlow::paintLines):
- rendering/render_image.cpp: (WebCore::RenderImage::RenderImage): (WebCore::RenderImage::~RenderImage): (WebCore::RenderImage::setContentObject): (WebCore::RenderImage::setImage): (WebCore::RenderImage::paint): (WebCore::RenderImage::layout): (WebCore::RenderImage::calcReplacedWidth): (WebCore::RenderImage::calcReplacedHeight):
- rendering/render_image.h: (WebCore::RenderImage::image): (WebCore::RenderImage::getImage):
- rendering/render_line.cpp: (WebCore::InlineFlowBox::paintBackground): (WebCore::InlineFlowBox::paintBackgroundAndBorder):
- rendering/render_list.cpp: (RenderListItem::getAbsoluteRepaintRect): (RenderListMarker::paint): (RenderListMarker::setImage): (RenderListMarker::calcMinMaxWidth):
- rendering/render_list.h:
- rendering/render_object.cpp: (WebCore::RenderObject::mustRepaintBackgroundOrBorder): (WebCore::RenderObject::paintBorderImage): (WebCore::RenderObject::paintBorder): (WebCore::RenderObject::setImage):
- rendering/render_object.h:
- rendering/render_replaced.cpp: (WebCore::RenderWidget::paint):
- rendering/render_replaced.h:
- rendering/render_theme_mac.mm: (khtml::RenderThemeMac::setFontFromControlSize):
- 2:20 AM Changeset in webkit [12413] by
-
- 2 edits in trunk/WebKitTools
2006-01-27 Eric Seidel <eseidel@apple.com>
Reviewed by mjs.
run-webkit-test --leaks crashes (malloc logging runs out of memory)
http://bugzilla.opendarwin.org/show_bug.cgi?id=6869
- Scripts/run-webkit-tests: fix --leaks to not crash
- 1:39 AM Changeset in webkit [12412] by
-
- 2 edits in trunk
Somehow these ChangeLog entries weren't committed
- 1:35 AM Changeset in webkit [12411] by
-
- 3 edits in trunk/LayoutTests
- revised test for http://bugzilla.opendarwin.org/show_bug.cgi?id=6731 REGRESSION: change event fires at load time for checked items
- fast/forms/radio-button-no-change-event.html: Updated to add a second radio button.
- fast/forms/radio-button-no-change-event-expected.txt: Updated.
- 1:29 AM Changeset in webkit [12410] by
-
- 2 edits in trunk/WebCore
Reviewed by Maciej.
- fix another part of http://bugzilla.opendarwin.org/show_bug.cgi?id=6731 REGRESSION: change event fires at load time for checked items
Revised test: fast/forms/radio-button-no-change-event.html
- khtml/html/HTMLInputElementImpl.cpp: (WebCore::HTMLInputElementImpl::setChecked): Don't send a change event for a radio button getting unchecked (matches Gecko at least).
- 12:10 AM Changeset in webkit [12409] by
-
- 2 edits in trunk/WebKitTools
2006-01-26 Eric Seidel <eseidel@apple.com>
Reviewed by darin.
run-webkit-tests should produce a self-contained results directory
http://bugzilla.opendarwin.org/show_bug.cgi?id=6864
- Scripts/run-webkit-tests: copy failing items to results dir.
Jan 26, 2006:
- 10:55 PM Changeset in webkit [12408] by
-
- 4 edits in trunk/WebCore
2006-01-26 Eric Seidel <eseidel@apple.com>
Reviewed by mjs.
Leak fixes in SVG code.
http://bugzilla.opendarwin.org/show_bug.cgi?id=6588
- kcanvas/device/quartz/KRenderingPaintServerGradientQuartz.mm: (teardownShadingWithStyle): delete image and context when through. (KRenderingPaintServerGradientQuartz::KRenderingPaintServerGradientQuartz): zero out cached mask (KRenderingPaintServerGradientQuartz::~KRenderingPaintServerGradientQuartz): destroy cached mask image. (KRenderingPaintServerLinearGradientQuartz::setup): delete mask before making new (KRenderingPaintServerRadialGradientQuartz::setup): delete mask before making new
- ksvg2/misc/KCanvasRenderingStyle.cpp: (WebCore::sharedSolidPaintServer): forgot "static" keyword
- ksvg2/svg/SVGPolyElementImpl.cpp: (SVGPolyElementImpl::notifyAttributeChange): use RefPtr
- 10:05 PM Changeset in webkit [12407] by
-
- 4 edits in trunk/WebCore
Reviewed by Eric.
- various leaks on "fast" LayoutTests
http://bugzilla.opendarwin.org/show_bug.cgi?id=6819
http://bugzilla.opendarwin.org/show_bug.cgi?id=6823
http://bugzilla.opendarwin.org/show_bug.cgi?id=6824
It turns out these all had the same cause.
- page/Frame.cpp: (Frame::clear): If we have a document, make sure to cancel parsing, in case it has a tokenizer and parser that are keeping it alive. Use cancelParsing to avoid possibly accidentally firing the onload handler.
- khtml/xml/DocumentImpl.cpp: (WebCore::DocumentImpl::cancelParsing): Factored out of implicitOpen. (WebCore::DocumentImpl::implicitOpen): Call cancelParsing now.
- khtml/xml/DocumentImpl.h:
- 9:42 PM Changeset in webkit [12406] by
-
- 1 edit4 adds in trunk/LayoutTests
Reviewed by Tim Omernick.
- test case for REGRESSION: crash when setting a list with list-style: none to display: none http://bugzilla.opendarwin.org/show_bug.cgi?id=6860
(This bug was fixed independently by Dave Harrison but I'm landing the test
case anyway.)
- fast/lists/list-style-none-crash-expected.checksum: Added.
- fast/lists/list-style-none-crash-expected.png: Added.
- fast/lists/list-style-none-crash-expected.txt: Added.
- fast/lists/list-style-none-crash.html: Added.
- 5:35 PM Changeset in webkit [12405] by
-
- 1 edit4 adds in trunk/LayoutTests
Test case for <rdar://problem/4256504>
khtml::RenderBlock::skipWhitespace crashes upon interaction with
radio buttons
- 5:12 PM Changeset in webkit [12404] by
-
- 2 edits in trunk/WebCore
Reviewed by Hyatt.
Fix for <rdar://problem/4256504> khtml::RenderBlock::skipWhitespace
crashes upon interaction with radio buttons.
The old code held an assumption that positioned elements would
always have line boxes. This is not necessarily the case, so this
patch makes sure that the lines are dirtied appropriately.
- 3:56 PM Changeset in webkit [12403] by
-
- 1 edit2 adds in trunk
2006-01-26 Mitz Pettel <opendarwin.org@mitzpettel.com>
Test: fast/dom/HTMLImageElement/image-without-renderer-width.html
Reviewed and landed by Anders.
- fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6841 REGRESSION: WebKit (Thursday, January 26, 2006) crash in HTMLImageElementImpl::width()
- khtml/html/html_imageimpl.cpp: (WebCore::HTMLImageElementImpl::width): Return the image's dimensions only if it isn't null, but even if it hasn't finished loading. (WebCore::HTMLImageElementImpl::height): Ditto.
- 3:32 PM Changeset in webkit [12402] by
-
- 2 edits in trunk/WebKit
Reviewed by John Sullivan.
<rdar://problem/4422365>
- WebView.subproj/WebHTMLView.m: (-[NSArray addSuperviewObservers]): In addition to registering for frame/bounds change notifications, call -_frameOrBoundsChanged. It will check the current size/scroll against the previous layout's size/scroll. We need to do this here to catch the case where the WebView is laid out at one size, removed from its window, resized, and inserted into another window. Our frame/bounds changed notifications will not be sent in that situation, since we only watch for changes while in the view hierarchy.
I have verified that this does not cause unnecessary layouts while running the PLT.
- 2:45 PM Changeset in webkit [12401] by
-
- 2 edits in trunk/WebCore
Reviewed by Hyatt and Vicki.
Fix for broken layout tests.
- khtml/html/HTMLInputElementImpl.cpp: (WebCore::HTMLInputElementImpl::focus):
- 2:00 PM Changeset in webkit [12400] by
-
- 10 edits3 adds2 deletes in trunk
2006-01-26 Anders Carlsson <andersca@mac.com>
Reviewed by Eric.
- http://bugzilla.opendarwin.org/show_bug.cgi?id=6845 Autogenerate more classes
This makes DocumentType and DOMImplementation autogenerated.
- JSCore.cpp: Added.
- WebCore.xcodeproj/project.pbxproj: Add JSCore.cpp
- bindings/scripts/CodeGeneratorJS.pm: Add attributes for controlling how null values should be converted to and from DOMString objects.
- bindings/scripts/IDLParser.pm:
- bindings/scripts/IDLStructure.pm: Remove the extended attributes hash on attribute objects and use the one on the signature.
- khtml/ecma/kjs_dom.cpp: (KJS::DOMNodeProtoFunc::callAsFunction): For isSupported, just convert the version argument to an empty string if null is passed in. The reason for this change is that the autogenerated binding for DOMImplementation::hasFeature does the same thing. This also makes us behave like Mozilla and Opera.
(KJS::toDocumentType):
(KJS::getDOMNode):
Use JSDocumentType.
(KJS::getDOMDOMImplementation):
Use JSDOMImplementation.
- khtml/ecma/kjs_dom.h: Remove DOMDOMImplementation and DOMDocumentType
- khtml/xml/DOMImplementation.idl: Added.
- khtml/xml/DocumentType.idl: Added.
- 1:40 PM Changeset in webkit [12399] by
-
- 3 edits4 adds in trunk
2006-01-26 Alexander Kellett <lypanov@kde.org>
Reviewed by eseidel.
<text> elements with a scale on a parent <g> breaks nodeAtPoint
Fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6843
- kcanvas/RenderSVGText.cpp: (RenderSVGText::nodeAtPoint):
- 12:32 PM Changeset in webkit [12398] by
-
- 6 edits in trunk/WebCore
Reviewed by Geoff.
- fixed a couple hundred node leaks on the traversal layout tests http://bugzilla.opendarwin.org/show_bug.cgi?id=6828
There were three basic problems here, any one of which was enough
to cause the leak.
1) If the filter function for a NodeFilter, NodeIterator or TreeWalker
had that object in scope, it would cause a GC reference cycle. I fixed
this by making these JS functions get marked by the corresponding JS object,
as for XMLHttpRequest.
2) The TraversalImpl class that's a base class for NodeIteratorImpl and
TreeWalkerImpl did not have a virtual destructor, leading the Shared
template to call the base class destructor on deref.
3) ref/deref on the document were mismatched, I fixed all that sort of nonsense
by using RefPtr instead of manual ref/deref.
- khtml/dom/dom2_traversal.h: (DOM::NodeFilterCondition::mark): Add no-op virtual mark method, JS subclass will do something useful here.
- khtml/ecma/kjs_traversal.cpp: (KJS::DOMNodeIterator::mark): Mark the DOM-level filter. (KJS::DOMNodeFilter::mark): ditto (KJS::DOMTreeWalker::mark): ditto (KJS::JSNodeFilterCondition::JSNodeFilterCondition): reformatted. (KJS::JSNodeFilterCondition::mark): Mark the filter function.
- khtml/ecma/kjs_traversal.h:
- khtml/xml/dom2_traversalimpl.cpp: (DOM::NodeFilterImpl::NodeFilterImpl): diligently applied RefPtr (DOM::TraversalImpl::TraversalImpl): ditto (DOM::TraversalImpl::~TraversalImpl): ditto (DOM::NodeIteratorImpl::NodeIteratorImpl): ditto (DOM::NodeIteratorImpl::~NodeIteratorImpl): ditto (DOM::NodeIteratorImpl::setReferenceNode): ditto (DOM::TreeWalkerImpl::TreeWalkerImpl): ditto (DOM::TreeWalkerImpl::setCurrentNode): ditto
- khtml/xml/dom2_traversalimpl.h: (DOM::NodeFilterImpl::mark): call mark() on the codition (DOM::TraversalImpl::root): Use .get() (DOM::TraversalImpl::filter): ditto (DOM::NodeIteratorImpl::referenceNode): ditto (DOM::NodeIteratorImpl::document): ditto (DOM::TreeWalkerImpl::currentNode): ditto
- 12:18 PM Changeset in webkit [12397] by
-
- 2 edits in trunk/WebCore
Reviewed by Geoff.
- fixed ~40 DOM nodes leaked when running any two of the outerHTML DOM tests (actually just a false positive) http://bugzilla.opendarwin.org/show_bug.cgi?id=6822
- kwq/KWQPageState.mm: (-[KWQPageState clear]): Garbage collect after deleting a bunch of stuff that holds onto JS objects. This should avoid false positive leaks and is good in any case since page state deallocation is already deferred.
- 10:46 AM Changeset in webkit [12396] by
-
- 2 edits in trunk/WebCore
Reviewed by Vicki.
- fix for <rdar://problem/4422624> REGRESSION: crash loading webmail.mac.com
- khtml/html/HTMLInputElementImpl.cpp: (WebCore::HTMLInputElementImpl::focus): Nil check for renderer.
- 8:59 AM Changeset in webkit [12395] by
-
- 1 edit4 adds in trunk/LayoutTests
Adding test case for:
http://bugzilla.opendarwin.org/show_bug.cgi?id=6835
- editing/deleting/delete-listitem-001-expected.checksum: Added.
- editing/deleting/delete-listitem-001-expected.png: Added.
- editing/deleting/delete-listitem-001-expected.txt: Added.
- editing/deleting/delete-listitem-001.html: Added.
- 8:58 AM Changeset in webkit [12394] by
-
- 2 edits in trunk/WebCore
Reviewed by Timothy.
- fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6835 REGRESSION: WebKit crashes when loading a script on Wikipedia
Test added:
editing/deleting/delete-listitem-001.html
- rendering/render_list.cpp: (RenderListItem::setStyle): (RenderListItem::resetMarkerValue): (RenderListItem::getAbsoluteRepaintRect): (RenderListMarker::paint):
- 4:06 AM Changeset in webkit [12393] by
-
- 3 edits1 add in trunk/WebCore
Fix by Mitz Pettel, reviewed by Darin, landed by ap.
- fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6388 REGRESSION: Incomplete repaint when dragging the map on Google maps
- manual-tests/bugzilla-6388.html: Added.
- rendering/render_layer.cpp: (WebCore::RenderLayer::computeRepaintRects): Cache the object's absolute position before update. (WebCore::RenderLayer::updateLayerPositions): Use the cached position to determine if the object moved. If it did, do a full repaint.
- rendering/render_layer.h:
- 3:32 AM Changeset in webkit [12392] by
-
- 2 edits in trunk/WebCore
2006-01-26 Eric Seidel <eseidel@apple.com>
Adding additional test case from:
http://bugzilla.opendarwin.org/show_bug.cgi?id=6801
- svg/custom/percentage-rect2-expected.checksum: Added.
- svg/custom/percentage-rect2-expected.png: Added.
- svg/custom/percentage-rect2-expected.txt: Added.
- svg/custom/percentage-rect2.svg: Added.
- 3:11 AM Changeset in webkit [12391] by
-
- 1 edit4 adds in trunk/LayoutTests
2006-01-26 Eric Seidel <eseidel@apple.com>
Adding additional test case from:
http://bugzilla.opendarwin.org/show_bug.cgi?id=6801
- svg/custom/percentage-rect2-expected.checksum: Added.
- svg/custom/percentage-rect2-expected.png: Added.
- svg/custom/percentage-rect2-expected.txt: Added.
- svg/custom/percentage-rect2.svg: Added.
- 2:57 AM Changeset in webkit [12390] by
-
- 2 edits1 add in trunk/WebCore
Fix by Mitz Pettel, reviewed by Darin, committed by ap.
- fix http://bugzilla.opendarwin.org/show_bug.cgi?id=6278 REGRESSION: Incomplete repaint when table cell width changes during layout
- manual-tests/bugzilla-6278.html: Added.
- rendering/RenderTableSection.cpp: (WebCore::RenderTableSection::setCellWidths): If during layout a cell changes from not needing layout to needing layout, call repaintObjectsBeforeLayout on it.
- 2:53 AM Changeset in webkit [12389] by
-
- 9 edits4 adds in trunk
2006-01-26 Eric Seidel <eseidel@apple.com>
Reviewed by mjs.
SVG shows up blank in WebKit+SVG
http://bugzilla.opendarwin.org/show_bug.cgi?id=6620
- css/svg.css: default to width/height 100% for <svg>
- 1:53 AM Changeset in webkit [12388] by
-
- 1 edit in trunk/WebCore/khtml/xml/WheelEvent.idl
Force this to be re-generated
- 1:43 AM Changeset in webkit [12387] by
-
- 9 edits1 add in trunk/WebCore
2006-01-26 Anders Carlsson <andersca@mac.com>
Reviewed by Darin.
http://bugzilla.opendarwin.org/show_bug.cgi?id=6805
Support constants in IDL files
- bindings/scripts/CodeGeneratorJS.pm: If an interface has constants, generate a constructor object and add the constants as properties. Also add a getConstructor method to the interface object.
- bindings/scripts/CodeGenerator.pm:
- bindings/scripts/generate-bindings.pl: Add a --force-generation flag to force regeneration of files even though nothing has changed.
- khtml/ecma/kjs_events.cpp:
- khtml/ecma/kjs_events.h: Remove MutationEventConstructor and DOMMutationEvent, those are autogenerated now.
- bindings/js/JSEvents.cpp: Include JSMutationEvent.cpp
- khtml/ecma/kjs_window.cpp: (KJS::Window::getValueProperty):
- khtml/ecma/kjs_window.h: (KJS::Window::): Add MutationEvent property.
- khtml/xml/MutationEvent.idl: Added.
- 1:32 AM Changeset in webkit [12386] by
-
- 34 edits4 adds in trunk
2006-01-25 Alexander Kellett <lypanov@kde.org>
Reviewed by eseidel.
Percentage values for width, height are not rendered correctly.
http://bugzilla.opendarwin.org/show_bug.cgi?id=5331
Create items for even empty paths to prevent an empty render tree.
Calculate the viewport element on the fly rather than caching it,
unless there is no context from which to calculate it. Call
notifyAttributeChange() on layout to fix usage of viewport
percentages for lengths. Remove portion of code which appears to
do nothing but worsen the situation.
- kcanvas/device/quartz/KCanvasItemQuartz.h:
- kcanvas/device/quartz/KCanvasItemQuartz.mm: (KCanvasItemQuartz::layout):
- ksvg2/svg/SVGLengthImpl.cpp: (SVGLengthImpl::value):
- ksvg2/svg/SVGStyledElementImpl.cpp: (SVGStyledElementImpl::createRenderer):
- 1:23 AM Changeset in webkit [12385] by
-
- 2 edits in trunk/WebCore
2006-01-26 Eric Seidel <eseidel@apple.com>
Reviewed by hyatt.
Disable <animateTransform> to stop layout test crashes.
http://bugzilla.opendarwin.org/show_bug.cgi?id=6780
- ksvg2/misc/KSVGTimeScheduler.cpp: hack out animateTransform.
- 12:37 AM Changeset in webkit [12384] by
-
- 16 edits in trunk/WebCore
2006-01-26 Eric Seidel <eseidel@apple.com>
Reviewed by mjs.
Deploy RefPtr throughout more of WebCore
http://bugzilla.opendarwin.org/show_bug.cgi?id=6754
- bridge/mac/MacFrame.mm: (MacFrame::shouldClose): no need to call .get()
- khtml/html/HTMLElementImpl.cpp: (WebCore::HTMLElementImpl::createContextualFragment):
- khtml/html/HTMLSelectElementImpl.cpp: (WebCore::HTMLSelectElementImpl::remove):
- khtml/html/HTMLTextAreaElementImpl.cpp: (WebCore::HTMLTextAreaElementImpl::setDefaultValue):
- khtml/html/html_baseimpl.cpp: (WebCore::HTMLBodyElementImpl::HTMLBodyElementImpl): (WebCore::HTMLBodyElementImpl::~HTMLBodyElementImpl): (WebCore::HTMLBodyElementImpl::createLinkDecl): (WebCore::HTMLBodyElementImpl::parseMappedAttribute):
- khtml/html/html_baseimpl.h:
- khtml/html/html_headimpl.cpp: (WebCore::HTMLLinkElementImpl::HTMLLinkElementImpl): (WebCore::HTMLLinkElementImpl::~HTMLLinkElementImpl): (WebCore::HTMLLinkElementImpl::process): (WebCore::HTMLLinkElementImpl::setStyleSheet): (WebCore::HTMLLinkElementImpl::isLoading): (WebCore::HTMLStyleElementImpl::HTMLStyleElementImpl): (WebCore::HTMLStyleElementImpl::childrenChanged): (WebCore::HTMLStyleElementImpl::isLoading):
- khtml/html/html_headimpl.h: (DOM::HTMLLinkElementImpl::sheet): (DOM::HTMLStyleElementImpl::sheet):
- khtml/html/html_tableimpl.cpp: (WebCore::HTMLTableElementImpl::parseMappedAttribute): (WebCore::HTMLTableSectionElementImpl::insertRow): (WebCore::HTMLTableSectionElementImpl::deleteRow): (WebCore::HTMLTableRowElementImpl::deleteCell):
- khtml/xml/NodeImpl.cpp: (WebCore::NodeImpl::dispatchEvent): (WebCore::NodeImpl::dispatchGenericEvent): (WebCore::NodeImpl::dispatchWindowEvent): (WebCore::NodeImpl::dispatchMouseEvent): (WebCore::NodeImpl::dispatchKeyEvent): (WebCore::NodeImpl::createRendererIfNeeded):
- khtml/xml/NodeImpl.h:
- khtml/xml/dom_elementimpl.cpp: (WebCore::MappedAttributeImpl::clone): (WebCore::StyledElementImpl::StyledElementImpl): (WebCore::StyledElementImpl::createInlineStyleDecl): (WebCore::StyledElementImpl::destroyInlineStyleDecl): (WebCore::StyledElementImpl::getInlineStyleDecl):
- khtml/xml/dom_elementimpl.h: (DOM::MappedAttributeImpl::MappedAttributeImpl): (DOM::MappedAttributeImpl::style): (DOM::MappedAttributeImpl::decl): (DOM::MappedAttributeImpl::setDecl): (DOM::StyledElementImpl::inlineStyleDecl):
- khtml/xml/dom_position.cpp: (DOM::Position::Position): (DOM::Position::clear):
- khtml/xml/dom_position.h: (DOM::Position::node):