Timeline



May 10, 2004:

4:47 PM Changeset in webkit [6574] by mjs
  • 8 edits in trunk

WebCore:

Reviewed by Darin.

  • avoid redecoding animated images that are only used once for ~2.5% iBench speedup (WebCore part of fix)
  • khtml/rendering/render_image.cpp: (RenderImage::setPixmap): Maintain use count for pixmap. (RenderImage::~RenderImage): Likewise.
  • kwq/KWQPixmap.h:
  • kwq/KWQPixmap.mm: (QPixmap::increaseUseCount): Call WebCoreImageRenderer (QPixmap::decreaseUseCount): Call WebCoreImageRenderer
  • kwq/WebCoreImageRenderer.h: Prototype new protocol methods.

WebKit:

Reviewed by Darin.

  • avoid redecoding animated images that are only used once for ~2.5% iBench speedup (WebCore part of fix)
  • WebCoreSupport.subproj/WebImageRenderer.h:
  • WebCoreSupport.subproj/WebImageRenderer.m: (-[WebImageRenderer increaseUseCount]): (-[WebImageRenderer decreaseUseCount]): (-[WebImageRenderer retainOrCopyIfNeeded]):
4:42 PM Changeset in webkit [6573] by hyatt
  • 1 edit in trunk/WebCore/khtml/xml/dom_docimpl.cpp

Back out the optimization since it will break the ibench cheat.

4:34 PM Changeset in webkit [6572] by mjs
  • 2 edits in trunk/WebKit

Fix build.

  • WebView.subproj/WebFrame.m:
4:28 PM Changeset in webkit [6571] by hyatt
  • 1 edit in trunk/WebCore/khtml/xml/dom_docimpl.cpp

Fix build failure

4:21 PM Changeset in webkit [6570] by mjs
  • 3 edits in trunk/WebKit

Reviewed by Ken.

  • avoid messing with undo manager needlessly for ~1% HTML iBench speedup
  • WebCoreSupport.subproj/WebBridge.h:
  • WebCoreSupport.subproj/WebBridge.m: (-[WebBridge registerCommandForUndo:]): mark undo/redo item flag (-[WebBridge registerCommandForRedo:]): ditto (-[WebBridge clearUndoRedoOperations]): check flag before removing items, and clear it after removing them
4:20 PM Changeset in webkit [6569] by mjs
  • 5 edits in trunk/WebKit

Reviewed by Darin.

  • -[WebFrame childFrames] is so hot that a special internal version which avoids the copy and autorelease results in a .75% performance improvement on HTML iBench.
  • WebView.subproj/WebFramePrivate.h: Prototype new method.
  • WebView.subproj/WebFrame.m: (-[WebFrame _internalChildFrames]): New method, just returns internal value instead of copying.

(-[WebFrame _descendantFrameNamed:]): Use it
(-[WebFrame _textSizeMultiplierChanged]): likewise
(-[WebFrame _viewWillMoveToHostWindow:]): likewise
(-[WebFrame _viewDidMoveToHostWindow]): likewise
(-[WebFrame _saveDocumentAndScrollState]): likewise
(-[WebFrame _numPendingOrLoadingRequests:]): likewise
(-[WebFrame _checkLoadComplete]): Refactored this and it's two
helpers a little so we could get away with using
_internalChildFrames.
(-[WebFrame _checkLoadCompleteForThisFrame]): Renamed from
_isLoadComplete
(-[WebFrame _recursiveCheckLoadComplete]): renamed from (class
method) _recursiveCheckCompleteFromFrame:

  • WebView.subproj/WebDataSource.m: (-[WebDataSource _defersCallbacksChanged]): Use it (-[WebDataSource isLoading]): likewise
  • WebView.subproj/WebView.m: (-[WebView _frameForDataSource:fromFrame:]): likewise (-[WebView _frameForView:fromFrame:]): likewise
4:19 PM Changeset in webkit [6568] by hyatt
  • 3 edits in trunk/WebCore

Make sure that when finished parsing if all the right conditions are met, we just go ahead and updateRendering and
layout.

Reviewed by rjw

  • khtml/khtmlview.cpp: (KHTMLView::scheduleRelayout):
2:47 PM Changeset in webkit [6567] by mjs
  • 5 edits in trunk/WebCore

Reviewed by Dave.

  • some minor performance tweaks that together give .5% or so on the HTML iBench
  • khtml/css/css_valueimpl.h: don't use a bitfield for id and important, unpacking the bitfield is too expensive
  • kwq/KWQArrayImpl.h:
  • kwq/KWQArrayImpl.mm: (KWQArrayImpl::at): inline this method because it's so hot
  • kwq/KWQString.mm: (allocateHandle): inline this function because it's so hot (initializeHandleNodes): move this in the file so above can be inlined (allocateNode): ditto
2:29 PM Changeset in webkit [6566] by hyatt
  • 11 edits in trunk/WebCore

Move image loading to content.

Reviewed by kocienda

  • khtml/html/html_imageimpl.cpp: (m_imageComplete): (HTMLImageLoader::~HTMLImageLoader): (HTMLImageLoader::updateFromElement): (HTMLImageLoader::removedFromDocument): (HTMLImageLoader::dispatchLoadEvent): (HTMLImageLoader::notifyFinished): (HTMLImageElementImpl::HTMLImageElementImpl): (HTMLImageElementImpl::parseHTMLAttribute): (HTMLImageElementImpl::attach): (HTMLImageElementImpl::detach): (HTMLImageElementImpl::removedFromDocument): (HTMLImageElementImpl::currentImage):
  • khtml/html/html_imageimpl.h: (DOM::):
  • khtml/html/html_objectimpl.cpp: (HTMLObjectElementImpl::HTMLObjectElementImpl): (HTMLObjectElementImpl::removedFromDocument): (HTMLObjectElementImpl::attach):
  • khtml/html/html_objectimpl.h:
  • khtml/misc/loader.cpp: (CachedImage::ref):
  • khtml/rendering/render_image.cpp: (RenderImage::RenderImage): (RenderImage::setImage): (RenderImage::updateAltText):
  • khtml/rendering/render_image.h: (khtml::RenderImage::getImage):
  • khtml/rendering/render_object.cpp: (RenderObject::canvas):
  • khtml/xml/dom_docimpl.cpp: (DocumentImpl::dispatchImageLoadEventSoon): (DocumentImpl::removeImage): (DocumentImpl::dispatchImageLoadEventsNow):
  • khtml/xml/dom_docimpl.h:
11:05 AM Changeset in webkit [6565] by cblu
  • 2 edits in trunk/WebKit

Forgot to commit this copied header.

  • DOM.subproj/DOMExtensions.h:
9:47 AM Changeset in webkit [6564]
  • 4 copies
    2 deletes in tags/WebCore-139~1

This commit was manufactured by cvs2svn to create tag
'WebCore-139~1'.

9:47 AM Changeset in webkit [6563] by darin
  • 2 edits in trunk/WebCore

Reviewed by Vicki.

  • fixed build failure seen in B&I log (due to case sensitive file system?)
  • kwq/KWQLoader.h: Include "KWQCString.h" rather than "QCString.h". The latter was failing because the forwarding header is called "qcstring.h" and I guess B&I started building with a case sensitive file system. I could have changed this to "qcstring.h", but it seemed better to use the header's real name since we're inside KWQ here.

May 8, 2004:

9:55 PM Changeset in webkit [6562] by mjs
  • 3 edits in trunk/WebCore

Reviewed by Dave.

  • change things around so extremely hot TokenizerString::current() method doesn't have to do any work. It's cheaper to maintain the relevant state than to always compute it.
  • khtml/misc/stringit.h: (khtml::TokenizerString::TokenizerString): Initialize m_currentChar (khtml::TokenizerString::push): update m_currentChar (khtml::TokenizerString::advance): ditto (khtml::TokenizerString::current): return m_currentChar
  • khtml/misc/stringit.cpp: (khtml::TokenizerString::clear): update m_currentChar (khtml::TokenizerString::append): update m_currentChar (khtml::TokenizerString::prepend): update m_currentChar

May 7, 2004:

7:38 PM Changeset in webkit [6561] by gramps
  • 2 edits in trunk/WebCore

Reviewed by Dave.

Fixed minor, edge-case regression in KEYGEN RSA logic caused by Dave's
"atomic string" patch back in January.

  • khtml/html/html_formimpl.cpp: (HTMLKeygenElementImpl::encoding):
3:31 PM Changeset in webkit [6560]
  • 2 copies
    3 deletes in tags/JavaScriptCore-139~1

This commit was manufactured by cvs2svn to create tag
'JavaScriptCore-139~1'.

3:31 PM Changeset in webkit [6559] by vicki
  • 2 edits in trunk/JavaScriptCore

Reviewed by darin.

Turn off GC since it uses ppc only instructions (which breaks
the B&I build).

  • kjs/value.h: set USE_CONSERVATIVE_GC to 0
11:10 AM Changeset in webkit [6558] by vicki
  • 3 edits in trunk

set version number to 140u. the tree is open!

11:03 AM Changeset in webkit [6557]
  • 3 copies in tags/Safari-139

This commit was manufactured by cvs2svn to create tag 'Safari-139'.

11:03 AM Changeset in webkit [6556] by vicki
  • 6 edits in trunk

Safari-139 stamp

9:51 AM Changeset in webkit [6555] by mjs
  • 4 edits in trunk/WebCore

Reviewed by Darin.

  • two trivial changes for a 1% speedup, thanks to Shark
  • khtml/dom/dom_string.cpp:
  • khtml/dom/dom_string.h: (DOM::DOMString::~DOMString): inline this method because it's hot and the function call overhead was very costly.
  • WebCore.pbproj/project.pbxproj: Add -falign-loops=16 optimization flag, as suggested by Shark.
9:50 AM Changeset in webkit [6554] by mjs
  • 3 edits in trunk/JavaScriptCore

Reviewed by Darin.

  • add -funroll-loops=16 compiler option for approx .5% speedup on HTML iBench and .5-1% speedup on JS iBench.

May 6, 2004:

3:50 PM Changeset in webkit [6553] by cblu
  • 3 edits in trunk/WebCore

Added alpha to DOMRGBColor as an extension.

Reviewed by dave.

  • kwq/DOM-CSS.mm: (-[DOMRGBColor alpha]): new
  • kwq/DOMExtensions.h:
3:04 PM Changeset in webkit [6552] by hyatt
  • 2 adds in trunk/LayoutTests/fast/text/whitespace

Add layout test for pres and brs.

3:01 PM Changeset in webkit [6551] by hyatt
  • 2 edits in trunk/WebCore

Fix whitespace rendering for brs inside pres. The bug is 3640711.

Reviewed by kocienda

  • khtml/rendering/bidi.cpp: (khtml::RenderBlock::layoutInlineChildren): (khtml::RenderBlock::findNextLineBreak):
2:41 PM Changeset in webkit [6550] by cblu
  • 2 edits in trunk/WebKit
  • DOM.subproj/WebDOMOperations.h: improved a header doc comment
Note: See TracTimeline for information about the timeline view.