[webkit-changes] [126680] trunk/LayoutTests
Title: [126680] trunk/LayoutTests Revision 126680 Author zandober...@gmail.com Date 2012-08-24 22:33:44 -0700 (Fri, 24 Aug 2012) Log Message Unreviewed GTK gardening. Skipping a test added in r126611 that requires CSS image-set support. That feature is not currently enabled on the GTK port. * platform/gtk/TestExpectations: Modified Paths trunk/LayoutTests/ChangeLog trunk/LayoutTests/platform/gtk/TestExpectations Diff Modified: trunk/LayoutTests/ChangeLog (126679 => 126680) --- trunk/LayoutTests/ChangeLog 2012-08-25 04:55:10 UTC (rev 126679) +++ trunk/LayoutTests/ChangeLog 2012-08-25 05:33:44 UTC (rev 126680) @@ -1,3 +1,12 @@ +2012-08-24 Zan Dobersek + +Unreviewed GTK gardening. + +Skipping a test added in r126611 that requires CSS image-set support. +That feature is not currently enabled on the GTK port. + +* platform/gtk/TestExpectations: + 2012-08-24 Florin Malita ASSERTION FAILED: !attached() in WebCore::Node::attach() Modified: trunk/LayoutTests/platform/gtk/TestExpectations (126679 => 126680) --- trunk/LayoutTests/platform/gtk/TestExpectations 2012-08-25 04:55:10 UTC (rev 126679) +++ trunk/LayoutTests/platform/gtk/TestExpectations 2012-08-25 05:33:44 UTC (rev 126680) @@ -302,6 +302,7 @@ // CSS image-set support not yet enabled (needs ENABLE_CSS_IMAGE_SET). BUGWKGTK SKIP : fast/css/image-set-parsing.html = TEXT BUGWKGTK SKIP : fast/css/image-set-parsing-invalid.html = TEXT +BUGWKGTK SKIP : fast/css/image-set-setting.html = TEXT // For now, Web Audio API is disabled BUGWKGTK SKIP : webaudio = TEXT ___ webkit-changes mailing list webkit-changes@lists.webkit.org http://lists.webkit.org/mailman/listinfo/webkit-changes
[webkit-changes] [126679] trunk/Source/WebKit2
Title: [126679] trunk/Source/WebKit2 Revision 126679 Author a...@apple.com Date 2012-08-24 21:55:10 -0700 (Fri, 24 Aug 2012) Log Message REGRESSION: WebProcessProxy destructor is sometimes called recursively https://bugs.webkit.org/show_bug.cgi?id=94997 Reviewed by Dan Bernstein. * UIProcess/WebContext.cpp: (WebKit::WebContext::disconnectProcess): Make sure that we don't try to delete WebProcessProxy while the vector still contains it. Previously, we used RefPtr::clear, which zeroes out the value before destructing, but doing that in every RefPtr destructor would be bad for performance. Modified Paths trunk/Source/WebKit2/ChangeLog trunk/Source/WebKit2/UIProcess/WebContext.cpp Diff Modified: trunk/Source/WebKit2/ChangeLog (126678 => 126679) --- trunk/Source/WebKit2/ChangeLog 2012-08-25 03:42:51 UTC (rev 126678) +++ trunk/Source/WebKit2/ChangeLog 2012-08-25 04:55:10 UTC (rev 126679) @@ -1,3 +1,15 @@ +2012-08-24 Alexey Proskuryakov + + REGRESSION: WebProcessProxy destructor is sometimes called recursively +https://bugs.webkit.org/show_bug.cgi?id=94997 + +Reviewed by Dan Bernstein. + +* UIProcess/WebContext.cpp: (WebKit::WebContext::disconnectProcess): Make sure that +we don't try to delete WebProcessProxy while the vector still contains it. Previously, +we used RefPtr::clear, which zeroes out the value before destructing, but doing that +in every RefPtr destructor would be bad for performance. + 2012-08-24 Benjamin Poulain Unify Number to StringImpl conversion Modified: trunk/Source/WebKit2/UIProcess/WebContext.cpp (126678 => 126679) --- trunk/Source/WebKit2/UIProcess/WebContext.cpp 2012-08-25 03:42:51 UTC (rev 126678) +++ trunk/Source/WebKit2/UIProcess/WebContext.cpp 2012-08-25 04:55:10 UTC (rev 126679) @@ -457,7 +457,10 @@ m_pluginSiteDataManager->invalidate(); #endif -// This can cause the web context to be destroyed. +// The vector may have the last reference to process proxy, which in turn may have the last reference to the context. +// Since vector elements are destroyed in place, we would recurse into WebProcessProxy destructor +// if it were invoked from Vector::remove(). RefPtr delays destruction until it's safe. +RefPtr protect(process); m_processes.remove(m_processes.find(process)); } ___ webkit-changes mailing list webkit-changes@lists.webkit.org http://lists.webkit.org/mailman/listinfo/webkit-changes
[webkit-changes] [126678] trunk/Source/WebKit/mac
Title: [126678] trunk/Source/WebKit/mac Revision 126678 Author psola...@apple.com Date 2012-08-24 20:42:51 -0700 (Fri, 24 Aug 2012) Log Message Don't create a new ResourceRequest if delegate returns the same NSURLRequest we passed it https://bugs.webkit.org/show_bug.cgi?id=94962 Reviewed by Geoffrey Garen. If the NSURLRequest returned from the delegate callback is the same as the one we passed to it, then avoid creating a new ResourceRequest object. This reduces the number of calls to ResourceRequest::doUpdateResourceRequest(). * WebCoreSupport/WebFrameLoaderClient.mm: (WebFrameLoaderClient::dispatchWillSendRequest): Modified Paths trunk/Source/WebKit/mac/ChangeLog trunk/Source/WebKit/mac/WebCoreSupport/WebFrameLoaderClient.mm Diff Modified: trunk/Source/WebKit/mac/ChangeLog (126677 => 126678) --- trunk/Source/WebKit/mac/ChangeLog 2012-08-25 02:35:00 UTC (rev 126677) +++ trunk/Source/WebKit/mac/ChangeLog 2012-08-25 03:42:51 UTC (rev 126678) @@ -1,3 +1,17 @@ +2012-08-24 Pratik Solanki + +Don't create a new ResourceRequest if delegate returns the same NSURLRequest we passed it +https://bugs.webkit.org/show_bug.cgi?id=94962 + +Reviewed by Geoffrey Garen. + +If the NSURLRequest returned from the delegate callback is the same as the one we passed to +it, then avoid creating a new ResourceRequest object. This reduces the number of calls to +ResourceRequest::doUpdateResourceRequest(). + +* WebCoreSupport/WebFrameLoaderClient.mm: +(WebFrameLoaderClient::dispatchWillSendRequest): + 2012-08-23 Mark Hahnenberg Change behavior of MasqueradesAsUndefined to better accommodate DFG changes Modified: trunk/Source/WebKit/mac/WebCoreSupport/WebFrameLoaderClient.mm (126677 => 126678) --- trunk/Source/WebKit/mac/WebCoreSupport/WebFrameLoaderClient.mm 2012-08-25 02:35:00 UTC (rev 126677) +++ trunk/Source/WebKit/mac/WebCoreSupport/WebFrameLoaderClient.mm 2012-08-25 03:42:51 UTC (rev 126678) @@ -346,8 +346,13 @@ if (redirectResponse.isNull()) static_cast(loader)->increaseLoadCount(identifier); +NSURLRequest *currentURLRequest = request.nsURLRequest(); +NSURLRequest *newURLRequest = currentURLRequest; if (implementations->willSendRequestFunc) -request = (NSURLRequest *)CallResourceLoadDelegate(implementations->willSendRequestFunc, webView, @selector(webView:resource:willSendRequest:redirectResponse:fromDataSource:), [webView _objectForIdentifier:identifier], request.nsURLRequest(), redirectResponse.nsURLResponse(), dataSource(loader)); +newURLRequest = (NSURLRequest *)CallResourceLoadDelegate(implementations->willSendRequestFunc, webView, @selector(webView:resource:willSendRequest:redirectResponse:fromDataSource:), [webView _objectForIdentifier:identifier], currentURLRequest, redirectResponse.nsURLResponse(), dataSource(loader)); + +if (newURLRequest != currentURLRequest) +request = newURLRequest; } bool WebFrameLoaderClient::shouldUseCredentialStorage(DocumentLoader* loader, unsigned long identifier) ___ webkit-changes mailing list webkit-changes@lists.webkit.org http://lists.webkit.org/mailman/listinfo/webkit-changes
[webkit-changes] [126677] branches/chromium/1229
Title: [126677] branches/chromium/1229 Revision 126677 Author cev...@google.com Date 2012-08-24 19:35:00 -0700 (Fri, 24 Aug 2012) Log Message Merge 126131 BUG=143551 Review URL: https://chromiumcodereview.appspot.com/10875067 Modified Paths branches/chromium/1229/Source/WebCore/dom/ContainerNodeAlgorithms.h branches/chromium/1229/Source/WebCore/dom/Node.h branches/chromium/1229/Source/WebCore/html/HTMLFrameElementBase.cpp branches/chromium/1229/Source/WebCore/html/HTMLFrameElementBase.h Added Paths branches/chromium/1229/LayoutTests/fast/frames/iframe-onload-and-domnodeinserted-expected.txt branches/chromium/1229/LayoutTests/fast/frames/iframe-onload-and-domnodeinserted.html Diff Copied: branches/chromium/1229/LayoutTests/fast/frames/iframe-onload-and-domnodeinserted-expected.txt (from rev 126131, trunk/LayoutTests/fast/frames/iframe-onload-and-domnodeinserted-expected.txt) (0 => 126677) --- branches/chromium/1229/LayoutTests/fast/frames/iframe-onload-and-domnodeinserted-expected.txt (rev 0) +++ branches/chromium/1229/LayoutTests/fast/frames/iframe-onload-and-domnodeinserted-expected.txt 2012-08-25 02:35:00 UTC (rev 126677) @@ -0,0 +1,11 @@ +This test ensures that any tree mutation in the load event handler cannot harm the tree consistency. + +On success, you will see a series of "PASS" messages, followed by "TEST COMPLETE". + + +PASS loadEventFired is true +PASS unless crash. +PASS successfullyParsed is true + +TEST COMPLETE + Copied: branches/chromium/1229/LayoutTests/fast/frames/iframe-onload-and-domnodeinserted.html (from rev 126131, trunk/LayoutTests/fast/frames/iframe-onload-and-domnodeinserted.html) (0 => 126677) --- branches/chromium/1229/LayoutTests/fast/frames/iframe-onload-and-domnodeinserted.html (rev 0) +++ branches/chromium/1229/LayoutTests/fast/frames/iframe-onload-and-domnodeinserted.html 2012-08-25 02:35:00 UTC (rev 126677) @@ -0,0 +1,26 @@ + + + + + + +