New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 800449 link

Starred by 1 user

Issue metadata

Status: WontFix
Owner:
Out until 24 Jan
Closed: Jan 2018
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Linux
Pri: 1
Type: Bug



Sign in to add a comment

Multiple webkit_layout_tests failing DCHECK in SVGElementProxy.cpp on linux_chromium_rel_ng

Project Member Reported by jbudorick@chromium.org, Jan 9 2018

Issue description

e.g.

16:50:23.650 22891   [1:1:0108/165023.250140:FATAL:SVGElementProxy.cpp(45)] Check failed: Lifecycle().GetState() <= DocumentLifecycle::kCompositingClean || Lifecycle().GetState() >= DocumentLifecycle::kPaintClean. 
16:50:23.650 22891   #0 0x00000312decc base::debug::StackTrace::StackTrace()
16:50:23.650 22891   #1 0x00000314c71c logging::LogMessage::~LogMessage()
16:50:23.650 22891   #2 0x000005b560b2 blink::SVGElementProxy::IdObserver::ContentChanged()
16:50:23.650 22891   #3 0x000005b562e2 blink::SVGElementProxySet::NotifyContentChanged()
16:50:23.650 22891   #4 0x0000058fd835 blink::LayoutSVGResourceContainer::NotifyContentChanged()
16:50:23.650 22891   #5 0x0000058fe473 blink::LayoutSVGResourceContainer::MarkAllClientsForInvalidation()
16:50:23.650 22891   #6 0x0000058fe7f1 blink::LayoutSVGResourceContainer::MarkForLayoutAndParentResourceInvalidation()
16:50:23.650 22891   #7 0x000005912571 blink::SVGResourcesCache::ClientWillBeRemovedFromTree()
16:50:23.650 22891   #8 0x0000058fc6d5 blink::LayoutSVGContainer::RemoveChild()

from https://chromium-swarm.appspot.com/task?id=3af129c7ee461710

Appears to have started around 3 PM PST on 2018-01-08.
 

Comment 1 by clamy@chromium.org, Jan 10 2018

Components: Internals>Sandbox>SiteIsolation
Owner: nasko@chromium.org
Status: Assigned (was: Untriaged)
From the log, it appears it's the same test that is failing (virtual/gpu/fast/canvas/canvas-filter-origin-clean.html) when SiteIsolation is enabled. I've ran 2000 iterations of the test locally without being able to reproduce. I'm assigning nasko@ from SiteIsolation since that happens when SiteIsolation is turned on, but I think this should probably be closed.
I had been seeing in webkit_layout_tests rather than site_per_process_webkit_layout_tests yesterday. I don't see examples of that this morning, though. Closing it out SGTM.

(site_per_process_webkit_layout_tests still appears to have other issues.)
Status: WontFix (was: Assigned)
Isn't still happening. Closing.

Sign in to add a comment