CRASH at WebCore::RenderView::updateVisibleViewportRect
<rdar://problem/23997530>
Created attachment 273382 [details] Patch
Comment on attachment 273382 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=273382&action=review > Source/WebCore/html/HTMLMediaElement.cpp:796 > + if (renderer() && renderer()->isRegisteredForVisibleInViewportCallback()) > renderer()->unregisterForVisibleInViewportCallback(); Would it be bad to just unconditionally call unregisterForVisibleInViewportCallback()? > LayoutTests/media/video-crash-invisible-autoplay-display-none.html:11 > + run('internals.setMediaElementRestrictions(video, "InvisibleAutoplayNotPermitted")'); This run() thing is weird. Does it just log and eval?
(In reply to comment #3) > Comment on attachment 273382 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=273382&action=review > > > Source/WebCore/html/HTMLMediaElement.cpp:796 > > + if (renderer() && renderer()->isRegisteredForVisibleInViewportCallback()) > > renderer()->unregisterForVisibleInViewportCallback(); > > Would it be bad to just unconditionally call > unregisterForVisibleInViewportCallback()? Not really, as unregister...() checks whether it's actually registered itself as well. > > LayoutTests/media/video-crash-invisible-autoplay-display-none.html:11 > > + run('internals.setMediaElementRestrictions(video, "InvisibleAutoplayNotPermitted")'); > > This run() thing is weird. Does it just log and eval? Yep. I like that it shows what's going on in the results.
Committed r197989: <http://trac.webkit.org/changeset/197989>
*** Bug 155344 has been marked as a duplicate of this bug. ***