Bug 115135 – WebProcess is crashing on http://achicu.github.io/css-presentation when direct pattern compositing is enabled

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[WebKit Commit Bot]

[Qt][WK2] WebProcess is crashing on http://achicu.github.io/css-presentation
Requested by rafaelbrandao on #webkit.

[Sergio Correia (qrwteyrutiyoup)]

This regression seems to be caused by https://bugs.webkit.org/show_bug.cgi?id=109588, or at least git bisect thinks so :)

[Sergio Correia (qrwteyrutiyoup)]

Created attachment 199685 [details]
GDB Stack trace

[Noam Rosenthal]

Created attachment 201285 [details]
Patch

[Noam Rosenthal]

Created attachment 201287 [details]
Patch

[Benjamin Poulain]

This is only common code. Why is this marked [Qt][EFL]?

[Noam Rosenthal]

(In reply to comment #5)
> This is only common code. Why is this marked [Qt][EFL]?
I think when the bug was reported it wasn't clear that the problem was in common code, since that common code is currently activated only in coordinated graphics.

[Noam Rosenthal]

Created attachment 201294 [details]
Patch

[Antti Koivisto]

Comment on attachment 201294 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=201294&action=review

> Source/WebCore/rendering/RenderLayerBacking.cpp:1477
> +    CachedImage* cachedImage = style->backgroundLayers()->image()->cachedImage();
> +    ASSERT(cachedImage);
> +
> +    RefPtr<Image> image = cachedImage->image();

This assertion seems unnecessary. Calling without null checking already communicates that the pointer is expected not to be null and we are going to get a clean crash anyway if it not.

[Noam Rosenthal]

Created attachment 202825 [details]
Patch for landing

[Noam Rosenthal]

Created attachment 202826 [details]
Patch for landing

[WebKit Commit Bot]

Comment on attachment 202826 [details]
Patch for landing

Clearing flags on attachment: 202826

Committed r150643: <http://trac.webkit.org/changeset/150643>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.