[Webkit-unassigned] [Bug 117733] [WK2][GTK] ASSERTION in WebKit::LayerTreeHostGtk::invalidate

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=117733





--- Comment #17 from Alberto Garcia <agarcia at igalia.com>  2013-07-09 02:45:19 PST ---
(In reply to comment #16)
> The bits in WebKitWebView look good to me, but I think any fixes for
> LayerTreeHost should be split out into another patch. I think given
> your comments earlier that the approach of returning null isn't a
> safe one (as the assertion you added attests). I think we are just
> going to have to make sure that we do not return an invalid
> LayerTreeHost.

The situation here is that, with the changes in WebKitWebView, the
scenario where LayerTreeHost fails to get the GL context is not
supposed to happen, ever.

So that change alone would be enough to fix the crash.

The other part of the patch just tries to fix the inconsistencies in
LayerTreeHost.

If we still want to return a non-null LayerTreeHost we can do
something like what I proposed in my first patch (comment 4). The code
already deals with the scenario where glContext() returns null, it was
only the assertion that was causing problems (that's why it only
crashed in debug builds).

Still I think it's a good idea to detect when we're incorrectly
attempting to enter AC mode, because that's not supposed to happen.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.