[Webkit-unassigned] [Bug 138093] [GTK] Move RedirectedXCompositeWindow from platform to WebKit2 layer

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=138093

--- Comment #10 from Carlos Garcia Campos <cgarcia at igalia.com> ---
(In reply to comment #8)
> Comment on attachment 240542 [details]
> Cleanup diffs to make it easier to review
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=240542&action=review
> 
> I like the reorganization overall and think it's an improvement, but there
> are some issues I have with the new design.
> 
> > Source/WebKit2/UIProcess/gtk/RedirectedXCompositeWindow.cpp:143
> > -    Display* display = GDK_DISPLAY_XDISPLAY(gdk_display_get_default());
> > +    ASSERT(supportsXDamageAndXComposite(display));
> 
> I think it's a bug to put this into an ASSERT. This means that this check is
> only run for debug builds and will crash when run on an XServer that does
> not support these extensions. Instead we should fail gracefully in this case.
> 
> > Source/WebKit2/UIProcess/gtk/RedirectedXCompositeWindow.cpp:223
> >      if (!m_needsNewPixmapAfterResize && m_surface)
> > -        return m_surface.get();
> > +        return;
> 
> One issue here is that you can no longer detect a failure to create the
> surface after a resize. I think that the old design handled this case better
> and this can now hide subtle failures.

Note that current code, doesn't event check the returned value is not NULL before using it :-)

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20141030/1938879e/attachment-0002.html>