[webkit-reviews] review granted: [Bug 80629] [chromium] Canvas acceleration is not restored after context loss : [Attachment 130925] Patch

[bugzilla-daemon at webkit.org]

James Robinson <jamesr at chromium.org> has granted Stephen White
<senorblanco at chromium.org>'s request for review:
Bug 80629: [chromium] Canvas acceleration is not restored after context loss
https://bugs.webkit.org/show_bug.cgi?id=80629

Attachment 130925: Patch
https://bugs.webkit.org/attachment.cgi?id=130925&action=review

------- Additional Comments from James Robinson <jamesr at chromium.org>
View in context: https://bugs.webkit.org/attachment.cgi?id=130925&action=review


R=me but looking at the code I'm pretty sure you gotta flip the makeCurrent and
extension query around or it'll crash. Try inducing a failure in
MaybeInitializeGL() to test that theory out

> Source/WebCore/platform/graphics/gpu/SharedGraphicsContext3D.cpp:40
> +	   if (m_context &&
((m_context->getExtensions()->getGraphicsResetStatusARB() !=
GraphicsContext3D::NO_ERROR) || !m_context->makeContextCurrent()))

nit: don't we have to make the context current before querying the extensions?
i think these have to be flipped around. obviously if we can't make the context
current we shouldn't try to look at its extensions bits either.