[Webkit-unassigned] [Bug 137793] Calling glReadPixels with BGRA format on an NVIDIA machine with an opaque context returns the wrong alpha values.

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Sat Oct 18 23:40:02 PDT 2014


https://bugs.webkit.org/show_bug.cgi?id=137793

--- Comment #16 from Roger Fong <roger_fong at apple.com> ---
(In reply to comment #15)
> Comment on attachment 240053 [details]
> patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=240053&action=review
> 
> > Source/WebCore/platform/graphics/opengl/GraphicsContext3DOpenGL.cpp:72
> > +    // NVIDIA drivers have a bug where calling readPixels in BGRA can return the wrong values for the alpha channel when the alpha is off for the context.
> 
> Is this true for all NVIDIA drivers, historically? Is this likely to change?
> Should we be detecting the wrongness somehow instead of just flipping things
> around like this?

We'd then have to make two queries to glReadPixels then which seems unfortunate.
I've tested on 3 devices, dating back to the 2009 macbook pro card.

> 
> > Source/WebCore/platform/graphics/opengl/GraphicsContext3DOpenGL.cpp:78
> > +        src.rowBytes = width*4;
> 
> Spaces around the star, please.
> 
> > Source/WebCore/platform/graphics/opengl/GraphicsContext3DOpenGL.cpp:84
> > +        dest.rowBytes = width*4;
> 
> Ditto.
Whoops missed those, I'll add those in.

-- 
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/20141019/5491065b/attachment-0002.html>


More information about the webkit-unassigned mailing list