[Webkit-unassigned] [Bug 60770] [Qt] QtWebKit bridge assignToHTMLImageElement() results in invalid Image for WebGLRenderingContext

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=60770





--- Comment #4 from Andrew Wason <rectalogic at rectalogic.com>  2011-05-18 08:29:27 PST ---
(In reply to comment #3)
> The patch looks more like a big workaround.
> 
> If an origin has to be set, I would rather see the origin of the frame.

OK, I think using the frame origin would also fix bug #59972

> The code of WebGLRenderingContext.cpp might also be bogus.
> Is there no way to reproduce the same issue by manipulating the image directly or with Canvas?

A similar problem (bug #59972) can be reproduced with Canvas, but this bug is specific to how WebGLRenderingContext is validating images.

WebGLRenderingContext::validateHTMLImageElement() was added to fix bug #58501. I think validateHTMLImageElement() could check HTMLImageElement::complete() and that width/height are non-zero, instead of looking at the response url. This seems to be what WebCore::CanvasRenderingContext2D::drawImage() is doing.

So should I use the frame origin (and add tests), or change WebGLRenderingContext::validateHTMLImageElement()? Setting the origin has the advantage of also fixing bug #59972

Maybe this bug should be fixed by changing validateHTMLImageElement() to check complete/width/height instead of response url, and a separate patch to set the image origin to fix bug #59972?

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