Bug 72423 – [Chromium] Crash executing the WebGL conformance tests

• Status: RESOLVED
• Resolution: FIXED
• Component: WebGL

[Jeff Timanus]

During WebGLRenderingContext tear down, the WebGLLayerChromium is attempting to stop a timer for a NULL context.

See chromium tracking issue here:  crbug.com\104367

[Jeff Timanus]

Created attachment 115248 [details]
Patch

[Jeff Timanus]

Created attachment 115273 [details]
Patch

[Jeff Timanus]

Created attachment 115276 [details]
Patch

[Kenneth Russell]

Comment on attachment 115276 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=115276&action=review

Looks good to me; minor comment nit. Could you upload a final patch? Also, did you get an email about what the style failure is? (Did you upload with webkit-patch upload? That checks style before uploading.)

> Source/WebCore/ChangeLog:4
> +        [chromium] During tear down, prevent the WebGLLayerChromium instance
> +        from attempting to stop a timer for a NULL context.

Does this need to be one line?

> Source/WebCore/platform/graphics/chromium/WebGLLayerChromium.cpp:131
> +    // The GraphicsContext3D used by the layer is the context associated with
> +    // with the drawing buffer. If the drawing buffer is changing, make sure

with with -> with

[Jeff Timanus]

Created attachment 115281 [details]
Patch

[Jeff Timanus]

Comment on attachment 115276 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=115276&action=review

>> Source/WebCore/ChangeLog:4
>> +        from attempting to stop a timer for a NULL context.
> 
> Does this need to be one line?

Done.

>> Source/WebCore/platform/graphics/chromium/WebGLLayerChromium.cpp:131
>> +    // with the drawing buffer. If the drawing buffer is changing, make sure
> 
> with with -> with

Thanks for catching that.

Done.

[Jeff Timanus]

(In reply to comment #4)
> (From update of attachment 115276 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=115276&action=review
> 
> Looks good to me; minor comment nit. Could you upload a final patch? Also, did you get an email about what the style failure is? (Did you upload with webkit-patch upload? That checks style before uploading.)

Yes, I used webkit-patch upload to upload all of these patches.  I also ran check-webkit-style before uploading.  It reported no errors.  I have yet to receive the style failure notification, so I'm not sure what is up.

Fortunately, the most recent patch seems to be passing the style check.

> 
> > Source/WebCore/ChangeLog:4
> > +        [chromium] During tear down, prevent the WebGLLayerChromium instance
> > +        from attempting to stop a timer for a NULL context.
> 
> Does this need to be one line?
> 
> > Source/WebCore/platform/graphics/chromium/WebGLLayerChromium.cpp:131
> > +    // The GraphicsContext3D used by the layer is the context associated with
> > +    // with the drawing buffer. If the drawing buffer is changing, make sure
> 
> with with -> with

[Kenneth Russell]

Comment on attachment 115281 [details]
Patch

Looks good. r=me

[WebKit Review Bot]

Comment on attachment 115281 [details]
Patch

Clearing flags on attachment: 115281

Committed r100401: <http://trac.webkit.org/changeset/100401>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.