Bug 90497

Summary: [Qt] When building WebKit with Qt5+minimal, WebGL should be enabled
Product: WebKit Reporter: Noam Rosenthal <noam>
Component: New BugsAssignee: Noam Rosenthal <noam>
Status: RESOLVED INVALID    
Severity: Normal CC: jesus, jturcotte, menard, vestbo
Priority: P2    
Version: 528+ (Nightly build)   
Hardware: Unspecified   
OS: Unspecified   
Attachments:
Description Flags
Patch vestbo: review-

Description Noam Rosenthal 2012-07-03 14:42:01 PDT
[Qt] When building WebKit with Qt5+minimal, WebGL should be enabled
Comment 1 Noam Rosenthal 2012-07-03 14:47:39 PDT
Created attachment 150671 [details]
Patch
Comment 2 Jesus Sanchez-Palencia 2012-07-03 14:51:07 PDT
Comment on attachment 150671 [details]
Patch

If they can't be joined under one condition check, couldn't they be checked at least at the same place?
Comment 3 Noam Rosenthal 2012-07-03 15:05:33 PDT
(In reply to comment #2)
> (From update of attachment 150671 [details])
> If they can't be joined under one condition check, couldn't they be checked at least at the same place?

No, because the one for Qt4 has to be done before applying the defaults, and the one for Qt5 after we apply the defaults.
Comment 4 Tor Arne Vestbø 2012-07-03 15:43:18 PDT
Comment on attachment 150671 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=150671&action=review

> Tools/ChangeLog:6
> +        WebGL-related code (GraphicsContext3D etc.) is now used as part of TextureMapperGL's path.

We should untangle this so that you can build with ENABLE_WEBGL=0 and still have TextureMapperGL work.

> Tools/qmake/mkspecs/features/features.prf:200
> +    root_project_file: message("WebGL code is part of the rendering  with Qt5. Overriding --no-webgl option.")

This seems upside-down to me? We shouldn't try to fix the entanglement by entangling it more.
Comment 5 Noam Rosenthal 2012-07-03 16:07:40 PDT
Sure, we can go down that r(In reply to comment #4)
> (From update of attachment 150671 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=150671&action=review
> 
> > Tools/ChangeLog:6
> > +        WebGL-related code (GraphicsContext3D etc.) is now used as part of TextureMapperGL's path.
> 
> We should untangle this so that you can build with ENABLE_WEBGL=0 and still have TextureMapperGL work.

Sure, we can go down that route, though I'm trying to keep things simple rather than support too many untested permutations of enabled/disabled features.
Comment 6 Noam Rosenthal 2012-07-03 16:24:18 PDT
Discussed it on IRC, I will work on moving the shared stuff to behind a USE() flag, will be a lot cleaner than ENABLE(WEBGL)
Comment 7 Tor Arne Vestbø 2012-07-03 16:26:03 PDT
Comment on attachment 150671 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=150671&action=review

>>> Tools/ChangeLog:6
>>> +        WebGL-related code (GraphicsContext3D etc.) is now used as part of TextureMapperGL's path.
>> 
>> We should untangle this so that you can build with ENABLE_WEBGL=0 and still have TextureMapperGL work.
> 
> Sure, we can go down that route, though I'm trying to keep things simple rather than support too many untested permutations of enabled/disabled features.

If that's the plan then we should remove ENABLE(WEBGL) completely, but that requires discussion on webkit-dev. That's effectively what this patch does, by enabling it without really using it.
Comment 8 Tor Arne Vestbø 2012-07-03 16:26:18 PDT
(In reply to comment #6)
> Discussed it on IRC, I will work on moving the shared stuff to behind a USE() flag, will be a lot cleaner than ENABLE(WEBGL)

Sounds good!