[Webkit-unassigned] [Bug 102501] Flickering after transitions on Apple HTML5 demo

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=102501


Noam Rosenthal <noam at webkit.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
 Attachment #175807|review?                     |review-
               Flag|                            |




--- Comment #8 from Noam Rosenthal <noam at webkit.org>  2012-11-26 09:30:36 PST ---
(From update of attachment 175807)
View in context: https://bugs.webkit.org/attachment.cgi?id=175807&action=review

The contract between GraphicsLayerTextureMapper and TextureMapperLayer is the following:
- m_state should mirror the state of GLTM after the last sync.
- m_opacity and m_transform should hold the transform and opacity to paint with, including animations.

This patch breaks that contract; We should think harder to find a way to fix the flicking problem without breaking that contract. I don't have a precise idea right now but I'd give it some thought.

>>> Source/WebCore/platform/graphics/texmap/TextureMapperLayer.cpp:495
>>> +        graphicsLayer->setOpacity(m_state.opacity);
>> 
>> It seems a bit unusual for TextureMapperLayer to update GraphicsLayerTextureMapper. Is there a cleaner way to do this?
> 
> This is the only way to sync TextureMapperLayer and GraphicsLayerTextureMapper known to us at the moment, although it appears to be a bit messy. If there's a better way to do that could you please share your thoughts then? If it's critical we could try searching for other ways. But nevertheless we'd like to point that GraphicsLayerTextureMapper doesn't get passed to flushCompositingStateSelf as const, and that's why it's prone to changes inside this method.

Are you sure this would work with fill modes? I have a feeling this kind of code was there before and had some issues...

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