Bug 103298 – Coordinated Graphics: Refactor code managing a backing store in LayerTreeRenderer.

• Status: RESOLVED
• Resolution: FIXED
• Component: Layout and Rendering

[Dongseong Hwang]

Coordinated Graphics: Refactor code managing a backing store in LayerTreeRenderer.

[Dongseong Hwang]

*** Bug 103215 has been marked as a duplicate of this bug. ***

[Dongseong Hwang]

Created attachment 176079 [details]
Patch

[Dongseong Hwang]

It is strange. In new bug, this patch is purple too.

[Noam Rosenthal]

Looking at the changelog, seems like it's depending on the surface bug, which is not yet in.

[Dongseong Hwang]

Created attachment 176139 [details]
Patch

[Dongseong Hwang]

Created attachment 176152 [details]
Patch

[Dongseong Hwang]

(In reply to comment #4)
> Looking at the changelog, seems like it's depending on the surface bug, which is not yet in.

Oops! You're right! EWS is smart!

[Dongseong Hwang]

Created attachment 176164 [details]
Patch

[Noam Rosenthal]

Comment on attachment 176164 [details]
Patch

Before this is committed, are the assert regressions from the previous patches all taken care of? I don't want to run too fast.

[Dongseong Hwang]

(In reply to comment #9)
> (From update of attachment 176164 [details])
> Before this is committed, are the assert regressions from the previous patches all taken care of? I don't want to run too fast.

I'm in progress.

In the detail, I'll make GraphicsLayerTextureMapper traverse TextureMapperLayer tree. After that, TextureMapperLayer::flushCompositingState does not traverse recursively anymore.

After that, I'll fix Bug 103171.
LayerTreeRenderer::setLayerState can synchronize with TextureMapperLayer at that time.

All will be finished today.

[Zeno Albisser]

(In reply to comment #9)
> (From update of attachment 176164 [details])
> Before this is committed, are the assert regressions from the previous patches all taken care of? I don't want to run too fast.

I am a bit concerned about the pace here as well. Lately i see a lot of patches going in in this area. And some of these touch quite integral parts of the architecture for several ports.
I am also quite confident that some of the patches cause regressions. Not all of these are caught by LayoutTests unfortunately.
It would be nice if we would at least get some time to check for such issues and to fix them.

[Noam Rosenthal]

(In reply to comment #11)
> (In reply to comment #9)
> > (From update of attachment 176164 [details] [details])
> > Before this is committed, are the assert regressions from the previous patches all taken care of? I don't want to run too fast.
> 
> I am a bit concerned about the pace here as well. Lately i see a lot of patches going in in this area. And some of these touch quite integral parts of the architecture for several ports.
> I am also quite confident that some of the patches cause regressions. Not all of these are caught by LayoutTests unfortunately.

I agree. I might have been a bit hasty in my reviews; My sense was that the changes so far have been going in the right direction and that Huang was diligently following up on issues, but now we should go slower and if some regressions still exist either fix them or roll back.

[Dongseong Hwang]

(In reply to comment #12)
> (In reply to comment #11)
> > (In reply to comment #9)
> > > (From update of attachment 176164 [details] [details] [details])
> > > Before this is committed, are the assert regressions from the previous patches all taken care of? I don't want to run too fast.
> > 
> > I am a bit concerned about the pace here as well. Lately i see a lot of patches going in in this area. And some of these touch quite integral parts of the architecture for several ports.
> > I am also quite confident that some of the patches cause regressions. Not all of these are caught by LayoutTests unfortunately.
> 
> I agree. I might have been a bit hasty in my reviews; My sense was that the changes so far have been going in the right direction and that Huang was diligently following up on issues, but now we should go slower and if some regressions still exist either fix them or roll back.

I think so. My mind has been a bit urgent in that I want to improve coordinated graphics ASAP. From now, I'll take more time to test and do my best to fix regressions.

[WebKit Review Bot]

Comment on attachment 176164 [details]
Patch

Clearing flags on attachment: 176164

Committed r135956: <http://trac.webkit.org/changeset/135956>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.