Bug 72753 – [WK2][Qt] Move Accelerated Composite animations to UIProcess

• Status: RESOLVED
• Resolution: FIXED
• Component: WebKit2

[Igor Trindade Oliveira]

SSIA

[Igor Trindade Oliveira]

Created attachment 115864 [details]
Patch

Proposed patch version 1.

[Noam Rosenthal]

Comment on attachment 115864 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=115864&action=review

Awesome! See comments, they are mainly nitpicks.

> Source/WebCore/ChangeLog:6
> +        Add helper method to synchronize animations.

... in TextureMapper.

> Source/WebCore/platform/graphics/texmap/TextureMapperNode.cpp:987
> +void TextureMapperNode::syncAnimationsRecursively(GraphicsLayerTextureMapper* graphicsLayer)

The argument might not be necessary. See next comment.

> Source/WebCore/platform/graphics/texmap/TextureMapperNode.cpp:994
> +        Vector<GraphicsLayer*> children = graphicsLayer->children();

You probably want to use the TextureMapperNode's children and not the graphicsLayer's children, otherwise you might be syncing animations for items that aren't in the synchronized tree yet.

> Source/WebCore/platform/graphics/texmap/TextureMapperNode.cpp:1000
> +        }

early return.

> Source/WebKit2/UIProcess/LayerTreeHostProxy.h:129
> +    void updateSceneGraphUpdateTimerFired(Timer<LayerTreeHostProxy>*);

I believe the convention is didFireViewportUpdateTimer

> Source/WebKit2/UIProcess/LayerTreeHostProxy.h:130
> +    Timer<LayerTreeHostProxy> m_sceneGraphUpdateTimer;

Let's use viewport instead of scenegraph

> Source/WebKit2/WebProcess/WebCoreSupport/WebGraphicsLayer.cpp:283
> +#if PLATFORM(QT)

Why QT only? This seems cross-platform. Nobody else uses WebGraphicsLayer yet so don't worry :)

> Source/WebKit2/WebProcess/WebCoreSupport/WebGraphicsLayer.cpp:306
> +#if PLATFORM(QT)

ditto

> Source/WebKit2/WebProcess/WebCoreSupport/WebGraphicsLayer.cpp:318
> +#if PLATFORM(QT)

ditto

[Igor Trindade Oliveira]

Created attachment 115897 [details]
Patch

Patch v2. Implement all suggestions proposed by Noam.

[Igor Trindade Oliveira]

Created attachment 115907 [details]
Patch

Patch v2. Improve Changelog

[Noam Rosenthal]

Comment on attachment 115907 [details]
Patch

Awsomeness.

[WebKit Review Bot]

Comment on attachment 115907 [details]
Patch

Clearing flags on attachment: 115907

Committed r100834: <http://trac.webkit.org/changeset/100834>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.

[Csaba Osztrogonác]

Many tests started to crash after this patch:
http://build.webkit.sed.hu/results/x86-32%20Linux%20Qt%20Release%20WebKit2/r100834%20%2815660%29/results.html

Unfortunately it made run-webkit-test exit before finishing. :(

Could you check it please as soon as possible?

[Igor Trindade Oliveira]

Ok Ossy, i am investigating right now


(In reply to comment #8)
> Many tests started to crash after this patch:
> http://build.webkit.sed.hu/results/x86-32%20Linux%20Qt%20Release%20WebKit2/r100834%20%2815660%29/results.html
> 
> Unfortunately it made run-webkit-test exit before finishing. :(
> 
> Could you check it please as soon as possible?

[Csaba Osztrogonác]

Rolled out by http://trac.webkit.org/changeset/100859

[Igor Trindade Oliveira]

Created attachment 116076 [details]
Patch

Proposed patch. This patch fixes the crashes.

[Noam Rosenthal]

Comment on attachment 116076 [details]
Patch

Cool, Please commit yourself when you can watch the bots :)

[WebKit Review Bot]

Comment on attachment 116076 [details]
Patch

Attachment 116076 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/10363332

New failing tests:
inspector/extensions/extensions-events.html

[Noam Rosenthal]

Igor, should we close this bug?

[Igor Trindade Oliveira]

Committed r100922: http://trac.webkit.org/changeset/100922

(In reply to comment #14)
> Igor, should we close this bug?