Bug 144782 – 3D-transformed video does not display on platforms without accelerated video rendering

• Status: RESOLVED
• Resolution: FIXED
• Component: Media

[[e-mail]]

I checked same behavior on EFL port.
I checked same behavior on GTK port with disable GSTREAMER_GL and ENABLE_THREADED_COMPOSITOR.

Video tag does not get RenderLayerBacking when MediaPlayerPrivate::supportsAcceleratedRendering() returns false which means no using accelerated video decoding.
With changing CSS Style by javascript, video tag gets RenderLayerBacking.
But there  is no way to set GraphicsLayer::setDrawsContent(true) for updating video content.

Reproduce: http://codepen.io/anon/pen/qdbBep

[[e-mail]]

Created attachment 252676 [details]
Set GraphicsLayer::m_drawsContent = true

[Simon Fraser (smfr)]

Comment on attachment 252676 [details]
Set GraphicsLayer::m_drawsContent = true

View in context: https://bugs.webkit.org/attachment.cgi?id=252676&action=review

> Source/WebCore/ChangeLog:7
> +

There's no explanation here of why being CSS transformed from JavaScript matters, and what the code change is doing.

Is this testable?

[Alexey Proskuryakov]

> Reproduce: http://codepen.io/anon/pen/qdbBep

The video at this URL fails to play in shipping Safari, so it's difficult to see what the regression is. Could you please clarify how to reproduce the problem?

[[e-mail]]

Ok, I'm on the way to make the test case in layouttest.

[[e-mail]]

Created attachment 252702 [details]
patch

[[e-mail]]

Created attachment 252704 [details]
patch

[[e-mail]]

Created attachment 252707 [details]
patch

[Simon Fraser (smfr)]

Comment on attachment 252707 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=252707&action=review

The code change seems OK, but the test will break everyone.

> LayoutTests/media/video-transformed-by-javascript-expected.txt:12
> +          (drawsContent 1)

This result will be platform-specific, because this layer will only be drawsContent=1 on platforms that don't have supportsAcceleratedRendering().

> LayoutTests/media/video-transformed-by-javascript.html:29
> +    <video id="videoRotate" width="480" height="270" type="video/mp4" src="test.mp4"/>

test.mp4 doesn't exist in that directory. You need to use the functions in media-file.js to load a media type that works on all platforms, like the other tests in this directory.

[[e-mail]]

Comment on attachment 252707 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=252707&action=review

>> LayoutTests/media/video-transformed-by-javascript-expected.txt:12
>> +          (drawsContent 1)
> 
> This result will be platform-specific, because this layer will only be drawsContent=1 on platforms that don't have supportsAcceleratedRendering().

Agreed. Do you have any recommendation like pixel-level comparisons?

>> LayoutTests/media/video-transformed-by-javascript.html:29
>> +    <video id="videoRotate" width="480" height="270" type="video/mp4" src="test.mp4"/>
> 
> test.mp4 doesn't exist in that directory. You need to use the functions in media-file.js to load a media type that works on all platforms, like the other tests in this directory.

Yes, you are right.

[[e-mail]]

Created attachment 252839 [details]
patch

[Gyuyoung Kim]

Comment on attachment 252839 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=252839&action=review

> Source/WebCore/testing/Internals.cpp:2372
> +    return element->player()->supportsAcceleratedRendering();

It seems that element->player() needs to be checked if it is null or not.

[[e-mail]]

Comment on attachment 252839 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=252839&action=review

>> Source/WebCore/testing/Internals.cpp:2372
>> +    return element->player()->supportsAcceleratedRendering();
> 
> It seems that element->player() needs to be checked if it is null or not.

Ok, thanks.

[[e-mail]]

Created attachment 252848 [details]
patch

[[e-mail]]

Created attachment 252850 [details]
patch

[[e-mail]]

Ping...

[Gyuyoung Kim]

(In reply to comment #3)
> > Reproduce: http://codepen.io/anon/pen/qdbBep
> 
> The video at this URL fails to play in shipping Safari, so it's difficult to
> see what the regression is. Could you please clarify how to reproduce the
> problem?

Daegyu, is there any other way to reproduce this bug ?

Simon, we need to fix this bug. So could you take a look this patch again ?

[[e-mail]]

http://codepen.io/anon/pen/pJEROO

I have tested above url on EFL. Video is not shown when you click "Flip Video" button.

I have tested above url on GTK without accelerated video rendering. Video is not shown when you click "Flip Video" button.

I have tested above url on Mac. Video is shown correctly.

[Simon Fraser (smfr)]

Comment on attachment 252850 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=252850&action=review

> LayoutTests/media/video-transformed-by-javascript.html:22
> +                if (window.internals.supportsAcceleratedRendering(document.getElementById("videoRotate")) == false) {
> +                    layerTreeAsText = window.internals.layerTreeAsText(document);
> +                    if (layerTreeAsText.indexOf('(drawsContent 1') == -1)
> +                        document.getElementById('outText').innerText = "FAIL";
> +                }
> +                testRunner.notifyDone();

I'm not sure it's worth going to the trouble of adding internals.supportsAcceleratedRendering(). Why not just land different results for different platforms?

[[e-mail]]

Created attachment 254042 [details]
patch

[[e-mail]]

Created attachment 254346 [details]
patch

[Gyuyoung Kim]

Comment on attachment 254346 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=254346&action=review

Simon already set r+ed. Please only request cq? after fixing Simon's comment.

> LayoutTests/ChangeLog:21
> [phone]  daegyu lee  <[e-mail]>

Duplicated changelog.

[[e-mail]]

Created attachment 254624 [details]
patch

[WebKit Commit Bot]

Comment on attachment 254624 [details]
patch

Rejecting attachment 254624 [details] from commit-queue.

[e-mail] does not have committer permissions according to http://trac.webkit.org/browser/trunk/Tools/Scripts/webkitpy/common/config/contributors.json.

- If you do not have committer rights please read http://webkit.org/coding/contributing.html for instructions on how to use bugzilla flags.

- If you have committer rights please correct the error in Tools/Scripts/webkitpy/common/config/contributors.json by adding yourself to the file (no review needed).  The commit-queue restarts itself every 2 hours.  After restart the commit-queue will correctly respect your committer rights.

[Gyuyoung Kim]

Comment on attachment 254624 [details]
patch

cq=me. win-ews seems to be broken by other patch.

[WebKit Commit Bot]

Comment on attachment 254624 [details]
patch

Clearing flags on attachment: 254624

Committed r185402: <http://trac.webkit.org/changeset/185402>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.