Bug 135291 – [MSE] YouTube playback stalls & readyState drops to HAVE_CURRENT_DATA at end of stream with unbalanced buffered SourceBuffers

• Component: New Bugs
• Status: RESOLVED FIXED

[Jer Noble]

[MSE] Playback stalls & readyState drops to HAVE_CURRENT_DATA at end of stream with unbalanced buffered SourceBuffers

[Jon Lee]

<rdar://problem/17715503>

[Jer Noble]

Created attachment 235525 [details]
Patch

[Jon Lee]

Comment on attachment 235525 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=235525&action=review

> Source/WebCore/Modules/mediasource/SourceBuffer.cpp:1462
> +        MediaTime start = length ? virtualRanges->end(length - 1) : MediaTime::zeroTime();

Is the added check necessary since an invalid index will return zeroTime() anyway?

[Build Bot]

Comment on attachment 235525 [details]
Patch

Attachment 235525 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-queues.appspot.com/results/5917561690521600

New failing tests:
media/track/add-and-remove-track.html

[Build Bot]

Created attachment 235535 [details]
Archive of layout-test-results from webkit-ews-09 for mac-mountainlion-wk2

The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews.
Bot: webkit-ews-09  Port: mac-mountainlion-wk2  Platform: Mac OS X 10.8.5

[Jer Noble]

Created attachment 235539 [details]
Patch

[Sam Weinig]

Comment on attachment 235539 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=235539&action=review

> Source/WebCore/ChangeLog:3
> +        [MSE] Playback stalls & readyState drops to HAVE_CURRENT_DATA at end of stream with unbalanced buffered SourceBuffers

Please add the radar for future Sam's sake and also change the title to explain why this is important (aka Youtube is broken).

> Source/WebCore/platform/graphics/PlatformTimeRanges.h:54
>  
> +    MediaTime start(unsigned index) const;
>      MediaTime start(unsigned index, bool& valid) const;
> +    MediaTime end(unsigned index) const;
>      MediaTime end(unsigned index, bool& valid) const;
> +    MediaTime duration(unsigned index) const;
> +    MediaTime maximumBufferedTime() const;

You don't have to do this now, but you should eventually switch these to return Optional<MediaTime> if they can fail.

[Jer Noble]

Comment on attachment 235539 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=235539&action=review

>> Source/WebCore/ChangeLog:3
>> +        [MSE] Playback stalls & readyState drops to HAVE_CURRENT_DATA at end of stream with unbalanced buffered SourceBuffers
> 
> Please add the radar for future Sam's sake and also change the title to explain why this is important (aka Youtube is broken).

Will do.

>> Source/WebCore/platform/graphics/PlatformTimeRanges.h:54
>> +    MediaTime maximumBufferedTime() const;
> 
> You don't have to do this now, but you should eventually switch these to return Optional<MediaTime> if they can fail.

Neat, I learned something new today.

[Jer Noble]

Committed r171624: <http://trac.webkit.org/changeset/171624>