[MSE] Further fixes for "fast forward" playback after seeking in YouTube behavior.

[Jer Noble]

[MSE] Further fixes for "fast forward" playback after seeking in YouTube behavior.

[Jer Noble]

Created attachment 235984 [details]
Patch

[Eric Carlson]

Comment on attachment 235984 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=235984&action=review

> Source/WebCore/Modules/mediasource/SourceBuffer.cpp:1211
> +        if (presentationTimestamp > MediaTime::createWithDouble(m_source->currentTime())) {

Should this be >= instead of > ?

[Jer Noble]

(In reply to comment #2)
> (From update of attachment 235984 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=235984&action=review
> 
> > Source/WebCore/Modules/mediasource/SourceBuffer.cpp:1211
> > +        if (presentationTimestamp > MediaTime::createWithDouble(m_source->currentTime())) {
> 
> Should this be >= instead of > ?

Good point.  In fact, this should probably be >, but checking against frameEndTimestamp. I.e., if the frame spans the currentTime (where the start time is <= but the end time is >), we should add the sample to the decode queue.

[Jer Noble]

Committed r172000: <http://trac.webkit.org/changeset/172000>

[Jer Noble]

Rolled out in r172007 <http://trac.webkit.org/r172007>. Re-opening.

[Jon Lee]

<rdar://problem/17856763>