Bug 135164 – [MSE][Mac] In SourceBufferPrivateAVFObjC::abort(), support reseting parser to the last appended initialization segment.

• Component: New Bugs
• Status: RESOLVED FIXED

[Jer Noble]

[MSE][Mac] In SourceBufferPrivateAVFObjC::abort(), support reseting parser to the last appended initialization segment.

[Jer Noble]

Created attachment 290022 [details]
Patch

[Jer Noble]

rdar://problem/28438637

[Eric Carlson]

Comment on attachment 290022 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=290022&action=review

Looks like you forgot a few files :-)

> Source/WebCore/ChangeLog:17
> +        and will block until the the previous append() operation completes.

Nit: the the

> Source/WebCore/ChangeLog:20
> +        SourceBufferPrivateAVFObjC to be reset after an abort(), so make that ivar a @property. Rather than passing a

Nit: "a @property" -> "an @property"

[Jer Noble]

Created attachment 290026 [details]
Patch

[Eric Carlson]

Comment on attachment 290026 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=290026&action=review

> Source/WebCore/ChangeLog:17
> +        and will block until the the previous append() operation completes.

Nit: "the the"

> Source/WebCore/ChangeLog:20
> +        SourceBufferPrivateAVFObjC to be reset after an abort(), so make that ivar a @property. Rather than passing a

Nit: "a @property" -> "an @property"

> Source/WebCore/platform/graphics/avfoundation/objc/SourceBufferPrivateAVFObjC.mm:675
> +    dispatch_group_wait(m_isAppendingGroup.get(), DISPATCH_TIME_FOREVER);

Is it really safe to wait forever?

[Jer Noble]

(In reply to comment #5)
> Comment on attachment 290026 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=290026&action=review
> 
> > Source/WebCore/ChangeLog:17
> > +        and will block until the the previous append() operation completes.
> 
> Nit: "the the"

Ok.

> > Source/WebCore/ChangeLog:20
> > +        SourceBufferPrivateAVFObjC to be reset after an abort(), so make that ivar a @property. Rather than passing a
> 
> Nit: "a @property" -> "an @property"

Ok.

> > Source/WebCore/platform/graphics/avfoundation/objc/SourceBufferPrivateAVFObjC.mm:675
> > +    dispatch_group_wait(m_isAppendingGroup.get(), DISPATCH_TIME_FOREVER);
> 
> Is it really safe to wait forever?

Now that I think of it, no, it's not.  The parsing thread may be blocked on the main thread providing it a AVStreamSession for EME-backed media elements. Apart from that, however, it should be fine.

[Jer Noble]

Created attachment 290034 [details]
Patch for landing

[Jer Noble]

Created attachment 290040 [details]
Patch for landing

[WebKit Commit Bot]

Comment on attachment 290040 [details]
Patch for landing

Clearing flags on attachment: 290040

Committed r206518: <http://trac.webkit.org/changeset/206518>

[Ahmad Saleem]

Seems landed but never changed to fixed - https://github.com/WebKit/WebKit/commit/f533a6787f5fe06616ced4d44333dd16f24eb59d