[Bug 768510] appsrc: handle pushing custom segments

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=768510

--- Comment #8 from Seungha Yang <sh.yang at lge.com> ---
(In reply to Sebastian Dröge (slomo) from comment #7)
> Review of attachment 331151 [details] [review]:
> 
> Thanks, this looks like a good start :) We should double-check if the
> seamless segment API of basesrc does the right thing though, it was
> implemented for a completely different use-case. That's why it has the
> "seamless" in the name. This here is not really for anything seamless
> anymore.

Thanks for your detailed review.
Frankly, the "seamless" is confusing me because it's used only dvb related
element, (I've never use it....). Can I ask why it's called seamless? (in the
sense the SEGMENT for seamless playback?)

What I've studied about previous use-cases, and they are
- DVDNAV_NAV_PACKET event ==> might pts discont. happen. For the "seamless"
playback, it need to push new segment without seek/flush
- DVDNAV_HOP_CHANNEL ==> obviously "non-seamless"
Both cases are use case of the seamless API if my understanding is correct. So
I though this API can be used for seek/flush independent SEGMENT event.

I don't think that my use case of this patch is not much different.
- DASH multi-period causes PTS discont, so need new SEGMENT without seek/flush.
- Playing with non-zero PTS is possible by using new SEGMENT (without
seek/flush). 

I might be wrong, so please correct me


> I'm not sure there should be a new property for this. It should be safe to
> just always do that if the sample contains a valid segment

I agree with you. I'll do rework.

> @@ +804,3 @@
> +   * - Application has been notified that it can control pipeline timeline
> +   *   (by setting 'driven-by-application' property)
> +   * - Both appsrc segment and the custom segment have TIME format
> 
> BYTES format and BYTES segment and other compatible pairs should also work
> though

I'll try to support BYTES format and BYTES segment, but could you explain more
about compatible pairs?

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.