[Bug 764733] qtdemux: Regression in YouTube TV tests in WebKit MSE after fix for #760779

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=764733

--- Comment #1 from Seungha Yang <sh.yang at lge.com> ---
(In reply to Xabier Rodríguez Calvar from comment #0)
> In our MSE implementation for WebKitForWayland we found that this patch
> causes regressions at the tests for YouTube TV.
> 
> We are using qtdemux to parse the data appended to the SourceBuffer and it
> seems that they delay in exposing the streams introduced by this patch is an
> issue.
> 
> For us the optimal behavior here would be to remove this patch but I guess
> Seungha Yang wouldn't be very happy with it. Any thoughts?
> 
> +++ This bug was initially created as a clone of Bug #760779 +++
> 
> In case of push mode, qtdemux expose streams after got moov box.
> We can not guarantee that moov box has sample data such as sample duration
> and the number of sample in stbl box in fragmented format case.
> If moov has no sample data, this patch will not expose streams
> until get the first moof.
> This patch depends on https://bugzilla.gnome.org/show_bug.cgi?id=760505

I think revert this patch is reasonable (I'm ok :) ). I'm sorry that MSE case
was not fully considered when I prepared the patch. Dear Xabier Rodríguez
Calvar, I'll make another solution to cover MSE & normal fragmented case. So,
please feel free to request reverting this patch.

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