[Bug 681198] [renegotiation] Reconfigure events on complex pipelines (multiple capsfilters/converters)

[GStreamer (bugzilla.gnome.org)]

https://bugzilla.gnome.org/show_bug.cgi?id=681198
  GStreamer | gstreamer (core) | 0.11.x

--- Comment #9 from Tim-Philipp Müller <t.i.m at zen.co.uk> 2012-08-06 22:28:19 UTC ---
> I'm thinking it's probably downstream elements that should keep on using the
> pre-reconfigure state until the source has re-configured...

Why ? It seems to me that continuing to use the pre-reconfigured state in a
downstream element(I'm thinking converters for example) if one *could*
reconfigure instead means:

a) data is *guaranteed* to get dropped, because we know we'll get a
not-negotiated, and we don't know how much data there is between the most
upstream element and us (could be queues with seconds worth of data).

b) we don't know if upstream will be able to renegotiate successfully at all,
which means we might actually drop all remaining data

c) we might not even ever error out then

d) we don't know when upstream fails to reconfigure, so downstream doesn't know
that it should try itself now

But perhaps I'm missing something somewhere.

-- 
Configure bugmail: https://bugzilla.gnome.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
You are the assignee for the bug.