[Bug 766612] Does not take address pool configuration into account for sending unicast UDP

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=766612

--- Comment #24 from Sebastian Dröge (slomo) <slomo at coaxion.net> ---
(In reply to Xavier Claessens from comment #22)

> 3) Do 2) then also do the complicated task of not having any udpsink
> elements until SETUP just in case the first client that connects is an mcast
> one that wants another destination/port. If we do this, and since in (2) we
> always have a tee anyway, I would go with a fakesink rather than blocking
> the stream. It turned out to be way too complicated, I still haven't
> succeeded in making unit tests pass with that solution.

You mean you didn't succeed just blocking and having no sinks? How does it fail
and why?


In general, I think we should only allocate UDP ports when needed (that is:
SETUP). This means that before we need some kind of sinks or the pad blocks.

Then during SETUP we should take into account what the clients tell us and set
up things accordingly, possibly adding new sinks/sources as needed. And during
teardown shut those down again depending on the still existing clients.


Having always a tee and a fakesink for the sinks is a possible option, but it
will have to be optimized away at a later time. The tee requires queues, and
people were complaining in the past already about the performance impact of
that (which is why we only add the tee and stuff when needed currently).

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