[gst-devel] spider restart

[Martin Soto]

On Mon, 2004-06-28 at 16:50, Giuliano Pochini wrote:
> On 28-Jun-2004 Johan Dahlin wrote:
> > ...
> > Well, nobody's stopping you (or anyone else) from writing a working
> > autoplugger for 0.8 right?
> 
> Yes, but I have no time.  I want to use gstreamer as an user, without
> caring too much about its internals and its development. My project
> is an audio CD player that supports at least all free file formats.
> I'll buy and EPIA motherboard and a hi-end sound card for that. Once
> the cd player software will work fine, I'll start developing a plugin
> to apply some effects and I will learn enough to be able to help
> hunting bugs. I don't want to spend time for gstreamer itself right
> now because my project is going on slow enough.

Well, development manpower is always limited. People will usually do
their best to help you, but there are no warranties. I'd also like to
see my project (a DVD player) go faster, but already had to make a few
fixes and improvements to GStreamer (and spend time learning the
internals and doing the work) in order to get things to work the way I
want. It is like that with every free software project you rely on:
either you accept all limitations as they come, or you fix them
yourself. Whining and screaming won't help at all...

> I'm currently stuck with these major problems:
> 
> - Spider reusability (ok, I'll look at libgstplay).
> - Seeking.
> - Random audio skips with alsasink (I'm trying to narrow
> this down before filing a bug report because it's strange.
> Alsasink sometimes stops the substream, sets again the
> hw_params and restarts the substream. Perhaps is it an
> incorrect out-of-data condition detection ?).

Are you using kernel 2.6, by the way? Under relatively moderate load
conditions, I've been noticing skips in all Alsa based applications,
including XMMS. An interesting thing is that playing through the ALSA
OSS emulation works flawlessly. It seems that the sound interface
returns funny values some times, because applications seem to be
confused. XMMS, for example, jumps back and forth before settling down,
and I wouldn't be surprised to know that the behavior you're seeing in
alsasink has the same cause.

I brought the issue some days ago to the Alsa mailing list, but no one
seemed to have a satisfactory answer. You may search the archives from
the last days, thought. I haven't checked in the last couple of weeks.

Cheers,

M. S.
-- 
Martin Soto <soto at informatik.uni-kl.de>