[icecast] Problem with ices on OpenBSD 2.9 w/ Icecast 1.3.10
Nick Ludlam
nick at ivision.co.uk
Tue Jul 10 14:38:11 UTC 2001
From: "Brendan Cully" <brendan at icecast.org>
> On Tuesday, 10 July 2001 at 14:10, Nick Ludlam wrote:
> > From: "Brendan Cully" <brendan at icecast.org>
> > > On Tuesday, 10 July 2001 at 03:21, Nick Ludlam wrote:
> > > > From: "Brendan Cully" <brendan at icecast.org>
> > > > > That handles the segfault, I don't know why the stream wouldn't be
> > > > > sent fast enough though...
> > > >
> > > > Tried with winamp, sonique and freeamp. All exhibit the same problem of
> > > > stuttering play, and eventually this drops out of the icecast log:
> > > >
> > > > -> [10/Jul/2001:03:04:57] Kicking client 14 [212.135.236.195] [Too many errors (client not receiving data fast enough)]
> > [listener],
> > > > connected for 35 seconds, 257939 bytes transfered. 0 clients connected
> > > >
> > > > and I know it's not bandwidth-based problems. This happens on a local
> > > > network as well as my ADSL, both of which are good enough to stream
> > > > at 12Kb/sec.
> > >
> > > Ah, I found this one while playing with ices on solaris. The problem
> > > seems to be a truncation error in libshout's timing/timing.c, in
> > > timing_sleep, which results in ices not sleeping at all between
> > > chunks. I hope you're experiencing the same problem, because I posted
> > > a patch a couple days ago. Check out the CVS or wait for 0.2.1.
> >
> > Unfortunately I am using the CVS version. Is there any way for me to trace
> > where it's mistiming things?
>
> I guess you could put some printfs in shout_sleep and/or timing_sleep.
> Could you first double-check that timing_sleep in
> ices/libshout/timing.c looks like this:
>
> void timing_sleep(long sleeptime)
> {
> struct timeval sleeper;
>
> sleeper.tv_sec = sleeptime / 1000;
> sleeper.tv_usec = (sleeptime % 1000) * 1000;
>
> select (1, NULL, NULL, NULL, &sleeper);
> }
>
> and that you are not using an installed, older version of
> libshout. ldd ices and check which libshout it is using. Ideally
> libshout won't show up at all. You might try building
> --without-libshout just to make sure.
Apart frrom the arg being a long long, it's identical. I checked it out of CVS
yesterday. libshout doesn't show up at all when ldd'ing ices, so it's properly
statically linked.
Nick
--- >8 ----
List archives: http://www.xiph.org/archives/
icecast project homepage: http://www.icecast.org/
To unsubscribe from this list, send a message to 'icecast-request at xiph.org'
containing only the word 'unsubscribe' in the body. No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.
More information about the Icecast
mailing list