Hi,
Thanks for the clock answer, now I understand why I had the two wallclock
error (wallclock_OSS and wallclock_oss).
For the rest. I'm struggling with an UMTS/3G connection which gives me
errors lately. They are hard to reproduce, that's the reason that I use the
route del way to test and debug.
I did some testing on a virtual server and with the latest SVN I get the
following error (didn't change the script that worked on rev8097)
Line 42, char 1: Unknown error
Fatal error: exception Cry.Error(1)
Current liq file:
set("log.file",true)
set("log.file.path","/home/fitpc2/liquidsoap.log")
set("log.level",5)
set("log.stdout",true)
def log_connect ()
route1 = get_process_output("route -n")
log(label="log_connect",level=5,route1)
signal1 = get_process_output("comgt sig -d /dev/ttyHS2")
log(label="log_connect",level=5,signal1)
end
def log_disconnect ()
route2 = get_process_output("route -n")
log(label="log_disconnect",level=5,route2)
signal2 = get_process_output("comgt sig -d /dev/ttyHS2")
log(label="log_disconnect",level=5,signal2)
end
input = input.oss()
clock(id="icecast",
output.icecast(
%mp3(mono, samplerate=22050, bitrate=24),
host="xoxo",
port=80,
mount="xoxo",
password="xoxo",
restart=true,
on_connect=log_connect,
on_disconnect=log_disconnect,
mksafe(buffer(input))))
output.file(
%mp3(mono),
"/home/fitpc2/record-%Y-%m-%d-%H-%M-%S.mp3",
input)
log_clocks(delay=10.,interval=5.,"/home/fitpc2/clocks.log")
----- end liq file ------
btw in rev8097 this script gave me the following notice: liquidsoap.liq At
line 24 char 6 - line 34 char 25: this value should be an active source, or
unit.
On my production PC this gave me strange behaviour:
- starting via init.d caused a few lines in the log (disclaimer) and hanging
of the program which i could kill (the reason why the production pc is back
to 1.0.0-beta1)
- from command line starting was no problem, but with notice.
Starting on the virtual server works both ways, but with notice.
grts, Cees
On Thu, Dec 30, 2010 at 7:35 PM, David Baelde <[email protected]>wrote:
> Hi,
>
> Concerning your clock question:
>
> 2010/12/29 Cees van Egmond <[email protected]>:
> > Having the Clock on the output.icecast instead on the input. Is there a
> > special reason? The 1.0.0-beta1 code isn't working on the current svn
> > version.
>
> The reason for the difference is that now, the OSS/ALSA/etc inputs
> force their own clocks (this makes sense, the soundcard API guides the
> synchronization). So instead of putting the input in one clock and
> leaving the output in the default one, we have to write it the other
> way around: leave the input in the OSS clock and put the output in a
> separate clock.
>
> I think there's more detail on the clock page of our doc. Notably an
> explanation saying that as long as you use buffer() the clock() isn't
> strictly really needed -- but it's very preferable to leave it.
>
> I'm sorry this is a bit short, but I'm leaving for a week of holidays.
>
> Cheers,
> --
> David
>
------------------------------------------------------------------------------
Learn how Oracle Real Application Clusters (RAC) One Node allows customers
to consolidate database storage, standardize their database environment, and,
should the need arise, upgrade to a full multi-node Oracle RAC database
without downtime or disruption
http://p.sf.net/sfu/oracle-sfdevnl
_______________________________________________
Savonet-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/savonet-users
