Hi guys,
It is possible to use custom authentication functions, which could be
useful here: write a function that, when asked too often, says no all
the time for a while. But this will still not be a very lightweight
mechanism, perhaps not enough to avoid flood attacks. If fail2ban does
not work
On Fri, Jun 25, 2010 at 9:58 AM, David Baelde david.bae...@ens-lyon.org wrote:
I confirm the bug, I could reproduce it. A crossfade on a rotate of
two playlist doesn't seem to get ready when it should, and so never
plays. A crossfade on just one playlist works fine, however. I don't
have time
On Thu, Jun 24, 2010 at 1:25 PM, Kerozen kero...@gmail.com wrote:
In my playlists, I have a bunch of m4a files, but although I compiled
liquidsoap with faac/faad support, m4a files are handled by mad, who
fails at efficiently playing them, just passing to liq a short
high-pitched audio glitch
Hi Azrael,
This list is getting really interesting! I'll only write about clocks
first, there's a lot to say.
On Thu, Jun 24, 2010 at 1:25 PM, Kerozen kero...@gmail.com wrote:
Error when initializing source at line 99 char 15 - line 102 char 2:
cannot unify two dependent clocks: cross_5074[],
Yo,
On Thu, Jun 24, 2010 at 2:43 PM, Kerozen kero...@gmail.com wrote:
What I tried was pretty much this:
playlist_1 = rotate([musique,jingles])
playlist_2 = rotate([musique_2,jingles])
radio = switch(track_sensitive=true,[
({(2w) and (20h29-21h00)}, playlist2),
({ true }, playlist1)
])
On Thu, Jun 24, 2010 at 6:30 PM, Andrew and...@instantofficecenter.com wrote:
1) Is there any way to make liquidsoap shutdown if it gets disconnected from
an icecast server?
If you don't get a shutdown with an output.icecast that doesn't have
restart=true, then it might be a bug. Could you
On Tue, Jun 22, 2010 at 12:50 AM, Cees van Egmond
cees.vanegm...@gmail.com wrote:
I think I had the same error a few day's ago in the fail of network
connection-mailing.
There I've set the following outputs:
output.icecast(%mp3(stereo=false, samplerate=22050, bitrate=24),live)
Hi,
Welcome to our list! If you haven't subscribe, I encourage you to do so.
2010/6/21 karl klammer p.nidero...@gmail.com:
is there a way to force a playlist to load and start immediately?
Yes, you should use the server command playlist_id.reload when your
script regenerates the playlist. This
Hey,
2010/6/21 karl klammer p.nidero...@gmail.com:
thanks a lot for the fast answer!
yes, the reloading behavior from the svn-version you describe sounds like
the solution for this problem!
i just compiled from the trunk, but i'm not able to get this working as
expected :(
Thanks for the
If you're using the telnet interface to trigger the reload, then you
could create a new command as follows, to use instead of
playlist_id.reload:
# insert this after the definition of your playlist source,
# assumed to be named pl below
def reload_and_skip(arg)
Good news: SVN and tarballs work the same, especially if you built
from the -full tarballs.
See http://savonet.sourceforge.net/doc-svn/build.html, it's still valid.
David
--
ThinkGeek and WIRED's GeekDad team up for the
I confirm that smooth_add doesn't mess with time flows (clocks) so it
is supported. It just mixes two streams, but does not accelerate any.
What's not supported (i.e. may behave unexpectedly, and soon will be
impossible by default) is cross(), smartcross() and derived operators.
Crossing a track
Hi Andrew,
You've just bumped into a novelty from the SVN version. Usually they
go unnoticed, but not here it, and it deserves a little explanation.
On Mon, Jun 21, 2010 at 8:49 PM, Andrew and...@instantofficecenter.com wrote:
Is it possible to have an Ogg stereo stream and an MP3 mono stream
Yo,
On Sun, Jun 20, 2010 at 10:39 AM, Cees van Egmond
cees.vanegm...@gmail.com wrote:
Now I have and CPU (Atom Z510) load of 70% , previously 55%.
Could you show us your script? There are a few operators that we need
to optimize again, in particular conversions. Now, running two threads
and one
So, did you manage to get liquidsoap's output from darkice using jack,
and encoding it in AAC+?
It doesn't seem like a very simple way to go, but if it works that's
interesting.
David
--
ThinkGeek and WIRED's GeekDad
2010/6/20 Mike Owens mikeow...@gmail.com:
2010/06/20 13:44:54 [decoder:3] Unable to decode /tmp/program.mp3 as
{audio=2;video=0;midi=0}!
In your case I think this was because it's a mono MP3. With the SVN
version, conversions are not done implicitly by decoders anymore. If
your script asks for
On Tue, Jun 15, 2010 at 2:36 PM, kosnickx kosni...@gmail.com wrote:
i saw that frequently there are GET requests to liquidsoap
Where did you see that? Wiresharking a liquidsoap running harbor?
i believe it has to do with asking for metadata
so i was wondering
is there any way i can have
Any news on that issue? Liquidsoap does not require any specific
version, as far as I know.
2010/6/13 budi prakosa i...@deadmediafm.org:
ERR: jack_wrapper.c::JACK_Error(793) Cannot connect to server socket err =
No such file or directory
This would look like your jack server is not running
In the SVN version you have arbitrary references, so you could do:
boolvar = ref false
...
boolvar := true
...
In 0.9.2 and before, however, you can only achieve such a thing with
interactive.float() or interactive.string(), converting the boolean
to/from those types.
In case you don't know
Hey,
Thank you for the details. It all seems to make sense. I also agree
that liquidsoap should be a good solution for that kind of need. Now
we just need to figure out what's wrong here.
On Mon, Jun 14, 2010 at 1:25 PM, Patrick Bohnet h...@yahoo.com wrote:
I tried to rewrite the script, piece
2010/6/15 Scott Robinson scott.robin...@studentradio.org.uk:
Just a quick question regarding the liquidsoap telnet interface!
How can i go about telling it to switch to a different source?
The switch() operator has no builtin server commands. You can however
use an interactive.float() or string
Hi Patrick,
I don't see an obvious solution right away, but here are a few advices.
On Sat, Jun 12, 2010 at 8:19 PM, Patrick Bohnet h...@yahoo.com wrote:
i get skipping and this is my log (it should eb playing from the local
playlist)
http://pastebin.com/V5d9ar7z
Your log shows that
On Sun, Jun 13, 2010 at 9:35 AM, kosnickx kosni...@gmail.com wrote:
as for your suggestion
i believe (and correct me if i am wrong) that in this way every person
should be aware of the mount point now playing so that he/she can use the
other one (the free one) every time he/she decides to
On Sun, Jun 13, 2010 at 10:05 AM, kosnickx kosni...@gmail.com wrote:
well i just changed my script and i am waiting until its time to reproduce
the situation
i will let you know as soon as i have some results
...
well i just reproduced the situation (the time had to be xx:00) and it just
Hi Nick,
The way I understand your script, there is no restriction on when/who
can connect to your many input.harbor(), the control is only at the
level of the switch which assigns each harbor to its time interval.
It seems that all you need to do is to add track_sensitive=false to
your switch,
You used PACKAGES, which means that liquidsoap was configured to use
ocaml-faad/src, without checking that it actually contains a built
lib. Check that this directory contains a cma file, or better just run
configure and make there, see if there are any errors that you might
have missed when
On Thu, Jun 10, 2010 at 12:23 AM, alexander lunyov sol...@gmail.com wrote:
radio = fallback (track_sensitive=false,[live,music,single])
radio = skip_blank(radio)
radio = amplify(1.,override=replay_gain,radio)
Yes it's slightly silly but you need to add another amplify:
radio =
Hey,
2010/6/8 Cees van Egmond cees.vanegm...@gmail.com:
I've added one testing manner by stopping the icecast server manually after
1min and restarting after another 1min.
This is what I tried on my laptop (everything local, nothing virtual)
and it worked fine.
Both way's showed that
You need to amplify().
David
--
ThinkGeek and WIRED's GeekDad team up for the Ultimate
GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
lucky parental unit. See the prize list and enter to win:
Hi Cees,
I hope you don't mind if I Cc your feedback to the list. We should
stay on-list as much as possible.
On Mon, Jun 7, 2010 at 3:23 PM, Cees van Egmond
cees.vanegm...@gmail.com wrote:
The test has been done as follows:
0:00 start Liquidsoap on VM
1:00 disconnect virtual lan cable
2:00
On Mon, Jun 7, 2010 at 3:56 PM, Romain Beauxis to...@rastageeks.org wrote:
Our binding needs to be updated as soon as possible. In its current version,
it will segfault when using the new libao.
Good timing: my liquidsoap started segfaulting today (I did an update
of my debian yesterday) but
Hi Cees,
Here is how to adapt your script to SVN, using clocks to avoid that
when icecast lags, no output is read from the soundcard, causing a gap
in the backup.
On Sun, May 30, 2010 at 2:44 PM, Cees van Egmond
cees.vanegm...@gmail.com wrote:
set(log.file,true)
I am very interested in finding the cause of that problem. I will run
some tests next week (I'll travel back to the US soon and don't have
time right now). As you point out, the problem is more noticeable with
a slow CPU, and I happen to have a tiny server. I'll be able to test
liquidsoap SVN and
Sounds good to me, I'm impressed! You forgot to show us how you
increment count_jingle. I expected to see it within the transition
itself. Do you use on_track or similar instead?
--
Hi,
2010/5/23 budi prakosa i...@deadmediafm.org:
songs_crossfade = smart_crossfade(conservative=true, songs)
songs_jingle = cross(crossjingle(jingles), songs)
radio = rotate(weights=[1,4],[jingles, songs])
This cannot work: After you crossfade in the first line, your stream
is
On Sat, May 22, 2010 at 12:31 PM, Davit Barbakadze jaya...@gmail.com wrote:
Thanks, that was it :)
Also having compiled almost all libraries (except lastfm one - it gave
me an error, so I just removed it from PACKAGES file and rerun a
bootstrap just in case), I got error on final stage -
Thanks for letting us know.
In non-conservative mode, a crossfade will create a burst of
computation at the end of a file. This burst would have to be
abnormally big to stop the streaming long enough to trigger a cry
disconnection, so I'm still puzzled about what happened to you. (With
After you launch a liquidsoap daemon, you don't have any log. But do
you have a liquidsoap process? is it normally active?
I'm wondering whether does fork behave the same in BSD and Linux with
respect to threads (only fork the current thread or fork everything).
That being said, I don't see how
Hi,
On Sat, May 22, 2010 at 8:57 AM, Davit Barbakadze jaya...@gmail.com wrote:
Trying to build liquidsoap on fc12 and constantly getting configure:
error: libao not found when trying to configure ocaml-libao. Although
I got it installed: yum install libao and:
You usually need the dev version
In the liquidsoap source directory (tarball or SVN) you'll find
liquidsoap/scripts/ask-liquidsoap.pl / .rb, it's a script (available
in Perl and Ruby language) to execute liquidsoap commands via telnet.
For example, use [ask-liquidsoap id.queue] to see the content of a
queue, [ask-liquidsoap
This is a known but elusive bug for Ogg streams. For AAC+ I'd say that
it's no surprise: I believe the external encoder is restarted when new
metadata arises, which is obviously time consuming. For Ogg streams,
there's no good reason. I have recently made some experiments [1] to
observe precisely
FYI I have created a ticket for that feature request. I encourage
everybody to give suggestions about design and think about
applications:
http://savonet.rastageeks.org/ticket/352
--
It just occurred to me, after having used it for a long time, that the
name add_timeout is misleading.
2010/5/18 budi prakosa i...@deadmediafm.org:
add_timeout(10.,{ submit_lastfm(m);(-1.) })
This does not trigger submit_lastfm with a timeout of 10 seconds, but
triggers the function
By the way, the custom solution doesn't have to be a complete script
plugged into liquidsoap via request.dynamic(), it can be a simple
script / shell command used as a switch predicate:
switch([({ test_process(whatever shell test) }, source), etc])
Hi Seann,
Thank you for your message. Several people have asked for such as
thing. I'm willing to add that feature to liquidsoap SVN, namely
adding dates (day, month, year) to our scheduling predicates (I've
been reluctant because it doesn't fit with some current implementation
choice, but I
On Sat, May 15, 2010 at 11:26 PM, Mike Owens mikeow...@gmail.com wrote:
I've been looking at it more, and I can't prove the difference comes
from the two revisions. I don't know what it was. I ran it in under
daemontools sometimes, and by init.d (daemon mode) others, and
different versions of
Hi,
Thank you very much for the extra info: it seems that your problem is
due to a rare external reason, independent of the format.
On Tue, May 11, 2010 at 11:28 AM, Andrew and...@instantofficecenter.com wrote:
I have liquidsoap encoding an AAC+ and Ogg stream. Interestingly enough, the
AAC+
No time for a long answer: the SVN is radically different from the
0.9.x and earlier versions. Check out the doc
(http://savonet.sourceforge.net/doc-svn/external_encoders.html) and
the mailing list on how to adapt your scripts to the SVN version. Let
me just add that it's worth the effort: we want
This issue was indeed very easy to reproduce, and it's no surprise
once you know what the bug was. Anyway, it should be fixed with r7307.
--
___
Savonet-users mailing list
Hi,
This could be a valid bug report, but we have to figure out if it's a
general problem in our icecast client code, or if it's specific to
external encoders and in particular AAC (which I suspect). The use of
an external binary for encoding, whose output is then sent to icecast,
implies some
Hi,
On Fri, May 7, 2010 at 3:00 PM, Henry Favretto e...@gmx.ch wrote:
Well, I'm pretty sure what we see in the log is actually passed to
festival because I can clearly hear the voice saying backslash in my
example :) (the trailing numbers are being omitted however).
I have checked the source
Hi,
Thank you for the report, this is a bug. Here are the relevant lines:
let r = Unix.read fd s 0 len in
(* TODO: recursive read ? *)
assert (len = r) ;
You encountered a partial read, and that is not supported in the code.
Our bad, this module has not been tested enough. (Blame
I don't have time to develop a full example, but you should be able to
use a reference to a float (in liquidsoap SVN) or an interactive.float
(in =0.9.2) to keep track of the beginning of the current track.
Update it using on_track, and you can create a server command to query
it. I'm not sure if
I have just committed revision 7299, which brings the possibility to
mark primary queue requests as expired. In short, this is used after
reloading playlists, so the queue is flushed from its old requests as
fast as possible. Tell me how it works for you, I think you won't
even notice the as fast
On Sun, May 2, 2010 at 9:41 AM, Octavio Rossell octa...@gnu.org.ve wrote:
radio = add([blank(),radio])
radio = rewrite_metadata([(title,metadata)],radio)
This should not change anything. As I said, radio contains a stereo
stream, you add stereo silence, no problem. The only way to have your
Hi Cees,
I believe that your problem is due to the load and wait caused by
icecast reconnect attempts. Those things slow down the whole streaming
process, including the live input. If you interfere too much with a
live input it empties its buffer, or fails to fill it in time, which
results in
On Tue, Apr 27, 2010 at 6:07 PM, Romain Beauxis to...@rastageeks.org wrote:
In fact, upon configuring the package, if festival is not detected, false is
used instead.
I agree that it is not meaningful. Perhaps we could detect it at run time and
echo a error message if not found..
Yeah, it's
Hi,
You're not missing anything obvious, we'll have to figure out why the
script returns an error. The speech synthesis script is located at
$(libdir)/liquidsoap/$(libs_dir_version)/liquidtts where $(libdir) is
/usr/lib or /usr/local/lib and $(libs_dir_version) is 0.9.2 for you.
You could have a
Hi,
On Mon, Apr 19, 2010 at 12:04 AM, Welle303 Admin p...@welle303.de wrote:
I think I have found the reason: LIQ simply doesn't restart
when it is told to do so. I do think so, because the log doesn't
have any entry about restarting, when it should have neither
at 4:00 nor at 6:05). But when
Hi,
I think I understood your problem.
(1) The playlists don't reload as soon as you expect. The
one-reload-per-round means that liquidsoap won't reload before it
plays the last file of the round, which involves playing at least one
file.
(2) I looked up the source code, the message unresolved
On Wed, Apr 14, 2010 at 10:50 AM, PyroPeter abi1...@googlemail.com wrote:
But I really don't get why someone would require a certain user for
make-install?
The most important use of the liquidsoap user is to run liquidsoap
services not as root. As a result, we also need to install
Currently this list is only accessible though sourceforge (not really
great for browsing) and mail-archive.com (easy archives from the
beginning of the list). I have nothing against making it accessible
through gmane, and Sam said he would see how to register us. However,
I suspect he won't bother
Hi,
On Mon, Apr 12, 2010 at 12:20 PM, em...@mccurly.de em...@mccurly.de wrote:
is there a way to let multiple harbor clients get mixed together?
[...]
we do not want 2 harbor ports, or any portredirection.
Then no, there's no way. Liquidsoap can do anything with sources, but
if you don't want
Hi,
It's nice to start talking concretely about that interesting project!
As you're pointing out, the first question is to see whether
liquidsoap is the right tool, and I'll try to address this first. Then
I'll deal with how to schedule files, which is in fact unrelated to
video.
On Mon, Apr 12,
Sorry, I read we do not want to harbor mount points, which would
have been a problem. Two different harbor sources (on two different
mount points, on the same port) are no problem, and you can do
whatever you want with those sources, they are sources like any other.
(Note that with shoutcast
Hi,
Your script looks good. In fact it's really nice, with several cool
features, a pleasure to read ;) I don't see any obvious reason why the
playlist wouldn't start. You're right that it is very puzzling: if it
can play when you start liquidsoap, why can't it switch to it when the
time comes?
Hi Petter,
I'm not sure of what you want exactly, and where you are right now.
But let me confirm that we'd be happy to help you experiment with
liquidsoap SVN: getting the video in a production state is our goal
right now.
2010/3/19 Petter Reinholdtsen p...@hungry.com:
It seems liquidsoap
Hi,
2010/4/9 Аркадий Левин poison...@gmail.com:
How change info eg name, description etc.. on output.icecast.mp3?
Those parameters are static per connection (icecast says so, not us).
So you'd need to reconnect to change them anyway. Now, it is currently
impossible to disconnect, change the
Octavio wrote:
Now I need to put a random JINGLE playlist _after_ this 5 seconds and
_before_ playlist. It can be used to indentifying the radio or to
announce the incomming playlist.
This can be scripted pretty much as you said it :) Transitions are a
good place for such jingles. Create a
On Fri, Apr 9, 2010 at 4:47 PM, Octavio Rossell octa...@gnu.org.ve wrote:
sequence([blank(duration=5.),jingles,podcasts])
Why this will play only one track from the playlist?
That's what sequence does: plays only one track of each source, except
the last one. This was already happening before
Hi,
On Fri, Apr 2, 2010 at 4:58 PM, Luciano A. Ferrer lafer...@gmail.com wrote:
No, this is extremous fail. sudo? in build() ? Adding accounts,
also in build() ?
This PKGBUILD is absolutely inacceptable.
I strongly recommend not to use it.
If this software does not build without a
Sorry, the type system warns you that you're getting a string back
from server.execute (it should be OK in that case) and that you're
not doing anything with it. You must be explicit that you don't care
about it: add ignore(...) around server.execute(...) or put reply=
just before it.
On Tue, Mar 30, 2010 at 1:32 PM, Subflow.FREE isof...@free.fr wrote:
Only 1 file shouldn't be a problem. What isn't proper in the
behavior? it doesn't reload at all? I bet it reloads fine but too
late, so you get again the old file before hearing the change. If
so... I'm not sure if it can be
I haven't any experience with AAC streams (decoded with Faad). I'd
suspect that network errors would yield a different exception (you
could try to shutdown the network) but I'm not sure. If you can, also
try with another codec.
Also, are you using the SVN version? 0.9.1, 2?
HTH
--
David
2010/3/30 Subflow.FREE isof...@free.fr:
labomix = playlist(reload=10800, /home/subflow/labo.lst)
unfortunately this playlist don't reload properly
I guess it's because there is only 1 file in it.
Only 1 file shouldn't be a problem. What isn't proper in the
behavior? it doesn't reload at all? I
I confirm that I have tested the following solution:
def delay_fallback(a,b)
def immediate(a,b)
b
end
def delayed(a,b)
sequence([blank(duration=10.),b])
end
fallback(track_sensitive=false,[a,b],transitions=[immediate,delayed])
end
I ran
Hi,
This is an informal bug report and fix: I have noticed that
input.http() is broken on SVN -- more precisely, ogg stream decoding
is broken. It seems that a not-so recent modification was incorrect,
or interacted badly with something else since then. In any case the
following quick fix exists,
Hi,
On Wed, Mar 24, 2010 at 3:47 PM, Romain Beauxis to...@rastageeks.org wrote:
s = add([blank(),s])
The first line will add a blank source to s. The behaviour of the add operator
is to relay tracks and metadata from the first source only. Hence, the
resulting source will have only one track
I didn't understand at first that the metadata loss was wanted. Yes,
it works. It's just not as simple as it should be. One of my points
still holds: use add(normalize=false,[blank(),s]) to avoid that the
volume gets divided by two.
--
David
Hi,
2010/3/27 Octavio Rossell octa...@gnu.org.ve:
radio=fallback(track_sensitive=false, live,silence,music])
radio does not work: silence remains and no music on playlist starts.
You probably want to put silence last in the list, so that you music
playlist is chosen instead of it when it's
Hi mccurly,
Thanks for the update. This is the same problem as last time. The only
new bit of information that I noticed is the involvement of
smart_cross, which is a possible suspect for this kind of problem. I
don't have much time these days (until almost mid-April, I'm afraid)
to investigate
Hi,
On Sun, Mar 21, 2010 at 10:52 AM, kosnickx kosni...@gmail.com wrote:
Is there any way to reload the script without having to kill the
process
The short answer is no, for a number of reasons that we should maybe
put in an FAQ.
But the long answer is that you don't need it. If it's only for
Hi,
I'm sure which script(s) you used, and what you're trying to achieve,
but here are a few comments.
2010/3/13 Shashank Singh shash...@awaaz.me:
def smooth_add() =
d = 1. # delay before mixing after beginning of mix
p = 0.2 # portion of normal when mixed
end
This function takes no
On Mon, Mar 8, 2010 at 4:30 PM, Romain Beauxis to...@rastageeks.org wrote:
An issue with the new ocaml-cry module that replaces the shout library has
been recently discovered.
This issue was related to the data sent to the icecast server and could result
in a higher load and delay when
After some investigations on IRC with Maxwell, here are some conclusions.
The key is that Maxwell is running liquidsoap is a pretty constrained
environment, with only 200M allowed for a process, and liquidsoap
has a higher memory consumption than that (~500M). The bad::alloc
exception was raised
By the way, the restricted environment is called cpanel, if that can
be useful for future searches.
--
Download Intel#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
Hi,
On Thu, Mar 4, 2010 at 8:55 PM, Maxwell Rathbone
mrathb...@orbitalgrooves.com wrote:
So I tried to set it up to run in daemon mode by doing:
liquidsoap -d shoutcast.liq
When I run it like this, I get a command prompt, and I can tune in to my
stream.. but I hear nothing but dead air. I'm
Hi Maxwell,
I'm sorry to hear about your unusually bad experience. Let's try to
get it straightened out.
On Tue, Mar 2, 2010 at 10:56 PM, Maxwell Rathbone
mrathb...@orbitalgrooves.com wrote:
Took me about
half a day to finally get LiquidSoap to compile properly *without* the
following
Hi list,
Gabriel Afana has been using liquidsoap for quite some time. He runs
the DJ radio http://www.plurlife.com and developed a web-based
interface for it. Gabriel has recently cleaned up his code, called
Dromey, to make it available to the world at
http://www.dromey.com/source/dromey.zip. I
On Mon, Mar 1, 2010 at 9:31 AM, fruity fru...@freaknet.org wrote:
sorry to ask but, what's up with django-washtub?
No, it's actually a good question. I hesitated to talk about it, but I
felt this could wait a bit. Personally, I don't do much web
development: enough work on liquidsoap, and I only
On Mon, Feb 22, 2010 at 8:55 PM, Nick Coons m...@nickcoons.com wrote:
Does it support all of the various clients that Icecast does (i.e.
shoutcast, ices, darkice, etc)? If so, I think I'd like to pursue in
this direction.
Yes, input.harbor follows the same protocol so it should support all
On Sun, Feb 28, 2010 at 1:22 PM, Nick Coons m...@nickcoons.com wrote:
I would like to create commercials as a playlist that is generated
from a script, and regenerated each time it's queued up to be played.
The script will be responsible for returning the correct number of MP3s
to make sure
2010/2/23 em...@mccurly.de em...@mccurly.de:
the termination aggain happened when: a live harbor source has disconnected.
Thank you for the report. I still have no clue (and haven't ran any
test, which would be difficult since the crash occurs quite rarely)...
I'm puzzled about your mentioning
Thanks for the feedback. In the future, input.oss and other similar
sources will require their own clock to be used, i.e. the default
behavior will be the good one for you.
--
Download Intel#174; Parallel Studio Eval
Try
On Wed, Feb 24, 2010 at 9:58 AM, David Baelde david.bae...@ens-lyon.org wrote:
My feeling is that I won't call this story a bug but a funny behavior.
A funny behavior should be viewed as a bug, but here I meant
complex behavior
Hi Romaric,
I have (finally) tested your script, and I confirm the surprising
behavior on the SVN version. There are essentially two mechanisms in
liquidsoap that replay metadata. I had in mind the low-level one,
but I forgot the specific mechanism attached to switching operators
which seems to
wrote:
Ok, It seems to work as needed!
Many thanks
Romaric
-Message d'origine-
De : David Baelde [mailto:david.bae...@gmail.com]
Envoyé : mardi 23 février 2010 16:56
À : Romain Beauxis
Cc : savonet-users@lists.sourceforge.net; Romaric Petion
Objet : Re: [Savonet-users] Issue
On Tue, Feb 23, 2010 at 2:32 PM, David Baelde david.bae...@ens-lyon.org wrote:
I've opened a ticket containing my remarks from a previous mail. As I
was digging more into this issue I found out that the replay_metadata
feature seemed to be already implemented as desired. But you don't
skip
Hi Nick,
First, welcome to liquidsoap and this list!
On Sat, Feb 20, 2010 at 6:59 PM, Nick Coons m...@nickcoons.com wrote:
I'll be running a 24/7 talk station, which will have commercials
throughout. Generally, the shows will be anywhere from 1-3 hours in
length, and be interrupted for
4 seconds of buffer seem pretty high. I'd imagine that the catchup
plays a role here: if liquidsoap gets late, oss probably drops some
data (ring buffer too small), and when liquidsoap tries to catchup by
going super fast, oss either blocks him, or feed random data.
The conclusion is what Romain
801 - 900 of 1142 matches
Mail list logo
