From: Kris L. <te...@ur...> - 2009-01-26 15:08:58
|
David Baelde wrote: >Indeed, these messages are harmless as long as they don't appear too >often. Even then, they just say that ALSA does not get data fast >enough to produce a clean continuous output, which has no effect at >all on the icecast output -- which probably matters most for you. >Concerning the silly ID src_XXXX, I noticed it on unbuffered ALSA >outputs by default. I didn't bother to fix... I should stop being >lazy. However, if I set the id explicitly output.alsa(id="foo",..) I >don't get them anymore. Could you tell me what version of liquidsoap >you have and how exactly you setup you ALSA input/output(s)? Thanks that fixed it fine. >You can execute arbitrary code in the transition. Suppose we have a >fallback as follows: >fallback([foo,bar]) > >This is the same as: >def f(a,b) > sequence([a,b]) >end >fallback(transitions=[f,f],[foo,bar]) > >Now if we want a notification, let's write in the logs before building >the compound source: >def f(a,b) > log("transitioning") > sequence([a,b]) >end >fallback(transitions=[f,f],[foo,bar]) >In your case, you might have a more complex transition already setup >in your fallback; you can edit it in the same way. Instead of log() >you can use system() to call a script of yours. >Finally, I'll show how to have different notifications without >duplicating the code: >def f(s,a,b) > log(s) > sequence([a,b]) >end >fallback(transitions=[f("foo"),f("bar")],[foo,bar]) >By partially applying arguments to f(), we specialized it to print >different things. The logging part of this works fine. However calling sequence seems to be giving an issue. As sequence seems to be track sensitive it waits for the end of a track before skipping back to the ALSA input. In fact it doesn't seem to come back from Last.FM at all. Any work around to this? |