On Sat, 02.05.09 10:59, Erich Boleyn (erich at uruk.org) wrote: > Clearly this is broken behavior. Indeed. In your case it is. > When looking through the code in 'module-rtp-send.c', I see the following: > > pa_source_output_new(&o, m->core, &data, PA_SOURCE_OUTPUT_DONT_INHIBIT_AUTO_SUSPEND); > > > Now, I'm thinking of just removing that line in 'module-rtp-send.c', > but the real questions are: > > Why is that code there in the first place? It clearly breaks my > case of using an audio source. It does. The reason I added this flag for the RTP streams is for a different use case: when you connect the module to the monitor source of a null sink with the purpose that this null sink then becomes a dedicated RTP sink you can play your music to. In this case it is a good idea if the sink (and hence the monitor source and hence the RTP stream) can be suspended when no real application is actually connected to the sink. Seems I didn't think about your uses case here. Stupid me ;-) We probably should make that flag controllable via a module parameter, which defaults to off. I must admit that my use case is less obvious than yours, so your desired behaviour should be the default. Hmm, could you please file a bug about this so that I don't forget about this? (Even better: attach a ready made patch! ;-)) Lennart -- Lennart Poettering Red Hat, Inc. lennart [at] poettering [dot] net http://0pointer.net/lennart/ GnuPG 0x1A015CC4
