'Twas brillig, and Mike Dickson at 16/11/09 19:39 did gyre and gimble: > On Fri, 2009-11-13 at 15:37 +0000, Colin Guthrie wrote: >> 'Twas brillig, and Mike Dickson at 13/11/09 14:21 did gyre and gimble: >>> Hmm, ok, very interesting... >>> >>> The test code I have (and probably the Vivox SLVoice client which is the >>> real goal) opens an output stream first and then tries to open the >>> capture device. It looks like it initializes a "mainloop" for both >>> streams. Is that correct behavior or a problem? I'll do some debugging >>> around this and see if I can't ferret out a problem. >> There should only really be one mainloop for a given PA context. If they >> start two contexts (which means two connections to PA then there would >> probably need to be two mainloops. I'm not really sure tho' as I don't >> really know how the integration works. > > I've done some debugging around this and have more data. The code does > initialize 2 mainloops for the 2 channels. It opens the output channel > first and this is successful. The hang happens when opening the capture > session. I see an AUTHORIZING event come back and then it hangs. The > mainloop thread is still running. > > Interesting to note, if I switch the order and open the capture session > first it will hang in exactly the same way when opening the output > side. > > Any suggestion on how to debug what's happening in the server and coming > back to the openal library? While I'm not 100% sure that using two contexts from the same main loop is not supported, I don't really think it makes sense to do that anyway. I'd look at refactoring things so that one (static) mainloop/context is used. Col -- Colin Guthrie gmane(at)colin.guthr.ie http://colin.guthr.ie/ Day Job: Tribalogic Limited [http://www.tribalogic.net/] Open Source: Mandriva Linux Contributor [http://www.mandriva.com/] PulseAudio Hacker [http://www.pulseaudio.org/] Trac Hacker [http://trac.edgewall.org/]
