On Mon, 2013-06-24 at 19:23 +0200, Max Kellermann wrote: > On 2013/06/24 18:35, Alexander Sulfrian <alexan...@sulfrian.net> wrote: > > oh thats true. Maybe we want simply lock this region. I reworked the > > patch and add a static Mutex to the PulseOutput struct. So that the lock > > should affect all pulse outputs but nothing else. > > That's even worse: if one Pulse server does not respond, all others > have to wait, too.
During pa_context_new the pulse client does not connect to the pulse server. It only creates a few memory structures and load the client config. > > Hope that this is fine. If not please give me advice how to make the > > code thread safe. > > Don't use environment variables to pass data to a library. If > libpulse doesn't support that, we're out of luck, and your feature > cannot be implemented properly. Okay... Than I will open a bug report for libpulse. Currently the environment variables are the only way I see, because the client-conf API is not exposed in libpulse. Thanks anyway, Alex ------------------------------------------------------------------------------ This SF.net email is sponsored by Windows: Build for Windows Store. http://p.sf.net/sfu/windows-dev2dev _______________________________________________ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team