On Debian when only "pulseaudio-utils" package is installed but not "pulseaudio" Xpra logs the following on attach:
2014-08-20 20:30:18,888 signal_safe_exec(['/usr/bin/pactl', 'list'],None) stdout='' 2014-08-20 20:30:18,888 signal_safe_exec(['/usr/bin/pactl', 'list'],None) stderr='Connection failure: Connection refused pa_context_connect() failed: Connection refused ' 2014-08-20 20:30:18,895 signal_safe_exec(['/usr/bin/pactl', 'list'],None) stdout='' 2014-08-20 20:30:18,895 signal_safe_exec(['/usr/bin/pactl', 'list'],None) stderr='Connection failure: Connection refused pa_context_connect() failed: Connection refused ' 2014-08-20 20:30:18,901 signal_safe_exec(['/usr/bin/pactl', 'list'],None) stdout='' 2014-08-20 20:30:18,901 signal_safe_exec(['/usr/bin/pactl', 'list'],None) stderr='Connection failure: Connection refused pa_context_connect() failed: Connection refused ' 2014-08-20 20:30:18,909 signal_safe_exec(['/usr/bin/pactl', 'list'],None) stdout='' 2014-08-20 20:30:18,909 signal_safe_exec(['/usr/bin/pactl', 'list'],None) stderr='Connection failure: Connection refused pa_context_connect() failed: Connection refused '
Those warnings appears even if "speaker" and/or "pulseaudio" are disabled in "xpra.conf" as well as when I pass "--no-pulseaudio" in command line. This did not happen in 0.13.9...
I never thought of testing something like this! (how does it even work?)
r7363 fixes this for me - will backport.
Can you confirm?
I did not test speaker forwarding on 0.14 yet (I'll open another bug if there are any problems). I only noticed errors in console hence opened this bug...
Fortunately with 0.13.* "pulseaudio" is not required on client for speaker forwarding. I'll test audio forwarding on 0.14.1 soon.
r7363 seems to fix this issue, thank you.
That's right, we need pulseaudio server side for sound capture on *nix, on the client side we can use any of gstreamer's native plugins (alsa, oss, pulseaudio, ...)
Backport was in r7367. Closing!
FYI: I'm not saying the sound is unbreakable, but it does get tested a lot... crashes would have been spotted and reproduce by now, there must be something fishy going on.
this ticket has been moved to: https://github.com/Xpra-org/xpra/issues/648