Comment 3 for bug 986774

Since pasuspender seems to handle manually sent SEGVs to mumble, there does not seem to be a straightforward way to get apport-retrace of this...

Backtrace: /usr/lib/libg15daemon_client.so.1(+0xd59) [0x7f292c3a3d59]
Backtrace: /lib/x86_64-linux-gnu/libpthread.so.0(+0xfcb0) [0x7f292aa4ccb0]
Backtrace: /lib/x86_64-linux-gnu/libpthread.so.0(pthread_cond_wait+0xc4) [0x7f292aa48d84]
Backtrace: /usr/lib/x86_64-linux-gnu/libQtCore.so.4(_ZN14QWaitCondition4waitEP6QMutexm+0x1bb) [0x7f2928ef34db]
Backtrace: /usr/lib/x86_64-linux-gnu/libQtCore.so.4(_ZN7QThread4waitEm+0xa8) [0x7f2928ef2be8]
Backtrace: mumble() [0x5b17f2]
Backtrace: mumble() [0x5b1829]
Backtrace: mumble() [0x520d75]
Backtrace: mumble() [0x43eb19]
Backtrace: /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xed) [0x7f29282cb76d]
Backtrace: mumble() [0x442fc9]
End of Backtrace.

TEST CASE:
- configure mumble to use alsa for sound input & output, text-to-speech off
- run "pasuspender mumble" and choose quit from the menu

Expected result: program quits and you get back to shell prompt.
Actual result: mumble gets in an infinite loop

Reproduced only with alsa, oss not tested. Does not occur when both input & output are set to pulseaudio and mumble is ran without pasuspender.

Additional notes:
- run "killall /usr/bin/speech-dispatcher" between repeated test runs to return to "pristine state"
- (another?) infinite loop also occurs if you try to change both input and output at same time back to pulseaudio, doing the change one at time avoids that infinite loop.