[pulseaudio-discuss] Pulse Jack module, dies sometimes on mpc toggle

[Ng Oon-Ee]

Colin Guthrie wrote:
> 'Twas brillig, and Ng Oon-Ee at 30/04/09 08:27 did gyre and gimble:
>> Hi Lennart, yes it is a segfault. I can reproduce it regularly here 
>> on my laptop, will try in my office desktop when I get to work in a 
>> couple of hours time. Basically I start up pulseaudio (just with 
>> pulseaudio -vvvvv > pulselog.txt 2>&1), then my script uses pactl 
>> list to get a list of modules (checking whether alsa is loaded, in 
>> this case it isn't since its commented out from my default.pa), then 
>> starts up the Jack server with jackd and calls pactl load-module on 
>> both module-jack-sink/source. After that, I run paplay 
>> /usr/share/sounds/alsa/Rear_Center.wav, and pulse dies AFTER playing 
>> the sound. I include with this email my gdb stack trace, at the end 
>> of which I'm left in gdb and quit using q, not sure what the *etc, 
>> etc* in the wiki means.
>
> You'll need to install the debugging symbols for your pulse build. If 
> you build from source, simply don't strip the files (should be 
> default) of if you use a distro package they (typically) provide a way 
> to install these as separate "debug" packages.
>
> Col
>
>
Okay, I should have thought of that myself, sorry. Will redo it later, 
for now I've reached the office and have a log of a working 
configuration, where the same behaviour does NOT trigger a segfault. I'm 
including two files, pulselog.txt and pulselog2.txt. Both were obtained 
using `pulseaudio -vvvvv > pulselog.txt 2>&1`, followed by loading of 
module-jack-source/sink and paplay of a sound file. In the case of 
pulselog2.txt (the working configuration, run on my office desktop), I 
have done paplay twice, then multiple `mpc toggle` calls, something 
which also triggers segfault on my laptop.

Will revert when I have proper gdb stacktrace.