16/06/2011 05:54:03 PM UTC, Boris Dušek: post #6774

Reported at

Contains backtrace using debugging symbols (I am attaching it), which shows the cause of abort being NULL value of pa_simple passed to pa_simple_write as first argument. pa_simple_write does an assert to make sure pa_simple is not NULL. This fails and hence the abort.

This is with 0.7.1-6ubuntu1 i386. Speech dispatcher logfile attached is attached in Launchpad, espeak.log was empty.
Attachment: Stacktrace.txt (2.337 kB)

16/06/2011 05:55:26 PM UTC, Boris Dušek: post #6775

Attaching stacktrace of all threads (previously attached has stacktrace only of the ABORTing thread).
Attachment: ThreadStacktrace.txt (5.312 kB)

06/04/2012 04:32:24 PM UTC, Joanmarie: post #13688

Is this the same bug as this downstream bug I reported?

I am seeing this occur very frequently with Orca.

06/04/2012 06:06:27 PM UTC, Boris Dušek: post #13690

I guess that is so, since in both cases the stacktrace is the same (except for missing spd_audio_play, but that probably just got optimized out) and the first argument to pa_simple_write in the crashing thread is NULL, and that is the cause of the crash since pulseaudio aborts on that (that parameter must never be NULL). Perhaps it's some initialization issue where in some rare code path (enabled by some unexpected scheduling order of threads etc.) the variable that should hold the non-NULL value is not set to such a value and retains its initial NULL value. But I have not found the cause.
Total records: 4
Note: You need to log in before you can post comments.