[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [pygame] BUG: reproducible hard hang (i.e. SIGKILL only) in pygame!



Hey Steven thanks for the info. Do you have bitbucket access? Are you familiar with the new bugtracker?

On Apr 5, 2013 8:52 PM, "Steven Boswell II" <ulatekh@xxxxxxxxx> wrote:
Hello!  For the longest time, I've encountered a frequent hard hang (i.e. SIGTERM won't end the process, only SIGKILL) in pykaraoke.  Tonight I traced it to pygame.mixer.music.play().

I managed to write a short Python script, and create a short MP3 (i.e. 50k) that can reproduce it.  They're attached to this e-mail.
I also figured out that it only happens on multiprocessor systems; if I boot my computer with "maxcpus=1", or run it on a single-core machine, it doesn't happen.
On rare occasions, it'll succeed on a multiprocessor system, and the process will exit normally, but most of the time it'll hang hard.

I'm running this on Fedora Core 17, but I've encountered this regularly with previous versions of Fedora Core, too.
Currently, I'm running with the latest version of python and pygame available from the FC17 yum repos, i.e. python-2.7.3-7.2.fc17.i686 and pygame-1.9.1-11.fc17.i686 .
It happens with SDL_mixer built with smpeg as well as with libmad.  I'm using SDL_mixer-1.2.12-3.fc17 (also from the FC17 yum repo), built from source with "--enable-music-mp3-mad-gpl" added to the %configure line in the .spec file.
But given how long I've encountered this problem, I doubt the exact software versions are that relevant.

Please let me know if there's anything else I can provide to help track down this problem.  Hopefully the attached files are enough.
Thanks in advance!

Steven Boswell