Hi Tom, Thanks for doing these valgrind runs:
On 24-Mar-08, at 10:22 PM, Tom Jennings wrote: > Oops, my bad, send huge post. Here 'tis as urls. > > The three identical runs: > http://wps.com/temp/mixxx-svn-1895.8912 > http://wps.com/temp/mixxx-svn-1895.8955 > http://wps.com/temp/mixxx-svn-1895.9025 > > With change to path, including mixxx stderr: > http://wps.com/temp/mixxx-svn-1895.9088 > http://wps.com/temp/mixxx-svn-1895.9088.mixxx-stderr > > .mixxx.cfg > http://wps.com/temp/mixxx.cfg > > > So I ran recent mixxx under valgrind, and got some odd behavior, > but also some big (megabyte) leaks. > Some of these big leaks are probably occurring at exit. I haven't seen Mixxx leak a crippling large amount of memory in ~2 hour sessions (not the best vote of confidence though). Ideally, we should try to be as paranoid as possible about memory leaks in the library and reader/mp3/ ogg code. > I'm sudo'd, else mixxx can't read skins (I know I should find > that...) > > $ valgrind --log-file=mixx-svn-1895 mixxx > > Oddly, when I do this mixxx comes fully up, then immediately > quits "normally" eg. Control-Q clean exit. As soon as the GUI > is fully drawn, the dialog "writing song database" appears, > then it quits. Yeah, this is the same problem I have. I thought it was something I was doing wrong. Does this happen every time for you, or is there something you can do to mitigate it? We need a valgrind run where a few songs are loaded and played so we can detect memory corruption inside the mixing engine... > > > To introduce a known change, I changed my collection (in > .mixxx.cfg) to a subdir of what it was (/home/tomic/Music to > /home/tomic/Music/Psy). I got the 'scanning collection' dialog, > and then immediately the "writing song database" dialog, which > terminated mixxx before the scan completed. mixxx stderr output from > that run attached. > > [EMAIL PROTECTED]:~$ uname -a > Linux oOo 2.6.22-14-rt #1 SMP PREEMPT RT Tue Feb 12 09:57:10 UTC 2008 > i686 GNU/Linux > > I'll attach my .mixxx.xfg also though I doubt that's useful here. > > Four runs, in PID order, three identical environments, the last > the path change above. These runs give me something useful to get started with, thanks again. If we can figure out how to get Mixxx to not exit after it starts under valgrind, then we should be able to get even more useful data. Thanks, Albert ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Mixxx-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/mixxx-devel
