abrt 1.1.1 detected a crash.
Attached file: backtrace
cmdline: opencv_haartraining -data trainout -vec Human.Buttocks.Nude.vec -bg Human.Buttocks.Nude.neg -nstages 22 -nsplits 5 -maxtreesplits 12 -minhitrate 0.98 -maxfalsealarm 0.5 -npos 769 -nneg 6594 -w 26 -h 26 -mem 1024 -mode ALL -minpos 32.0417 -nonsym
reason: Process /usr/bin/opencv_haartraining was killed by signal 11 (SIGSEGV)
release: Fedora release 13 (Goddard)
It appears that this may be the same bug I reported before. It was "fixed" by setting the library to be single threaded (no OpenMP). I have been using that version (provided as a test) and changed to the mainline Fedora one today. The crash is there and OpenMP is back on.
I cannot say whether or not this has anything to do with my other bug reports for opencv package.
How to reproduce
1. Start the trainer
Created attachment 425014 [details]
One of my five trainers is running, the other three aren't. It seems the ones
that aren't have splits off the 0 node. The one that is didn't do a split until
later in the cascade.
Thank you again.
Hi, and sorry for the late reply.
Can you test with newer release (either OpenCV 2.1 from fc14 or OpenCV 2.2 from fc15 ?).
Can you attach/reference the files used to test opencv_haartraining.
*** Bug 572451 has been marked as a duplicate of this bug. ***
I am afraid that at the moment I am unable to duplicate this in F15 (I will keep trying). Also, the data is about 2 gigabytes of data. Not something you want in bugzilla.
Note that there is still an opencv update that request feedbacks in testing.
(backporting important bugfix)
I'm closing this bug, please reopen if the proble appears for f15.
Thx for the report.
Thank you for fixing all of this.