I might be misunderstanding your question, but I created the original bug report (at
http://debbugs.gnu.org/cgi/bugreport.cgi?bug=14378) via report-emacs-bug, and included a repro from emacs -Q. I dumped an OS X console crash log that may be obscuring the information that report-emacs-bug includes, but it's there toward the bottom of the email.
I haven't done a backtrace with gdb before but when I get some time I'll look into it.
If I'm misunderstanding or there's more I can do to help, please let me know!
I'll include the repro here as well:
* Repro from emacs -Q:
0. Obtain magit. I'm using 924a56fe24c3c4acf55daa7266b2e9ad4fc30abb (master
HEAD as I write this), but I have also seen this happen with earlier
versions for several weeks.
1. (add-to-list 'load-path "/path/to/magit")
2. (require 'magit)
3. Open a file that is in a git repo. For example, I can repro using my own
.emacs.d, https://github.com/sandinmyjoints/.emacs.d. It's not very big
(184 commits).
4. Do command `magit-status`
5. Hit `l` for magit-log
6. Hit `l` for short log
7. Hit `q` to bury log.
8. Hit `q` to bury magit-status.
9. Repeat steps 4-8 until crash.
With my normal .emacs.d, I can repeat it only 2-3 times before crash. With
emacs -Q, it takes more like 5-8. Step 6 seems to get a little slower each
time it is run.
Thanks,
William