Description of problem: A rebuild of MySQL in Rawhide (outside of mock) causes the following segmentation faults to me which looks *very* strange, because this is make test/check or so?! Mar 10 22:50:52 tux kernel: factorial[31459]: segfault at 00020b41 eip 0804adc8 esp bf97e190 error 4 Mar 10 22:50:52 tux kernel: factorial[31463]: segfault at 00020b41 eip 0804adc8 esp bfca64c0 error 4 Mar 10 22:50:52 tux kernel: factorial[31471]: segfault at 00020b41 eip 0804adc8 esp bffa87d0 error 4 Mar 10 22:50:52 tux kernel: factorial[31478]: segfault at 00020b41 eip 0804adc8 esp bff937b0 error 4 Mar 10 22:50:52 tux kernel: factorial[31485]: segfault at 00020b41 eip 0804adc8 esp bfb73b80 error 4 Version-Release number of selected component (if applicable): mysql-5.0.51a-1 How reproducible: Everytime, see above. Actual results: Segmentation fault during package rebuild Expected results: No segmentation fault during package rebuild Additional info: IMHO this bug report is somehow related with my bug #427707.
Interesting. I had seen those messages in my own machine's logs, but failed to connect them with building mysql. The problem appears to be that by default mysql builds everything (in particular libmysys.a) with -DDBUG_OFF, which means that dbug can't possibly work, and yet the build sequence tries to use dbug anyway as part of an example program that is run to provide sample output for use in the dbug manual. Since this manual isn't actually installed, the failure has zero effect on the resulting package. I suppose we could patch the Makefile to not bother building the manual, but it barely seems worth it. The "| cat" hacks in the makefile commands that invoke the sample program have no apparent functional use. I suspect they are there to prevent make from noticing the program failure, which means that somebody at mysql AB already kluged around this in a very obscure way ... Anyway, I'm afraid you're out of luck if you hoped this would lead to solving any other problems. It certainly isn't a blocker.
Changing version to '9' as part of upcoming Fedora 9 GA. More information and reason for this action is here: http://fedoraproject.org/wiki/BugZappers/HouseKeeping
Ping?
Fixed in rawhide.