On Sat, 2009-08-15 at 01:07 -0400, Tom Lane wrote: > I've been pursuing, with increasing frustration, the seemingly simple > goal of getting mysql to rebuild in rawhide since the mass rebuild. > What is frustrating me is that I can't reproduce the problem outside > koji where I might have a shot at debugging it. I had a similar experience with OOo. Worked locally, worked in mock, failed on koji. Turned out that one of the build tools created during the build used strcpy on overlapping source and destination pointers which on koji for some reason is far more likely to blow up than locally. Perhaps due to the different underlying kernels, maybe for some other reason, either way, perhaps attempting to run the final tests where it blows up on koji under valgrind locally might be helpful. C. -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
