Jakub Jelinek <jakub@xxxxxxxxxx> writes: > On Thu, Apr 09, 2009 at 04:26:09PM -0400, Tom Lane wrote: >> Apparently there's more than one problem with glibc. I was complaining >> about the "experimental malloc" code, which really now was not the time >> to enable. But based on Ulrich's last remarks there may be independent >> issues on i586 arising from the platform selection itself. > If you have issues with the new malloc code, first verify you aren't > allocating anything from within async signal handlers. Hmm ... I'm not anywhere near familiar enough with the mysql code to swear on the spur of the moment that they don't do that; except that surely any such code would be broken on most platforms, not only bleeding-edge glibc. In any case I think it's fair to ask what the heck we were doing swapping in experimental code in a core library post-beta. regards, tom lane -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
