This is more for future reference in case anyone encounters this problem. I reported a bug <https://rt.cpan.org/Ticket/Display.html?id=80474> on 10/29. Tim Bunce has updated the bug report today with a possible fix that looks good to me (no segfaults on 5.12.3 or 5.16.2).
Matt Matt On Fri, Oct 26, 2012 at 3:44 PM, Matthew Musgrove <mr.musk...@gmail.com>wrote: > I didn't think that it would matter but it still happened after > downgrading DBI to 1.621. > > Here's a full backtrace. > > Matt > > > > On Fri, Oct 26, 2012 at 3:30 PM, Matthew Musgrove <mr.musk...@gmail.com>wrote: > >> Could this be related to bug 75614? >> https://rt.cpan.org/Public/Bug/Display.html?id=75614 >> >> Matt >> >> >> >> On Fri, Oct 26, 2012 at 3:24 PM, Matthew Musgrove >> <mr.musk...@gmail.com>wrote: >> >>> So once again we're seeing segmentation and general protection faults >>> when using DBI with threads. I've tested it with Perl 5.12.3, 5.14.3, and >>> 5.16.1. The backtraces keep pointing to DBI.xs line 3265: >>> >>> #0 0x00002aaaaaaddd8f in XS_DBI_dispatch (my_perl=0x2ed7310, >>> cv=0x3448528) at DBI.xs:3265 >>> 3265 if ((i = DBIc_DEBUGIV(imp_xxh))) { /* merge handle into >>> global */ >>> >>> Has anyone else seen this? >>> >>> I don't even know how to begin to recreate the failure outside of this >>> one daemon process. >>> >>> I'll try reverting to DBI 1.61 (I can't go further back that than or >>> I'll hit bug 77137 again) and see it is still an issue. >>> >>> Matt >>> >> >> >
