On Thu, 26 Jun 2003, Cosimo Streppone wrote: > Hi Rudy, > > I saw your last post on dbi-dev regarding your request of a > backtrace file (25/06 04:26). I already received such a post and I replied > to you on 25/06 10:56. >
I know. My email has been a bit screwey this past few days -- Changed DNS servers and all email from/to the lists got backed up, so the message I sent to you is just making the list today.... I am getting confused here just trying to keep incoming email straight right now :-(. > I'm reporting it below (see dbi-dev): Thank you, again. > So I did further investigations. > I recompiled my perl 5.6.1 from same clean source package (no rpm), > enabling the -DDEBUG switch at config time (I thought it was not enabled in > my previous perl version). > > Installed DBI 1.35, DBD::Pg 1.30_2 and everything worked! > Note that `perl -V' output is the same as before, nothing changes except > the compile date. If relevant, `ldd' on the two executables shows no difference. > > Everything else is same as before, postgres db version (7.3.3), > dbi (1.35), dbd-pg, libs, ... > I can reproduce the error simply switching the symlink from `/opt/perl-5.6.1-new' > to `/opt/perl-5.6.1', where i have the two installations of perl. > At this point, I don't know if more investigation is useful. I am sure that there is a bug somewhere, but it seems to be fairly isolated; I am just worried about what happens when people try to use this in production envionment. We need a release, but I have seen this too, I can get a segfault on some compiles, but for me it only shows up in clean-up after the croak() for SQL_BINARY not being supported. Tim, Any thoughts on this? > > I have also a (failed) make test report on 5.8.0. > I don't know if this can be useful... Maybe... But it would be nice to try and fix this before a release, anyway. Rudy
