Good news and bad news...
Well, PostgreSQL does, in fact, work when compiled without optimisations.
That's the good news. I can post a regression report if needed.
Bad news is that something happens during packaging that prevents it from
working. I would love to guess the wrapper, but I don't think that's the
case anymore. I believe the fault is in the procedure to make it conform
to the FSSTD. I did notice quite a few errors during packaging, but
didn't think much of them since they occurred in the secondary packages
(Perl support, etc).
Basically, we can do one of many things here. I'm tempted to file a bug
against the package, BUT there is a new version on the horizon (slated for
7 July, if I remember correct). Plus, nailing this one down could be VERY
tough since I'm not sure why it's dying in the first place on the packaged
version. If I had to take a guess, I would say that it somehow
misinstalls the templates for new db's, thereby forcing postmaster to
exceed some memory bounds looking through/for a corrupted template.
Either way, if anyone NEEDS postgresql for the time being, I highly
recommend hand-building it and install it in /usr/local or something until
we get this worked out.
In the meantime, I'll bug the maintainer to see what he thinks of the
situation. If he can help us get this resolved somehow, someone should
send him some beer :-)
PS to Bart: I guess this means that we can use pgsql after all to work on
that organisation that we've been chatting about. It will take a bit more
work, but I'd be happy to send you some patches so you can build this