Subject: Re: PostgreSQL weirdness and C++
To: None <port-pmax@NetBSD.ORG>
From: Jonathan Gapen <jagapen@students.wisc.edu>
List: port-pmax
Date: 04/16/1998 15:57:38
Le 15-Avr-98, Jonathan Gapen a =E9crit:

>    I've built the latest PostgreSQL (6.3.1) on my pmax, but....

    Just a note:  I had compiled a few kernels before adding the SysV
compatibility options to the configuration.  After running 'config' and '=
make
depend', the correct objects were built and linked into the new kernel, b=
ut
apparently some other object file(s) that required a change were not re-b=
uilt.
 Zapping re-building from scratch fixed matters, and now both xanim and
PostgreSQL work with shmget and semget, respectively.
    On another note, PostgreSQL fails all the regression tests that requi=
re
dlopen(), as that call fails.  (And dlerror() doesn't say why.)  Are thes=
e
calls supported?  If so, what might cause them to fail?