As NCC’s alpha geek, I never turn down a challenge and I’ve got an interesting one this week. I have to take a collection of perl scripts designed to run on a Linux/Apache system and get them running on a Dell Laptop running Windows NT. The preferred method of simply defenestrating the Laptop and installing Linux isn’t acceptable because it will be used in demostrations requiring the use of a Netscape Navigator client that looks like the Windows version. In particular the drop-down boxes as rendered by Motif seem to confuse people who’ve never used a computer with a real OS. So the only choice appears to be running the whole system on NT.
Yesterday, I did more experimenting with PostgreSQL. First I upgraded to the most recent release to fix a problem with pgaccess, the tcl-based GUI provided with the package. The eventual goal is to set up an interface between Apache and PostreSQL via mod_perl. Unfortunately, it appears that the mod_perl binary is defective in the Red Hat 6.0 for Sparc release. I installed it and then configured Apache to use it in the usual way – which consists of adding two lines to httpd.conf to load the modules and then uncommenting the mod_perl stuff in srm.conf. Apache segfaulted on startup. I grabbed the updated RPMs from the Red Hat ftp server that look like they’ll be used in the 6.1 release but the results were the same. Since we don’t plan on running Apache and mod_perl on the Sparc in the production system, I decided to stop short of attempting a comlete rebuild from the source RPMs. But if anybody has any insight into what’s causing Apache to dump core on startup, I’d be curious.