I decided to look at some system statistics (DISK IO to be specific), so I
reran the regression test on my RH5.2 box. I was surprised when it failed
after createlang was called for plpgsql. It should be a quick fix for Jan.
Either fix createlang to exit with a 0 status on language already installed:
143c143
< exit 1
---
> exit 0
Or fix it to exit with some other status and test for that status in the
Regression GNUMakefile.
Hope this helps,DEJ