Tom Lane <tgl@sss.pgh.pa.us> writes:
> *Should* there be a default location established by configure?
I think so.
> I think I'd prefer that to having the dependence on an environment var
> ... but on the other hand we couldn't really remove the envar without
> breaking a lot of installations...
Keep the environment variable and let it override the default.
The test suite could use a sane default for PGLIB too. It seems
strange to compile and install postgres into a default location and
then have `make runtest' fail due to an environment variable not being
set. All the more so because the README file doesn't mention it.