Justin Clift <justin(at)postgresql(dot)org> writes:
> Just compiled PG 7.3.1 on Solaris 8 x86, and it failed the GEOMETRY
> regression test with one of those "15th decimal place is one digit
> different from expected" types of errors.
I don't think we really care anymore about that; CVS tip has a modified
geometry test that suppresses low-order-digit differences.
If you care to try CVS tip on that machine, it'd be useful to know if
it passes cleanly.
regards, tom lane