Christopher Kings-Lynne <chriskl(at)familyhealth(dot)com(dot)au> writes:
> Seems like a reasonable thing to submit.

I thought it would be a waste of time (both human and mechanical).
The existing usage of upper/lower in the regression tests is plenty
to detect basic problems.  As for advanced problems, the only
interesting aspect of upper/lower is locale-dependent behavior, which
the main regression tests are singularly unsuited to test for.

It might be interesting to expand src/test/locale/, but that stuff
needs far more work than just addition of upper/lower tests ...

			regards, tom lane

In response to

• Re: upper() / lower() regression test case needed? at 2004-06-04 06:44:34 from Christopher Kings-Lynne

pgsql-hackers by date

Next:	From: Greg Stark	Date: 2004-06-04 14:25:41
	Subject: Re: Converting postgresql.conf parameters to kilobytes
Previous:	From: Tom Lane	Date: 2004-06-04 13:43:53
	Subject: Re: [HACKERS] pg_dump --comment?