On Feb 20, 2010, at 11:03 AM, Bruce Momjian wrote:
>> There is also the "must fix" issue with pg_regress.
>
> Why? My pg_regress runs just fine. If you are talking about 3rd party
> modules, I suggested conditional Makefile rules.
Because you can either make the simple change to pg_regress that David Fetter sent yesterday and have things continue
towork, or you can break a slew of third-party module test suites (and possibly modules, as well) and make a lot of
otherpeople do a lot more work, not to mention email -hackers and ask WTF happened because they may well not know.
I think that not changing pg_regress is more work for third-party module maintainers *and* more work for the Pg
communitywhen those maintainers come asking what happened and for advice on how to fix it.
Best,
David