Fabien COELHO <coelho@cri.ensmp.fr> writes:
> I agree that using TAP test if another simpler option is available is not
> a good move.
> However, in the current state, as soon as there is some variation a test
> is removed and coverage is lost, but they could be kept if the check could
> be against a regexp.
I'm fairly suspicious of using TAP tests just to get a regexp match.
The thing I don't like about TAP tests for this is that they won't
notice if the test case prints extra stuff beyond what you were
expecting --- at least, not without care that I don't think we usually
take.
I've thought for some time that we should steal an idea from MySQL
and extend pg_regress so that individual lines of an expected-file
could have regexp match patterns rather than being just exact matches.
I'm not really sure how to do that without reimplementing diff(1)
for ourselves :-(, but that would be a very large step forward if
we could find a reasonable implementation.
Anyway, my opinion about having TAP test(s) for psql remains that
it'll be a good idea as soon as somebody submits a test that adds
a meaningful amount of code coverage that way (and the coverage
can't be gotten more simply). But we don't need a patch that is
just trying to get the camel's nose under the tent.
regards, tom lane