On Wed, Aug 28, 2013 at 12:09:05AM -0400, Peter Eisentraut wrote:
> On Tue, 2013-08-27 at 09:44 -0400, Tom Lane wrote:
> > Boszormenyi Zoltan <zb@cybertec.at> writes:
> > > When adding regression tests, can you please add intentional
> > > syntax error cases to exercise all the new ereport()s?
> >
> > Please do not add test cases merely to prove that. Yeah, you should
> > probably have exercised each error case in devel testing, but that does
> > not mean that every future run of the regression tests needs to do it too.
>
> I disagree. The next person who wants to hack on this feature should be
> given the confidence that he's not breaking behavior that the last guy
> put in.
+1. I wouldn't make full error-outcome test coverage a condition of patch
acceptance. However, when an author chooses to submit high-quality tests with
that level of detail, our source tree is the place to archive them. I share
Tom's desire for a Makefile target that completes quickly and checks only
those behaviors most likely to break, but not at the cost of letting deep test
coverage dissipate in a mailing list attachment or in the feature author's
home directory.
--
Noah Misch
EnterpriseDB http://www.enterprisedb.com