[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: [PATCH] Don't use XFAIL for unwritten tests; also rm redundant test runs

From: Julian Foad <julianfoad_at_btopenworld.com>
Date: 2006-03-26 23:36:33 CEST

Ben Collins-Sussman wrote:
> On 3/26/06, Julian Foad <julianfoad@btopenworld.com> wrote:
>>"XFail" is intended to mean that a test is failing because of a known bug or
>>deficiency in Subversion. I noticed that some tests were using it as a way of
>>inserting a place-holder for a test which might one day be written. This patch
>>removes those place-holders.
>>There's one possible ill effect: the test index numbers of some existing tests
>>are changed by this patch.
> Why not just Skip() the test-stubs, rather than outright delete them?
> Wouldn't that be a lot less disruptive?

I don't think that having some of the test numbers change can cause a
significant problem. It's not like we ever don't know which version of the
test suite is being run.

If we think it is a problem, we ought to make a conscious decision and document
that we Don't Do That. (While I generally try to avoid changing the numbers of
existing tests, I believe I have, within the last few months, approved at least
one patch that inserted a new test between existing tests, and seen more than one.)

The output "SKIP: schedule_tests.py 1: schedule: add some files" implies to me
that the test was skipped because it was inappropriate on this platform for
some reason, not that the named test is in fact just an idea for a test that
doesn't actually exist. I think if we wanted a way of listing a place-holder
for a test that doesn't actually exist, we should invent something specifically
for the purpose rather than using "Skip".

- Julian

To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sun Mar 26 23:37:02 2006

This is an archived mail posted to the Subversion Dev mailing list.