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

Re: svn commit: r29753 - in trunk/subversion/tests/cmdline: . svntest

From: <svnlgo_at_mobsol.be>
Date: Mon, 10 Mar 2008 10:44:56 +0100

Quoting Senthil Kumaran S <senthil_at_collab.net>:

> Julian Foad wrote:
>> Why should this command return a failure code on one RA layer and a success
>> code on another? It seems to me that Subversion is wrong, not the test.

>> We could say, "This is only a regression test, so it should test for the
>> current behaviour." I don't much like that: I would prefer either for it to
>> ignore the exit code, or preferably for us to fix Subversion right now to be
>> consistent.

Absolutely. That's why I prefer to refer to our python test suite as "acceptance
tests", as testing for regressions is only one aspect of what they're supposed
to test.

> C. Michael Pilato wrote:
> >> If we do what your patch does, then please insert a comment saying
> >> "### This is merely testing for the current behaviour of exit codes,
> >> not the desired behaviour."
> >
> > Or perhaps mark the test as XFail() for particular RA layers?
> I ve marked these tests as XFail() on 'ra_dav' in r29819.
I'd rather have we fix the issue instead of silencing the test like this,
especially for the unlock_wrong_token test that we already had in 1.5 and
passes there.

The XFail wrapper means the test can now fail for any reason, not only because
of an incorrect exit code. We might introduce a new issue in the lock code
which we'll never notice this way.

I'll try to have a look tonight.


This message was sent using IMP, the Internet Messaging Program.

To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-03-10 10:45:11 CET

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