patch for subunithelper
jelmer at samba.org
Sat Nov 20 12:21:51 MST 2010
On Sat, 2010-11-20 at 22:11 +0300, Matthieu Patou wrote:
> On 20/11/2010 21:25, Jelmer Vernooij wrote:
> > On Sat, 2010-11-20 at 13:15 +0300, Matthieu Patou wrote:
> >> On 15/11/2010 02:29, Jelmer Vernooij wrote:
> >>> Hi Matthieu,
> >>> On Sun, 2010-11-14 at 14:59 +0300, Matthieu Patou wrote:
> >>>> can you review this patch it's design to make filter-subunit.py to exit
> >>>> with a non zero error code if we have only 1 testsuite who has failed
> >>>> but without any failed test (happens if the initialization of the
> >>>> testsuite has failed, ie. because a module is missing).
> >>>> An example of subunit output that shows this combination is also
> >>>> attached to the email.
> >>> I think in this case we should just change the error code if there was a
> >>> failed testsuite without failed tests, rather than making up nonexistant
> >>> tests. Creating tests will cause havoc for test runners, since they will
> >>> not be able to request this test be run manually.
> >> Is this version of the patch better for you ?
> >> I tried not to pollute the test output with fake/non existing errors.
> > This seems to be the wrong patch.
> Oups !
> Try this one.
My comments from last time still apply. Please don't increment the test
error count in this situation, it's for tests, not testsuites.
This patch only seems to change testsuite-failures to testsuite-errors.
How is that fixing the problem you're seeing? Shouldn't we simply change
the code to exit(1) instead of exit(0) if any testsuite-failure or
testsuite-error's were seen?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 836 bytes
Desc: This is a digitally signed message part
More information about the samba-technical