On Fri Aug 02 05:20:56 2013, ETHER wrote:
Show quoted text> On Thu, Aug 01, 2013 at 04:03:48PM -0400, Kenichi Ishigaki via RT
> wrote:
> > <URL:
https://rt.cpan.org/Ticket/Display.html?id=86952 >
> >
> > Ah, finally got it. It's not Test::Kwalitee itself but tests for your
> > versions of Test::Kwalitee that have been broken.
> > Test::Kwalitee 1.02-1.11 also work fine. Only their tests are broken
> > due to the recent changes (and will be broken again if newer versions
> > of MCA change something).
>
> No, it's not just TK's tests that are failing. The issue that this
> ticket
> was originally addressing was indeed problems with the metrics that TK
> was
> expecting, vs. what MCA was actually providing (you opened a few PRs
> in
> github regarding those).
>
> > Fix your fragile tests. Test::Kwalitee 1.01 (by chromatic) still
> > works fine (and passes its tests) with the latest MCA.
>
> v1.01 still passes its tests because it wasn't really testing anything
> :)
> I don't think it would pass its tests against MCA 0.88 because of the
> number of metrics that were added and removed (again what you referred
> to
> in the github PRs).
>
> I'll be further amending TK's tests to be more flexible as to what MCA
> provides, but it needs to make *some* assumptions regarding what
> metrics
> are available, so it can check that some metrics actually ran
> properly.
TK shouldn't expect anything. Partly because Module::CPANTS::Kwalitee is pluggable, and partly because it is not TK who should decide which metric is core and which is extra etc.
Anyway, I reject this ticket because MCA should continue to work with all the installed versions of TK. If MCA 0.88+ doesn't work with installed versions of TK, that's a bug and should be fixed.
It's another story that some of the older versions of TK now fail their tests. That's their issues, and is now (hopefully) fixed in the latest version. If someone does need to install older versions of TK for whatever reasons, skip tests, and it should work.