[rspec-users] Running specs in reverse

David Chelimsky dchelimsky at gmail.com
Sun Nov 11 16:56:10 EST 2007


On Nov 11, 2007 3:50 PM, Scott Taylor <mailing_lists at railsnewbie.com> wrote:
>
>
> On Nov 11, 2007, at 12:47 PM, David Chelimsky wrote:
>
> > On Nov 11, 2007 11:01 AM, Alvin Schur <a.schur at nucleus.com> wrote:
> >>
> >>> On Nov 08, 2007, at 6:07 pm, Alvin Schur wrote:
> >>>
> >>>
> >>>> My goal is to detect inter-dependencies sooner than later...
> >>>>
> >>>
> >>>
> >>> I have to say I've NEVER run specs backwards.  Am I sitting on a
> >>> time
> >>> bomb?  Are there subtle traps that can create inter-dependencies
> >>> between specs?  To look at my specs I would not assume that running
> >>> order matters - I don't think I've ever used before(:all) for
> >>> example.
> >>>
> >>> Ashkey
> >>>
> >> Our team uses Test::Unit and numerous inter-dependencies have
> >> crept in.
> >> Our team is also slowly moving to rspec.
> >>
> >> Part of resolving the inter-dependencies is:
> >>
> >> 1) education
> >> 2) early detection
> >>
> >> I don't want to be the "dependency police" particularly if rspec
> >> can do
> >> the job automatically for everyone on the team.
> >
> > Well - I don't think rspec should do this implicitly - that might
> > cause quite a bit of confusion for unsuspecting devs trying to debug.
> >
> > I command line arg like --toggle_reverse might work - but you'll have
> > to submit a feature request in the tracker and probably a patch if you
> > want to see it any time soon. Very low priority on my list.
>
> I really like this idea, especially having had the following experience:
>
> My coworker had no failing tests with rake spec, but I was having
> failing tests with Autotest.  He had introduced something into before
> (:all) which stored some data into the database.  Oddly enough - the
> tests weren't failing for him, although they were failing in reverse
> for me- showing a dependency between the specs.  The worst part is
> that he was convinced this was a non-problem (because they were
> passing for him).
>
> I would implement this patch, if I had some good idea where I should
> store the state of the last test run (without adding an extra text
> file to my project).

You've got to maintain state between processes. How else would you do
it besides a text file?

>
> Scott
>
>
>
>
>
> _______________________________________________
> rspec-users mailing list
> rspec-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rspec-users
>


More information about the rspec-users mailing list