Re: Test.pm & t/force_todo refactoring idea

On Apr 22, 2005, at 11:20 AM, Andy Dougherty wrote:
> On Fri, 22 Apr 2005, Nathan Gray wrote:
> > On Thu, Apr 21, 2005 at 02:24:29PM -0400, Stevan Little wrote:
> > > On Apr 21, 2005, at 12:00 PM, Nathan Gray wrote:
> > > > I don't quite understand the global/env variable idea.  Can you 
> > > > explain
> > > > that in more detail?
> > >
> > > It could be something as simple as this:
> > >
> > > %*ENV<TODO_ALL_FOR_RELEASE> = 1;
> > >
> > > It is really just a hack to allow for quick TODO-ing of all failing
> > > tests in the suite. I am not sure how well it will work out to be
> > > honest.
> >
> > So if TODO_ALL_FOR_RELEASE is true, we just have 'proclaim' treat all
> > failures as TODO.  Sounds effective enough.
>
> I'm probably misunderstanding, but this sounds *too* effective.
> Wouldn't it also mask unexpected failures?

Yes it would, but it would be a tool only used for release (if at all). 
And Pugs common practice it to mask all failures  in the release.

> That is, it's reasonable to have a set of "known and expected 
> failures", but
> you do want to be sure to learn about unexpected failures.
>
> I don't see any way to avoid specifying exactly which tests are 
> expected
> to fail.
>
> --
>     Andy Dougherty              doughera@xxxxxxxxxxxxx