Not at all.
Unlike CPAN Testers, this is designed with standalone components.
So as long as you can create a Distribution::Harness object and run it,
it should spit out PITA::Report XML content/files for you, without any
of the system image stuff needed.
So if you wrap up your tests as something that _looks_ like a CPAN dist
(directory layout, Makefile.PL etc) then it should work.
And I'm hoping we can implement some math in the PITA::Report class too,
so that you can just $Report1 += $Report2 to merge the two together.
So in your case, you need Distribution::Harness + PITA::Report, and the
rest is irrelevant.
As for the schema, it's _potentially_ a little richer than you need, as
it offers the chance for a smarter analyzer later on to re-evaluate the
conclusions, but it does have all of the TAP output in raw form in it as
well. Should be just what you need.
Adam K
Michael Peters wrote:
Adam Kennedy wrote:
Michael
There's existing work happening in this area you may want to get
involved in. We even have a draft XML schema that does exactly what you
are talking about.
Go read http://ali.as/pita/
Definitely an interesting project. It is much wider in scope than my little
idea. There are a couple of reasons that I think might make it too much, or too
different from what I'm trying to do (at least initially).
While I want to test module packages I would also like to test whole
applications. I don't want the developers to have to do anything special with
their dev machines, or smoke machines, just send their reports to my app. This
makes it a lot simpler.
But I am interested in the XML schema that you have so far. Is it tied heavily
to PITA, or mainly focused around the data output from TAP tests?
