Re: [cclib-devel] Jaguar data files

[Noel O'B. <bao...@gm...>]

On 31/10/2007, Karol Langner <kar...@kn...> wrote:
> On Wednesday 31 October 2007 07:35, Noel O'Boyle wrote:
> > > Done. The cclib package is now almost 50% slimmer, so that might be
> > > something to still include in the 0.8 final.
> > > Still to do:
> > > 1. Support reading multiple files in regressions.py, plus general
> > > cleanup. 2. Allow unittests to be run in regressions.py, for those
> > > used-to-be-unittests and maybe other files.
> > > 3. Update further to Jaguar7.0 when we get the liscence.
> >
> > Regarding 2, wouldn't it be easier just to have testall.py run those
> > tests if basicJaguar4.2 is present? This would be a lot easier to
> > implement and I don't see any downside.
>
> You mean "Jaguar4.2" now. That would work, the coverage is the same, and I
> think both solutions are easy to implement. The question is just, how
> important is keeping the most-recent unittests separate from the regressions.

I don't think it's a good idea to keep them too separate. Otherwise,
failures will only be discovered at a later point...

> I see one drawback: if you parse test these files with testall.py and then go
> on to run regression.py, you will actually parse them twice!

Regression.py parses everything twice already, because there are files
in the basic folders which are not tested by "testall.py".

> > Regarding 1, we should probably make a subfolder of tests, and put
> > everything that testall.py imports into that.
>
> I agree; also, for clarity I think the current 'testall' should be renamed
> to 'testparser' since thats what it does. In my view, a 'testall' script
> would run testparser, testmethod, whatever other testsuits there may be and
> possibly regressions, and print the results in **very** condensed form.

Sure.

Noel