[pytest-dev] regarding parametrized ids

[Floris Bruynooghe]

Hello Holger,

On 16 December 2013 05:31, holger krekel <holger at merlinux.eu> wrote:
> Hi Floris,
>
> i think the docstring is wrong: "argvalues" is not part of @fixture signature.

Oops, copied that from MetaFunc.parameterize().  Fixed it to refer to params.

> Btw, Anatoly suggests even core committers (including me) should propose
> PRs instead of committing directly for any non-trivial change.  I'd like
> to try this out for a while (probably hardest for me :).  What do you think?

I'm all for giving this a go.  But maybe we'd want to discuss some
workflow on pytest-dev, e.g. what branching style to use etc.
Personally I like mq but bitbucket's support for it is a bit poor (no
pull requests).  So far I've just created a throw-away repo per pull
request to avoid creating branches but that gets a bit cumbersome I
guess.

What would also be great is if we could aromatically have the full tox
tests run for each pull request, maybe this is possible with some
devpi interaction?  Or this is already possible with the travis setup
(but I thought that had incomplete coverage)?

I'm happy to try and help towards this and I'm pretty sure we can help
donate a vm to run tests on (we already have a cpython buildslave
running on sparc), or we might be able to get some rackspace dontated
boxes (I heard they do this for free software projects and they'll
have better connectivity)?


Regards,
Floris


-- 
Debian GNU/Linux -- The Power of Freedom
www.debian.org | www.gnu.org | www.kernel.org