Re: Unit test framework for gnome-games?

["Vincent Povirk" <madewokherd+d41d gmail com>]

Winning the game with dogtail would of course require the game to play
out the same way every time. This might be done in aisleriot by
creating a test game that doesn't shuffle the cards. But I can't see
any similar solutions for the other games.

Aisleriot games could also be tested with some guile script that does
what the C part of aisleriot does, but I can't see this being worth
it. There's just not enough tampering that goes on in the scheme part
of aisleriot.

Then again, if we start on ambitious projects like making the hint
function more visual there will be a lot of tampering all at once.

On 10/30/06, Thomas Mills Hinkle <tmhinkle gmail com> wrote:
> > Someone is working on implementing unit tests for gtk+, which sounds
> > like a good idea ( http://blogs.igalia.com/dape/):
>
>
> There was a lengthy discussion of testing at the GNOME Summit here in Boston
> -- the notes are here:
>  http://live.gnome.org/TestingUsingAtSpi
>
> I'm not sure if those folks have moved forward on any of their goals or not.
> If so, we might end up trying to plug-in to whatever framework they end up
> developing. It looks like Dogtail may be the framework they use, so that
> would be good.  As a python guy, dogtail looks pretty good to me.
>
> What I imagine with dogtail is having a directory full of scripts that do
> common tasks -- open game, new game, win game, enter score, etc. -- and make
> sure the results are as expected.
>
> For me, this is slightly different than what I would think of with
> unit-testing, which I would tend to use to test each module of a system and
> make sure it functions as it should.
>
> These aren't mutually exclusive but for me they're different.
>
> Tom
>
>
> _______________________________________________
> Games-list mailing list
> Games-list gnome org
> http://mail.gnome.org/mailman/listinfo/games-list


--
Vincent Povirk