Is it time to start regression testing yet?

[C. Florian Ebeling]

On Wed, Jun 10, 2009 at 9:37 AM, Jordan K. Hubbard<jkh at apple.com> wrote:
> Just one suggestion, as we all pile on Florian's design:  Can we please try
> to do our best to resist the temptation of allowing the perfect to become
> the enemy of the "good enough?"   I too see a lot of issues with Florian's
> portmill system, but I also know that things tend to evolve rather rapidly
> once they're pressed into actual service, so I'm OK with it not being the Be
> All and End All of build/test frameworks.  Given the amount of time which
> has passed without having anything at all (6 years?), I think we should be
> willing to start pretty much anywhere, just so long as we start!

I consider the system as it is the absolute minimal thing that offers
any utility. It's just a list, and some background mechanics watching
svn and wrapping around MPAB. Obviously there is so much more that one
could add. But I'm also quite fond of agile mindset, which values
simplicity. Simplicity is quite a high goal though, and it is not the
simplicity of missing features that I'm after ;) But a simplistic
system that actually exists and runs is more important to me than a
perfect system which only exists as a project outline.

I really only mentioned my little app once online it quickly leaked to
the dev list. I was stuck for quite some time with a mysterious
macports regression in 1.7.1 which went away with trunk MacPorts, and
then it hit everybody here. So hold your breath. I will open a feature
wishlist thread when I have time to think of new features.

Currently I'm discussing the possibilities of moving everything to
dedicated hardware at MacForge with Bill Siegrist. Once those things
stabilize, new features can go in as well. Oh, and as always: patches
are welcome :)

Florian


-- 
Florian Ebeling
Twitter: febeling
florian.ebeling at gmail.com