Re: Darcs supports pre-checkin hooks was Re: Controlled anarchy

Am Freitag, 21. Oktober 2005 22:54 schrieb Shae Matijs Erisson:

Jacques Carette writes:

...

I would strongly recommend against 'allow everyone to just commit' without the presence of a large automated test suite which is used to (automatically) reject code that breaks a test.

Darcs supports 'run test before checkin' and if we had darcs, people would contribute tests too.

Well, I appreciate everybody's enthusiasm for darcs, but the statement above is highly misleading and *not* a valid reason at all to change to darcs: We could easily have this CVS already (and of course with Subversion etc.), see: http://ximbiot.com/cvs/manual/cvs-1.12.13/cvs_18.html#SEC190 Our main problem here is not technical, it is simply a lack of a decent test suite for all the library packages in the repository. And I totally agree with Jacques here: As long as we don't have this test suite, I'll furiously object to 'allow everyone to just commit'. This will simply not work and will actually turn away much more people from Haskell due to the resulting instability and API volatility than attract people to our beloved language. I see a dire need for more Haskell *maintainers*, not for more developers or brand new shiny version control systems. With "maintainers" I mean people accepting/testing/merging patches, discussing with people about APIs, kicking coders to write tests for their code, collecting opinions and writing down API proposals, keeping existing APIs stable/sane/usable, etc. This is a lot of work, it's difficult, time-consuming and much less fun than coding, but we need those people. Darcs won't help with this at all... Don't get me wrong: I think that darcs is a great tool and did a lot to improve Haskell's visibility and acceptance, but it is not a solution for the problems we have. People have always a tendency to propose technical solutions for non-technical problems, this is exactly what I'm observing currently on this list... Cheers, S.