Re: guile-gnome's version control

[Andy Wingo]

Hey Greg,

I definitely agree that whatever system we end up with has to be usable.
GNU arch grew to be a nightmare in that regard.

On Mon, 2006-08-14 at 08:34 -0400, Greg Troxel wrote:
> 3. The build system (and module splitting, if any) should be
>    completely decoupled from the revision control system.

Perhaps. Certainly our =RELEASE-ID stuff sucked. You still need to be
able to tell if a subdir is in VC or not; I suppose we can detect that
using characteristics other than version control.

> 4. VC tool should have stable storage representation, and we should
>    expect it to be around and in common use in 5 years.

See for example tailor, http://progetti.arstecnica.it/tailor. These days
there's strong support for converting changesets between anything; I
don't see this as a huge issue. It's more "can we leave from this system
with our history", and I don't currently see an issue with any system.

> 5. VC/hosting should allow for long-term project survivability without
>    depending on an individual.

Yes.

>   This means multiple committers, and
>    multiple admins (those who can approve committers), perhaps backed
>    by an organization with meta-root privs.

This is how you would do it in a centralized system. In a distributed
system I don't see this as necessary. If I died or something you could
always get GNU to assign someone else. Guile-gnome is a GNU project by
the way.

For the rest of your points, I understand your concern. We had a bad
experience with arch. But I don't think it's worth giving up on the idea
of distributed version control.

Instead of addressing all of your points, lemme finish my next mail :)

Cheers,

Andy.
-- 
http://wingolog.org/