[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Arch and Guile
From: |
Andreas Rottmann |
Subject: |
Arch and Guile |
Date: |
Mon, 01 Mar 2004 15:27:38 +0100 |
User-agent: |
Gnus/5.1006 (Gnus v5.10.6) Emacs/21.3 (gnu/linux) |
Thien-Thi Nguyen <address@hidden> writes:
> This fork-thing is quite a f**ckup IMHO.
>
> it's probably better to call it a forced-inconvenience than a fork.
> in any case, it has demonstrated the delicate nature of using cvs for
> revision control.
>
Using Arch would also make stuff easier for the "mainstream" Guile
developers to maintain their stable and development branches: They
could have an guile--fixes--1.6 branch, wich they'd simple star-merge
into HEAD.
> perhaps you could ask Miles Bader how he set up the arch-cvs gateway
> for emacs, and do the same for guile.
>
I already have a bit of experience with tla-cvs-sync from
tla-tools. For running a gateway in both directions, one needs Guile
CVS write access. I could set up a gateway that tracks CVS and hosts
additional stuff from various sources, however. A potential problem
with that is that merges into CVS (if to happen) should happen via a
tool such as tla-cvs-sync and not via ordinary patches.
> i would like to learn how to use such a gateway once it is available
> and there are docs for the newbie (me). btw, i'm glad to host such
> docs or links to them in <http://www.glug.org/docbits/>.
>
Such a gateway is very simple to use: You have someone who is the
"Gatekeeper" between the Arch and CVS world. This person then merges
stuff form other people's branches into the dedicated "CVS" branch,
which is synced with CVS in both directions.
Due to the distributed nature of Arch, such a Gatekeeper could have
"lieutenants", which are responsible accumulating the "good" stuff
from several people into their "integration" branches. This means
Gatekeeper doesn't (need to) become a bottleneck, even if there are a
a lot of people contributing on the Arch side.
I think that I'm not suitable as a Gatekeeper, since I'm not a Guile
(core) hacker and haven't assigned copyright (yet). It would be
really, really terrific if someone of the core hackers would take this
responsibility. I volunteer helping with
* Setting up the Gateway. If I get positive feedback, I might set up a
CVS-tracking-but-not-commit gateway as described above. If the stuff
in the Arch world is considered interesting enough for the Guile
core hackers, one of them might want to take over
Gatekeepership. Actually, this sounds like a good intermediate
plan. Thoughts?
* As a (first? ;-) lieutenant, leaving the Gatekeeper only with a
single branch to deal with.
> previously i had thought it better to wait until there is available
> an implementation of the arch "wire protocol" (whatever that may be)
> in guile scheme, but perhaps that is too hard line...
>
I'm not totally clear what you mean here, and how that relates to an
Arch<->CVS gateway. Anyway, you might be interested in my ITLA
thingy[0].
Cheers, Andy
--
Andreas Rottmann | address@hidden | address@hidden | address@hidden
http://yi.org/rotty | GnuPG Key: http://yi.org/rotty/gpg.asc
Fingerprint | DFB4 4EB4 78A4 5EEE 6219 F228 F92F CFC5 01FD 5B62
- Re: Extended -e syntax, Andreas Rottmann, 2004/03/01
- Re: Extended -e syntax, Thien-Thi Nguyen, 2004/03/01
- Arch and Guile,
Andreas Rottmann <=
- ITLA (was Re: Arch and Guile), Tom Lord, 2004/03/01
- Re: ITLA, Andreas Rottmann, 2004/03/01
- Re: ITLA, Tom Lord, 2004/03/01
- Re: ITLA, Andreas Rottmann, 2004/03/02
- Re: ITLA, Tom Lord, 2004/03/02
- Re: ITLA, Andreas Rottmann, 2004/03/02
- Re: ITLA, Miles Bader, 2004/03/02
- Re: ITLA, Neil Jerram, 2004/03/04
- Re: ITLA, Andreas Rottmann, 2004/03/04
- Archive of library modules for Guile, Andreas Rottmann, 2004/03/07