gnu-arch-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Gnu-arch-users] Using arch for offline development


From: Andrzej Wąsowski
Subject: [Gnu-arch-users] Using arch for offline development
Date: Thu, 13 May 2004 00:08:11 +0200
User-agent: Gnus/5.1006 (Gnus v5.10.6) Emacs/21.3 (gnu/linux)

Hi,

I am currently evaluating tla as a replacement for my CVS (in managing
single person projects mostly, maintaining my homedir, etc). I am not
primarily interested in the most glamour advances of modern SCM tools
over CVS (like atomic commits, moving files, etc), but in a possiblity
of making commits when being offline.

My current setup is based on a local CVS repository stored on laptop's
hard driver. This repository is backed up daily to some server to ensure
data safety. I am not entirely happy with this setup as with a portable
machine daily backups are often skipped (bcs I am not online at expected
time, or the machine is off, etc). On the other hand I dislike solutions
like backing up with anacron (which invariably heavily slows down the
machine when I turn it on---most often this is precisely the time when I
need it quickly) or using a remote CVS repository (which prevents me
from commiting, when I should, enforcing commiting when I am online
instead).

I thought I could try to use distributed features of tla to get a safer
and more convenient setup, when I would have my principle repository
(archive) on a network server, but used a local copy when offline (and I
should be able to commit locally when offline). I am not sure though,
how to exactly set this up, mostly due to an overwhelming choice in
constructs.

Should I keep the main branch on the server and do development in the
local "laptop-branch" commiting back to the server when I am online?
Would it replicate my local changesets on the remote server? Or will it
just make slightly less detailed log entries (one entry subsuming all
the applied patches)? Also the problem here seems to be that I am
actually working on single development line, but have to maintain two,
with different revisions, etc.

Alternatively should I just have the main archive (and branch) on my
laptop and use mirroring on the server to achieve backups? Is such a
mirrored archive a proper archive in a sense that if the laptop gets
stolen/broken/etc I can easily turn the mirror into my main repository?

Or perhaps should it be another way around (having a local mirror of the
remote archive on the server)? The mirror facility is not very well
explained in the guide as it is now. Am I only allowed to perform read
only operations on the mirror?

Operationally, what I would really like most is something like using a
remote archive, with posibility to commit localy when disconnected
(storing commites locally until I am connected again and they can be
propagated to the archive). I am not sure I can achieve this effect with
arch using various goodies branching/mirroring + may be some
scripting. 

Can enybody guide me in that? I will be most grateful.

Andrzej





reply via email to

[Prev in Thread] Current Thread [Next in Thread]