[darcs-users] branching

Keith Irwin keith.irwin at gmail.com
Tue Oct 5 00:32:30 UTC 2004


On Tue, 5 Oct 2004 03:22:39 +0300, Aggelos Economopoulos
<aoiko at cc.ece.ntua.gr> wrote:
> On Tuesday 05 October 2004 00:22, Keith Irwin wrote:

> > My assumption is that a "branch" is simple a copy (get) from a
> > self-designated central repo,

> Yes, though nobody forces you to have a star topology (i.e. with /one/
> 'central' repo). No, I've never tried keeping more than one "bleeding edge"
> repos for a significant amount of (development) time, so I don't know if that
> works well.

Well, I'm experimenting. ;)  The thing is, I might work on, say, four
machines: my laptop, one of two workstations at work, and one at home.
 With a central repo, I can always checkin to the repo before I switch
to a different machine, then update from the repo before I start work
on a new machine.

With darcs, I start to think I have to remember where the latest is at
any time and get myself to thinking about the endless permutations.

With darcs source itself, I can just "darcs pull"  every morning to
see what's changed and then carry on.  From my perspective as a
user/non-developer of darcs itself, I might as well be using the CVS
central repo mode.

> > but then I quickly get messed up
> > thinking about how you keep track of these branches outside of knowing
> > where all these check outs reside.  If I have a branch on a laptop and
> > the laptop crashes, I'm sunk, eh?
 
> And if you have another branch on your desktop and it's disk crashes how are
> you any better? I suppose that, if only for your convenience, you'ld keep all
> your repos in sync (with push/pull or rsync), on all the machines you use for
> hacking, which is a more optimistic answer to your question.

Right. My assumption is that at least one of those machines is
properly backed up. If I check in each session's worth of work, then I
can recover from that "central repo".  I think I'm thinking about this
all wrong, though.

> Heh, that's a  poll question :-)

Poll?
 
> Personally I use:
> 
> ~/devel/myproj/myproj
> ~/devel/myproj/myproj-mybranch
> ~/devel/myproj/myproj-tmp
> etc

But each one of those dirs are a complete source tree, right? 
devel/myproj isn't the top of the tree . . . right?

Hm.

Maybe the wiki needs a "how I use darcs" section.  As soon as I use it
for a few weeks, maybe I'll write up something so that everyone else
will say, "you fool!" and with the resulting clue-by-four applied to
my skull, I'll have a revelation.

Keith




More information about the darcs-users mailing list