[darcs-users] Re: questions on development process

Eric S. Johansson esj at harvee.org
Sun Aug 8 13:44:49 UTC 2004


David Roundy wrote:

> On Sat, Aug 07, 2004 at 04:26:46PM -0400, Eric S. Johansson wrote:
> Right, darcs push/pull will only move recorded changes.  If you want to
> move unrecorded changes, however, you *could* temporarily record them, and
> then unrecord them after the transfer.  This would be a bit of trouble, and
> you'd still have to be careful that the unrecorded changes only exist at
> any given time in one of the two repos (to avoid conflicts with yourself).
> 
> Or if you don't mind recording your changes more often, you could just
> permanently record your changes whenever you feel the need to go from one
> computer to the other.  This would definitely be the safest way to go.
> 
> The third darcs-related alternative (as opposed to just using unison or
> rsync for everything) would be to use unison or rsync for the working
> directory, but not for the recorded changes, which you'd push manually.
> You'd need to do the push/pull before using rsync on the working directory,
> but at least you'd be safe from losing any recorded changes due to trying
> to rsync the wrong way.
> 
> I generally go with the second option.  I work in a number of different
> repositories on different computers, but each change I finish and record in
> a given repository.

maybe this is a bad attitude but I feel that but files should not go 
into a repository unless they are fully functional.  I think this comes 
from my experience in working with groups were too many times my work 
has been halted by the fact that somebody checked-in broken code.

this means that usually I have a fairly substantial body of work in the 
"work in progress" files around the repository.  So it sounds like to me 
that my better option is the unison bidirectional synchronization.  That 
way I also get copies of files I haven't checked in yet and may never 
want to.

>>Third: creating repositories.  this is still a bit of a mystery to me. 
>>Do I just want to copy to the public HTTP site the _darcs directory?  Or 
>>do I want to copy what's below that?  Because to copy what I think of as 
>>the repository (_darcs + "work in progress" files) would reveal too much.
> 
> 
> I always make the working directory also available, but on the server it's
> just identical to the recorded version, since I update things with
> send+apply.  If you need to use rsync to transfer to the public site, you
> certainly could leave out the working directory.

I will experiment with that and see.  I was also thinking about 
automatic triggers for updating the repository whenever I tag it for 
release but it's probably not a good idea.  but maybe adding rsync as an 
option for pushing.  On the other hand, my earlier suggestion of 
separating the communications functionality from darcs could make adding 
new forms of repository propagation easier.  It also might simplify 
darcs by allowing you to merge push/send as well as pull/get.

---eric

-- 
Speech recognition in use.  It makes mistakes, I correct most





More information about the darcs-users mailing list