Distributed SCM (was: Re: compiling cvsup broken on -current)

[Peter Schuller]

> >I also understand that the Linux development-model is a bit different 
> >from the BSD-model, might be wrong here but don't they have separate 
> >maintainers for different parts of the kernel who gets to descide what 
> >goes in and not. Or was that how it used to be? Anyway, my point is 
> >that different models have different requirements on the SCM.
> 
> That might be, but it doesn't have to do with the distributed nature of 
> an SCM.  The pros I see is that I can develop and have my local patches 
> versioned until I feel that it's ready to go into the tree.

Another advantage of having a distributed SCM, that often gets lost in
these debates IMO, is that it makes it *MUCH* simpler for us mere
mortals who have no access to the official repository to maintain our
own changes in our own private branches, while still keeping things in
synch with the official codebase. I don't know how many times I have
wished this were possible with project that are using CVS/svn.

With SCM:s like darcs (which is sadly not scalable enough at the
moment to be practical with a tree the size of DragonFly) these things
become a breeze. Does Git provide sherry picking of patches and the
ability to maintain private branches while not loosing any version
control power? (I have no idea, not having used Git.)

I dream of a future where I can maintain my own repository
(dfly/freebsd/whatever), apply my own patches and sherry pick other
3rd party patches from a number of unrelated repositories. :)

-- 
/ Peter Schuller, InfiDyne Technologies HB

PGP userID: 0xE9758B7D or 'Peter Schuller <peter.schuller at xxxxxxxxxxxx>'
Key retrieval: Send an E-Mail to getpgpkey at xxxxxxxxx
E-Mail: peter.schuller at xxxxxxxxxxxx Web: http://www.scode.org