[darcs-users] best practice for tracking a vendor branch with darcs?

[Trent W. Buck]

"Stephen J. Turnbull" <stephen at xemacs.org> writes:

> I don't think Darcs supports tree-less (not just no-pristine, no
> working tree, either) branches, but if it does that would help with
> this problem.

http://darcs.net/issue431 (needs volunteer)

And just because nobody has mentioned it yet, quilt(1) is a VCS-agnostic
equivalent of hg queues and bzr looms.  It is useful for maintaining a
series of small patches against an rapidly-moving upstream.

You could conceivably use it as the loom/queue, with Darcs essentially
as just the transport mechanism (with appropriate pre- and post-hooks).
If you know in advance that your workflow is going to trigger the
exponential blowout on darcs-2, this might be an alternative.  I haven't
tried it (with Darcs) yet myself.