[darcs-users] Re: Moving a tag?

[Eric S. Johansson]

BARBOUR Timothy wrote:

> Keep the current stable release of darcs in a separate repo, and only push
> new features to that repo after they have been tested in the unstable
> version for some time. Have a series of repos, one for each actively
> maintained release (at present maybe just the last stable release), then a
> repo for the next pending stable release, then the trunk. Bug-fixes can
> probably be spread around everywhere that they are applicable (without
> dragging new features into releases). New features should be pushed from the
> trunk to the pending stable release after sufficient testing. When it is
> time to release the pending stable release, it is added to the stable
> release series, and is copied (via darcs) to make the repo for the *next*
> pending stable release. As above, patches and tags for releases should be
> pushed into the trunk, so that the releases can be accessed from there.
> 
> I think the above processes could be automated. Pure bug-fixes could be
> automatically pushed everywhere (provided we can identify them), and new
> feature patches could be automatically pushed to the pending release after
> they reach a certain age. The creation of a release could be triggered by a
> human running a script, or from a cron job.

I'm inclined to agree with you.  Having been in situations were I have needed to manually propagate a bug fix across 7 OS releases (pre revision control) I would love to be able to create a branch and then apply a bug fix to a branch.  After applying the bug fix, waive the electronic magic wand and have it reveal where it would slot in smoothly and where it would need some work.  I wouldn't be happy with entirely automatic updates.  I really want to say "where", "how easily", and "go ahead".

But yes, that would be quite interesting. 

---eric