On Thu, May 14, 2009 at 14:16:02 +1000, Trent W. Buck wrote: > Discussion of transplant is in http://bugs.darcs.net/issue938. > > My impression that it was intended for exceptional cases where a > retroactive continuity change will avoid an exponential merge or > conflictingnons thingy, rather than for regular use. I guess we'll > find out when someone implements it :-)
There are three types of darcs transplant-y ideas I have heard of: My personal use case is as part of a darcs first-aid tool: recovering from corruption introduced, for example, by an older version of darcs; or maybe dealing with poor decisions of the past. These would be cases for which there is no meaningful fully, automated darcs repair command. The darcs transplant command would be used for automating the recovery process as much as possible: fix the damage by hand, then transplant your patches over to the repaired repository. The more popular use case is as a mechanism to help with long lived branches as Simon mentioned. A stronger of this is to have a mechanism that keeps tracks of equivalences between transplanted patches so that patches could still conceivably flow back and forth. Finally, another idea is that of amend-record on steroids, which I think means being able to amend a patch on which other patches depend and adjusting them accordingly. Things to keep in mind: this doesn't necessarily need to be one single "darcs transplant" command. You could imagine commands like "amend group" and "amend ungroup" to break patches apart playing a role. Also, working out a good user interface for these kind of transplant-y jobs (and others) could be a good way to contribute to darcs if you have a good rough mental model of darcs and you're not in a position to do any deep hacking. -- Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow> PGP Key ID: 08AC04F9
pgpeq7yTB10Wr.pgp
Description: PGP signature
_______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
