Hi Michael, On 31/03/2012 12:42, Michael Hendricks wrote:
> * I don't think darcs rebase command, as implemented, is a good way > forward for darcs (I've been compiling feedback after using the > branch for the last couple weeks and may send that later today) It would be great to have this feedback fairly soon given my current proposal to merge rebase in to mainline, which is something that's been on the radar of at least some of the core team for a while, although probably only mentioned on IRC/in person discussions up till now. > * it seems odd to reword a patch with amend-record when there are no > dependencies but use rebase when there are. The user shouldn't have > to think about patch dependencies (that's why they're using Darcs :-) > > That last point may be the most meaningful to me. Git has two commands > for this (git commit --amend vs git rebase -i, depending on whether the > commit you're rewording is the most recent). I've spent many hours > trying to explain the distinction to frustrated developers. We would > all have been happier if I could have said "Run `darcs amend --edit > --none` and follow the prompts, it'll do just what you want" I think it definitely makes sense for amend-record to do a rebase if necessary; we just have to make it clear to users that they are changing the identity both of the patch being amended and the patches that it depends on. BTW the general concept of rebase was just as a souped-up amend-record, so I'm certainly happy if we can unify them more generally. To some extent the current separation is precisely because it's been developed separately. Cheers, Ganesh _______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
