Excerpts from Rob's message of Fri May 15 08:25:24 +0200 2009: > 2009/5/15 Stephen J. Turnbull <[email protected]>: > > Simon Michael writes: > > > > > Nit pick: shouldn't this just be called darcs rebase ? > > > > No. If you want, you could call Darcs itself "rebase", because > > commuting patches is fundamentally what rebase is about. David would > > probably have to sue you for defamation, though. :-) > > > > Transplant is IMHO a bad name, partly because Darcs fundamentally > > doesn't think in terms of a history DAG, so "moving (some part of) a > > branch" doesn't really make sense.[1] There's also the problem that > > Mercurial's use of the term does conflate rebase and cherrypicking. > > > > I think "cherry-pick" is really closest to the flavor of what's > > proposed here. The main problem I can see is that in other VCSes, a > > cherry-pick nearly completely destroys the ability to detect that a > > patch has been previously applied. That connotation would not be true > > in Darcs. > > > > Footnotes: > > [1] It's a *set* of related patches, not a *sequence* of commits as > > understood in other VCSes. > > What about calling it "borrow"? I think a name like this conveys that > it will take the content of the patch(es) but not the entire (branched) > context leading up to the patch(es).
What about calling it "darcs replay", it would cover at least one use case. The command would be able to "record as new patches" a set of patches from a given repository. In combination with "darcs amend-record" on steroids it would cover most important use cases IMHO. Best regards, -- Nicolas Pouillard _______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
