|Re: [jgit-dev] [egit-dev] JGit/EGit merging|
On 12/22/2011 11:05 AM, Christian Halstrick wrote: > Hi, > > here is my personal view on it: as you can see in  there are > multiple things which the recursive merge strategy adds compared to > our current resolve strategy. One is handling merge situations where > you have multiple common parents (e.g. criss-cross merges). I would > say that this aspect is easy to do and that we can put in on our todo > list (christmas holidays bring soo many free hours ...). > I currently can't estimate how much effort it is to bring the second > aspect of recursive merge strategy: rename detection! That's the > aspect I think you are interested in. I guess it's not much but I > don't know yet. Somebody else knows? not that i want to push on something, just as a side-note: we're bringing our >600 plugins project on GIT on 9th of january 2012. so hurry! :D no, joke aside: i think it would be a really cool feature to have, because renaming is something quite common, _and_ in addition, merging rename/modify becomes a _lot_ more likely as soon as you use gerrit. criss-cross merges in turn are something that become less likely when using gerrit as i understand it :) Regards, Markus > >  http://schacon.github.com/git/git-merge.html#_merge_strategies > > Ciao > Chris > > On Wed, Dec 21, 2011 at 17:34, Markus Duft <markus.duft@xxxxxxxxxx> wrote: >> I think not having a recursive merge is a really bad thing, as the no.1 killer feature of git doesn't work. I read somewhere (lost the exact URL; the comment was by a jgit committer i think) that implementing a recursive merge now that jgit has rename following, should be "fairly trivial". does fairly trivial mean i can expect this soon?
Back to the top