I tried it after using Emacs Magit for about six or seven years, and jujutsu is really easier to use than git and useful if one wants a tidy public history of changes (Linus Torvalds recommendations on that linked here). Plus it is fully compatible to git as backend - other contributors will not even note you are using it.
I use plain git when a project wants to use some tool that itself calls git commands that modify the repo state. You can use a colocated repo in this case (where jj and git commands both work) and nothing will break, but it can mess up your graph, creating duplicate commits which you then have to fix. I’ve seen this with Gentoo’s pkgdev for example.
Git blame and some other graph operations are also just faster right now which is why I sometimes use them in large repos over jj’s equivalent.
Well, the thing that this seeks to improve on is the crazy complexity of advanced git commands, which gave rise to several humorous mentions on XKCD and even satire man pages like this:
https://git-man-page-generator.lokaltog.net/
… and if you think that you know most of git well, then, quick, what happens when you merge two branches of a repo which has several changed submodules in each branch? Is this deterministically resolved, and if so, how?
Oh my god this is amazing. Thanks for this.
I would assume a merge conflict if the submodules were changed in both branches from the base… but it’s probably not that simple, is it? I’ve never tested it.