The creator of Jujutsu has joined ERSC
ersc.io
ersc.io
JJ is just so low friction. The seamless and painless shuffling of commits, working up and down the commit chain without explicit rebasing, unchanging change-ids. The min identifiers in change-ids being highlighted. Being able to make commits and worry about the message later. Snapshots of changes while you are working serving as a tool agnostic undo.
Painless and sensible cli args that are vastly more consistent than git.
Superior conflict handling.
You can “jj undo” anything and it works perfectly.
It’s fabulous.
> "@--+::"
I have no idea what that means and I’ve never had to type anything remotely like it.
The rev specification language is pretty elaborate but you don’t need to know it.
In your opinion, is jj losing steam? How do you feel about it requiring a CLA to contribute to, which would enable the developers to go closed-source and keep all the contributions of the community? Are you afraid of vendor lock-in?
IMO that's almost fundamentally impossible. The whole point is that it makes maintaining a "clean" linear commit history easier than git, so by the time you push anything the observable differences are already gone. At best you just see a repo without any WIP-style commits, which can be done using more effort and git alone.
Or if it were a family tree, "all descendants of your parents and their cousins" (ignoring that family trees are way more constrained than DAGs).
I have no idea why you would use that revset; I assumed it was just an example of revset's complexity and power.
As a pragmatic matter, unless I'm scripting with revsets, I personally stick to a few go-to aliases.
But I don't tell my team what they have to use as long as it's compatible. The last thing I want to do is try to teach them a different VCS when they've spent time learning how to get by with git.
I've told them about JJ, what I like, what pain points it solves for me. They are all adults, they can make their own decisions about it.
JJ is for me. It helps me unwind some rather elaborate problems we've had with rebasing, or dealing with conflicts from bringing some old support branch up to date, or back porting some particular thing. It helps me create pretty clean MRs, break things up into smaller reviewable chunks, etc. You can do all this with git, but it's easier in JJ. In particular if you have to do shuffling and rebasing and conflict arise. JJ lets you solve conflicts asynchronously to rebasing. Git stops you in your tracks. There's some rerere thing that is supposed to help but give me a break. Git's a train wreck of dumpster fires. But it's still been basically better than everything else since svn.
Until JJ IMHO. Valid arguments can be made for other things like maybe mercurial or even fossil. But JJ is the most pain-free vcs evolution I've seen.
Mercurial's big failing was making too many critical things optional. It's the same screw-up riscv is going through now in the chip architecture space.
- undo
- shuffling commits around with squash, split, and rebase is much better than git’s interactive rebase
- make commits without having to come up with a branch name (I might make three versions of the same change in parallel to see how they compare)
I see git is working on adding some of this under the history command. The revset language is really not esoteric: @ is head, @—- is two behind head. That’s about it.
Part of it is that the sorts of shuffling I used to avoid because they were a pain in git (so I didn’t feel I needed them improved) are so easy in jj that I do them all the time.
Can you explain this to me? I feel like Git is pretty easy there.
- select oldest commit to modify
- move the commits around with a mouse or the cursor
- close the editor to apply
Alternatively I can add changes to older commits by recording them on top (--fixup) and tell git to auto apply them (--autosquash). I can also tell git to do the first thing automatically (git absorb, I believe it's inspired by jj).
But say you’re in the middle of working on something and you wish you had a commit you made last week on an experimental branch on the current branch before the last commit you made. That’s jj rebase -r oldercommit --before @-, without interrupting your work. I don’t like to think about how I’d do that with git.
I rebase a lot using Fork (git gui) and it’s really easy to drag and drop reorder commits, rename, apply as fixups or squash in. I can see how jj makes sense if you’re in the command line, though. And splitting changes from commits in git is really annoying, but it only comes up once in a blue moon for me.
With git, if you do this, everything after is left orphaned. Even if you have a GUI that does this for you, now rebase conflicts become an enormous pain.
This is the best pitch for jj I have seen yet. This should practically be the first text on their website.
git -c sequence.editor="sed -i '1ip oldercommit'" rebase -i @~
All of that is automatic with jj.
How many do I want?
> How do find the correct arguments
How do you find the correct arguments for jj?
> and git checkouts to do before?
Why would I do a checkout before? I don't want to modify the worktree? Maybe I in fact also want to alter the worktree, but that's unrelated.
> Are you sure you remembered to do everything?
Am I sure I have all my files ordered? No. Does it matter? Also no. Are you sure you have no bug in your commits in JJ?
If I want git to alter some set of commit chains, I can tell it to. Actually I never needed to do that, because I don't work on several thousand branches at the same time. I prefer it to not alter unrelated branches automatically, just because some earlier commit changed that is in both. Such things actually undermines the trust I have in a tool, because it does things I haven't told it to do, even if I'm aware it does these things.
> What if there are merge commits?
Then I resolve them. Merge conflicts occur, because there is some actual conflicting change and often it is also semantic. These don't go a way by changing the VCS. On a theoretic basis, these require outside information(=decisions) that is not there yet. There are more often semantic conflicts, that are not syntactic, then there are the other way around. If you are referring to doing the same merge conflicts again, I can tell Git to resolve them automatically too, but it actually occurred too often, that this is not actually what I want, so I actually dislike that feature now.
git cherry-pick + git rebase -i (to swap commit order)?
?
But if the cherry pick doesn’t cleanly apply at the current HEAD, then you have to remember to either do the git rebase -i first and pause at the appropriate place to cherry pick it if that works (I think it should? though I also recall rebase only letting you pause before a commit so you lose your commit message, but that’s probably a me problem) or else maybe detach your head and start doing surgery because otherwise you’ll be resolving conflicts in two different directions as you cherry pick and then rebase, and at that point I’m usually going back to git reflog to try to find the last point where history made sense. Or I guess you could just remember that you can introduce arbitrary existing commit refs into an interactive rebase like 1718627440 did. But I said I didn’t want to think about this anymore.
If you don't want to do that, you can tell git to do it automatically too? I actually thought I would like that, so I used it for a week. I did not like it at all.
> But if the cherry pick doesn’t cleanly apply at the current HEAD
If the diff you want to commit isn't applicable to the tree you want to apply it to, it's not going to work, no matter the VCS you use.
> then you have to remember to either do the git rebase -i first and pause at the appropriate place to cherry pick
You say that like these would be separate operations, but for me these are very much not, I guess like the fact that you need to supply both '-r oldercommit' and '--before @-' are for you. I think this is the point where I actually don't understand your view. Like, we both need to supply two parameters, we need to because this is the operation we want to do. You write two parameters free form, I write one free and select the other from a list. There is a difference, I actually think selecting from a list can be more convenient in some cases, less in others. But I don't get why you write like it would be crazy work.
> though I also recall rebase only letting you pause before a commit so you lose your commit message, but that’s probably a me problem
It does only let you handle full commits, if that's what you mean, if you want to split a commit you need to provide information how. But that it doesn't sound like you mean that.
> so you lose your commit message
You can apply commit metadata independently of the tree you want to commit, so that definitely occurs never.
> maybe detach your head
I bet you don't have issues with detaching the head in JJ, so why do you in Git, it's just a normal state.
> start doing surgery because otherwise you’ll be resolving conflicts in two different directions as you cherry pick and then rebase
Yes, but that is not because you use Git, but because you wrote the change against some other commit first, which you don't do in your JJ example. You could do that in Git as well and then you only have one set of merges to resolve.
> I’m usually going back to git reflog to try to find the last point where history made sense
Which is git rebase --abort or git reset @{1}. The latter is always the same, this is as silly as saying I can't remember whether it's jj undo or jj revert.
> But I said I didn’t want to think about this anymore.
Yet you do in JJ.
Because it left you fewer opportunities to experience the ecstasy of typing git commands into a terminal? Or because it didn't work so well?
I believe this to be the same with the index, that is praised as unnecessary and "overcomed" in JJ.
I am super curious as to what you mean specifically, but I also know that getting into details can be annoying, so if you have some specifics I'd love to hear more about them. I loved git's index so this isn't coming from a place of "the index sucks and you shouldn't care about it" it's "I think jj executes the index better than git does actually". But also if you don't, that's fine by me too, if you like a tool you should continue to use it.
(I never set autosquash either)
If you care to learn more, there's tons of posts on hn and the broader web, so I won't bother sharing any links. I will say, though, that you should use it with jjui, which is a tremendous TUI for jj - you don't even really need to learn any of the actual jj syntax because it just handles it all for you.
Similar to you, 95% of the time I do the simple use case that I've memorized. The problem is that even after a decade of using Git, when I get into those 5% where something goes wrong, I have no idea how to actually solve it. None of it makes any sense in my head, and I have to search for a bunch of magical commands to run, or nuke things to get back to a good state.
With JJ, it just makes sense. It's intuitive and clear. I have been able to recover without any help. Just the undo alone is basically a life saver. Rebasing and merge conflict resolution is so much simpler.
Yes I probably won't use it to its full potential, but still, the little I do use makes 10x more sense than Git ever did.
It wasn’t. I was wrong.
The staging area is just a commit like any other, only due to git’s design it has to be special-cased everywhere. In jj it’s just a commit.
I do all my work on an unnamed commit on the tip of a branch. As I complete bits, I extract out cohesive sets of changes into new commits (split) or as patches of earlier commits (squash).
Git is a DAG of commits, each of which is a node in the DAG. There's also a staging area that isn't a node in the DAG, and a working copy that also isn't a node in the DAG. The head nodes of every branch in the DAG are always named in git, and every node in the DAG also has a description. Nodes are immutable, but the edges between nodes are mutable so you can edit a node by replacing it with a new node & updating the edges.
JJ is a DAG of commits. Each commit is a node. Nodes have stable identifiers, automatically generated. Nodes may have descriptions. Nodes may also have names (bookmarks) which get used for compatibility with git's branch names. Nodes are mutable, but by default nodes pushed to a remote will be immutable.
btw, did Martin leave google?
Martin left Google for this, yes.
That being said, there's a few things that lead to this perception: the first is that it used to be under Google's GitHub account, but is now under its own org. The second is that contributing to jj does require signing Google's CLA. That is something that is desired to change, but that's up to Google. Due to the license, the community could fork it, and still have the rights to use the code, but that would be pretty aggressive, and so everyone involved would rather work together on this than do so. The move to its own GitHub org was one example of movement in that direction, but I'm not aware of an actual timeline to remove the CLA just yet. The project will figure it out over time.
(I have a commit bit but am not a jj maintainer, so the above is my own impression from being involved in the community for the past few years.)
Maybe if Martin could comment - his @google.com email still appears in the repo readme, I'm assuming he doesn't have access to that box anymore - was the repo move and ownership coordinated with Google prior to his leaving?
I get why you may want to color this as non-Google-owned but it seems to very much still be Google-owned.
My point is not that it is “not Google owned” but to point out the specifics, which really matter. “Google owned” to me implies that they own the copyright to contributions, which they very much do not. That doesn’t mean that they have no control over it at all, of course, but the specific details matter a lot.
(Also, as I just said in another comment, I forgot about trademark: if the community did want to fork, they'd have to rename.)
In theory there is now nothing stopping us (the community/project leaders) from just... Moving the repo somewhere else. We could do that. It would be a pretty sudden alienation of other Google contributors though and probably piss off their legal department -- 99% of them just submit regular fixes like any user, because they are ordinary users of JJ. They just happen to work at Google. It would just be stirring up shit for nothing. It isn't an unthinkable outcome, but it isn't one we're just going to "do" for no reason or because it would make people posting on forums happier.
Realistically a non-ownership-assigning CLA on a non-copyleft Apache 2.0 project like in this case doesn't actually mean anything or confer any power. It's all just paperwork that serves as a legal ass covering mechanism because big companies use lawyers. It's like checking the "Yes I am over 18 years old" box on a website. I say this as someone who hates the stupid CLA bot and am a contributor.
As far as actual leadership goes Google does not "own" anything in that regard, nor the day to day work, nor can it really force any of the community to do anything. There's 1 maintainer from Google out of 9 total and +95% of major contributions don't come from Googlers, ERSC, or even any particular company. They don't contribute most of the code, they don't lead most of the community, they are vastly dwarfed by the community in total size, contribution count, etc. What "ownership" or leverage do they have beyond a Github org in this setup? It's annoying but it isn't the same thing as being able to exercise meaningful levers of power, in my experience, and so most of us are just fine with it. Even if there was some stupid bullshit like a disagreement over implied trademark rights, the community could just rename and move on and Google would have nothing. That is where the real power lies.
You have to get out from under Google. This would be a red flag during due diligence for many investors.
I'm not saying Google is going to fuck you over but why not just get yourself into a place where you don't need to be writing replies like this, and we don't need to be having this conversation.
And also, I would strongly recommend acquiring trademark rights as soon as possible. If a foundation is formed later you can decide what to do with them, but having them protected is a huge deal.
There's nothing starting me from creating jjhub.com tomorrow and that is a real potential threat for you.
It's not an issue for investors, because again, Google does not own this, and we aren't even modifying jj at this point, and are unlikely to need to. The IP is very clear and not an issue.
> There's nothing starting me from creating jjhub.com tomorrow and that is a real potential threat for you.
That's fine! There's a lot of people already doing this, today. Competition is natural.
Do you have any idea if this is coming to jj soon?
I wonder how guaranteed that is, can imagine GH deciding to 'clean' it (how much data can you put there theoretically? Or to close a range of potential security issues) and suddenly relatively niche but genuine uses like jj's change-id are unintentionally not preserved.
The change-id is exactly as much part of the commit object as the author/committer name/timestamp, description, parent commit ID, tree, and participates in the commit hashing as those do.
Not really. I remember my early years with git and the favt was that I never needed more than clone, add, commit, pull and push. While I’ve done some mistakes that got ne to learn more, especially with creating branches and undoing. I’ve never needed a lot, even when I started using GUI which exposed more concepts.
Why? Because I have no understanding of version management and how it’s useful in a dev workflow/release process. I was just using it for checking in work.
Since then, I’ve read the “Pro Git” book, learned how devs and teams handle versioning and devel a good understanding of how git can help me in my coding process and general software development. And it’s very good at what it does.
At the team level, it’s guided by the release process, configuration management, which version is canonical. At the programmer level, it’s usually guided by how to switch between task, how easy to explore an idea and save the resulting experiment, how to reset the code to a know state and how to replay a previous changes on top of new changes.
So you discern what you want to do (which is independent of the tool), the learn how to use git to do them. If you start from git, you’re going to be confused, just like someone opening autocad with no knowledge of engineering drawing.
Git is quite powerful, but the cli is a train wreck of complexity and inconsistency. Learning the options to one git command means you’ve learned the options to exactly one git command. No other git command is required to use the same terminology. Personally, I was stuck in “git basics land” (all the basic commands you list) until I adopted Magit in Emacs. IMO, the Magit team deserves a medal for making git usable at an advanced level.
I don’t have an opinion about jj yet, but no one can hold up git as a pinnacle of great source control. Yes, it’s powerful, but it’s a UI train wreck.
You forgot the “because it's actually kind of terrible from a UX standpoint”. My stance is against that. I haven’t learned git in depth because I never knew any other workflow than code and check in the changes. There’s not a lot of guides on how version control can help in the software development process.
It wasn’t until I got involved into OSS that I learned more which in turns give me the motivation to use and learn about git in depth. I’m also using magit (after a tour in various GUI) but for me it is to git what vi is to ex. Direct interaction instead of a command prompt.
I'm not sure I understand how "I use a specific text editor with a specialized interface into git that's superior" is a counterargument to the first-party tool having horrible UX. That seems basically the exact same as what I was saying about jj, except via emacs instead of a CLI. Either way, we've opted out of the actual experience of the tool in favor of using an alternative tool that operates on the version control.
It requires the same mental model as git.
The repository is a store, you check out the previously saved instance of a file, a group of files, or a subtree. In that regards, switching branches and restoring files is actually the same thing. A commit stores whole files and branches are pointers to commits (which update themeselves).
Maybe you can explain how is it leaky based on your understanding of version control?
EDIT: GitHub had probably the first UX where I actually liked what I got, compared to various earlier git based ones, or the horrible CVS and SVN ones where clicking on a file name definitely didn't do what you expected.
I barely remember Bitbucket from the mercurial era but I do recall that while it was better than some, it was still worse than GitHub, whose "here's default branch's HEAD, plus auto-opened and formatted README" was really a game changer
I also have used plenty of SCM systems since mid-90's.
They have less to walk away from then.
TBH learning git in depth was a hindrance for me learning jj - for my first few weeks I kept instinctively looking for complex solutions to simple problems and being confused, when the answer was always "use the tools you already know"
As a concrete example - git has one set of commands for managing commits, a different set of commands for managing the index, a third set of commands for managing stashes, a fourth set of commands for managing the working directory, and a fifth set of commands for managing conflicts -- in jj all of those things are a single concept with a single set of commands, equally (if not more) powerful, but also simpler
We have git.
jujutsu works with git.
git can do everything jujutsu can do (otherwise, jujutsu couldn't work with git).
Thus, jujutsu is a UX / new steering wheel.
ERSC is trying to be a GitHub competitor with what surplus value?
Don't get me wrong, GitHub has a slew of its own problems, but I have not seen any commentary on how ERSC addresses any of the downfalls GitHub has too. We're jumping from one frying pan into another, currently-cooler one?
What about SourceHut? Codeberg? They could offer the same steering wheels.
Edit: Thank you to everyone who replied. The enterprise/teams infrastructure angle makes sense.
If I had to guess, then I would say ERSC is trying to build a `jj` forge based on this new hire and that would be amazing imo.
While jj works with git, it doesn't have to. jj has the ability to work with many backends. Git is the popular one in the open source world, but that doesn't mean that other ones are not possible. And that's the core tech we have built.
With the rise of agents, more and more companies are running into the limits of what Git can do. It was never designed for large monorepos, and so there are performance issues, but also features that it's missing, that are important once you start trying to scale things up to an entire organization. Companies like Meta and Google ran into these things well before the present moment, but you don't have to be a FAANG these days to be running into this sort of stuff.
We also understand that incremental adoption is really important, especially for something as fundamental as source control. So our tech also has a git ingress/egress that will work with the core functionality, but more advanced things that git simply can't do will end up requiring using a non-git protocol. jj is the bridge between those two worlds.
I hope that makes sense! We have been in psudeo-stealth for a while now, this is the start of us coming out of our shell a bit and actually talking about what we're doing.
Normal Git/JJ is fine for most people, and most open source projects. It's just when you get to the scale of commercial software companies (100+ employees) you run into the classic monorepo/submodule problem.
Monorepos are currently the best solution - they make testing tractable, you can do cross-project changes, Git's support for them isn't buggy as hell. But they have obvious downsides - performance is the big one, but also integration with open source projects, permission control, etc.
So there's this big problem space that really only matters to companies where as far as I know there are currently no good open source or even commercial solutions. The only people that have (maybe) solved it are the mega-corps like Facebook and Google who have made their own private solutions. I don't know exactly what ERSC are going to offer but it's surely going to be in that space and also probably not going to be super relevant to private individuals.
You mention performance issues with Git, but there are performance issues with this blog post.
No one has agreed to this premise. You have invented it out of nowhere.
I’m not saying not to try or that there’s no value in turning the wheel once more. But I used to be excited about these kinds of things and now I’m not because I’ve seen the cycle a few times now.
But there are so many hooks into the Git workflow today, and I’m struggling to find a “user story” that would not be met by custom Git tooling. UX could be added to Git objects à la Gerrit with custom CLIs, custom Gitoxide ODB and (eventually) ref storage backends can help the server scale, and things like promisor remotes and partial clones can keep things smaller on the client side.
I’m curious what ERSC is able to share publicly about where the scaling fails with the Git specification? Is it the wire format, or the relative speed of upstreaming new things to Git? Or is the side-channel approach for storing rich data in Git objects insufficient in some way?
P.S. I wrote a blog post[1] about this a bit ago. I’m sure I’m missing something. Since I wrote it, I’ve wondered if the lack of deferred conflicts in-tree would be a scaling limiter.
Yes, it's important to realize that git is multiple things: the on-disk repo format, the CLI, and the wire protocol. Each has their own challenges, just like any project, but they can be tackled independently. But at the end of the day, there just are fundamental limits to the way that git does things that make total sense in the context Linus developed git for, but just do not in other ones. Some of those are performance related, but some of those are things like how the kernel is an open source project, and so per-directory ACLs are something that git wouldn't ever bother supporting, but are valuable for companies. We'll be talking about these things in the future, so I'm just going to be hand-wavy about it right now. But we do think that git is a really great system that has served us all well for 20 years, but just like how git displaced systems before it due to significantly re-thinking the details, for some users, making meaningful improvements over git will also require a similar significant re-think. That's just the way of technological development.
But that's good to hear.
Initially we won't be offering our own CI directly, you'll bring your own. We use Buildkite internally, so we'll probably share that integration first, but fundamentally you can do whatever you'd like. We won't require you to use whatever first-party offering we end up with.
I don't know how experienced you are with jj, but one thing I can say for sure is jj offers significant ux improvements over git and they plan to have a backend eventually. we might see the first implementation from ersc. and from their website I see they are already taking things like agents into account from day one. so we might be in the verge of getting something that scales better with modern day workload?
FWIW I think it's a bit more nuanced than that. As a trivial counter example Git does not currently have a way to store conflicts inside commits[1], which is one of our major features. The way this works is that we basically store the sides of the conflicting trees in the Git object database (and another conflicted tree pointing to all of them), but the actual data structure representing the conflict as Jujutsu understands it is stored elsewhere. Git will be very confused looking at a conflicted tree object created by Jujutsu -- it shouldn't break but it doesn't understand what it's seeing. So, if a object can be stored in the database that is correct but Git can't really "view" it properly, does "git" (or capital-G Git) support it or not? Depends on who you ask, I guess. The storage layout being intimately tied to Git "as a thing", part of its identity, is one of the most interesting properties it has.
The way I see it is this. Jujutsu uses the Git storage layout on disk. But basically all of the algorithms and UX between the two are completely different. That is where most of the real differences are. To the extent we integrate with lots of fiddly Git details (eg tracking branches, remote integration etc) so that `git` commands work -- that's more necessity to make the UX nice. We could absolutely use the on disk Git format but in a way that was inscrutible to the `git` CLI, it just obviously would not be very nice.
As a disclosure I work at ERSC but I am also one of the maintainers of Jujutsu; this post is written with my maintainer hat on.
[1] At Git Merge last year, Elijah Newren prototyped and demo'd a really early and cool example of storing first class conflicts in Git, based on his work on ort-merge. So this might not be true forever :)
Yes, I like how jj does it!
Did a big `git rebase -i` yesterday and I felt stressed up, almost a bit angry, when there were some conflicts and I wasn't sure if I'd resolve all of them before it was time to leave, and work more later from a different laptop.
But with jj, no stress: I can just check in the conflicts, and later, on the other laptop: check out the conflicts and continue resolving.
(I just wish there was Git submodule support :- ))
I use JJ locally and for those that call it a "porcelain" or a "git UI" are missing a lot. Sure you can use it that way, but what I've found is that it facilitates (and perhaps encourages) a pretty different workflow than I'd use with just plain Git. I could pull off what I do in JJ in just plain Git... but it wouldn't at all feel natural and you can really see this when you start using a JJ driven repository with Git tools (including just the Git command line tools).
Personally, and as a mere version control user, I prefer the JJ model of version control much more than Git's. The mechanics of the UI was very easy to get under my fingers (the essentials in about a day), but getting the ideas and really getting an understanding/appreciation for the conceptual differences and their ramifications on workflow took a month or two of regular working and trying things. For me, it was worth the effort.
Just like TypeScript with JS, QUIC with UDP, GraphQL with HTTP, Android with Linux, and so on
> git can do everything jujutsu can do (otherwise, jujutsu couldn't work with git)
Replace "jujutsu"/"git" by "typescript"/"js", "quic/udp", etc and you'll understand.
assembly can do everything [insert high level language] can do. why not just write everything in that?
For anyone curious the big thing with jj is you can undo. Basically if you were running a rebase and bungled a commit sha or branch name, no worries undo your way out, if you forgot to push but abandoned a commit not issues undo it. Same with all jj commands.
Delta db takes this a step further and everything you edit is a change so your actual edits are also all saved.
I find non destructive-ness for as long as possible approach very nice, since there aren't as many constraints on us not being able to do it nowadays, we had limited storage and stuff in 2000s perhaps but in 2026 even with the killer SSD prices it's fairly cheap to buy more SSD, we can definitely waste 100MB per project storing redundant history we may not ever need.
Same way in photoshop non destructive editing took over, especially useful for humans, but works for AI and LLMs too, they can undo their mistakes as well.
¹ https://drive.google.com/file/d/1dVzug1lHoOxdbFu8gcCJCu-G_uV...
Points of no return.
You worked on multiple things, you sync with the upstream, then boom now you have conflicts in one or more of your local branches.
Instead of immediately presenting you a conflict resolution view (usually 3 pane vimdiff) and you'll have to resolve at the sync time, the commits are simply marked as "have conflict", and you can continue working on other things until you have time to deal with it.
And for a chain of commits that all have conflicts, chances are fixing the first conflict automatically resolves all other conflicts in the chain, thanks to the auto-rebasing feature.
What happens if one is unlucky and the conflicts are not automatically resolved from changes to the first commit? Does that potentially imply more work overall than if one was using a VCS that forced conflict resolution earlier, like git?
Why is git's "forced conflict resolution" better, assuming the same workflow (local change+sync upstream) would produce the same # of conflicted commits?
The main advantages are that you can do the changes piecemeal, in any order, at any time, and by jumping around in the commit history however you need or want to.
Sometimes the best way to solve a conflict is by changing something before the conflict, or by moving pieces from one commit to another. Git doesn’t allow this, you can only do: fully fix commit 1, fully fix commit 2, fully fix commit 3, and so on.
2. Yeah submodules aren't supported but since you can use jj in a git repo you can commit git stuff via git like submodules and then enjoy the world of jj. Since submodule interactions should be rare imho. I use it for a bunch of projects with submodules.
For instance, if you are interactively using `git bisect` and you mark commits as good or bad, and you accidentally mark a commit incorrectly, you have to do something like:
git bisect log > /path/to/file.txt
git bisect reset
# modify file.txt to "look correct"
git bisect replay /path/to/file.txt
Another example is when you do something like screw up an interactive rebase. Let's say you rebase 20 commits and then you get a conflict on commit 8. You fix the file conflicts, and continue. You accidentally solve the merge incorrectly, continue and get another conflict -- but only realize your mistake after you start solving it. The reflog can't save you here. You have to completely abandon the rebase and start over. (This specific example might be handled better these days).
I think the biggest thing about `jj undo` is that it works everywhere. You can undo rebases, merges, conflict resolutions, copies, deletions, whatever. The secret behind it all is that internally, jj is architected in a way where implementing a feature looks like you are working with a transactional database. You actually have `begin_transaction()` and `commit()` methods in the codebase that will make changes to the commit graph visible in an atomic way. When a command like `jj rebase` happens, all of the changes it makes are inside a transaction and committed at once. Every operation in the repo is a transaction, and it all goes into a log, which records the effects of a transaction -- very much like a database system!
So "undo" just means "undo the effects recorded in a transaction" and that is about all. And so it works for everything! And this design is very easy to intuitively understand and program against, as a maintainer, along with our other high level internal APIs. Any developer can easily write code that Just Does The Right Thing and the user can undo it and it's no big deal. When I develop and work on Jujutsu myself -- like I'm actively developing new features or prototyping ideas -- I almost always _use my own jj repository_ as a test repo while testing my builds.
In contrast, Git does not have one unified "transactional" layer for things like this. But not all is lost, there has been work on 'git undo' and it was implemented by... Someone who is now a Jujutsu maintainer[2]. :)
[1] Technically we do not yet have "step by step" bisect with good/bad yet (only "automatic" bisect that is one-shot), so that is something Git can do we can't do at all right now, but bear with me. :')
Since jj snapshots the commit you're editing a lot, you can often even jj undo an accidental edit to a just-created file that in git would have still been uncommitted.
If you haven't tried jj yet, take it for a spin and see how you like it.
jj's data store is git. You can use jj while other people on your project just use git. There's no migration of data or history. You can use Github and all of the various git tooling that exists today.
As such, the initial cost of switching to jj is zero for individuals, teams, or projects. That's huge.
jj's ergonomics are significantly nicer. Fewer ways to shoot yourself in the foot, much better merge resolution, etc.
Where can I read on jj merge resolution strategies? LLM links to Arch Linux man page and other random stuff.
The jj tutorial is useful: https://docs.jj-vcs.dev/latest/tutorial/
It doesn't sound like much, but before I started contributing to Jujutsu I was a user of https://sapling-scm.com/ -- Sapling is fantastic. It stores data on disk in git format, but back then it was not exposed to the user by default -- the .git directory is hidden away. It actually ended up causing a lot of friction that random `git` commands would not work, or tools had other various git-based behaviors. These days, sapling has a "dotgit" mode (partially inspired by jj I assume) that puts the .git directory in the root dir, so `git` commands still work.
Beyond that though, speaking Git over the network only has other complications on the client side. In particular Git clients and servers negotiate what objects each side has (given what the client requests), and from that negotiation derives a list of needed objects to give the client, and then sends a packfile for it. This is all relatively expensive to do on demand, actually, so it would make the interactivity for network operations much worse if you need to wholly translate your storage into packed objects, etc. And doing it efficiently would require you keeping a cache around that is basically a git database anyway, so you might as well just go ahead and use it. Not the end of the world in terms of downsides, but a trade-off that adds some baggage.
There's no network effect headwinds, it's interoperable with git (I assume most people using jj have an upstream repo that's on git).
I have only played a little with jj and hg, but I felt that jj has a more modern UX, learning from years of Git existence (and even influencing Git, like the new `git history` command [1]). On the other hand, hg was created in the same month as Git, both trying to improve the experience of the VCSs available at the time (april 2005).
[1] https://lore.kernel.org/git/20250819-b4-pks-history-builtin-...
jj borrowed quite a few things from mercurial.
So, jj can be thought at least partially as a spiritual successor. Meta's Sapling is in a similar position. From this perspective, the key feature of both is transparent git repo and git remote support (hg can speak git as a foreign language, but it's limited)
And if your muscle memory doesn't allow an immediate transition, give yourself 2 weeks. That's how long it took for me at least.
Invest in custom tooling and helper scripts.
It will be good to see what jj will look like with more funded dev work, but I'm always a little worried about financial incentives mixing with the tools I use for the long term. I guess the saving grace is that I don't really need more upgrades to jj or jjui as it stands. I'm pretty much content with the features and so I could just save this copy of the repo for future reference.
As far as large assets goes, I have my own VCS-ish system which I just integrate with jj, but it would be nice to see a non-git backend handle large assets better as well.
As a side note, jj has rules around company representation in maintainership, and so when this initially happened a few months back, ERSC actually became over-represented in the maintainers group, as we also employ several other maintainers. The project decided to add more maintainers to keep the balance, rather than ask folks to step down.
We very much want to properly support the upstream project, and we want to make sure that it continues to be awesome for everyone, not just our customers.
The CLA means Google also has a license to the code to do what they will with it, but it does not do copyright assignment. They do not own your contributions.
> von Zweigbergk will continue to be a core maintainer of JJ as an open source project under the Apache 2.0 license.
Most of the maintainer stuff is also overhyped in a way (OSS maintainer-heads know this). The maintainer group is largely structured to just keep things moving; not really a privilege but a responsibility -- making sure things get merged, doing technical reviews, ensure builds are unstuck, and hitting the green approval button. Our own community members also do code review, find bugs, triage things, answer questions, design features, help get users unstuck, etc. We are very much not alone, and I have great confidence in the constitution and strength of the community itself as someone who has been around the block a few times.
This is going to sound silly and "guy approves of himself" adjacent, but I've been doing OSS maintainer stuff for like 2 decades at this point, and jj is truly a gem as a project. It is vibrant, and the community is generally speaking insanely talented (huge number of git powerusers), and we have a weird dogfood advantage where people use the tool to contribute right back to it. Martin did an awesome job setting the tone of the project just through his behavior before many of us joined years ago. He's a naturally good project lead.
As I said elsewhere: I work for ERSC and I am a Jujutsu maintainer. This is just me speaking candidly on my own behalf.
Once you start thinking in jj it becomes frustrating having think back to use GitHub as a code repo.
When do you need something like this?
⋆˚.⋆⋆˚.⋆
When I do need to advance a branch name, `jj bookmark advance` is easy enough. And it works great with the model that `jj git push` pushes all out-of-sync branches. In this model, advancing a bookmark is the reliable signal that I’m ready to share work, not simply “having authored a commit”.
We liked our old look, but one funny thing about how things are going in the world: our old color scheme was too close to what happens when you ask Claude to generate a site, and so even though we paid a human to build it, we didn't want to communicate slop, especially because that's not what we did!
Nick Noble is our designer, he put in a bunch of work on this, and I really like it too.
There will be a lot more of it soon :D
Small hint: A moving background layer makes reading supremely annoying, and for some people even (virtually) impossible.
Any other reccs?
I've filed a bug internally and we'll make sure to take a look. We don't use these animations in the product, only the company site, so this is the first time it's hitting a broader audience. I appreciate the report and sorry about that!
Looking forward to using the VCS stuff you'll create :- )
The things I am most interested in right now are jj and llms. And the intersection between the two. I'm thinking a lot these days about how agentic development is changing our industry, with source control being one part of that.
Edit: a fix is in review
Rust is the default language for all of our non-frontend web stuff, yes.
Is it the animation at the top?
Good choice in my opinion, as I hadn't even noticed the animation existed until I looked for something that might make the scrolling slow.
I opened up Activity Monitor and CPU use looked normal but it showed the GPU was pegged at 100% as long as that page was open.
(Martin joined back then, but we are posting about it today, so we should have the post date as today)
I wrote about why I joined the company back when I did previously here: https://steveklabnik.com/writing/i-see-a-future-in-jj/
In short: we are building a next generation source control system for the enterprise.
We should be respecting https://developer.mozilla.org/en-US/docs/Web/CSS/Reference/A... . If we do not, that's a bug, and please let me know what browser/OS/whatever else combo you have if that is an issue.
love the website though, i hope i one day will be able to make something that elegant!
Oh no! Looks like the ram shortage wasn’t enough, SSDs are next! /s