To me, it makes jujutsu look like the Nix of VCSes.
Not meaning to offend anyone: Nix is cool, but adds complexity. And as a disclaimer: I used jujutsu for a few months and went back to git. Mostly because git is wired in my fingers, and git is everywhere. Those examples of what jujutsu can do and not git sound nice, but in those few months I never remotely had a need for them, so it felt overkill for me.
Tbf you wouldn't use/switch to jj for (because of) those kind of commands, and are quite the outlier in the grand list of reasons to use jj. However the option to use the revset language in that manner is a high-ranking reason to use jj in my opinion.
The most frequent "complex" command I use is to find commits in my name that are unsigned, and then sign them (this is owing to my workflow with agents that commit on my behalf but I'm not going to give agents my private key!)
jj log -r 'mine() & ~signed()'
# or if yolo mode...
jj sign -r 'mine() & ~signed()'
I hadn't even spared a moment to consider the git equivalent but I would humbly expect it to be quite obtuse.
Actually, signing was one of the annoying parts of jujutsu for me: I sign with a security key, and the way jujutsu handled signing was very painful to me (I know it can be configured and I tried a few different ways, but it felt inherent to how jujutsu handles commits (revisions?)).
If you need to type in a password to unlock your keychain (e.g. default behavior for gpg-agent), then signing commits one at a time constantly is annoying.
Does "own" try to sign working copy snapshot commits too? That would greatly increase the number and frequency of signatures.
No, jj is super simple in daily use, in contrast with git that is a constant chore (and any sane person use alias). This include stuff that in git is a total mess of complexity like dealing with rebases. So not judge the tool for this odd case.
I don’t understand how people can remember all these custom scripting languages. I can’t even remember most git flags, I’m ecstatic when I remember how to iterate over arrays in “jq”, I can’t fathom how people remember these types of syntaxes.
I am convinced that the vast majority of professionals simply don't bother to remember and, ESPECIALLY WITH GIT, just look stuff up every single time the workflow deviates from their daily usage.
At this point perhaps a million person-years have been sacrificed to the semantically incoherent shit UX of git. I have loathed git from the beginning but there's effectively no other choice.
That said, the OP's commands are useful, I am copying them (because obviously I won't ever memorize them).
> At this point perhaps a million person-years have been sacrificed to the semantically incoherent shit UX of git. I have loathed git from the beginning but there's effectively no other choice.
Yes! We mostly wouldn’t tolerate the complexity and the terrible UX of a tool we use everyday--but there's enough Stockholm Syndrome out there where most of us are willing to tolerate it.
> I am convinced that the vast majority of professionals simply don't bother to remember and, ESPECIALLY WITH GIT, just look stuff up every single time the workflow deviates from their daily usage.
I wrote a cheat sheet in my notes of common commands, until they stuck in my head and I haven't needed it now for a decade or more. I also lean heavily on aliases and "self-documenting" things in my .bashrc file. Curious how others handle it. A search every time I need to do something would be too much friction for me to stand.
indeed, I held off for a while but finally caved because I got sick of seeing commits with `git commit -m .` littered in there. These are personal projects so I'm the only one dev-ing on them, but still so nice to have commit messages.
I’ve recently been looking into some tools that provide quick or painless help like pop up snippets with descriptions and cheat sheets, got any recommendations?
I don't, I will google things and fiddle, then put it in a git alias (with a comment on what it does and / or where I got it from) and push it to my private dotfiles repo, taking it with me between computers and projects.
You research it once, use it and then remember that it has "ancestor" in the command somewhere and then use ctrl + R to dig up something from your shell history.
Some things are idioms that one repeats so often they just stick, e.g. I use "grep.... | cut -c x-y | sort | uniq -c | sort -nr" to quickly grep frequency of some events from a log file.
Don't feel bad - no one remembers them all, we just remember a few idioms we use...
jj's template and revset languages are very simple syntactically, so once you're comfortable with the few things you do use often it's just a question of learning about the other existing functions (even if only enough to know to look them up), which slot right in and compose well with everything else you know (unlike flags which typically have each their own system).
Or, perhaps better yet, defining your own functions/helpers as you go for things you might care about, which, by virtue of having been named you, are much easier to remember (and still compose nicely).
If I look something up twice, I record it in Obsidian. If I need it more than a couple of times, I'll probably make an alias, a script or a mask [1] file. Autocomplete and autosuggest are essential to my workflow. And good history search.
If you don't have to codedive new projects all the time, there's zero reason to memorize these. If your job is to look at new codebases all the time, you probably learn to remember these commands pretty quickly.
a project isn’t dying because of no commits. Rather it’s stable
I often feel I need to setup bots to make superfluous commits just to make it look like my useful and stable repos are “active”
One example (not mine) a a qr-code generator library. Hasn’t been updated in 10 years. It’s perfect as is. It just provides the size and the bits. You convert those bits to any representation you want. It has no need to be updated
It's rare, I think, for a project to have such a well defined and singular purpose that has not changed in 10 years nor have any bugs been discovered or its dependencies changed underneath it.
It's not impossible, of course, but if I saw even a qr library that hadn't changed in 10 years I would worry that it wouldn't build on current systems (due to dependencies) and that nobody was actually using it (due to lag of bug reports).
I have several of those projects. I avoid dependencies as much as possible, striving to only use things which I know ship with my target OS. I code for a level of correctness and longevity. That benefits everyone, including myself.
A QR (or barcode) library is exactly the type of thing I’d assume would still work fine, since there’s nothing new to do, the parsing rules don’t change, it’s a static, known, solved problem.
> A QR (or barcode) library is exactly the type of thing I’d assume would still work fine, since there’s nothing new to do, the parsing rules don’t change, it’s a static, known, solved problem.
I agree with you - and yet the barcode library I used recently for a variable-data-printing project was last updated 13 hours ago, despite having been around since 2008!
Well said. Even an awesome library with no bugs that has no external dependencies still depends on the stdlib. For a while, before we were using containers, we even had the issue on Mac dev machines especially, where a half dozen Rubygems would crash while building its C extensions if your Mac OS version wasn’t just what the author expected, due to changes in the compiler shipped by Apple. So a MacOS major update might on its own functionally break a gem, even if the gem itself was designed well and you were using the same Ruby version.
This might be true for libraries or utilities that have a well-defined scope and no dependencies, but that's not what the article is focused on. When considering a company's main product, it's usually never done and patterns of activity—and especially changes in those patterns—can give you insight into potential issues.
> a project isn’t dying because of no commits. Rather it’s stable
Agreed. Assuming there are no open issues and PRs. When I find a project, if the date of the last commit is old, I next look at the issues and PRs. If there are simple-to-deal-with issues (e.g. a short question or spam) and easy-to-merge PRs (e.g. fixing a typo in the README) which have been left lingering for years, it’s probably abandoned. Looking at the maintainer’s GitHub activity graph should provide more clues.
> I often feel I need to setup bots to make superfluous commits just to make it look like my useful and stable repos are “active”
I have never done it, but a few times thought about making a “maintenance” release to bump the version number and release date, especially since I often use a variant of calendar versioning.
I don't want to program git, I want to get stuff done so I would reject using that tool and do what the article author did running tried and true pipeable Linux/UNIX commands. It's also the same reason why I dislike Gradle and use Maven, I don't want to program my build I want to define and run my build.
But the git commands in the article is also programming of the same kind, just using more terse, more obscure language. All the shell pipelines are sort, uniq, and grep.
A language that properly maps to the data model, and has readable identifiers is a boon. Git is a database, a database needs a proper query language.
Hah someone really looked at jq (?) and thought: "yes, more of this everywhere". I feel jq is like marmite (edit: aka vegemite, i.e. "you either love it or you hate it")
It's really not that bad, although the jq comparison might be apt. You have such primitives you need to understand, and then everything just fits together nicely. I find this much easier to write and understand than git's cryptic format strings.
I love how the author thinks developers write commit messages.
All joking aside, it really is a chronic problem in the corporate world. Most codebases I encounter just have "changed stuff" or "hope this works now".
It's a small minority of developers (myself included) who consider the git commit log to be important enough to spend time writing something meaningful.
AI generated commit messages helps this a lot, if developers would actually use it (I hope they will).
This is a team lead/CTO problem. A good leader will be explicit in their expectations that developers write good commit messages. I've certainly had good leaders that expect this.
Yes, and a culture problem, too. I guess I've been blessed that I've mostly only worked for "grown up" companies, but I've never encountered a workplace where people didn't write useful commit messages. At least one line description of the work done, but often multiple lines of valuable context. Only the junior devs had to be told to do it, but once they got into the habit, everyone understood why we do it and it was no big deal.
If I joined a company where people committed their code with "stuff" or "made some changes" or "asdfhlfo;ejfo;ae," that would be a red flag that I might have joined the wrong company, and I'd start to wonder what else the developers here do carelessly.
Indeed. If you can't spend two minutes (MAX) writing a sentence or two explaining what the commit is for, then what are we doing as developers? Commits are for future you and your future team. They are a tool. Please, use them.
Good commit messages would be nice but honestly I would be over the moon if our pull requests would be approved within a week without having to ping one or more people.
I once tutored an intern. Who thought he was The Best Programmer On Earth (didn't we all at that age?).
He refused to use revision control, it slowed him down.
So we told him to commit at least once every day, with a relevant commit message, or else fail his internship.
He worked 21 more days. There were 21 commits: "17:00, time to go home".
This reads like the intern was left to his own devices and his output not checked at all for three weeks straight. Actual tutoring would have surfaced the issue after 1 or 2 days tops.
I don't agree. These things actually matter. A developer who isn't told otherwise is just going to do whatever they feel like, so if there is nothing or no-one enforcing the standards, then the failure isn't on the individual developer, it is on the team lead. Someone needs to be setting the standards.
In the company I work for, there is a team that has isolated itself to some extent from other teams and works at a furious pace to keep their particular section of the business happy. We're lucky enough that they spun up their own repo to do their work on, so they don't actually impact other teams, but if the quality of the commit messages is anything to go by, I am 100% certain they're going to end up in a huge mess, if they aren't already. The team lead encourages this, and certainly doesn't care about commit messages etc.
Developers who care about other developers tend to write better quality code, because they care what other developers think of them. If you care about other developers, you will most likely write decent quality commit messages too.
I have seen over many years the types of developers who only care about moving their own code into production as fast as possible and getting to the next thing. There is a very high correlation with a mess at the end, which they inevitably won't have to tidy up because they'll be doing the next thing. These types of developers hate owning stuff in production, so they don't do it, so they don't actually care how maintainable their "clever" code is. I am very certain that a number of people reading this will be those types of developers.
Useless? So you never use “git annotate” or your IDE to see who wrote a line of code whose purpose puzzles you, and go to the commit message to see what they were trying to accomplish? This is invaluable to me as long as commit messages are clear.
As a manager, one of the first things I do is make sure that the PR titles (the PR text becomes the commit messages in squash-merging workflows) at minimum begin with a ticket number. Then we can later read both the intention and the commentary on it.
> Useless? So you never use “git annotate” or your IDE to see who wrote a line of code whose purpose puzzles you, and go to the commit message to see what they were trying to accomplish?
Personally no, the code is the "truth". If I need more I'm going to open a dialog with the author, not spend time trying to interpret a 7 word commit message, "good" or otherwise.
The code can only convey what is being done (and then, in some cases, only superficially). It can't convey what decisions were made, what alternatives were discarded, what business motivations may have led to that code.
And for old enough code, the author may not be available, or more likely doesn't remember.
I partially disagree. Technical leadership at the micro/mid level should be able to set and enforce standards like "you must have semi-meaningful or meaningful commit messages." If and only if they set those standards, and the team does not follow them, then we can say that either the leadership is lacking, or there is a structural barrier/disincentive to following the rules. Within that framework, I do think using process-smells like this is valid for judging technical leadership.
To the point of other commenters however, I wouldn't lay something this micro at the foot of the CTO in all but the smallest of organizations.
It's a stretch to lay at the CTOs feet, but not the team lead or even Head/VP of Engineering IMHO. It's also easy to "enforce" if you're already doing peer review (which you definitely should be, even if not required for compliance).
In codebases where PRs are squashed on merge, the commit messages on the main branch end up being the PR body description text, and that's actually reviewed so tends to be much better I find.
And in every codebase I've been in charge of, each PR has one or more issue # linked which describe every possible antagonizing detail behind that work.
I understand this isn't inline with traditional git scm, but it's a very powerful workflow if you are OK with some hybridization.
I personally find this to be a substantially better pattern. That squashed commit also becomes the entire changeset - so from a code archeology perspective it becomes much easier to understand what and why. Especially if you have a team culture that values specific PRs that don’t include unrelated changes. I also find it thoroughly helpful to be able to see the PR discussions since the link is in the commit message.
I agree, much as it's a loss for git as a distributed system (though I think that ship sailed long ago regardless). As far as unrelated changes, I've been finding that another good LLM use-case. Like hey Claude pull this PR <link> and break it up into three new branches that individually incorporate changes A, B, and C, and will cleanly merge in that order.
One minor nuisance that can come up with GitHub in particular when using a short reference like #123 is that that link breaks if the repo is forked or merged into something else. For that reason, I try to give full-url references at least when I'm manually inserting them, and I wish GitHub would do the same. Or perhaps add some hidden metadata to the merge commit saying "hey btw that #123 refers to <https://github.com/fancy-org/omg-project/issues/123>"
Our small team has a lot of commit messages like this. For a while, we had a guy on the team who had come from a site that expected more. The pet peeve he brought along was that commit messages end with a period (my guess is that someone at their previous work place had reasoned that forcing periods encouraged developers to actually write meaningful sentences). When I look at that period of development, I see lots of messages like “stuff changed.” And “more stuff changed.” And then it goes back to just “stuff changed” around the time they moved on.
Only two of the five depend on commit messages. Churn, authorship, and velocity work regardless. Even teams with terrible hygiene write "fix" when something breaks.
Is it really a small minority? I have never worked on a project that didn't have commit messages that at least tried to be descriptive (sometimes people fail at it but its very different to an outright "changed stuff").
I don't remember any friend mentioning to me them encountering a work project where the messages were totally neglected either.
Never seen that in any company I worked at either and I can’t believe professional developers seem to think that it would be ok to write meaningless commit messages. That’s just so sloppy.
tbh I'm not convinced that a git log history should be treated as a group journal because it's not.
relying on git commit messages assumes they're correct by convention since there is no technical constraint to enforce it. and it assumes no work in progress commits, sometimes it's just necessary to hit the save button real quick or move a workspace from one device to another.
my point is:
git is a way of storing and loading files at its core.
Only two of the five insights are based on commit messages and the author acknowledges that they won't work in projects without message discipline. But the remaining ones will give you valuable insights even into the most lazy project department.
It's because the vast majority of commit messages are never read by anyone, and there's other ways to fund out what happened in the handful of cases where you would need to.
> One caveat: squash-merge workflows compress authorship. If the team squashes every PR into a single commit, this output reflects who merged, not who wrote. Worth asking about the merge strategy before drawing conclusions.
I abhor squash merging for this and a few other reasons. I literally have to go out of my way to re-check out a branch. Someone who wants to use my current branch cannot do so if I merge my changes a month later, because the squash rewrites history, and now git is very confused. I don't get the obsession with "cleaning up the history" as if we're all always constantly running out of storage over 2 more commits.
For me the benefit is that I can revert or cherry-pick things one entire PR at a time, and I don't have to care if the author implemented their PR with a bunch of small "work in progress" commits.
And GitHub at least sets the author of the squashed commit as the one who opened the PR, not the one who merged it.
I can definitely see where it wouldn't work well for other workflows but I've had it work well on several teams and it seems easier than trying to get everyone to clean up their commits into nice, clean, well-titled histories before putting up a PR.
I ran these commands on a number of codebases I work on and I have to say they paint a very different picture than the reality I know to be true.
> git shortlog -sn --no-merges
Is the most egregious. In one codebase there is a developer's name at the top of the list who outpaced the number 2 by almost 3x the number of commits. That developer no longer works at the company? Crisis? Nope, the opposite. The developer was a net-negative to the team in more ways than one, didn't understand the codebase very well at all, and just happened to commit every time they turned around for some reason.
Everything in context. This is one of many reasons I'm a proponent of squash-and-merge. If a change really needs more than one permanent commit, it should probably be split up or if absolutely necessary should be on a feature branch maintaining similar process. Under this process, feature branches are not squashed.
This leaves developers to commit locally and comment as much or little as they like.
I'm not saying more commits = bad developer. In my example that happened to be the case but not because they had a lot of commits but because they were bad at their job. I was just trying to warn that taking these git snippets at face-value does not paint the full picture.
If someone came to me and said "I ran these and I see XXX was the most prolific committer and they left X months ago, what will be do???" I'd have to work hard not to laugh.
Since these snippets are self-described as ways to get familiar with the code/projects I wanted to provide the counter point. Most of those snippets do not at all paint the real picture and for all the repos I tested it on they paint the opposite of reality.
I know these codebases like the back of my hand, the purported purpose of these snippets is to better understand the codebase, I can tell you they don't work for anything I tested them on. Maybe they work for other codebases but the sample size I have access to says they don't work for me.
I haven't finished the article yet but I think your point is an important one, and that's to run the commands with a context in mind. The article seems to be coming from the perspective of somebody who is brand new to the project, and as your experience indicates, interviewing teams and leads before running those commands might add more understanding to what they're telling you.
> The 20 most-changed files in the last year. The file at the top is almost always the one people warn me about. “Oh yeah, that file. Everyone’s afraid to touch it.”
The most changed file is the one people are afraid of touching?
I've just tried this, and the most touched files are also the most irrelevant or boring files (auto generated, entry-point of the service etc.) in my tests.
I just tried it too and it basically just flagged a handful of 1500+ line files which probably ought to be broken up eventually but arent causing any serious problems.
If it's (like in my case) dependency management, localization or config files, breaking them up will likely only cause more issues. Make sure that it's an actual improvement before breaking things up.
Plotting Churn against Complexity is far more useful than merely churn.
It shows places that are problematic much better. High churn, low complexity: fine. Its recognized and optimizef that this is worked on a lot (e.g. some mapping file, a dsl, business rules etc). Low churn high complexity: fine too. Its a mess, but no-one has to be there.
But both? Thats probably where most bugs originate, where PRs block, where test coverage is poor and where everyone knows time is needed to refactor.
In fact, quite often I found that a teams' call "to rewrite the app from scratch" was really about those few high-churn-high-complexity modules, files or classes.
Complexity is a deep topic, but even simple checks like how nested smt is, or how many statements can do.
This command needs a warning. Using this command and drawing too many conclusions from it, especially if you’re new, will make you look stupid in front of your team mates.
I ran this on the repo I have open and after I filtered out the non code files it really can only tell me which features we worked on in the last year. It says more about how we decided to split up the features into increments than anything to do with bugs and “churn”.
I found it interesting, that Git itself has built in similarity notion... when it packs objects, it groups files by path+size, runs delta cmpression to find which are close.
Not really strong enough in a post about what to do in a codebase you’re not familiar with. In that situation you’re probably new to the team and organisation and likely to get off on the wrong foot with people if you assume their code “hurts”.
The post is “here’s what I do”, not “here’s what you should do and then confront the team about the results.” It’s just showing you a quick way to get some insights. It’s not even guaranteeing it’s accurate, just showing you some things you might be able to draw some quick conclusions on.
I’m not sure why HN attracts this need to poke holes in interesting observations to “prove” they aren’t actually interesting.
Better for people to know they're just blindly copying tools and parroting their output as if it's automatically meaningfully. Any warning against that should be built into the individual, for their own sake
Definitely not in my experience. The most changed are the change logs, files with version numbers and readmes. I don't think anyone is afraid of keeping those up to date.
Yes. Because the fear is butressed with necessity. You have to edit the file, and so does everyone else and that is a recipe for a lot of mess. I can think back over years of files like this. Usually kilolines of impossible to reason about doeverything.
Curious - why write it as a function in presumably .gitconfig and not just a git-summary script in your path? Just seems like a lot of extra escapes and quotes and stuff
I wouldn't trust "commit counts." The quality and content of a "commit" can vary widely between developers. I have one guy on my team who commits only working code that has been thoroughly tested locally, another guy who commits one line changes that often don't work, only to be followed by fixes, and more fixes. His "commits" have about 1/100th of the value of the first guy.
My comment still seems relevant? Do frequent commits to correct mistakes imply more "value" than infrequent, but well tested, commits, or what? I don't think it is a reliable signal.
> One caveat: squash-merge workflows compress authorship. If the team squashes every PR into a single commit, this output reflects who merged, not who wrote. Worth asking about the merge strategy before drawing conclusions.
In my experience, when the team doesn't squash, this will reflect the messiest members of the team.
The top committer on the repository I maintain has 8x more commits than the second one. They were fired before I joined and nobody even remembers what they did. Git itself says: not much, just changing the same few files over and over.
Of course if nobody is making a mess in their own commits, this is not an issue. But if they are, squash can be quite more truthful.
Solid list. I'd add git log --all --oneline --graph pretty early on — gives you a quick sense of how active different branches are and whether this is a "one person commits everything" project or actually distributed. Helped me a ton on a job where I inheritied a monolith with like 4 years of history.
The git blame tip is underrated. People treat it like a gotcha tool but its maybe the fastest way to find the PR/ticket that explains a weird decision.
Yeah, this one demonstrates a particularly pernicious view of software development. One where growth, no matter how artificial, is the only sign of success.
If you work with service oriented software, the projects that are "dying" may very well be the most successful if it's a key component. Even from a business perspective having to write less code can also be a sign of success.
I don't know why this was overlooked when the churn metric is right there.
Whenever we initiated a new (internal) SW project, it had to go through an audit. One of the items in the checklist for any dependency was "Must have releases in the last 2 years"
I think the rationale was the risk of security vulnerabilities not being addressed, but still ...
That was my question too. I have plenty of projects I've worked on where they rarely get touched anymore. They don't need new features and nothing is broken.
Sometimes you need to bump a dependency version, adjust the code to a changed API endpoint, or update a schema. Even if the core features stay the same, there's some expected maintenance. I'd still call that being worked on, in a sense that someone has to do it.
Technically you're correct that change frequency doesn't necessarily mean dead, but the number of projects that are receiving very few updates because they're 'done' is a fraction of a fraction of a percent compared to the number that are just plain dead. I'm certain you can use change frequency as a proxy and never be wrong.
That sort of project exists in an ocean of abandoned and dead projects though. For every app that's finished and getting one update every few years there are thousands of projects that are utterly broken and undeployable, or abandoned on Github in an unfinished state, or sitting on someone's HDD never be to touched again. Assuming a low change frequency is a proxy for 'dead' is almost always correct, to the extent that it's a reasonable proxy for dead.
I know people win the lottery every week, but I also believe that buying a lottery ticket is essentially the same as losing. It's the same principle.
With respect, this is a myopic view. Not all software is an "app" or a monolith. If you use a terminal, you are directly using many utilities that by this metric are considered dying or dead.
Rather than using an LLM to write fluffy paragraphs explaining what each command does and what it tells them, the author should have shown their output (truncated if necessary)
and it touches in detail what exactly commit standards should be, and even how to automate this on CI level.
And then I also have idea/vision how to connect commits to actual product/technical/infra specs, and how to make it all granular and maintainable, and also IDE support.
I would love to see any feedback on my efforts. If you decide to go through my entire 3 posts I wrote, thank you
Most good projects end up solving a problem permanently and if there is no salary to protect with bogus new features it is then to be considered final?
These were interesting but I don't know if they'd work on most or any of the places I've worked. Most places and teams I've worked at have 2-3 small repos per project. Are most places working with monorepos these days?
For "what changes the most", in my project it's package.json / lock (because of automatic dependency updates) and translation / localization files; I'd argue that's pretty normal and healthy.
For the "bus factor", there's one guy and then there's me, but I stopped being a primary contributor to this project nearly two years ago, lol.
> If the team squashes every PR into a single commit, this output reflects who merged, not who wrote.
Squash-merge workflows are stupid (you lose information without gaining anything in return as it was easily filterable at retrieval anyway) and only useful as a workaround for people not knowing how to use git, but git stores the author and committer names separately, so it doesn't matter who merged, but rather whether the squashed patchset consisted of commits with multiple authors (and even then you could store it with Co-authored-by trailers, but that's harder to use in such oneliners).
Can you explain to me (an avid squash-merger) what extra information do you gain by having commits that say "argh, let's see if this works", "crap, the CI is failing again, small fix to see if it works", "pushing before leaving for vacation" in the main git history?
With a squash merge one PR is one commit, simple, clean and easy to roll back or cherry-pick to another branch.
These commits reaching the reviewer are a sign of either not knowing how to use git or not respecting their time. You clean things up and split into logical chunks when you get ready to push into a shared place.
In some cases, reviewing PR diffs commit-by-commit (and with the logs as the narration of the diff-by-diff story) is a substantial improvement over reviewing the entire PR diff. Concrete examples...
* A method or function that has code you realize needs to be shared...the code may need to be moved and also modified to accommodate its shared purpose. Separating the migration from any substantive modifications allows you to review the migration commit with the assistance of git's diff.colorMoved feature. It becomes easier to understand what changes are due to the migration, and what changes were added for more effective sharing.
* PRs sometimes contain mechanical work that is easy to review in isolation. Added or removed arguments, function renames, etc. No big deal if it's two or three instances, but if it's dozens or hundreds of instances, it's easier for the humans to review all of those consistent changes together, rather than having them mixed in with other things one has to reason about.
* Sometimes a flow of commits can help follow a difficult chain of reasoning. PR developer claims that condition X can never occur, but the code is complex enough that it's difficult to verify. However, by transforming the code in targeted ways that are possible to reason about, the complexity might be reducible to the point where the claim becomes obvious. One frequent example I see of this is of function/method arguments that are actually unnecessary, but it wasn't obvious until after some code transformations.
Because it's a useful abstraction. If you only look at PRs and don't ever care about commits, why are they even being sent to reviewer in the first place? Just send a diff file.
Having atomic commits lets you actually benefit from having them. Suddenly you don't have to perform weird dances with interconnected PRs with dependencies as "PR too big" is not such a problem anymore as long as commits are digestible; you can have things property bisectable; you can preserve shared authorship; you can range-diff and have a better view on what and how changed between review passes, and so on...
The unit of change is commit, and PRs group commits you want someone to pull. If you don't want or need any of that, you're just sending a patch file in a needlessly elaborate way.
> If you only look at PRs and don't ever care about commits, why are they even being sent to reviewer in the first place? Just send a diff file.
This is in fact what hg does with amending changesets and yes it works far better. Keep PRs small and atomic and you never need to worry about what happens intra-pr. If you need bigger units of work that's what stacking is for.
Stacking is good for expressing dependencies, but isn't helpful when you need to make several distinct changes that aren't necessarily needed unless you take them all in. What's the value in having a separate PR that introduces a framework that you later use in another PR when you may not actually want to merge it if the latter one doesn't end up being merged as well?
A PR is a group of commits, just utilize that when you need it.
>Swap to diff and start reading through the changes
this forces the reviewer to view the entire diff at once, which can greatly increase the cognitive load vs. being able to view diffs of logical units of work
for tiny PRs it may not matter, but for substantial PRs it can matter a lot
You can, but instead you can also just squash merge in one click. And avoid that people merge there dozens of fixes if you allow anything but squash merge.
Nothing is destroyed by a force push. It just overwrites a single pointer, and even keeps its old value in reflog.
Things that aren't referenced by anything anymore will eventually get garbage collected and actually destroyed, but you can just keep a reference somewhere to prevent that from happening if you need. Or even disable garbage collection completely.
Looks like people's fears about git come just from not knowing what it does.
You can't use the remote reflog to revert what you force pushed, can you? But I agree that having your local reflog means you're never totally lost. I still just make a branch before major edits so I can go back.
if you mean better messages, it's not really that. those junk messages should be rewritten and if the commits don't stand alone, merged together with rebase. it's the "logical chunks" the parent mentioned.
it's hard to say fully, but unless a changeset is quite small or otherwise is basically 0% or 100%, there are usually smaller steps.
like kind of contrived but say you have one function that uses a helper. if there's a bug in the function, and it turns out to fix that it makes a lot more sense to change the return type of the helper, you would make commit 1 to change the return type, then commit 2 fix the bug. would these be separate PRs? probably not to me but I guess it depends on your project workflow. keeping them in separate commits even if they're small lets you bisect more easily later on in case there was some unforseen or untested problem that was introduced, leading you to smaller chunks of code to check for the cause.
If the code base is idempotent, I don't think showing commit history is helpful. It also makes rebases more complex than needed down the line. Thus I'd rather squash on merge.
I've never considered how an engineer approaches a problem. As long as I can understand the fundamental change and it passes preflights/CI I don't care if it was scryed from a crystal ball.
This does mean it is on the onus of the engineer to explain their change in natural language. In their own words of course.
Commits don't show "how an engineer approaches a problem". Commits are the unit of change that are supposed to go into the final repository, purposefully prepared by the engineer and presented for review. The only thing you do by squashing on merge is to artificially limit the review unit to a single commit to optimize the workflow towards people who don't know how to use git. Personally I don't think it's a good thing to optimize for.
Preserving commit history pre-merge only seems useful if I had to revert or rebase onto an interstitial commit. This is at odds with treating PRs as atomic changes to the code base.
I might have not stated my position correctly. When I mean "squash on merge", I mean the commit history is fully present in the PR for full scrutiny. Sometimes commits may introduce multiple changes and I can view commit ranges for each set of changes. But it takes the summation of the commits to illustrate the change the engineer is proposing. The summation is an atomic change, thus scrutinizing terms post-merge is meaningless. Squashing preserves the summation but rids of the terms.
Versioned releases on main are tagged by these summations, not their component parts.
Haha, good luck working with a team with more than 2 people. A good reviewer looks at the end-state and does not care about individual commits. If im curious about a specific change i just look at the blame.
> A good reviewer looks at the end-state and does not care about individual commits.
Then I must be a bad reviewer. In a past job, I had a colleague who meticulously crafted his commits - his PRs were a joy to review because I could go commit by commit in logical chunks, rather than wading through a single 3k line diff. I tried to do the same for him and hope I succeeded.
And then someone comments on a thing, they change it and force-push another "clean" history on top and all of your work is wasted because the PR is now completely different =)
You can, but most of us work in Github and having a PR to dump commits to is very easy and convenient. Then, when you get some feedback on your PR, you can dump more commits and it's very easy for the reviewer to see what has changed since the last time they reviewed it.
I feel like what you're arguing is that you should clean up your commits before anyone else sees them. Fair. But you could also clean it up right before merging to main. It's not that different, except the latter is much less annoying, particularly when going back and forth with people.
I know this is a very github centric workflow, but that's where most engineers work now, and it's nice and easy. This wouldn't work for eg: contributing to linux, but that's not what most of us do.
Why is it "wade through" if there are 10 clearly distinct but dependent commits, but comfortable if it's 10 stacked PRs instead? They are basically the same thing, presented ever so slightly differently.
I think in most teams I've worked with, the majority of developers (> 85%) barely undestand what Git is doing or what things mean inside GitHub, have never seen commit history as a graph, have never run something like "git log --oneline --graph --decorate" or "--format", and have never heard of "git range-diff" which is very useful for following commit/PR/unit changes.
Personally I review using "git" itself, so I see the graph structure either way, and there's little difference between stacked PRs, commit chains in a single PR, or even feature branches, from that point of view. Even force-push branch updatea aren't difficult to review, because of the reflog and "git range-diff". The differences are mainly in what kinds of behaviour the web-based tooling promotes in the rest of the team, which does matter, and depends on the team.
I agree with you if you're using Graphite instead of GitHub. Having a place to give feedback and/or approval on the individual "units" (commits in a PR, or PRs in a stack) is useful, grouping dependent but distinct changes is useful, and diff'd commit evolution within each unit PR in response to back-end-forth review feedback is useful in some collaborative settings. Though, if you know "git range-diff" and reflog, that shows diff'd commit evolution quite well.
In GitHub, people are confused by stacked PRs both conceptually and due to the GitHub UX around them. Most times when I've posted a stacked PR to a GitHub project, other people didn't realise it was stacked, and occasinally someone else has merged the tip of a stack made by me, and been surprised to see all the dependent PRs merged automatically as a side effect. Usually before they get to reviewing those other PRs :-)
People understand commit sequences in a PR, though I've rarely seen people treat the individual commits as units for review when using GitHub, unfortunately. In the Linux kernel world where Git was born, the PR flow is completely different from GitHub: Their system tends to result in feedback on individual commits. It also encourages better quality feedback, with less nitpicking, and better quality commits.
Some of these commits even get reviewed by different maintainers before being merged, which is common when a patchset touches several subsystems at once.
Part of new feature you had working in an intermediate commit, but broke somewhere along the way and is not working in your last commit when you squashed.
If you catch it early enough, I suppose it's in your reflog, but otherwise you're screwed.
It sounds like a silly example, but I bet most developers have run into this at some point.
With mercurial/jujutsu, you get the best of both worlds: The "argh, let's see if this works" commits are what I call "microcommits", and the squashed versions are the real/public commits. With jujutsu, you get both. Your log shows only the "real" commits (equivalent of squashing all the commits between that and the prior "real" commit). But if you want to drill down into the microcommits, the information is always there.
Let's acknowledge the reality. Many people use git not just for version control, but for backup ("Let me commit this so I don't lose it"). Let's ensure the VC tool supports both and doesn't force you to pick one over the other.
If someone uses git commits like the save function of their editor and doesn't write messages intended for reading by anyone else, it makes sense to want to hide them
For other cases, you lose the information about why things are this way. It's too verbose to //comment on every like with how it came to be this way but on (non-rare in total, but rare per line) occasion it's useful to see what the change was that made the line be like this, or even just who to potentially ask for help (when >1 person worked on a feature branch, which I'd say is common)
> If someone uses git commits like the save function of their editor
I use it like that too and yet the reviewers don't get to see these commits. Git has very powerful tools for manipulating the commit graph that many people just don't bother to learn. Imagine if I sent a patchset to the Linux Kernel Mailing List containing such "fix typo", "please work now", "wtf" patches - my shamelessness has its limits!
Seems like a lot of extra effort (save, add, commit, come up with some message even if it's a prayer to work now) only to undo it again later and create a patch or alternate history out of the final version. Why bother with the intermediate commits if you're not planning for it to be part of the history?
Git is a version control system. It does not care about what it versions.
When I work on something, I commit often and use the commit graph as a undo tool on steroids. I can see what I tried, I can cherry-pick or revert stuff while experimenting, I can leave promising but unfinished stuff to look at later, or I can just commit to have a simple way to send stuff to CI, or a remote backup synced between machines.
Once I'm done working on something, it's time to take a step back, look at the big picture, see how many changes my experiments have actually yielded, separate them, describe and decide whether they go to review together or split in some way, as sometimes working on a single thing requires multiple distinct changes (one PR with multiple commits), but sometimes working in a single session yields fixes for multiple unrelated issues (several PRs). Only then it gets presented to the reviewer.
It just happens that I can do both these distinct jobs with a single tool.
Because I might want to go back to this current messy state but I don't want to commit it like this (hardcoded test strings, debug logs, cutted corners to see if something works, you name it).
I simply commit something like "WIP: testing xy" and if its working and properly implemented i can squash/rebase/edit the commit message and force push it to my feature branch.
Using a Git client like Gitkraken makes this incredibly easy, takes seconds.
This way I can leverage version control without committing bogus states to the final PR.
If the team is using a PR workflow, the PR is a working place to produce one single commit. The individual commits are just timestamped changes and comments. Think of it as the equivalent of annotated diff in mailing list conversation.
You gain the extra information by having reasonable commit messages rather than the ones you mentioned. To fix CI you force push.
Can you explain to me what an avid squash-merger puts into the commit message of the squashed commit composed of commits "argh, let's see if this works", "crap, the CI is failing again, small fix to see if it works", and "pushing before leaving for vacation" ?
The squashed commit from the PR -> main will have a clean title + description that says what was added.
Usually pretty close to what the PR title + description are actually, just without the videos and screenshots.
Example:
feat(ui): Add support for tagging users
* Users can be tagged via the user page
* User tags visible in search results (configurable)
etc..
I don't need to spend extra time cleaning up my git commits and force-pushing on the PR branch, losing context for code reviews etc. Nor does anyone have to see my shitty angry commits when I tried to figure out why Playwright tests ran on my machine and failed in the CI for 10 commits.
> "argh, let's see if this works", "crap, the CI is failing again, small fix to see if it works", "pushing before leaving for vacation"
These are all bad commits IMHO. Aside from the CI one, I understand that message. I have commits like that on personal projects but for professional projects I'd be frustrated if people were committing messages like that.
Personally I'm a "one commit" type of guy, I don't like committing things in a broken state even on a side branch unless I have to (to share the code or test a CI). Occasionally I will make multiple commits at the very end to make review easier or once I have everything working but I want to try something different but I have a bunch of options of saving code that don't involving committing:
- Stash
- Shevle (IDEA)
- Backblaze
- Time Machine
- Local History (IDEA)
The idea of committing WIP before leaving for a vacation just feels so wrong to me.
I once worked for someone who wanted developers to commit code before the end of every day as a safety measure. His reasoning was in case the developer's computer died or similar. I found that silly at the time and still do now. That's what backups are for, I dislike when people use git as a backup like that in a professional setting.
Squash merge is the only reasonable way to use GitHub:
If you update a PR with review feedback, you shouldn’t change existing commits because GitHub’s tools for showing you what has changed since your last review assume you are pushing new commits.
But then you don’t want those multiple commits addressing PR feedback to merge as they’re noise.
So sure, there’s workflows with Git that doesn’t need squashing. But they’re incompatible with GitHub, which is at least where I keep my code today.
Is it perfect? No. But neither is git, and I live in the world I am given.
If your goal here is to have linear history, then just use a merge commit when merging the PR to main and always use `git log --first-parent`. That will only show commits directly on main, and gives you a clean, linear history.
If you want to dig down into the subcommits from a merge, then you still can. This is useful if you are going back and bisecting to find a bug, as those individual commits may hold value.
You can also cherry pick or rollback the single merge commit, as it holds everything under it as a single unit.
This avoids changing history, and importantly, allows stacked PRs to exist cleanly.
Yes, I think people who are anti squash merge are those who don't work in Github and use a patch based system or something different. If you're sending a patch for linux, yes it makes sense that you want to send one complete, well described patch. But Github's tooling is based around the squash merge. It works well and I don't know anyone in real life who has issues with it.
And to counter some specific points:
* In a github PR, you write the main commit msg and description once per PR, then you tack on as many commits as you want, and everyone knows they're all just pieces of work towards the main goal of the eventually squashed commit
* Forcing a clean up every time you make a new commit is not only annoying extra work, but it also overwrites history that might be important for the review of that PR (but not important for what ends up in main branch).
* When follow up is requested, you can just tack on new commits, and reviewers can easily see what new code was added since their last review. If you had to force overwrite your whole commit chain for the PR, this becomes very annoying and not useful to reviewers.
* In the end, squash merge means you clean up things once, instead of potentially many times
Forcing a single commit per PR is the issue imo. It's a lazy solution. Rebase locally into sensible commits that work independently and push with lease. Reviewers can reset to remote if needed.
The author is talking about the case where you have coherent commits, probably from multiple PRs/merges, that get merged into a main branch as a single commit.
Yeah, I can imagine it being annoying that sqashing in that case wipes the author attribution, when not everybody is doing PRs against the main branch.
However, calling all squash-merge workflows "stupid" without any nuance.. well that's "stupid" :)
I don't think there's much nuance in the "I don't know --first-parent exists" workflow. Yes, you may sometimes squash-merge a contribution coming from someone who can't use git well when you realize that it will just be simpler for everyone to do that than to demand them to clean their stuff up, but that's pretty much the only time you actually have a good reason to do that.
* git merge ALWAYS does a merge and git pull ALWAYS does a fast forward.
* git log --first-parent is the default. Have a git log --deep if you want to go down into branches.
If you use a workflow that always merges a PR with a merge commit, then git log --first-parent gives you a very nice linear history. I feel like if this was the default, so many arguments about squashing or rebasing workflows wouldn't be necessary to get our "linear history", everyone would just be doing merges and be happy with it. You get a clean top level history and you can dig down into the individual commits in a merge if you are bisecting for a bug.
Do people actually share PR as in different people contributing to the same branch?
Also I can understand not squashing if the contribution comes from outside the organization. But in that case, I would expect a cleaned up history. But if every contribution is from members of the team, who can merge their own PR, squash merge is an easy way to get a clean history. Especially when most PR should be a single commit.
We do. If we are building out a feature, none of its code is merged into main until it is complete (if this is a big feature, we milestone into mergeable and releasable units).
The feature is represented by a Story in Jira and a feature branch for that story. Subtasks in jira are created and multiple developers can pick up the different subtasks. There is a personal branch per subtasks, and PRs are put up against the feature branch. Those subtasks are code reviewed, tested, and merged into the feature branch.
In the end, it is the feature branch that is merged (as a single merge commit and complete unit) into main, and may well have had contributions from multiple people.
I get your POV, but I’ve always considered that long-lived branches in the canonical repo (the one in the forge) other than the main one should be directly related to deployable artifacts. Anything else should be short-lived.
There can be experiment on the side that warrants your approach, but the amounts of merge going back and forth would make this hard to investigate (especially when blaming) I would prefer to have one single commit with a message that describe every contribution.
I think the point is that if you have to squash, the PR-maker was already gitting wrong. They should have "squashed" on their end to one or more smaller, logically coherent commits, and then submitted that result.
It’s not “having to squash”. The intent was already for a PR to be a single commit. I could squash it on my end and merge by rebasing, but any alteration would then need to be force-pushed. So I don’t bother. I squash-merge when it’s ready and delete the branch.
Squash-merge is entirely fine for small PRs. Cleaning up the commits in advance (probably to just squash them to one or two anyway) is extra work, and anything that discourages people from pushing often (to get the code off their local machine) needs to be well-justified. Just review the (smallish!) total outcome of all the commits and squash after review. A few well-placed messages on the commit, attached to relevant lines, are more helpful and less work than cleaning up the commit history of a smallish PR.
For really large PRs, I’m more inclined to agree with you, but those should probably have their own small-PR-and-squash-merge flow that naturally cleans up their git history, anyway.
I categorically disagree that squash-merge is “stupid” but agree there are many ways to skin this cat.
How does not squash merging deal with the fact that branches disappear when merging? What I mean is that the information "this commit happened in the context of this PR or this overarching goal" goes missing. When you squash, you use the one central unit of information management in Git: the commit.
Having the commit graph easy to filter means exactly that you don't have to sift through hundreds of commits for no reason. What else did you think it would mean?
I just finished¹ building an experimental tool that tries to figure out if a repo is slopware or not just by looking at it's git history (plus some GitHub activity data).
The takeaway from my experiment is that you can really tell a lot by how / when / what people commit, but conclusions are very hard to generalize.
For example, I've also stumbled upon the "merge vs squash" issue, where squashes compress and mostly hide big chunks of history, so drawing conclusions from a squashed commit is basically just wild guessing.
(The author of course has also flagged this. But I just wanted to add my voice: yeah, careful to generalize.)
Out of curiosity, I ran the 5 command on my project's public git tree. The only informative one was #4 ("Is This Project Accelerating or Dying") - it showed cliffs when significant pieces of logic were decoupled and moved to other repos.
Trusting the messages to contain specific keywords seems optimistic. I don't think I used "emergency" or "hotfix" ever. "Revert" is some times automatically created by some tools (E.g. un-merging a PR).
For the stuff I've worked on, if you want to know about bugfixes and emergency releases, you'd go to Jira where those values are formalized as fields. Someone else in the comments here had a suggestion which just looks for the word "fix" which would definitely capture some bugfix releases, but is more likely to catch fixes that were done during development of a feature.
These commands are very useful, but adapting them to the codebase makes a huge difference.
For most, I added some filters and slightly changed the regex, and it showed the reality of the codebase (I already knew the reality, I just wanted to see if it matched, and it did).
My team usually uses "Squash and merge" when we finish PRs, so I feel that would skew the results significantly as it hides 99% of the commit messages inside the long description of the single squashed merge commit.
Some readme files include changelogs. But aside from that I think this can still net some useful information. I like to look at the most recently changed files in a repo as well.
Fair point. I skip lockfiles, changelogs, and generated code. The first application file on the list is the one that matters. Should have been explicit about that in the post.
It’s easy enough to filter those out with grep. It still is relatively meaningless. If the team incrementally adds things then it’s just going to show what additions were made. It isn’t churn at all.
superficial. If I have to unfuck the backend 10 times a week in our API adapter, then these commands will show me constantly changing the API adapter, although it's the backend team constantly fixing their own bugs
This way I can see right away which branches are 'ahead' of the pack, what 'the pack' looks like, and what is up and coming for future reference ... in fact I use the 'gss' alias to find out whats going on, regularly, i.e. "git fetch --all && gss" - doing this regularly, and even historically logging it to a file on login, helps see activity in the repo without too much digging. I just watch the hashes.
I'm so used to magit, it seems kind of primitive to pipe git output around like this.
Anyway, I can glean a lot of this information in a few minutes scrolling through and filtering the log in magit, and it doesn't require memorizing a bunch of command line arguments.
No searching the codebase/commits for "fuck" and shit"? That will give you an idea what what was put in under stressful circumstances like a late night during a crunch.
> The 20 most-changed files in the last year. The file at the top is almost always the one people warn me about. “Oh yeah, that file. Everyone’s afraid to touch it.”
I've got my Emacs set up to display next to every file that is versioned the number of commits that file has been modified in (for the curious: using a modified all-the-icons-ivy-rich + custom elisp code + custom Bash scripts I wrote and it's trickier than it seems to do in a way that doesn't slows everything down). For example in the menu to open a file or open a recently visited file etc.: basically in every file list, in addition to its size, owner, permissions, etc. I also add the number of commits if it's a versioned file.
I like the fix/bug/broken search in TFA to see where the bugs gather.
I was curious what information I could glean from these for some popular repos. Caveat: I'm primarily an low-level embedded developer so I don't interface with large open source projects at the source level very often (other than occasionally the linux kernel). I chose some projects at random that I use.
*Mainline linux*
Most changed files: pretty much what I expected for 1 and 2... the "cutting edge" of Linux development over other OSes -- bpf and containers. The bpf verifier and AMD GPU driver might get a boost in this list due to sheer LoCs in those files (26K and 14K respectively). An intel equivalent of amdgpu_dm is #21 in the list (drivers/gpu/drm/i915/display/intel_display.c) and nvidia is nowhere to be seen (presumably due to out-of-tree modules/blobs?).
10399 Christoph Hellwig -> I only know his name because of drama last year regarding rust bindings to DMA subsystem
8481 Mauro Carvalho Chehab -> I also know his name from the classic "Mauro, shut the fuck up!" Linus rant
8413 Takashi Iwai -> Listed as maintainer for sound subsystem, I think he manages ALSA
8072 Al Viro -> His name is all over bunch of filesystem code
Buggy files: Intel comes out on top of GPU drivers this time (twice). Along with KVM for x86(64), the main allocator, and BTRFS.
Buggy files: DWARF debuginfo generation, x86 heuristics tables, RS6000(?!) heuristic tables. I had to look up RS6000, it's an IBM instruction set from the 90s lol. cp-tree.h is an interesting file, it seems be the main C(++) AST datastructures.
*xfwm4*
Most changed files: the list is dominated by *.po localizations. I filtered these out. Even after this, I discovered there is very little active development in the last few years. If I extend to 4 years ago, I get:
1. src/client.c - Realizing this project is too "small" to glean much from this. client.c is just the core X client management code. Makes sense.
2. src/placement.c - Other core window management code.
This has not told me much other than where most of the functionality of this project lies.
Bus factor: Pretty huge. Not really an issue in this case due to lack of development I guess.
Files with bug commits: Very similar distribution to most changed files. Not enough datapoints in this one to draw any big conclusions.
I think these massive open projects (excl xfwm) are generally pretty consistent code quality across the heavily trodden areas because of the amount of manpower available to refactor the pain points. I've yet to see an example of "god help you if you have to change that file" in e.g. linux, but I have of course seen that situation many times in large proprietary codebases.
Big projects tend to self-correct. These commands hit differently on private codebases with 3-10 contributors, where high-churn usually means one person patching the same thing repeatedly.
Jujutsu equivalents, if anyone is curious:
What Changes the Most
Who Built This Where Do Bugs Cluster Is This Project Accelerating or Dying How Often Is the Team Firefighting Much more verbose, closer to programming than shell scripting. But less flags to remember.To me, it makes jujutsu look like the Nix of VCSes.
Not meaning to offend anyone: Nix is cool, but adds complexity. And as a disclaimer: I used jujutsu for a few months and went back to git. Mostly because git is wired in my fingers, and git is everywhere. Those examples of what jujutsu can do and not git sound nice, but in those few months I never remotely had a need for them, so it felt overkill for me.
Tbf you wouldn't use/switch to jj for (because of) those kind of commands, and are quite the outlier in the grand list of reasons to use jj. However the option to use the revset language in that manner is a high-ranking reason to use jj in my opinion.
The most frequent "complex" command I use is to find commits in my name that are unsigned, and then sign them (this is owing to my workflow with agents that commit on my behalf but I'm not going to give agents my private key!)
I hadn't even spared a moment to consider the git equivalent but I would humbly expect it to be quite obtuse.Actually, signing was one of the annoying parts of jujutsu for me: I sign with a security key, and the way jujutsu handled signing was very painful to me (I know it can be configured and I tried a few different ways, but it felt inherent to how jujutsu handles commits (revisions?)).
The only reasonable way to use signing in jj is with the sign-on-push config https://docs.jj-vcs.dev/latest/config/#automatically-signing... rather than as commits are made
Why? I have my signing behavior set to own and I haven't noticed any issues, but I don't actually rely on signatures for much.
If you need to type in a password to unlock your keychain (e.g. default behavior for gpg-agent), then signing commits one at a time constantly is annoying.
Does "own" try to sign working copy snapshot commits too? That would greatly increase the number and frequency of signatures.
Ah, I use my SSH key to sign my commits and I don't have a password on my SSH key.
> Does "own" try to sign working copy snapshot commits too?
Yes
No, jj is super simple in daily use, in contrast with git that is a constant chore (and any sane person use alias). This include stuff that in git is a total mess of complexity like dealing with rebases. So not judge the tool for this odd case.
[delayed]
I don’t understand how people can remember all these custom scripting languages. I can’t even remember most git flags, I’m ecstatic when I remember how to iterate over arrays in “jq”, I can’t fathom how people remember these types of syntaxes.
I am convinced that the vast majority of professionals simply don't bother to remember and, ESPECIALLY WITH GIT, just look stuff up every single time the workflow deviates from their daily usage.
At this point perhaps a million person-years have been sacrificed to the semantically incoherent shit UX of git. I have loathed git from the beginning but there's effectively no other choice.
That said, the OP's commands are useful, I am copying them (because obviously I won't ever memorize them).
> At this point perhaps a million person-years have been sacrificed to the semantically incoherent shit UX of git. I have loathed git from the beginning but there's effectively no other choice.
Yes! We mostly wouldn’t tolerate the complexity and the terrible UX of a tool we use everyday--but there's enough Stockholm Syndrome out there where most of us are willing to tolerate it.
> I am convinced that the vast majority of professionals simply don't bother to remember and, ESPECIALLY WITH GIT, just look stuff up every single time the workflow deviates from their daily usage.
I wrote a cheat sheet in my notes of common commands, until they stuck in my head and I haven't needed it now for a decade or more. I also lean heavily on aliases and "self-documenting" things in my .bashrc file. Curious how others handle it. A search every time I need to do something would be too much friction for me to stand.
I just use Claude Code as a terminal for git these days. It writes up better commit messages than I would write anyway. No more "git commit -m fix"
indeed, I held off for a while but finally caved because I got sick of seeing commits with `git commit -m .` littered in there. These are personal projects so I'm the only one dev-ing on them, but still so nice to have commit messages.
I’ve recently been looking into some tools that provide quick or painless help like pop up snippets with descriptions and cheat sheets, got any recommendations?
Cheaters: https://brettterpstra.com/projects/cheaters
Navi is good for generating personal cheatsheets:
https://github.com/denisidoro/navi
But for Git, I can't recommend lazygit enough. It's an incredible piece of software:
https://github.com/jesseduffield/lazygit
I've found tldr to be useful
https://github.com/tldr-pages/tldr
I don't, I will google things and fiddle, then put it in a git alias (with a comment on what it does and / or where I got it from) and push it to my private dotfiles repo, taking it with me between computers and projects.
You research it once, use it and then remember that it has "ancestor" in the command somewhere and then use ctrl + R to dig up something from your shell history.
Some things are idioms that one repeats so often they just stick, e.g. I use "grep.... | cut -c x-y | sort | uniq -c | sort -nr" to quickly grep frequency of some events from a log file.
Don't feel bad - no one remembers them all, we just remember a few idioms we use...
jj's template and revset languages are very simple syntactically, so once you're comfortable with the few things you do use often it's just a question of learning about the other existing functions (even if only enough to know to look them up), which slot right in and compose well with everything else you know (unlike flags which typically have each their own system).
Or, perhaps better yet, defining your own functions/helpers as you go for things you might care about, which, by virtue of having been named you, are much easier to remember (and still compose nicely).
Same, but now with AI I don't have to remember that anymore
If I look something up twice, I record it in Obsidian. If I need it more than a couple of times, I'll probably make an alias, a script or a mask [1] file. Autocomplete and autosuggest are essential to my workflow. And good history search.
[1] https://github.com/jacobdeichert/mask
> I don’t understand how people can remember all these custom scripting languages.
We can't.
Why do you think the `man` command exists?
Nobody does. One person figures it out, then writes a blog post, and we all Google for it. Even git’s man pages are long and sometimes cryptic.
Yeah especially with git. All I know is pull, add, commit, push. Everything else I have to look up.
You add them into your GIT config file as shortcuts?
If you have multiple machines (/must have), just apply your user config to current machine?
If you don't have to codedive new projects all the time, there's zero reason to memorize these. If your job is to look at new codebases all the time, you probably learn to remember these commands pretty quickly.
a project isn’t dying because of no commits. Rather it’s stable
I often feel I need to setup bots to make superfluous commits just to make it look like my useful and stable repos are “active”
One example (not mine) a a qr-code generator library. Hasn’t been updated in 10 years. It’s perfect as is. It just provides the size and the bits. You convert those bits to any representation you want. It has no need to be updated
It's rare, I think, for a project to have such a well defined and singular purpose that has not changed in 10 years nor have any bugs been discovered or its dependencies changed underneath it.
It's not impossible, of course, but if I saw even a qr library that hadn't changed in 10 years I would worry that it wouldn't build on current systems (due to dependencies) and that nobody was actually using it (due to lag of bug reports).
I have several of those projects. I avoid dependencies as much as possible, striving to only use things which I know ship with my target OS. I code for a level of correctness and longevity. That benefits everyone, including myself.
A QR (or barcode) library is exactly the type of thing I’d assume would still work fine, since there’s nothing new to do, the parsing rules don’t change, it’s a static, known, solved problem.
> A QR (or barcode) library is exactly the type of thing I’d assume would still work fine, since there’s nothing new to do, the parsing rules don’t change, it’s a static, known, solved problem.
I agree with you - and yet the barcode library I used recently for a variable-data-printing project was last updated 13 hours ago, despite having been around since 2008!
Well said. Even an awesome library with no bugs that has no external dependencies still depends on the stdlib. For a while, before we were using containers, we even had the issue on Mac dev machines especially, where a half dozen Rubygems would crash while building its C extensions if your Mac OS version wasn’t just what the author expected, due to changes in the compiler shipped by Apple. So a MacOS major update might on its own functionally break a gem, even if the gem itself was designed well and you were using the same Ruby version.
This might be true for libraries or utilities that have a well-defined scope and no dependencies, but that's not what the article is focused on. When considering a company's main product, it's usually never done and patterns of activity—and especially changes in those patterns—can give you insight into potential issues.
> a project isn’t dying because of no commits. Rather it’s stable
Agreed. Assuming there are no open issues and PRs. When I find a project, if the date of the last commit is old, I next look at the issues and PRs. If there are simple-to-deal-with issues (e.g. a short question or spam) and easy-to-merge PRs (e.g. fixing a typo in the README) which have been left lingering for years, it’s probably abandoned. Looking at the maintainer’s GitHub activity graph should provide more clues.
> I often feel I need to setup bots to make superfluous commits just to make it look like my useful and stable repos are “active”
I have never done it, but a few times thought about making a “maintenance” release to bump the version number and release date, especially since I often use a variant of calendar versioning.
I don't want to program git, I want to get stuff done so I would reject using that tool and do what the article author did running tried and true pipeable Linux/UNIX commands. It's also the same reason why I dislike Gradle and use Maven, I don't want to program my build I want to define and run my build.
But the git commands in the article is also programming of the same kind, just using more terse, more obscure language. All the shell pipelines are sort, uniq, and grep.
A language that properly maps to the data model, and has readable identifiers is a boon. Git is a database, a database needs a proper query language.
I can't remember all of this, does anyone know of any LLM model trained on CLI which can be run locally?
If you copy those commands into a file and use that file to prompt the “sh” LLM.
That works until you need a small variation of any of these commands and you’re lost.
Try https://cheat.sh/
Not a model, but a product: warp.dev
Hah someone really looked at jq (?) and thought: "yes, more of this everywhere". I feel jq is like marmite (edit: aka vegemite, i.e. "you either love it or you hate it")
It doesn't seem any more egregious than something like:
`git log --color --graph --pretty=format:'%Cred%h%Creset -%C(yellow)%d%Creset %s %Cgreen(%cr) %C(bold blue)<%an>%Creset' --abbrev-commit --`
Which is something I see a lot of people alias in Git for viewing logs.
It's really not that bad, although the jq comparison might be apt. You have such primitives you need to understand, and then everything just fits together nicely. I find this much easier to write and understand than git's cryptic format strings.
Disclaimer: I love jq too :)
I love how the author thinks developers write commit messages.
All joking aside, it really is a chronic problem in the corporate world. Most codebases I encounter just have "changed stuff" or "hope this works now".
It's a small minority of developers (myself included) who consider the git commit log to be important enough to spend time writing something meaningful.
AI generated commit messages helps this a lot, if developers would actually use it (I hope they will).
This is a team lead/CTO problem. A good leader will be explicit in their expectations that developers write good commit messages. I've certainly had good leaders that expect this.
Yes, and a culture problem, too. I guess I've been blessed that I've mostly only worked for "grown up" companies, but I've never encountered a workplace where people didn't write useful commit messages. At least one line description of the work done, but often multiple lines of valuable context. Only the junior devs had to be told to do it, but once they got into the habit, everyone understood why we do it and it was no big deal.
If I joined a company where people committed their code with "stuff" or "made some changes" or "asdfhlfo;ejfo;ae," that would be a red flag that I might have joined the wrong company, and I'd start to wonder what else the developers here do carelessly.
Indeed. If you can't spend two minutes (MAX) writing a sentence or two explaining what the commit is for, then what are we doing as developers? Commits are for future you and your future team. They are a tool. Please, use them.
Good commit messages would be nice but honestly I would be over the moon if our pull requests would be approved within a week without having to ping one or more people.
Some of my favorite are the perhaps well-meaning but totally misguided log of what files they changed.
I once tutored an intern. Who thought he was The Best Programmer On Earth (didn't we all at that age?). He refused to use revision control, it slowed him down.
So we told him to commit at least once every day, with a relevant commit message, or else fail his internship.
He worked 21 more days. There were 21 commits: "17:00, time to go home".
This reads like the intern was left to his own devices and his output not checked at all for three weeks straight. Actual tutoring would have surfaced the issue after 1 or 2 days tops.
I think it's a stretch to measure leadership quality on something so minor, a lot of teams find them pretty useless no matter how good they are.
I don't agree. These things actually matter. A developer who isn't told otherwise is just going to do whatever they feel like, so if there is nothing or no-one enforcing the standards, then the failure isn't on the individual developer, it is on the team lead. Someone needs to be setting the standards.
In the company I work for, there is a team that has isolated itself to some extent from other teams and works at a furious pace to keep their particular section of the business happy. We're lucky enough that they spun up their own repo to do their work on, so they don't actually impact other teams, but if the quality of the commit messages is anything to go by, I am 100% certain they're going to end up in a huge mess, if they aren't already. The team lead encourages this, and certainly doesn't care about commit messages etc.
Developers who care about other developers tend to write better quality code, because they care what other developers think of them. If you care about other developers, you will most likely write decent quality commit messages too.
I have seen over many years the types of developers who only care about moving their own code into production as fast as possible and getting to the next thing. There is a very high correlation with a mess at the end, which they inevitably won't have to tidy up because they'll be doing the next thing. These types of developers hate owning stuff in production, so they don't do it, so they don't actually care how maintainable their "clever" code is. I am very certain that a number of people reading this will be those types of developers.
Useless? So you never use “git annotate” or your IDE to see who wrote a line of code whose purpose puzzles you, and go to the commit message to see what they were trying to accomplish? This is invaluable to me as long as commit messages are clear.
As a manager, one of the first things I do is make sure that the PR titles (the PR text becomes the commit messages in squash-merging workflows) at minimum begin with a ticket number. Then we can later read both the intention and the commentary on it.
> Useless? So you never use “git annotate” or your IDE to see who wrote a line of code whose purpose puzzles you, and go to the commit message to see what they were trying to accomplish?
Personally no, the code is the "truth". If I need more I'm going to open a dialog with the author, not spend time trying to interpret a 7 word commit message, "good" or otherwise.
The code can only convey what is being done (and then, in some cases, only superficially). It can't convey what decisions were made, what alternatives were discarded, what business motivations may have led to that code.
And for old enough code, the author may not be available, or more likely doesn't remember.
Fine, but none of that is in a normal commit message, lets be real...
And what of the original author is not there anymore?
The world will not end. I’ll get there.
I partially disagree. Technical leadership at the micro/mid level should be able to set and enforce standards like "you must have semi-meaningful or meaningful commit messages." If and only if they set those standards, and the team does not follow them, then we can say that either the leadership is lacking, or there is a structural barrier/disincentive to following the rules. Within that framework, I do think using process-smells like this is valid for judging technical leadership.
To the point of other commenters however, I wouldn't lay something this micro at the foot of the CTO in all but the smallest of organizations.
It's a stretch to lay at the CTOs feet, but not the team lead or even Head/VP of Engineering IMHO. It's also easy to "enforce" if you're already doing peer review (which you definitely should be, even if not required for compliance).
I particularly love when the “CTO” is also the main offender.
In codebases where PRs are squashed on merge, the commit messages on the main branch end up being the PR body description text, and that's actually reviewed so tends to be much better I find.
And in every codebase I've been in charge of, each PR has one or more issue # linked which describe every possible antagonizing detail behind that work.
I understand this isn't inline with traditional git scm, but it's a very powerful workflow if you are OK with some hybridization.
I personally find this to be a substantially better pattern. That squashed commit also becomes the entire changeset - so from a code archeology perspective it becomes much easier to understand what and why. Especially if you have a team culture that values specific PRs that don’t include unrelated changes. I also find it thoroughly helpful to be able to see the PR discussions since the link is in the commit message.
I agree, much as it's a loss for git as a distributed system (though I think that ship sailed long ago regardless). As far as unrelated changes, I've been finding that another good LLM use-case. Like hey Claude pull this PR <link> and break it up into three new branches that individually incorporate changes A, B, and C, and will cleanly merge in that order.
One minor nuisance that can come up with GitHub in particular when using a short reference like #123 is that that link breaks if the repo is forked or merged into something else. For that reason, I try to give full-url references at least when I'm manually inserting them, and I wish GitHub would do the same. Or perhaps add some hidden metadata to the merge commit saying "hey btw that #123 refers to <https://github.com/fancy-org/omg-project/issues/123>"
If developers don't write commit messages, that's a culture problem. At my company we demand that of each other.
Our small team has a lot of commit messages like this. For a while, we had a guy on the team who had come from a site that expected more. The pet peeve he brought along was that commit messages end with a period (my guess is that someone at their previous work place had reasoned that forcing periods encouraged developers to actually write meaningful sentences). When I look at that period of development, I see lots of messages like “stuff changed.” And “more stuff changed.” And then it goes back to just “stuff changed” around the time they moved on.
Only two of the five depend on commit messages. Churn, authorship, and velocity work regardless. Even teams with terrible hygiene write "fix" when something breaks.
As noted, authorship does not if commits are squashed, which seems to be common (I never do it).
> Even teams with terrible hygiene write "fix" when something breaks.
They might not include anything but the Jira ticket number, if the environment is truly lacking.
> It's a small minority
Is it really a small minority? I have never worked on a project that didn't have commit messages that at least tried to be descriptive (sometimes people fail at it but its very different to an outright "changed stuff").
I don't remember any friend mentioning to me them encountering a work project where the messages were totally neglected either.
Never seen that in any company I worked at either and I can’t believe professional developers seem to think that it would be ok to write meaningless commit messages. That’s just so sloppy.
I personally use git commit -m "." for: "Just snapshots this state real quick" on a feature branch.
main branch is advanced on PR level, with squashed commits.
So the "." should never make it to main, and have PR description as commit message.
Sometimes it could be just a ticket number/title
tbh I'm not convinced that a git log history should be treated as a group journal because it's not.
relying on git commit messages assumes they're correct by convention since there is no technical constraint to enforce it. and it assumes no work in progress commits, sometimes it's just necessary to hit the save button real quick or move a workspace from one device to another.
my point is: git is a way of storing and loading files at its core.
Bad commit messages always fail PR review for me. It requires will and discipline, but it's not that hard.
Commit messages are often squashed after merging a feature branch
Only two of the five insights are based on commit messages and the author acknowledges that they won't work in projects without message discipline. But the remaining ones will give you valuable insights even into the most lazy project department.
> AI generated commit messages
git log --oneline and a sprinkle of your personal sauce on .claude goes a long way :)
I love how the commentator thinks a developer makes decisions based on commit messages.
Random, subjective, or written in a state of mental exhaustion commit messages.
I also love the switcheroo the author made: git not logs. But hey :)
It's because the vast majority of commit messages are never read by anyone, and there's other ways to fund out what happened in the handful of cases where you would need to.
git blame's are amazing, and rely on good comments.
> One caveat: squash-merge workflows compress authorship. If the team squashes every PR into a single commit, this output reflects who merged, not who wrote. Worth asking about the merge strategy before drawing conclusions.
I abhor squash merging for this and a few other reasons. I literally have to go out of my way to re-check out a branch. Someone who wants to use my current branch cannot do so if I merge my changes a month later, because the squash rewrites history, and now git is very confused. I don't get the obsession with "cleaning up the history" as if we're all always constantly running out of storage over 2 more commits.
For me the benefit is that I can revert or cherry-pick things one entire PR at a time, and I don't have to care if the author implemented their PR with a bunch of small "work in progress" commits.
And GitHub at least sets the author of the squashed commit as the one who opened the PR, not the one who merged it.
I can definitely see where it wouldn't work well for other workflows but I've had it work well on several teams and it seems easier than trying to get everyone to clean up their commits into nice, clean, well-titled histories before putting up a PR.
I ran these commands on a number of codebases I work on and I have to say they paint a very different picture than the reality I know to be true.
> git shortlog -sn --no-merges
Is the most egregious. In one codebase there is a developer's name at the top of the list who outpaced the number 2 by almost 3x the number of commits. That developer no longer works at the company? Crisis? Nope, the opposite. The developer was a net-negative to the team in more ways than one, didn't understand the codebase very well at all, and just happened to commit every time they turned around for some reason.
Everything in context. This is one of many reasons I'm a proponent of squash-and-merge. If a change really needs more than one permanent commit, it should probably be split up or if absolutely necessary should be on a feature branch maintaining similar process. Under this process, feature branches are not squashed.
This leaves developers to commit locally and comment as much or little as they like.
Also - be careful of automated workflow that uses a single persons credential.. skews this by a lot.
So that person, on one central codebase at a company I work for, is me.
Assuming I'm not ego-mad, I like to think this is because I built the project from the ground up before handing it over to the rest of the team.
These days other people commit more often than I do, but my name is still dominant, and probably will be for some time.
I'm not saying more commits = bad developer. In my example that happened to be the case but not because they had a lot of commits but because they were bad at their job. I was just trying to warn that taking these git snippets at face-value does not paint the full picture.
If someone came to me and said "I ran these and I see XXX was the most prolific committer and they left X months ago, what will be do???" I'd have to work hard not to laugh.
Since these snippets are self-described as ways to get familiar with the code/projects I wanted to provide the counter point. Most of those snippets do not at all paint the real picture and for all the repos I tested it on they paint the opposite of reality.
I know these codebases like the back of my hand, the purported purpose of these snippets is to better understand the codebase, I can tell you they don't work for anything I tested them on. Maybe they work for other codebases but the sample size I have access to says they don't work for me.
I haven't finished the article yet but I think your point is an important one, and that's to run the commands with a context in mind. The article seems to be coming from the perspective of somebody who is brand new to the project, and as your experience indicates, interviewing teams and leads before running those commands might add more understanding to what they're telling you.
> The 20 most-changed files in the last year. The file at the top is almost always the one people warn me about. “Oh yeah, that file. Everyone’s afraid to touch it.”
The most changed file is the one people are afraid of touching?
Just like that place that's so crowded nobody goes there anymore.
I've just tried this, and the most touched files are also the most irrelevant or boring files (auto generated, entry-point of the service etc.) in my tests.
Yeah same thing happens with lockfiles and CI configs. You end up filtering out half the list before it tells you anything useful.
I just tried it too and it basically just flagged a handful of 1500+ line files which probably ought to be broken up eventually but arent causing any serious problems.
If it's (like in my case) dependency management, localization or config files, breaking them up will likely only cause more issues. Make sure that it's an actual improvement before breaking things up.
Plotting Churn against Complexity is far more useful than merely churn.
It shows places that are problematic much better. High churn, low complexity: fine. Its recognized and optimizef that this is worked on a lot (e.g. some mapping file, a dsl, business rules etc). Low churn high complexity: fine too. Its a mess, but no-one has to be there. But both? Thats probably where most bugs originate, where PRs block, where test coverage is poor and where everyone knows time is needed to refactor.
In fact, quite often I found that a teams' call "to rewrite the app from scratch" was really about those few high-churn-high-complexity modules, files or classes.
Complexity is a deep topic, but even simple checks like how nested smt is, or how many statements can do.
This command needs a warning. Using this command and drawing too many conclusions from it, especially if you’re new, will make you look stupid in front of your team mates.
I ran this on the repo I have open and after I filtered out the non code files it really can only tell me which features we worked on in the last year. It says more about how we decided to split up the features into increments than anything to do with bugs and “churn”.
I found it interesting, that Git itself has built in similarity notion... when it packs objects, it groups files by path+size, runs delta cmpression to find which are close.
Very different from just counting commits - https://vectree.io/c/delta-compression-heuristics-and-packfi...
Good thing that the article contains that warning, then.
Not really strong enough in a post about what to do in a codebase you’re not familiar with. In that situation you’re probably new to the team and organisation and likely to get off on the wrong foot with people if you assume their code “hurts”.
The post is “here’s what I do”, not “here’s what you should do and then confront the team about the results.” It’s just showing you a quick way to get some insights. It’s not even guaranteeing it’s accurate, just showing you some things you might be able to draw some quick conclusions on.
I’m not sure why HN attracts this need to poke holes in interesting observations to “prove” they aren’t actually interesting.
It’s a bit reductive to call it poking holes. The author shared his valuable knowledge and I shared mine.
These commands are just about what files to start looking at to understand new codebase.
Better for people to know they're just blindly copying tools and parroting their output as if it's automatically meaningfully. Any warning against that should be built into the individual, for their own sake
Definitely not in my experience. The most changed are the change logs, files with version numbers and readmes. I don't think anyone is afraid of keeping those up to date.
Yes. Because the fear is butressed with necessity. You have to edit the file, and so does everyone else and that is a recipe for a lot of mess. I can think back over years of files like this. Usually kilolines of impossible to reason about doeverything.
Yeah, the truth is going to be a lot more subtle than this.
The LLM that wrote the copy is an idiot.
This is such obvious LLM slop.
Could be also that a frequently edited file had most opportunity to be broken. And it was edited by the most random crowd.
In my case, it's .github/CODEOWNERS.
Nobody is afraid of changing it.
Why does github owners need frequent change? Do members in you team change so often?
I have a summary alias that kind of does similar things
EDIT: props to https://github.com/GitAlias/gitaliasCurious - why write it as a function in presumably .gitconfig and not just a git-summary script in your path? Just seems like a lot of extra escapes and quotes and stuff
It's a very old config that I copied from someone many years ago, agree that it's a bit hard to parse visually.
Not the poster, but one theory: so you only need to copy one file. Portability.
Looks like the above assumes a POSIX shell, so one could argue a dedicated script would actually be more portable.
Looks nice. Unfortunately I don't have log-of-count-and-email, log-of-count-and-day or churn
+1 Same.
Edit. https://github.com/mattrighetti/dotfiles/blob/master/.gitcon...
You could make a local `man` page.
Some nice ideas but the regexes should include word boundaries. For example:
git log -i -E --grep="\b(fix|fixed|fixes|bug|broken)\b" --name-only --format='' | sort | uniq -c | sort -nr | head -20
I have a project with a large package named "debugger". The presence of "bug" within "debugger" causes the original command to go crazy.
Good catch, that's better
I wouldn't trust "commit counts." The quality and content of a "commit" can vary widely between developers. I have one guy on my team who commits only working code that has been thoroughly tested locally, another guy who commits one line changes that often don't work, only to be followed by fixes, and more fixes. His "commits" have about 1/100th of the value of the first guy.
The author does not look at counter values but rather at how the values changes. That reveals dynamics.
My comment still seems relevant? Do frequent commits to correct mistakes imply more "value" than infrequent, but well tested, commits, or what? I don't think it is a reliable signal.
Interesting ideas, but some to me seem very overgeneralizef, e.g.:
> How Often Is the Team Firefighting
> git log --oneline --since="1 year ago" | grep -iE 'revert|hotfix|emergency|rollback
> Crisis patterns are easy to read. Either they’re there or they’re not.
I disagree with the last two quoted sentences, and also, they sound like an LLM.
> One caveat: squash-merge workflows compress authorship. If the team squashes every PR into a single commit, this output reflects who merged, not who wrote. Worth asking about the merge strategy before drawing conclusions.
In my experience, when the team doesn't squash, this will reflect the messiest members of the team.
The top committer on the repository I maintain has 8x more commits than the second one. They were fired before I joined and nobody even remembers what they did. Git itself says: not much, just changing the same few files over and over.
Of course if nobody is making a mess in their own commits, this is not an issue. But if they are, squash can be quite more truthful.
Solid list. I'd add git log --all --oneline --graph pretty early on — gives you a quick sense of how active different branches are and whether this is a "one person commits everything" project or actually distributed. Helped me a ton on a job where I inheritied a monolith with like 4 years of history.
The git blame tip is underrated. People treat it like a gotcha tool but its maybe the fastest way to find the PR/ticket that explains a weird decision.
> Is This Project Accelerating or Dying > > git log --format='%ad' --date=format:'%Y-%m' | sort | uniq -c
If the commit frequency goes down, does it really mean that the project is dying? Maybe it is just becoming stable?
Yeah, this one demonstrates a particularly pernicious view of software development. One where growth, no matter how artificial, is the only sign of success.
If you work with service oriented software, the projects that are "dying" may very well be the most successful if it's a key component. Even from a business perspective having to write less code can also be a sign of success.
I don't know why this was overlooked when the churn metric is right there.
Bad memories at my former big tech company.
Whenever we initiated a new (internal) SW project, it had to go through an audit. One of the items in the checklist for any dependency was "Must have releases in the last 2 years"
I think the rationale was the risk of security vulnerabilities not being addressed, but still ...
For this command in particular, one can add a cheap bar chart with awk:
git log --format='%ad' --date=format:'%Y-%m' | sort | uniq -c | awk '{printf $2" "; for (i=1;i<=$1;i++){printf "-";} print ""; }'
That was my question too. I have plenty of projects I've worked on where they rarely get touched anymore. They don't need new features and nothing is broken.
Is it fair to say they are being "worked on" if nothing is being done?
Sometimes you need to bump a dependency version, adjust the code to a changed API endpoint, or update a schema. Even if the core features stay the same, there's some expected maintenance. I'd still call that being worked on, in a sense that someone has to do it.
Technically you're correct that change frequency doesn't necessarily mean dead, but the number of projects that are receiving very few updates because they're 'done' is a fraction of a fraction of a percent compared to the number that are just plain dead. I'm certain you can use change frequency as a proxy and never be wrong.
> I'm certain you can use change frequency as a proxy and never be wrong.
I (largely) wrote a corporate application 8 years ago, with 2 others. There was one change 2 years ago from another dev.
Lots of programs are functionally done in a relatively short amount of time.
"Accelerating or Dying" sounds like private equity's lazy way to describe opportunity, not as a metric to describe software.
That sort of project exists in an ocean of abandoned and dead projects though. For every app that's finished and getting one update every few years there are thousands of projects that are utterly broken and undeployable, or abandoned on Github in an unfinished state, or sitting on someone's HDD never be to touched again. Assuming a low change frequency is a proxy for 'dead' is almost always correct, to the extent that it's a reasonable proxy for dead.
I know people win the lottery every week, but I also believe that buying a lottery ticket is essentially the same as losing. It's the same principle.
With respect, this is a myopic view. Not all software is an "app" or a monolith. If you use a terminal, you are directly using many utilities that by this metric are considered dying or dead.
Projects become more stable with time? Since when?
Something something Red Queen's race
Or you hired someone who squashes or doesn't commit every single change.
Rather than using an LLM to write fluffy paragraphs explaining what each command does and what it tells them, the author should have shown their output (truncated if necessary)
I also feel this reads like an AI slop, but at least I learned 5 commands. Not too bad.
To me all of these are symptoms of the problem that I outlined in my recent blog post: https://news.ycombinator.com/item?id=47606192
and it touches in detail what exactly commit standards should be, and even how to automate this on CI level.
And then I also have idea/vision how to connect commits to actual product/technical/infra specs, and how to make it all granular and maintainable, and also IDE support.
I would love to see any feedback on my efforts. If you decide to go through my entire 3 posts I wrote, thank you
These are actually fun to run. Just checked from work who makes most commits and found I have as many commits in past 2 years as 3 next people.
That probably isn’t a good sign
Nice set of commands! I would suggest using --all flag with git log though - scans through all branches and not just the current one
Dying or stabilizing?
Most good projects end up solving a problem permanently and if there is no salary to protect with bogus new features it is then to be considered final?
Can't resist making it as a git command https://github.com/zdk/git-critique
Instead of focusing on the top 20 files, you can map the entire codebase with data taken from git log using ArcheoloGit [1].
[1]: https://github.com/marmelab/ArcheoloGit
These were interesting but I don't know if they'd work on most or any of the places I've worked. Most places and teams I've worked at have 2-3 small repos per project. Are most places working with monorepos these days?
I can't speak for most, but the past few places I consulted or worked at used monorepos.
Jesus I've seen what you've done for others and want that for myself.
For "what changes the most", in my project it's package.json / lock (because of automatic dependency updates) and translation / localization files; I'd argue that's pretty normal and healthy.
For the "bus factor", there's one guy and then there's me, but I stopped being a primary contributor to this project nearly two years ago, lol.
Biggest life changer for me has been:
git clone --depth 1 --branch $SomeReleaseTag $SomeRepoURL
If you only want to build something, it only downloads what you need to build it. I've probably saved a few terabytes at this point!
> If the team squashes every PR into a single commit, this output reflects who merged, not who wrote.
Squash-merge workflows are stupid (you lose information without gaining anything in return as it was easily filterable at retrieval anyway) and only useful as a workaround for people not knowing how to use git, but git stores the author and committer names separately, so it doesn't matter who merged, but rather whether the squashed patchset consisted of commits with multiple authors (and even then you could store it with Co-authored-by trailers, but that's harder to use in such oneliners).
Can you explain to me (an avid squash-merger) what extra information do you gain by having commits that say "argh, let's see if this works", "crap, the CI is failing again, small fix to see if it works", "pushing before leaving for vacation" in the main git history?
With a squash merge one PR is one commit, simple, clean and easy to roll back or cherry-pick to another branch.
These commits reaching the reviewer are a sign of either not knowing how to use git or not respecting their time. You clean things up and split into logical chunks when you get ready to push into a shared place.
Why would the reviewer look at the commit messages instead of the code?
1. Open PR page in whatever tool you're using
2. Read title + description to see what's up
3. Swap to diff and start reading through the changes
4. Comment and/or approve
I've never heard anyone bothering to read the previous commit messages for a second, why would they care?
In some cases, reviewing PR diffs commit-by-commit (and with the logs as the narration of the diff-by-diff story) is a substantial improvement over reviewing the entire PR diff. Concrete examples...
* A method or function that has code you realize needs to be shared...the code may need to be moved and also modified to accommodate its shared purpose. Separating the migration from any substantive modifications allows you to review the migration commit with the assistance of git's diff.colorMoved feature. It becomes easier to understand what changes are due to the migration, and what changes were added for more effective sharing.
* PRs sometimes contain mechanical work that is easy to review in isolation. Added or removed arguments, function renames, etc. No big deal if it's two or three instances, but if it's dozens or hundreds of instances, it's easier for the humans to review all of those consistent changes together, rather than having them mixed in with other things one has to reason about.
* Sometimes a flow of commits can help follow a difficult chain of reasoning. PR developer claims that condition X can never occur, but the code is complex enough that it's difficult to verify. However, by transforming the code in targeted ways that are possible to reason about, the complexity might be reducible to the point where the claim becomes obvious. One frequent example I see of this is of function/method arguments that are actually unnecessary, but it wasn't obvious until after some code transformations.
Because it's a useful abstraction. If you only look at PRs and don't ever care about commits, why are they even being sent to reviewer in the first place? Just send a diff file.
Having atomic commits lets you actually benefit from having them. Suddenly you don't have to perform weird dances with interconnected PRs with dependencies as "PR too big" is not such a problem anymore as long as commits are digestible; you can have things property bisectable; you can preserve shared authorship; you can range-diff and have a better view on what and how changed between review passes, and so on...
The unit of change is commit, and PRs group commits you want someone to pull. If you don't want or need any of that, you're just sending a patch file in a needlessly elaborate way.
> If you only look at PRs and don't ever care about commits, why are they even being sent to reviewer in the first place? Just send a diff file.
This is in fact what hg does with amending changesets and yes it works far better. Keep PRs small and atomic and you never need to worry about what happens intra-pr. If you need bigger units of work that's what stacking is for.
Stacking is good for expressing dependencies, but isn't helpful when you need to make several distinct changes that aren't necessarily needed unless you take them all in. What's the value in having a separate PR that introduces a framework that you later use in another PR when you may not actually want to merge it if the latter one doesn't end up being merged as well?
A PR is a group of commits, just utilize that when you need it.
>Swap to diff and start reading through the changes
this forces the reviewer to view the entire diff at once, which can greatly increase the cognitive load vs. being able to view diffs of logical units of work
for tiny PRs it may not matter, but for substantial PRs it can matter a lot
What if the shared place is the place where you run a bunch of CI? Then you push your work early to a branch to see the results, fix them etc.
You can do whatever you want with stuff nobody else looks at. I do too.
I meant "shared place" as an open review request or a shared branch rather than shared underlying infrastructure. Shared by people's minds.
You can always force-push a cleaned up version of your branch when you are ready for review, or start a new one and delete the WIP one.
You can, but instead you can also just squash merge in one click. And avoid that people merge there dozens of fixes if you allow anything but squash merge.
I hate (and fear) force-pushing and "cleaning up" git history as much as other people dislike squash-merging =)
It just feels wrong to force push, destroying stuff that used to be there.
And I don't have the time or energy to bisect through my shitty PR commits and combine them into something clean looking - I can just squash instead.
Nothing is destroyed by a force push. It just overwrites a single pointer, and even keeps its old value in reflog.
Things that aren't referenced by anything anymore will eventually get garbage collected and actually destroyed, but you can just keep a reference somewhere to prevent that from happening if you need. Or even disable garbage collection completely.
Looks like people's fears about git come just from not knowing what it does.
You can't use the remote reflog to revert what you force pushed, can you? But I agree that having your local reflog means you're never totally lost. I still just make a branch before major edits so I can go back.
What are examples of better ones. I don’t get the let me show the world my work and I’m not a fan of large PR
if you mean better messages, it's not really that. those junk messages should be rewritten and if the commits don't stand alone, merged together with rebase. it's the "logical chunks" the parent mentioned.
it's hard to say fully, but unless a changeset is quite small or otherwise is basically 0% or 100%, there are usually smaller steps.
like kind of contrived but say you have one function that uses a helper. if there's a bug in the function, and it turns out to fix that it makes a lot more sense to change the return type of the helper, you would make commit 1 to change the return type, then commit 2 fix the bug. would these be separate PRs? probably not to me but I guess it depends on your project workflow. keeping them in separate commits even if they're small lets you bisect more easily later on in case there was some unforseen or untested problem that was introduced, leading you to smaller chunks of code to check for the cause.
If the code base is idempotent, I don't think showing commit history is helpful. It also makes rebases more complex than needed down the line. Thus I'd rather squash on merge.
I've never considered how an engineer approaches a problem. As long as I can understand the fundamental change and it passes preflights/CI I don't care if it was scryed from a crystal ball.
This does mean it is on the onus of the engineer to explain their change in natural language. In their own words of course.
Commits don't show "how an engineer approaches a problem". Commits are the unit of change that are supposed to go into the final repository, purposefully prepared by the engineer and presented for review. The only thing you do by squashing on merge is to artificially limit the review unit to a single commit to optimize the workflow towards people who don't know how to use git. Personally I don't think it's a good thing to optimize for.
Preserving commit history pre-merge only seems useful if I had to revert or rebase onto an interstitial commit. This is at odds with treating PRs as atomic changes to the code base.
I might have not stated my position correctly. When I mean "squash on merge", I mean the commit history is fully present in the PR for full scrutiny. Sometimes commits may introduce multiple changes and I can view commit ranges for each set of changes. But it takes the summation of the commits to illustrate the change the engineer is proposing. The summation is an atomic change, thus scrutinizing terms post-merge is meaningless. Squashing preserves the summation but rids of the terms.
Versioned releases on main are tagged by these summations, not their component parts.
Haha, good luck working with a team with more than 2 people. A good reviewer looks at the end-state and does not care about individual commits. If im curious about a specific change i just look at the blame.
> A good reviewer looks at the end-state and does not care about individual commits.
Then I must be a bad reviewer. In a past job, I had a colleague who meticulously crafted his commits - his PRs were a joy to review because I could go commit by commit in logical chunks, rather than wading through a single 3k line diff. I tried to do the same for him and hope I succeeded.
And then someone comments on a thing, they change it and force-push another "clean" history on top and all of your work is wasted because the PR is now completely different =)
Why are those not just separate PRs? Or if they really needed to be merged at once - they should still be separate PRs but on a feature branch
Why have PRs - groups of commits to pull - then if all you need is a single patch file?
You can, but most of us work in Github and having a PR to dump commits to is very easy and convenient. Then, when you get some feedback on your PR, you can dump more commits and it's very easy for the reviewer to see what has changed since the last time they reviewed it.
I feel like what you're arguing is that you should clean up your commits before anyone else sees them. Fair. But you could also clean it up right before merging to main. It's not that different, except the latter is much less annoying, particularly when going back and forth with people.
I know this is a very github centric workflow, but that's where most engineers work now, and it's nice and easy. This wouldn't work for eg: contributing to linux, but that's not what most of us do.
Split the PR rather than force me to wade through your commit history. Use graphite or something else that allows you to stack PRs.
Why is it "wade through" if there are 10 clearly distinct but dependent commits, but comfortable if it's 10 stacked PRs instead? They are basically the same thing, presented ever so slightly differently.
I think in most teams I've worked with, the majority of developers (> 85%) barely undestand what Git is doing or what things mean inside GitHub, have never seen commit history as a graph, have never run something like "git log --oneline --graph --decorate" or "--format", and have never heard of "git range-diff" which is very useful for following commit/PR/unit changes.
Personally I review using "git" itself, so I see the graph structure either way, and there's little difference between stacked PRs, commit chains in a single PR, or even feature branches, from that point of view. Even force-push branch updatea aren't difficult to review, because of the reflog and "git range-diff". The differences are mainly in what kinds of behaviour the web-based tooling promotes in the rest of the team, which does matter, and depends on the team.
I agree with you if you're using Graphite instead of GitHub. Having a place to give feedback and/or approval on the individual "units" (commits in a PR, or PRs in a stack) is useful, grouping dependent but distinct changes is useful, and diff'd commit evolution within each unit PR in response to back-end-forth review feedback is useful in some collaborative settings. Though, if you know "git range-diff" and reflog, that shows diff'd commit evolution quite well.
In GitHub, people are confused by stacked PRs both conceptually and due to the GitHub UX around them. Most times when I've posted a stacked PR to a GitHub project, other people didn't realise it was stacked, and occasinally someone else has merged the tip of a stack made by me, and been surprised to see all the dependent PRs merged automatically as a side effect. Usually before they get to reviewing those other PRs :-)
People understand commit sequences in a PR, though I've rarely seen people treat the individual commits as units for review when using GitHub, unfortunately. In the Linux kernel world where Git was born, the PR flow is completely different from GitHub: Their system tends to result in feedback on individual commits. It also encourages better quality feedback, with less nitpicking, and better quality commits.
Sometimes I have to go back and fix a bug that appeared during another branch. Having the original commits helps me bisect it.
Not often, but given that it costs me nothing to have it all in my tree, I'd rather have it than not.
You review code not to verify the actual output of the code, but the code itself. For bugs, for maintainability. Commit hygiene is part of that.
I have no troubles working on big FLOSS projects where reviews usually happen at the commit level :)
So if a PR consists of 20 commits, they review every single commit linearly without looking at the end result first?
Yes, and in some projects 20 commits is not even a big PR, more like "regular sized". The LKML's first page is now full of PRs with around 20 commits, here's a random one as an example: https://lore.kernel.org/netdev/20260408121252.2249051-1-dhow...
And here's a slightly smaller one which isn't about "miscellaneous fixes": https://lore.kernel.org/netdev/20260408122027.80303-1-xuanzh...
Some of these commits even get reviewed by different maintainers before being merged, which is common when a patchset touches several subsystems at once.
Trivial and not too silly example:
Part of new feature you had working in an intermediate commit, but broke somewhere along the way and is not working in your last commit when you squashed.
If you catch it early enough, I suppose it's in your reflog, but otherwise you're screwed.
It sounds like a silly example, but I bet most developers have run into this at some point.
With mercurial/jujutsu, you get the best of both worlds: The "argh, let's see if this works" commits are what I call "microcommits", and the squashed versions are the real/public commits. With jujutsu, you get both. Your log shows only the "real" commits (equivalent of squashing all the commits between that and the prior "real" commit). But if you want to drill down into the microcommits, the information is always there.
Let's acknowledge the reality. Many people use git not just for version control, but for backup ("Let me commit this so I don't lose it"). Let's ensure the VC tool supports both and doesn't force you to pick one over the other.
If someone uses git commits like the save function of their editor and doesn't write messages intended for reading by anyone else, it makes sense to want to hide them
For other cases, you lose the information about why things are this way. It's too verbose to //comment on every like with how it came to be this way but on (non-rare in total, but rare per line) occasion it's useful to see what the change was that made the line be like this, or even just who to potentially ask for help (when >1 person worked on a feature branch, which I'd say is common)
> If someone uses git commits like the save function of their editor
I use it like that too and yet the reviewers don't get to see these commits. Git has very powerful tools for manipulating the commit graph that many people just don't bother to learn. Imagine if I sent a patchset to the Linux Kernel Mailing List containing such "fix typo", "please work now", "wtf" patches - my shamelessness has its limits!
Seems like a lot of extra effort (save, add, commit, come up with some message even if it's a prayer to work now) only to undo it again later and create a patch or alternate history out of the final version. Why bother with the intermediate commits if you're not planning for it to be part of the history?
Git is a version control system. It does not care about what it versions.
When I work on something, I commit often and use the commit graph as a undo tool on steroids. I can see what I tried, I can cherry-pick or revert stuff while experimenting, I can leave promising but unfinished stuff to look at later, or I can just commit to have a simple way to send stuff to CI, or a remote backup synced between machines.
Once I'm done working on something, it's time to take a step back, look at the big picture, see how many changes my experiments have actually yielded, separate them, describe and decide whether they go to review together or split in some way, as sometimes working on a single thing requires multiple distinct changes (one PR with multiple commits), but sometimes working in a single session yields fixes for multiple unrelated issues (several PRs). Only then it gets presented to the reviewer.
It just happens that I can do both these distinct jobs with a single tool.
Because I might want to go back to this current messy state but I don't want to commit it like this (hardcoded test strings, debug logs, cutted corners to see if something works, you name it).
I simply commit something like "WIP: testing xy" and if its working and properly implemented i can squash/rebase/edit the commit message and force push it to my feature branch. Using a Git client like Gitkraken makes this incredibly easy, takes seconds.
This way I can leverage version control without committing bogus states to the final PR.
If the team is using a PR workflow, the PR is a working place to produce one single commit. The individual commits are just timestamped changes and comments. Think of it as the equivalent of annotated diff in mailing list conversation.
You gain the extra information by having reasonable commit messages rather than the ones you mentioned. To fix CI you force push.
Can you explain to me what an avid squash-merger puts into the commit message of the squashed commit composed of commits "argh, let's see if this works", "crap, the CI is failing again, small fix to see if it works", and "pushing before leaving for vacation" ?
The squashed commit from the PR -> main will have a clean title + description that says what was added.
Usually pretty close to what the PR title + description are actually, just without the videos and screenshots.
Example:
feat(ui): Add support for tagging users
* Users can be tagged via the user page * User tags visible in search results (configurable)
etc..
I don't need to spend extra time cleaning up my git commits and force-pushing on the PR branch, losing context for code reviews etc. Nor does anyone have to see my shitty angry commits when I tried to figure out why Playwright tests ran on my machine and failed in the CI for 10 commits.
> "argh, let's see if this works", "crap, the CI is failing again, small fix to see if it works", "pushing before leaving for vacation"
These are all bad commits IMHO. Aside from the CI one, I understand that message. I have commits like that on personal projects but for professional projects I'd be frustrated if people were committing messages like that.
Personally I'm a "one commit" type of guy, I don't like committing things in a broken state even on a side branch unless I have to (to share the code or test a CI). Occasionally I will make multiple commits at the very end to make review easier or once I have everything working but I want to try something different but I have a bunch of options of saving code that don't involving committing:
- Stash
- Shevle (IDEA)
- Backblaze
- Time Machine
- Local History (IDEA)
The idea of committing WIP before leaving for a vacation just feels so wrong to me.
I once worked for someone who wanted developers to commit code before the end of every day as a safety measure. His reasoning was in case the developer's computer died or similar. I found that silly at the time and still do now. That's what backups are for, I dislike when people use git as a backup like that in a professional setting.
Why are those commits ending in the PR? Just unprofessional to work like that.
git bisect gets more useful because it will pin a smaller set of changes
Squash merge is the only reasonable way to use GitHub:
If you update a PR with review feedback, you shouldn’t change existing commits because GitHub’s tools for showing you what has changed since your last review assume you are pushing new commits.
But then you don’t want those multiple commits addressing PR feedback to merge as they’re noise.
So sure, there’s workflows with Git that doesn’t need squashing. But they’re incompatible with GitHub, which is at least where I keep my code today.
Is it perfect? No. But neither is git, and I live in the world I am given.
If your goal here is to have linear history, then just use a merge commit when merging the PR to main and always use `git log --first-parent`. That will only show commits directly on main, and gives you a clean, linear history.
If you want to dig down into the subcommits from a merge, then you still can. This is useful if you are going back and bisecting to find a bug, as those individual commits may hold value.
You can also cherry pick or rollback the single merge commit, as it holds everything under it as a single unit.
This avoids changing history, and importantly, allows stacked PRs to exist cleanly.
Git bisect is one of the important reasons IMO to always squash-merge pull requests: Because the unit of review is the pull request.
I think this is all Github's fault, in the end, but I think we need to get Github to change and until then will keep using squash-merges.
Yes, I think people who are anti squash merge are those who don't work in Github and use a patch based system or something different. If you're sending a patch for linux, yes it makes sense that you want to send one complete, well described patch. But Github's tooling is based around the squash merge. It works well and I don't know anyone in real life who has issues with it.
And to counter some specific points:
* In a github PR, you write the main commit msg and description once per PR, then you tack on as many commits as you want, and everyone knows they're all just pieces of work towards the main goal of the eventually squashed commit
* Forcing a clean up every time you make a new commit is not only annoying extra work, but it also overwrites history that might be important for the review of that PR (but not important for what ends up in main branch).
* When follow up is requested, you can just tack on new commits, and reviewers can easily see what new code was added since their last review. If you had to force overwrite your whole commit chain for the PR, this becomes very annoying and not useful to reviewers.
* In the end, squash merge means you clean up things once, instead of potentially many times
Forcing a single commit per PR is the issue imo. It's a lazy solution. Rebase locally into sensible commits that work independently and push with lease. Reviewers can reset to remote if needed.
The author is talking about the case where you have coherent commits, probably from multiple PRs/merges, that get merged into a main branch as a single commit.
Yeah, I can imagine it being annoying that sqashing in that case wipes the author attribution, when not everybody is doing PRs against the main branch.
However, calling all squash-merge workflows "stupid" without any nuance.. well that's "stupid" :)
I don't think there's much nuance in the "I don't know --first-parent exists" workflow. Yes, you may sometimes squash-merge a contribution coming from someone who can't use git well when you realize that it will just be simpler for everyone to do that than to demand them to clean their stuff up, but that's pretty much the only time you actually have a good reason to do that.
I really, really wish git changed two defaults:
If you use a workflow that always merges a PR with a merge commit, then git log --first-parent gives you a very nice linear history. I feel like if this was the default, so many arguments about squashing or rebasing workflows wouldn't be necessary to get our "linear history", everyone would just be doing merges and be happy with it. You get a clean top level history and you can dig down into the individual commits in a merge if you are bisecting for a bug.Do people actually share PR as in different people contributing to the same branch?
Also I can understand not squashing if the contribution comes from outside the organization. But in that case, I would expect a cleaned up history. But if every contribution is from members of the team, who can merge their own PR, squash merge is an easy way to get a clean history. Especially when most PR should be a single commit.
We do. If we are building out a feature, none of its code is merged into main until it is complete (if this is a big feature, we milestone into mergeable and releasable units).
The feature is represented by a Story in Jira and a feature branch for that story. Subtasks in jira are created and multiple developers can pick up the different subtasks. There is a personal branch per subtasks, and PRs are put up against the feature branch. Those subtasks are code reviewed, tested, and merged into the feature branch.
In the end, it is the feature branch that is merged (as a single merge commit and complete unit) into main, and may well have had contributions from multiple people.
I get your POV, but I’ve always considered that long-lived branches in the canonical repo (the one in the forge) other than the main one should be directly related to deployable artifacts. Anything else should be short-lived.
There can be experiment on the side that warrants your approach, but the amounts of merge going back and forth would make this hard to investigate (especially when blaming) I would prefer to have one single commit with a message that describe every contribution.
I think the point is that if you have to squash, the PR-maker was already gitting wrong. They should have "squashed" on their end to one or more smaller, logically coherent commits, and then submitted that result.
It’s not “having to squash”. The intent was already for a PR to be a single commit. I could squash it on my end and merge by rebasing, but any alteration would then need to be force-pushed. So I don’t bother. I squash-merge when it’s ready and delete the branch.
Squash-merge is entirely fine for small PRs. Cleaning up the commits in advance (probably to just squash them to one or two anyway) is extra work, and anything that discourages people from pushing often (to get the code off their local machine) needs to be well-justified. Just review the (smallish!) total outcome of all the commits and squash after review. A few well-placed messages on the commit, attached to relevant lines, are more helpful and less work than cleaning up the commit history of a smallish PR.
For really large PRs, I’m more inclined to agree with you, but those should probably have their own small-PR-and-squash-merge flow that naturally cleans up their git history, anyway.
I categorically disagree that squash-merge is “stupid” but agree there are many ways to skin this cat.
How does not squash merging deal with the fact that branches disappear when merging? What I mean is that the information "this commit happened in the context of this PR or this overarching goal" goes missing. When you squash, you use the one central unit of information management in Git: the commit.
Having the tree easy to filter doesn't matter if it returns hundreds of commits you have to sift through for no reason.
Having the commit graph easy to filter means exactly that you don't have to sift through hundreds of commits for no reason. What else did you think it would mean?
This is good stuff. Why I never think of things like this is beyond me. Thanks
The last sentence of the article is "Here’s what the rest of the week looks like." and then it just stops. Am I missing something?
It’s phrased in a confusing way, seems to be a mistake. It becomes clearer when you click the link in the sentence before.
It is probably referring to the article linked in the "codebase audit" text
These are some helpful heuristics, thanks.
This list is also one of many arguments for maintaining good Git discipline.
I just finished¹ building an experimental tool that tries to figure out if a repo is slopware or not just by looking at it's git history (plus some GitHub activity data).
The takeaway from my experiment is that you can really tell a lot by how / when / what people commit, but conclusions are very hard to generalize.
For example, I've also stumbled upon the "merge vs squash" issue, where squashes compress and mostly hide big chunks of history, so drawing conclusions from a squashed commit is basically just wild guessing.
(The author of course has also flagged this. But I just wanted to add my voice: yeah, careful to generalize.)
¹ Nothing is ever finished.
Out of curiosity, I ran the 5 command on my project's public git tree. The only informative one was #4 ("Is This Project Accelerating or Dying") - it showed cliffs when significant pieces of logic were decoupled and moved to other repos.
Ages ago, google released an algorithm to identify hotspots in code by using commit messages. https://github.com/niedbalski/python-bugspots
Ages ago google wrote an algorithm to detect hotspots by using commit messages, https://github.com/niedbalski/python-bugspots
Trusting the messages to contain specific keywords seems optimistic. I don't think I used "emergency" or "hotfix" ever. "Revert" is some times automatically created by some tools (E.g. un-merging a PR).
For the stuff I've worked on, if you want to know about bugfixes and emergency releases, you'd go to Jira where those values are formalized as fields. Someone else in the comments here had a suggestion which just looks for the word "fix" which would definitely capture some bugfix releases, but is more likely to catch fixes that were done during development of a feature.
These commands are very useful, but adapting them to the codebase makes a huge difference.
For most, I added some filters and slightly changed the regex, and it showed the reality of the codebase (I already knew the reality, I just wanted to see if it matched, and it did).
My team usually uses "Squash and merge" when we finish PRs, so I feel that would skew the results significantly as it hides 99% of the commit messages inside the long description of the single squashed merge commit.
For more insights on Git, check out https://github.com/nolasoft/okgit
I created a small TUI based on the article https://github.com/mikaoelitiana/git-audit
You beat me to it! I envisioned creating some aliases but you exceeded it by building a TUI. Good job Claude! LOL ;)
> The 20 most-changed files in the last year. The file at the top is almost always the one people warn me about.
What a weird check and assumption.
I mean, surely most of the "20 most-changed files" will be README and docs, plus language-specific lock-files etc. ?
So if you're not accounting for those in your git/jj syntax you're going to end up with an awful lot of false-positive noise.
Why would you touch the README file hundreds of times a year?
You're right about package.json, pnpm-lock etc though, but those are easy to filter out if the project in question uses them.
> Why would you touch the README file hundreds of times a year?
You're right, perhaps I should have said CHANGELOG etc.
Although some projects e.g. bump version numbers in README or add extra one-liner examples ....
Some readme files include changelogs. But aside from that I think this can still net some useful information. I like to look at the most recently changed files in a repo as well.
Fair point. I skip lockfiles, changelogs, and generated code. The first application file on the list is the one that matters. Should have been explicit about that in the post.
It’s easy enough to filter those out with grep. It still is relatively meaningless. If the team incrementally adds things then it’s just going to show what additions were made. It isn’t churn at all.
Before, I ask AI "is this project maintained" done.
I put it into a gist :)
https://gist.github.com/aeimer/8edc0b25f3197c0986d3f2618f036...
Nice timing. I was just today needing some of the info that these commands surface. Serendipitous!
This is a great list of commands to quickly understand a repository. Thank you for sharing.
What's the subversion equivalents to these commands?
superficial. If I have to unfuck the backend 10 times a week in our API adapter, then these commands will show me constantly changing the API adapter, although it's the backend team constantly fixing their own bugs
So you value more rushed descriptions of changes than actual changes. Nice
Great tips, added to notes.txt for future use ..
Another one I do, is:
This way I can see right away which branches are 'ahead' of the pack, what 'the pack' looks like, and what is up and coming for future reference ... in fact I use the 'gss' alias to find out whats going on, regularly, i.e. "git fetch --all && gss" - doing this regularly, and even historically logging it to a file on login, helps see activity in the repo without too much digging. I just watch the hashes.I'm so used to magit, it seems kind of primitive to pipe git output around like this.
Anyway, I can glean a lot of this information in a few minutes scrolling through and filtering the log in magit, and it doesn't require memorizing a bunch of command line arguments.
Nice! Will probably adopt this, seems to give a great overview!
No searching the codebase/commits for "fuck" and shit"? That will give you an idea what what was put in under stressful circumstances like a late night during a crunch.
thank you - these are useful
Step 6: grep the thread count on the squash-merge debate to determine if the team has unresolved interpersonal conflict.
Just looking at how often a file changes without knowing how big the file is seems a bit silly. Surely it should be changes/line or something?
Sure, normalizing by size would be more precise. But this is a quick gut check to know which files to look at first, not a metric.
blog posts are just comments that would have been torn apart if only posted on a forum, now masquerading as important universal edicts
This should be renamed to "Git commands that I run as a new hire to get metrics I'll forget on day 2".
> The 20 most-changed files in the last year. The file at the top is almost always the one people warn me about. “Oh yeah, that file. Everyone’s afraid to touch it.”
I've got my Emacs set up to display next to every file that is versioned the number of commits that file has been modified in (for the curious: using a modified all-the-icons-ivy-rich + custom elisp code + custom Bash scripts I wrote and it's trickier than it seems to do in a way that doesn't slows everything down). For example in the menu to open a file or open a recently visited file etc.: basically in every file list, in addition to its size, owner, permissions, etc. I also add the number of commits if it's a versioned file.
I like the fix/bug/broken search in TFA to see where the bugs gather.
I was curious what information I could glean from these for some popular repos. Caveat: I'm primarily an low-level embedded developer so I don't interface with large open source projects at the source level very often (other than occasionally the linux kernel). I chose some projects at random that I use.
*Mainline linux*
Most changed files: pretty much what I expected for 1 and 2... the "cutting edge" of Linux development over other OSes -- bpf and containers. The bpf verifier and AMD GPU driver might get a boost in this list due to sheer LoCs in those files (26K and 14K respectively). An intel equivalent of amdgpu_dm is #21 in the list (drivers/gpu/drm/i915/display/intel_display.c) and nvidia is nowhere to be seen (presumably due to out-of-tree modules/blobs?).
Bus factor: obviously none. The top 4 Buggy files: Intel comes out on top of GPU drivers this time (twice). Along with KVM for x86(64), the main allocator, and BTRFS. *GCC*Most changed files: IR autovectorization code, riscv heuristics tables, and C++ template handling (pt.c is "paramaterized types").
Buggy files: DWARF debuginfo generation, x86 heuristics tables, RS6000(?!) heuristic tables. I had to look up RS6000, it's an IBM instruction set from the 90s lol. cp-tree.h is an interesting file, it seems be the main C(++) AST datastructures. *xfwm4* Most changed files: the list is dominated by *.po localizations. I filtered these out. Even after this, I discovered there is very little active development in the last few years. If I extend to 4 years ago, I get: 1. src/client.c - Realizing this project is too "small" to glean much from this. client.c is just the core X client management code. Makes sense. 2. src/placement.c - Other core window management code.This has not told me much other than where most of the functionality of this project lies.
Bus factor: Pretty huge. Not really an issue in this case due to lack of development I guess.
Files with bug commits: Very similar distribution to most changed files. Not enough datapoints in this one to draw any big conclusions.I think these massive open projects (excl xfwm) are generally pretty consistent code quality across the heavily trodden areas because of the amount of manpower available to refactor the pain points. I've yet to see an example of "god help you if you have to change that file" in e.g. linux, but I have of course seen that situation many times in large proprietary codebases.
This is better than the post itself. Showing real output from a real repo.
Big projects tend to self-correct. These commands hit differently on private codebases with 3-10 contributors, where high-churn usually means one person patching the same thing repeatedly.
More AI slop.
Wtf is happening to this website
git commands I run before reading any code:
git rm -rf .