Utter Disregard for Git Commit History | Hacker News Reader

Utter Disregard for Git Commit History

zachholman.com

14 pointskiyanwang10 comments

Personally I like to push as soon as it's not broken to have something to revert too. This also means e.g. function and design might be in different commits. Bundeling more together makes the commits easier to read, but leaves me without version history in between steps. Is there a trick to have both?

I shudder to think what kind of names he probably uses for local variables.

If your commit messages are meaningless, why not make them empty? Like this: `git commit -m "" --allow-empty-message`.

It's like you didn't actually read the same article that I did. Nowhere does the author say that commit messages are meaningless. "foo", followed by "revert foo" carries a LOT of meaning. It means you tried something and it didn't work. If you leave that out of your history you risk not remembering that you tried something and that it didn't work. You don't have to look at the revert commit. For all other purposes, the two commits may as well not exist, but together they carry extremely useful meta information.

Disregarding the low level commit messages and using commit clusters for features allows you to easily track when and how this happens. Using single atomic feature commits does not.

How is "Revert foo" different from "Revert 52e8ace0571ec9ac9b669bc5acfd0bcea537805d"?

> Nowhere does the author say that commit messages are meaningless

He doesn't. I say it. His commit messages are meaningless. To understand what's going on, you need to read his diffs, not commit messages. In his workflow, only reverts and merges provide any context, and for them git creates meaningful commit messages automatically.

It's different from rebasing to squash many commits into one. The article has nothing to do with hashes. Foo is also easier to parse mentally than umpty digits of hexadecimal vomit.

> It's different from rebasing to squash many commits into one.

What is different?

My point is, if you don't write meaningful commit messages for some commits, you probably don't need any commit messages for those commits at all. It reduces noise for people who read your `git log`. If you make a commit and then immediately revert it, instead of "foo", it's better to write at least "tried ...; didn't work".

The posted article is about whether to squash commits or not. The posted article is comparing the view of the commit as ultimate arbiter of feature history or the commit cluster (aka Pull Request) as ultimate arbiter of feature history. Putting details below the fold on a single feature commit makes them 99% invisible.

The best solution, which the posted article suggests, is for git to natively support commit clusters. So that you can have just feature information and also meta information that isn't 99% invisible.

TL,DR: Pedantic People ranting about missing commit history. Slackers declaring it a art and "look into the diff". Not present- pragmatics training NN on the diffs, so that changelog messages get autogenerated.

This might be the least apt TLDR I've ever read. How did you come up with it?

Reply on news.ycombinator.com