Today I built a script that installs local Git hooks to help maintain Conventional Commits and branch naming conventions. It works globally or per project.
If you’re interested, hit me up :)
#Git #DevOps #Automation #CICD #OpenSource #ConventionalCommits #GitLab
Good commit messages keep the history clean, make it easier to work together and are even machine-readable <3
https://www.conventionalcommits.org/en/v1.0.0/
With this standard, commits can be automatically turned into nice, clear changelogs.
Tác giả đề xuất hệ thống đặt tên commit Git mới, chi tiết hơn Conventional Commits. Nó phân loại rõ ràng thay đổi "Hướng người dùng" (ví dụ: feat, fix) và "Nội bộ" (ví dụ: refactor, chore). Mục tiêu là giúp commit message rõ ràng và dễ hiểu hơn. Bạn nghĩ sao?
#Git #LậpTrình #Commit #ConventionalCommits #PhátTriểnPhầnMềm
#Git #Programming #CommitNaming #SoftwareDevelopment #DevOps
https://www.reddit.com/r/programming/comments/1pui0hc/commit_naming_system/
Just published a guide on using changelogen + a custom AI slash command to auto-generate and clean up your `CHANGELOG.md` from Conventional Commits. Duplicate issues gone, contributor names fixed, semantic bumps handled.
So... I use #gitmoji for my #ConventionalCommits, and I absolutely love it. I've been using semantic-release-gitmoji [1] in order to build a `CHANGELOG.md` file out of the commit history, but it doesn't seem to be particularly well-maintained. I'm currently bumping up against a security vulnerability that I can't patch (reliably) because of a transient dependency, so I'm hoping somebody out there has an alternative approach. A cursory web search turned up bupkis.
I'm not necessarily married to semantic-release, but I like it quite a bit and would prefer to continue using it if possible. (I've used release-please in the past and thought it was decent, as well, so if there's a solution involving that, maybe I'd be up for it.)
Does anybody else build changelogs from conventional commits using gitmoji? If so, what's your stack?
Every changelog autogenerated from a commit history using #ConventionalCommits is its own argument against ‘Conventional Commits’.
I'm not thrilled with #Drupal adopting #ConventionalCommits. For one thing, it doesn't properly spec the grammar of the description part. https://www.conventionalcommits.org/en/v1.0.0/
I run from anything with npm in the docs, so I built my own Conventional Commits linter—convcommitlint— simple, open source, GitHub Action included! All human-made, even the blog post. 😅
🔗 https://github.com/coolapso/convcommitlint/
📝 https://blog.coolapso.sh/en/posts/convcommitlint/
Keep your Git History Clean and Nice! dev.to/janderssonse... #opensource #git #conventionalcommits #rebase
A clean Git history with Git R...
Using conventional commits at work and someone puts everything as a ‘chore’, but like, I get it.
It took me weeks to finalize this blog post. Not because it was so difficult, but because I was so lazy! 😅 Anyway, it’s finally done. 🎉 Check out my Conventional Commit Types Cheatsheet https://www.bavaga.com/conventional-commit-types-cheatsheet/ #conventionalcommits #cheatsheet #devtips
PSA: I just updated https://thi.ng/monopub (a tool for synchronized publishing of packages in a monorepo) to add a new option to truncate generated changelogs with a configurable cut-off date (i.e. to exclude older versions). I've set the default cut off to 1st of January 3 years ago. As a result this saves around 1MB of just changelog files per full #ThingUmbrella release (for 200 packages). Older version history can of course still be obtained via Git...
Has someone attempted to create a version of #ConventionalCommits based on using git trailers instead of the commit message's subject?
I mean, there've always been people complaining that this kind of semantic metadata wouldn't belong into the free-form subject
A reddit user also suggested using git notes for this purpose, but compared to git trailers, they have some trade-offs in functionality
#amCoding #amProgramming #softwareDevelopment #softwareEngineering
Convention over customization.
This is why I love projects like Keep a Changelog, Conventional Commits and Semantic Versioning. They allow for some neat processing by other tools which just wouldn't ever be possible if everyone did their own, custom thing.
#git #ConventionalCommits #SemanticVersioning #Neovim #NeovimLazy
I hereby propose that the Conventional Commits standard is expanded by the commit type `tmp:`
> This commit type is used for placeholder work that is intentionally messy, unsquashable, and unreviewable. This type of commit is only meant to exist as a transient solution in the codebase, and the reasons for its presence are not subject to questioning. It serves a functional but impermanent purpose.
Characteristics:
- Cannot be squashed in a clean-up process.
- Often contains quick, messy, or experimental code.
- Reserved for situations where there’s an unspoken reason for its existence, known only to the contributor.
@mih
So what I hear is that you have a desire on commit messages themselves. I'm with ya, my take is that commit messages should contain the story of the patch in some format that suits the project, and sure, if a project chooses #ConventionalCommits, I'll go with that.
The Linux docs on submitting patches says something similar, BTW, but is a bit more elaborate (https://www.kernel.org/doc/html/latest/process/submitting-patches.html#describe-your-changes).
@levitte Thanks! For me #ConventionalCommits are an approach to ensure that all relevant information is in the repository -- not in a PR or an email that is detached and will be lost eventually, even when the sources are preserved. They can provide a narrative, tags, references in a non-intrusive, yet fully integrated fashion.
Still have to learn how they feel in practice, though.
@mih
The only two variants I know of that aren't #github style (some modicum of centralized forges) are email conversations (like #sourceforge promotes) and #radicle.
Not sure how #ConventionalCommits answer your request, but perhaps I misunderstand what you're looking for. Care to expand on that?
