Contributing guidelines

Thank you (again) for your interest in the project!

Talk to us before committing to a change

We recommend you come and talk to us in the channel and/or open an issue before you start working on a feature, to see if it aligns with our goals.

We'll do our best to review and discuss changes with you but we're also humans with other activities, please be patient with us too.

General guidelines

The library is still in development, and while this is the case we adopt a fail-fast strategy, which is also a reason for the choice of the language.

The earlier we catch bugs and fix them, the less chances they have to confuse users of our library, or make end-users give up on software developed using this library. This also helps improving other software in the ecosystem.

We will not accept changes (code or otherwise) created with the aid of "AI" tooling. "AI" models are trained at the expense of underpaid workers filtering inputs of abhorrent content, and does not respect the owners of input content. Ethically, it sucks.

Nor will we accept changes (code or otherwise) that move xmpp-rs crates toward enterprise / surveillance / other toxic capitalist pursuits.

Keep commits short and meaningful

To help with reviews, to facilitate reverts, reading and grepping through commit history, bisects: it is important for us that changes come in small and meaningful commits.

That is, don't bundle all the things in a single commit, if a change can be added via different commits standalone and still make sense, then it's probably a good candidate for splitting.

Meta

Do not forget to update changelogs and other crate metadata where necessary.

Signing commits (git commit -S) and adding DCO bits (git commit -s) are welcome but not mandatory.

Ensure CI passes

Code changes should include documentation. They should also include tests where appropriate, and pass the existing test suite.

CI should pass to submit your changes. This is done by ensuring cargo fmt and cargo test pass (in the workspace). Please run cargo fmt as part of each of your commits.

More thorough tests can be done locally with act or forgejo-runner exec, which is also what is run in the CI. We require docker to be setup for this to work.

Making a release

We are using the semver versioning system.

For each crate you are releasing:

• Ensure there aren't urgent changes in the buffer waiting to be released too. Even though increasing a number is cheap, the process to get there isn't doing itself
• Have a run of cargo clippy if it hasn't been done (TODO: Add to CI)
• Ensure tests, docstrings and examples pass (TODO: Add --all-features to CI.)
• Ensure dependencies are up-to-date.
• Update the version attribute in the Cargo.toml file.
• Update the ChangeLog file with the version number, the release date etc.
• Rework the ChangeLog if necessary. Add missing entries, merge duplicates, unify wording, make sure everything is where it should be, add meta information such as merge requests, etc.
• Send all of this as a merge request.
• If the change is significant enough, prepare a blog article about it.

For the person with commit access

• git tag -s releases
• cargo publish
• Publish blog article
• Promote on social media