matrix-rust-sdk

# Contributing to `matrix-rust-sdk`

## Chat rooms

In addition to our [main Matrix room], we have a development room at

[#matrix-rust-sdk-dev:flipdot.org]. Please use it to discuss potential changes,

the overall direction of development and related topics.

[main Matrix room]: https://matrix.to/#/#matrix-rust-sdk:matrix.org

[#matrix-rust-sdk-dev:flipdot.org]: https://matrix.to/#/#matrix-rust-sdk-dev:flipdot.org

## Testing

You can run most of the tests that also happen in CI also using

`cargo xtask ci`. This needs a few dependencies to be installed, as it also runs

automatic WASM tests:

```bash

rustup component add clippy

cargo install cargo-nextest typos-cli wasm-pack

```

If you want to execute only one part of CI, there are a few sub-commands (see

`cargo xtask ci --help`).

Some tests are not automatically run in `cargo xtask ci`, for example the

integration tests that need a running synapse instance. These tests reside in

`./testing/matrix-sdk-integration-testing`. See its

[README](./testing/matrix-sdk-integration-testing/README.md) to easily set up a

synapse for testing purposes.

### Snapshot testing

You can add/review snapshot tests using [insta.rs](https://insta.rs)

Every new struct/enum that derives `Serialize` `Deserialise` should have a

snapshot test for it. Any code change that breaks serialisation will then break

a test, the author will then have to decide how to handle migration and test it

if needed.

And for an improved review experience it's recommended (but not necessary) to

install the `cargo-insta` tool:

Unix:

```shell

curl -LsSf https://insta.rs/install.sh | sh

```

Windows:

```shell

powershell -c "irm https://insta.rs/install.ps1 | iex"

```

Usual flow is to first run the test, then review them.

```shell

cargo insta test

cargo insta review

```

### Intermittent failure policy

While we strive to add test coverage for as many features as we can, it

sometimes happens that the tests will be intermittently failing in CI (such

tests are sometimes called "flaky"). This can be caused by race conditions

of all sorts, either in the test code itself, but sometimes in the underlying

feature being tested too, and as such, it requires some investigation, usually

from the original author of the test.

Whenever such an intermittent failure happens, we try to open an issue to track

the failures, adding the

[`intermittent-failure`](https://github.com/matrix-org/matrix-rust-sdk/issues?q=is%3Aissue%20state%3Aopen%20label%3Aintermittent-failure)

label to it, and commenting with links to CI runs where the failure happened.

If a test has been intermittently failing for **two weeks** or more, and no one

is actively working on fixing it, then we might decide to mark the test as

`ignored` until it is fixed, to not cause unrelated failures in other

contributors' pull requests and pushes.

## Pull requests

Ideally, a PR should have a _proper title_, with _atomic logical commits_, and

each commit should have a _good commit message_.

A _proper PR title_ would be a one-liner summary of the changes in the PR,

following the same guidelines of a good commit message, including the

area/feature prefix. Something like `FFI: Allow logs files to be pruned.` would

be a good PR title.

(An additional bad example of a bad PR title would be `mynickname/branch name`,

that is, just the branch name.)

## Writing changelog entries

We use towncrier to generate changelogs from individual fragment files. Each

change must be documented as a changelog fragment in the appropriate crate:

```text

/crates/<crate-name>/changelog.d/

```

Each changelog fragment must include **both the pull request number and the

fragment type** in its filename:

```text

<PR number>.<fragment type>.md

```

For example:

```text

4357.added.md

4357.fixed.md

4357.changed.md

```

The pull request link is automatically added during changelog generation, so it

must not be included manually.

We use standard towncrier fragment types to categorize changes. Common types

include:

- `added` – new features or functionality

- `changed` – changes in existing behavior

- `fixed` – bug fixes

- `removed` – removed features or APIs

- `security` – security-related fixes

- `internal` – changes that do not affect users (refactoring, CI, tooling)

Choose the type that best matches the nature of the change.

A well-written changelog entry should be understandable even to those who may

not be deeply familiar with the project. Provide enough context to ensure

clarity and ease of understanding.

A couple of examples of bad changelog entry would look like:

```markdown

- Fixed a panic.

```

```markdown

Added the Bar function to Foo.

```

A good example of a changelog entry could look like the following:

```markdown

Use the inviter's server name and the server name from the room alias as

fallback values for the via parameter when requesting the room summary from the

homeserver. This ensures requests succeed even when the room being previewed is

hosted on a federated server.

```

For security-related changelog entries, please include the following additional

details alongside the pull request number:

- Impact: Clearly describe the issue's potential impact on users or systems.

- CVE Number: If available, include the CVE (Common Vulnerabilities and

  Exposures) identifier.

- GitHub Advisory Link: Provide a link to the corresponding GitHub security

  advisory for further context.

```markdown

Use a constant-time Base64 encoder for secret key material to mitigate

side-channel attacks leaking secret key material

[CVE-2024-40640](https://www.cve.org/CVERecord?id=CVE-2024-40640),

[GHSA-j8cm-g7r6-hfpq](https://github.com/matrix-org/vodozemac/security/advisories/GHSA-j8cm-g7r6-hfpq)).

```

## Commit message format

Commit messages should be formatted as Conventional Commits. In addition, some

git trailers are supported and have special meaning (see below).

### Conventional commits

Conventional Commits are structured as follows:

```text

<type>(<scope>): <short summary>

```

The type of changes which will be included in changelogs is one of the

following:

- `feat`: A new feature

- `fix`: A bugfix

- `doc`: Documentation changes

- `refactor`: Code refactoring

- `perf`: Performance improvements

- `ci`: Changes to CI configuration files and scripts

The scope is optional and can specify the area of the codebase affected (e.g.,

olm, cipher).

### Security fixes

Commits addressing security vulnerabilities must include specific trailers for

vulnerability metadata, which should also be reflected in the corresponding

changelog entry.

The metadata must be included in the following git-trailers:

- `Security-Impact`: The magnitude of harm that can be expected, i.e.

  low/moderate/high/critical.

- `CVE`: The CVE that was assigned to this issue.

- `GitHub-Advisory`: The GitHub advisory identifier.

Please include all the fields that are available.

Example:

```text

fix(crypto): Use a constant-time Base64 encoder for secret key material

This patch fixes a security issue around a side-channel vulnerability[1]

when decoding secret key material using Base64.

In some circumstances an attacker can obtain information about secret

secret key material via a controlled-channel and side-channel attack.

This patch avoids the side-channel by switching to the base64ct crate

for the encoding, and more importantly, the decoding of secret key

material.

Security-Impact: Low

CVE: CVE-2024-40640

GitHub-Advisory: GHSA-j8cm-g7r6-hfpq

```

## Review process

To streamline the review process and make it easier for maintainers to review

your contributions, follow these basic rules:

1. Do not force push after a review has started. This helps maintainers track

   incremental changes without confusion and makes it easier to follow the

   evolution of the code.

2. Do not mix moves and refactoring with functional changes. Keep these in

   separate commits for clarity. This ensures that the purpose of each commit is

   clear and easy to review.

3. Each commit must compile. If commits don’t compile, git bisect becomes

   unusable, which hampers the debugging process and makes it harder to identify

   the source of issues.

4. Commits should only introduce test failures if they are proving that a bug

   exists. New features should never introduce test failures. Test failures

   should only be used to demonstrate existing bugs, not as part of adding new

   functionality.

5. Keep PRs on topic and small. Large PRs are harder to review and more prone to

   delays. Create small, focused commits that address a single topic. Use a

   combination of [git add] -p or [git checkout] -p to split changes into

   logical units. This makes your work easier to review and reduces the chance

   of introducing unrelated changes.

[git add]: https://git-scm.com/docs/git-add#Documentation/git-add.txt---patch

[git checkout]: https://git-scm.com/docs/git-checkout#Documentation/git-checkout.txt---patch

### Addressing review comments using fixup commits

So you posted a PR and the maintainers aren't quite happy with it. Here are some

guidelines to make the maintainers life easier and increase the chances that

your PR will be reviewed swiftly.

1. Use [fixup] commits. When addressing reviewer feedback, you can create fixup

   commits. These commits mark your changes as corrections of specific previous

   commits in the PR.

Example:

```shell

git commit --fixup=<commit-hash>

```

This command creates a new commit that refers to an existing one, making it

easier to rebase and squash later while showing reviewers the history of fixes.

For extra points, link to the fixup commit in the thread where the change was

requested.

2. After all requested changes were addressed, feel free to re-request a review.

   People might not notice that all changes were addressed.

3. Once the PR has been approved, rebase your PR to squash all the fixup

   commits, the [autosquash] option can help with this.

```shell

git rebase main --interactive --autosquash

```

[fixup]: https://git-scm.com/docs/git-commit#Documentation/git-commit.txt---fixupamendrewordltcommitgt

[autosquash]: https://git-scm.com/docs/git-rebase#Documentation/git-rebase.txt---autosquash

## Sign off

In order to have a concrete record that your contribution is intentional

and you agree to license it under the same terms as the project's

license, we've adopted the same lightweight approach that the [Linux

Kernel](https://www.kernel.org/doc/Documentation/SubmittingPatches),

[Docker](https://github.com/docker/docker/blob/master/CONTRIBUTING.md),

and many other projects use: the DCO ([Developer Certificate of

Origin](http://developercertificate.org/)). This is a simple declaration that

you wrote the contribution or otherwise have the right to contribute it to

Matrix:

```text

Developer Certificate of Origin

Version 1.1

Copyright (C) 2004, 2006 The Linux Foundation and its contributors.

660 York Street, Suite 102,

San Francisco, CA 94110 USA

Everyone is permitted to copy and distribute verbatim copies of this

license document, but changing it is not allowed.

Developer's Certificate of Origin 1.1

By making a contribution to this project, I certify that:

(a) The contribution was created in whole or in part by me and I

    have the right to submit it under the open source license

    indicated in the file; or

(b) The contribution is based upon previous work that, to the best

    of my knowledge, is covered under an appropriate open source

    license and I have the right under that license to submit that

    work with modifications, whether created in whole or in part

    by me, under the same open source license (unless I am

    permitted to submit under a different license), as indicated

    in the file; or

(c) The contribution was provided directly to me by some other

    person who certified (a), (b) or (c) and I have not modified

    it.

(d) I understand and agree that this project and the contribution

    are public and that a record of the contribution (including all

    personal information I submit with it, including my sign-off) is

    maintained indefinitely and may be redistributed consistent with

    this project or the open source license(s) involved.

```

If you agree to this for your contribution, then all that's needed is to

include the line in your commit or pull request comment:

```text

Signed-off-by: Your Name <your@email.example.org>

```

Git allows you to add this signoff automatically when using the `-s` flag to

`git commit`, which uses the name and email set in your `user.name` and

`user.email` git configs.

If you forgot to sign off your commits before making your pull request and are

on Git 2.17+ you can mass signoff using rebase:

```text

git rebase --signoff origin/main

```

## AI policy

This policy is a copy of the [Forgejo's AI agreement][Forgejo].

### Terminology

This does not necessarily reflect the official or commonly used terminology.

Software and services that heavily rely on large language model technology to

generate their outcomes are referred to as _Artificial Intelligence_ (AI).

Examples of products that fit this definition: GitHub Copilot, ChatGPT, Claude

Sonnet, DeepSeek, Llama and Gemini.

There is a distinction between _general_ and _narrow_ AI, all the aforementioned

examples fall under general AI as they were not trained to execute a specific

well-defined task. Narrow AI is trained to be used for specific well-defined

tasks where the problem space is known in advance.

_Vibe coding_ is the practice where AI creates a code change (feature, bugfix,

tests, refactor) with a human that describes what needs to be implemented.

_AI agents_ are AIs that are configured to perform interactions or make changes

with little to no human supervision.

### Agreement

1. If content was made with the help of AI, you **must** convey that this is

   the case. This includes content that you authored but was motivated by a

   suggestion of AI.

2. If at any point you used AI's work in your contribution you should make

   an effort to **verify** that you can submit this under the license of the

   repository.

3. The **accountability** of using AI in a contribution lies with the person

   that makes that contribution.

4. All communication, that includes: commit messages, pull request messages,

   documentation, code comments and issues (and comments on issues/pull

   requests), that is intended to be read by people to understand your thoughts

   and work **must not** have been generated with AI. We exclude machine

   translation and tooling that helps with grammar and spelling check.

5. Using general AI for review is **forbidden**. If the change contains changes

   to the user experience it has to be approved by a human reviewer.

6. It is **not allowed** to use AI in an autonomous-looking way to contribute to

   the Matrix Rust SDK. This also applies when someone engages in _vibe coding_

   or uses so-called _agent mode_.

[Forgejo]: https://codeberg.org/forgejo/governance/src/branch/main/AIAgreement.md