tokio/docs/contributing/pull-requests.md

Pull Requests

Pull Requests are the way concrete changes are made to the code, documentation, and dependencies in the Tokio repository.

Even tiny pull requests (e.g., one-character pull request fixing a typo in API documentation) are greatly appreciated. Before making a large change, it is usually a good idea to first open an issue describing the change to solicit feedback and guidance. This will increase the likelihood of the PR getting merged.

Cargo Commands

Due to the extensive use of features in Tokio, you will often need to add extra arguments to many common cargo commands. This section lists some commonly needed commands.

Some commands just need the --all-features argument:

cargo build --all-features
cargo check --all-features
cargo test --all-features

NOTE: there are some features that are not supported in every system, so you might need to specify which features you want to pass to cargo (e.g., cargo check --features=full,io-uring)

Ideally, you should use the same version of clippy as the one used in CI (defined by env.rust_clippy in ci.yml), because newer versions might have new lints:

cargo +1.88 clippy --all --tests --all-features

When building documentation, a simple cargo doc is not sufficient. To produce documentation equivalent to what will be produced in docs.rs's builds of Tokio's docs, please use:

RUSTDOCFLAGS="--cfg docsrs --cfg tokio_unstable" RUSTFLAGS="--cfg docsrs --cfg tokio_unstable" cargo +nightly doc --all-features [--open]

This turns on indicators to display the Cargo features required for conditionally compiled APIs in Tokio, and it enables documentation of unstable Tokio features. Notice that it is necessary to pass cfg flags to both RustDoc and rustc.

There is a more concise way to build docs.rs-equivalent docs by using cargo docs-rs, which reads the above documentation flags out of Tokio's Cargo.toml as docs.rs itself does.

cargo install --locked cargo-docs-rs
cargo +nightly docs-rs [--open]

The cargo fmt command does not work on the Tokio codebase. You can use the command below instead:

# Mac or Linux
rustfmt --check --edition 2021 $(git ls-files '*.rs')

# Powershell
Get-ChildItem . -Filter "*.rs" -Recurse | foreach { rustfmt --check --edition 2021 $_.FullName }

The --check argument prints the things that need to be fixed. If you remove it, rustfmt will update your files locally instead.

You can run loom tests with

cd tokio # tokio crate in workspace
LOOM_MAX_PREEMPTIONS=1 LOOM_MAX_BRANCHES=10000 RUSTFLAGS="--cfg loom -C debug_assertions" \
    cargo test --lib --release --features full -- --test-threads=1 --nocapture

Additionally, you can also add --cfg tokio_unstable to the RUSTFLAGS environment variable to run loom tests that test unstable features.

You can run miri tests with

MIRIFLAGS="-Zmiri-disable-isolation -Zmiri-strict-provenance" \
    cargo +nightly miri test --features full --lib --tests

Performing spellcheck on tokio codebase

You can perform a spell-check on the Tokio codebase. For details of how to use the spellcheck tool, feel free to visit https://github.com/drahnr/cargo-spellcheck

# First install the spell-check plugin
cargo install --locked cargo-spellcheck

# Then run the cargo spell check command
cargo spellcheck check

If the command rejects a word, you should backtick the rejected word if it's code related. If not, the rejected word should be put into spellcheck.dic file.

Note that when you add a word into the file, you should also update the first line which tells the spellcheck tool the total number of words included in the file

Tests

If the change being proposed alters code (as opposed to only documentation for example), it is either adding new functionality to Tokio or it is fixing existing, broken functionality. In both of these cases, the pull request should include one or more tests to ensure that Tokio does not regress in the future. There are two ways to write tests: integration tests and documentation tests. (Tokio avoids unit tests as much as possible).

Tokio uses conditional compilation attributes throughout the codebase, to modify rustc's behavior. Code marked with such attributes can be enabled using RUSTFLAGS and RUSTDOCFLAGS environment variables. One of the most prevalent flags passed in these variables is the --cfg option. To run tests in a particular file, check first what options #![cfg] declaration defines for that file.

For instance, to run a test marked with the 'tokio_unstable' cfg option, you must pass this flag to the compiler when running the test.

$ RUSTFLAGS="--cfg tokio_unstable" cargo test -p tokio --all-features --test rt_metrics

Integration tests

Integration tests go in the same crate as the code they are testing. Each sub crate should have a dev-dependency on tokio itself. This makes all Tokio utilities available to use in tests, no matter the crate being tested.

The best strategy for writing a new integration test is to look at existing integration tests in the crate and follow the style.

Fuzz tests

Some of our crates include a set of fuzz tests, this will be marked by a directory fuzz. It is a good idea to run fuzz tests after each change. To get started with fuzz testing you'll need to install cargo-fuzz.

cargo install --locked cargo-fuzz

To list the available fuzzing harnesses you can run;

$ cd tokio
$ cargo fuzz list
fuzz_linked_list

Running a fuzz test is as simple as;

cargo fuzz run fuzz_linked_list

NOTE: Keep in mind that by default when running a fuzz test the fuzz harness will run forever and will only exit if you ctrl-c or it finds a bug.

Documentation tests

Ideally, every API has at least one [documentation test] that demonstrates how to use the API. Documentation tests are run with cargo test --doc. This ensures that the example is correct and provides additional test coverage.

The trick to documentation tests is striking a balance between being succinct for a reader to understand and actually testing the API.

Same as with integration tests, when writing a documentation test, the full tokio crate is available. This is especially useful for getting access to the runtime to run the example.

The documentation tests will be visible from both the crate-specific documentation and the tokio facade documentation via the re-export. The example should be written from the point of view of a user that is using the tokio crate. As such, the example should use the API via the facade and not by directly referencing the crate.

The type level example for tokio::time::timeout provides a good example of a documentation test:

/// Create a new `Timeout` set to expire in 10 milliseconds.
///
/// ```rust
/// use tokio::time::timeout;
/// use tokio::sync::oneshot;
///
/// use std::time::Duration;
///
/// # async fn dox() {
/// let (tx, rx) = oneshot::channel();
/// # tx.send(()).unwrap();
///
/// // Wrap the future with a `Timeout` set to expire in 10 milliseconds.
/// if let Err(_) = timeout(Duration::from_millis(10), rx).await {
///     println!("did not receive value within 10 ms");
/// }
/// # }
/// ```

Lines that start with /// # are removed when the documentation is generated.

Benchmarks

You can run benchmarks locally for the changes you've made to the tokio codebase. Tokio currently uses Criterion as its benchmarking tool. To run a benchmark against the changes you have made, for example, you can run;

cd benches

# Run all benchmarks.
cargo bench

# Run all tests in the `benches/fs.rs` file
cargo bench --bench fs

# Run the `async_read_buf` benchmark in `benches/fs.rs` specifically.
cargo bench async_read_buf

# After running benches, you can check the statistics under `tokio/target/criterion/`

You can also refer to Criterion docs for additional options and details.

Commits

It is a recommended best practice to keep your changes as logically grouped as possible within individual commits. There is no limit to the number of commits any single Pull Request may have, and many contributors find it easier to review changes that are split across multiple commits.

That said, if you have a number of commits that are "checkpoints" and don't represent a single logical change, please squash those together.

Note that multiple commits often get squashed when they are landed (see the notes about commit squashing).

Commit message guidelines

A good commit message should describe what changed and why.

1. The first line should:

• contain a short description of the change (preferably 50 characters or less, and no more than 72 characters)
• be entirely in lowercase with the exception of proper nouns, acronyms, and the words that refer to code, like function/variable names
• start with an imperative verb
• not have a period at the end
• be prefixed with the name of the module being changed; usually this is the same as the M-* label on the PR

Examples:

• time: introduce Timeout and deprecate Deadline
• codec: export Encoder, Decoder, Framed*
• ci: fix the FreeBSD ci configuration

2. Keep the second line blank.

3. Wrap all other lines at 72 columns (except for long URLs).

4. If your patch fixes an open issue, you can add a reference to it at the end of the log. Use the Fixes: # prefix and the issue number. For other references use Refs: #. Refs may include multiple issues, separated by a comma.

   Examples:

   • Fixes: #1337
   • Refs: #1234

Sample complete commit message:

module: explain the commit in one line

Body of commit message is a few lines of text, explaining things
in more detail, possibly giving some background about the issue
being fixed, etc.

The body of the commit message can be several paragraphs, and
please do proper word-wrap and keep columns shorter than about
72 characters or so. That way, `git log` will show things
nicely even when it is indented.

Fixes: #1337
Refs: #453, #154

Opening the Pull Request

From within GitHub, opening a new Pull Request will present you with a template that should be filled out. Please try to do your best at filling out the details, but feel free to skip parts if you're not sure what to put.

Discuss and update

You will probably get feedback or requests for changes to your Pull Request. This is a big part of the submission process so don't be discouraged! Some contributors may sign off on the Pull Request right away, others may have more detailed comments or feedback. This is a necessary part of the process in order to evaluate whether the changes are correct and necessary.

Any community member can review a PR and you might get conflicting feedback. Keep an eye out for comments from code owners to provide guidance on conflicting feedback.

Once the PR is open, do not rebase the commits. See Commit Squashing for more details.

Commit Squashing

In most cases, do not squash commits that you add to your Pull Request during the review process. When the commits in your Pull Request land, they may be squashed into one commit per logical change. Metadata will be added to the commit message (including links to the Pull Request, links to relevant issues, and the names of the reviewers). The commit history of your Pull Request, however, will stay intact on the Pull Request page.