title | intro | versions | type | topics | shortTitle | |||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Building and testing Rust |
You can create a continuous integration (CI) workflow to build and test your Rust project. |
|
tutorial |
|
Build & test Rust |
{% data reusables.actions.enterprise-github-hosted-runners %}
This guide shows you how to build, test, and publish a Rust package.
{% data variables.product.prodname_dotcom %}-hosted runners have a tools cache with preinstalled software, which includes the dependencies for Rust. For a full list of up-to-date software and the preinstalled versions of Rust, see AUTOTITLE.
You should already be familiar with YAML syntax and how it's used with {% data variables.product.prodname_actions %}. For more information, see AUTOTITLE.
We recommend that you have a basic understanding of the Rust language. For more information, see Getting started with Rust.
{% data reusables.actions.workflow-templates-get-started %}
{% data variables.product.prodname_dotcom %} provides a Rust workflow template that should work for most basic Rust projects. The subsequent sections of this guide give examples of how you can customize this workflow template.
{% data reusables.repositories.navigate-to-repo %} {% data reusables.repositories.actions-tab %} {% data reusables.actions.new-starter-workflow %}
-
The "Choose a workflow" page shows a selection of recommended workflow templates. Search for "Rust".
-
Filter the selection of workflows by clicking Continuous integration.
-
On the "Rust - by {% data variables.product.prodname_actions %}" workflow, click Configure.
{%- ifversion ghes %}
If you don't find the "Rust - by {% data variables.product.prodname_actions %}" workflow template, copy the following workflow code to a new file called rust.yml
in the .github/workflows
directory of your repository.
name: Rust
on:
push:
branches: [ "main" ]
pull_request:
branches: [ "main" ]
env:
CARGO_TERM_COLOR: never
jobs:
build:
runs-on: ubuntu-latest
steps:
- uses: {% data reusables.actions.action-checkout %}
- name: Build
run: cargo build --verbose
- name: Run tests
run: cargo test --verbose
{%- endif %}
- Edit the workflow as required. For example, change the version of Rust.
- Click Commit changes.
{% ifversion fpt or ghec %}
The rust.yml
workflow file is added to the .github/workflows
directory of your repository.
{% endif %}
{% data variables.product.prodname_dotcom %}-hosted runners include a recent version of the Rust toolchain. You can use rustup to report on the version installed on a runner, override the version, and to install different toolchains. For more information, see The rustup book.
This example shows steps you could use to setup your runner environment to use the nightly build of rust and to report the version.
- name: Temporarily modify the rust toolchain version
run: rustup override set nightly
- name: Output rust version for educational purposes
run: rustup --version
You can cache and restore dependencies using the Cache action. This example assumes that your repository contains a Cargo.lock
file.
- name: Cache
uses: {% data reusables.actions.action-cache %}
with:
path: |
~/.cargo/registry
~/.cargo/git
target
key: {% raw %}${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}{% endraw %}
If you have custom requirements or need finer controls for caching, you should explore other configuration options for the cache
action. For more information, see AUTOTITLE.
You can use the same commands that you use locally to build and test your code. This example workflow demonstrates how to use cargo build
and cargo test
in a job:
jobs:
build:
runs-on: ubuntu-latest
strategy:
matrix:
BUILD_TARGET: [release] # refers to a cargo profile
outputs:
release_built: {% raw %}${{ steps.set-output.outputs.release_built }}{% endraw %}
steps:
- uses: {% data reusables.actions.action-checkout %}
- name: Build binaries in "{% raw %}${{ matrix.BUILD_TARGET }}{% endraw %}" mode
run: cargo build --profile ${% raw %}{{ matrix.BUILD_TARGET }}{% endraw %}
- name: Run tests in "${% raw %}{{ matrix.BUILD_TARGET }}{% endraw %}" mode
run: cargo test --profile ${% raw %}{{ matrix.BUILD_TARGET }}{% endraw %}
The release
keyword used in this example corresponds to a cargo profile. You can use any profile you have defined in your Cargo.toml
file.
Once you have setup your workflow to build and test your code, you can use a secret to login to crates.io and publish your package.
- name: Login into crates.io
run: cargo login {% raw %}${{ secrets.CRATES_IO }}{% endraw %}
- name: Build binaries in "release" mode
run: cargo build -r
- name: "Package for crates.io"
run: cargo package # publishes a package as a tarball
- name: "Publish to crates.io"
run: cargo publish # publishes your crate as a library that can be added as a dependency
If there are any errors building and packaging the crate, check the metadata in your manifest, Cargo.toml
file, see The Manifest Format. You should also check your Cargo.lock
file, see Cargo.toml vs Cargo.lock.
After a workflow completes, you can upload the resulting artifacts for analysis or to use in another workflow. You could add these example steps to the workflow to upload an application for use by another workflow.
- name: Upload release artifact
uses: {% data reusables.actions.action-upload-artifact %}
with:
name: {% raw %}<my-app>{% endraw %}
path: {% raw %}target/${{ matrix.BUILD_TARGET }}/<my-app>{% endraw %}
To use the uploaded artifact in a different job, ensure your workflows have the right permissions for the repository, see AUTOTITLE. You could use these example steps to download the app created in the previous workflow and publish it on {% data variables.product.github %}.
- uses: {% data reusables.actions.action-checkout %}
- name: Download release artifact
uses: {% data reusables.actions.action-download-artifact %}
with:
name: {% raw %}<my-app>{% endraw %}
path: ./{% raw %}<my-app>{% endraw %}
- name: Publish built binary to {% data variables.product.github %} releases
- run: |
gh release create --generate-notes ./{% raw %}<my-app>/<my-project>#<my-app>{% endraw %}