Skip to content

Latest commit

 

History

History
115 lines (70 loc) · 6.18 KB

CONTRIBUTING.md

File metadata and controls

115 lines (70 loc) · 6.18 KB

Contributing

Wrangler is an open source project because we believe that folks should have access, insight, and the opportunity to contribute to their developer tools. Wrangler is also a product delivered by Cloudflare, so it's important to clarify how we think about issue triage and contributions.

If you want to learn about how to get started developing Wrangler, click here

People

Wrangler is owned by the Cloudflare Workers Team and maintained by community members like you! The core maintainers are @EverlastingBugstopper and @ashleymichal, and everything that gets merged must be approved by at least one team member.

Developing Wrangler

Get started

To get started with developing Wrangler, we recommend that you first get up to speed with Wrangler. Then, get up to speed with the basics of Rust. (You'll need to install rustup where we're going).

Build Wrangler from source

To build Wrangler from source, fork the repo, clone it, and cd wrangler. To run the test suite, run cargo test. To quickly test a command, for example wrangler init, run cargo run -- init. This will compile a "debug" binary, then run it, passing in all arguments after --.

It is important to note that many Wrangler commands depend on a wrangler.toml in the current working directory. It is advised to keep a test project somewhere with an active wrangler.toml to exercise your code.

Run Wrangler during development

There are many ways to build and execute a development version of Wrangler, in addition to using cargo run:

cargo build will produce a local binary at ./target/debug/wrangler that you can execute just by calling it by path. cargo install --debug --path . will replace any globally installed wrangler with the one you've just built from source.

You can read more about cargo here, just find something that works for you.

Module System

Each of the commands supported by Wrangler have entrypoints in ./src/commands. It's useful to understand the module system if you will be adding new commands or need to refactor/organize imports.

Notable external libraries

Command-line argument parsing (clap)

The primary framework we use for developing Wrangler is called clap, which provides fast and structured argument parsing. This is how different features are exposed to users, and most of that logic lives in main.rs.

API calls (cloudflare-rs)

When developing a new feature for Wrangler, it's quite common to need to make API calls. The way we do this is by submitting a PR to cloudflare-rs and releasing a new version of that library to depend on. There are some legacy endpoints that we use our own client for, but the goal is to eventually move everything to cloudflare-rs. All endpoint calls should be made with the clients in ./src/http.

Figure out where to start

If you're working on a specific issue, make sure there is buy-in from the Wrangler team before starting, and feel free to ask where you should start. We're more than happy to help!

Requirements for merging a PR

Passing tests

When adding features or fixing bugs, we'd love if you would add a test! There are two types of tests in Wrangler, integration and unit tests. To execute tests, you can run cargo test. You can read more about testing here. All tests must pass when submitting a new PR, this is enforced by our GitHub Actions runners, which run each test on Windows, MacOS, and Linux.

Proper formatting

You must run cargo fmt on your code before CI will allow you to merge your PR.

Primary Issue Triage

Within 3 days, any incoming issue should be triaged. Triage involves:

  • reading the issue and requesting any further information
  • always thank the user for submitting
  • assigning appropriate labels

Labelling

  • each issue should have a status label and a category label, and they should be kept up to date
    • once design has been settled for an issue, please label with status - PR welcome
  • each issue from non-team members should be labelled user report (issue templates assign this automatically)
  • subject labels and other call to actions are nice to have
  • if an issue seems easy to tackle, please label with good first issue so new contributors can use it to ramp up

Assignment

  • if the issue will require a large amount of back and forth between the reporter and the team assign a single team member to manage the conversation

Product Issue Triage

Once a week, the team holds the Wrangler Contributors meeting. This is where we assign work and update our plans for the milestones and releases.

Assignment and Milestones

  • assign all issues for the next two releases a milestone
  • assign all issues for the current milestone a person to take point

Pull Request Triage

Within 3 days, all incoming Community PRs should be triaged. If a team member opens a PR it should be triaged immediately upon open by the PR author.

Labelling

  • All work-in-progress PRs should be labelled work in progress and the title should be annotated [WIP] for easy scanning. No WIP PRs will be reviewed until the annotations are removed.
  • All PRs that need to be reviewed should be labelled needs review until they have received all required reviews.
  • All PRs should be labelled with a changelog label: BREAKING, feature, fix, maintenance, docs, N/A
  • All PRs that are ready for review should be tagged with the appropriate release milestone

Merging

  • All PRs should be merged with a Merge Commit. We recommend that folks rebase into a small number of task driven commits. This is enforced more heavily for team members than community members. Be reasonable.
  • All PRs should be labelled with the current milestone before merging. If a PR for an issue labelled with a different milestone is to be merged, update the issue milestone as well.