Skip to content

Latest commit

 

History

History
378 lines (261 loc) · 17.2 KB

CONTRIBUTING.md

File metadata and controls

378 lines (261 loc) · 17.2 KB
Raw

Contributing

Thanks for taking the time to join our community and start contributing. These guidelines will help you get started with the Contour project. Please note that we require DCO sign off.

Read this document for additional website specific guildlines: Site Contribution Guidelines. Guidelines in this document still apply to website contributions.

If you want to get more insight into how the Contour maintainer team approaches R&D, this page captures how we work on Contour.

Building from source

This section describes how to build Contour from source.

Prerequisites

  1. Install Go

    Contour requires Go 1.16 or later.

Fetch the source

Contour uses go modules for dependency management.

  1. Fork the repo

  2. Create a local clone

git clone [email protected]:YOUR-USERNAME/contour.git

Building

To build Contour, run:

make

This uses a go install and produces a contour binary in your $GOPATH/bin directory.

Running the unit tests

Once you have Contour building, you can run all the unit tests for the project:

make check

To run the tests for a single package, change to package directory and run:

go test .

Lint

Before making a commit, it's always a good idea to check the code for common programming mistakes, misspellings and other potential errors. The lint checks can be run by invoking the make lint task:

make lint

Note: The lint tasks require the codespell application. Be sure to install version 2.0 or newer before running the lint tasks.

Local Development/Testing

It's very helpful to be able to test out changes to Contour locally without building images and pushing into clusters.

To accomplish this, Envoy can be run inside a Kubernetes cluster, typically a kind cluster. Then Contour is run on your local machine and Envoy will be configured to look for Contour running on your machine vs running in the cluster.

  1. Create a kind cluster
kind create cluster --config=./examples/kind/kind-expose-port.yaml --name=contour
  1. Deploy Contour & Deps to cluster:
kubectl apply -f examples/contour

Note: The Contour Deployment/Service can be deleted if desired since it's not used.

  1. Find IP of local machine (e.g. ifconfig or similar depending on your environment)

  2. Edit Envoy Daemonset & change the xds-server value to your local IP address

kubectl edit ds envoy -n projectcontour

Change initContainers: to look like this updating the IP and removing the three envoy cert flags:

 initContainers:
  - args:
    - bootstrap
    - /config/envoy.json
    - --xds-address=<YOUR_IP_ADDRESS>
    - --xds-port=8001
    - --xds-resource-version=v3
    - --resources-dir=/config/resources

  1. Change your Contour code.

  2. Build & start Contour allowing Envoy to connect and get its configuration.

make install && contour serve --kubeconfig=$HOME/.kube/config --xds-address=0.0.0.0 --insecure

  1. Test using the local kind cluster by deploying resources into that cluster. Many of our examples use local.projectcontour.io which is configured to point to 127.0.0.1 which allows requests to route to the local kind cluster for easy testing.

  2. Make more changes and repeat step #6.

Contribution workflow

This section describes the process for contributing a bug fix or new feature. It follows from the previous section, so if you haven't set up your Go workspace and built Contour from source, do that first.

Before you submit a pull request

This project operates according to the talk, then code rule. If you plan to submit a pull request for anything more than a typo or obvious bug fix, first you should raise an issue to discuss your proposal, before submitting any code.

Depending on the size of the feature you may be expected to first write a design proposal. Follow the Proposal Process documented in Contour's Governance.

Issue and PR Triage

Project Board

In addition to maintaining the project repositories, project maintainers are responsible for maintaining the project tracking board. This board is intended to organize work for the team as well as provide visibility into the status and priority of tracks of work, specific Issues, and PRs. The board is used in conjunction with Issue and PR labels.

Priority of cards on the board flows from left to right, with the leftmost column representing what is currently being worked on. Within a column, priority flows from top to bottom, with the topmost cards having the highest priority.

The leftmost column should represent what is planned for the upcoming release. The "Investigating" column contains longer term items that may need more information, feedback from Issue reporters, etc. Further columns represent decreasing priority, with "Prioritized Backlog" containing cards that are coming soon, all the way to the "Unprioritized" column which contains items currently not yet sorted.

Notes for maintainers and contributors

  • If you are looking for work to pick up:
    • Look to the leftmost columns of the project board
    • New contributors can use this shortcut link to find good beginner Contour issues to work on and this shortcut link to find Operator issues
  • When a new Issue or PR is added, add it to the project board and make a best judgement on relative priority so we have a starting place in triage
  • When moving items between columns, please add a comment to the Issue or PR detailing why so we have context in triage sessions

Community Meeting Triage

The weekly Contour community meeting provides some time for the maintainer team and community to collaborate on Issue and PR triage.

Community members can add links to specific items they would like to discuss to the meeting notes. This time will be used for clarification, potential to bump priority in the queue of work items for the team, and the ability for contributors to provide more context to their contributions.

In addition, the meeting will be used to go over untriaged issues, longer-term items, and current progress on items in-flight as needed.

Procedural notes

  • Once an issue has been discussed, remove and add the appropriate labels and pull the item into the correct column
  • After triage, Issues and PRs that are actively being worked on should be assigned to someone. This could be:
    • the person working on the issue or PR or
    • a maintainer who is shepherding a new contributor or large PR
  • Issues and PRs that are not being actively worked on should have comments updated with the current context, then be unassigned

Commit message and PR guidelines

  • Have a short subject on the first line and a body. The body can be empty.
  • Use the imperative mood (ie "If applied, this commit will (subject)" should make sense).
  • There must be a DCO line ("Signed-off-by: David Cheney [email protected]"), see DCO Sign Off below.
  • Do not merge commits that don't relate to the affected issue (e.g. "Updating from PR comments", etc). Should the need to cherrypick a commit or rollback arise, it should be clear what a specific commit's purpose is.
  • Put a summary of the main area affected by the commit at the start, with a colon as delimiter. For example 'docs:', 'internal/(packagename):', 'design:' or something similar.
  • PRs must be labelled with a release-note/category label, where category is one of major, minor, small, docs, or infra, unless the change is really small, in which case it may have a release-note/not-required category. PRs may also include a release-note/deprecation label alone or in addition to the primary label.
  • PRs must include a file named changelogs/unreleased/PR#-githubID-category.md, which is a Markdown file with a description of the change. Please see changelogs/unreleased/<category>-sample.md for sample changelogs.
  • If main has moved on, you'll need to rebase before we can merge, so merging upstream main or rebasing from upstream before opening your PR will probably save you some time.
  • Pull requests must include a Fixes #NNNN or Updates #NNNN comment. Remember that Fixes will close the associated issue, and Updates will link the PR to it.

Commit message template

<packagename>: <imperative mood short description>

<longer change description/justification>

Updates #NNNN
Fixes #MMMM

Signed-off-by: Your Name <[email protected]>

Sample commit message

internal/contour: Add quux functions

To implement the quux functions from #xxyyz, we need to
florble the greep dots, then ensure that the florble is
warbed.

Fixes #xxyyz

Signed-off-by: Your Name <[email protected]>

Merging commits

Maintainers should prefer to merge pull requests with the Squash and merge option. This option is preferred for a number of reasons. First, it causes GitHub to insert the pull request number in the commit subject which makes it easier to track which PR changes landed in. Second, it gives maintainers an opportunity to edit the commit message to conform to Contour standards and general good practice. Finally, a one-to-one correspondence between pull requests and commits makes it easier to manage reverting changes and increases the reliability of bisecting the tree (since CI runs at a pull request granularity).

At a maintainer's discretion, pull requests with multiple commits can be merged with the Create a merge commit option. Merging pull requests with multiple commits can make sense in cases where a change involves code generation or mechanical changes that can be cleanly separated from semantic changes. The maintainer should review commit messages for each commit and make sure that each commit builds and passes tests.

Import Aliases

Naming is one of the most difficult things in software engineering. Contour uses the following pattern to name imports when referencing packages from other packages.

thing_version: The name+package path of the thing and then the version separated by underscores

Examples:

contour_api_v1 "github.com/projectcontour/contour/apis/projectcontour/v1"
contour_api_v1alpha1 "github.com/projectcontour/contour/apis/projectcontour/v1alpha1"
envoy_v3 "github.com/projectcontour/contour/internal/envoy/v3"
xdscache_v3 "github.com/projectcontour/contour/internal/xdscache/v3"

Pre commit CI

Before a change is submitted it should pass all the pre commit CI jobs. If there are unrelated test failures the change can be merged so long as a reference to an issue that tracks the test failures is provided.

Once a change lands in main it will be built and available at this tag, ghcr.io/projectcontour/contour:main. You can read more about the available contour images in the tagging document.

Build an image

To build an image of your change using Contour's Dockerfile, run these commands (replacing the repository host and tag with your own):

docker build -t ghcr.io/davecheney/contour:latest .
docker push ghcr.io/davecheney/contour:latest

or, you can use the make helper, like so:

REGISTRY=ghcr.io/davecheney VERSION=latest make push

This will push to :latest in ghcr.io/davecheney obviously you'll also need to replace the repo host with your own here too. If you don't specify VERSION, make push will push to a git hash tag (the output of git rev-parse --short=8 --verify HEAD).

Verify your change

To verify your change by deploying the image you built, take one of the deployment manifests, edit it to point to your new image, and deploy to your Kubernetes cluster.

Contour testing

This section provides some useful information and guidelines for working with Contour's tests.

Glossary

Config/Data Categories

  • Kubernetes Config: HTTPProxy, Ingress or Gateway API config that Contour watches and converts to Envoy config.
  • DAG: The internal Contour representation of L7 proxy concepts. Kubernetes config is first converted to DAG objects before being converted to Envoy config.
  • Envoy Config: Configuration that can be provided to Envoy via xDS. This is Contour's final output, generated directly from the DAG.

Test Categories

  • Unit Test: A Go test for a particular function/package. In some cases, these test more than one package at a time.
  • Feature Test: A Go test in internal/featuretests that tests the translation of Kubernetes config to Envoy config, using a Contour event handler and xDS server.
  • End-To-End (E2E) Test: A Go test in test/e2e that performs a full end-to-end test of Contour running in a cluster. Typically verifies the behavior of HTTP requests given a Kubernetes HTTPProxy, Ingress or Gateway API config.

Summary of Major Test Suites

The following table describes the major test suites covering the core Contour processing pipeline (Kubernetes config -> DAG -> Envoy config). In general, changes to the core processing pipeline should be accompanied by new/updated test cases in each of these test suites.

Test Suite Description
internal/dag/builder_test.go (specifically TestDAGInsert* functions) Tests conversion of Kubernetes config to DAG objects.
internal/dag/status_test.go Tests invalid Kubernetes (HTTPProxy) configs, verifying their status/conditions.
internal/envoy/v3/*_test.go Tests conversion of DAG objects to Envoy config.
internal/xdscache/v3/*_test.go (specifically the Test[Cluster|Listener|Route|Secret]Visit functions) Tests conversion of Kubernetes config to Envoy config.
internal/featuretests/v3/*_test.go Tests conversion of Kubernetes config to Envoy config, using a ~full Contour event handler and xDS server.
test/e2e/[httpproxy|gateway|ingress] E2E tests with Contour running in a cluster. Verifies behavior of HTTP requests for configured proxies.

DCO Sign off

All authors to the project retain copyright to their work. However, to ensure that they are only submitting work that they have rights to, we are requiring everyone to acknowledge this by signing their work.

Since this signature indicates your rights to the contribution and certifies the statements below, it must contain your real name and email address. Various forms of noreply email address must not be used.

Any copyright notices in this repository should specify the authors as "The project authors".

To sign your work, just add a line like this at the end of your commit message:

Signed-off-by: David Cheney <[email protected]>

This can easily be done with the --signoff option to git commit.

By doing this you state that you can certify the following (from https://developercertificate.org/):

Developer Certificate of Origin
Version 1.1

Copyright (C) 2004, 2006 The Linux Foundation and its contributors.
1 Letterman Drive
Suite D4700
San Francisco, CA, 94129

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.