8.1 KiB
Contributing
When contributing to this repository, please first discuss the change you wish to make via issue, email, or any other method with the owners of this repository before making a change.
Code of Conduct
Please note we have a Code of Conduct, please follow it in all your interactions with the project.
Filing an Issue
-
Look for a similar issue that may already exist.
-
If none exists, create a new issue and provide as much description and context as possible.
-
Apply any relevant labels to the issue.
Pull Request Process
-
Find the issue you want to work on. If no relevant issue exists, see above and file an issue, then continue to the next step.
-
Assign the issue to yourself.
-
Create a branch in your local fork with the issue number in the name.
-
Implement the code changes described by the issue. All Python code changes should be accompanied by any relevant tests, either by modifying existing ones or adding new ones. Front-end changes (HTML/CSS/JS) that involve React code should create or update any relevant JS tests. Infrastructure changes (Docker related things) do not require tests.
-
When all your changes are complete to your satisfaction:
- Run
make check
to run the full test, linting, formatting, and coverage suites. - Fix any linting issues by running
make code_format
or by hand if the auto formatter is unable to fix it. - Fix any broken tests
- Ensure coverage remains at 100% by adding new tests for any uncovered lines.
-
When all tests and checks are passing, commit all your changes into a single commit and follow the Git Commit Guidelines
-
Push your branch up to your fork and submit a pull request on to main. Add any additional information you'd like to the pull request body, including descriptions of changes, screenshots of any UI changes, special instructions for testing, etc.
-
Find a reviewer. If you're not sure who should review, please contact us on #nimbus-dev on the Mozilla Slack if you have access to it.
-
Any PRs that require changes to deployment infrastructure ie environment variable changes, dependent services, etc also require a review by a member of the operations team. Please include any relevant instructions about the changes in the PR. File a Bugzilla ticket with a link to the PR so that operations can track the change.
-
If you receive feedback that requires changes to your pull request, make the changes locally, run
make check
again to ensure all tests and linting are passing, and then create a new commit that describes what feedback was addressed. This commit can be formatted however you like, it will be squashed before it is merged into main. -
When your pull request is approved, it can be closed by using the 'Squash and Merge' button to squash all of the commits into a single one that refers to both the issue and the pull request and contains any additional descriptive information.
-
Thank you for submitting changes to Experimenter :D
Git Commit Guidelines
Subject
The subject should follow the this pattern:
type(scope): Description
Type
One of the following
- feat: A new feature
- fix: A bug fix
- docs: Documentation only changes
- style: Changes that do not affect the meaning of the code (white-space, formatting, missing semi-colons, etc)
- refactor: A code change that neither fixes a bug or adds a feature
- perf: A code change that improves performance
- test: Adding missing tests
- chore: Changes to the build process or auxiliary tools and libraries such as documentation generation
Scope
One of the following:
- project: Anything that affects the entire project
- nimbus: Anything scoped only to Nimbus experiment frontend
- cirrus: Anything scoped to Cirrus.
- schemas: Anything scoped to Nimbus schemas.
- legacy: Anything scoped only to Legacy experiment frontend
- visualization: Anything scoped only to Analysis Visualization
Description
- use the imperative, present tense: "change" not "changed" nor "changes"
- don't capitalize first letter
- no dot (.) at the end
Body
The body should describe the purpose of the commit, so that it's clear why this change is being
made. To assist in writing this along with the footer, a git commit template (saved as ~/.gitmessage
)
can be used:
Because
* Reason
This commit
* Change
Fixes #github_issue_number
Just as in the subject, use the imperative, present tense: "change" not "changed" nor "changes". Commits are expected to follow this format.
Dependency Management
Dependencies are automatically updated by Dependabot which is now integrated
directly into GitHub. Each week Dependabot will create a large number of individual PRs that update
each dependency in each of the Experimenter and
Experimenter Docs repos. To merge those into the main
branch, use the following process:
Merge Dependabot PRs
Manually
- Dependabot will create many individual PRs against the
main
branch, which must pass all CI checks and be approved before merging. Approve each PR with the following comment and if they pass CI they will merge automatically:
@dependabot squash and merge
All done!
Automatically
- Install and configure the Github CLI
- From your local Experimenter repo run
make dependabot_approve
Failed Dependabot PRs
If a Dependabot PR fails the CI checks you can either investigate the failure and see if it can be resolved quickly/easily, or close it altogether.
Security Warnings
Dependabot will also produce Security Advisories for packages that have registered CVE numbers. These can not be resolved automatically. To resolve the security warnings:
-
Copy the Remediation version from the security warning into the
"resolutions"
section ofexperimenter/package.json
, example:"resolutions": { "postcss": "^7.0.36",
-
Update the
yarn.lock
file by runningyarn install
-
Commit your changes in a PR titled
chore(deps): Security <list affected packages>
-
Create a PR and request review
-
Merge when approved
Continuous Deployment Process
When a PR is merged into main it will automatically be deployed to the stage instance and if that is successful then it will automatically be deployed to production.
If the deployment fails at one of the stage/production steps, ops will automatically be paged.
If the deployment succeeds but the change inadvertently breaks stage/production, there are two options:
-
Revert the change using a revert commit
-
Create a branch from main called
revert-#123
where#123
is the number of the issue, and then revert the commit withgit revert <squashed commit hash>
. Fix any potential conflicts or lint/test failures and finish the revert commit. -
If the original PR included any database migrations, they must be preserved and reverted by creating a new subsequent migration
-
git checkout origin/main -- experimenter/experimenter/experiments/migrations/XXXX_migration_file.py
-
make makemigrations
-
git add .;git commit -m 'Reverted migration'
- Push the branch and create a PR as normal following the above PR process
-
-
Fix the change following the normal issue/PR processes above
Because all changes will go to stage and prod automatically, you can use stage to validate your changes.
Updating an existing fork
-
Go to your cloned directory and add the main repository as your upstream
git remote add upstream https://github.com/mozilla/experimenter
-
Fetch the latest changes from the main repository
git fetch upstream
-
Merge the changes from the main repository into your forked branch
git merge upstream/main