Skip to content

wemake-services/kira-release

Repository files navigation

Kira Release Bot

wemake.services kira-family test Conventional Commits Dependencies Status

Automate routine work with releasing new versions.

Part of the @kira bots family.

Release steps

So, when the time to release comes our bot:

  1. calculates a new semantic version since the last tagged release,
  2. summarizes the release notes from these commits,
  3. creates a signed GitLab release,
  4. creates a changelog entry, commits it to the project,
  5. tags a new release with the newly calculated version number,
  6. optionally uploads new docker images to GitLab registry

Internally we use semantic-release. With multiple plugins.

How does it work?

Every commit is validated to be compatible to conventional-changelog format (here's our format just for example):

/^(revert: )?(feat|fix|docs|refactor|chore)(\(.+\))?:.{1,50}refs #\d+/

However, we do not enforce any particular format with this project. You are free to choose any existing or create your own.

We also recommend to enforce this format in Gitlab's push rules.

In this case I won't be able to push incorrect commit messages:

» git push
Counting objects: 3, done.
Delta compression using up to 4 threads.
Compressing objects: 100% (3/3), done.
Writing objects: 100% (3/3), 692 bytes | 692.00 KiB/s, done.
Total 3 (delta 2), reused 0 (delta 0)
remote: GitLab: Commit message does not follow the pattern '/^(revert: )?(feat|fix|docs|refactor|chore) ...'

Gitlab CI setup

It is recommended to use this bot as a part of your CI. There are two possible ways to release your software:

  1. On every push to master (we consider master branch protected by default)
  2. Manually by setting up manual CI jobs

Extra settings

They are not required, but may help you:

  1. Invite @kira-bot to your project
  2. All tags must be protected, only @kira-bot is allowed to create them
  3. master branch must be protected, only @kira-bot is allowed to push

CI variables

You are required to set:

  • GITLAB_TOKEN secret variable in CI configuration

You can also optionally set:

  • KIRA_RELEASE_SKIP_DOCKER to 'true', so your docker image deploy will be skipped
  • KIRA_INCLUDE_COMMENTS one of ['all', 'issue', 'mergeRequest', 'off'] values when and where to post a comment on release, default: 'off'
  • KIRA_RELEASE_ASSETS to any assets string, so it will upload these files to GitLab release
  • KIRA_RELEASE_EXEC_CONFIG string with json config to pass to @semantic-release/exec. Related docs on why would you possibly need this
  • KIRA_RELEASE_REPLACE_CONFIG string with json config of how to bump version numbers in a project definition file, example: KIRA_RELEASE_REPLACE_CONFIG='{ "project": "package.json", "from": "\"version\": \".*\"", "to": "\"version\": \"${nextRelease.version}\"" }'
  • KIRA_RELEASE_BRANCHES to semantic release branches default: ['master', 'main']
  • GITLAB_URL to change the GitLab location, see semantic-release-gitlab docs, default: 'https://gitlab.com'

Note: You might want to use $$ to escape $ char in several environments like GitLab CI configuration file

Container release

We use five ebv variables to publish pre-built containers into the registry. Basically, this is ready to work with GitLab CI and GitLab Container registry, but you can modify it to work with any registry / CI. For example, we use GitHub Container Registry in this project's CI (see .github/workflows/test.yml).

Variables:

  • CI_REGISTRY: what registry you use: registry.gitlab.com, ghrc.io, or any other
  • CI_REGISTRY_USER: username to use for login
  • CI_JOB_TOKEN: the token we use to auth, can be any token that your registry supports
  • CI_PROJECT_NAMESPACE: organization name to use in image name
  • CI_PROJECT_NAME: project name to use in image name

Running

Copy-paste our .gitlab-ci.yml file. And modify it to match your project.

Credis

Special thanks to: