Links

Directory based affected targets

If you don't use Bazel like systems to auto-generate the affected targets, you can instead manually define targets based on your directory structure. There are two ways to do it.

Automated (preferred)

A GitHub action can be created to automatically assign affected targets to reach PR.
name: Aviator Queue PR
on:
pull_request:
types: [ labeled ]
jobs:
path_board:
if: ${{ github.event.label.name == 'av-ready' }}
runs-on: ubuntu-latest
steps:
- uses: dorny/paths-filter@v2
id: filter
with:
filters: |
workflows:
- '.github/workflows/*'
billing:
- billing/.*
tests:
- tests/*
other:
- *.txt
- run: |
TARGETS=${{ toJSON(steps.filter.outputs.changes) }}
PR_NUMBER=${{ github.event.number }}
curl -X POST -H "Authorization: Bearer ${{ secrets.AVIATOR_API_KEY }}" \
-H "Content-Type: application/json" \
-d '{
"action": "queue",
"pull_request": {
"number": '"$PR_NUMBER"',
"repository": {"name": "testrepo", "org": "ankitjaindce"},
"affected_targets": '"$TARGETS"'
}
}' https://api.aviator.co/api/v1/pull_request/
A few things to note:
  • We use path-filters action to capture all the file path changes in the given PR.
  • You may also want to store the API access token as AVIATOR_API_KEY in GitHub secrets. or replace the secrets above
  • filters are the various directory packages that you can define as various affected targets. It accepts glob format and you can define more than 1 path per affected target. You may also optionally specify this in a separate file. Read more on the path-filters action documentation.
  • This will queue the PR in Aviator. You can also want to setup a similar GH action to dequeue when the label is removed.
  • This label av-ready should be separate from the label defined in Aviator’s configuration.

Manually specify targets

If you prefer to do this manually, you can also specify the affected targets as a Slash command comment in your PR. The targets can be represented as a comma separated list in the /aviator merge command:
/aviator merge --targets=frontend,api,android
Although Manual approach is much simpler, it is prone to human errors and is a less preferred approach.