name: Lint & Test on: workflow_call jobs: lint-test: runs-on: ubuntu-latest env: # Dummy values for required bot environment variables BOT_API_KEY: foo BOT_SENTRY_DSN: blah BOT_TOKEN: bar REDDIT_CLIENT_ID: spam REDDIT_SECRET: ham REDIS_PASSWORD: '' steps: - name: Checkout repository uses: actions/checkout@v5 - name: Install uv uses: astral-sh/setup-uv@v6 with: enable-cache: true cache-dependency-glob: "uv.lock" activate-environment: true - name: Install dependencies run: uv sync --frozen - name: Run pre-commit hooks run: SKIP=ruff pre-commit run --all-files # Run `ruff` using github formatting to enable automatic inline annotations. - name: Run ruff run: "ruff check --output-format=github ." - name: Run tests and generate coverage report run: pytest -n auto --cov -q env: # Use "sys.monitoring" based coverage backend for better speed (see #3075) COVERAGE_CORE: sysmon # Prepare the Pull Request Payload artifact. If this fails, we # we fail silently using the `continue-on-error` option. It's # nice if this succeeds, but if it fails for any reason, it # does not mean that our lint-test checks failed. - name: Prepare Pull Request Payload artifact id: prepare-artifact if: always() && github.event_name == 'pull_request' continue-on-error: true run: cat $GITHUB_EVENT_PATH | jq '.pull_request' > pull_request_payload.json # This only makes sense if the previous step succeeded. To # get the original outcome of the previous step before the # `continue-on-error` conclusion is applied, we use the # `.outcome` value. This step also fails silently. - name: Upload a Build Artifact if: always() && steps.prepare-artifact.outcome == 'success' continue-on-error: true uses: actions/upload-artifact@v4 with: name: pull-request-payload path: pull_request_payload.json