[zephyr] Fix coordinator loop crash causing silent pipeline hangs #6045

	name: Marin - Tests
	on:
	push:
	branches:
	- main
	pull_request:


	concurrency:
	group: ${{ github.workflow }}-${{ github.event.pull_request.number \|\| github.ref }}
	cancel-in-progress: true

	jobs:
	changes:
	runs-on: ubuntu-latest
	permissions:
	contents: read
	pull-requests: read
	outputs:
	should_run: ${{ steps.filter.outputs.relevant }}
	steps:
	- uses: actions/checkout@v4
	- uses: dorny/paths-filter@v3
	id: filter
	with:
	filters: \|
	relevant:
	- 'lib/marin/**'
	- 'tests/**'
	- 'uv.lock'
	- 'pyproject.toml'
	- '.github/workflows/marin-unit-tests.yaml'

	marin-tests:
	needs: changes
	if: needs.changes.outputs.should_run == 'true'
	runs-on: ubuntu-latest
	timeout-minutes: 15
	steps:
	- name: Checkout code
	uses: actions/checkout@v3

	- name: Set up Python 3.12
	uses: actions/setup-python@v4
	with:
	python-version: "3.12"

	- name: Install uv
	uses: astral-sh/setup-uv@v6
	with:
	version: "0.7.20"
	enable-cache: true

	- name: Install dependencies
	run: \|
	python -m pip install --upgrade pip
	uv sync --package marin --extra cpu --extra dedup --group test --frozen

	- name: Set up Node.js 20.10.0
	uses: actions/setup-node@v4
	with:
	node-version: "20.10.0"

	- name: Test with pytest
	env:
	HF_TOKEN: ${{ secrets.HF_TOKEN }}
	# Avoid Ray propagating `uv run` to workers (can trigger slow per-worker venv setup in CI).
	RAY_ENABLE_UV_RUN_RUNTIME_ENV: "0"
	run: \|
	# Ensure we select the CPU torch wheels (and JAX CPU) for unit tests on GitHub runners.
	PYTHONPATH=tests:. uv run --package marin --extra cpu --frozen pytest -n 2 --dist=worksteal --durations=5 --tb=short -m 'not slow and not tpu_ci' -v tests/

Provide feedback