Files
computer-playbook/.github/workflows/call-validation.yml
T
kevinveenbirkenbachandClaude Opus 5.5 a5f00ca777 fix(ci): let a cancel reap jobs guarded by always()
Closing a PR left "Cancel: PR Runs on Close" (run 36385268627) on
pending for hours. The canceller joins the PR orchestrator's
concurrency group, so GitHub cancelled the orchestrator run within a
second. But a cancel re-evaluates the job-level if: of every live job
and spares each one whose condition still holds. The deploy chunks
started with always(), which always holds: test-workspace died at
06:11:59, while Chunk 0 kept deploying and its ~60 queued matrix jobs
kept waiting for runners. The group stays occupied until the occupant
is terminal, so the canceller parked behind the very run it had
already cancelled.

The chunks, test-runner-smoke, the fork-prereqs gate and the release
job now guard with !cancelled(). Outside a cancel it matches always():
they still run after a failed or skipped need. On a cancel it turns
false and the job is cancelled with the run.

done and report-main-failures keep always(). They exist to report a
cancelled run, and a done skipped on cancel would count as a passing
required check.

The new lint walks every workflow whose concurrency cancels in
progress, plus everything it calls transitively, and rejects a
job-level always() outside that allowlist. It found the release job,
which a line-based grep had missed. It also fails when an allowlisted
job stops needing the exemption, and when its scope no longer reaches
the deploy workflows, so it cannot go silently empty. Against the
pre-fix tree it reports all six offenders.

Behaviour change: manually cancelling a release run now stops the
retag too. cron-release-highest re-dispatches the release.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 09:07:24 +02:00

205 lines
7.9 KiB
YAML

# The gate every entry point passes before it spends a runner on anything else.
# What it checks depends on the event and on the inputs it is handed, not on
# which workflow called it: a pull request gets its branch prefix and its merge
# base asserted, a run that carries a selection gets that selection checked
# against the branch it would deploy from, and a fork PR waits here for the
# privileged run's images. Nothing downstream may start before this is green,
# so a refusal fails the whole run instead of surfacing as one red job beside a
# chain that keeps going.
#
# Under pull_request_target the token is privileged and the head belongs to the
# fork, so the code under test is not checked out at all there and the branch
# prefix is validated from the calling ref's own tooling. Only the merge
# assertion and the selection check read the head, and both are reachable on
# pull_request and workflow_dispatch alone.
name: "✅ CI: Validation"
on:
workflow_call:
inputs:
pr_scope:
description: "🔎 Detected PR scope. Set: the branch prefix is validated against it. Empty: not a PR entry, skipped."
required: false
type: string
default: ""
whitelist:
description: "🎯 Selection to validate against this branch. Empty: nothing to check."
required: false
type: string
default: ""
priority:
description: "⭐ Priority line to validate against this branch. Empty: nothing to check."
required: false
type: string
default: ""
mode:
description: "🚀 Deploy modes the run draws from; narrows what a pinned mode may be."
required: false
type: string
default: "auto"
tor:
description: "🧅 Tor axis of the run; decides whether a pinned onion state is reachable."
required: false
type: string
default: "auto"
distros:
description: "🐧 Distro pool of the run; narrows what a pinned distro may be."
required: false
type: string
default: ""
filesystem:
description: "🗄️ Filesystem pool of the run; narrows what a pinned filesystem may be."
required: false
type: string
default: ""
lifecycles:
description: "♻️ Lifecycle envelope; a role outside it has no row to match."
required: false
type: string
default: ""
fork_prereqs:
description: "🍴 Wait for the privileged fork run's images. Off for entries that build their own."
required: false
type: boolean
default: false
trusted_pr:
description: "🛡️ True when the PR carries the maintainer Trusted label; skips the fork-prereq wait."
required: false
type: boolean
default: false
image_wait_attempts:
description: "🔁 Polling attempts while waiting on prebuilt or mirrored images."
required: false
type: string
default: "1980"
image_wait_sleep_seconds:
description: "⏱️ Polling interval in seconds while waiting on images."
required: false
type: string
default: "10"
permissions:
contents: read
jobs:
validation:
name: ✅ Validation
runs-on: ubuntu-latest
steps:
- name: ⬇️ Checkout the calling ref for its tooling
uses: actions/checkout@v7.0.0
with:
path: .base-tooling
persist-credentials: false
- name: ⬇️ Checkout the code under test
if: ${{ github.event_name != 'pull_request_target' }}
uses: actions/checkout@v7.0.0
with:
ref: ${{ github.event.pull_request.head.sha || github.sha }}
fetch-depth: ${{ inputs.pr_scope != '' && '0' || '1' }}
path: pr-head
persist-credentials: false
- name: 🧭 Validate the branch prefix against the detected scope
if: ${{ inputs.pr_scope != '' }}
working-directory: .base-tooling
env:
PR_HEAD_REF: ${{ github.event.pull_request.head.ref }}
PR_SCOPE: ${{ inputs.pr_scope }}
run: ./scripts/meta/resolve/pr/branch_prefix.sh
- name: 🔀 Assert the base branch is fully merged into the PR head
if: ${{ github.event_name == 'pull_request' && inputs.pr_scope != '' }}
working-directory: pr-head
env:
INFINITO_MAIN_REMOTE: origin
INFINITO_MAIN_BRANCH: ${{ github.event.pull_request.base.ref }}
run: bash "${GITHUB_WORKSPACE}/.base-tooling/scripts/git/assert/main_merged.sh"
- name: 🔧 Set up Python
if: ${{ inputs.whitelist != '' || inputs.priority != '' }}
uses: actions/setup-python@v6
with:
python-version: "3.12"
# Exception: no pip cache. The steps below run the PR head's own code,
# which would write a cache that a default-branch run later restores.
- name: 📦 Install project (host venv)
if: ${{ inputs.whitelist != '' || inputs.priority != '' }}
working-directory: pr-head
shell: bash
run: make install
- name: 🎯 Validate the selection against this branch
if: ${{ inputs.whitelist != '' || inputs.priority != '' }}
working-directory: pr-head
shell: bash
env:
INPUT_WHITELIST: ${{ inputs.whitelist }}
INPUT_PRIORITY: ${{ inputs.priority }}
INPUT_MODE: ${{ inputs.mode }}
INPUT_TOR: ${{ inputs.tor }}
INPUT_DISTROS: ${{ inputs.distros }}
INPUT_FILESYSTEM: ${{ inputs.filesystem }}
INPUT_LIFECYCLES: ${{ inputs.lifecycles }}
run: ./scripts/github/validate/selection.sh
wait-fork-prereq-run:
name: ⏳ Wait for fork run
if: ${{ inputs.fork_prereqs }}
runs-on: ubuntu-latest
env:
WAIT_REQUIRED: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.repo.full_name != github.repository && !inputs.trusted_pr }}
steps:
- name: 🏷️ Mark fork prerequisites as ready outside fork PRs or for trusted PRs
if: ${{ env.WAIT_REQUIRED != 'true' }}
shell: bash
run: |
set -euo pipefail
echo "Fork prerequisites are implicitly ready (non-fork PR or '🛡️ Trusted' label)."
- name: ⬇️ Checkout trusted workflow repo (helpers)
if: ${{ env.WAIT_REQUIRED == 'true' }}
uses: actions/checkout@v7.0.0
with:
ref: ${{ github.event.pull_request.base.sha }}
persist-credentials: false
- name: 🔑 Login to GHCR for CI image checks
if: ${{ env.WAIT_REQUIRED == 'true' }}
uses: docker/login-action@af1e73f918a031802d376d3c8bbc3fe56130a9b0 # v4.4.0
with:
registry: ghcr.io
username: ${{ github.actor }}
password: ${{ secrets.GITHUB_TOKEN }}
- name: ⏳ Wait for privileged fork CI images
if: ${{ env.WAIT_REQUIRED == 'true' }}
shell: bash
env:
GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
WORKFLOW_FILE: entry-pr-change-orchestrate.yml
TARGET_EVENT: pull_request_target
PR_NUMBER: ${{ github.event.pull_request.number }}
PR_HEAD_SHA: ${{ github.event.pull_request.head.sha }}
PR_UPDATED_AT: ${{ github.event.pull_request.updated_at }}
INFINITO_DISTROS: "arch debian ubuntu fedora centos"
IMAGE_TAG: ci-${{ github.sha }}
WAIT_ATTEMPTS: ${{ inputs.image_wait_attempts }}
WAIT_SLEEP_SECONDS: ${{ inputs.image_wait_sleep_seconds }}
run: ./scripts/meta/wait/images.sh
fork-prereqs-ready:
name: 🚦 Gate Dependencies
needs: wait-fork-prereq-run
if: ${{ !cancelled() && inputs.fork_prereqs }}
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v7.0.0
- name: ✅ Assert privileged prerequisites are ready
shell: bash
env:
RESULT: ${{ needs.wait-fork-prereq-run.result }}
run: ./scripts/github/gate/fork_prereqs.sh