PaperClipAI/.github/workflows/release-smoke.yml

name: Release Smoke

on:
  workflow_dispatch:
    inputs:
      paperclip_version:
        description: Published Paperclip dist-tag to test
        required: true
        default: canary
        type: choice
        options:
          - canary
          - nightly
          - beta
          - latest
      host_port:
        description: Host port for the Docker smoke container
        required: false
        default: "3232"
        type: string
      artifact_name:
        description: Artifact name for uploaded diagnostics
        required: false
        default: release-smoke
        type: string
      qualification_source_sha:
        description: Optional immutable source SHA for provider-free native package qualification
        required: false
        default: ""
        type: string
  workflow_call:
    inputs:
      paperclip_version:
        required: true
        type: string
      host_port:
        required: false
        default: "3232"
        type: string
      artifact_name:
        required: false
        default: release-smoke
        type: string
      qualification_source_sha:
        required: false
        default: ""
        type: string

permissions:
  contents: read

jobs:
  # The Docker smoke below can never exercise the background-service leg of
  # onboarding: containers have no service manager, so v2026.824.0 shipped a
  # service install that crash-looped on a missing shim while every
  # golden-path check stayed green. Run the same published artifact directly
  # on the runner VM's systemd and require the installed service to end up
  # serving.
  smoke_service:
    if: inputs.qualification_source_sha == ''
    runs-on: ubuntu-latest
    timeout-minutes: 30

    steps:
      - name: Checkout repository
        uses: actions/checkout@v7

      - name: Setup Node.js
        uses: actions/setup-node@v7
        with:
          node-version: 24

      - name: Start a user systemd session
        # The hosted runner has no login session for the runner user, so
        # `systemctl --user` cannot reach a user manager until lingering
        # starts one. Export the session address for the steps below.
        run: |
          sudo loginctl enable-linger "$(id -un)"
          uid="$(id -u)"
          for _ in $(seq 1 30); do
            [[ -S "/run/user/$uid/bus" ]] && break
            sleep 1
          done
          [[ -S "/run/user/$uid/bus" ]]
          {
            echo "XDG_RUNTIME_DIR=/run/user/$uid"
            echo "DBUS_SESSION_BUS_ADDRESS=unix:path=/run/user/$uid/bus"
          } >> "$GITHUB_ENV"

      - name: Onboard with the background service
        env:
          PAPERCLIPAI_VERSION: ${{ inputs.paperclip_version }}
          DATA_DIR: ${{ runner.temp }}/service-smoke-data
          SMOKE_CLEANUP: "false"
        run: ./scripts/service-onboard-smoke.sh

      - name: Capture service diagnostics
        if: always()
        run: |
          {
            systemctl --user --no-pager status paperclipai.service || true
            journalctl --user -u paperclipai.service --no-pager || true
          } > "$RUNNER_TEMP/paperclipai-service.log" 2>&1

      - name: Upload service diagnostics
        if: always()
        uses: actions/upload-artifact@v7
        with:
          name: ${{ inputs.artifact_name }}-service
          path: ${{ runner.temp }}/paperclipai-service.log
          retention-days: 14

  smoke:
    if: inputs.qualification_source_sha == ''
    runs-on: ubuntu-latest
    timeout-minutes: 45

    # Fixed here rather than read back out of the harness, so the `always()`
    # diagnostics steps below still know the container's name when the launch
    # step is the thing that failed. Reading it back was why a failing smoke
    # uploaded no Docker logs at all.
    env:
      SMOKE_CONTAINER_NAME: release-smoke-onboard

    steps:
      - name: Checkout repository
        uses: actions/checkout@v7

      - name: Setup pnpm
        uses: pnpm/action-setup@v6
        with:
          version: 9.15.4

      - name: Setup Node.js
        uses: actions/setup-node@v7
        with:
          node-version: 24
          cache: pnpm

      - name: Install dependencies
        run: pnpm install --no-frozen-lockfile

      - name: Verify runner Chrome
        # Release smoke also runs headless on GitHub's Ubuntu image, so use the
        # runner's preinstalled Chrome instead of a Playwright browser download.
        run: google-chrome --version

      - name: Launch Docker smoke harness
        run: |
          HOST_PORT="${{ inputs.host_port }}" \
          DATA_DIR="$RUNNER_TEMP/release-smoke-data" \
          PAPERCLIPAI_VERSION="${{ inputs.paperclip_version }}" \
          SMOKE_READY_TIMEOUT_SECONDS=420 \
          SMOKE_DETACH=true \
          SMOKE_METADATA_FILE="${{ runner.temp }}/release-smoke.env" \
          SMOKE_LOG_FILE="${{ runner.temp }}/docker-onboard-smoke.log" \
          ./scripts/docker-onboard-smoke.sh
          set -a
          source "${{ runner.temp }}/release-smoke.env"
          set +a
          {
            echo "SMOKE_BASE_URL=$SMOKE_BASE_URL"
            echo "SMOKE_ADMIN_EMAIL=$SMOKE_ADMIN_EMAIL"
            echo "SMOKE_ADMIN_PASSWORD=$SMOKE_ADMIN_PASSWORD"
            echo "SMOKE_DATA_DIR=$SMOKE_DATA_DIR"
            echo "SMOKE_IMAGE_NAME=$SMOKE_IMAGE_NAME"
            echo "SMOKE_PAPERCLIPAI_VERSION=$SMOKE_PAPERCLIPAI_VERSION"
          } >> "$GITHUB_ENV"

      - name: Run release smoke Playwright suite
        env:
          PAPERCLIP_RELEASE_SMOKE_BASE_URL: ${{ env.SMOKE_BASE_URL }}
          PAPERCLIP_RELEASE_SMOKE_EMAIL: ${{ env.SMOKE_ADMIN_EMAIL }}
          PAPERCLIP_RELEASE_SMOKE_PASSWORD: ${{ env.SMOKE_ADMIN_PASSWORD }}
          PAPERCLIP_PLAYWRIGHT_CHANNEL: "chrome"
        run: pnpm run test:release-smoke

      - name: Capture Docker logs
        if: always()
        run: |
          log_file="${{ runner.temp }}/docker-onboard-smoke.log"
          # A live container has the fuller story, so prefer it. When the
          # harness already tore the container down it wrote this file on its
          # way out, and that copy is kept rather than clobbered.
          if docker inspect "$SMOKE_CONTAINER_NAME" >/dev/null 2>&1; then
            docker logs "$SMOKE_CONTAINER_NAME" >"$log_file" 2>&1 || true
          fi
          # Never leave the upload with nothing to say. An absent log reads as
          # a missing artifact; a file saying the container was gone reads as
          # the diagnosis it is.
          if [[ ! -s "$log_file" ]]; then
            echo "No Docker logs captured: container '$SMOKE_CONTAINER_NAME' left no log dump and is no longer present." >"$log_file"
          fi
          echo "Captured $(wc -l <"$log_file") log lines to $log_file"

      - name: Upload diagnostics
        if: always()
        uses: actions/upload-artifact@v7
        with:
          name: ${{ inputs.artifact_name }}
          path: |
            ${{ runner.temp }}/docker-onboard-smoke.log
            ${{ runner.temp }}/release-smoke.env
            tests/release-smoke/playwright-report/
            tests/release-smoke/test-results/
          # The capture step above guarantees the log file, so an empty upload
          # means the diagnostics wiring itself broke — which is worth failing
          # over rather than burying in a warning nobody reads.
          if-no-files-found: error
          retention-days: 14

      - name: Stop Docker smoke container
        if: always()
        run: docker rm -f "$SMOKE_CONTAINER_NAME" >/dev/null 2>&1 || true

  qualification_daemons:
    if: inputs.qualification_source_sha != ''
    uses: ./.github/workflows/release-verify.yml
    with:
      ref: ${{ inputs.qualification_source_sha }}
      runner_binaries_only: true

  qualification_pack:
    if: inputs.qualification_source_sha != ''
    needs: qualification_daemons
    name: Assemble and verify public Linux packages
    runs-on: ubuntu-latest
    timeout-minutes: 45
    steps:
      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
        with:
          ref: ${{ inputs.qualification_source_sha }}
          persist-credentials: false
      - uses: pnpm/action-setup@0977fd99725f1db4007ccb2928dbb4e90d06cc86 # v6
        with:
          version: 9.15.4
      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
        with:
          node-version: 24
      - uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4
        with:
          pattern: ${{ needs.qualification_daemons.outputs.runner_artifact_prefix }}-*
          path: ${{ runner.temp }}/release-runner-inputs
          merge-multiple: true
      - name: Bind the immutable Linux producer source
        env:
          SOURCE_SHA: ${{ inputs.qualification_source_sha }}
        run: |
          set -euo pipefail
          [[ "$SOURCE_SHA" =~ ^[a-f0-9]{40}$ ]] || exit 1
          [[ "$(git rev-parse HEAD)" == "$SOURCE_SHA" ]] || exit 1
      - name: Build normal workspace outputs and public CLI
        run: |
          pnpm install --no-frozen-lockfile
          pnpm build
          bash scripts/build-npm.sh --skip-checks --skip-typecheck
          mv cli/package.dev.json cli/package.json
      - name: Assemble exact-source daemons before server packing
        env:
          SOURCE_SHA: ${{ inputs.qualification_source_sha }}
        run: |
          node packages/paperclip-runner/scripts/stage-release-runner-binaries.mjs \
            assemble-daemons "$RUNNER_TEMP/release-runner-inputs" "$SOURCE_SHA"
          cp -R packages/paperclip-runner/dist/bin/. server/dist/vendor/paperclip-runner/bin/
      - name: Verify clean public installation and retain the same package bytes
        env:
          PAPERCLIP_RUNNER_PUBLIC_PACKS_DIR: ${{ runner.temp }}/public-runner-packs
        run: node scripts/verify-grok-npm-install.mjs
      - name: Locate completed package assembly
        id: pack_identity
        if: ${{ !cancelled() }}
        run: |
          if [[ -s "$RUNNER_TEMP/public-runner-packs/manifest.json" ]]; then
            echo "available=true" >> "$GITHUB_OUTPUT"
          fi
      - name: Retain completed package assembly even after a later Linux failure
        if: ${{ !cancelled() && steps.pack_identity.outputs.available == 'true' }}
        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4
        with:
          name: ${{ inputs.artifact_name }}-public-packages
          path: ${{ runner.temp }}/public-runner-packs/
          if-no-files-found: error
          retention-days: 14

  qualification_native:
    if: ${{ !cancelled() && inputs.qualification_source_sha != '' && needs.qualification_daemons.result == 'success' }}
    needs: [qualification_daemons, qualification_pack]
    name: Clean installed package (${{ matrix.target }})
    runs-on: ${{ matrix.host }}
    timeout-minutes: 30
    strategy:
      fail-fast: false
      matrix: ${{ fromJSON(needs.qualification_daemons.outputs.native_consumer_matrix) }}
    steps:
      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
        with:
          ref: ${{ inputs.qualification_source_sha }}
          persist-credentials: false
      - uses: pnpm/action-setup@0977fd99725f1db4007ccb2928dbb4e90d06cc86 # v6
        with:
          version: 9.15.4
      - uses: actions/setup-node@820762786026740c76f36085b0efc47a31fe5020 # v7
        with:
          node-version: 24
      - uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4
        with:
          name: ${{ inputs.artifact_name }}-public-packages
          path: ${{ runner.temp }}/public-runner-packs
      - name: Install only the existing verifier tooling
        run: pnpm install --ignore-scripts --no-frozen-lockfile
      - name: Verify host identity and actual clean npm consumer
        env:
          SOURCE_SHA: ${{ inputs.qualification_source_sha }}
          TARGET: ${{ matrix.target }}
        run: |
          set -euo pipefail
          [[ "$(git rev-parse HEAD)" == "$SOURCE_SHA" ]] || exit 1
          [[ "$(node -p 'process.platform + "-" + process.arch')" == "$TARGET" ]] || exit 1
          mode=--mac-consumer
          if [[ "$TARGET" == linux-arm64 ]]; then mode=--linux-arm64-consumer; fi
          node scripts/verify-grok-npm-install.mjs "$mode" \
            "$RUNNER_TEMP/public-runner-packs" "$SOURCE_SHA" "$RUNNER_TEMP/native-public-install.json"
      - uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4
        if: always()
        with:
          name: ${{ inputs.artifact_name }}-${{ matrix.target }}
          path: ${{ runner.temp }}/native-public-install.json
          if-no-files-found: warn
          retention-days: 14