open-swe/.github/workflows/ci.yml

name: CI

permissions:
  contents: read

on:
  push:
    # dev as well as main so every dev HEAD carries the full CI signal that
    # the dev->main promotion gate (check-dev-green.sh) reads. PR checks alone are
    # not enough: an admin-merge can land a red PR onto dev.
    branches: ["main", "dev"]
  pull_request:
  workflow_dispatch:

concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: true

jobs:
  lint:
    name: Lint
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v7
      - uses: astral-sh/setup-uv@20cfd1bf945f4377ade1205e4dbc17946fc9a30d # v10.0.1
      - name: Install dependencies
        run: uv sync --locked --extra dev
      - name: Run lint
        run: make lint

  format:
    name: Format check
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v7
      - uses: astral-sh/setup-uv@20cfd1bf945f4377ade1205e4dbc17946fc9a30d # v10.0.1
      - name: Install dependencies
        run: uv sync --locked --extra dev
      - name: Run format check
        run: make format-check

  typecheck:
    name: Typecheck
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v7
      - uses: oven-sh/setup-bun@0c5077e51419868618aeaa5fe8019c62421857d6 # v2
        with:
          bun-version: "1.3.14"
      - name: Install UI dependencies
        working-directory: ui
        run: bun install --frozen-lockfile
      - name: Run Typecheck
        working-directory: ui
        run: bun run typecheck

  unit-tests:
    name: Unit tests
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v7
      - uses: astral-sh/setup-uv@20cfd1bf945f4377ade1205e4dbc17946fc9a30d # v10.0.1
      - name: Install dependencies
        run: uv sync --locked --extra dev
      - name: Run unit tests
        run: make test

  e2e:
    name: Playwright E2E
    runs-on: ubuntu-latest
    timeout-minutes: 30
    steps:
      - uses: actions/checkout@v7
      - uses: astral-sh/setup-uv@20cfd1bf945f4377ade1205e4dbc17946fc9a30d # v10.0.1
      - uses: actions/setup-node@v7
        with:
          node-version: "24"
      - uses: oven-sh/setup-bun@0c5077e51419868618aeaa5fe8019c62421857d6 # v2
        with:
          bun-version: "1.3.14"
      - name: Install Python deps (langgraph dev runtime)
        run: uv sync --locked
      - name: Install Playwright + Chromium
        working-directory: tests/e2e
        run: |
          npm ci
          sudo rm -f /etc/apt/sources.list.d/azure-cli.list /etc/apt/sources.list.d/microsoft-prod.list
          npx playwright install --with-deps chromium
      # Playwright's webServer boots `langgraph dev`; globalSetup builds the real
      # ui/ SPA. The fake LLM/GitHub/Slack boundaries need no secrets.
      - name: Run E2E
        working-directory: tests/e2e
        # Playwright's globalSetup runs the real `bun run build`, whose vite bundle
        # exceeds Node's default ~2 GB heap.
        # The runner has ~16 GB, so lift the heap cap.
        env:
          NODE_OPTIONS: "--max-old-space-size=8192"
        run: npx playwright test
      - name: Upload Playwright report
        if: ${{ !cancelled() }}
        uses: actions/upload-artifact@v7
        with:
          name: playwright-report
          path: |
            tests/e2e/playwright-report
            tests/e2e/test-results
          retention-days: 7

  docker-build:
    name: Docker build smoke
    runs-on: ubuntu-latest
    timeout-minutes: 20
    steps:
      - uses: actions/checkout@v7
      # Smoke-build the sandbox image so a bad Dockerfile pin (e.g. an
      # apt "nodejs=${NODEJS_VERSION}" that no longer resolves) fails a check
      # instead of only surfacing at sandbox-provision time. No push, no
      # registry credentials: this only proves the image builds.
      - name: Build sandbox image (no push)
        run: docker build --pull -t open-swe-sandbox:ci .

  triage-ledger:
    name: Triage ledger up to date
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v7
      # docs/upstream-sync/triage.md is GENERATED from triage.jsonl. Fail if a
      # ledger edit forgot to regenerate it (`make triage-render`). Stdlib-only
      # Python, so no uv sync / deps needed.
      - name: triage.md is up to date with triage.jsonl
        run: make triage-check

  ui-lockfile:
    name: ui bun.lock in sync
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v7
      - uses: oven-sh/setup-bun@0c5077e51419868618aeaa5fe8019c62421857d6 # v2
        with:
          bun-version: "1.3.14"
      - name: ui/package.json and ui/bun.lock agree
        working-directory: ui
        run: bun install --frozen-lockfile