playwright-postgresql-e2e.yml

#  Copyright 2021 Collate
#  Licensed under the Apache License, Version 2.0 (the "License");
#  you may not use this file except in compliance with the License.
#  You may obtain a copy of the License at
#  http://www.apache.org/licenses/LICENSE-2.0
#  Unless required by applicable law or agreed to in writing, software
#  distributed under the License is distributed on an "AS IS" BASIS,
#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#  See the License for the specific language governing permissions and
#  limitations under the License.

# This workflow executes end-to-end (e2e) tests using Playwright with PostgreSQL as the database.
# For more information see: https://github.com/actions/setup-java/blob/main/docs/advanced-usage.md#apache-maven-with-a-settings-path

name: Postgresql PR Playwright E2E Tests
on:
  merge_group:
  workflow_dispatch:
  # Note: paths-ignore removed — the workflow always triggers so that the
  # required status check (playwright-summary) always completes. Path filtering
  # is handled inside the workflow via dorny/paths-filter so PRs without
  # relevant changes skip the expensive jobs while still reporting a passing status.
  pull_request_target:
    types:
      - labeled
      - opened
      - synchronize
      - reopened
      - ready_for_review

permissions:
  contents: read

concurrency:
  group: playwright-ci-pr-postgresql-${{ github.event.pull_request.number || github.run_id }}
  cancel-in-progress: ${{ github.event_name != 'pull_request_target' || github.event.action != 'labeled' || github.event.label.name == 'safe to test' }}

jobs:
  check-changes:
    runs-on: ubuntu-latest
    outputs:
      e2e: ${{ steps.filter.outputs.e2e }}
      docker-compose: ${{ steps.filter.outputs.docker-compose }}
    steps:
      - uses: dorny/paths-filter@v4
        id: filter
        with:
          filters: |
            e2e:
              - 'openmetadata-service/**'
              - 'openmetadata-ui/**'
              - 'openmetadata-spec/**'
              - 'openmetadata-integration-tests/**'
              - 'ingestion/**'
              - 'bootstrap/**'
              - 'conf/**'
              - 'docker/development/**'
              - 'openmetadata-clients/**'
              - 'openmetadata-airflow-apis/**'
              - 'openmetadata-mcp/**'
              - 'openmetadata-sdk/**'
              - 'openmetadata-shaded-deps/**'
              - 'openmetadata-ui-core-components/**'
              - 'openmetadata-k8s-operator/**'
              - 'common/**'
              - 'openspec/**'
              - 'pom.xml'
              - 'Makefile'
            docker-compose:
              - 'docker/development/docker-compose.yml'
              - 'docker/development/docker-compose-postgres.yml'

  build:
    needs: check-changes
    runs-on: ubuntu-latest
    if: |
      !github.event.pull_request.draft &&
      (github.event_name == 'merge_group' || github.event_name == 'workflow_dispatch' ||
       needs.check-changes.outputs.e2e == 'true' || needs.check-changes.outputs.docker-compose == 'true') &&
      (github.event_name != 'pull_request_target' || github.event.action != 'labeled' || github.event.label.name == 'safe to test')
    steps:
      - name: Wait for the labeler
        uses: lewagon/wait-on-check-action@v1.7.0
        if: ${{ github.event_name == 'pull_request_target' }}
        with:
          ref: ${{ github.event.pull_request.head.sha }}
          check-name: Team Label
          repo-token: ${{ secrets.GITHUB_TOKEN }}
          wait-interval: 90

      - name: Verify PR labels
        uses: jesusvasquez333/verify-pr-label-action@v1.4.0
        if: ${{ github.event_name == 'pull_request_target' }}
        with:
          github-token: "${{ secrets.GITHUB_TOKEN }}"
          valid-labels: "safe to test"
          pull-request-number: "${{ github.event.pull_request.number }}"
          disable-reviews: true

      - name: Checkout
        uses: actions/checkout@v4
        with:
          ref: ${{ github.event_name == 'merge_group' && github.sha || github.event.pull_request.head.sha }}

      - name: Setup JDK 21
        uses: actions/setup-java@v4
        with:
          java-version: '21'
          distribution: 'temurin'

      - name: Cache Maven Dependencies
        uses: actions/cache@v4
        with:
          path: ~/.m2
          key: ${{ runner.os }}-maven-${{ hashFiles('**/pom.xml') }}
          restore-keys: |
            ${{ runner.os }}-maven-

      - name: Install antlr cli
        run: sudo make install_antlr_cli

      - name: Build with Maven
        run: mvn -DskipTests clean package

      - name: Upload Maven build artifact
        uses: actions/upload-artifact@v4
        with:
          name: openmetadata-build
          path: openmetadata-dist/target/openmetadata-*.tar.gz
          retention-days: 1

  detect-changes:
    needs: check-changes
    runs-on: ubuntu-latest
    if: |
      !github.event.pull_request.draft &&
      (github.event_name == 'merge_group' || github.event_name == 'workflow_dispatch' ||
       needs.check-changes.outputs.e2e == 'true' || needs.check-changes.outputs.docker-compose == 'true') &&
      (github.event_name != 'pull_request_target' || github.event.action != 'labeled' || github.event.label.name == 'safe to test')
    outputs:
      spec_only_mode: ${{ steps.compute.outputs.spec_only_mode }}
      changed_specs: ${{ steps.compute.outputs.changed_specs }}
      matrix: ${{ steps.compute.outputs.matrix }}
    steps:
      - name: Checkout
        if: ${{ github.event_name == 'pull_request_target' }}
        uses: actions/checkout@v4
        with:
          ref: ${{ github.event.pull_request.head.sha }}
          fetch-depth: 0
          filter: blob:none

      - name: Get all changed files
        id: all-changes
        if: ${{ github.event_name == 'pull_request_target' }}
        uses: tj-actions/changed-files@9426d40962ed5378910ee2e21d5f8c6fcbf2dd96 # v47.0.6

      - name: Get changed runnable spec files
        id: changed-specs
        if: ${{ github.event_name == 'pull_request_target' }}
        uses: tj-actions/changed-files@9426d40962ed5378910ee2e21d5f8c6fcbf2dd96 # v47.0.6
        with:
          path: openmetadata-ui/src/main/resources/ui
          files: |
            playwright/e2e/**/*.spec.ts
          files_ignore: |
            playwright/e2e/**/SystemCertificationTags.spec.ts
            playwright/e2e/**/IntakeForm.spec.ts

      - name: Compute spec-only mode and matrix
        id: compute
        env:
          ALL_FILES: ${{ steps.all-changes.outputs.all_changed_files }}
          SPEC_FILES: ${{ steps.changed-specs.outputs.all_changed_files }}
        run: |
          FULL_MATRIX='{"shardIndex":[1,2,3,4,5,6],"shardTotal":[6]}'
          SPEC_MATRIX='{"shardIndex":[1],"shardTotal":[1]}'

          ALL_COUNT=$(echo "$ALL_FILES" | wc -w)
          SPEC_COUNT=$(echo "$SPEC_FILES" | wc -w)

          if [ "$SPEC_COUNT" -gt 0 ] && [ "$ALL_COUNT" -eq "$SPEC_COUNT" ]; then
            echo "spec_only_mode=true" >> "$GITHUB_OUTPUT"
            echo "changed_specs=$SPEC_FILES" >> "$GITHUB_OUTPUT"
            echo "matrix=$SPEC_MATRIX" >> "$GITHUB_OUTPUT"
            echo "Spec-only mode: ${SPEC_COUNT} spec(s) changed"
          else
            echo "spec_only_mode=false" >> "$GITHUB_OUTPUT"
            echo "changed_specs=" >> "$GITHUB_OUTPUT"
            echo "matrix=$FULL_MATRIX" >> "$GITHUB_OUTPUT"
            echo "Full suite mode (total=${ALL_COUNT}, runnable-specs=${SPEC_COUNT})"
          fi

  playwright-ci-postgresql:
    needs: [build, detect-changes]
    runs-on: ubuntu-latest
    if: ${{ !cancelled() && needs.build.result == 'success' && needs.detect-changes.result == 'success' }}
    environment: test
    permissions:
      contents: read
      id-token: write # Required for GitHub OIDC (Flakiness.io reporter)
    env:
      # Playwright logs the admin user in many times (performAdminLogin per test, parallel
      # workers, retries). The production default of 5 active sessions per user would evict the
      # long-lived storageState session the page fixtures rely on and 401 every request. Raise
      # the cap for E2E only; docker compose reads this for ${AUTHENTICATION_MAX_ACTIVE_SESSIONS_PER_USER:-5}.
      AUTHENTICATION_MAX_ACTIVE_SESSIONS_PER_USER: "10000"
    strategy:
      fail-fast: false
      matrix: ${{ fromJSON(needs.detect-changes.outputs.matrix) }}
    steps:
      - name: Free Disk Space (Ubuntu)
        uses: jlumbroso/free-disk-space@main
        with:
          tool-cache: false
          android: true
          dotnet: true
          haskell: true
          large-packages: false
          swap-storage: true
          docker-images: false

      - name: Checkout
        uses: actions/checkout@v4
        with:
          ref: ${{ github.event_name == 'merge_group' && github.sha || github.event.pull_request.head.sha }}

      - name: Download Maven build artifact
        uses: actions/download-artifact@v4
        with:
          name: openmetadata-build
          path: openmetadata-dist/target

      - name: Setup Openmetadata Test Environment
        uses: ./.github/actions/setup-openmetadata-test-environment
        with:
          python-version: "3.10"
          args: "-d postgresql -s true"
          ingestion_dependency: "all"

      - name: Setup Node.js
        uses: actions/setup-node@v4
        with:
          node-version-file: "openmetadata-ui/src/main/resources/ui/.nvmrc"

      - name: Install dependencies
        working-directory: openmetadata-ui/src/main/resources/ui/
        run: yarn --ignore-scripts --frozen-lockfile
      - name: Install Playwright Browsers
        run: npx playwright@1.57.0 install chromium --with-deps
      - name: Run Playwright tests
        id: run-tests
        working-directory: openmetadata-ui/src/main/resources/ui/
        run: |
          if [ "$SPEC_ONLY" = "true" ]; then
            echo "🔹 Spec-only mode: running changed specs → ${CHANGED_SPECS}"

            PROJECT_ARGS=""

            # DataAssetRulesEnabled runs standalone; its deps (setup) are resolved automatically
            if echo "$CHANGED_SPECS" | tr ' ' '\n' | grep -q "DataAssetRulesEnabled\.spec\.ts"; then
              PROJECT_ARGS="$PROJECT_ARGS --project=DataAssetRulesEnabled"
            fi
            # DataAssetRulesDisabled depends on DataAssetRulesEnabled — Playwright resolves this automatically
            if echo "$CHANGED_SPECS" | tr ' ' '\n' | grep -q "DataAssetRulesDisabled\.spec\.ts"; then
              PROJECT_ARGS="$PROJECT_ARGS --project=DataAssetRulesDisabled"
            fi
            # SearchRBAC depends on the full DataAssetRules chain — Playwright resolves this automatically
            if echo "$CHANGED_SPECS" | tr ' ' '\n' | grep -q "SearchRBAC\.spec\.ts"; then
              PROJECT_ARGS="$PROJECT_ARGS --project=SearchRBAC"
            fi
            # DomainIsolation uses serial workers and its own testMatch
            if echo "$CHANGED_SPECS" | tr ' ' '\n' | grep -q "DomainIsolation/"; then
              PROJECT_ARGS="$PROJECT_ARGS --project=DomainIsolation"
            fi
            # Regular chromium specs: anything not in a special-project category.
            # Also add Basic because chromium has grepInvert:[@basic], so @basic-tagged tests
            # in the changed files would be silently skipped without it.
            if echo "$CHANGED_SPECS" | tr ' ' '\n' | grep -qvE "(DataAssetRulesEnabled\.spec\.ts|DataAssetRulesDisabled\.spec\.ts|SearchRBAC\.spec\.ts|DomainIsolation/)"; then
              PROJECT_ARGS="$PROJECT_ARGS --project=chromium --project=Basic"
            fi

            [ -z "$PROJECT_ARGS" ] && PROJECT_ARGS="--project=chromium --project=Basic"

            npx playwright test $PROJECT_ARGS $CHANGED_SPECS

          elif [ "${{ matrix.shardIndex }}" -eq "1" ]; then
            echo "🔹 Running DataAssetRules-only tests on shard 1"

            # The testMatch pattern ensures only DataAssetRules*.spec.ts files run
            npx playwright test \
              --project=setup \
              --project=DataAssetRulesEnabled \
              --project=DataAssetRulesDisabled \
              --project=Basic \
              --project=SearchRBAC \
              --project=DomainIsolation \

          else
            # Shards 2-6 run common chromium tests equally distributed (5-way sharding)
            CHROMIUM_SHARD=$(( ${{ matrix.shardIndex }} - 1 ))
            echo "🔹 Running common chromium tests on shard ${CHROMIUM_SHARD}/5"
            npx playwright test \
              --project=chromium \
              --grep-invert @dataAssetRules \
              --shard=${CHROMIUM_SHARD}/5
          fi

        env:
          SPEC_ONLY: ${{ needs.detect-changes.outputs.spec_only_mode }}
          CHANGED_SPECS: ${{ needs.detect-changes.outputs.changed_specs }}
          PLAYWRIGHT_IS_OSS: true
          PLAYWRIGHT_SNOWFLAKE_USERNAME: ${{ secrets.TEST_SNOWFLAKE_USERNAME }}
          PLAYWRIGHT_SNOWFLAKE_PASSWORD: ${{ secrets.TEST_SNOWFLAKE_PASSWORD }}
          PLAYWRIGHT_SNOWFLAKE_ACCOUNT: ${{ secrets.TEST_SNOWFLAKE_ACCOUNT }}
          PLAYWRIGHT_SNOWFLAKE_DATABASE: ${{ secrets.TEST_SNOWFLAKE_DATABASE }}
          PLAYWRIGHT_SNOWFLAKE_WAREHOUSE: ${{ secrets.TEST_SNOWFLAKE_WAREHOUSE }}
          PLAYWRIGHT_SNOWFLAKE_PASSPHRASE: ${{ secrets.TEST_SNOWFLAKE_PASSPHRASE }}
          PLAYWRIGHT_PROJECT_ID: ${{ steps.cypress-project-id.outputs.CYPRESS_PROJECT_ID }}
          PLAYWRIGHT_BQ_PRIVATE_KEY: ${{ secrets.TEST_BQ_PRIVATE_KEY }}
          PLAYWRIGHT_BQ_PROJECT_ID: ${{ secrets.PLAYWRIGHT_BQ_PROJECT_ID }}
          PLAYWRIGHT_BQ_PRIVATE_KEY_ID: ${{ secrets.TEST_BQ_PRIVATE_KEY_ID }}
          PLAYWRIGHT_BQ_PROJECT_ID_TAXONOMY: ${{ secrets.TEST_BQ_PROJECT_ID_TAXONOMY }}
          PLAYWRIGHT_BQ_CLIENT_EMAIL: ${{ secrets.TEST_BQ_CLIENT_EMAIL }}
          PLAYWRIGHT_BQ_CLIENT_ID: ${{ secrets.TEST_BQ_CLIENT_ID }}
          PLAYWRIGHT_REDSHIFT_HOST: ${{ secrets.E2E_REDSHIFT_HOST_PORT }}
          PLAYWRIGHT_REDSHIFT_USERNAME: ${{ secrets.E2E_REDSHIFT_USERNAME }}
          PLAYWRIGHT_REDSHIFT_PASSWORD: ${{ secrets.E2E_REDSHIFT_PASSWORD }}
          PLAYWRIGHT_REDSHIFT_DATABASE: ${{ secrets.TEST_REDSHIFT_DATABASE }}
          PLAYWRIGHT_METABASE_USERNAME: ${{ secrets.TEST_METABASE_USERNAME }}
          PLAYWRIGHT_METABASE_PASSWORD: ${{ secrets.TEST_METABASE_PASSWORD }}
          PLAYWRIGHT_METABASE_DB_SERVICE_NAME: ${{ secrets.TEST_METABASE_DB_SERVICE_NAME }}
          PLAYWRIGHT_METABASE_HOST_PORT: ${{ secrets.TEST_METABASE_HOST_PORT }}
          PLAYWRIGHT_SUPERSET_USERNAME: ${{ secrets.TEST_SUPERSET_USERNAME }}
          PLAYWRIGHT_SUPERSET_PASSWORD: ${{ secrets.TEST_SUPERSET_PASSWORD }}
          PLAYWRIGHT_SUPERSET_HOST_PORT: ${{ secrets.TEST_SUPERSET_HOST_PORT }}
          PLAYWRIGHT_KAFKA_BOOTSTRAP_SERVERS: ${{ secrets.TEST_KAFKA_BOOTSTRAP_SERVERS }}
          PLAYWRIGHT_KAFKA_SCHEMA_REGISTRY_URL: ${{ secrets.TEST_KAFKA_SCHEMA_REGISTRY_URL }}
          PLAYWRIGHT_GLUE_ACCESS_KEY: ${{ secrets.TEST_GLUE_ACCESS_KEY }}
          PLAYWRIGHT_GLUE_SECRET_KEY: ${{ secrets.TEST_GLUE_SECRET_KEY }}
          PLAYWRIGHT_GLUE_AWS_REGION: ${{ secrets.TEST_GLUE_AWS_REGION }}
          PLAYWRIGHT_GLUE_ENDPOINT: ${{ secrets.TEST_GLUE_ENDPOINT }}
          PLAYWRIGHT_GLUE_STORAGE_SERVICE: ${{ secrets.TEST_GLUE_STORAGE_SERVICE }}
          PLAYWRIGHT_MYSQL_USERNAME: ${{ secrets.TEST_MYSQL_USERNAME }}
          PLAYWRIGHT_MYSQL_PASSWORD: ${{ secrets.TEST_MYSQL_PASSWORD }}
          PLAYWRIGHT_MYSQL_HOST_PORT: ${{ secrets.TEST_MYSQL_HOST_PORT }}
          PLAYWRIGHT_MYSQL_DATABASE_SCHEMA: ${{ secrets.TEST_MYSQL_DATABASE_SCHEMA }}
          PLAYWRIGHT_POSTGRES_USERNAME: ${{ secrets.TEST_POSTGRES_USERNAME }}
          PLAYWRIGHT_POSTGRES_PASSWORD: ${{ secrets.TEST_POSTGRES_PASSWORD }}
          PLAYWRIGHT_POSTGRES_HOST_PORT: ${{ secrets.TEST_POSTGRES_HOST_PORT }}
          PLAYWRIGHT_POSTGRES_DATABASE: ${{ secrets.TEST_POSTGRES_DATABASE }}
          PLAYWRIGHT_AIRFLOW_HOST_PORT: ${{ secrets.TEST_AIRFLOW_HOST_PORT }}
          PLAYWRIGHT_ML_MODEL_TRACKING_URI: ${{ secrets.TEST_ML_MODEL_TRACKING_URI }}
          PLAYWRIGHT_ML_MODEL_REGISTRY_URI: ${{ secrets.TEST_ML_MODEL_REGISTRY_URI }}
          PLAYWRIGHT_S3_STORAGE_ACCESS_KEY_ID: ${{ secrets.TEST_S3_STORAGE_ACCESS_KEY_ID }}
          PLAYWRIGHT_S3_STORAGE_SECRET_ACCESS_KEY: ${{ secrets.TEST_S3_STORAGE_SECRET_ACCESS_KEY }}
          PLAYWRIGHT_S3_STORAGE_END_POINT_URL: ${{ secrets.TEST_S3_STORAGE_END_POINT_URL }}

          # Recommended: pass the GitHub token lets this action correctly
          # determine the unique run id necessary to re-run the checks
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

      - uses: actions/upload-artifact@v4
        if: ${{ !cancelled() }}
        with:
          name: playwright-report-${{ matrix.shardIndex }}
          path: openmetadata-ui/src/main/resources/ui/playwright/output/playwright-report
          retention-days: 5

      - name: Upload test results (screenshots, traces)
        uses: actions/upload-artifact@v4
        if: ${{ !cancelled() }}
        with:
          name: playwright-test-results-${{ matrix.shardIndex }}
          path: openmetadata-ui/src/main/resources/ui/playwright/output/test-results
          retention-days: 5

      - name: Upload results JSON for summary
        uses: actions/upload-artifact@v4
        if: ${{ !cancelled() }}
        with:
          name: playwright-results-json-${{ matrix.shardIndex }}
          path: openmetadata-ui/src/main/resources/ui/playwright/output/results.json
          retention-days: 1
          if-no-files-found: ignore

      - name: Clean Up
        run: |
          cd ./docker/development
          docker compose down --remove-orphans
          sudo rm -rf ${PWD}/docker-volume

  playwright-summary:
    if: always()
    needs: playwright-ci-postgresql
    runs-on: ubuntu-latest
    permissions:
      contents: read
      pull-requests: write # Posts the consolidated PR comment
    steps:
      - name: Download all results JSON
        if: needs.playwright-ci-postgresql.result != 'skipped'
        uses: actions/download-artifact@v4
        with:
          pattern: playwright-results-json-*
          path: results

      - name: Post consolidated PR comment and gate on results
        uses: actions/github-script@v7
        with:
          github-token: ${{ secrets.GITHUB_TOKEN }}
          script: |
            const upstreamResult = '${{ needs.playwright-ci-postgresql.result }}';
            if (upstreamResult === 'skipped') {
              // Distinguish the two reasons build/playwright can be skipped:
              // 1. A non-"safe to test" label was added — build skips intentionally for fork security.
              //    Tests will run once "safe to test" is applied or code is pushed.
              // 2. No relevant paths changed — nothing to test.
              const eventAction = context.payload.action;
              const labelName   = context.payload.label?.name ?? '';
              if (eventAction === 'labeled' && labelName !== 'safe to test') {
                console.log(`Playwright E2E tests pending — label "${labelName}" added but only "safe to test" triggers E2E. Tests will run when "safe to test" is applied or code is pushed.`);
              } else {
                console.log('Playwright E2E tests not required for this PR (no relevant paths changed).');
              }
              return;
            }
            if (upstreamResult === 'cancelled') {
              core.setFailed('Playwright E2E tests were cancelled.');
              return;
            }

            const fs = require('fs');
            const path = require('path');

            const runId = '${{ github.run_id }}';
            const repo = context.repo;
            const prNumber = context.payload.pull_request?.number;

            const artifactUrl = `https://github.com/${repo.owner}/${repo.repo}/actions/runs/${runId}`;
            const commentMarker = '<!-- playwright-summary -->';

            // Collect results from all shards
            const shardResults = [];
            const resultsDir = 'results';
            if (fs.existsSync(resultsDir)) {
              for (const dir of fs.readdirSync(resultsDir).sort()) {
                const jsonPath = path.join(resultsDir, dir, 'results.json');
                if (!fs.existsSync(jsonPath)) continue;

                const shardNum = dir.replace('playwright-results-json-', '');
                const report = JSON.parse(fs.readFileSync(jsonPath, 'utf8'));

                const allTests = [];
                function collectTests(suite, filePath) {
                  const file = suite.file || filePath || '';
                  for (const spec of (suite.specs || [])) {
                    for (const test of (spec.tests || [])) {
                      const results = test.results || [];
                      const lastResult = results[results.length - 1] || {};
                      const firstResult = results[0] || {};
                      allTests.push({
                        title: spec.title,
                        file: file,
                        status: test.status,
                        retries: results.length - 1,
                        error: lastResult.error?.message || firstResult.error?.message || '',
                      });
                    }
                  }
                  for (const child of (suite.suites || [])) {
                    collectTests(child, file);
                  }
                }
                for (const suite of (report.suites || [])) {
                  collectTests(suite, '');
                }

                shardResults.push({
                  shard: shardNum,
                  genuine: allTests.filter(t => t.status === 'unexpected'),
                  flaky: allTests.filter(t => t.status === 'flaky'),
                  passed: allTests.filter(t => t.status === 'expected'),
                  skipped: allTests.filter(t => t.status === 'skipped'),
                });
              }
            }

            if (shardResults.length === 0) {
              core.setFailed('No Playwright results found — shards may have been cancelled or failed to upload artifacts.');
              return;
            }

            // Aggregate totals
            const totalPassed = shardResults.reduce((s, r) => s + r.passed.length, 0);
            const totalFailed = shardResults.reduce((s, r) => s + r.genuine.length, 0);
            const totalFlaky = shardResults.reduce((s, r) => s + r.flaky.length, 0);
            const totalSkipped = shardResults.reduce((s, r) => s + r.skipped.length, 0);

            const lines = [commentMarker];

            if (totalFailed > 0) {
              lines.push(`## 🔴 Playwright Results — ${totalFailed} failure(s)${totalFlaky > 0 ? `, ${totalFlaky} flaky` : ''}`);
            } else if (totalFlaky > 0) {
              lines.push(`## 🟡 Playwright Results — all passed (${totalFlaky} flaky)`);
            } else {
              lines.push(`## ✅ Playwright Results — all ${totalPassed} tests passed`);
            }
            lines.push('');
            lines.push(`✅ ${totalPassed} passed · ❌ ${totalFailed} failed · 🟡 ${totalFlaky} flaky · ⏭️ ${totalSkipped} skipped`);
            lines.push('');

            // Per-shard summary table
            lines.push('| Shard | Passed | Failed | Flaky | Skipped |');
            lines.push('|-------|--------|--------|-------|---------|');
            for (const r of shardResults) {
              const status = r.genuine.length > 0 ? '🔴' : r.flaky.length > 0 ? '🟡' : '✅';
              lines.push(`| ${status} Shard ${r.shard} | ${r.passed.length} | ${r.genuine.length} | ${r.flaky.length} | ${r.skipped.length} |`);
            }
            lines.push('');

            // Genuine failures detail
            const allGenuine = shardResults.flatMap(r => r.genuine.map(t => ({ ...t, shard: r.shard })));
            if (allGenuine.length > 0) {
              lines.push('### Genuine Failures (failed on all attempts)');
              lines.push('');
              for (const t of allGenuine.slice(0, 30)) {
                const shortFile = t.file.replace(/.*playwright\/e2e\//, '');
                lines.push(`<details><summary>❌ <code>${shortFile}</code> › ${t.title} (shard ${t.shard})</summary>`);
                lines.push('');
                lines.push('```');
                lines.push(t.error.substring(0, 1000));
                lines.push('```');
                lines.push('</details>');
                lines.push('');
              }
              if (allGenuine.length > 30) {
                lines.push(`... and ${allGenuine.length - 30} more failures`);
                lines.push('');
              }
            }

            // Flaky tests
            const allFlaky = shardResults.flatMap(r => r.flaky.map(t => ({ ...t, shard: r.shard })));
            if (allFlaky.length > 0) {
              lines.push(`<details><summary>🟡 ${allFlaky.length} flaky test(s) (passed on retry)</summary>`);
              lines.push('');
              for (const t of allFlaky.slice(0, 30)) {
                const shortFile = t.file.replace(/.*playwright\/e2e\//, '');
                lines.push(`- \`${shortFile}\` › ${t.title} (shard ${t.shard}, ${t.retries} ${t.retries === 1 ? 'retry' : 'retries'})`);
              }
              if (allFlaky.length > 30) {
                lines.push(`- ... and ${allFlaky.length - 30} more`);
              }
              lines.push('');
              lines.push('</details>');
              lines.push('');
            }

            lines.push(`📦 [Download artifacts](${artifactUrl})`);
            lines.push('');
            lines.push('<details><summary>How to debug locally</summary>');
            lines.push('');
            lines.push('```bash');
            lines.push('# Download playwright-test-results-<shard> artifact and unzip');
            lines.push('npx playwright show-trace path/to/trace.zip    # view trace');
            lines.push('```');
            lines.push('</details>');

            const body = lines.join('\n');

            // Post PR comment only for pull_request_target events
            if (prNumber) {
              let existingComment = null;
              for await (const response of github.paginate.iterator(
                github.rest.issues.listComments, { ...repo, issue_number: prNumber, per_page: 100 }
              )) {
                const found = response.data.find(c =>
                  c.user?.login === 'github-actions[bot]' && c.body?.includes(commentMarker)
                );
                if (found) { existingComment = found; break; }
              }

              // Also clean up any old per-shard comments from previous runs
              for await (const response of github.paginate.iterator(
                github.rest.issues.listComments, { ...repo, issue_number: prNumber, per_page: 100 }
              )) {
                for (const c of response.data) {
                  if (c.user?.login === 'github-actions[bot]' && /Playwright Shard \d/.test(c.body || '') && !c.body?.includes(commentMarker)) {
                    await github.rest.issues.deleteComment({ ...repo, comment_id: c.id });
                    console.log(`Deleted old per-shard comment ${c.id}`);
                  }
                }
              }

              if (existingComment) {
                await github.rest.issues.updateComment({ ...repo, comment_id: existingComment.id, body });
              } else {
                await github.rest.issues.createComment({ ...repo, issue_number: prNumber, body });
              }
            }

            // Gate: fail the step (and therefore the job) when genuine test failures exist.
            // This is the single source of truth — the same parsed results that drive the
            // PR comment also determine whether playwright-summary passes or fails.
            if (totalFailed > 0) {
              core.setFailed(`${totalFailed} Playwright test(s) failed.`);
            }