Files
pulse-triage[bot] f1bdc44ea5 test: isolate shell integration runs and scope cleanup
Shared image tags, container names and host ports let concurrent checkout qualification collide or inspect another runtime. Give each shell integration invocation its own identities and allocated listeners, and scope failure cleanup to those resources so results cannot silently qualify unrelated bytes.

Contract-Neutral: Integration harness isolation and documentation only; no user-visible frontend, product contract or release surface changes.
Change-source: pulse-maintainer
2026-09-05 11:36:38 +01:00

5.7 KiB

Quick Start Guide - Update Integration Tests

This guide will help you get the update integration tests running quickly.

Prerequisites

  • Docker and Docker Compose
  • Node.js 24 and npm
  • Go 1.25+ (for building mock server)

Setup (One-time)

cd tests/integration
./scripts/setup.sh
```text

This will:
- Install npm dependencies
- Install Playwright browsers
- Build Docker images for test environment

## Running Tests

### Run All Tests

```bash
npm test

This runs all test suites with appropriate configurations.

Run Specific Test Suite

# Happy path only
./scripts/run-tests.sh happy

# Bad checksums
./scripts/run-tests.sh checksums

# Rate limiting
./scripts/run-tests.sh rate-limit

# Network failures
./scripts/run-tests.sh network

# Stale releases
./scripts/run-tests.sh stale

# Frontend validation
./scripts/run-tests.sh frontend

Interactive Mode

# Open Playwright UI
npm run test:ui

# Debug mode
npm run test:debug

# Run in headed browser
npm run test:headed

Manual Docker Control

# Start test environment
npm run docker:up

# View logs
npm run docker:logs

# Stop environment
npm run docker:down

# Rebuild images
npm run docker:rebuild

Accessing Test Services

While the test environment is running:

For the repo-local managed dev runtime, use npm run dev from the repo root and browse http://127.0.0.1:5173 instead. :7655 in this quick start is the backend-served test UI inside the docker environment, not the canonical hot-dev browser shell.

Viewing Test Results

After running tests:

# View HTML report
npm run test:report

# Reports are saved to:
# - playwright-report/ (HTML report)
# - test-results/ (screenshots, videos)

Test Scenarios

1. Diagnostic Smoke Test (00-diagnostic.spec.ts)

  • Ensures the containerized stack boots and the UI renders.

2. Core E2E Flows (01-core-e2e.spec.ts)

  • First-run setup wizard (fresh instance)
  • Login/logout + authenticated state
  • Alerts thresholds create/delete
  • Settings persistence across refresh
  • Add/delete a Proxmox node (test-only)

Troubleshooting

Tests failing to start

# Check Docker is running
docker ps

# Rebuild images
npm run docker:rebuild

# Check logs
npm run docker:logs

Port conflicts

If ports 7655 or 8080 are in use:

# Find and stop conflicting processes
lsof -i :7655
lsof -i :8080

Clean slate

# Remove all test containers and volumes
docker-compose -f docker-compose.test.yml down -v

# Clean Docker
docker system prune -f

# Reinstall
./scripts/setup.sh

CI Integration

The full Playwright suite, including the update-flow spec (tests/79-update-flow.spec.ts), runs via .github/workflows/test-e2e.yml on every push and PR that touches:

  • internal/**
  • frontend-modern/**
  • tests/integration/**
  • Dockerfile

Success Criteria

All test scenarios pass reliably Tests catch checksum validation issues (like v4.28.0) Frontend UX regressions are blocked Tests run in CI on every relevant PR

Architecture

┌─────────────────┐     ┌──────────────────┐     ┌─────────────────────┐
│  Playwright     │────▶│  Pulse Server    │────▶│  Mock GitHub API    │
│  (Browser UI)   │     │  (Test Instance) │     │  (Controlled        │
│                 │     │                  │     │   Responses)        │
└─────────────────┘     └──────────────────┘     └─────────────────────┘

The mock GitHub server provides controllable responses for testing different scenarios via environment variables:

  • MOCK_CHECKSUM_ERROR=true - Return invalid checksums
  • MOCK_NETWORK_ERROR=true - Simulate network failures
  • MOCK_RATE_LIMIT=true - Enable aggressive rate limiting
  • MOCK_STALE_RELEASE=true - Mark releases as stale

Writing New Tests

  1. Add test file to tests/ directory
  2. Use helpers from tests/helpers.ts
  3. Follow existing test patterns
  4. Update run-tests.sh if new environment config needed
  5. Update CI workflow if needed

Example:

import { test, expect } from '@playwright/test';
import { ensureAuthenticated, navigateToSettings } from './helpers';

test('my new test', async ({ page }) => {
  await ensureAuthenticated(page);
  await navigateToSettings(page);

  // Your test logic here
});

Getting Help

  • Check the main README for detailed information
  • Review existing test files for examples
  • Check Docker logs for service issues
  • Review Playwright documentation: https://playwright.dev

Worktree-safe shell runner

scripts/run-tests.sh builds unique per-invocation image tags and uses a unique Compose project and container names. Host ports are allocated by Docker on loopback and discovered after startup; this runner does not honour fixed port or base-URL overrides. Its EXIT/INT/TERM/HUP cleanup removes only that invocation's stack, volumes and image tags. SIGKILL or host failure can leave resources behind; the printed pulse-e2e-… project identifies those resources for manual cleanup.

This isolation applies to the shell runner, not direct npm test, setup.sh or hand-written Compose commands, which retain legacy defaults. Do not use those defaults concurrently for checkout-specific qualification.