Enhancement: Improve server process cleanup and readiness checks

[justin808]

Background

While implementing a multi-mode E2E test runner for React on Rails demos (shakacode/react_on_rails-demos#17), we discovered some improvements that could benefit the server process management in cypress-on-rails.

Suggested Improvements

1. Add Warning Logging for Process Group Kill Failures

Current behavior: When Process.kill raises Errno::EPERM, there's silent fallback which makes debugging difficult.

Suggestion: Add warning logs with process group ID and error details:

def send_term_signal
  if @server_pgid
    Process.kill('TERM', -@server_pgid)
  else
    safe_kill_process('TERM', @server_pid)
  end
rescue Errno::ESRCH, Errno::EPERM => e
  puts "Warning: Failed to kill process group #{@server_pgid}: #{e.message}, trying single process"
  safe_kill_process('TERM', @server_pid)
end

2. Improve Server Readiness Check

Current behavior (assumed): May accept any non-error response as "ready"

Suggestion: Be more precise about what constitutes a "ready" server:

def server_responding?
  url = "http://localhost:#{@port}"
  response = Net::HTTP.get_response(URI(url))
  # Accept 200-399 (success and redirects), reject 404 and 5xx
  (200..399).cover?(response.code.to_i)
rescue Errno::ECONNREFUSED, Errno::EADDRNOTAVAIL, SocketError
  false
end

Rationale: A 404 response means the server is running but the app isn't ready. Accepting only 2xx-3xx status codes provides a more accurate health check.

Benefits

• Better debugging: Process termination issues become visible in logs
• More reliable tests: Server readiness checks won't pass prematurely
• Fewer flaky tests: Proper health checks reduce timing-related failures

Implementation Note

These improvements were tested in production use with Playwright tests across multiple server configurations. See the implementation at: https://github.com/shakacode/react_on_rails-demos/pull/17/files

Would you be interested in a PR implementing these improvements?