feat: unified CI/CD pipeline and production Docker deployment

[Omdeepb69]

This PR directly implements GSoC 2026 Objectives 1 (First Deployment)
and 2 (CI/CD Integration).

Problem
The existing CI setup had two separate fragmented workflow files that
did not install GDB or g++, meaning backend tests involving compilation
would fail on any CI runner. The docker-compose.yml used a deprecated
version field, had no healthchecks, and the backend had no production
Dockerfile — only a dev one.

Changes

CI/CD:

• Replaced run-test-for-server.yml and run-test-for-webapp.yml with a
  single unified ci.yml
• Backend job installs gdb and g++ so real compilation tests run in CI
• pytest runs with --cov-fail-under=50 to enforce coverage threshold
• Frontend job runs npm ci and npm run build to catch build failures early
• pip and npm caching for faster CI runs
• Both jobs run in parallel

Docker:

• New gdbui_server/Dockerfile using python:3.11-slim with gunicorn
  for production serving (not Flask dev server)
• Installs gdb and g++ in the container
• HEALTHCHECK instruction so the container reports ready status
• Rewrote docker-compose.yml: removed deprecated version field, added
  service-based networking, healthchecks, restart policies
• webapp service depends_on gdbui_server with service_healthy condition
  so the frontend only starts after the backend is confirmed ready

Note on VITE_API_BASE_URL: The docker-compose build arg sets this to
the internal service name (gdbui_server:10000) which works for
server-side rendering scenarios. For browser-based SPA deployment,
this should be set to the public hostname at build time — documented
in .env.production.