Precompiled headers

[aceinetx]

Needs testing
This PR speeds up compilation times (from 10m 30s avg. to 8m 30s avg. according to 2 ci/cd runs) by adding precompiled headers.

[mentatbot]

If you would like me to review this PR, either tag me in a comment or check this box:
✅ Review PR

_{You can disable automatic comments on my settings page}

[mentatbot]

I'm getting to work. See my steps and track my cost here.

• Wake on any new activity.
  _{I always see new activity while actively working. When paused, I'll only respond when tagged unless this box is checked.}

[mentatbot]

PR Review: Precompiled Headers Implementation

I've reviewed the implementation of precompiled headers in the codebase and it looks well-executed with a clear performance benefit. Adding precompiled headers is a great way to improve build times, especially for C++ projects with many headers.

Strengths

1. Performance Improvement: The 2-minute reduction in build time (from 10m 30s to 8m 30s) is a significant improvement that will benefit developers.

2. Clean Implementation:

   • The PCH files are well-organized with standard headers that rarely change.
   • The src/pch.hpp has comprehensive coverage of std library headers.
   • The test/pch.hpp is appropriately minimal with just what's needed for tests.

3. Build System Changes: The CMake integration is correctly done with target_precompile_headers() for all relevant targets.

4. Workflow Improvement: Updating the GitHub workflow to use actions/checkout@v4 is a good practice over manual git commands.

Minor Issues

1. Empty Header File: src/Emulator/Core/CPU/Interrupts/InterruptHandler.hpp appears to be empty. Is this intentional?

2. A Few Redundant Includes: Some files like src/Emulator/Misc/print.hpp include headers directly that are already in the PCH. While not a functional issue, it's slightly inconsistent.

3. Optimization Potential: The fmt-related includes could potentially be further optimized to ensure they're only in the PCH to avoid duplicate inclusion.

Conclusion

This is a solid PR that achieves its goal of improving build times. The implementation follows best practices for precompiled headers, and the approach taken is clean and maintainable.

The PR is ready to merge from my perspective, as the issues noted are minor and don't affect functionality.