Over the last few days, I managed to get DOS 5.5 booting correctly. HIMEM.SYS handling is now working as expected and no longer breaks the startup sequence.

I also started testing Windows 2.0/286. The boot process is now successful, but VGA modes still cause display corruption in some cases, and graphics output is occasionally incorrect.

Given how many resources I burned trying to fix Windows 2.0 graphics, I will now shift focus back to MS-DOS. This was an experimental step forward: now I know the boot path works, but the real center of gravity is VGA compatibility.

In the last few days, the codebase has also grown quickly, so I had to start a small refactoring pass. Over the next days, I plan to focus on that refactoring work and on stabilizing smaller features.