I am still not up to speed, but the technology is progressing.
The virtual machine now makes some architectural sense, running inside a host that can now pause execution for inspection purposes. Not that there's much to see so far, except for that test pattern, and that the flash attributes stop flashing when the vm is paused.
Pleasingly, I was able to implement the video ram in native format at an acceptable speed, so those test patterns are actually bytes poked into place. It's not quite like working in Z80, but far closer than using Pygame libraries.
The next step is to implement renderers for the bitmap font, backgrounds, and sprites. These are simple enough where raw data is concerned, but compression and decompression will involve repaying the technical debt of code which I wrote to handle such matters in my previous game.
Then, after rather less hassle with passing keyboard commands, and moderate hassle of adding a placeholder title screen, I shall be ready to tackle the unknown hassle of implementing character movement. My ambition by the next entry is to have something that the player can control, and the challenge is to do so with code that is equally prepared to marshal npcs.
This amounts to stunningly dull news for most backers, but don't worry - I'll let Mr Biffo know as soon as there's the slightest risk of something fun happening.
Email: comments at arbitraryfiles.com