It’s been a number of years since I had a chance to look at this project, but recently I updated the codebase to the latest version (sync with upstream ScummVM) and found that the videos no longer worked.
It turned out to be a pretty simple fix (see commit), but it rekindled my interest in the project.
I added some additional debug output and cleaned up the script debug loop so that there wasn’t as much spamming of commands that were waiting for another event.
Also, thanks to @xesf, we can now actually proceed up and down stairs via fixes he made to the script processor.
The path-finding in the original engine is a lot smarter and more forgiving than our implementation.
We can currently pick up an item and remove the graphic from the screen, but there’s no way to retrieve/use items once collected.
clean up graphical glitches
There are still a number of clipping issues, as well as incorrect behaviour when Max is standing still (wrong animation might play).
Encounters load, but you can’t interact with keywords yet.
There was quite a bit of work done on this, but we still need to verify that everything is being restored properly on load.