Crosscut

There's an obvious way to address the problem of the code database in the new prototype: Just include it from the beginning. And that's what I've been doing.

I had initially planned to implement the "checkout" model, because that seems easier and allows for reusing established editors. I had also planned to include a debugger again (more on that later), but with a very minimalist command-line interface that can render code and parse simple commands.

But before I had the opportunity to go down that road, I looked at my plans for this debugger CLI and had another one of my "how hard could it be" moments: Could I just implement a whole editor like this? Well, not exactly, but maybe with a Vim-like interface? I already started regretting that, but now I seem to be over that first hump. We'll see how it goes.