2020-07-15

Yesterday, I got managed to get a better version of my last-use analysis to work. It still doesn't handle aliasing properly, but I think it's good enough that I can adapt my linear-scan to use it instead of the old last-use. I also really need to write the pseudo-code for my linear scan for Cosmin, today.

So, after figthting with scoping for a bit, I think linear scan works a lot better now. I'll write some more about the status tomorrow. And write the pseudo-code…