Eldred Habert

I don't know. The risk of adding false statements to Pan Docs always exists, and is a regression of sorts. I'd like to fix them from the get-go if the...

SameBoy's source is one of the best references, since it's accurate, well-tested, and *often* well commented; but since it's code, it's still subject to interpretation. That's why we prefer citing...

Have you been able to find any test ROMs? If not, that's fine, then SameBoy would be quite sufficient.

Note that the PR is, by itself, complete, but that the design of both `INCBIN` functions sucks; I will refactor it to perform bulk reads **and** updates of the various...

There are CI failures, and Sylvie pointed me out some further improvements this PR could use, so I've marked it as draft for the time being.

Yay! All green! The claimed performance improvements are yet to be measured, though.

Since the CPU is able to load only individual data pieces, and I believe that operation is not slower than loading the “native size”, I don't think using smaller data...

If the runtime is on par, how about the memory consumption?

I thought this could be implemented without too much hassle by making numbers in RGBASM a `{ value: u32, precision: u8 }`. (Integers simply are `q0`.) This requires no changes...

I don't see a *principled* way either, it's just that I've found it somewhat tedious to have to keep thinking about that info as I've been manipulating fixed-points (e.g. to...