hellerve 2 hours ago

That’s quite fun, thanks! (Not super far yet)

For some of these lookup tables, I wonder why we wouldn’t just use a macro to generate them (the deterministic ones like Huffman encodings, not the Monte Carlo ones) instead of hardcoding and then using the algorithm in the test. But maybe I’m just being a macro fiend again.

joonfjp 2 hours ago

I imagine we'll start to find magic constants everywhere..