I reckon there's some public spec doc for this to make implementation relatively easy

I've come across some interesting 'tricks' like converting a char to its corresponding integer by subtracting '0' or uppercasing a string using bitmask [0xffdf] 😉

currently going through the tokenizer. lots of interesting easter eggs. the logic itself is pretty straightforward and simple

hi @eatonphil.bsky.social,
why isn't h2 ever mentioned in db design conversations? is there anything wrong with its design/implementation?