eggbug enjoyerso i was implementing format() for Crucible and i decided i wanted to tackle formatting integers and floats by hand so they could be constexpr. integers i came up with a probably-not-super-great-but-good-enough implementation in a couple hours. then i decided to look up algorithms for printing floats.
as it turns out, there have been entire academic papers written on the subject of "printing floating point numbers quickly and accurately" (literal title of one paper)
i found a pretty good summary of the general idea in this article, and while it's definitely not the most complicated code i've ever dealt with, it's amazing how a different format can spiral out into so much complexity, even with the same number of bits!
