1. 2

  2. 4

    There’s no documentation linked, no explanation of use-cases, features, etc. What’s the point of this submission?

    1. 2

      It’s just neat how it uses dependent types. For example, here is an image where the size of the pixel matrix is typed, but using a previously decoded value: https://github.com/yeslogic/fathom/blob/main/formats/image.fathom#L12

      Isn’t that kinda neat? I don’t think this is possible using Typical, Avro, Binrep or Kaitai. Wuffs might be able to do it since it has refinement types. But OTOH, it also looks more complicated.