This library does not document whether 16 bit architectures are considered supported, yet blockdata::script::read_uint returns usize for what can go above 2^16.

Either it should be documented that 16-bit platforms are not supported or the above mentioned function should be fixed - return a different type or an error. I'm not sure which one it should be, so I'm asking here.

Same could go for 32-bit although there I would b surprised if it was unsupported and I don't know of any code that breaks besides #658 which has an easy, obvious fix.