Erlang's bitstring syntax has some weird endianness behaviour: it's biased to big-endian and can do weird things with non-byte-aligned little-endian strings. Here's an example based on the old 26-bit ARM program counter - https://fanf.dreamwidth.org/77938.html
Some years ago there was some nice work on adding bitdata support to Haskell, aimed at the kinds of things operating systems need to do with hardware (rather than serialization or network protocols) - http://web.cecs.pdx.edu/~mpj/pubs/bitdata.html and http://web.cecs.pdx.edu/~mpj/pubs/bytedata.html - these take the approach of bitdata being for layout of bitfields inside words, and bytedata being for detailed control of the memory layout of fields in a structure.
More recently I fleshed out the ideas at the end of that old blog post, with some thoughts on how to handle differing endianness in a more principled/algebraic way - https://fanf.dreamwidth.org/132484.html
Some years ago there was some nice work on adding bitdata support to Haskell, aimed at the kinds of things operating systems need to do with hardware (rather than serialization or network protocols) - http://web.cecs.pdx.edu/~mpj/pubs/bitdata.html and http://web.cecs.pdx.edu/~mpj/pubs/bytedata.html - these take the approach of bitdata being for layout of bitfields inside words, and bytedata being for detailed control of the memory layout of fields in a structure.
More recently I fleshed out the ideas at the end of that old blog post, with some thoughts on how to handle differing endianness in a more principled/algebraic way - https://fanf.dreamwidth.org/132484.html