Boost.Multiprecision support

[zxffffffff]

The precision issues with float and double have long been a challenging aspect of programming, particularly when JSON is used for network transmission, Java's BigDecimal often gets converted to number, but there isn't a well-suited solution on the C++ side.

I've typically used a workaround involving json_dump() -> parse_string() -> cpp_dec_float_50, which is rather inelegant. It would be ideal if a JSON library, known for its ease of use, could natively support this.

[nlohmann]

Looking into this.

(Also linking this to our existing tracking item for user-defined/custom number types, since arbitrary-precision support like Boost.Multiprecision would build on the same underlying mechanism.)

---

This reply was drafted by Claude Code on behalf of @nlohmann, as part of a review of open discussions.