json::to_bson does not handle unsigned 64 bit integers

[Regfare]

Description

Trying to convert a nlohmann::json object which contains an unsigned int of 64 bit length breaks BSON conversion.
I do consider this to be a bug, as BSON should be capable of handling those values [1] , but the corresponding header-bits are not set in write_bson_unsigned. Only entry headers 0x10 (for int32) and 0x12 (for int64) are written, but 11 (for uint_64) is not.

Reproduction steps

This bug is triggered whenever one uses the json::to_bson function on any object that contains an uint64_t, that really uses all 64 bits.

Expected vs. actual results

I expect BSON to generated, but I get an error message (see below)

Minimal code example

#include <map>
#include <string>
#include <nlohmann/json.hpp>

int main(){
  std::map<size_t,std::string> m={{14631357879723465589u /*really just any large enough uint64_number*/,"To_Big"}};
  nlohmann::json j;
  j["map"]=m;
  auto bson=nlohmann::json::to_bson(j); //can not be converted
}

Error messages

terminate called after throwing an instance of 'nlohmann::detail::out_of_range'
  what():  [json.exception.out_of_range.407] integer number 14631357879723465589 cannot be represented by BSON as it does not fit int64

Compiler and operating system

Clang-13 on Mint 21 64bit, Clang Trunk (via godbolt)

Library version

3.11.2 (develop branch via godbolt) , 3.10.5

Validation

• The bug also occurs if the latest version from the develop branch is used.
• I can successfully compile and run the unit tests.

[nlohmann]

It seems I was confused with the alias "timestamp" when I implemented BSON. You are right - it makes no sense to not support uint64 numbers.

An extension should be straightforward, but to do it properly with tests may take some time.

[anshulchou-123]

@nlohmann Can I raise a PR for this?

[nlohmann]

Sure!

[slowriot]

This is resolved by #4380 which has been open since May - @nlohmann, any chance of a review? Thanks!

[nlohmann]

Sorry for the delay. I added some comments.

[codenut]

Hello @nlohmann @slowriot, what is the status of the PR for this issue? I am available to help if needed. Thanks!

[nlohmann]

Fixed by merging #4590.