SDAM requires us to downcase domain names of MongoDB servers:

https://github.com/mongodb/specifications/blob/master/source/server-discovery-and-monitoring/server-discovery-and-monitoring.rst#hostnames-are-normalized-to-lower-case

Our decision was to translate uppercase ASCII chars to lowercase and leave non-ASCII chars unchanged.

The C Driver uses mongoc_lowercase for this. I'm concerned that a multibyte UTF-8 character could be corrupted by its algorithm:

void
mongoc_lowercase (const char *src, char *buf /* OUT */)
{
   for (; *src; ++src, ++buf) {
      *buf = tolower (*src);
   }
}

It should probably use bson_utf8_next_char to advance through the string, instead of using "++".