According to https://msdn.microsoft.com/en-us/library/windows/desktop/dd317756(v=vs.85).aspx mariadb_compiled_charsets array contains wrong data for codepage field for some collations. latin1 being probably most important.

Also for "auto" charset detection on Windows(mariadb_get_charset_by_name("auto")) I guess it's better to always use GetACP(), and not GetConsoleCP() in case of GetConsoleWindow(). At least from c/odbc point of view. Since GetConsoleCP() returns "OEM" codepage(usually 437 in US/Europe) and not default ANSI codepage(1252 in same places)

Suggested fix: https://github.com/MariaDB/mariadb-connector-c/pull/47