I'm using the mingw compiler in a MSYS2 UCRT64 environment.
I dont have a lot of experience with the Win32 api but I've got it narrowed down to the FindFirstFile/FindNextFile of which probably the narrow versions are being used.
These seem to truncate the individual charachters by dropping the extra byte instead of doing something sensible like throwing an error.
nob.h should probably be explicitly using the wide versions of these functions and converting them to utf-8.
I will try to work out a sensible solution.
I'm using the mingw compiler in a MSYS2 UCRT64 environment.
I dont have a lot of experience with the Win32 api but I've got it narrowed down to the FindFirstFile/FindNextFile of which probably the narrow versions are being used.
These seem to truncate the individual charachters by dropping the extra byte instead of doing something sensible like throwing an error.
nob.h should probably be explicitly using the wide versions of these functions and converting them to utf-8.
I will try to work out a sensible solution.