Rewriting as C++ would only help a little, so it proabaly could read NULs, but it is still required to be (convertible to) UTF-8.

Just a query, why is it so limiting? What languages have embedded NULs?

Geany is a lightweight IDE which edits text files, and can build applications from them, thats its use-case, and then Unicode covers all text. Its use-case does not include examining and patching binary files.

If you have a use-case for reading/modifying binary it would be better to use the right tool, one that is designed to handle unformatted binary.