To be fair it doesn't really have strings 'built in' particularly well either. It directly manipulates memory patches, and the string support that is there is mostly a very simple abstraction on that.

If you want support for unicode etc, there are libraries.

But then that brings you compatibility issues, overhead ...

If you want to avoid development overhead, absolutely, C is a terrible language to pick.

I'm not sure what you mean by compatibility issues in this context?

I'm thinking about differen string representations used by different libraries. Is this not true?

Honestly all I've encountered default strings (and people shy away from those now) and LibICU stuff. I think ICU has become a defacto standard but I have been out of C for a few years.

This tends to be important for strings you're going to show to the user (or input you gather from them), maybe most of the C in the world isn't aimed at interfaces like that...

Actually the best way to deal with unicode in almost all situations is to use UTF-8, and dealing with UTF-8 in C is trivial.

For other encodings, well you need to decode explicitly (or just go for wide chars). Shouldn't come as a surprise, and not a big deal either.