@b4n, could you look at this?

LGTM.

char vs. unsigned char is a bit subtle, especially because an implementation is actually allowed to have char itself as unsigned. Anyway, most API taking a string expect a pointer to char, not unsigned char. OTOH, most API taking single characters expect an int with the value of an unsigned char.

In theory, unsigned char makes more sense so you get meaningful values for bytes > 0x7f (e.g. above ASCII) -- having e.g. byte 0xff as 255 looks more sensible than getting -1.
But in practice it doesn't matter so much, because not only as we see most APIs take chars, and APIs that take ints properly handle the conversion anyway, but also we generally don't deal with non-ASCII characters directly anyway.

So I guess that when we want to use string.h API it makes more sense to use char.