Question

In the struct dirent can the member d_name have characters that are unicode but not ascii? I tested and it works with ‘é’ but I’m not sure if that’s because ‘é’ is in ascii. What type is d_name, if it’s char then I’m hooped but if it’s wchar_t I’m saved.

Answer 1

The Open Group Base Specifications state that struct dirent shall have the following member of unspecified size:

char d_name[];

In conforming implementations it is of type char[] but that doesn’t necessarily mean that it can’t store Unicode characters; the specification does not require the support of multiple character sets (or characters beyond the portable character set) but it doesn’t forbid it either ('é' by the way, is not in ASCII).

In practice, the supported character set depends on your locale. In C terms, the string in d_name is stored as “multi byte” characters.