Fix invalid wide strings (UTF-16) handling

[TheBestTvarynka]

Usually, we multiply the chars count by 2 to calculate the total bytes in the wide string. For example,

sspi-rs/ffi/src/sspi/sec_winnt_auth_identity.rs

Line 323 in bef5fd2

domain: raw_str_into_bytes((*auth_data).domain as *const _, (*auth_data).domain_length as usize * 2),

Read this thread for more details: #304 (comment)

In the scope of this issue, we need to fix it.

[CBenoit]

We could also consider using the widestring crate for a UTF-16 string type if necessary.