On Twitter @LunarLambda pointed out that Clang allows Hangul whitespace Unicode
characters in identifiers, which allows users to write very confusing
programs: https://twitter.com/LunarLambda/status/1110097030423240705
Clang warns about similar whitespace Unicode characters. Add the Hangul
half-width and full-width whitespace characters to the set that Clang
warns about.
N.B.: Clang warns about Japanese space character <U+3000>, but in a
different way, because that character is not a valid identifier
character according to the C++11 standard. So Clang emits a warning that
it will treat the Japanese <U+3000> as whitespace. This is different
from the Korean Hangul whitespace character, which is a valid identifier
character according to the C++11 standard. For this reason, Clang warns
the character will be treated as an identifier character, not as a
whitespace character -- so in sum, Clang's behavior is slightly
different for the Japanese whitespace character compared to the Korean
Hangul one.