Handling of overlong character literals

[bkamins]

After the change of Char representation it is possible to get overlong characters from a String, e.g.:

julia> c = String([0xc0,0xa0])[1]
'\xc0\xa0': [overlong] Unicode U+0020 (category Zs: Separator, space)

However, if you try to write this as a character literal you get:

julia> '\xc0\xa0'
' ': ASCII/Unicode U+0020 (category Zs: Separator, space)

so they are converted to their not-overlong equivalents.

I know it is a corner case, but I suppose that this behavior is not intentional.
So the question is is this a behavior we want (or we want character literal to be parsed exactly as written)?

[JeffBezanson]

Duplicate of #25072