While Qt expects const char* to be UTF-8, existing APIs treat a single char as a L1 character, incl. QChar(char) and QString::arg(char). Even char8_t literals are only allowed to contain US-ASCII content, therefore are a subset of L1. Fix the detection of encoding to treat all 1-byte characters as L1 and the rest as UTF-16. This restores consistency with overload sets including QChar, and the intention of the original author. While treating char(0xE4) as a malformed UTF-8 sequence can theoretically be argued for, that behavior would be useless in practice. This change does not affect sliced()/first()/last(), nor front() and back(). [ChangeLog][QtCore][QAnyStringView] Fixed a bug where a QAnyStringView constructed from `char` would produce an invalid UTF-8 sequence instead of a valid Latin-1 one, as expected from the equivalent QChar constructor. Fixes: QTBUG-125730 Change-Id: I8cacdd4af4b101db1c2c5797c49fe35a30abdc79 Reviewed-by: Ahmad Samir <a.samirh78@gmail.com> Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>