> I wonder if the javascript version is faster because it's less safe or because the JIT compiler identifies properly that unicode-awareness is useless and specializes the code.

Most JS VMs have several different string types internally. Looking at V8's source code (https://github.com/v8/v8/blob/master/src/builtins/builtins-s...), it looks like there's a very happy ASCII fast path that makes toLowerCase really fast on ASCII. Actually, it even looks as if it's processing the characters 8 bytes at a time (https://github.com/v8/v8/blob/master/src/strings/string-case...).