Remove inherent `ascii_ctype` methods from `str` and `[u8]`

Remove inherent `ascii_ctype` methods from `str` and `[u8]`

This has been discussed in #39658. It's a bit ambiguous how those
methods work for a sequence of ascii values. We prefer users writing
`s.iter().all(|b| b.is_ascii_...())` explicitly.

The AsciiExt methods still exist and are implemented for `str`
and `[u8]`. We will deprecated or remove those later.