Replace `String.fromUtf8Unchecked` with `String.fromUtf8?`

[Kha]

Unlike String.fromUtf8Unchecked, this one should check for valid UTF-8. In fact, since we have to decode the string anyway to figure out its character length, we might as well remove the former function.

[Vtec234]

However all String.Pos-based functions immediately work for an Utf8Unchecked without any extra decoding because the UTF8 size is just the ByteArray size and can be cached from the start.

[Kha]

I'm not quite sure what you mean. String eagerly, always caches the number of code points.

[Vtec234]

My bad, indeed there is an eager utf8_strlen call in the constructor and the "unchecked" is not any faster.