Reintroduce an Exception type for invalid UTF-8

[gregw]

Introduce Utf8StringBuilder.InvalidUtf8Exception as a substitute for the removed Utf8Appendable.NotUtf8Exception.

[joakime]

Can we get by with a nested cause instead of reintroducing the top level exception?

[joakime]

I would prefer something like this instead ...

value = buffer.takeCompleteString(() -> new IllegalArgumentException(
new CharacterCodingException("Invalid value: Bad UTF-8")));

[gregw]

The problem is that CharacterCodingException does not take a message. It doesn't even take a cause, so we really have to extend it and force the toString and message, otherwise we don't know what the character coding problem is.

So now I have:

    public static class Utf8CharacterCodingException extends CharacterCodingException
    {
        @Override
        public String getMessage()
        {
            return "Invalid UTF-8";
        }

        @Override
        public String toString()
        {
            return "%s@%x: Invalid UTF-8".formatted(CharacterCodingException.class.getSimpleName(), hashCode());
        }
    }

    public static class Utf8IllegalArgumentException extends IllegalArgumentException
    {
        public Utf8IllegalArgumentException()
        {
            super(new Utf8CharacterCodingException());
        }
    }

So any exception will always have a CharacterCodingException in its causal chain. It will also have the "Invalid ITF-8" message reported.

[joakime]

Almost feels like we should use java.nio.charset.MalformedInputException instead (which extends from CharacterCodingException).

Like the JDK itself does in the java.lang.String class

https://github.com/openjdk/jdk/blob/jdk-17%2B35/src/java.base/share/classes/java/lang/String.java#L1238-L1241