The point of ProbabilisticMap is to avoid doing expensive values.Contains(char) checks for every input character by applying a bloom filter first. This works well for IndexOfAny, but with IndexOfAnyExcept we still have to check each character since we don't have an "inverse bloom filter".

For an input value that isn't in the set, the bloom filter will save us from doing the verification step, but that will by definition happen only once per IndexOfAnyExcept call. The overhead of checking it for all the characters that are in the set is wasted.

This PR avoids doing those bloom filter checks for IndexOfAnyExcept paths.

In practice IndexOfAnyExcept in IndexOfAnyCharValuesProbabilistic now isn't doing anything interesting with the ProbabilisticMap, but that's just a convenient place to share some of the logic that ProbabilisticMap needs anyway.