In almost every place we use {Last}IndexOf{Any} for a given character or set of characters, we can use {Last}IndexOfAnyExcept for the negated counterpart.  This leads to cleaner generated code, and for longer searches, benefits from vectorization.  This includes:
- In set loops of a small number of negated chars, finding the end of the loop with IndexOfAnyExcept.
- In loops followed by a set of a small number of negated chars, using LastIndexOfAnyExcept to find the next location to which to backtrack.
- In Singleline .*? lazy loops followed by a set of a small number of negated chars, using IndexOfAnyExcept to find the next location to backtrack to.
- In unbounded atomic one and set loops, using IndexOfAnyExcept to find the end of the loop

The two places we don't currently use IndexOf* but don't use the *Except variants are:
- As part of TryFindNextPossibleStartingPosition.  There appear to be relatively few patterns in the wild this would benefit, and we'd need custom logic for finding the set of negated starting characters.
- As part of backtracking in a non-Singleline lazy loop.  It needs to search for the target plus a \n, which works well when the target is a non-negated char (as it just adds the char to the set of things to search for) but doesn't when working with a negated char (as then different elements of the search have different polarities).