Not disagreeing there. Fast NFA implementations are very nice for matching the regular subset and falling back to a more general algorithm for the non-regular c

Not disagreeing there. Fast NFA implementations are very nice for matching the regular subset and falling back to a more general algorithm for the non-regular cases :)