By default, this filter ignores any tokens in the wrapped {@code TokenStream} once the limit has been exceeded, which can result in {@code reset()} being called prior to {@code incrementToken()} returning {@code false}. For most {@code TokenStream} implementations this should be acceptable, and faster then consuming the full stream. If you are wrapping a {@code TokenStream} which requires that the full stream of tokens be exhausted in order to function properly, use the #LimitTokenPositionFilter(TokenStream,int,boolean) consumeAllTokens option.