On 15.06.2012 23:03, bearophile wrote:
A lazy Knuth-Morris-Pratt that works on a Input Range:
http://ideone.com/dUs5B
Do you have suggestions for improvements of the code? Maybe do I have to
turn it into a Forward Range if the first range is a Forward one? Is
something similar to this useful for Phobos?
Yes, definitely just decouple table preparation and searching range
itself. It's common to use KMP and its ilk to do a lot of series of
searches for the same needle.
Bye and thank you,
bearophile
--
Dmitry Olshansky