hola / challenge_mail_filter Goto Github PK

I just tried to fix my implementation to pass the correctness tests and I noticed that you does not test large rule sets. All tests were run at 100k×100 configuration. I believe this does not check scalability of the algorithm design at all. Many fast implementation will fail if you make rule set large enough. This is unfortunate, because I was trying to remove polynomial dependency of rules number in the first place.
Frankly, I thought the scalability was the goal of challenge.

Programmer who created this benchmark have no conscience and enough competence or just lazy man.

#5
Please check it and merge my pull request.

I think the better is to calculate time per message, not the average of sum.

For example T = t1 + t2/8
Where t2 for 800k messages and t1 for 100k