06:35 <yasserhussain111> hello

18:04 <yxhuvud> Hmm. Shouldn't Enumerable.html#each_cons use a Deque instead of an array internally?

18:09 <oprypin> yxhuvud, the `count` in each_cons is almost always tiny, so it would make no difference. but it should certainly be allowed to pass a `Deque` for `reuse`.

18:11 <oprypin> i'd change it to `if reuse.is_a?(Bool)`

18:12 <oprypin> or actually, this could be two different methods

18:12 <yxhuvud> oprypin: with 25_000_000 elements, the difference is big even with only 2 elements.

18:13 <oprypin> it will probably be slower with deque in that case, lol

18:13 <oprypin> try and see

18:21 <yxhuvud> I did, and deque seems twice as fast at that size, for small inputs and has constant speed on bigger counts, while the array implementation balloons. I guess I'll do some more organized benchmarking tomorrow.

18:21 <yxhuvud> *inputs -> counts. input stay constant at 21M

18:23 <oprypin> yxhuvud, is `a, b = pair` part of your test?

18:23 <oprypin> cuz i cant believe that it's actually faster

18:28 <yxhuvud> well, I compare it against a needle I'm looking for and it prints out the index for when it finds it. the index is 20288091, FWIW.

18:29 <oprypin> please paste the code

18:29 <oprypin> im just genuinely curious

18:32 <yxhuvud> https://gist.github.com/yxhuvud/38f437fd8e7dfdd2b5811674a1b3978c <- comment at the end as necessary (contains AoC spoilers)

18:42 <oprypin> yxhuvud, just send a pull request to make it accept not only Array for `reuse`. but i think changing the default will be very controversial

18:42 <yxhuvud> can you repeat it?

18:44 <oprypin> you can just re-read it O_o what do you mean?

18:44 <yxhuvud> the results I had, not your sentence :P

18:45 <oprypin> oh.. well i looked at the program and i was just disappointed it's not a Benchmark

18:46 <yxhuvud> Well, it was the code I had available. time is not very precise, but can still be repeatable enough to show a difference.