From: "Gulyás Attila" <toraritte@...>
Date: 2013-02-28T03:02:27+09:00
Subject: Re: Please, help (GCD) greatest common divisor.

Thanks for taking the time to looking into it!
I could figure out the boolean chain given the precedence but I
thought that I am missing something obvious with "--p" because I
couldn't twist more out of it in irb. ( ... and yes, my first thought
was that p is being decremented but that made no sense:)

I always found "inject" hard to grasp (my mind is like Pratchett's
A'tuin when it comes to abstraction) and your example gives a new
angle to chew on as I've never seen this approach before. Thanks!

If you're interested in obfuscated code: http://mamememo.blogspot.hu/
(Yusuke Endoh's blog)
I saved this for later because it's way over my head and he is writing
one twisted code:)
Thanks again.
Attila

On Wed, Feb 27, 2013 at 6:19 PM, Bartosz Dziewoński <matma.rex@gmail.com> wrote:
> Replying to the gist.
>
>> ### II. "inject"-part
>> # Breaking it down with [6,8]
>> gcd[[6,8]]      #=> 2
>>
>> gcd=->(_){_.sort_by(&:-@).inject{|p,q|q.nonzero?&&gcd[[q,p-q]]||--p}}
>> # ( I am sure that I am missing something here )
>> # 1st iteration: [8,6] ->                       6 && gcd[[6,2]] || --8
>> # 2nd iteration: [6,2] ->                  6 && 2 && gcd[[2,4]] || --6
>> # 3rd iteration: [4,2] ->             6 && 2 && 2 && gcd[[2,2]] || --4
>> # 4th iteration: [2,2] ->        6 && 2 && 2 && 2 && gcd[[2,0]] || --2
>> # 5th iteration: [2,0] -> 6 && 2 && 2 && 2 && nil && gcd[[0,2]] || --2
>> # On the last iteration everything evaluates to nil before the ||.# (One
>> remaining question: why --p?)
>
>
> #inject on a two-element array is essentially the same as splatting it in
> two separate variables. The following two snippets do the same thing:
>
>     result = [1, 2].inject{|a, b| do_stuff_with(a, b) }
>
>     a, b = *[1, 2]
>     result = do_stuff_with(a, b)
>
> "a && b || c" is the same as "if a; b; else c; end" (or "a ? b : c"). This
> is a little abuse of the behavior of boolean operators :)
>
> "--p" is just for laughs. While this looks like decrementation in C / C++,
> it parses as "-(-p)" as is the same as simply "p".
>
> --
> Matma Rex
>