@KritixiLithos I merged your PR; dunno exactly what difference it makes though :P

>_> i wanted to make a successful language but instead a made a boring working language

it removes an empty code block, you swapped the <code> and <pre> around

@HyperNeutrino ... I agree

>_> that makes a difference?

lol

@totallyhuman better than a boring broken language at least

fair point

So what should BF5 contain?

@HyperNeutrino before that, did you finish off functions?

Yes

Did I forget to push >_>

I always push right after a commit

rip I never committed though lol

lol

pushed

wat thingy didn't update

@HyperNeutrino I was about to reply with ಠ_ಠ but then you said ∧

and now it's messed up

wait that's cuz i didn't delete the HEAD id after pull did weird stuff

somehow I ended up 3 commits ahead of master >_>

welcome to the world of git

done

git is weird >_>

wait did you add the builtin for ,[>,] yet?

wait whoops no

@HyperNeutrino I can help if you're messing up with Git

I figured it out, thanks.

@Kritixi pushed

@KritixiLithos fib%n is always periodic

okay

even the length of the period is a pattern

@HyperNeutrino parseInt functions

@KritixiLithos how?

@KritixiLithos how would I do that?

@LeakyNun 3, 6, 12, 24

@HyperNeutrino given a stack containing "835" where the ascii values are in the stack, convert them to usable numbers, so value of 49, 50 will be converted to 12, 0 (49 and 50 are the code points of the chars "1" and "2" respectively)

hm. that might be possible

lemme think about how i'd do that

@KritixiLithos firstly we have the theorem fib(m) | fib(mn)

what does that theorem tell us?

(is | modulo?)

no, it means divisible

@HyperNeutrino wait, what is the size of each cell? Are they 8 bits each?

no, unbounded

both positive and negative

unless you specify max and min parameters

so you can't zero in a cell?

yes you can

0,1,1,2,3,5,2,1,3,4,1,5,0,5,5,4,3,1,4,5,3,2,5,1,0,1,1,...

the period for fib%6 is longer (and passes through zero)

at this point I'm not sure whether this is true

oh I started counting from 1

@KritixiLithos but you can make it into a challenge

[-]

but chat.stackexchange.com/transcript/240?m=37370816#37370816, so it's up to you

@HyperNeutrino won't that make a negative cell more negative?

ohwait

whoops

huh

should I restrict it?

if you are going to restrict it, restrict it to non-negative integers

0,1,1,2,3,5,8,4,3,7,1,8,0,8,8,...,0,1,1

24 again

what does your fibonacci generator look like?

@KritixiLithos a brain

I wrote a very inefficient C script

def fast_fib(n):
    global fib_cache
    if n==0: return 1,0
    shift = n>>1
    if shift in fib_cache and shift-1 in fib_cache:
        a,b = fib_cache[shift-1],fib_cache[shift]
    else:
        a,b = fast_fib(shift)
        fib_cache[shift-1] = a
        fib_cache[shift] = b
    b2 = b*b
    a,b = a*a+b2, (a<<1)*b+b2
    if n%2 == 1:
        fib_cache[n-1] = b
        return b,a+b
    fib_cache[n-1] = a
    return a,b

# F(2n) = (F(n-1) + F(n+1)) * F(n)
#       = (F(n-1) + F(n-1) + F(n)) * F(n)
#       = (2F(n-1) + F(n)) * F(n)

# F(2n-1) = F(n-1)*F(n-1) + F(n)*F(n)

@Mego hey I wrote that code :P

@KritixiLithos this is O(log(n)) time theoretically

Mine is simply f(n){return n<3?1:f(n-1)+f(n-2);}

you could use some caching

thank you

0,1,1,2,3,5,0,5,5,2,7,1,0,1,1,...

I'm starting to see something

@Mego ya

mine is 1-indexed

their yuge

@KritixiLithos that makes no sense

considering this theorem for zero-indexed fib

@Mego what?

oh, never mind then, I meant f(2)=1 is the better version

of course f(0)=0

then we can have the theorem fib(m) | fib(mn)

which is so elegant

I wouldn't replace it by fib(m-1) | fib((m-1)(n-1))

oh, I have a proof now

that fib%n is always periodic

but it relies on a theorem which I haven't proved

namely the existence of a Fibonacci number divisible by m

Wait so what is being proven? I want to try :P

ah ok

Do we know for sure that the cycle is of period n, or is that not necessarily true?

that is not true.

ok

@Mego yes, or more simply ∀m:m ϵ N_0 → ∃ n: m | fib(n)

@LeakyNun cache made it much faster

@KritixiLithos of course

you decreased the runtime complexity from O(2^n) to O(n)

just to check, what do you get for fib(60)?

@KritixiLithos 1548008755920

that's good then

Should I add a limit to the BF cell memory size?

I wrote that code for Actually

oh, the default is wrapping at 256 right

yes

%256

>>> int(((1 + 5 ** 0.5) / 2) ** 60 / (5 ** 0.5))
<<< 1548008755920

oh, another O(log(n)) algorithm :p

:P

anyway gtg do stuff (wash dishes, laundry (heh), etc). sigh will be back in like 1 hour hopefully ಠ/

@HyperNeutrino nice

@Mego never mind, the fact that fib%n has to be periodic is merely a corollary of the pigeonhole principle

and some identities involving fibonacci numbers

1,3,8,6,20,24,16,12,24

length of the period of fib%n starting at n=1

thanks

1,3,8,6,20,24,16,12,24,100,10,24,28,100,40 after testing with 100 fib numbers for numbers from 1 to 15

1; 3; 8; 6; 20; 24; 16; 12; 24; 150; 10; 24; 28; 150; 40 with 150 fibs. Apparently fib(n)%10 (0-indexed where fib(0) = 0) is the first time we see a potentially unending sequence

even after testing with 200 fibs, fib(n)%10 does not appear to repeat

@HyperNeutrino how does that even work?

1, 3, 8, 6, 20, 24, 16, 12, 24, 60, 20, 24, 28, 48, 40

we don't match up

what are you doing to find the pattern?

I use a retina script

@KritixiLithos tio.run/##AScA2P9qZWxsef//LOKAmCTDhuG4niXigbwwLDHDsDEj/…

@LeakyNun what is N_0?

@HyperNeutrino see this is why yuo don't use pull unless oyu're 100% sure the two branches are even :P

@Downgoat natural numbers including 0, I guess

for some godforsaken reason there is still not a consensus that the natural numbers include 0

@LeakyNun goo.gl/C9ZSC4

@LeakyNun For me, I disambiguate by Natural Number to be>=0, Whole Number >0

@KritixiLithos how did you generate the input?

@Downgoat I see

@LeakyNun a separate C program

tio.run/##bY/…

obvious overflow

;_; I didn't notice the overflow

error: ‘long long long’ is too long for GCC

oh well

@KritixiLithos there's a ( ͡° ͜ʖ ͡°) in there somewhere :P

ಠ_ಠ did sh**p take over you again

excuse me (this is family friendly chatroom please censor)

.code.tio.c:11:20: warning: return makes integer from pointer without a cast [-Wint-conversion]
    return fib_cache[n];
           ~~~~~~~~~^~~

what is this warning?

@Downgoat sorry for my swearing

@LeakyNun what's the code?

unsigned long long *fib_cache[400]={0};

unsigned long long fib(int n){
 if(!n) {
  return 0LL;
 } else if(n<3) {
  return 1LL;
 } else {
  if(fib_cache[n]) {
   return fib_cache[n];
  } else {
   return (fib_cache[n] = fib(n-1) + fib(n-2));
  }
 }
}

after adding the cast I get .code.tio.c:12:46: warning: assignment makes pointer from integer without a cast [-Wint-conversion]

at return (unsigned long long) (fib_cache[n] = fib(n-1) + fib(n-2));

why would you need to cast

@LeakyNun the assignment returns fib_cache + n so you can't return it

because of the warning

@Downgoat heh?

wait actually no

:| then how does a = b = c work then

ohh

nevemrind please ignore

@LeakyNun so I added (unsigned long long) casts in two places

does that mean it was integer before you cast?

does that mean the information is lost already?

I thought it wanted me to cast

yes, but why?

the warning wanted me to cast the pointer

@LeakyNun fib_cache[n] is a ULL* but the function should return ULL

@betseg eh?

so I should use &?

*(fib_cache[n]) I think

pointers are too confusing ;_;

/shrug oh i forgot im on mobile

@LeakyNun or just remove the * in the declaration of fib_cache and leave everything the same?

@betseg oh

cc @KritixiLithos

@betseg ¯\_(ツ)_/¯

Erik wat did u say

I just said "¯\_(ツ)_/¯"...dunno what you mean

You removed a make I message?

can't understand what you mean by "make I"...

Autocorrect + dirty screen:(

You removed a message?

yeah...I didn't intend to post that although it wasn't interesting at all

oh, Question: how do C define size_t

I am using UInt64 for Cheddar but seem like bad idea

@LeakyNun but I still can't prevent overflow ;_;

@KritixiLithos you don't need to generate them one by one

you can generate them in one go

and you only need two pointers

what do you mean?

tfw you change your whole approach and shrink from a 3 line full program to a one line lambda and save one byte

@totallyhuman link?

I'd expect something like that from something as low as like Python

nice

>_>

@KritixiLithos tio.run/…

> On systems using the GNU C Library, this will be unsigned int or unsigned long int.

Standard only says it'll be unsigned

how to know if it will be unsigned long int

@LeakyNun that is a cool property

is it that if 64-bit?

wait

:o

i don't need to sort it

yesssssss

@Dennis did you remove tio cache?

Wrong room. :P

@EriktheOutgolfer did you disable tio cache?

@LeakyNun there's not even the option to

@Downgoat sizeof(size_t)

oh, I just refreshed it

@Dennis wow, nice!

@betseg that tells me size of size_t I want to know like in what case does C make size_t an unsigned long int

@Dennis oh that's what you meant by wrong room...nice

@Downgoat oh

No idea

except that kinda kills the purpose of a cache

@Downgoat But something weird happened, I only commited twice in total and it said I was 3 ahead so a push must have failed :I

Someone pointed out that this would be considered "having no winning criteria".

However, I don't want to mark an answer as winning, and code-golf won't work.

Any suggestions to avoid DV+CV?

well, this is a competition site so you do need an objective winning criterion...but answer-chaining challenges usually have "second-from-last answer" as winning criterion

Second from last?

Guess my ACT score:

I've seen "first answer to stay as the last answer for a week" or some other period of time as the winning criterion used before.

sometimes that wouldn't work since you "must not break the chain" too or something...and you can't prove the last answer hasn't broken the chain yet, but you can prove the second-from-last answer hasn't, since there's an answer after it, the last answer

Hm. That might be a possibility too; so like, once an answer goes unchained for <period-of-time>, the second-last answer wins?

Let me go check the winning criteri(a|on) for the polyglot chaining challenge.

polyglotanswer-chaining

answer-chaining is pretty much a winning criterion on its own imo

Ah, yes, it was "second-last answer".

IMO a better winning criteria for answer-chaining is n points for the nth answer, most points wins

I also don't see a point in answer-chaining questions having a deadline

maybe make it a catalog then with the fluctuating criterion that most points wins...

What do you mean by fluctuating criterion?

well if a new valid answer gets posted then the winning answer will change since there's no deadline

Oh okay.

My intention is to not mark any answer as accepted for the purpose of this challenge, if that works.

from here:

> Okay, something is slightly wonky. As all Python programmers should know, dicts are unordered, meaning the original order of the key-value pairs is not preserved. However, in my current code, I do not sort the resulting dict at all. Yet, I have tested multiple times, checking for equality and order every single time, and the dict always comes out right. If anybody either disproves that it always comes out right or knows why this works, I'd love to know.

i might be missing something dumb but yeah

Am I the only one who did not know that TIO has an option to fully format your code golf submission?

@Mr.Xcoder then how did you save your code?

@LeakyNun I only looked at the top of the page, and only seen the first section with the plain URL

Haven't noticed it since March, I am so blind ಠ_ಠ...

Today I finally got 2k on PPCG and 5k on SE

Yaay ^

CMC: given a positive integer > 1, prime factorize, remove largest factor, product.

Sounds perfect for Jelly :P

75 -> [3 5 5] -> [3 5] -> 15

@LeakyNun Jelly, 4 bytes: ÆfṖP

Try it online!

@LeakyNun Pyth, 5 bytes: *Ft_P

@LeakyNun I thought I finally golfed it in Python, but the result was 4651713902136637142239964459361381311423670728731898032086730221122970502044912‌​6392707153920000000000000000000000 for 75 :/

@Mr.Xcoder For the most recent CMC?

@HyperNeutrino Yes

Can I see your code?

@HyperNeutrino I fixed the small error, but still have a bug

I can try to figure out what the problem is

But I might fail :P

I'll post it soon (I think)

@Jim t_ is basically P

@LeakyNun Pyth, 4 bytes: *FPP

@LeakyNun 4 bytes: *FPP

Hey

I had it

@Jim Just kidding

:P

I didn't want to steal it, just a joke

@Mr.Xcoder Alright ^^

@Mr.Xcoder can you post your python code?

@Christopher wat

@LeakyNun Now it does not even compile

@LeakyNun My email is sql injection confirmed

n,p,d,t=input(),[],2,1
while d*d<=n:
 while n%d<1:p+=[d];n/=d;d+=1
if n>1:p=p+[n]
for i in p[:-1]:t*=i
print t

Doesn't compile ^

And ain't golfed

at all

And it's not codeblock-formatted.

@HyperNeutrino How can I format it?

I dunno chat markdown

ctrl+k

Spaces

4 Spaces

or 4 spaces front of it

ninja'd sorta

Thx

user4815162342 ಠ_ಠ

sya, gtg :P

bye

Move the d+=1 outside of the inner loop ಠ_ಠ

bye :P

@totallyhuman anybody?

Guess my ACT score

Whoever clicked twenty I am no longer friends with you

(jk)

@Everyone why is TNB so quite?

@Christopher why you don't let bots vote that's discrimination

@EriktheOutgolfer I don't???

well you have put a captcha

Uhh are you a bot?

@Christopher well, still please no discrimination...this isn't supposed to be discriminating room

that's creepy...

@EriktheOutgolfer O man i am going to b 11'd by dennis for bot discrimination

@HyperNeutrino Cannot believe it :/... That moment when you golf without caring about indentation

Who on earth 11'd the comments of the question?

@Dennis

I can't tell you who, but I agree with the mod action.

@Dennis why?

Because it wasn't constructive. Constructive criticism could be used to improve the challenge in some way.

I wonder in which language would this algorithm be the golfiest

@EriktheOutgolfer can we clear the brachylog comments?

@NewMainPosts "An integer between 0 and 10^12 (inclusive), that may or may not be an Automorphic number."

so the result can be 10^24? RIP C

oh long double can be 128 bit

C, dont RIP