@Phoenix brb seeing if can decompile x86 to LLVM

@Downgoat ?

> Why not just use erb

I still don't get it

The word "use"?

That doesn't have drug connotations in my mind

erb mean weed

wat

...k

@Downgoat apidock.com/ruby/ERB

Question: if you do a type intersection should you take a root conformant base onto which you use as your primary type list?

" Yes "

@ATaco will jsd support routing

what do you mean by routing?

e.g. / -> Thing A, /goat -> Thing B

like: reacttraining.com/react-router

Yes, it already kind of does. Just not particularly well.

CMC: A003602: if n = (2k-1)*2^m then a(n) = k.

I'll hopefully never release jsd into the public anyway, not till I write something better

@LeakyNun putt, 3 bytes: ฒ

@Downgoat what the hell

:D

link to language?

idk if I still have >_<

unicode 3602

:| okay I never upload source: github.com/vihanb/Putt

I see what you did there

@ATaco Why not?

I want jsd

@LeakyNun :)

@Mendeleev why not webpack?

idk what that is

oh it's basically rails but JavaScript

also because JSX you can have HTML in JS, and JS in Pug

@Mendeleev if I release jsd I'm doing it right. Which isn't how I'm doing it right now

@ATaco but what if I want to merge constrained candidates

" Yes "

I am thinking of posting codegolf.meta.stackexchange.com/a/12994/71001 unless anyone has any last minute comments?

@ATaco 1) why in quote 2) that is not binary question

The spaced quote was supposed to imply a delayed and slow response, indicating that I don't understand the question at all.

CMC: Get the Prime Factors of a number without builtins.

posted.. fingers crossed there is no silly mistake

@ATaco Python, 64 bytes TIO

n=int(input())
p=2
while~-n:
	if n%p<1:print(p);n//=p
	else:p+=1

@LeakyNun oh hi

@Arthur hi

the challenge is up in case you were interested

I mean the ppcg one

Oh, that's more efficient than what I wrote.

@Arthur I saw it, thanks

it will be interesting to see what people come up with

@ATaco what did you write?

Not at all golfed, but:

function primeFactorise(n)
	local factors = {}
	for i=2, math.sqrt(n) do
		if n % i == 0 then
			local a = i
			local b = math.floor(n / i)
			a = primeFactorise(a)
			b = primeFactorise(b)
			for k, v in pairs(a) do
				factors[#factors+1] = v
			end
			for k, v in pairs(b) do
				factors[#factors+1] = v
			end
			break;
		end
	end
	return #factors==0 and {n} or factors
end

facts = primeFactorise(1325687)
for k, v in pairs(facts) do
	print(v)
end

Which is a recursive implementation.

I feel there is always a risk that a challenge either has a trivial optimal solution or is way too hard so that all the answers are the same. I hope this is somewhere in the middle

@ATaco why would you use local

Why wouldn't I? Local variables are indexed first.

As I said, this isn't golfed.

@ATaco What do you mean?

Oh, of course you should replace math.floor(n / i) by n//i.

There is no justification for converting to float and then back to integer

The way Lua's memory works, in the current scope it first sees {MY_LOCAL_VARIABLES, MY_PARENTS_LOCALS, THEIR_PARENTS_LOCALS, ..., GLOBALS}

oh wait, I thought it is python, sorry

brain fart

That was added in 5.3

oh...

local sqrt = math.sqrt
local pairs = pairs

function primeFactorise(n)
	local factors = {}
	for i=2, sqrt(n) do
		if n % i == 0 then
			local a = i
			local b = n // i
			a = primeFactorise(a)
			b = primeFactorise(b)
			for k, v in pairs(a) do
				factors[#factors+1] = v
			end
			for k, v in pairs(b) do
				factors[#factors+1] = v
			end
			break;
		end
	end
	return #factors==0 and {n} or factors
end

facts = primeFactorise(1325687)
for k, v in pairs(facts) do
	print(v)
end

That offers a small but noticeable speed boost.

I would use table.insert(factors,v) instead of factors[#factors+1] = v

That actually takes slightly longer, firstly due to the fact that table.insert needs to index table, and secondly due to the fact that table.insert is internally implemented almost identically to factors[#factors+1] =, but with a function call now wrapped around it. I'd be better off keeping track of the table's size.

oh, ok.

I don't see how my version is faster at all

My version's slower, because it needs to confirm that each prime is prime, there's a few wasted cycles as the end of each prime.

I see

Although my test implementation seems much less efficient... Perhaps I'm wrong.

On 3000 tests, it preforms 5 times as slow.

Ah, it's exponentially less efficient.

At 30000, it took 30 times as long as the other tests.

I wonder...

Well I don't understand what's going on well enough to tell you why this algorithm's faster.

if anything is unclear in my question please ping me. I will leave this window open

I have a 72MB text file. How do I upload it?

Oh, google drive.

VPS?

Here's a neat thing, the prime factors will always be in ascending order.

That is a lot of numbers.

local sqrt = math.sqrt
local pairs = pairs

function primeFactorise(n)
	local factors = {}
	local size = 2
	for i=2, sqrt(n) do
		if n % i == 0 then
			factors[1] = i
			for k, v in pairs(primeFactorise(n // i)) do
				factors[size] = v
				size = size + 1
			end
			return factors
		end
	end
	return {n}
end

A much improved version.

It occurred to me that i had to always be prime.

has anyone here heard the binding of isaac soundtrack?

if you haven't heard it, you probably should because it is good

7 votes 0 answers ;_;

classic

@ZacharyT Ubuntu 14.04

@LeakyNun Currently golfing down 126 bytes of C. :)

@Doorknob nice

Hope it doesn't take 10 minutes to do a(20)

It can do a(10000) almost instantly, although more digits than that starts to slow it down.

so is this like reverse lucky numbers?

@Doorknob nice

@LeakyNun Now you have 8 votes and 2 answers :)

quite certain your answer is wrong @Mr.Xcoder

Why?

I see no self reference

And small divisors

Better test it.

meaning that it would only work for small numbers

Again, Let's test it

alright, just let me test it

You should be right

It disagrees with mine, anyway.

First at a(20), for which it gives 6 but should give 3 (according to the test data and my answer).

I tested it myself, and got wrong result for over a(20)

I removed it

Now I've got 8 votes and 1 answer.

Sad :P

Well, this is a really cool old challenge. I want to see more competition here ;;

@ATaco APL (Dyalog), 30 bytes: {⍬≡f←i/⍨0=⍵|⍨i←2↓⍳⍵:⍵⋄∇⍵÷⎕←⊃f}

So I spent several hours code golfing at work today, and became enlightened that it's not really important to formally have code golf as part of my job description.

@feersum You mean that you can just golf anyway, even though it isn't part of your official job?

Google Hangouts removes Transparent UI

local sqrt = math.sqrt
local pairs = pairs

function primeFactorise(n)
	for i=2, sqrt(n) do
		if n % i == 0 then
			return i, primeFactorise(n // i)
		end
	end
	return n
end

Further optimized. I think I've nearly hit the point where this is just the other algorithm, strangely.

@ATaco yes, but slower

because every time you check from 2 again

you should start at i

This is running magnitudes faster, for some reason.

try starting at i

That can't work based on the algorithm.

?

Because it's often the case that it needs to return a conga line of 2, 2, 2

I think you mistook me

Perhaps.

would storing the list of prime numbers be any faster?

I mean, the smallest possible prime factor of n//i is i

not 2

@KritixiLithos 1. pointless 2. doesn't work for larger numbers

I mean, once you got a list of them, you only need to iterate through this list and bot go through every integer from 2 to sqrt(n)

I know

but I'm trying to do it without storing a list of prime factors

I still don't understand what you mean by, "Start at i"

well, for optimizations we can check 2 3 and 6n-1 and 6n+1

then how is it pointless?

I still don't understand what you mean by that.

Oh, wait now I get you.

what does he mean?

I don't get it

Yep, tests show a noticable increase in speed.

local sqrt = math.sqrt
local pairs = pairs

function primeFactorise(n, i)
	for i=i or 2, sqrt(n) do
		if n % i == 0 then
			return i, primeFactorise(n // i, i)
		end
	end
	return n
end

@ATaco how did you implement it? using a global i?

Just passing i down the chain.

interesting

Could you show me your tests?

local sqrt = math.sqrt
local pairs = pairs

function primeFactorise(n, i)
	for i=i or 2, sqrt(n) do
		if n % i == 0 then
			return i, primeFactorise(n // i, i)
		end
	end
	return n
end

function primeFactorise2(n)
	for i=2, sqrt(n) do
		if n % i == 0 then
			return i, primeFactorise2(n // i)
		end
	end
	return n
end

print(primeFactorise(12))

local TESTS = 300000

local c = os.clock()
for i=1, TESTS do
	primeFactorise(i)
end
print(os.clock()-c)

local c = os.clock()
for i=1, TESTS do
	primeFactorise2(i)

oh cool

But seriously, 12?

You should use a larger test

Probably, but I like 12.

It covers the basic functionality.

seriously

ooh, I get you now

For convenience, 64 does correctly return 2 2 2 2 2 2

Also, the outputs are:

1.507
1.626

Where test 2 isnt optimized, shows a confident increase in speed.

Seriously, you should optimize for large inputs.

Well my tests are up to 300,000 so I'd say I've got plenty of testing room at large numbers.

oh wait, sorry, I thought the test was on 12 lol

I'm not just running 12 300,000 times, don't worry.

@ATaco how does , work?

What do you mean? It's a thousands digit separator.

rust is the bane of my life... .Can anyone compile codegolf.stackexchange.com/a/127417/9206 . I get bpaste.net/show/2d84321ed6d7

@ATaco no, I mean in return i, primeFactorise2(n // i)

Oh, that's a lua thing, because return values are kind of like arrays.

It appends the results of primeFactorise to the return.

@ATaco yes, that's the basic, but it's stranger than that

so could you expand on that

In this case, Although I can't see it much different, it's returning i, aswell as, the return value of primeFactorise2, which will similarly return i aswell as it's next factor, until n is prime, in which it will simply return n. This then chains back to make a sort of pseudo array, which is faster than working with tables recursively.

@Lembik For finding out which lib you're missing: stackoverflow.com/a/21647591

And I think that's the GNU MP Bignum Library

That would be the package libgmp10 on Ubuntu I think

Not forever

eventuslly they will figure out how to work with each other and create multi cellular organisms

@betseg D-Bus?

@SIGSEGV so a cell is a biological quine

@Mayube right, or is it a forkbomb? MUAHAHAHAHA

actually yeah, 2^n cells in n time

%0|%0

:(){ :|:&};:

A cell just isn't very radiation hardened, is all.

Anyone else having problems with the userscript disabling Stack Snippets in the editor?

Good News for Both Python Golfers and Non-Python Golfers, For the former, you have a great library in here, and For the latter, you have a reason to learn Python

I think we can use the Google Assistant SDK with api.ai using node

btw how do programs get mouse input (like click, position, etc) on console

or something like that

@KritixiLithos um no

Thats actions

Its different

ooh

@Jim Thanks.. I will try that

That thing is embedding your program on the assistant, while this is embedding the assistant on your soft/hardware

Speaking of actions

boom

Heres a challenge officially from google

I'm trying out Kotlin but IntelliJ triggers me

It's sooooooo slow

And it eats up your battery like a hungry child wolfing down cake

@Qwerp-Derp sounds like you need a pc that isn't from 2008, I've never had any problems with IDEA performance-wise. Is it because your machine doesn't have a dedicated GPU? If your computer can't support hardware acceleration IDEA ends up using the CPU to render the UI, which can make it pretty slow. (Which is why you shouldn't use IDEA on a VM without some amount of GPU passthrough)

@Mayube I'm on a 2012 Macbook Air running Windows

well there's your problem

LOL, that's so true

IntelliJ was built for desktops I think, with laptops they're not that good

I can count at least 3 problems with that sentence

Because it's slow and it drains your battery

@Qwerp-Derp my IntelliJ takes 5 minutes to start

@KritixiLithos that sounds like windows start times

I wanted to just do my Java development in VIM like I do with everything else, but I quickly discovered that working in Java without an IDE is an act of insanity, so I use IdeaVIM with a custom color scheme instead

@Mayube Programming in Java with vim is not different than C

I don't particularly like C

who does?

C is still wayy beter than assembly for embedded systems

Dennis Ritchie?

ppcg is awesome.. more accurately the quality of some of the answers is awesome!

I am still trying to find an easy way to get fast simd complex arithmetic! gcc and clang are terrible at autovectorizing complex math it turns out

someone should make an optimized library

just saying :)

Is there a way to simplify a AND NOT (b AND c)?