@Adám Nope, not until they have a user account here. Apparently the SO account on its own isn't enough to trigger, which is odd. I don't have their Chat Account info so I can't add them. Have them join here first, then I can extract their chat ID and add them that way

@ThomasWard Sorry, yeah, I know about that, but didn't notice I was looking at an SO account.

finger wags at you

hmm, for some reason the diamond mods only show the diamond, not their nick, if they only speak one line of text

meanwhile I see "Thomas ♦︎"

Apparently it's screen width issue again

GLORY UNTO CODE GOLF

FOR CODE GOLF IS IMPORTANT

in names without spaces like HyperNeutrino only a diamond shows but it seems like either the last word, all but the first word, or a certain number of words depending on width get chopped off

@UnrelatedString better? :P

YES IT IS

actually, I see "Wheat ♦︎" if I scroll up, so it's just wrapping for longer names for me

yeah screen width is a pain

I see full name, then as I shrink the window it becomes Thomas ♦︎, then Thoma ♦︎, down to T ♦︎

yup

so that settles that it's width based

But I can't reproduce only single diamond being shown

all you need to know is the Diamonds are power :P :)

@Bubbler what browser

Chrome on Windows

I'm on Firefox on macOS 10.14

Unrelate

d?

That's some kind of inconsistent UI

Or is that an imperative verb?

SE chat doesn't even try to be responsive.

@Adám i see a nonstandard theme :P

wags finger at you again

@ThomasWard github.com/dzaima/PPCGDark

as i said nonstandard theme which may be the reason the name is gone :P

looks like it floats right and Firefox pushes the name down "to the next line" if it won't fit

ahh

yeah that checks out

That's FF-specific then

and i guess chrome does the same thing but without splitting on spaces

actually safari 13 seems to have the same problem, just tried it

(yes there's an update pending, not sure why it didn't update automatically)

@Adám maybe thats why the mobile version looks completely different

@Razetime no it looks completely different because trying to use the full version is a real pain on a small screen and the old version didn't seem to fit very well

I'm not sure why I'm getting downvotes on this

reddit.com/r/adventofcode/comments/kj53l1/…

CMQ: How do I get a truthy value when comparing the list type and iter type in python?

I tried list is iter and iter is list but they don't work

Oh wait iter isn't a type

oh lol

nvm then

@Lyxal please everyone ignore this message

That must be list == [*iter] in py3 (does it work in py2?)

or is it a totally different question

Probably a totally different question

I wanted a way to check if a type is iterable (i.e. list, range would return True while int, float, str, etc would return False)

(str isn't considered an iterable in this particular instance)

str is also iterable

then you have a problem

well what do you expect when implementing type cohesion not exactly implemented by python?

maybe one way to do it is to check if it can be consumed by iter, then it is not specifically a string

try: iter(x);return x is not str
except: return False

erm, that doesn't work either

I ended up just explicitly checking for range and list

Tomorrow I'll have to do days 20 and 23, I never finished them

@RedwolfPrograms that's a lot to do in one day

20, 23 and 25

Well, 8 hours with nothing to do should be enough. AoC isn't until 2300 in my time zone

Lel, AOC is 1600 for me

Hopefully today's won't be super difficult so I only have to do 3 instead of 4

This one better not be related to stupid crab games or my answer will just be step on the fricking thing

@RedwolfPrograms plot twist, that's the correct answer

Making python take implicit input if a list doesn't have enough items without creating a custom class is harder than you'd think

At least that's my current experience

Especially if the source the implicit input is taken from is unique to the list

Has anyone solved all aocs in brainfuck?

Presumably not

Probably not originally, but it'd be not-too-hard to go through a few transpilations to more and more tarpit-like languages until you go from a python or JS solution to BF

If it takes a long time for naive solutions in languages with standard data structures (e.g. python), then how much longer would things take in a language which only works with cells?

BF solutions would probably have to run for quite a while to get an answer

Unless someone makes a CPU that runs BF directly

or someone has access to a supercomputer

Besides, I doubt anyone would have enough patience to parse the various input formats supplied by AOC

Input formats which would have benefited from being placed in the sandbox

loool

Have the writers of the problems never heard of our list of things to avoid when writing challenges?

Was just mentioning that a few days ago, lol. Their standards must be a lot lower than ours.

10m

If there is another sea monster I will beat it to death with a whip

I mean, one can easily convert Brainfuck to ASM

Thus, most CPUs run brainfuck, one step removed.

Hmm yeah maybe it's possible with the asm2bf project

Predictions about today's suffering challenge?

@Lyxal reindeer

Yeah, but it'd still be faster if the hardware literally ran brainfuck

@RedwolfPrograms yeah good luck with that lol

I mean, only if you had the resources to build a chip on par with Intell or AMD, but poured those resources into optimising the limited subset of 8 instructions

But even then, the main optimization one would use is... more instructions.

Assuming the chip would be an 8 bit chip (As to handle the math), we'd be using a tiny fraction of possible instruction space

Another option would be to handle multiple instructions simultaneously with "Merged instructions"

"Given a list of seats, output all space cards where the fields aren't allergic to sea-monsters and fit together when rotated 82 degrees anti-clockwise. The answer is the index of the cup immediately left of the adapter with the most joltage multiplied by the three numbers that are found at random places in the weird input format."

There, I wrote a challenge for AOC

That'll probably be day 25

MOAR MATH PLZ

ur input iz big number. fin root of it in base a dquadrillion or sth and put it in box. haf fun

3

CMC: Given two numbers, a and b, Find the digital root of a in base b.

@RedwolfPrograms 05AB1E, 3 bytes

Probably not the right time to post a cmc

lol ^

@RedwolfPrograms mathematica probably has a built-in for that

The right time to CMC is whenever. The beauty of cmcs

We've gone from competitively typing numbers to mocking AOC problems

1m

I think this has progressed rather nicely

prediction: dynamic programming

and by prediction i mean uneducated yolo guess

Something that takes 69 years to run

10sec

GL&HF

> hex

fml

WHY HEX GRID

okay not too bad

A hex grid is just a square grid that's half shifted.

I swear it's athat stupid crab again

wait what counts as east of a hexagon?

Who's solving in hexagony

euler's identity is failing me

30/5

nice

dammit... i typod one character in part 1

Already!??!?!!?!?!?!?!?!?!

i practiced hex grids xd

...oh

because i fucking used 8 directions

lol f

dammit i lost to beta by 1 second

if i didn't make that typo i could've gotten easily like 15/3 or so

what is this input format?/1

it's just annoying xddd it doesn't make the problem any harder

floating point errors and a complex type that can't be rounded itself ftw

hex grids are actually extremely easy to express in complex numbers but i can't really give a hint other than just the answer

Wrong :/

i started so late

hex grids are fake square grids btw

complex numbers are shit

input format is ezpz its litearlly like a9 char regex

Imagine using a language that supports Complex numbers amirite

goddamnit

game of life again ugh

xd

so my prediction was not at all right :P

i am wishing i had maybe implemented an actual grid now

my prediction was wrong as well

@HyperNeutrino it does for me

sorry i mean like. conceptually

it's just annoying format especially for some languages

I have no idea why my answer is wrong lol

It fails the example but my code is practically impossible to debug

send?

In a second

i was about to fix my complex based solution but then i realized that each tick was quadratic in grid size so...

gist.github.com/RedwolfPrograms/…

Oh, it's because I;m dumb

I'm just incrementing and decrementing them over and over again

your grid encoding doesn't work

Yeah, I feel dumb now :p

rip :p

Still not right :/

Did you mean the encoding itself wouldn't work, or just my stupid implementation?

the encoding itself

Oh

for sanity check, NE SE W should reset to 0

Oh hey, I don;t need to work with integers

Got it!

ehh 500/350

good enough for 5 min late start

pog \o/

More cellular automata?

that's not bad. pretty good actually for 5m

@RedwolfPrograms smh

ew but ok

if you don't use integer grid ur just bad

wait what if not int grid then how

Wait, so is the result of p1 the initial state of the cells?

@RedwolfPrograms assuming you're still on part 1, you'll wish you did on part 2

@HyperNeutrino ±.5 is ew

@RedwolfPrograms yes.

@UnrelatedString wrong but ok

i used a 2D array in js for p1 and p2

arr[[i,j]]

my point exactly

im so big brain ngl

not really... if you have any neighbor list and arbitrary access mechanism p2 is the same logic for state transition

i'm completely scrapping my p1 solution because floating point complex numbers are not very good for ceullular automata

@UnrelatedString wrong

i could literally trivially switch to ±.5

since guess what

floats can represent .5 exactly

js moment?

first of all why's that and second of all if you're using a real hex grid with irrationals then that's gonna not work

oh like that

well... js dont have 2d array

all arrays/objs have string keys

and that's the secret

2d array wouldnt work with negative indices anyway

and a dict would work with float tuples fine of course. so no problem no matter what language

wait what does your array store

also does a[2][3] not exist like

js doesn't have 2d arrays ?? what

I mean, it has a[i][j]

But I don't think it has a[i,j]

oh it has jagged arrays but not multidim yea

Keep in mind, indexing nil is an error

oh.

@HyperNeutrino a['2,3'] === a[[2,3]]

huh okay

but what does a[[i,j]] mean in your program

Oh my

it is coords

what else

oh js doesn't have sets does it

Yeah, [[i,j]]just converts to ["i,j"]

it does

I didn't know that existed, that's great.

im just lazy

a[[i,j]] just indicates whether something is black or not?

^^ oh nice

you can't index into sets so i don't see why i would use those

that makes sense tho

eh. i suppose

but like

or well im just lazy is what it is

set insertion, membership, and deletion is O(1)

just like hashmaps so.

An issue arises when one uses that format with strings

My p2 is going to take like 10m hope it;s correct

As l[["a,b", "c"]] is equivilent entirely to l[["a","b","c"]]

oh true. wait so does js even have arrays or are they secretly just hashmaps or wat

Secretly Hashmaps.

Secretly Hashmaps of strings, to make it worse.

._.

i hate this so much

nvm it's getting quadratically slower after each iteration

My p2 is too slow

yikes what how

I;m storing a list of alive cells

Actually, a hashmap with alive/dead

(arrays are just Objects with more information stored in them, they're less efficient than just using an object)

does each iteration give the right ans tho

No idea

I'm going to try a more efficient approach since this one might take a few heat deaths of the universe

Ah yes, the good ol' fashioned "Works theoretically"

ok now I can finally use th input yay

Finished 182/938

The slowdown is constant this time!

And I'm not seeing how one can write a slow part 2 algorithm. I spent little effort optimizing mine and it completed in several seconds.

i can think of a fast-ish way and a slightly slower approach but quadratic speed difference sounds like a bug probably

even if you scan the entire possible space it's linear slowdown i think. but i could be wrong

You'd be amazed how inefficiently I can code if I set my mind to it 🕶

true

Mine gets way slower with more active cells

my code has decided that the example collapses to the point of having exactly two black tiles forever

@HyperNeutrino due to hidden classes it does have arrays. spec wise theyre objects

...I just realized finding the unique items in an array of arrays does nothing at all and I;m iterating through tens of thousands of repeated items

wait not that was my real input

oh okay. js is confusing (to me) xd

@UnrelatedString maybe it does way after 100 iterations

it just completely died on the example

@RedwolfPrograms why th

How'd I guess 53 would be wrong...

75/1038

Frick, I didn't read the GoL rule correctly

twice

or three times

...my hex grid is completely wrong judginf rom that it thinks the example starts with 17 black

How big were y'all's answers? Is 53 reasonable? 530? 5300?

mine was 4k-5k

Mine is between 3k and 4k

Oh, so 53 is probably wrong

i can't get top 10

pain

f

i'm just proud i got #25 yesterday lmao

ey nice

I was relatively fast on part 1 because I knew one of the hex grid tricks

my total global score is 90 now

nice

nice

but truly didn't expect a hex GoL

i did

yeah i knew about how to encode hex grids with the +2 thing cuz uh

but that didn't stop me from making something unsalvageably dumb

i practiced a hex grid problem from a previous AOC when building my template lmfao

+2 thing? where the hell does a +2 come in

east/west is +/-2, 0, and then the other four are +/-1 +/-1

oh like that

...of course

I really need to start actually using paper

yeah. idk if it's actually the best but it seems fine

and I still got the algorithm right pretty early, but didn't realize the GoL rule was wrong until I re-re-reread the instructions, losing 20 or so minutes

and my grid still starts with the wrong number

f

Trivia: in Factor, f is the only falsy object, and it has insanely many overloads for empty containers

part2 yay! :) success = brutally inefficient coding + patience

nifty

In funky2, the only falsey values are nil and 0.

true and false don't exist, which is probably a mistake.

Sanity check: The rule can be written as black_tiles_adjacent == 2 || is_black_tile && black_tiles_adjacent == 1, right?

WRONG

okay i finally stopped not actually moving west at all for the west ones

wait, black_tiles_adjacent is what?

sum of surrounding six?

Wait, really? (That's for if it should become black, not if it switches) (yes to your question)

I'm still confused with the rule apparently

I think it's right then

yeah it took 30 seconds on a piece of paper to figure out why my old encoding doesn't work at all

I tend to sum all neighbors plus self when implementing any kind of GoL, so I usually get something different from others

this is the rare exception because i already had the directions somewhere

Can someone tell me if they see a problem with this?

Can't figure out why it's wrong

It seems to keep giving me less alive cells then there should be

Wait

I just noticed one of the function calls said there were seven adjacent active cells and...oh

duh

uhh

Was it counting itself?

I'm counting the cell if it's alive in the list of alive adjacent ones lol

It looks like it's not the right way to count neighbors in hex

wait. is this happening on an infinite grid?

Yes

sure, why not

aaaa

so you need to expand every turn like you did in previous GoL tasks

Okay, now it's slightly less incorrect

@Razetime it's only 100 steps, so 300x300 is as good as infinite

300x300 is good for correctness, but probably too large for speed

Hexagonal grids have no gliders, that i could find.

Wait, there's a 5 step glider

Which means a tile can only stray ~100/5 tiles away from start

Anyone see any other issues with my code?

I replaced the test function with (o, n) => (n - o) == 2 || o && (n - o) == 1, which fixes the counting itself problem

Still wayyyy to low though

I stupid

Mistyped an array index

@RedwolfPrograms are you sure that formula gives you only 6 neighbours?

I think, as long as my coordinate system thing works how I think

Probably it is counting self?

does AOC usually have this many games of life

I already corrected for that

so, what should [0,0]'s neighbours be? [[1,0],[-1,0],[0.5,0.5],[-0.5,0.5],[0.5,-0.5],[-0.5,-0.5]]?

No, [[1,0],[-1,0],[0.5,1],[-0.5,1],[0.5,-1],[-0.5,-1]]

    [x[0] + 1, x[1]], [x[1] - 1, x[1]],
    [x[0] + 0.5, x[1] - 1], [x[0] - 0.5, x[1] - 1],
    [x[0] + 0.5, x[1] + 1], [x[0] - 0.5, x[1] + 1]

(Ignore the mistyped index on that first line, fixed it already)

And the code is still wrong?

Wait, code was casting to string on the .s(","). Fixed, now it's working for test case

haha I got 38*69*

(I got to 8th place in CGCC leaderboard, just because Redwolf is 3 stars behind)

lol

Not for long, although it's not like it'll move me up very far

YES!

It's a competition between you and me. I don't think either of us will catch pppery

@RedwolfPrograms ah, i see. and [0,0] is a neighbour of itself?

(and you might not get enough points to get 8th place back. The golds are slowly filling up)

Another victory limp to the finish line sort of victory not a waste of time accidental win reason to switch to Python

Well, JS and Python are almost equally usable (not considering libs, which you can solve by creating your own lib for speedcoding anyway)

@HyperNeutrino please do leak the secret imaginary number tech

At least python protects you from spending 40 minutes debugging type conversion issues

You don't need complex numbers per se, you can do just well with number pairs

I finished it, just want to try using imaginary numbers

Plus python has imaginary numbers and arbitrary precision ints by default

complex numbers are just a convenience

A pretty nice one for fast programming

and more of a convenience when you have a sense of orientation than when you're doing something like this

This jay foad guy is really fast with apl

i ended up doing a shit ton of roundtripping between complex and integer pairs and it took like 5 minutes to iron out the kinks lmao

@Razetime That jay foad guy is on APL Wiki

oh wow

no wonder lol

@UnrelatedString so you reworked it multiple times to use both?

And APL can indeed save a lot of time when you know how to solve it

yep, he's 70th global

(in other words, without enough APL problem solving experience, it's a pain)

@Razetime the which

@Bubbler story of my life this year

it's literally just E = 2, W = -2, and then everything else is +/-1 +/-i

which is why I advocate the APL bounty, especially 200+s and solving more complex problems

standard shifted rect grid and then use complex cuz OP

@Bubbler i am trying lol

> He also developed a technique for transposing Boolean matrices with multiple-of-8 dimensions

what

you do E=2, W=-2, NE=1+i, NW= -1+i, SE=1-i, SW= -1-i

lol

@Razetime probably some clever bitwise trick that's more efficient than general transposition

@Razetime It's just a matrix with 8n rows and 8m cols

ah right