Advent of Code

Razetime

01:54

@emanresuA i tried rotation matrices and failed so i just got every permutation of the point and every place that could be negated

then i got 48 different possibilities

emanresu A

02:06

... aren't there 24?

Razetime

02:19

yep

@emanresuA i did not have the patiennce to find the relevant 24

it was fast enough to not be bad even with 2n point rotations

emanresu A

Can you just ignore rotations when matching and align the absolute distances, or does that lead to problems?

Razetime

you need rotations

because you need to figure out what on earth the probe positions are

emanresu A

when matching

As in you just figure out "this set of beacons has absolute differences of [435.32, 1932.4, 953.2], and so does this set, so they must share these points"

Then figure out where #2 is based on that

Can you do that?

Razetime

hm i've no clue

but you can try

idk how you'd align the points based on that

hyper-neutrino

02:39

i don't think there's any way to do it without rotations

and even if there is

it would far exceed the difficult of just doing the rotations

how i did rotations is just for each of the four x-axis rotations, look at all of the four y-axis rotations, and then consider z-axis rotate +1 and -1

and that covers the 24 exactly

there are many ways of going about it

emanresu A

04:40

20m

15m

hyper-neutrino

quiet today huh

tjjfvi

8m

7m

Redwolf Programs

Phew, made it just in time

Had to take care of 5 pets in 20 minutes

tjjfvi

advent of pet ownership?

6m

Redwolf Programs

Well, nobody said I own the pets

Two cats, who are mine and also the rest of my family's, same with one dog, one dog my family's taking care of for a friend, and my sister's hamster

They're gone on a trip

The Fifth Marshal

04:55

5m

Redwolf Programs

This feels directed at me lol

hyper-neutrino

impossible

Redwolf Programs

4

adventofrealizingicanttype

emanresu A

Adventofruntimelongerthancodingtime

(Lanternfish p2)

tjjfvi

@RedwolfPrograms fine, advent of pet sitting

2.75m

hyper-neutrino

04:58

2m

Redwolf Programs

Any time you write awful code, think about the fact that in some slice of time, that code is the culmination of the entirety of the universe's existence

tjjfvi

what

1m

emanresu A

glhfaag

Redwolf Programs

glhf

emanresu A

glhfnr

(good luck have fun ninja redwolf)

Wait that's cursed

p1 always has the same parity on their score

& p2's parity alternates

Unrelated String

05:13

that was
very confusing to read, for some reason

p2 looks fun at least

emanresu A

I was incrementing the dice by 1 each turn instead of 3

I'm not even sure how to approach p2

hyper-neutrino

2/68

that's... a massive diff and a half

emanresu A

Also how do you three-sided-die?

hyper-neutrino

i somehow managed to drop the "roll the die thrice" part from my p2 solution...

cost me about 50pts

i got so caught up in figuring out how to make p2 efficient that i didn't even realize a) i am fucking stupid b) you don't really need efficiency

Redwolf Programs

Wait you don't

Uh oh

hyper-neutrino

05:20

oh yeah i guess that's a hint for p2 now that lb's (basically) closed - you don't really need a particularly smart approach

emanresu A

I've figured it out I think

(Just use a dict with count of every state, right?)

hyper-neutrino

in fact my attempt to make it smart cost me about 5-10 minutes of doing nothing

because it was a completely incorrect approach in the first place

@emanresuA yes, memoization is really all you need

Unrelated String

^

hyper-neutrino

oh

i reclaimed #4

The Fifth Marshal

Done 60/221

Unrelated String

05:26

nice

The Fifth Marshal

I incorrectly interprteted the dice as d2s rather than d3s in my solution

Which took me a while to fix, and another while to manually transcribe in how many universes every total occurs

Total:
3 in one universe
4 in 3 universes
5 in 6 universes
6 in 7 universes
7 in 6 universes
8 in 3 universes
9 in one universe

hyper-neutrino

oh i just did [x + y + z for x in [1, 2, 3] for y in [1, 2, 3] for z in [1, 2, 3]]

Unrelated String

^^^

The Fifth Marshal

That would work, but that idiom didn't occur to me

hyper-neutrino

and did not bother doing multiplication on the substates

i just did memoization

The Fifth Marshal

05:28

Interesting that we came up with so different approaches to the same problem

btnlq

@TheFifthMarshal 1110^3 = 1367631000

The Fifth Marshal

What does that have to do with anything?

Redwolf Programs

Dang it I made a dumb mistake on p2

Bubbler

To save memory, adding a die is equivalent to windowed sums

Redwolf Programs

But I did get top 10 on p1!

hyper-neutrino

05:30

pog!

Redwolf Programs

I'm gonna end up with 9/1000 or something like that lol

Idk what I'm doing wrong on p2, I'm miscounting universes somewhere

Unrelated String

same

Redwolf Programs

Ohhh hang on I think I see where maybe

Unrelated String

I literally lost 5 minutes to unary minus precedence

i was using an unbelievably shit way to work with state

that involved [::-1**t]

and i forgot that had to be [::(-1)**t]

Bubbler

lol

Unrelated String

05:35

nah that had to be more than five minutes actually

Redwolf Programs

Oh my god I'm an idiot

Unrelated String

more like
15 minutes?

Redwolf Programs

I typed 1 instead of 0

tjjfvi

@UnrelatedString JS gives:

Uncaught SyntaxError: Unary operator used immediately before exponentiation expression. Parenthesis must be used to disambiguate operator precedence

Unrelated String

nice

wish python did that lmao

Bubbler

05:37

JS's operator precedence is very sus

Python's too

Unrelated String

1058/513

Redwolf Programs

Okay, finally mine should be working

Bubbler

C one is relatively simple if you ignore bitwise ops: unary postfix > unary prefix > any infixes

Redwolf Programs

Just for fun someone should calculate the results for all 100 possible inputs

Day       Time  Rank  Score       Time  Rank  Score
 21   00:03:33     9     92   00:39:05   664      0

I'm back on the global leaderboard too! :D

Top 10 for p1, on the global leaderboard after days of totally throwing, and then totally threw p2. Kind of sums up this this year's AoC for me so far :p

Bubbler

throw aoc;

3

Razetime

05:42

is this problem as easy as it looks?

Redwolf Programs

Yep

Razetime

oh well then

Redwolf Programs

I overthought p2 then made a tiny type that killed me, but it wasn't too bad either

emanresu A

Help please, my code's infinite looping

Redwolf Programs

You can't use arrays as map keys

It uses the reference to the array as the key

So you'll need to .join(",") them

Actually wait

nvm

You're not using the map how I though you were

emanresu A

05:52

It sets a to a blank map, iterates through states, iterates through all possible dice rolls and adds to those in a, then sets states to a

Redwolf Programs

Wait, actually, you do need to use a string as a map key. The way you're using the map, you'll end up with a zillion copies of everything, all with 0 or 1 as the count

emanresu A

Oh ok

Redwolf Programs

I really wish JS had something like Python's tuples-as-dict-keys

tjjfvi

github.com/tc39/proposal-record-tuple

Redwolf Programs

Ooh, that's cool. Don't like that # syntax, though.

emanresu A

05:56

@RedwolfPrograms Tried that, still broken

Redwolf Programs

(p1 + v + v2 + v3) % 10 || 10

Are you storing the positions 0-indexed or 1-indexed?

tjjfvi

@RedwolfPrograms Better than the other ([|1,2,3|]) IMO

Redwolf Programs

I kind of like how [?...] and {?...} look, [?...] is what one of my languages uses for tuples

Not sure if it would conflict with any existing syntax though

emanresu A

Shouldn't\

Redwolf Programs

@RedwolfPrograms If you're storing them 1-indexed, you need to do ((p1 - 1) + ...) % 10 + 1

Which it seems like you are, since you provide the inputs as 4 and 8 instead of 3 and 7

emanresu A

06:01

Done, still infinite loooping

tjjfvi

It should finish in a universe or two

Redwolf Programs

Can you post your updated code?

Oh nvm

It's there

emanresu A

Yeah

Redwolf Programs

Well it finishes now

Rather quickly, too

console.log(states.size); lets you monitor the progress

btnlq

swap p1 and p2 in 19 and 25 lines

emanresu A

06:08

@btnlq Oh thanks

@RedwolfPrograms What browser are you using?

Redwolf Programs

Chrome 94

emanresu A

Oh wait nvm

It finished, just took about 10s

hyper-neutrino

huh

emanresu A

Except now it's giving the wrong results...

hyper-neutrino

mine only takes around .2 seconds to run and it doesn't use any crazy optimizations just basic memoization

emanresu A

06:12

slowest-code

Maybe I wrote it in Vyxal by mistake

[238441549296251, 238417507056364] for the example input...

hyper-neutrino

send?

emanresu A

Still in the same gist

Redwolf Programs

I've confirmed the Map is working correctly, no weird JS typing stuff breaking that

hyper-neutrino

the amount of code duplication is bothering me

anyway what if you move L3 to L7

actually wait

i don't understand how on earth your player toggle is working

Redwolf Programs

k = a.get(f) || 0
a.set(f,count+k)

This seems incorrect

emanresu A

06:15

@hyper-neutrino It isn't

That might be the bug, thx

hyper-neutrino

just use recursion

emanresu A

IT WORKS!

Thanks hyper

And redwofl

And btlnq

Redwolf Programs

redw on floor laughing

emanresu A

Lol

Well, 1575/1786

Razetime

07:00

hm part 2

bad recursive algo possible?

hyper-neutrino

07:28

definitely

just memoize

Razetime

no memoization

boo

ran a whole of 10 seconds lol

hyper-neutrino

oh

Razetime

i really wanted an array based solution for this.. since it's a dp there should be a way to array-ify it is some form

emanresu A

08:05

¯\_(ツ)_/¯ You could use a 4d array

emanresu A

10:28

CMC: Guess who wrote this:

while um:
    r=l()
    if r:
        print(f'got {r}')
    else:
        print('what the fuck')

user

16:15

@emanresuA ur mom

emanresu A

19:43

@UnrelatedString Mom? Is that you?

Transcript for