advent-of-code

[sullen adder]

now I'm waiting on the submission rate limit because I misread one word ._.

[vast jetty]

at least yall have something to show

ok

i can now find closest pairs

great fucking progress ig.

[sullen adder]

gl :-)

[vast jetty]

guys

why am i stuck on part 1

is that insanely dumb of me

[sullen adder]

i am also stuck on part 1, so no :-)

[mossy basin]

nope! you'll get it!

[humble copper]

todays was really dumb

it wasnt clear up to what point i needed to stop

i got caught up so many times cuz i was doing it for the amount specified in the example, which was different from the amount required for the actual solution

[sullen adder]

yeah, i misread that part at first too 😮‍💨

[paper folio]

aaahhhh i got part one

[humble copper]

at the very least i like that an O(n^2) solution in a lot of cases can be considered fast enough

[vast jetty]

ill go touch grass

fuck the leaderboard for today

ill do it

just later

[vast jetty]

yeah idk everything i can think of requires loops

multiple loops

[uncut mason]

Better than me. I'm just going to ignore day 8 til the heat death of the universe

[vast jetty]

i can think of like an n^3 or n^4 solution which is def not good enough

so fuck it

[vast jetty]

lol im sure its not that bad

i prolly just cant think of the right data structure to use

and am instead forcing functionality onto lists that they dont have

[uncut mason]

It's not clear when to stop connecting nodes

[humble copper]

too much leetcode has rotted my brain. turns out, a lot of problems can be solved by throwing more resources at them

[paper folio]

woot woot I got pt2 too

time to refactor and make it pretty

[worthy horizon]

do you have some samples?

[mossy basin]

it was this, he sent me the generator

[worthy horizon]

oh, like a bomb fuse

[mossy basin]

yeah, it's pretty cool

[worthy horizon]

do you have it as text?

[mossy basin]

i think i deleted the generator, lemme check

[worthy horizon]

smh get more floppies

[mossy basin]

https://gist.github.com/algmyr/c3e0884ad6fdd980634dbd033fcb19dc

all my floppies are on the cloud

had to change the print to go to file though

[worthy horizon]

size 10 makes my code take about 69 ms, size 20 is about 1 second

so it does check out

[mossy basin]

i don't remember what size i used in the visualization... like 15?

[cyan sleet]

and size 20 is only around 245**2 large

for reference for my solution in crystal

10: 0.047s
20: 0.166s
40: 0.586s
80: 2.211s
```linear in the size of the grid

also, kat wtf?

[magic oxide]

@cyan sleet how do you read a line in zig

i have been here for an hour

i'm starting to remember why i added that node to the flowchart last year

[cyan sleet]

fn read_line(reader: *std.io.Reader, line: *std.io.Writer.Allocating) !void {
    line.clearRetainingCapacity();
    _ = try reader.streamDelimiter(&line.writer, '\n');
    _ = reader.toss(1);
}

fn solve(reader: *std.io.Reader, ally: std.mem.Allocator) !struct { u64, u64 } {
    var line = std.Io.Writer.Allocating.init(ally);
    defer line.deinit();

    try read_line(reader, &line);  // Reads one line into the writer
    line.written()  // The line as bytes

    ...
}

pub fn main() !void {
    ...

    // Set up buffered reader for stdin.
    var stdin_buffer: [1024]u8 = undefined;
    var stdin_reader = std.fs.File.stdin().reader(&stdin_buffer);
    const stdin = &stdin_reader.interface;

    const result = try solve(stdin, ally);
}

it's a bit spread out

but at least the read logic itself is in one place

[magic oxide]

ah you read it into the writer

ok

omg it prints

finally

i'm just going to presume p2 is asking for the total number of tachyon beams where each split is a unique beam

[lime fulcrum]

||yeah||

[cyan sleet]

Idris 1 is deprecated.
Idris 2 is still a work in progress

[woven solar]

spent a good 10 minutes today on somehow misreading the problem statement and thinking I need to connect 12 boxes in the test, rather than 10 🥴

[cyan sleet]

As such, documentation is no longer available through the website. If you still require access to the Idris 1 documentation, the following Internet Archive (aka. Wayback Machine) snapshots are available

AAAAAA

[vast jetty]

Alright

I think I’ll get back to it in a bit

[cyan sleet]

god compiling idris2 is slow

oh god, don't run the tests 😭

:(((((((

[magic oxide]

how large were your ints?

[cyan sleet]

u64

[cyan sleet]

finally

[feral hazel]

whoops shadowed a variable lol```
Error: While processing right hand side of main. Can't solve constraint between: length junctionBoxes and length junctionBoxes.

[hollow summit]

you know what
after all, it's better to leave a day unsolved rather than to submit a dishonest solution (ie written by llm)

😃

ill figure out this ||DSU|| thing sooner or later

[cyan sleet]

DSU isn't actually so complicated, it's a very obvious structure that you put some optimizations on top of to make it perform well

basically you have the idea of representatives of a set, to start with everything is their own representative, which we can represent as arrows to themselves

in practice, this is probably just represented as [0, 1, 2, 3]

0 points to 0, 1 to 1, ...

[whole schooner]

https://youtu.be/vq5u09x2Kzo

[cyan sleet]

when we want to take the union of two sets we just re-point one of the arrows, e.g. merging A and B could be done by putting an arrow from B to A (or A to B)

[feral hazel]

-# (non spoiler chat?)

[whole schooner]

https://github.com/cheran-senthil/PyRival/blob/master/pyrival/data_structures/DisjointSetUnion.py

This is my go-to implementation of dsu

[cyan sleet]

To query which set something belongs to you just follow the links until something points to itself

[hollow summit]

how does this look like in code?

like, the re-pointing

[half swan]

union of disjoint sets

[cyan sleet]

it's not really a problem specific thing, so it's hopefully fine

[bright spindle]

Pointing is generally represented by normal variable references

[bright spindle]

||Well, it's super relevant to day 8...||

[whole schooner]

Pretty easy. Just maintain a parent array. For each index i, parent[i] can store the index of the parent of the set containing i.

[cyan sleet]

representative = [0, 1, 2, 3]
representative[1] = 0

the issue with doing it naively is that you can end up with long chains, worst case O(n)

but there is a very simple idea to make it amortized O(log n)

[whole schooner]

Isn't this where path compression comes in?

[cyan sleet]

while during traversal you are figuring out where all of the nodes in the path should end up at the end, so we could just update them to point there (often done in the return part in a recursion)

[magic oxide]

oh god i have to implement a functional dsu

please tell me there's mutability

[cyan sleet]

so if we queried C we would update C, D, B to all point to A directly

and we end up with the now more optimized

so later queries got cheaper

there are other optimizations you can do, but this is enough to get you going

in particular what you can do is to always merge a smaller set into a larger set, this on its own also produces amortized O(log n) operations, but you can combine it with the process discussed above to get something even better

[cyan sleet]

iirc if you drop path compression it should be fairly pure

and you still get O(log n) operations

[magic oxide]

i implemented a dsu in purescript last year as well

i'm getting flashbacks

[hollow summit]

if im understanding correctly, ||each day 8 circuit is a node here, right?||

[half swan]

||each junction box is a node||

[hollow summit]

||what are the circuits then?||

[bright spindle]

||The sets||

[cyan sleet]

||every junction box (node) starts out as its own circuit (set)||

and then you ||merge circuits||

[bright spindle]

||And any group of nodes that are connected by pointers is a set||

[half swan]

||Initially there are as many circuits as junction boxes, and then you start connecting the closest junction boxes together||

[hollow summit]

||so basically i merge all the closest nodes?||

[cyan sleet]

you ||try to merge|| until you have ||successfully merged whatever many times||

[half swan]

||except if they're already connected fully, but using a DSU you don't have to worry about it the DSU logic takes care of that||

[cyan sleet]

you need to check that you don't ||count combining things already in the same set already||

but that's easy

[half swan]

yeah

[cyan sleet]

if ||find(x) == find(y)|| then ||they are in the same set||

[hollow summit]

||how do i actually get the list of sets/circuits? like, iterate through every node and see which other nodes point to it?||

[cyan sleet]

if you add in ||tracking sizes|| for the sake of ||union by size|| you already have the information you need

but you can also just ||iterate over all the nodes and count how many times find(x) appears is to get the sizes of each component||

i.e. you could do something like ||Counter(find(x) for x in nodes)||

[woven solar]

#advent-of-code message
evergreen message

[magic oxide]

has anyone finished idris yet?

[cyan sleet]

no

[feral hazel]

no

[magic oxide]

understandable

[cyan sleet]

I haven't managed to read the input

I can read one line and parse that

but finding docs is...hard

[feral hazel]

i've read it but i'm having issues with fins and nats and vects and lists

and imports bringing everything into scope

[magic oxide]

how do you navigate the docs

where are the arrows

[woven solar]

If you run out of memory while attempting to link Idris, try using ld.gold
wow

[feral hazel]

i'm mostly using these
https://www.idris-lang.org/Idris2/base/
https://idris-community.github.io/idris2-tutorial/

[magic oxide]

ok thx

i was using the first one, but i was on a specific page and i didn't know how to get to the home page

[feral hazel]

||https://idris-community.github.io/idris2-tutorial/Tutorial/Traverse.html|| might help if you can already parse one line

[cyan sleet]

you know, maybe I should be lazy and just use their csv module

[feral hazel]

fair lol

[cyan sleet]

the idris2 repl is so bad compared to the original

[magic oxide]

oh

i haven't even looked at the puzzle yet

[cyan sleet]

not even readline integration/completions

[magic oxide]

i have syntax highlighting at the very least

[feral hazel]

i just want a repl that accepts arrow keys 😭

[cyan sleet]

the idris 1 repl supported that

and tab completion

[feral hazel]

implementation Show JunctionBox where
  show = joinBy ", " . (\[x,y,z] => [x,y,z]) . map show . components

[woven solar]

I tried installing idris1 to see what y'all are struggling with, but the suggested solution cabal install idris only works for cabal <= 3.4, which is too ancient to get from apt 🥴

[cyan sleet]

idris1 is deprecated, don'cha know?

[woven solar]

it has been deprecated for like 4 months!

[feral hazel]

there's a toList lmao i thought i tried that lol

[magic oxide]

tf does Error: Undefined name unsafePerformIO. mean

am i performing unsafe IO or do i have an undefined variable name?

[half swan]

I have decided to join you in suffering

[cyan sleet]

I should probably install the lsp for my own sanity

[magic oxide]

same

[feral hazel]

same

[magic oxide]

the only valid program i have written is hello world

[cyan sleet]

note for others: the plugin espects a idris2 project idris2 --init

[magic oxide]

i have managed to read and print a file

i already have this many tabs

ok for my own sanity i'm just gonna quickly do a python sol first

[terse mesa]

Today is crazy, I just got my first star and have been working on it since half past nine.

[hollow summit]

congratulations!🎉

[static lily]

what's this about

[half swan]

https://adventofcode.com/ a little bit of fun

[tender swallow]

a series of programming puzzles that runs every December. we have a dedicated space in this server to discuss them and (if you want) have some friendly competition among ourselves.

[static lily]

oo

interesting

[magic oxide]

cursed syntax highlighting

[half swan]

deep inhale

[hidden musk]

compiling hole

[hollow wharf]

WAIT

I JUST WOKE UP

AND ITS A 3d PROBLEM

LMAO

[half swan]

the 3d bit isn't the major thing

[hollow wharf]

ohh reading more yeah lets ee

[half swan]

||all you need to know is to compute straight line distance||

[magic oxide]

yeah

it could be 4d and it'd still be like the same problem

[hollow wharf]

okay

now that i have read it

it doesn't really matter

its just distance between two

then make circuits of two pair

[woven oriole]

Mfw by the time the next Sunday will come aoc would have ended alr

[unborn carbon]

lmao

i forgot about thus

this

[terse mesa]

I should have been studying for exams, but priorities have to be set

[frigid mural]

so real, I had an exam this morning that I did AOC ahead of, and exams tomorrow and the day after (and Friday) but here we are

[half swan]

TIL there's comments in the problems 👀

[hollow wharf]

wt d fuck

[hollow wharf]

how do one even see it lmao

[half swan]

highlighting the text

not all text has it

[half swan]

try hovering over it

[blissful granite]

İm code

[hollow wharf]

i didn't work

but i found it regardless

[half swan]

chromium or gecko?

this might be the one time there's a chromium L

[woven solar]

it doesn't highlight for me either, weird. maybe it's disabled until you solve aoc2025?

[hollow wharf]

firefox actually

[half swan]

damn

I'm on zen

not the engine's fault then

[hollow wharf]

i did solve it though

and its not

[minor cave]

The CIA wishes it had our level of AoC redactions

[hollow wharf]

this is from day 7

nice easters lmfao

[half swan]

can we have the epstein files?
we have epstein files in pydis
epstein files in pydis:

[woven solar]

It looks like it's missing in the styles. Back in 2023 there was a CSS style for article [title] highlighting those, and in 2025 it's missing.

[hollow wharf]

this is actually so funny

LOL

[viral grotto]

For day 2 are the examples correct for part 1?

[hollow wharf]

yh

[magic oxide]

@minor cave i'd like to say
me from 2 days ago kinda wished to have some functional languages in the roulette
i would like to retract that wish, functional languages are terrible except gleam

[terse mesa]

I still have exams until the 22nd, so it's probably good for my grades that it's only 12 puzzles.

[woven solar]

oh, it's an inline stylesheet. so probably it's intentionally not highlighted, and will be once you solve aoc 2025

[viral grotto]

95-115 has one invalid ID, 99.
998-1012 has one invalid ID, 1010.```
How so 998-1012 has 1010????

[hollow wharf]

day 1 comments,

[hollow wharf]

between 998 and 1012

comes 1010

its a range

[viral grotto]

Ohh okok thanks

[hollow wharf]

this is nice

lmao

[viral grotto]

Wait was that part of the puzzle??

Or I am just thick

[half swan]

yes

[viral grotto]

Don't understand this but will try after

[half swan]

wait what are you asking about

[hollow wharf]

its like comments hidden but yeah

[half swan]

the range or the comments

[hollow wharf]

around this part

[hollow wharf]

comments

[magic oxide]

relatable
i have an exam in 6 hours and i'm sitting here idris'ing

[hidden musk]

what stickie actually meant was functional functional langs

[viral grotto]

The range

[half swan]

I gave up :)

[magic oxide]

if you give me polyml i'd be happy

[hidden musk]

ocaml 👍

[magic oxide]

that's about it

[half swan]

idris was insistent that a List of points was actually a singular point and not a list of points

[magic oxide]

i have about 2 hours of experience with ocaml

[tender swallow]

they did mention that these were ranges a few times...

[viral grotto]

Lmaoo

[magic oxide]

i would give up but it's been 3 hours

[half swan]

https://tenor.com/bOyec.gif

[magic oxide]

and i'm almost there

[minor cave]

Lmfaoooo, I was tempted to add Gleam into the rotation. I might still.

I think I've managed to pick really incompatible languages with the puzzles for the day except for 1 day

[hidden musk]

which 👁️

[half swan]

wait how'd you solve p1 without this knowledge nvm I misread

[hidden musk]

or, maybe it's that the languages you picked are incompatible with most problems 🥴

[minor cave]

:)

[vast jetty]

still not done p1

[minor cave]

I think y'all thought the crystal day was fine? Or maybe the zig one?

[vast jetty]

alright man

[minor cave]

I think C would've been fine for most other days... except the day rolled it for

[vast jetty]

i cant do this one apparently

i suck today

[hidden musk]

ah yeah zig was kinda chill. parsing was easy, so it was just fast. no complex data structures needed

[magic oxide]

zig was not chill for me

i struggled with a lot of docs being outdated

once i got past parsing it was fine

me fr

[vast jetty]

why cant i do today's problem?

granted ive only been at it for like

2 ish hours

but not evn gettting part 1 is horrendous

[minor cave]

Problems get harder

[hidden musk]

yeah. solving the problem itself was fine

[vast jetty]

eh i did yesterday[s immediately after getting up while literally brushing

not making much progress with today's is

:<

[magic oxide]

sometimes you might just not be familiar with a specific algorithm/data structure

bruh

[vast jetty]

hm yeah

[woven solar]

i also made this mistake tbh. I did for a,b in data: for x in [a,b] and confusedly stared at the problem description when it didn't work

[hollow wharf]

bruh

😭

[vast jetty]

yesterday was quite easy

at least part 1

[hidden musk]

me too 😔

[hidden musk]

I'm glad we're finally numbering our debug prints

[vast jetty]

the issue for me was forgetting to include the end point

[magic oxide]

i've always numbered my debug prints

gotta know which joever it was

[vast jetty]

wtf is joever

[magic oxide]

tho this is the first time i've actually got a printed joever

[viral grotto]

For part 2 of day 2,
In
1188511880-1188511890
1188 is repeated twice in all numbers in this range!

So y only consider 1188511885?

[magic oxide]

it's not just any substring is repeated twice

the entire number has to be some number repeated twice

or actually if you're talking part 2 then not necessarily twice

[viral grotto]

Alr alr that's make it easier(not really )

[magic oxide]

but the entire number still has to be one specific number, repeated some number of times

[vast jetty]

(at least twice)

(one doesnt count as a reptition)

[sharp elk]

it does though

[viral grotto]

only took me half an hour

and the solution sucks ass

should have used regx or something

[vast jetty]

throw it into #1445278429372219422

alr ooookay

im getting somewhere with this crap

[viral grotto]

ohh mb is this spoiler..

[vast jetty]

lol

accidentally pasted my fucking input in

:D

ok fuck me

okay what could i be doing wrong

it works on the example input

[fossil vigil]

Bro can anyone help me

[vast jetty]

is there a thing today with confusing example input

[fossil vigil]

#career-advice

Probably i sent the wrong channel

Sory

[magic oxide]

not really, do you wanna move to #aoc-solution-hints?

[woven solar]

well, today I misread 10 as 12 somehow

besides that, not really

[magic oxide]

oh I guess make sure you're doing 1000 and not 10

[tender swallow]

make sure you're changing the number of connections to make
it's 10 for example and 1000 for real input

[vast jetty]

i did do thaaat hm

[vast jetty]

sure

[tender swallow]

failing that there is more debugging you can do
I have a test case posted in #aoc-solution-hints

[vast jetty]

ok lets try that

[iron roost]

wasn't aware day one resets the inputs it gives you if you get it wrong😭

i was trying for so long and was sure code worked

[vast jetty]

it doesnt?

[iron roost]

mine was different i swear

maybe i just copied it in wrong the first time

i used ctrl A but i guess it didn't copy it in right cause the second list is way longer i've just noticed

[woven solar]

I wonder if duckdb is in the language roulette

or, like, SQL

[iron roost]

can sql solve these i'm not sure of its limits

[vast jetty]

trust me

theres people doing day 7 in excel

[iron roost]

my mouth dropped

what am i looking at

[half swan]

day 7

it's not as intimidating as it looks

[half swan]

sql can probably solve some of these

[iron roost]

wouldn't even be surprised i saw someone code minecraft in power point last time

[half swan]

there are technically no limits in (modern) sql, it is turing complete

but I don't know how feasible it would be to solve any given problem practically

[iron roost]

thats interesting something i'd have to look into more at some point since i only really know its most important basics (even though i'd never attempt these tasks in sql lol)

[half swan]

procedures might help a fair bit

[half swan]

has been going for a while now

[hollow wharf]

Ah

[half swan]

since 25/11

[hollow wharf]

Dang

[vast jetty]

some great commit names today

[hollow wharf]

Works on example

But fails on actual

😭 ☠️

[hollow wharf]

What’s ur approach

[vast jetty]

follow the saga in #aoc-solution-hints

[naive wharf]

can anyone spoiler p2 for me?

[iron roost]

truly in the trenches with "bleh" 😭

[vast jetty]

in any case ill post both parts when im done

[half swan]

how did you do p1

[tender swallow]

you solved p1?

[naive wharf]

I just mean the question, sorry

I didn't do p1 yet

[tender swallow]

it's not a huge leap at all

[magic oxide]

||continue joining, find the first connection that causes everything to be connected||

[tender swallow]

once you do p1 you should be able to do p2 directly

[naive wharf]

good, so ||actually tracking 'groups'|| is the way to go

[half swan]

yes

curious, what other ideas did you have?

(I can't think of another option)

[naive wharf]

I should write my python like I do my C...
"objects? no! index hell."

[naive wharf]

ok I did that
||```py
i2c = [] # index -> coords
c2i = {} # coords -> index
i2g = [] # index -> group
g2i = {} # group -> indices

[half swan]

i2c will never not be Inter-Integrated Circuit in my mind lmao

[naive wharf]

OK I actually do not know why I made c2i

[frigid mural]

ohh that's what it stands for

No idea why I'd never actually checked despite using it

[tender swallow]

<@&831776746206265384>

[frigid mural]

!compban 1443489344273776836

[marsh current]

:incoming_envelope: :ok_hand: applied ban to @final mesa until <t:1765564442:f> (4 days).

[naive wharf]

what happened here

[tender swallow]

crypto spam got removed

[vast jetty]

ok so

part 2

we're continuing to pair closest points?

until we get one contiguous circuit?

im not sure what we're pairing lol

is it asking what the last closest pair is

or what pair closes the already mapped circuit?

[hidden musk]

yes

[vast jetty]

ok

lets see

bruh

my code thinks everything is connected in example input after pairing 23 numbers

which is ok ig

but then it thinks the list formed has like 18 elements

which isnt believable

idk whats going on

argh

this shouldnt be hard at all

[rapid wolf]

wow Zig was an experience, I can see why this was chosen for that one financial database and why it is as niche as it is.

at least the day itself wasn't ||too difficult||.

[dense osprey]

part 1 is just ||dsu|| right?

i feel like the only solution i can think of is extremely suboptimal

i dont want to implement it

[hidden musk]

every day uses ||data structures and ulgorithms||. but yes

[vast jetty]

oh

hm

ok

lets see

[sharp elk]

300 ms solution for both parts is good enough?

Probably the sorting is the slowest step

[vast jetty]

tfw the discpline that is made up of ||data structures and algorithms|| requires ||data structures and algorithms||

||ulgorithms*||

[frigid mural]

yeah, naive sorting took about 300ms for me too

[sharp elk]

are y'all making your own implementations?

[frigid mural]

I have in the past, this time I just borrowed scipy's

[sharp elk]

same

laziness at it's peak

[vast jetty]

ok

ffs

im gonnna be here a while

come to think of it, shouldve written a ||binary search||

ykw

i cant be fucked to write one

imma do it manually

works

[tiny crater]

Hey guys Im new to Python (and coding lets say). Im doing the 1st task and Im stuck on how to read the text from each line.
I made an if statement using an array to read whether to move left or right.
But now Im stuck with how to read the numbers after since they vary from 1-3 digits.

[vast jetty]

we typically read from a txt file

you can use with open

then you have methods like read(), readlines()

stuff

this is what i have as a basic template in my aoc files:

with open('d4.txt') as input:
    puzzle_input = [i.strip() for i in input.readlines()]

then you can do whatever you want in there

add your stuffs

lets say you wanna split at '-', you can use ||the .split() method||

my parsing for d1 is:

||```py
with open("d1.txt") as input:
puzzle_input: list[str] = [i.strip() for i in input.readlines()]

instructions: list[tuple[str, int]] = [(i[0], int(i[1:])) for i in puzzle_input]

its stored like that just cuz its convenient

for me

https://tenor.com/view/i-hope-that-was-helpful-corey-vidal-i-hope-that-helps-you-hope-that-was-useful-to-you-hope-that-was-of-value-gif-20766062

[tiny crater]

Yes thx I was looking for this
||(i[0], int(i[1:])||

[vast jetty]

there you go

do spoiler it tho

double brackets

are those brackets

[tiny crater]

yes thx

[vast jetty]

pipes

whatever

[austere kettle]

am i understanding correctly that for star 2 we need to start from the circuits we got in the end of star 1? (after connecting 10 first pairs). And then continue until we get one circuit?

day 8 (today)

[vast jetty]

eh

sure

point is the rules are the same

[frigid mural]

Yeah, you can restart the computation if that really tickles your fancy but otherwise...

[vast jetty]

just keep pairing

[austere kettle]

but then isnt the first connection already exactly one circuit?

just A--B

[vast jetty]

no

[frigid mural]

it sure isn't a circuit that spans all the points

[vast jetty]

the rest are unconnected

[austere kettle]

ohhh

i thought all points we already have

alright

thanks

[vast jetty]

good on you for finding that tho

[austere kettle]

thats why i thought we need to start from end of 1

okay thanks guys

[vast jetty]

im sure some people had that in their first iteration and had to change their code

i forgot how to write ||binary searches|| tbh

[frigid mural]

what would you be ||binary searching|| over?

[vast jetty]

just ||run the computation a bunch of times, and see when the first time it returns a single connected list is||

[austere kettle]

okay nice, got the second star

thanks again guys

good night

[vast jetty]

otherwise its tooooo many iterations

at least the way i implemented it

i had to do it manually tbh

but for final code i should write the thing

i wonder how yall did it without that

[frigid mural]

How would ||you be able to skip computation in this binary search case||?

[vast jetty]

ig youd have faster connection making

[vast jetty]

well youd just have to do fewer.

[frigid mural]

I don't see any way of ||skipping iterations|| here

[hollow wharf]

Will roulette ever land on python

😭

[frigid mural]

If you were here for last year, you would know 👀

[hidden musk]

the invisible hand of the kat has been known to manipulate the wheel on certain days

[hollow wharf]

Lmfao

I was not

[frigid mural]

It stops being stockatstic, some would say

[hollow wharf]

😭

[minor cave]

:)

The house always wins

[vast jetty]

not for the computations

for how many times you have to do it

ok so i just ||have a function that does the bloody computation for n pairs||

[hollow wharf]

LMFAO those are wise words

[frigid mural]

yes that's what I'm referring to

[hollow wharf]

It’s everywhere

Today for me

[vast jetty]

and i didnt wanna modify the function to ||keep pairing until everything is paired||

[vast jetty]

so like for part 1:

[frigid mural]

Ok, but then if you do that your runtime gets an extra log(n) tacked on for fun

[vast jetty]

now i can just ||binary search over the possible number of pairs|| and do that

im doing a lot of repeat computations tho yea

[vast jetty]

yep

ig its cuz of how im handling the pairing

alr lemme try to change it-

very scary

[vast jetty]

presenting more great commit names

[hollow wharf]

LMFAO

[frigid mural]

total sanity loss, a tale in 6 acts

[vast jetty]

compare this to yesterday's

seriously...

[bright spindle]

I guess it's pretty expected depending on how you implement the solution

that part 2 would be fast

[warm gazelle]

i might be missing something but in day 8 part 1 why do they only make the 10 closest connections? for the actual input data do i make the 500 closest connections (n / 2) or only 10?

After making the ten shortest connections, there are 11 circuits

[ivory vigil]

It's an arbitrary limit. 10 for the example, 1000 for the real input.

Your list contains many junction boxes; connect together the 1000 pairs of junction boxes which are closest together.

[warm gazelle]

thanks, not sure how i missed this

[humble copper]

that difference should have been bolded i think

like how a lot of other important info on the prose is bolded

cuz it took forever to actually find that

[ivory vigil]

The 1000 is in bold

[humble copper]

was it?

didnt seem it when i tried it last night

[hollow wharf]

💀

[humble copper]

💀

i might just be blind then

[hollow wharf]

tbh it was night

might as well be cuz of late

[ivory vigil]

I believe I'm legally blind in my left eye!

Oh. nvm. I don't meet the legal US requirements.

[hollow wharf]

bruh

[ivory vigil]

Visual Acuity: 20/200 or worse in the better eye with corrective lenses.
I was told (incorrectly) that if you need -10 corrective lenses, you're legally blind. Google says it's the corrected vision that counts, not the without-glasses vision.

I also have no idea how 20/200 translates to prescriptions.

[hollow wharf]

isaac put on the glasses because

[ivory vigil]

60 seconds!!

[hollow wharf]

AOC is just about to start

YEAH

LMFAO

[ivory vigil]

lol. I always have glasses on unless I'm sleeping

[hollow wharf]

of course you would not have it when u going to sleep 😭

no shit

https://tenor.com/view/sad-kitten-sad-cat-sad-kitty-kitten-kitty-gif-12913660166773409941

[hollow wharf]

i hope there is no more parsing

i really don't wanan do it tonight

im too sleepy for it

[ivory vigil]

(Or swimming or in the shower)

I sorta liked the parsing. But they're less common.

[hollow wharf]

Im guessing

more algorithmic

problems

[ivory vigil]

5 seconds!

[shut trellis]

<@&518565788744024082> Good morning! Day 9 is ready to be attempted. View it online now at https://adventofcode.com/2025/day/9. Good luck!

[hollow wharf]

and there

bruh

this is a path finding thing

[ivory vigil]

Uh.... p1 was easy.

And now I'm stuck.

[magic oxide]

is the carbon compiler even finished

[dense drum]

fr 💀

[vast jetty]

well

[dense drum]

easier than some of the previous ones imo 😭

[vast jetty]

oh well

p2 is interesting

[dense drum]

fr

[vast jetty]

i shit you not

[dense drum]

LMAOO

[vast jetty]

one of my best times

ignore yesterday uh

that was bad

[dense drum]

me too 😭

ignore me not doing part 2 😔

[vast jetty]

at least i got it done

[dense drum]

real

got busy with stuff but i want to go back and do 7-8

[vast jetty]

my part 2's can take some time

but still only around 2 hours

except stupid d8

[dense drum]

LOL 😭

[minor cave]

From my brief look into it there should be a stable enough compiler for it. If not I have a backup language

[magic oxide]

i shall look into it after i finish my exam

though someone else will hopefully have done that before then

[mossy basin]

never even heard of carbon

[half swan]

google's c++ successor thingy

Ultimately, the largest rectangle you can make in this example has area 50. One way to do this is between 2,5 and 11,1:
am I crazy or is this wrong?

!e print(abs(2 - 11) * abs(5 - 1))

[marsh current]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

36

[magic oxide]

you're ||off by one, there's 11 - 2 + 1 tiles between 11 and 2||

[half swan]

If I had a nickel for every time I hit an ||off by one|| I'd be rich

[mossy basin]

inclusive ranges, amirite

[hidden musk]

i am a fan of this backup language

[vast jetty]

||If I had one (1) nickel for every time i hit an OBE, I'd have no nickels||

[pine tiger]

get WA
add additional conditions to shrink answer space
get higher answer than before
AC

[uncut mason]

Damn giant inputs 😝

[feral hazel]

oh wow carbon version 0.0.0-0.nightly.2025.12.09 released 3 hours ago

[hidden musk]

😔

[feral hazel]

enable execution :D

[hidden musk]

oh right, forgot the error message:

<source>:2:16: error: `package` declarations must end with a `;`
package sample api;
               ^~~

<source>:2:1: error: semantics TODO: `handle invalid parse trees in `check``
package sample api;
^~~~~~~~~~~~~~~~~~~
Compiler returned: 1

but apparently it just works if i don't modify anything 👍

[vast jetty]

alrighty

we're getting somewhere

[humble copper]

it hurts so much to see so many libraries that do exactly what i need to do to solve part 2

and force myself to not use any of them

[vast jetty]

!e

print('..............'.count('.'))

[marsh current]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

14

[vast jetty]

always

ok

great to have a visualisation

[humble copper]

new york?

[sullen adder]

wow, there's only three days left 😔

[vast jetty]

lets see if im almost done

[viral grotto]

#1445278429372219422 message
How could I've made this better

[vast jetty]

whoops

[woven solar]

now trying to solve day9 in julia

once again I try julia and once again I have to wait multiple seconds for the REPL to load, etc. I wonder if I'm doing something wrong

[vast jetty]

!e

print(len('.......X.'))

[marsh current]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

9

[vast jetty]

ok

lets see if it works with actual input tho lol

[naive wharf]

let me guess
part 2 has something to do with counting the tiles in the rectangle

[vast jetty]

ok not confident at all

yeah no ||floodfill|| just wont work ig

ok im gonna think about the part 2 later

[cyan sleet]

Oh shit, we're doing carbon?

[hidden musk]

yeah it's honestly not that bad

[woven oriole]

fuck variable shadowing

i am running out of variables and shadowing like 1e9 times and not realizing that i am shadowing

[woven solar]

ERROR: MethodError: no method matching filter(::var"#star##6#star##7"{…}, ::Base.Generator{…})
I'm doing Julia instead of Carbon 🥴

[terse mesa]

Before I throw myself at todays puzzle, will I need as much time today as I did yesterday?

[ivory vigil]

More

[terse mesa]

well, I don't have much time today, but in that case, the rest will have to wait.

[feral hazel]

https://carbon.compiler-explorer.com/z/fq95f31jd segfault!!!!!

no consts for you

[vast jetty]

more prolly

[terse mesa]

The first one was easy, but the second seems to take a while.

[terse mesa]

this what the first one looks like

[cyan sleet]

Damn on-call pages blocking AoC work

[summer tusk]

yippee got my python solution down to 30ms

for both parts + parsing

[hollow summit]

while my very efficient part 2 solution runs... how are y'all doing? 😃

my solution got slimed by oom killer

umm.... time to think of another solution

[naive wharf]

Am I correct in assuming the 'curve' never intersects itself?

[bright spindle]

||If you mean the boundary, yes.||

[naive wharf]

Hmmm... I see a n² log n solution I think

[terse mesa]

it took me 6 fucking hours

[sharp elk]

Idk why I assumed for the first part that if dist btw 2 points is maximum area will be maximum too , corrected it and part 2 has shocked me too much

[vast jetty]

hahahahhaa

indeed

i think i have a working algorithm in mind

lets see

might go out to eat rn so ill implement it later

[sharp elk]

Please explain

[magic oxide]

are you doing tuple(map(int, ...)) somewhere in your parsing code?

[sharp elk]

yes

My inputs are basically a list of pairs of 2 coordinates ((x1, y1), (x2,y2))

[magic oxide]

yeah, so when you do that, python doesn't know that your input specifically has 2d points, so the type is a tuple of unknown size of ints

[sharp elk]

oh understood so I've to change the parsing

[magic oxide]

i usually just put a pyright: ignore[whatever the rule is] in my parser for that

[magic oxide]

no the code is fine

the type checker just doesn't know your input looks like that

usually you try and narrow it down as early as possible

so I'd probably do like points: list[tuple[int, int]] = ... and then ignore the type warning there

[sharp elk]

k Thanks!

[warm berry]

!compban 1357282860972642314

[hollow wharf]

<@&831776746206265384>

[marsh current]

:incoming_envelope: :ok_hand: applied ban to @errant saddle until <t:1765643797:f> (4 days).

[hollow wharf]

Why are they banned for 4 days ?

[warm berry]

Compromised account

[hollow wharf]

No I mean why not for indefinite

Lmao

[warm berry]

If they regain control of their account they can re-join. It is very rare for them to join the server while still compromised

[hollow wharf]

Ohhh I see makes sense

[hollow wharf]

Are most of the scam accounts nowadays just normals peeps

Compromised

[warm berry]

Yeah

[hollow wharf]

Dang

[half swan]

is it just me or has the number of compromised accounts seen an uptick in the past week or so

[warm berry]

Probably coincides with holidays time off and increased activity on discord in general

[half swan]

hm, makes sense

[naive wharf]

this is how my data looks

I haven't done p2 yet aaa

can someone tell me whether the following is true:
||a rectangle is valid for p2 if at least two corners are red tiles, AND there is at least one red tile each that is

• to the top-right of the rectangle
• to the top-left of the rectangle
• to the bottom-right of the rectangle
• to the bottom-left of the rectangle
  where I count the red tiles that are just on the edge of the rectangle, and therefore at least two of these are fulfilled by the two corner red tiles||

[half swan]

||the other 2 corners need not be red at all they can be green too i.e there is a red tile somewhere if you extend x and y outward||

[naive wharf]

that's what I mean with 'to the x-y'

[vast jetty]

Prettt

Pretty

[half swan]

ah I misunderstood

yeah that's correct then

[naive wharf]

funny, I made a lucky guess with the scaling, all numbers are ||between 0 and 100.000||

[vast jetty]

Huh

I was getting stuff ||over a million||

Must have made some stupid error

[naive wharf]

I mean coordinates

but you can also obviously see the final number must be under 10.000.000.000

[vast jetty]

No i mean that too

I was trying to render

And it said out of range

Hmmm

Index error

[sharp elk]

how is 0.7 sec for day 9 part 2?

[blissful granite]

Whats that

[sharp elk]

time of code running

[blissful granite]

Hmmm what does code mean

[sharp elk]

you're trolling ryt?

[blissful granite]

Yes im trolling

[lyric sigil]

darn.

i might not finish day 9

[vast jetty]

goddamn great code

[humble copper]

🥹

3 more days to go :D

[bright spindle]

That's what I got as well

[signal summit]

guys parrt 2 i cannot get

day 1 part 2 i'm failing

[vast jetty]

holy shit dude

my code is gonna take forever to run

it gets the right inside for example input but itll take forever for actual input

idk if i should let it run

we're talking about half an hour here maybe

[cyan sleet]

@hidden musk how build a carbon binary?

[hidden musk]

there's an example on the github, you have to manually use the linker

[cyan sleet]

😭

[hidden musk]

#/usr/bin/env bash
VERSION="0.0.0-0.nightly.2025.12.09"

~/carbon_toolchain-${VERSION}/bin/carbon compile --output=generated.o generated.carbon \
    && ~/carbon_toolchain-${VERSION}/bin/carbon link --output=generated generated.o \
    && ./generated

[vast jetty]

ok i made it slightly faster

now itll take just some mins

hopefully its not wrong

[cyan sleet]

> bash build.sh main.carbon
ld.lld: error: undefined symbol: _CRange.Core
>>> referenced by main.carbon:13
>>>               main.o:(_CMake.Sieve.Main)
>>> referenced by main.carbon:13
>>>               main.o:(main)

ld.lld: error: undefined symbol: _CInclusiveRange.Core
>>> referenced by main.carbon:34
>>>               main.o:(main)
error: linker command failed with exit code 1 (use -v to see invocation)

do I need to reference the lib directory manually?

[hidden musk]

show code? though honestly i have no idea

[bright spindle]

I solved AoC 2015 day 19 part 2, so now I'm all set for ||this year's path-finding||

[cyan sleet]

I copied one of their examples and modified the "main" function

// Part of the Carbon Language project, under the Apache License v2.0 with LLVM
// Exceptions. See /LICENSE for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception

import Core library "io";
import Core library "range";

// Compute and return the number of primes less than 1000.

class Sieve {
  fn Make() -> Sieve {
    returned var s: Sieve;
    for (n: i32 in Core.Range(1000)) {
      s.is_prime[n] = true;
    }
    return var;
  }

  fn MarkMultiplesNotPrime[ref self: Self](p: i32) {
    var n: i32 = p * 2;
    while (n < 1000) {
      self.is_prime[n] = false;
      n += p;
    }
  }

  var is_prime: array(bool, 1000);
}

fn Run() {
  var s: Sieve = Sieve.Make();

  var number_of_primes: i32 = 0;
  for (n: i32 in Core.InclusiveRange(2, 999)) {
    if (s.is_prime[n]) {
      ++number_of_primes;
      Core.Print(n);
      s.MarkMultiplesNotPrime(n);
    }
  }
  Core.Print(number_of_primes);
}

[vast jetty]

made it even a bit more faster

[cyan sleet]

maybe their examples are outdated?

[vast jetty]

still gonna take forever

IT MIGHT BE WRONG lol

[hidden musk]

oh it's the range. i'm pretty sure that doesn't exist. @feral hazel played around with the ranges more i think

[cyan sleet]

do you have some minimal example file that works? 🥲

[vast jetty]

istg if after all this minutes of runtime its wrong ill cry

[hidden musk]

https://carbon.compiler-explorer.com/z/1PbKcnsen

[cyan sleet]

also, why does Print not accept a string?

[hidden musk]

https://github.com/carbon-language/carbon-lang/blob/trunk/core/io.carbon

[cyan sleet]

mfw no generics

[hidden musk]

oh actually, there is ReadChar, so maybe i don't need python today at all 🤔

[cyan sleet]

also, the language server is ass

that or I messed up some part of the setup

all the errors

[feral hazel]

oh yeah i couldn't get the built in range functions to work but the constructor for the actual object they create does work

https://carbon.compiler-explorer.com/z/3M483sr7a

[cyan sleet]

https://github.com/carbon-language/carbon-lang/blob/trunk/examples/advent2024/io_utils.carbon

[hidden musk]

oh these look helpful

[feral hazel]

ah the lang should have options but it still overloads the output lol
https://carbon.compiler-explorer.com/z/41YMeEcvd

todo

[hidden musk]

how is it 255 when it's an i32 🥴

[feral hazel]

oh what i didn't even notice that
it looks like it's converting the output to a u8

[vast jetty]

fucking shit

[hidden musk]

it must be a u8 at some point in the function itself i guess

[cyan sleet]

does carbon have dynamic arrays implemented?

I'm guessing no

and no type inference? 🥲

[vast jetty]

https://tenor.com/view/absolute-cinema-raccoon-absolute-cinema-gif-17862327649353748812

[vast jetty]

is 88 seconds good

[bright spindle]

It was my initial time.

But I got it down to 600 ms.

So...

[vast jetty]

..

insane

idk where my lag is

but im having to ||import cache|| so it couldnt be good

[bright spindle]

I used pyinstrument to profile, it's pretty neat

I also made some algorithmic changes

[vast jetty]

profile what

lemme just post the bd code

[bright spindle]

Profile the code, to see where most of the time is spent

[vast jetty]

oh

i cant even post

lmfao

too long

[bright spindle]

Pastebin

Or Github

[vast jetty]

yeah

[vivid dove]

part 2 day 9 takes so long to compute loll

[ivory vigil]

Depends how you did it. 0.4s for me.

[rancid creek]

my first pass involved calculating all red/green points and took like 5 minutes to run
my second pass was a 1 liner ||using shapely|| and finished instantly. 🙁

[covert arrow]

I got mine down to about 170ms

[cyan sleet]

60ms, but not python

I could probably get it down a decent amount more in a language that isn't kinda pain

[covert arrow]

I rewrote it in rust and it runs in ~9ms

[rancid creek]

I wrote mine in python and then goofed off. 🙂

[shut trellis]

<@&518565788744024082> Good morning! Day 10 is ready to be attempted. View it online now at https://adventofcode.com/2025/day/10. Good luck!

[half swan]

today might be one I'm not able to finish right away 👀

[vast jetty]

morning

i just got up

[vast jetty]

could you do the redt

[half swan]

wdym?

[vast jetty]

the last two have been brutal

[half swan]

yeah I managed to finish the rest, most it took me on any day was 5 hours I think?

this one I have 0 idea on how to even approach

[magic oxide]

finally a language that is not pain and suffering

[half swan]

||regex support, how do I make the groups be all of them and not just the last match?||

[vast jetty]

ok

this is

insanity

[vast jetty]

hmmmm?

[vast jetty]

it is all of them no

[half swan]

it should be?

[vast jetty]

ah the joltages arent working

[half swan]

but maybe I'm doing nested groups wrong?

[magic oxide]

put the + inside the group? is that what you mean?

you'll probably want to nest a non capturing subgroup

[vast jetty]

||why even regex this one though||

[magic oxide]

so like ||`((?:\d+,?)+)||

[vast jetty]

actually i take it back its prolly a good idea

[half swan]

yeah I felt like manual parsing will get a bit too long

[vast jetty]

but yeah i have no clue how to approach this one

i dont think brute force would work.

[magic oxide]

I'm not looking at it until I finish day 9

you guys have fun

[half swan]

alrighty then

[vast jetty]

go finish day 9\

i need to brush and have breakfast

ill be wondering about the problem while i do that

very interesting problem

i cant even think of what type of problem it is

other than just general 'optimisation'

[humble copper]

i know exactly what to do to get day 10 part 2 to work

the only problem is

its muuuuuuuuch easier to do with libraries

and ||i just had a final for one of my classes about this exact thing today|| 😭

[vast jetty]

i dont even really know how to make part 1 work

you could possibly semi brute force it

[waxen hound]

Tuff

[feral hazel]

woot finally got an answer for ||the first line in my input|| 😭

[sage nacelle]

I am best python developer with 10+ years experience.(
NumPy
pandas
SciPy
xarray
statsmodels
Dask
Matplotlib
Seaborn
Plotly
Bokeh
Altair
Pygal)

Who will co-work with me?

[mossy basin]

no one if you spam random channels with off-topic bullshit

[vast jetty]

...i feel like its a joke

[vast jetty]

show us your aoc resume

[sage nacelle]

DM me

[vast jetty]

[ivory vigil]

You missed shapely

[full dune]

No portfolio = no freinds

[vast jetty]

i wanna see their aoc stars

better have like 300 minimum

even i have 100

[naive wharf]

does the trivial O(2^n) solution work?

[ivory vigil]

for?

[naive wharf]

today

wow I think I have a moderately elegant solution for yesterday

[ivory vigil]

p1? Yes. p2? No.

[naive wharf]

n² log n 🔥

[humble copper]

p2 today is destroying me

conceptually i understand what to do, the implementation is the hard part

[woven oriole]

Importless aoc enthusiast 👍

[sweet cove]

I have 2

[vast jetty]

dw youll get there!

this is the breakdown

i started in 2022 lol

[sweet cove]

Yeah I'll probably start doing past years when I get stuck on this year eventually, I don't have too much time to spend on programming anyway

[uncut mason]

If you want a confidence boost, do 2015. It's not too hard.

[vast jetty]

alr finally got around to doing today's

part 1 done

make that 101

hmmmmmmmmm

interesting

very interesting

looks really annoying

[peak dock]

i love multiples of 10

[strange zinc]

i am all over the place 🙈

[uncut mason]

What's AoC++?

[keen trout]

Supporter https://adventofcode.com/2025/support

[limpid grail]

damn

[mortal bronze]

bruh u guys alr day 10 😩

I lagged on day 6

and when I start I had seg fault on the data set but not on test data 😩

[vast jetty]

its cuz youre doing it in non-python then

[mortal bronze]

lol

[vast jetty]

cpp?

just do em in python much better

day 6 i think was slightly annoying

[mortal bronze]

nah I wanna do string manipulation on cpp

[vast jetty]

oh yeah parsing

[hollow summit]

day 6 was fun

[vast jetty]

why cpp tho

the language roulette was c

[mortal bronze]

why not

[vast jetty]

why is there only star in this ascii art :(

[mortal bronze]

cus im more comfortable with OOP

[hollow summit]

python has oop

[mortal bronze]

buts it's not c++ :3

[vast jetty]

i cant imagine doing it in cpp

must be really annoying

afaik strings in general are annoying on cpp

[hollow summit]

shouldve been rust

[vast jetty]

do yall also have this

[mortal bronze]

try that on c 😀

its more fun on c 😀

[vast jetty]

hey im fine with my python

got it done in 22 lines

fine by me

[hollow summit]

day 6?

[vast jetty]

yea

[vast jetty]

brace yourself tho

d7 is only slightly annoying

d8 is annoying if you dont know the required algorithm

i only got through it through having faith in 'throw resources at it'

[hollow summit]

i am still stuck at day 9 part 2

[vast jetty]

i thought itd take like dayssss to run

but it ran in like under a second

whoops

[mortal bronze]

imagine

const char* a= "data";
const char* b= "data";
if (strcmp(a,b) == 0);

[vast jetty]

turns out i couldve just done it in like an hour or two then

[vast jetty]

oh its fun

can you get it to work on example input yet

[hollow summit]

my previous solution worked on example input (constructing a string grid 😢)

but it was super super super slow on the real

[vast jetty]

did you consider ||functools.cache|| yet?

[hollow summit]

it's not ||recursive||

[vast jetty]

i somehow forgot thats a super helpful thing

[vast jetty]

no, for the inside checks

||the rectangles you check for will have a bunch of repeating coords||

i had like ||16k hits|| or smth

[hollow summit]

bruh

i shouldve tried that

time to rewrite

[vast jetty]

why rewrite, just ||slap on the decorator||

[vast jetty]

||160k lol||

[hollow summit]

yeah but i did it yesterday and my vscode doesn't have the edit history to revert back to it

the solution was very bad though

[vast jetty]

shouldve ran git commit huh

[hollow summit]

why would i commit a non working solution

[vast jetty]

trust me mine is shit too

it runs in like

80 seconds?

but it does the job

[vast jetty]

just good practice if youre gonna tear down code

maybe youll find a way to reuse an old idea

[hollow summit]

in order to ||fill the polygon, i iterated the grid twice: top to bottom first, then left to right||
there is probably a better algorithm

i think its called ||flood fill||

[vast jetty]

you can always just branch from it

[vast jetty]

yeah no thats shit

the border of the actual inputh has like ||500k elements||

we're talking like ||25*25*10^6 elements on the inside||

[mortal bronze]

uhh the day 6 data looks weird.... don't tell me on part 2 it'll add trailing 0 ._.

[hollow summit]

how did you fill it

[vast jetty]

:')

[hollow summit]

did you even do the ||string grid approach|| yourself

[vast jetty]

i dont

come to #aoc-solution-hints

[vast jetty]

its just 4-5 lines

looks weird cuz of wrapping

[mortal bronze]

wdym?

lines of code??

[vast jetty]

no like 4-5 lines of input

i mean youve done part 1 right

youd know

its like this

the page looks weird cuz of wrapping

[mortal bronze]

well ofc if you do like 10 lines you'll need super beefy data type cus pretty sure long long won't be enough

[vast jetty]

yes

do long long long long

[mortal bronze]

you would probably need int256_t

lol

[vast jetty]

i have to go give a test

3 hours

then ill actually sit and think about d10 p2