advent-of-code

[placid lake]

ah yeah, those suck

[woven sable]

which i absolutely struggled with

i gave up last year bc i slept really early one day and got behind on aoc during the hardest puzzles bc it was like day 16

[ripe scroll]

On day3 part 1 what does adjacent even mean here, in the example they said 14 isn't adjacent, but it is diagonally adjacent to * . 58 is adjacent to +.

This is turning to be a english comprehension competition at this point.

[woven solar]

*lanternfish, not jellyfish, but yeah that one was fun

[woven sable]

that one and the snail fish

i hated the snail fish wtv so much

didn't they multiply or smth it threw me off

[woven sable]

i think they mean that it isn't directly adjacent (up, down, left & right), but it can be diagonally adjacent and still count too

[woven solar]

they did math

https://adventofcode.com/2021/day/18
snailfish math.

[ripe scroll]

but 58 is directly adjacent to + right (example)

on line 6

[lusty rover]

oh no
There's a screenshot somewhere of me at 100% RAM usage because of those stupid snailfish

[woven sable]

it's not touching tho

it'd be adjacent if it was +58

instead it's +.58..

[woven sable]

3rd worst puzzle for me that year

idk if i'll finish aoc this year bc i j don't have the time

2nd worst was the cuboids stuff and the worst was the scanner beacon thing it took me like 6 hours to find thensolurio

[ripe scroll]

Oh i get it, so any symbol just below a number is considered touching it?

[woven sable]

yeah as long as it's a cell away

[woven solar]

yeah, beacon scanner was a nightmare disguised as a day-19 task

cuboids? which one was that?

[woven sable]

like

..467...
.....+..
``` is touching

[ripe scroll]

Okay thanks for the explanation. The statement should've been more verbose

[woven sable]

day 22

splitting 3d cubes

[magic oxide]

my worst puzzle was by far amphipods

took me 2 weeks

[woven sable]

which one was that

i feel like i remember that one

[magic oxide]

2021 day 23

[woven sable]

my fav problems in 2021 were the flooding ones

[woven solar]

ah, right, reactor reboot. yeah, I also cheesed it with numpy and then got dumped into 3d-interval-intersection-hell in part2.

[woven sable]

i loved getting to use BFS & DFS and all

i honestly forgot what they were

but i'll relearn them eventually

[woven solar]

i never forgot bfs and dfs ever since my professor casually mentioned that the only difference between them is using a queue vs a stack to store nodes to be processed. this is how I thought of them ever since.

[magic oxide]

I learnt bfs 5 weeks into learning python

and dfs last year day 19

don't think I've ever used iterative dfs tho

[humble copper]

oh my god

i remember this

it was torture

[woven solar]

yeah, 2022 day 19 was pretty evil too. you needed a good heuristic or the search was way too slow.

[humble copper]

honestly i prefer 2021 day 19 over 2022

[woven sable]

im prob gonna give up on day 15 again tbh

[humble copper]

ive tried 2022's 3 separate times and still haven't gotten it to work

[magic oxide]

I spent a good while on day 19 before it finally clicked and I understood what I was doing

[placid lake]

day 19 was harddd

definitely had to memoize that one

[humble copper]

one of the days for 2021, cant remember which one it was

it had like a trajectory prediction thing

in all honesty i just bruteforced it on a large enough area

and it got the right answer in a really reasonable amount of time

oh day 17

[placid lake]

wow i’m looking at my past solutions for these oh gosh

hah i implemented lshift and rshift for day 17

that’s pretty funny

[humble copper]

my solution for part 2 of day 15 takes 5 hours to run

but it works

[placid lake]

finding the cycle for day 17 was really something

[placid lake]

constraint solver!

[woven solar]

my notes for day17 (rocks falling) say "don't know why this works, but it does"

[cyan sleet]

hopefully it's going to be less of a pain than ocaml

[humble copper]

idk why i do this to myself but ive made it a challenge every year to try to complete these without any imports

[placid lake]

hahahaha no way i’m doing that

sounds like hell

[humble copper]

not even stdlib imports

it absolutely is

the one year i completed a full calendar was my first one doing it

instant burnout on day 25 🤩

[woven sable]

aoc def gave me burnout in 2021

i think i stopped coding for a few months after

[long jacinth]

todays puzzle long

[ivory dagger]

I don't know if this event is something for me. I've been studying python for data science on datacamp for a couple of months to prepare for a bootcamp I will take. Tried the first couple of days but I never learned anything remotely close to this and at this point I wonder if I'm wasting my time.

[sinful pewter]

What is DIGIT_MAP?

[flat vortex]

Finished day 2 🥹🥹

I agree day 2 was fun lolol on day 1 there's a tricky test case :v

[hollow wharf]

eightwo?

another day another part 2 easier than 1

i belive thats just cause the parsing is already done

[random flower]

I wanted to automate receiving the puzzle input with a get request
But forgot we needed to login in order to get it at all

[kind oasis]

breaths in
breaths out
holy *&@#/ part 2 today is really not cooperating with my part 1 solution

xD

[sullen ruin]

That's how they work.
Part_1 usually introduces the problem, let's you write the correct parsing and makes sure, you understand.
Part_2 is where the meat of the problem really is.

[sour trout]

Does anybody have any methods toc hcek if a number is adjacent to a symbol?

[long jacinth]

your answer is not right
for first part, man how do i debug whats wong now lol

[hollow wharf]

print out everything you can for the example case

and see what goes wrong

[long jacinth]

example case produces the correct result

[hollow wharf]

well now you have a problem

[long jacinth]

that i do

ive made a map of all the correct parts, so its probably something im going to have to find manually

[long jacinth]

oh i think i got it

alright, second part seems easy

[halcyon pulsar]

Does anyone have advice on debugging? like spotting silly mistakes, slight logic errors, edge cases, etc?

[low condor]

The biggest one is to run against test input

Usually, but not always, you can shake out any issues by doing that

I write automated unit tests for my code. So with one click I can run all tests and show my code isn’t broken by a change I made

[long jacinth]

alright part 2 done, super easy

[halcyon pulsar]

Unfortunately the test input didn't contain any zeroes

[hidden musk]

functional fever i suppose

[low condor]

Yeah that’s unlucky

But the next thing is then to grab a section of your real input and test against that

[halcyon pulsar]

Yeah I grabbed the first line which had no symbols, and it returned 1500

So I was like hang on

[low condor]

Also, you can avoid bugs by writing more resilient code

[magic oxide]

it might be because I'm doing this with an added challenge, but today seems much harder than I would expect from day 3

[low condor]

I know it’s kind of like saying “if you are lost in the woods, just try not going into the woods to start with”

[long jacinth]

my solution has 193 LoC for todays, way more 'difficult' than previous days

[low condor]

But you can check for digits using str.isdigit() or using regex with \d

And these are more robust ways of detecting if you have a digit

[low condor]

I would agree with this yes

[halcyon pulsar]

I didn't want to deal with it being untyped + I checked for "123456789." and I didn't want an and

[low condor]

I still suspect the ante has been upped because of AI

[low condor]

But an and is more resilient

[halcyon pulsar]

How

[low condor]

Well because you don’t need to code in all the digits directly

Another option is string.ascii_digits

Oh lol - one sec

!d string.digits

[marsh current]

string.digits

string.digits```
The string `'0123456789'`.

[low condor]

You could have used this and add . on the end

In general you want to reduce the amount you have to hardcode

The more you hardcode the more potential sources of errors can creep in

[halcyon pulsar]

Ok good to know

[modern terrace]

i couldn't do this earlier because it was sunday

[sour trout]

Does anybody know any methods to check for adjancet numbers? i cant figure one decent.

[modern terrace]

||regex||

i have used it for the past 3 days

[sour trout]

hmm that might be it

||Also what pattern if possible?||

[modern terrace]

i used ||^/$ with slicing||

i actually just used ||regex|| for the ||.findall() feature||

[sour trout]

Hm, im gonna try and see.

Thanks

[random flower]

im on first day:
how is the last line interpreted as 77?

1abc2
pqr3stu8vwx
a1b2c3d4e5f
treb7uchet

there's only one 7. or is it because of this hint:
||```
...single two-digit number.

[halcyon pulsar]

if there is only one number then it is repeated

[random flower]

ok

[hollow wharf]

you take first number you find going from left and going from right

forming a 2 digit number

so in this case both from left and from right side the first number is 7

hence number will be 77

[low condor]

On each line, the calibration value can be found by combining the first digit and the last digit (in that order) to form a single two-digit number.

[random flower]

first time i read it, didn't make sense lmaoo

[low condor]

Advent of Reading Comprehension :)

[random flower]

my day1part1 didn't just compile in two minutes
my laptop is crying

[low condor]

I’m curious to see what your code looks like

[random flower]

nvm i was compiling the release build

[low condor]

Oh what language are you using?

[random flower]

trying it on rust 😄

[low condor]

Ah

It still takes that long to compile? Interesting

[random flower]

tbf im on a 7yo lapto xD

[low condor]

Ahh

That explains that

[jade atlas]

where can I see the full leaderboard? I just saw Petr Viktorin in the aoc leaderboard and I am curious to see which other core developers are taking part here

[edgy hearth]

is day 2 easy guys?

[hardy zenith]

today's AOC is cruel

any number adjacent to a symbol, even diagonally

[hardy zenith]

yes, but day 3 ramps up difficult again

[hollow wharf]

im scared of day 4

[edgy hearth]

im scared of falling too much behind

going to try and nuke all of 2 today

[hardy zenith]

oh 2 is much easier than you think

today however... it'll probably take me an hour

[hollow summit]

i think for day 3 i'll just convert input into a 2d array

so i can index stuff easier

[hardy zenith]

I just went ||yolo and used regex||

[lime fulcrum]

today seems kind of hard

[molten fulcrum]

[Day 2] ||Regex|| has always confused me a bit. If I'd like to find how many red, blue and green ones there are in a draw, for example:

• 3 blue, 4 red
• 1 red, 2 green, 6 blue
• 2 green

||How would I go about parsing all these with the same regex, and still know which part of the parsed data is blue, red and green? And how do you handle the different orders of items, as well as some items not being present in all draws?||

I'm not looking for a full solution, but rather how I should approach thinking about this problem in terms of ||regex||

[hardy zenith]

||I wouldn't use regex||

but also, you can use ||OR statements, like (blue|green|red)||

[molten fulcrum]

Yeah, I know how to ||match some things||, like the thing you sent, but I don't get how you'd extract useful information from it, like how many blue, red and green ones exist

[hardy zenith]

you can't solve the problem with one tool alone

[potent rivet]

guys ... day 03 ... went to grab bunch of correct numbers, but the sum not ok ... missing something here?
grab all numbers with adjacent to symbols ... i checked like 8 lines from the txt

[hardy zenith]

||you need to check the other lines as well. ALL adjacent values to the number||

[molten fulcrum]

Of course you need more than one tool to solve the whole problem, but that doesn't make the tool useless for extracting some information

[long jacinth]

i dont think today was difficult

just time consuming

i mean, maybe it was difficult if youre like.. golfing it lol

[low condor]

I think today is difficult if you don’t have a good approach, and I think more people than usual for a day 3 puzzle will not have that

But if you have a good approach then it’s fine

[magic oxide]

today is difficult for me because i overthink things

[bronze vault]

Can the same number have more than one adjacent symbol?

[earnest forge]

Theoretically yes but it doesn't seem to ever actually occur

[bronze vault]

ok thats greatr

[royal gull]

||With a regex, you can have multiple capturing groups. Each paren is a group, so you could capture: "(digits)space(blue|green|red)". The match object would have the digits in group(1) and colors in group(2). So, one regex, multiple matches, and each match has both digits and color||

[molten fulcrum]

Ah, that’s neat!

[hollow wharf]

Bro where all the ghost pings coming from 💀

Oh I have the role

[hollow wharf]

honestly that's a more efficient way than what i did

i just did this:
||```py
if color == 'red' and quantity > fewest_red:
fewest_red = quantity
elif color == 'green' and quantity > fewest_green:
fewest_green = quantity
elif color == 'blue' and quantity > fewest_blue:
fewest_blue = quantity

[halcyon pulsar]

I used 3 different regexs patterns for each colour

[pliant hornet]

Just sorted Day 2, without looking for hints and only a couple of googles!

[halcyon pulsar]

Or this

[hollow wharf]

day 3 looks tricky

[low condor]

overthinking in general is bad

[magic oxide]

by overthink i mean like
if i think there is a neat/clever way to do something that isn't obvious at first, i'll tunnel vision into that method for ages

[low condor]

fair

[magic oxide]

it's still bad tho i agree

[tepid cairn]

maybe I'll try solving some of the puzzles with gleam 👀

[proven orbit]

Does anyone know if the inputs and by extension solutions vary per the person? Just submitted Day 3 part 2 and was winding dowin watching a video og some of the "pros" and their solution number was different than mine?

[earnest forge]

Yes, there are multiple inputs

(sometimes, when you get a wrong answer, the submission page will tell you it's the right answer for someone else :P)

[proven orbit]

Oh that's devilish. Thanks!!

[short karma]

Yeah that sounds about right

[hardy zenith]

omg

part 2 is insane again today

[hardy zenith]

this year is definitely starting off harder than any AOC I can remember. part 1s weren't impossible, but part 2 definitely ramp hard

[low condor]

I wouldn't even say it's p1 vs p2

for day 3 i think the difficulty is part of the whole day, not just p2

[hardy zenith]

I felt part 1 was at least intuitively solvable ||if you know about sets|| at least

but that led me down the absolute wrong path for part 2

[low condor]

right - day 1 part 2 is the hardest day 1 puzzle

[naive stone]

how so?

[magic oxide]

i found part 1 of today significantly harder than part 2, but i think that was mostly my solution for part 1 being well suited for part 2

[long jacinth]

part 2 took me 5 minutes, part 1 took me 2 hours 👀

[hardy zenith]

||because I didn't preserve the position of the cogs so I could check neighboring numbers||

[naive stone]

oh i see

[hardy zenith]

so now I need to write another parsing function instead

I think that'll be an easier approach

very intuitive naming

[woven solar]

consider having todo!() in parse2, instead of doing -> !. the former won't be an error.

[hardy zenith]

you can't use todo in the return position

[woven solar]

yeah, i mean just don't specify the return type. or, I guess, ! would work too.

[hardy zenith]

on nightly it's not an error I think

[woven solar]

huh

[hardy zenith]

I don't think the never type is stable

[hidden musk]

seriously?

[hardy zenith]

plenty of stuff the compiler devs get to use that stable doesn't 😂

[earnest forge]

*Never as ! is unstable

Never as std::something::Infallible is perfectly stable

I forget where in std Infallible is but it is stable

[hardy zenith]

in python it's just called typing.NoReturn

but I don't think it has the same characteristics as never where anything can be cast into it

[hollow wharf]

in python you dont even have to typehint

[hidden musk]

we do have a Never, too. they're equivalent

!d typing.Never

[marsh current]

typing.Never

typing.Never```
The [bottom type](https://en.wikipedia.org/wiki/Bottom_type), a type that has no members.

This can be used to define a function that should never be called, or a function that never returns...

[hidden musk]

!d typing.NoReturn

[marsh current]

typing.NoReturn

typing.NoReturn```
Special type indicating that a function never returns.

For example:

```py
from typing import NoReturn

def stop() -> NoReturn:
    raise RuntimeError('no way')
```  `NoReturn` can also be used as a [bottom type](https://en.wikipedia.org/wiki/Bottom_type), a type that has no values. Starting in Python 3.11, the [`Never`](https://docs.python.org/3/library/typing.html#typing.Never) type should be used for this concept instead. Type checkers should treat the two equivalently...

[hardy zenith]

oh, I didn't know it was added in 3.11

that's very recent

[untold briar]

Dang problem today is not so easy anymore

This 2D problem is on a different level from the previous ones

[vital lotus]

probably to combat LLMs, took me three times the time to solve this one than the previous ones

[untold briar]

Took me 3 hours to solve them

Yeah about 3 times the previous ones for me too

[hardy zenith]

that's an interesting thought

the language has been a lot harder to understand in the prompts to me as well

[hidden musk]

much nicer than NoReturn imo

[vital lotus]

I didn't find it too complex to understand, but I haven't done the previous years (this is my first), so I don't really have a point of reference

[hardy zenith]

day1part2 was more what I was referring to with my statement

the sample set they provided did not have the edge-cases you needed to solve for

[rapid wolf]

yeah, that I don't like

[hidden musk]

i think the main struggle with d1p2 was not having clarification for the edge case

[hardy zenith]

but if they had it in the sample set, you would've found it very early because your tests would fail

[untold briar]

Yeah it feels that the problems are worded in a very roundabout way

[naive stone]

not informing you of every edge case is an intentional part of the difficulty. I have no problem with it

[untold briar]

With lots of flavor text (read: gibberish for LLMs) as well

[hardy zenith]

but it was poorly worded and the edge case wasn't possible to catch

[rapid wolf]

ehhh, having to manually invent test cases isn't something I find interesting.

[vital lotus]

I thankfully didn't run into that problem (i sort of guessed that it was the case), but that did happen for me on today's (my solution didn't account for when ||the number is on the edge of the board||, which wasn't in the sample)

[hardy zenith]

but it is

the very first number is

unless you mean ||the right edge|| but that should be intuitive

[vital lotus]

yeah, I forgot to ||reset the flags whenever the line reached its end, thankfully I realized quickly||

[untold briar]

For me I was having a bit of troubles with indexing

[hidden musk]

the problem isn't that there is an edge case, it's that it wasn't clear how it was meant to be handled

[naive stone]

what edge case are we talking about?

[hidden musk]

in day 1 part 2, ||the overlapping word digits||

[hardy zenith]

spoiler that please

[vital lotus]

||believe things like twone and nineight||

[untold briar]

18 probably

Oh those too

[hollow wharf]

||eightwo|| real one

[naive stone]

I'm not convinced that needed to be explicitly explained. I think it would have to be explained if ||overlapping|| made it so they weren't valid

[untold briar]

I didn't have troubles with overlapping ones though, since I found the digits independently from each other

[hardy zenith]

18 wasn't one of the edgecases

it was clear ||they only wanted you to replace one through nine||

[naive stone]

when i was writing my solution, i didn't even realize that was happening in the input, my code just handled it without me knowing

[untold briar]

||oneight||

[hardy zenith]

that's not 18

I thought you were indicating ||eighteen||

[vital lotus]

||They didn't test for that actually, I had zero in mine and I saw some other solutions that counted above nine and it still worked||

[hardy zenith]

you would get wrong answers if you replaced more parts

[untold briar]

Well no problem stated quite clearly that we were looking for digits, not numbers

[hardy zenith]

actualy wait

no you wouldn't, but it wasn't part of the question

digits are numbers

[untold briar]

On each line, the calibration value can be found by combining the first digit and the last digit (in that order) to form a single two-digit number.

Bruh digits =/= numbers

[hardy zenith]

yes they are

[woven sable]

fs

i had some really weird regex solution

i should've just iterated thru the string and put the string into buckets of 5, and see if it contained a digit and then convert it to its number form in the string

[hardy zenith]

ughh... another day where sample input parses correctly but the actual set doesn't

[low condor]

which day?

[hardy zenith]

3

[low condor]

Happy to chat more in #aoc-solution-hints but in genral I would say ||check boundary conditions - can you parse numbers and symbols at the edge of your input, etc.||

[hardy zenith]

hmm, it might be missing ||if a gear is at the end of the bound since that's kind of tricky||

see:
||```rs
for m in number_finder.find_iter(input.inner[y - 1]) {
let x_start = m.start();
let x_end = m.end();

                // check the start bound edge case
                let slice_start = if x_start == 0 {
                    0
                } else {
                    x_start - 1
                };
                // check the end bound edge case
                let slice_end = if x_end == input.x_max {
                    x_end
                } else {
                    x_end + 1
                };

                if (slice_start..slice_end).contains(&x) {
                    neighbors.push(m.as_str().parse().unwrap());
                }
            }

[low condor]

write a sample input with that case and see if it works

[hardy zenith]

ohhhhh

I'm dumb

that's supposed to be nested in the inner loop

got it

[sleek cave]

heh, so much javascript i forgot python syntax for a sec

python destructuring when

[hardy zenith]

it kind of does, but only for tuples

you can use dataclass with dataclasses.as_tuple

[sleek cave]

yea ik there's unpacking

[hardy zenith]

doing a decent pace for part 2 minus the first day when I had work lol

[simple raven]

is the star of the left a part of the tree graphic or am I missing something on day2 and 3?

[placid lake]

trebuchet

[ripe meteor]

Its you being loaded and then flung into the sky

[woven sable]

that was my issue, i had a ||gear at the very last line which prevented my code from working till i fixed it||

[median juniper]

how to get "Advent of Code" role?

and should i get it?

i think i see

hmm

im confused
what does &aoc link denballakh do? i dont see any changes neither in discord nor in AoC website

[median juniper]

this was a thing in python2

[short karma]

#roles

#roles message

[velvet anvil]

I have to study for a test tomorrow and i randomly decided to do AOC day 1 in regex... WTF is wrong with me

[short karma]

Average internet user

[hardy zenith]

my issue was I had a variable at the wrong nesting of loops

so I was pushing all the neighbors for a row into a given list

[hardy zenith]

tbh I used regex for day 1 too

[hearty sinew]

Is it ok to ask a clarification question regarding day 3 part 2?

[minor cave]

yup, you're welcome to ask here. If you need to get into more of the particulars you can ask in #aoc-solution-hints

[modest ermine]

i've used regex every day so far lol

[hearty sinew]

I wanted to clarify:

A gear can only have EXACTLY 2 adjacents? Anything with more than 2 can't be a gear?

[modest ermine]

correct

[hearty sinew]

Awesome. learned my lesson in day 1 about understanding the instructions and wanted to clarify

thank you!

[modest ermine]

yw!

dw that happens a lot to me where there are specific stipulations that i miss

[sleek cave]

i take it all back

im suffering again

even after reading up on solutions

i can't implement in python

[low condor]

Which day is this?

[sleek cave]

and these numbers are huge

aoc '21, reactor reboot

[low condor]

Ahhhh

Honestly there’s no shame in that

In 3 years of AOC it’s the toughest problem for me personally

[sleek cave]

im currently following

• this js solution
• this haskell solution - i have never seen haskell before but the language looks incredibly elegant

but when i try to implement it in python...

1. the intersection cuboids i store multiply exponentially, completely unfeasible
2. i get wrong answer for larger sample input 😩

ig if i may have some hints

i'd greatly appreciate :D

in spoilers ofc

[low condor]

There is one solution that sort of worked for me

I actually don’t know where to write AOC 2021 spoilers lol

[tiny field]

Anyone know hmac?

[sleek cave]

just here lol

with ||

oh

or #aoc-solution-hints

ping me :) i'm taking a short break and will be back to tackle this soon

[low condor]

Sure!

Let me look up what I wrote

[tiny field]

Anyone know hmac?

[low condor]

This is a server for Advent of Code. You probably want to ask your question in a help channel. More info is here: #❓｜how-to-get-help

[woven sable]

im pretty sure i took inspiration from your solution for that problem bc i was lost

or acc no that was the snailfish one

cuz i genuinely couldn't figure it out

[low condor]

It was DEFINITELY not reactor reboot

well, maybe it was

but it wasn't MY solution

[woven sable]

was reactor reboot the cube one

[low condor]

I had to look up a solution because I was completely and utterly lost

yes it was

"my" solution is now in #aoc-solution-hints if you want to look

[woven sable]

yeah i lowkey took inspiration from someone's solution for that one too

i was lost

lemme look

[low condor]

snailfish on the other hand I am proud of my solution for that

[woven sable]

ur code looks so clean wtf

[low condor]

hah

well i still don't understand it

I'm trying to read whatever the hell I wrote

right now

[woven sable]

me too i found my solution and idk wtf is going on

like idk what's going on ||https://github.com/Shom770/adventofcode-2021/blob/main/solutions/day_22.py||

[low condor]

you solved it though so that's very good :)

[bronze vault]

aoc is awesome

[simple raven]

there is a guy on aoc reddit who solved it in 21 lines of code, it's pretty crazy

[short karma]

Wow

[cinder tree]

does anyone know for day 3 part 2 if a possible edge case is: ||more than 2 adjacent, and if so, for example if I had 3 adjacent, should those all then be multiplied?||

[simple raven]

I haven't checked manually, but I would recommend accounting for it and the task is exactly 2 numbers near a star make a valid gear

[cinder tree]

oki thank ya

oh another question sorry

[simple raven]

do you want me to spoil whether there are more than 2 or not?

[cinder tree]

||what if line 2-4 has a number in line 4 for a gear, that is then also correct for like 4-6 for example?||

[cinder tree]

sure

although we have diff puzzle inputs, so maybe mine will if yours doesnt

[simple raven]

||no, all stars have only 1 or 2 numbers attached if my script is running well||

[simple raven]

wait the inputs are unique?

[cinder tree]

yea

everyone gets a different generated input list

[simple raven]

O.o I had no idea 😄 that's really nice

[cinder tree]

as far as I know

so you cant just cheat off someone else I guess

[simple raven]

is your first line .....180.........230..........................218.....189......415.......................322....507..................206..............111...?

[cinder tree]

nope my first line is ...............307............130..................969...601...186.........................................312....628..........878..........

[simple raven]

wow, okay so you should probably account for the edge case then 😄

[cinder tree]

roger

[cinder tree]

do you know about this one at all?

[simple raven]

I don't understand, can you provide a small example?

[cinder tree]

sure

||- in line 3 I have an * that's adjacent to a num ber in line 2: 455 and a number in line 4: 333.

• in line 5 I have an * that's adjacent to a num ber in line 4: 333 and a number in line 6: 111.

will I end up having to use line 4:333 twice? or is it excluded if it is already used? ||

[simple raven]

||the language of the part2 assignment makes you focus on the stars rather than the numbers, we are interested in gears which are represented as the stars, then we have the sentence "Its gear ratio is the result of multiplying those two numbers together." so I wouldn't expect you need to account for that and in your case you would use the same number twice for the two different gears||

[cinder tree]

alrighty, thanks a lot 🙂 then I can not over engineer this one

[simple raven]

actually wait...

[cinder tree]

ruh-oh

[simple raven]

let me check I said the right thing

[cinder tree]

legend

[simple raven]

I've just read my entire input and I don't have that edge case 😅 so I have no idea which is correct, but in my code ||I only consider each number once||

sorry

[simple raven]

based on how the assignment reads I think my previous suggestion was correct, I would expect my solution to fail for your edge case

@cinder tree sorry for spamming, once you solve it could you send me your input so I can check if my solution is correct by comparing against your answer?

[median juniper]

thank you 👍

i forgot that this thing exists

[hazy patrol]

Hi! I'm working on solving AoC in a single line of Python code. Does anyone have any tips? I made it through a dozen days last year and have Days 1-3 for this year! https://github.com/savbell/advent-of-code-one-liners

Mostly I'm wondering if anyone has any little-known tricks for shortening longer code/statements into a single-line expression! A few of the things I'm already doing are under the "Fun Stuff" in the Readme.

[jovial jay]

hardest day so far for you guys?

[simple raven]

that's crazy, good luck

[jovial jay]

to me, day 1 is the hardest. though i didn't do that in python

[simple raven]

I had to think about day3 the most so that's my choice

[hazy patrol]

Honestly same. I needed to chat with a friend to figure out what edge case(s) I was missing for part 2, while the other days I figured out on my own

[jovial jay]

the example didn't provide enough for the test. so i was left confused for hours

[hazy patrol]

Thanks! 🥰 I just hope to make it further than last year. I stopped once I found myself trying to implement A* in a single line...

[simple raven]

I'm not a programmer so seeing people do magic with regex is way beyond my comprehension and what you've already done is incredible to me

[minor cave]

#1180008645384216606 will probably have some tips for you!

[untold briar]

Is anyone able to use the /aoc join command mentioned in the aoc-faq channel?

[minor cave]

I see the error on our side. It's unclear why it's happening though. Did you get the board you were assigned to?

[untold briar]

No, I got the message that the application did not respond, only that.

[untold briar]

Oh nvm, I was being dumb, I used it in the wrong channel

[bright laurel]

advent of code should edit their inputs so it suits the sample input, because i'm not gonna inspect thousand of lines to see where the fuck is the erro

[random flower]

what the hell did i do
the result for day1 part2 changes when run a bunch of times
It shouldn't do this

im laughing so hard idfk what i did lmaoooo

[bright laurel]

you achieved true randomness

[random flower]

maybe it's because i used hashmaps for one thing

yep
the iterator gives back values in arbitrary order
gotta think of something else

[jovial jay]

I personally like visual debugging for that reason.

[bright laurel]

what is that

[jovial jay]

When you execute the code and see what it did at the a visual level. So you can see what is going on. I'm using G'MIC to do AoC, so when I execute my code, I get displays of images reflecting the output of my code.

[bright laurel]

my code gives the right answer for every single input, idk how and why aoc full input is not accepted

[jovial jay]

Day 2?

[bright laurel]

1 part2

[jovial jay]

Oh that. The simple example doesn't cover corner cases

[bright laurel]

which are

[jovial jay]

twone -> two one

[bright laurel]

right, doesn't work, thanks

solved, thanks again

[cyan sleet]

depending on hashmap ordering 🥴

[hidden musk]

that's quite impressive

oh it's rust, nevermind

if python that would be fun

[cyan sleet]

fun latent bugs

make your own

[earnest forge]

GL everyone

Hopefully no stupid bugs today 🙃

[sharp panther]

may the odds be ever in your favor

[earnest forge]

damn I used to know the response for that it's been too long since I read/watched hunger games lol

[sharp panther]

there's a response?

[earnest forge]

I could have sworn there was

[sharp panther]

i just like the phrase :p

[earnest forge]

Maybe I'm hallucinating

there isn't and I'm just dumb

[naive stone]

https://tenor.com/view/katniss-everdeen-jennifer-lawrence-the-mocking-jay-the-hunger-games-gif-8640100

[minor cave]

Language Roulette: Day 4
The language is ... Lisp

[mossy basin]

dope language

[sharp panther]

any one in particular?

[earnest forge]

(oh (no))

[median juniper]

why

())()()((((())

[mossy basin]

any flavor?

[median juniper]

stinky

[minor cave]

pick your fave flavor

[naive stone]

i knew i was gonna see parenthesis the second the language was chosen

[earnest forge]

Lmao

[woven sable]

what are the languages in the roulette?

[sharp panther]

they're all pinned (the ones so far, anyway)

[minor cave]

:) a secret

[muted aurora]

hanging out in VC0 if anyone wants to chill and talk about the AOC

[woven sable]

oo

[earnest forge]

*(knows (fisk) (see (brackets)))

[woven sable]

i hate lisp

sm

[mossy basin]

i love lisp, but i can't type parens very easily

[sharp panther]

someone need's to greenspun's tenth rule this one and implement their own lisp

[woven sable]

i get lost w all the parentheses

scheme is a flavor of lisp right

[sharp panther]

it's close enough

[woven sable]

in freshman year we had to code using scheme

😭

[covert arrow]

30s

glhf

[median juniper]

its 8am rn, i cant speak, my neighbour is sleeping

[woven solar]

lisp but [], () and <> all mean the same thing

[sharp panther]

{} feeling left out, ouch

[woven sable]

scheme is horrible tbh

cool lang tho

[shut trellis]

<@&518565788744024082> Good morning! Day 4 is ready to be attempted. View it online now at https://adventofcode.com/2023/day/4. Good luck!

[sharp panther]

oh nice, parsing work isn't bad

[earnest forge]

33/18 let's go

[magic oxide]

looks simple enough

damn

nice

what's p2?

not at home rn so can't do it

[earnest forge]

I feel like I probably shouldn't say that at least until the p2 leaderboard is full

[compact wasp]

we're so back

[magic oxide]

yeah ok

[earnest forge]

Oh you got 9/6 nice job

[compact wasp]

ty

[magic oxide]

Jesus 9/6

[earnest forge]

where's oliver ni

[magic oxide]

insane

[compact wasp]

near the bottom

83

[sharp panther]

done at 00:07:08, much better than yesterday's 00:11:37

[compact wasp]

did he leave the server?

[shadow musk]

[earnest forge]

||Instead of getting points you get bonus cards (copies of the next N cards where N is the number of matches), figure out how many you have in total||

[shadow musk]

def crazy():
    print("""Crazy?
I was crazy once. 
They put me in a room with a function.
A recursive function.
A recursive function with no base case.
That made me crazy.\n""")
    crazy()

crazy()

💀

[minor cave]

#1181099526598168708

[earnest forge]

Lmao

[magic oxide]

ah ic

[earnest forge]

43s for p1 is NUTS btw

[pine tiger]

that might be gpt
but i'm probably coping

[earnest forge]

Wait I'm 89th global now let's go

[covert arrow]

64/396 p2 choke again

[earnest forge]

Rip

[magic oxide]

nah today p1 looks doable in that time at least imo
I'm not good at writing quick code tho so idk

[covert arrow]

I didn't read and thought p2 was looking for ||the total score and not the number of cards||

[earnest forge]

I should really get used to using F5 to run my code lol it probably cost me several places

[compact wasp]

the time difference though lmao

[pine tiger]

ayo i accidentally left the pydis lb

can someone tell me how i can join again

[earnest forge]

/aoc join iirc

[pine tiger]

your guess is as good as mine

[compact wasp]

read from bottom to top

the top stuff is all useless

[magic oxide]

elves are pretty useless

[compact wasp]

honestly you could get this one just by reading the example

[naive stone]

feels like fate that lisp was chosen for today

[median juniper]

done

pretty confusing problem tbh

[earnest forge]

Also if your input library prints your input when it first fetches it (at puzzle unlock) you can start thinking about parsing while the puzzle loads

(e.g. for today, it's pretty clear that ||two splits + extracting ints|| is a good parse)

[covert arrow]

they've managed to fuck something up every year for 8 years straight, I would say they're more of a problem

[humble copper]

is there any way to get this linked to my github account? i made this current account with my gmail but i want it to be able to display my gh username and link to my gh page

[earnest forge]

If you have an AoC account with your github linked and any progress, contact eric

[median juniper]

toggle it in the settings

[earnest forge]

If not, make one, then IIRC there's a setting

Oh wait yeah you might be able to just link another oauth provider

[magic oxide]

yeah if your github account has no progress you can just migrate it across

I did that a few days ago myself

[earnest forge]

(in settings)

[feral hazel]

ah the example made me do a double take

Card 1: 41 48 83 86 17 | 83 86 6 31 17 9 48 53
In the above example, card 1 has five winning numbers (41, 48, 83, 86, and 17) […]
"wait what 41 isn't in the numbers"

[…] and eight numbers you have (83, 86, 6, 31, 17, 9, 48, and 53). Of the numbers you have, four of them (48, 83, 17, and 86) are winning numbers
(rereading it after risking it anyways and finishing) "oh"

[compact wasp]

yeah i had an int extractor and a multicharacter split thing prepared

[humble copper]

awesome tyyyy

[earnest forge]

(if the github-linked account has no progress you don't have to contact him)

[covert arrow]

I saw the example input and "Card 2 has two winning numbers" and I could tell what was going on

[compact wasp]

i only used the int extractor though bc i know there are 10 numbers on the left

[earnest forge]

Oh yeah I guess you could partition it like that

[humble copper]

it almost works but for some reason i cant get the username to link to the profile

ik theres a setting for it but it resets every time i click out

oh

nvm

theres a save button

💀

[earnest forge]

LMAO

[magic oxide]

💀

yeah that tripped me up as well when I was doing it

[humble copper]

there we go

[jovial jay]

Is there inherent disadvantage when it comes to timezones?

[earnest forge]

I mean being in the UK and having it at 5AM isn't great from a sleep POV

[magic oxide]

yeah probably

[median juniper]

doing this at 8am is pretty hard
(because i didnt sleep)

[magic oxide]

having it midday isn't very nice either

since I'm almost always out

[jovial jay]

Guess I'll accept never having a score

[median juniper]

midnight feels like a sweet spot

[magic oxide]

nighttime would be ideal imo

slightly before midnight

like 10-11 would be sweet

[median juniper]

yeah

[earnest forge]

Yeah I totally agree with that

[jovial jay]

I prefer at 9. That's ideal. It's 12:30 AM and I am in my bed.

Imagine coding a solution in Python from your phone

[earnest forge]

I mean QPython is a thing

[magic scaffold]

fumbled soo hard on part two lol

[deep ferry]

💀 so much space wasted already

[naive stone]

ugh, i'm so close with part 2. Works with the example but not my input. I have a headache though, not sure i can look at my screen much longer...

[woven sable]

this was a really confuisng problem

took me 10 minutes to read part 2

[worldly raven]

yeah, reading it at first was a little confusing, but once you understand it's pretty simple

[woven sable]

yeah

i still odn't understand it to an extent lmao

i finished but i kinda don't get it

like it hurts my head to think about it

[placid vessel]

Oh boy when I finally get home I will get creative

[median juniper]

turns out at my pace i would need ~8h per each AoC (whole year) if i wanted to solve previous years

[woven sable]

it took me 6h for one problem once in AOC in 2021 lmao

it really gets hard around day 15-20

[placid vessel]

I couldn't participate in 22 cause I went to work trip or smth

[woven sable]

i gave in 22

up

i j had no time with school

[placid vessel]

Now I am gonna complete it all

[placid vessel]

In a cursed way

[median juniper]

i will try to complete all AoC's until new year

[placid vessel]

My 25th solution probably gonna be written in hieroglyphs

[worldly raven]

i'm too lazy to go back and do the years i didn't finish

[cerulean epoch]

this is my first time doing aoc, so far the questions have been fairly straitforward. How hard does it get?

[woven sable]

check out day 18, day 19 and day 22 of 2021

notoriously hard problems

i would say u can get through it tho

[peak dock]

they get a lot worse

[woven sable]

where do u find this

[worldly raven]

https://adventofcode.com/events

[peak dock]

i swear ill finish 2022

[pine tiger]

lmao

[humble copper]

unpopular opinion: snailfish was fun

[pine tiger]

i'm sure you will!

[magic oxide]

I agree

[woven sable]

twins

[humble copper]

19 and 22 were from the pits of hell

[magic oxide]

snailfish was fun for me

[woven sable]

lowkey

[pine tiger]

what were those i kinda forgor

[woven sable]

it was hard though

[woven sable]

19

i hated it

so much

[magic oxide]

I didn't have much trouble with snailfish tbh

[humble copper]

beacons/scanners and reactor cubes

[woven sable]

i went to sleep at 4 am and woke up at 3 pm the next day and solved it bc i thought of the solutoin in my sleep 😭

[worldly raven]

i'm very inconsistent it turns out 😩

[peak dock]

oh god i remember beacons and scanners

[median juniper]

day22part1 doesnt sound very hard, just do the thing they described

[woven sable]

i'm pretty sure the hard part was part 2

i'll check

[pine tiger]

man i don't even remember what i did for those

[humble copper]

day 23 i stupidly decided to do without any code

just a whiteboard

[magic oxide]

oh god day 22 is cuboids

[pine tiger]

what was 23

[magic oxide]

no that was hard

amphipods

[peak dock]

naive solution needs unrealistic memory

[woven sable]

yikes 😭

this was part 2

[humble copper]

hes funny

[cerulean epoch]

I see... this is quite tedious to code

[flat sequoia]

I trolled

[humble copper]

2022 day 19 i cannot finish

no matter how i try

[flat sequoia]

I could not comprehend the problem for some reason

[humble copper]

not even pt 1 its like you'd think the most obvious solution would work

but it just doesnt

[scarlet wasp]

that's still on my todo-list as well, found it very difficult

[kindred badge]

what’s this

[median juniper]

100x100x100 = 1 million of items, no?

day 19 is scary

[compact wasp]

smh

[flat sequoia]

day 19 is the robots?

[woven sable]

the beacons

the worst

[humble copper]

yup

[flat sequoia]

There's not a particularly clever solution or anything you just have to prune the search space when you know it's no longer possible

[peak dock]

oh yeah part 1 was fine, part 2 is where it blows up

[woven sable]

those r my least fav problems

where the naive solution doesn't work

but i also learn the most so

[peak dock]

💀

im so sad i lost my solutions

[median juniper]

can you pls post part2 description here?

[flat sequoia]

the # timesteps just goes up from 24 to 32

[flat sequoia]

It's really the same solution though, a complete search

[peak dock]

https://paste.pythondiscord.com/BSEQ

hehe

[flat sequoia]

oh we're talking about different problems

[median juniper]

at this point it is easier to just count it by hand 💀

[humble copper]

underrated problem

2021 day 16 part 2

[woven sable]

day 16 was rlly cool

but i didn't understand it at all 💀

[peak dock]

i did 2021 day 23 by hand lmao

[woven sable]

was that the armapods

oh they're called amphipjods

i didn't get that one either lmao

[peak dock]

armpits

[pine ravine]

||for part b i had a 30 second run time using recusion, then i memoized it and went down to instant runtime, am proud||

[woven sable]

autocorrect 💀💀

[median juniper]

classic recursion

[median juniper]

i guess i will not be able to achieve this 😢

[ripe scroll]

what in the recursion is part 2 😦

[pine ravine]

recursion, bout 12 lines

i made that up

[magic scaffold]

DP problem

[weary kelp]

Day 4 is a bit early for a ||dp problem|| innit

[long jacinth]

well, todays easy!

[vestal token]

is it just me or was day 4 easier than day 3?

[magic scaffold]

still easy than day 1

[woven sable]

fs

[weary kelp]

True dat, weekend shenanigans I guess

[covert arrow]

wait until next year when december 1 is a sunday

[magic scaffold]

RE: Day 22: Monkey Map [2022]

[long jacinth]

oh is that what it is, are sundays tough?

[covert arrow]

puzzles on the weekend tend to be slightly harder

[long jacinth]

i see, its my first year this year so i have no prior experience, but i guess that does work great

[hazy ermine]

wait really?

[untold briar]

Man I hate it when my code runs on the small testcase but not the big bad input

[atomic cairn]

Ah yes the feeling of the coming 3 hours consisting at looking at your screen to see if the numbers are correct

[untold briar]

Yes yes, the period of pure despair

Which lasts until we find where things went wrong

[atomic cairn]

And then you fix it and it doesnt work

[untold briar]

Back to the period of pure despair, rinse and repeat until you got the approval message

[modern terrace]

i don't have one 👍

[cyan sleet]

I think the randomness in the roulette is broken 😛

3rd functional in a row

[magic oxide]

there is no random

kat is the roulette

[rapid wolf]

you can write imperative lisp if you want to, it is supported in common lisp

[cyan sleet]

also which lisp? pick whatever?

[magic oxide]

flip a coin

[tall silo]

Someone has probably already mentioned this but everyone who hasnt done any days in the python discord AoC leaderboard is tied at 69

[cyan sleet]

how many sides does your coin have?
https://en.wikipedia.org/wiki/List_of_Lisp-family_programming_languages

[magic oxide]

flip a coin 6 times

[cyan sleet]

the boring answer is to use common lisp

(is python a lisp?)

(it sure can process lists)

[peak dock]

(for (sure))

[gloomy kernel]

I'll do day 4 in rust

[wicked elm]

difficulty really be completely random

[hidden musk]

kinda weird tbh

I got lucky the past 2 days but now I've never used any lisp

[hidden musk]

maybe only 2 are needed

Today, the most widely known general-purpose Lisp dialects are Common Lisp and Scheme.

[cyan sleet]

I'm doing clisp and I hate everything

granted my baseline hate is high

[hidden musk]

i think i'll do clojure

i think people still use that

[winged tartan]

guys, if i want to talk about the implementation of a solution that im struggling with, do I ask in the python threads or in here?

[earnest oriole]

why are there only 10 people in the leaderboard

[earnest forge]

If you're talking about a particular private leaderboard on the aoc website, it's because pydis uses lots of them due to the limit of 200 users per private leaderboard

[rocky spire]

I'd say a thread to be safe

[earnest forge]

If you're talking about the command, it's because they removed the full leaderboard paste

[earnest oriole]

i joined this

[earnest forge]

That's for people doing the language roulette

[earnest oriole]

oh

damn

[earnest forge]

Possibly there's only 10 of them :P

[hidden musk]

😭

[earnest oriole]

how do i join the regular leaderboard

[earnest forge]

/aoc join

[earnest oriole]

here?

[lime fulcrum]

i mean does it count as ||dp when you just add to the number of that ticket you got?||

[modern terrace]

it's ||dp||??

[earnest forge]

(yes)

[modern terrace]

i didn't know that

[earnest forge]

(also yes)

[lime fulcrum]

is it? i just looped it somehow

[earnest oriole]

thanks

[lime fulcrum]

didn't know there is a special name for that

[earnest oriole]

idk dp and i did it

[earnest forge]

Assuming you ||kept a list of how many copies of each ticket you had and only looped through the input once||, yes that counts as dp

[earnest oriole]

i guess i used it then

[lime fulcrum]

||combined in the same list. just kept track of the number for each line||

[rocky spire]

I'd say ||dp|| is just a very fancy word for a ||recursive relation but you save the result for reuse later||

[earnest forge]

Technically I didn't say ||kept a separate list||, now did I :P

[hidden musk]

it's very hard to resist clicking

[rocky spire]

It's... well, spoilers, spoiling the idea of today's AoC
So if you don't mind that just click (?

[lime fulcrum]

sorry. should i delete?

[hidden musk]

no it's fine lol

I'm just saying

[rocky spire]

Few mortals can resist the temptation of the clickable black bars

[earnest forge]

||pop||||pop||||pop||||pop||||pop||||pop||
||pop||||pop||||pop||||pop||||pop||||pop||
||pop||||pop||||pop||||pop||||pop||||pop||
||pop||||pop||||pop||||pop||||pop||||pop||
||pop||||pop||||pop||||pop||||pop||||pop||
||pop||||pop||||pop||||pop||||pop||||pop||

[lime fulcrum]

nice

[rocky spire]

Dude my friend sent me that exact same message but bigger just a few days ago

[marsh current]

:incoming_envelope: :ok_hand: applied timeout to @earnest forge until <t:1701695412:f> (10 minutes) (reason: emoji spam - sent 100 emojis).

The <@&831776746206265384> have been alerted for review.

[modern terrace]

oopsies

[hidden musk]

bannable imo

[marble schooner]

100 is a lot, good lord

[magic oxide]

damn

[rocky spire]

Maybe he sent this but bigger & emoji

[modern terrace]

oh

minesweeper

that's a possible thing

[sinful pewter]

So 100 is the limit

[marble schooner]

!unmute @208918673178492929 just keep it ontopic please

[marsh current]

Bad argument

Could not convert "user" into UnambiguousMember or UnambiguousUser.
@208918673178492929 is not a User mention, a User ID or a Username in the format name#discriminator.

[hidden musk]

no, it's more like 10

[marble schooner]

!unmute 232948417087668235

[marsh current]

:incoming_envelope: :ok_hand: pardoned infraction timeout for @earnest forge.

[earnest forge]

Sorry lmao

(and yes it was minesweeper)

[sinful pewter]

Got it

[earnest forge]

It's 20, I checked

4x4 minesweeper anyone? /j

[peak dock]

we may or may not have a minesweeper pinned in ot1

[earnest forge]

Yeah (just checked) but it isn't fixed-width lmao

[sinful pewter]

Guys, slightly off topic. How do I share a PyCon event in this discord? PyCon Pakistan is happening after a for few years.

[hidden musk]

a few years?

[sinful pewter]

Sorry it after a few years

[marble schooner]

Reach out to @plucky beacon if you are one of the event organizers, otherwise, suggest it in #community-meta

[sinful pewter]

Got it, thank you

[sleek cave]

snow day :D

what a beautiful time to grind aoc from morning til dark

goal today is to get aoc '21 dirac dice and day 25 of '22 cus i missed that freebie lol

and maybe beacon scan

[tepid cairn]

lucky 😩

[young tree]

does greedy match with a negative lookahead not work?

[onyx flicker]

wym? why not? what's your pattern?

[hidden musk]

clojure is hard bruh

I just got part 1

[vestal breach]

i found todays easier cuz less regex

[unique osprey]

Question for day 3 p1 as I'm just getting around to it: ||do we count multiple occurrences of the same part number on the board indivually?||

[pine tiger]

yeah i think so(?)

[tepid cairn]

https://youtu.be/jlPaby7suOc?si=wpkjkt5SjczjDrlV

[unique osprey]

Correct. Thanks, found my bug(s)

[noble skiff]

Gotta say this is easily the hardest first 4 days I remember for AoC

[magic oxide]

1 and 3 were definitely anomalous imo

2 and 4 seemed standard

following the pattern, tmr is going to be very difficult

[simple raven]

day4 part1 was a nice break after yesterday 😄

[sinful pewter]

Top3 at work are separated by 1 point each. 1st one is the only female among the 10 of us. This year is a blast.

[cyan sleet]

&aoc lb

[shut trellis]

Wrong Channel

You are not allowed to use that command here. Please use the #aoc-bot-commands, #sir-lancebot-playground, #bot-commands channel(s) instead.

[hidden musk]

@shut trellis nerd

[magic oxide]

@long egret nerd

[long egret]

What

[noble skiff]

I'd agree about 2, but even 4 part 2) has a bit of a twist.
For comparison, last year's part 4 was literally just taking pairs of integer ranges, and seeing if a) one is a union of the other, then b) if they have non-zero intersection at all

very mechanical. For 4 part 2 you do need to be a bit careful propagating data forward and the most naive solutions will fail (and some already have)

I was mostly thinking of the overall trend though. It seems significant/consistent enough that it feels deliberate. I wonder if the goal is to make AoC harder as a whole, or just to make the start more challenging.

[hardy zenith]

my naive solution ran incredibly slow

I just used ||a mapping to buckets and filled the future buckets using a nested for loop||

[noble skiff]

yeah

[hardy zenith]

if I was smart, I'd ||cache the winning results so I wasn't re-calculating every time I needed to propagate, I bet @lru_cache would've solved that nicely||

though I'm not sure how that works with copying, maybe it doesn't work nicely

What is the report command again?

eh, can't find anything in the info channels, so I'll do a tag thing.

<@&831776746206265384> someone trying to solicit private sales off-server

user: 893177187723788318
message: 1181310928034214050

[sleek cave]

aoc '21 day 21: dirac dice

this is the most cursed thing i've written in some time

https://paste.pythondiscord.com/5JBA

my logic was so straightforward 😂

just combinatorics everything with DP

leading to 8 levels of indentation

solutions referenced: salt's and david bartram's, except i wanted to solve it another way lol

[torn sapphire]

@plain sluice don't post ads like this in this server

[torn sapphire]

Ty for the report - for future reference though, you can DM @plucky beacon for getting our attention for stuff like this too

[compact wasp]

i'm willing to bet it's a gpt countermeasure

so my guess is that the first 10 or so days will be more difficult than usual, and after that it will be around the same level as the previous years

[modern harness]

That is a really good point that I didn't think of despite noticing the problems seemed harder this year. They potentially even pre-screened these questions against chatGPT. Based on their announcement banning chatGPT, it seemed they might be mainly focused on preventing 100% automated solutions, which I wouldn't be surprised if the difficulty increase on the early problems does just that, though I haven't tried putting any of these into chatgpt, but I thought I recalled the fully automated solutions last year stopping at some point when they got a bit more challenging.

[noble skiff]

idk, I'd want additional evidence to believe that

there's fluctuations in the difficulty from year to year

but it's definitely possible

it would be very cool if the author goes on record about this at some point

I'm also not really sure (literally) to what degree human difficulty and AI difficulty match up.

[hidden musk]

we need john henry pt 2 for aoc problems

[noble skiff]

I've started benchmarking stuff

I'm debating whether I should benchmark solving both 1 and 2 together, or each individually

I know that if I do teh former I'm going to rewrite code in some pretty awkward ways to share stuff

day 4 is a pretty good example. counting matches for each line is significant, so you would want to share that between 1 and 2.
but dumping the number of matches into a vec is a needless heap allocation so you'd want to avoid that

so you'd need to have like, an "object" solver for part 1 and part 2, that has an update method, that recieves the number of matches for each line

[short karma]

bruh what the hail is this scartch card rule system XD

[minor cave]

Really fighting the urge to do d3p1 by hand.......

[sinful pewter]

I am trying to find the perfect regex

Technically doing it by hand

Considering the pain

Why do I do this to myself?

[royal gull]

If you're stuck or need a hint, ask in #aoc-solution-hints , happy to help with regexs.

[sinful pewter]

Often times, when I am feeling down, someone makes me feel elated to be part of such a vibrant community of nerds and geeks. Thank you so much. I will ask in #aoc-solution-hints. It's easy to isolate yourself.

[cyan sleet]

is there a 2d regex?

[sinful pewter]

I don't think so, haven't heard of such a thing

[sinful pewter]

But you are thinking about 2D so you are on the right path

[cyan sleet]

I was mostly just curious if there was such a thing

I did day 3 with quite different machinery while wrestling with haskell 🥴

[hidden musk]

😩

[terse comet]

エニラ joked about creating 2D regex lol
(...at least I think that it was only a joke?)

[earnest forge]

Tbh I don't see how that could possibly work

[cyan sleet]

a(?:↑b+)c

[earnest forge]

Lol

[terse comet]

idk, it would probably end up very esoteric, but I could imagine it being a thing

[earnest forge]

But now is C above the b+ or to the right of it?

[cyan sleet]

I imagined something like

.b.
.b.
abc

[earnest forge]

The only way I can really see that working is if you actually write it as a 2D string, but that would be painful to write in an actually fair way (you'd need some kind of defined padding character)

[cyan sleet]

(?:) for scoping directions 🥴

[terse comet]

looks like https://esolangs.org/wiki/Category:Two-dimensional_languages is a thing

[cyan sleet]

or maybe it just being a direction change makes more sense

[royal gull]

or maybe as a 1D string, but using line length modulo (offset?)

[cyan sleet]

abcba↑ba←a*↓c+

..aaa
..c.b
abcba
..c..

[earnest forge]

How about:

abcd((?vefgh)|(?^ijkl))

   l
   k
   j
   i
abcd

abcd
   e
   f
   g
   h

[cyan sleet]

↶↷ for relative turns

v for down is a brave choice

[earnest forge]

((?<), (?>), (?^), and (?v) are all disallowed sequences in 1D regex AFAIK)

[earnest forge]

What would be your (ASCII-only) choice?

[cyan sleet]

why do we need to restrict ourselves to ascii?

[earnest forge]

Because we're doing good and proper language design here

[cyan sleet]

?< ?> feels quite similar to lookaround

[earnest forge]

oh crap yeah ?<= is lookaround

Maybe ?{direction}

Heck it could even extend to 3D regex with * and X 🙃

ok and?

[cyan sleet]

?{NW}

[earnest forge]

Ehh ascii art arrows are better

[cyan sleet]

you want ?{^<} ?

[earnest forge]

no, no diagonals

^v<>*X only

(*X being for 3D)

[cyan sleet]

clearly I meant north-west 😛

N S W E NW ... SE L R B

[earnest forge]

Is this implying that the regex depends on your system compass

[cyan sleet]

I guess F for completeness

that would be a hilarious esolang feature

[hidden musk]

i guess we're looking down on the world then

[cyan sleet]

aren't we always?

[hidden musk]

only the haskellers

and occasionamly the ocamlers

[earnest forge]

And rustaceans 🙃

(but only because we are better)

[hidden musk]

rustaceans are descendants of the ocamlers

[cyan sleet]

"point your device northwest for the code to run as expected"

[earnest forge]

'works on my machine' to a whole new level

[cyan sleet]

it annoys me that on mobile 🧭 shows it pointing NW, looking at details shows it pointing NE

[lusty rover]

[earnest forge]

That sounds like an apple thing

[cyan sleet]

it's not

[earnest forge]

Oh just discord being very dumb

I know that they have to use apple emoji on ios for some reason, didn't realise it's just broken in that menu on android

[cyan sleet]

discord normally uses twemoji

but apparently not everywhere

[earnest forge]

I'm pretty sure that's new since the change to RN

[simple raven]

Day4Part2 was so hard for me 😦 Had to read through a couple solutions to understand how to write the recursion but I finally solved it

[earnest forge]

(and frankly the RN change sucked)

((if they're going to change the android codebase they should use flutter))

[cyan sleet]

I didn't even do a recursive solution

[earnest forge]

Me neither

[hidden musk]

me neither

i just reduced over the lines

[cyan sleet]

it was iterative in nature (but written recursively because lisp)

[simple raven]

hm... maybe I overcomplicated it then

[hidden musk]

oh? i didn't recurse at all

[cyan sleet]

I could have done a reduce I guess

with state being kept in the accumulator

[hidden musk]

yeah. clojure has these cool "reducer" things which are sorta similar to concurrent.futures.Executor.map

super easy to parallelize things by just adding another argument to the function call

[cyan sleet]

oh right, you have a nice lisp

[simple raven]

I saw someone ||make a list and then keep adding copies of the new cards to it while iterating over that list which eventually went over all the copies which I thought was neat||

[hidden musk]

that's pretty funny

[simple raven]

? is this a function or something?

[hidden musk]

in python it's functools.reduce

[cyan sleet]

you don't need to iterate over all copies anyway

[hidden musk]

i guess you could end up with a solution that touches all the copies somehow. but that would have a fun time complexity

[cyan sleet]

oh, people wrote such things in here

[simple raven]

oh I completely forgot about this function, I suppose it's not recursion but achieves the same

[earnest forge]

I can't help but feel like it's O(n^n) *(worst case) lmao

[cyan sleet]

took like 16 seconds, fixing it made it 250x faster

[simple raven]

my part2 solution runtime is 7.47s 🙂

[cyan sleet]

it's bounded by 2^n for sure

probably significantly less with our constraints

[hidden musk]

your branching factor is at most the number of winning numbers isn't it?

[earnest forge]

That's fair, m^n

[lusty rover]

I tried that, but I kept ending up in infinite loops
I ended up changing it slightly: https://github.com/shenanigansd/scratchpad/blob/main/events/advent_of_code/2023/04/python/aoc_2023_04.py

[hidden musk]

why do you have a Deck class 😩. they're not playing cards

[simple raven]

holy classes 😄 nice

[earnest forge]

It's not infinite it's just about 9 million

[lusty rover]

It was the first name that came to mind

[simple raven]

a collection of cards is a deck so it checks out

[cyan sleet]

if we assume that we add to every following card it's 2^n, and we don't allow even close to that

1 1 1 1 1 1
  2 2 2 2 2
    4 4 4 4
      8 8 8 ...

[earnest forge]

Oh yeah fair enough

[simple raven]

    @classmethod
    def from_cards(cls, cards: list[Card]):
        return cls(cards)
``` what's this for? you don't use it anywhere

[lusty rover]

Only 1M for me
But yeah, in retrospect, my list solution may have worked if I had just let it play out
I just assumed it was broken because I saw it printing out giant lists
I might try rebuilding it and seeing what happens

[cyan sleet]

I'm kinda curious what the expression is

for adding to the next k

[earnest forge]

Mb I meant 5.9 million

Misremembered the number

I'm surprised yours is as low as 1M though

[hidden musk]

huh. i got 1.1e7

[lusty rover]

Huh
You’re right
I guess it can just be removed then

[earnest forge]

Anyway time to go get 4h30 of core sleep before waking up for AoC tomorrow 🙃 o/ see you all later

[simple raven]

what's cls for anyway? I've seen it in some repos before

or like what does it stand for

[lusty rover]

Your puzzle answer was 5920640.

Ah, 5M
I misread as 500k

[simple raven]

I had 8805731, took a while to get done 😄

[lusty rover]

class
It operates on the class itself, instead of a specific instance