But you underestimate how large that is
#AdVENt of Code
solemn python
Not to mention having to calculate the actual value of those, bc the formula only deals with the count
sullen fiber
||binomial coeff|| in uiua is so pain
sullen fiber
But you underestimate how large that is
nah i think ur overestimating
for my input given that i have a way to iterate through them it reduces my iteration count from 5.5mil to 32k
||naive 2^k vs nCr counts||
putting this here in case i need it again later cus it was a pain to write
||```
Naive 2^k count
/+ ⁿ:2 ⊐≡/+ =@? ≡⊡ 0 Table
nCr count
Fac ← /×+1⇡
≡⍜°⊟(⊃(-:⊓⧻/+|⋅⧻) ∩°□) Table
≡(÷:⊃(Fac|×∩Fac ⊃(⋅∘|-:)) °⊟)
/+
@solemn python I KNOW
I KNOW HOW TO ITERATE
I KNOW
i'll write about it later
solar tide
I will delete them today
sullen fiber
i'll write about it later
||basically just read the "row partition" section of the article i linked earlier, im just gonna iterate through all the a sets that satisfy the system||
tranquil vapor
oh real
sullen fiber
i had part 2 spoiled for me so i alr know it wont be worth waiting
kinda glad i did tho cus i would have been annoyed otherwise
like eagely said it feels lazy
tranquil vapor
WHY DID U HAVE P2 SPOILED
hazy sinew
maybe if u post it to github and put the unlicense the copycat can blame you in case of a lawsuit
sullen fiber
simple wont license
hazy sinew
i think doing it like a normal person is faster
sullen fiber
its not
in uiua
uiua is NOT fast
imagine uiua is python but even python would be faster at a bruteforce
sullen fiber
WHY DID U HAVE P2 SPOILED
not fully just the jist
hazy sinew
uiua is NOT fast
it takes 200ms in kotlin
sullen fiber
it would take 30m for part 1 in uiua
hazy sinew
today is gonna take seconds as well then ig
sullen fiber
so ur saying a solution exists that between naive and what im doing rn?
hmm
i can think of another way but its basically the same just
hazy sinew
its literally ||part 1 with a cache||
hazy sinew
||recursive instead of iterative||
both work
hazy sinew
||a cache??||
yes
sullen fiber
||ngl if i wasnt doing this at a library i could use perf and find out what was making my impl slow||
||but yea that solution where u walk through the 2^k possibilities and check if they match, it is super slow in uiua||
wait i thought of an optimization
nope slower
fuck u uiua
wait i thought of another
hazy sinew
bro just ||bruteforce it with memoization pls||
sullen fiber
explain or die
tranquil vapor
|| im about to brute force my fist in your face with memoisation (you will never forget) ||
sullen fiber
TRUE
hazy sinew
explain or die
what did u do for part 1
sullen fiber
||count the ?s as k, iterate through 2^k as k-bits, use the bits as a mask to insert # or . into the ?s, count the partitions for each iteration and check if they match and if so increase accumulator||
||i dont actually use strings but yea its basically that||
also i just realized why it was slow and i made it a lot faster
from 30m to 3m
still kinda slow
hazy sinew
im trying to understand
ok thats smart
but idk how youd use it for part 2
i did ||iterate through each string recursively, check the first char, recursively add the function to total for the rest of the string||
||pass in the config and the group sizes, check first of config if its in ".?" then result += count(config without first char, sizes) if its in "#?" then check if there is a . or a ? after the group (substr until groupsize) and if yes add count (config substr until groupsize, group sizes)||
that was a terrible explanation
sullen fiber
nah it made sense i think
its basically the same as mine but recursive rather than iterative
i dont know if you can do more than one recursive call in uiua
recursion is experimental
tranquil vapor
@sullen fiber IS THIS YOUR ALT https://mk.absturztau.be/notes/9n67o737plebb937
daring marlinBOT
<@454072114492866560> IS THIS YOUR ALT https://mk.absturztau.be/notes/9n67o737p...
sullen fiber
thats just the uiua frontpage
mandelbrot so easy in uiua tho uiua so cracked
also you love large matrices displaying as images
it can also do this in ur terminal lol
@tranquil vapor https://www.uiua.org/
click the arrows next to run NOW
ok i was wrong @solemn python its so bad for part 2
im just gonna typescript today with eagle's idea
solemn python
ok i was wrong <@487443883127472129> its so bad for part 2
Yeah I know lmao
I was testing it on P2 not P1 and needed to go bigger than i128s lol
Factorials are crazy
hazy sinew
why tf is 🚓 🚑 🚒 right to left
tranquil vapor
<@343383572805058560> https://www.uiua.org/
SCARY
steady fog
@tranquil vapor learn Hebrew
sullen fiber
Is Hebrew rtl
solemn python
aaaaaa
i hate this math stuff
I dont know how any of it works and am just getting numbers with no clue how they relate to each other
solemn python
technically both
solemn python
pArt 1 looks like it could be solved using math
well both can, because its the same problem, one is just magnitudes larger
solemn python
as with like 1/2 the p2s this year
steady fog
what should first step of p1 be
solemn python
I have math
idk what to do with it
but I have numbers
solemn python
what should first step of p1 be
solve the problem
solemn python
especially considering its arbitrary precision arithmetic rather than fixed precision
but anyways
steady fog
the number is how many possibilities right
tranquil vapor
what should first step of p1 be
first step: parse input
steady fog
I did
solemn python
the number is how many possibilities right
first number is the amount of possibilities that the 1,2,2 stuff is valid
given the length of the springs
the 2nd is the amount of possible arrangements for the springs list in general
the 3rd is the amount of possible spring lists when replacing the ?s with every possible option
solemn python
ok i give up on the math for now
i just can't think of a way to calculate this without at least generating some of the spring list stuff
solemn python
should I solve this with ai
ok yeah I give up for real this time
there might be a way but I don't know enough about probability or math in general to figure this out
if you find it vap pls tell im curious
solemn python
b/c its in my search query
steady fog
of
tranquil vapor
ty the type of fella to google for of row
solemn python
so true
sullen fiber
I just found a polynomial time algorithm to do this
That's crazy
O(n^2) rather than O(2^n)
solemn python
IHATERECURSIONIHATERECURSIONIHATERECURSIONIHATERECURSIONIHATERECURSIONIHATERECURSIONIHATERECURSIONIHATERECURSION
im trying to do this the smart way but it is so hard to debug or even conceptualize recursion
delicate crane
Recursion is never viable for me because I cant get tail call optimization to work in js. (I don't know if it even exists)
sullen fiber
TMR I'll try to write the polynomial time algorithm in the morning
In rust
@solemn python
Or maybe tonight if I use my friends laptop
solemn python
time to delete my recursive function and try again for the 3rd time
ill get it eventually probably
ok well I looked at the megathread for a hint or two and now I think I figured it out
its actually pretty simple once I actually thought through what the function was supposed to do and added comments to literally every ifelse path
NOOOOOOO
i made it so far without worrying about the borrow checker and lifetimes
and when I go to cache shit I realize I can't add a slice to a map
native pewter
Recursion is never viable for me because I cant get tail call optimization to wo...
Update: As of July 22, 2023 Safari is the only browser that supports tail call optimization.
use bun maybe?
delicate crane
Good excuse to finally try bun
native pewter
guh fucking wasm has tco
warped dust
tail calls can always be rewritten as loops
delicate crane
guh fucking wasm has tco
So I should use porffor
sullen fiber
tail calls can always be rewritten as loops
This true simply do@delicate crane
delicate crane
but recursion less lines and less indent than loop
steady fog
@delicate crane have you done day 8
delicate crane
will not wait
delicate crane
did you just steal my code
steady fog
yes
delicate crane
steady fog
evil
delicate crane
steady fog
@delicate crane hi
delicate crane
unsteal my code
steady fog
never
solemn python
LOVE
steady fog
@solemn python hi
tyminky
@solemn python get ready
well i was gonna do gleam
but its dum and wont let me do a single file
haskell it is
delicate crane
I dislike this
teal mountain
i kept confusing rows and columns
teal mountain
i did now what i should've done from the start, just write one version and transpose
solemn python
p2 im just going to throw some code at it and hope it works
teal mountain
can confirm this is the correct intuition for p2
teal mountain
p1 took me 25 mins, p2 38
solemn python
<@487443883127472129> how bad is it
see above
teal mountain
it's not that bad tbh
solemn python
I should make a 2d array lib for rust
because the amount of times I have used 2d arrays is crazy
steady fog
I should make a 2d array lib for rust
do not
solemn python
i will
delicate crane
where is the reflection in this shit?
##.##.#.#..##
##.##.#.#...#
.#.###......#
.###.##..#..#
##.#.##....##
.#..###.###.#
.#...#...#.##
#.#.##.#...##
#.###.#.##.#.
#.#.#...####.
#.#.#...####.
#.###.#.##.#.
#.#.##.#...##
rows 10&11
delicate crane
ah
sullen fiber
because the amount of times I have used 2d arrays is crazy
Oh man do I have a language for you
solemn python
where is the reflection in this shit?
```
##.##.#.#..##
##.##.#.#...#
.#.###.......
question for p2 solvers: ||the smudge is second to last column, first row?||
delicate crane
..###..##..
.#.#.....##
#.##..##.#.
#.#..###..#
.#..##.##.#
.#....##..#
#.####.##.#
#.####.##.#
.#....###.#
.#..##.##.#
#.#..###..#
#.##..##.#.
.#.#.....##
..###..##..
..###..##..
where is the reflection in this one?
its not 7.5
I swear to fucking god if 14 and 15 counts as a center
steady fog
I see it
solemn python
I swear to fucking god if 14 and 15 counts as a center
it does
all it has to do is hit the edges
not have more than 2 lines reflected
delicate crane
I way overthought reflections
my code that worked is way simpler
for p2 solvers, what the actual fuck?
solemn python
workin on it
delicate crane
ngl p2 is easy edit of p1
solemn python
it probably would be if I wasn't so bad at programming
I know exactly how to do it, my setup is just horrifying
its only vaguely usable
delicate crane
||in your function for determining if a specifc center is actually a mirror
- keep track of smudgeFound (default to true for part 1)
- replace line equivilance checker with oine that will allow smudge only if !smudgeFound
- when equivilance checker uses a fuzzy, set smudgeFound to true
- only return line is mirror if smudgeFound
||
delicate crane
BUT WHY DOES MY CODE OUTPUT WRONG
||```js
const vert = findReflectionVert(pattern)
if (vert !== -1)
return vert
const horz = findReflectionHorz(pattern)
if (horz === -1) console.error(i) // <--- add this
return horz * 100
for me it was because I wasn't finding all the reflections, so add code to see if you have the same problemUiuable
sullen fiber
YES
delicate crane
dreeam of solution
sullen fiber
I already am having those dreams
delicate crane
I can't sleep once I see an AOC problem
brain will keep working on solution and I cant sleep until I solve it
solemn python
ok I need to sleep
I have 2 finals tomorrow
I don't want to have a backlog of p2s 
delicate crane
3ms for p1 + p2
p1: ~1.5ms
p2: ~1.5ms
delicate crane
gonna try and see what multithread does to that time because I want to tuck with the js lib I found yesterday
solemn python
ZLJDKLSJLKDSJKJKDSJKLSAJDSLAS
tip: it helps to read the problem before overcomplicating everything
delicate crane
I didn't read too much of p2, please explain your mistake
solemn python
I didn't realize all you needed for the answer was the ||smudge'd reflection||, not the rest
delicate crane
why would you keep bad data in your (very good) summary of all the patterns
I absolutely love how the outputs to the problems wouldnt help at all in most cases because how they are joined together
solemn python
if it can figure out how to fetch the example yeah I can just add --input example and it will run on example input and even comare it to the solution
real hinge
is part2 terrilble again
solemn python
it manages to get p1 example solution but not p2 for some reason
delicate crane
aaaaaaaaaaaaaaaaaaa
Are you sure your only ||accepting smudge corrected mirrors||?
solemn python
Are you sure your only ||accepting smudge corrected mirrors||?
I short-circuit when I find lines that ||are mirrored, minus one char||
delicate crane
No clue
delicate crane
@real hinge did you read my advice last time you asked?
real hinge
<@287555395151593473> did you read my advice last time you asked?
didnt see
lememe look
delicate crane
||```js
const vert = findReflectionVert(pattern)
if (vert !== -1)
return vert
...
.
delicate crane
You are telling me you don't in any part of your code, loop over each pattern and check for vertical/horizontal mirrors
real hinge
terrible
real hinge
You are telling me you don't in any part of your code, loop over each pattern an...
I check each group horizontal first
then rotate the input
and use same function
delicate crane
||question literally says to check vertical first and ignore horizontal for stuff that had vertical||
real hinge
question can explode
I am getting too low anyway
oh I found the problem
this has 2 mirrors
but first one is invalid
so it eplodes
delicate crane
Explod
real hinge
okay completed first one
it turns out I neeeded to trim the group 
it was treating empty line as a gorup
tranquil vapor
i fucking hate rows and columns in coding
cause if u use 2d array you need to arr[y][x]
tranquil vapor
at least tell me if its high or low
you're one off 
hazy sinew
i had to do it using lists (ugly) instead of grid because it for some reason was returning an answer lower by like 300 but correct for example and i couldnt figure out why
||
private fun findMirror(grid: List<String>, smudges: Int = 0) = (1 until grid.size).firstOrNull { grid.take(it).asReversed().zip(grid.drop(it)) { a, b -> a.zip(b).count { (a, b) -> a != b } }.sum() == smudges } ?: 0
private fun transpose(grid: List<String>) = List(grid.first().length) { i -> grid.joinToString("") { it[i].toString() } }
override fun solvePart1(input: File) = input.sdnl().sumOf { findMirror(it.lines()) * 100 + findMirror(transpose(it.lines())) }
override fun solvePart2(input: File) = input.sdnl().sumOf { findMirror(it.lines(), 1) * 100 + findMirror(transpose(it.lines()), 1) }
||
this would be a lot longer with grid, which is the opposite of what im trying to achieve with it
solemn python
Sleep with unsolved code problems
You see, I have finals today
Which is a nice motivator
solemn python
this has 2 mirrors
you will love p2
hazy sinew
this has 2 mirrors
i guess i had that too, multiple times but when i did list implementation instead of grid i accidentally ordered it correctly
real hinge
<:trolley:1024751352028602449> you will love p2
It doesnt seem too bad
steady fog
hope you fail 
native pewter
doesnt matter if i fail cause its just bonus points
either way, not very nice zt
das y ur in da hole
steady fog
NO
sullen fiber
||
```kotlin
private fun findMirror(grid: List<String>, smudges: Int = 0) = ...
oh wow
i dont like that this is arguably more concise than my uiua
delicate crane
i fucking hate rows and columns in coding
Just write something for rows, then rotate the array for columns
sullen fiber
||or just transpose the array, by swapping x and y||
@solemn python are u still curious about the polynomial time alg for yesterday
sullen fiber
u mean like how u did in part 1
real hinge
so variable changing applies vertically too
sullen fiber
variable changing 😨
sullen fiber
FIRE
hazy sinew
||or just transpose the array, by swapping x and y||
i tried that, it didnt work, so i did it with lists instead of custom grid, and it started working
rip leaderboard i guess
sullen fiber
thats what i did lol weird
hazy sinew
i mean it did work but not with my custom grid for some reason
i rewrote it with lists
and it magically started working
hazy sinew
i was getting like 300 and change below correct answer
hazy sinew
kotlin looks so hot
ikr thats why im learning python
hazy sinew
yes
sullen fiber
so fire
hazy sinew
js and umm uhh
sullen fiber
native i hope
hazy sinew
ive never met anyone that uses it
hazy sinew
native i hope
yes
sullen fiber
fire
sullen fiber
wdym
hazy sinew
its super slow in console
sullen fiber
wdym in console
hazy sinew
with graalvm on jvm it competes with c
hazy sinew
wdym in console
u can type kotlin in console
sullen fiber
like a REPL?
hazy sinew
and itll open up one of those input thingies
hazy sinew
whats a repl
sullen fiber
not
hazy sinew
did u know theres a function
inline fun Boolean.not() = !this
also theres 2 types of kotlin users
list.filter { it.isNotEmpty() }
and
list.filterNot { it.isEmpty() }
list.filter { it.isEmpty().not() }
list.filter { !it.isEmpty() } better
native pewter
u can type kotlin in console
hop on julia repl
native pewter
kotlin seems like a stupid language to repl
steady fog
also theres 2 types of kotlin users
```kotlin
list.filter { it.isNotEmpty() }
``...
function reference
list.filter(String::isNotEmpty) assuming it's a list of strings
sullen fiber
anyone else notice that it seems like it ends on day 14
like how its visually laid out
i wonder if the layout will change drastically tmr or if it just will extend
steady fog
sullen fiber
minmax ur stackframes
hazy sinew
i wonder if the layout will change drastically tmr or if it just will extend
no it it will just end on day 14
steady fog
nop
sullen fiber
no it it will just end on day 14
true
tranquil vapor
zt loves func references in crystal
steady fog
yop
tranquil vapor
/run ```cr
puts ["zt", "banana"].map &.upcase
split heartBOT
Here is your cr(0.36.1) output @tranquil vapor
["ZT", "BANANA"]
very good
split heartBOT
@hazy sinew I received kotlin(1.8.20) compile errors
file0.code.kt:1:1: error: expecting a top level declaration
println("Hello World!")
^
file0.code.kt:1:8: error: expecting a top level declaration
println("Hello World!")
^
file0.code.kt:1:9: error: expecting a top level declaration
println("Hello World!")
^
file0.code.kt:1:10: error: expecting a top level declaration
println("Hello World!")
^
file0.code.kt:1:22: error: expecting a top level declaration
println("Hello World!")
^
file0.code.kt:1:23: error: expecting a top level declaration
println("Hello World!")
^
Error: Unable to access jarfile code.jar
/run
fun main() =
println("Hello World!")
Here is your kotlin(1.8.20) output @hazy sinew
Hello World!
omg it works
/run
fun main() = "me when".map { it.toInt() }.reduce(Int::xor).let { it * it }.toString(16)
@hazy sinew I only received kotlin(1.8.20) error output
Error: Invalid or corrupt jarfile code.jar
/run
fun main() = println("me when".map { it.toInt() }.reduce(Int::xor).let { it * it }.toString(16))
@hazy sinew I only received kotlin(1.8.20) error output
Error: Unable to access jarfile code.jar
/run
fun main() =
println("me when".map { it.toInt() }.reduce(Int::xor).let { it * it }.toString(16))
@hazy sinew I only received kotlin(1.8.20) error output
Error: Unable to access jarfile code.jar
stupid bot
steady fog
agreed
native pewter
smart
tranquil vapor
hazy sinew
im trying to understand berlekamp-massey
wtf is he cooking
ok i get it
its not useful
real hinge
les goo
I finished day 12 part 2 too
ig its good
god now I feel relaxed
full gold
tranquil vapor
wont feel relax after dinosaur attack
real hinge
where will I get dinosour attack
real hinge
tranquil vapor
real hinge
🦖
sullen fiber
wow timescript is super cursed
@warped dust look
[number, ...number[]] is meh but D & {} is crazy..
warped dust
D & {} is just non nullable
sullen fiber
so
if u just do D, and u do new Tensor([5, 5], x => x)
instead of inferring from the first D, and using it for the second D
it cries that they arent equal
warped dust
TimScript
sullen fiber
so ig D & {} is like a "derived" type, so it actually uses the first D then
warped dust
fucked
sullen fiber
idk how it works really im just assuming cus thats how i fixed it
tranquil vapor
LOVE
@real hinge
fallocate -l 1G mantika.explosion
git add mantika.explosion
git commit -m "add big explosion"
git push
run anyway
real hinge
vscode cloud changes
when I finish aoc I just store changes in cloud
and pull in my main computer
hazy sinew
<@287555395151593473>
```sh
fallocate -l 1G mantika.explosion
git add mantika.e...
github cant handle 1gb
so actually its not a big explosion
sullen fiber
@warped dust CHALLENGE
real hinge
github cant handle 1gb
are you sure
sullen fiber
wait i should stop treating this thread as general chat
hazy sinew
yea but ur uploading 1 file
real hinge
for 1 file yeah possible
hazy sinew
obviously the linux kernel github is more than 1 gb
real hinge
too late
hazy sinew
for 1 file yeah possible
i tried uploading the 10mil input for day 4 and it told me to use github lfs and to also go fuck myself
real hinge
CADDY INSANE
I think it got ratelimited
not related my to my config issues just restart triggered it
I should have reloaded h
high lintel
gat damn
real hinge
I am so skilled
native pewter
not the shitpoop :sob
exotic meadow
proper totem
i was keeping up then the mental health exploded before i could finish day 12
maybe i will catch up over the weekend
native pewter
been too busy after the pipes 
tranquil vapor
i kinda have no motivation to do it rn
warped dust
gave up after day 4
steady fog
@warped dust return
sullen fiber
It's been a fun morning ritual walking to the library after waking up
I think imma keep doing them after waking up when I get home
A lot more enjoyable than doing at midnight
solemn python
just finished d13p2
only minor change to my program from last night which is nice
basically I was just ||not checking if a smudge for 2 specific lines matching would actually expand into a valid reflection||
so instead of ||returning from the iterator instantly I just stored if the current reflection in the picture was smudged and then find the one valid smudged reflection at the end after checking everything||
ok finished d12p2
holy shit ||dynamic programming|| is crazy
how is it so fast wtf
I was not expecting that
and my code is very bad too
steady fog
@solemn python
solemn python
ready for what? it to release? it already did
sullen fiber
Lolol
steady fog
is this the last day of aoc?
delicate crane
there are 25 innit?
steady fog
why does it just stop
solemn python
no, see how 15 is going downwards
steady fog
no
sullen fiber
LOL
solemn python
I think we go back down the islands
steady fog
NO MORE
delicate crane
ruby?
steady fog
i did crystal does that count as the same
solemn python
have you done elixir
delicate crane
I guess
steady fog
have you done elixir
nop
delicate crane
nim?
solemn python
do it
sullen fiber
This is like
delicate crane
raw wasm
sullen fiber
6 chars in Uiua
sullen fiber
I might do this in my phone rn
steady fog
insane
solemn python
@steady fog@steady fog
sullen fiber
Today is crazy Uiuable
solemn python
you love disassembled rust wasm
steady fog
@sullen fiber forbidden from using usuaia
sullen fiber
hazy sinew
YAY
real hinge
today looks cancer
steady fog
@real hinge hi
real hinge
hi
hazy sinew
14 minutes with rank 44
i shouldve been way faster on p2
i solved it exactly when everyone else did
#100 is 17:15
teal mountain
wasted like 5 mins on part 1 wondering why my shit wasn't working bc i hard coded ||10 - i for the load lol||
hazy sinew
this was basically 1 trillion elephants exactly
for part 1 i just ||made a function that rolls rocks to the north and that was faster than everyone but 7 people ig||
solemn python
p1 done
hazy sinew
for part 2 ||made the function go in all directions made a cache and iterated through the cycles, checked the cache and if a state is repeated calculate that for 1 billion and get the remaining cycles, then just break and iterate through the remaining cycles manually||
teal mountain
same
solemn python
I love a billion iterations
teal mountain
a trillion elephants
hazy sinew
It would have been if I did the lazy way 😭
wtf did u do to make it not compatible
sullen fiber
||Split columns into groups by #, number them, count O them mul by number||
teal mountain
for part 1 || i iterated column wise with a hashmap with indices of cube rocks, incrementing the highest index on round rocks||
steady fog
suprized
hazy sinew
i just used kotlin for it ||```kotlin
grid.withIndex().sumOf { (i, row) ->
row.count { it == 'O' } * (grid.size - i)
}
mantika cant spell
teal mountain
i'm so pissed i did part1 in basically 1 try but i hard coded some shit for the sample
steady fog
i have an interesting idea i wanna try
just need to figure out how to do it in elixir
sullen fiber
i just used kotlin for it ||```kotlin
grid.withIndex().sumOf { (i, row) ->
...
Yea but I didn't do gravity
hazy sinew
huh
teal mountain
i just used kotlin for it ||```kotlin
grid.withIndex().sumOf { (i, row) ->
...
literally here i hard coded ||grid.size for 10||
hazy sinew
lmao
sullen fiber
Lol rip zain
hazy sinew
i just used kotlin for it ||```kotlin
grid.withIndex().sumOf { (i, row) ->
...
wait that shouldnt work
wtf
why does that work
wait i thought only rocks that touched the north row directly or via other rocks should count, but i forgot to implement that and it ended up being right
steady fog
does elixir have a way to shorten this
Enum.map(fn m -> m |> String.split("") end)
surely theres something
dont feel like scouring documentation right now
sullen fiber
wait i thought only rocks that touched the north row directly or via other rocks...
its all rocks
Just after they shift
hazy sinew
yea i get it now
sullen fiber
Ok gravity is a lot shorter in Uiua than my stupid group thinf
steady fog
ok elixir has some cool concepts
i kinda like this
@hazy sinew we need to port to kotlin
hazy sinew
Ok gravity is a lot shorter in Uiua than my stupid group thinf
i originally was gonna sum all the rocks and then check if theyre touching the north or not but forgot to do that and got the right answer
hazy sinew
<@415132870252822539> we need to port to kotlin
kt prob alr has it
steady fog
pipe operator
solemn python
dang it, no bruteforcing for me :(
hazy sinew
pipe operator
example?
real hinge
dang it, no bruteforcing for me :(
Day 12?
sullen fiber
Omg today is real
steady fog
i was thinkling i could sort the values
but
that would also move the #
idk how to get around that
rustic orchid
maybe next year I will be able to actually do aoc
real hinge
Dynamic programming will get us through part 2 alhamdulillah
At some point it probably repeats
hazy sinew
1 billion actually isnt that long
should take about 1h40 to bruteforce in kotlin with an inefficient function
steady fog
guh i thought i was onto something
i was gonna transpose, then sort the values
it wouldve been so simple
sullen fiber
Idk how to do part 2 in a fast way
But I can do it in a funny way
Probably intended ig
real hinge
My guess would be try to find a point where it repeats
If it exists
Then modulo it with 1000000
Do the rest
solemn python
dang it, no bruteforcing for me :(
actually it seems much faster
might just let it run for the memes
oh wait nvm
im running it on the example lol
hah
but I might let the example run for the memes :)
steady fog
tymonky
solemn python
but I might let the example run for the memes :)
at 34% at 10ish mins
steady fog
@solemn python how did you do part 1
solemn python
at 34% at 10ish mins
well after ||adding an LRU cache|| it gets the wrong answer anyways
steady fog
@solemn python@solemn python@solemn python tell
hazy sinew
i was gonna transpose, then sort the values
it does work
steady fog
how
solemn python
well after ||adding an LRU cache|| it gets the wrong answer anyways
||guh I assume my "go any direction but north" logic is failing||
AoC is so hard to debug and I hate it
solemn python
hm I think I have a better idea for my dumb direction logic
||what if instead of having separate logic for each direction, I just flip the board, run the north code, then flip it back||
steady fog
how
hazy sinew
or maybe not sorting but bro is talking about diagonally flipping a matrix and then inverting the columns bro just make 4 for loops and a switch case and change the row or column that ur changing
its better to start writing than to think about how to do it simpler
you should only think when your bruteforce doesnt work
steady fog
except i dont know the lang
hazy sinew
you should only think when your bruteforce doesnt work
today it was obvious that it wouldnt work so u can get straight to cycle detection
hazy sinew
except i dont know the lang
for loops and indexing shouldnt be too hard right
steady fog
i guess
sullen fiber
or maybe not sorting but bro is talking about diagonally flipping a matrix and t...
Laughs in Uiua
real hinge
see you guys in a year
steady fog
NO ONE SANE USES UIAIAWE
sullen fiber
I'm sane tho
real hinge
wrong
sullen fiber
I'm super sane
hazy sinew
memoization will kick in after a hour
is that for today
real hinge
@steady fog 0.003%
real hinge
is that for today
yes
hazy sinew
cause today doesnt even take that long
real hinge
I believe that it will repeat after some time
hazy sinew
2h max
steady fog
mantika mastery
hazy sinew
I believe that it will repeat after some time
it repeats after 7 cycles bro
real hinge
I move them one by one
sullen fiber
Manti I promise it should be much sooner
hazy sinew
I move them one by one
so do i
sullen fiber
HUH
real hinge
it repeats after 7 cycles bro
how
hazy sinew
https://youtu.be/WCVOBKUNc38?si=1IY1m86UHSEYkfZo
idk for him it was 7
i didnt print for mine i just calculated the output
sullen fiber
For me it was 98
hazy sinew
maybe 7 is example
solemn python
||guh my rotations appear to work but my answer is wrong||
oh wait im an idiot as always
solemn python
I can't figure out how to best cache this
steady fog
tmpfs in ram formatted as zfs
solemn python
not what I meant
steady fog
too late
solemn python
holy
nice one, solution
it really does take time to kick it in
anyways who cares, my solution is bad but works, I should sleep
hazy sinew
how do rolling rocks and non rolling rocks make sense
if none of the rocks are attached and you tilt the plane all the rocks are gonna fall
you need the square rocks to be attached to the ground
but then it doesnt matter if theyre square
and it makes more sense for the rolling rocks to be square because if they get stuck on a square theyre just gonna roll to the side and continue if theyre round
native pewter
and it makes more sense for the rolling rocks to be square because if they get s...
maybe the cubes are concave
hazy sinew
maybe were tilting it exactly north and not part of a degree to the side
native pewter
noooo
hazy sinew
cuz its vscode
real hinge
good editor
native pewter
my dad used to use emacs apparently
native pewter
you use emacs
i dont
sullen fiber
but then it doesnt matter if theyre square
I imagined they were much bigger than the rocks like bricks
Or Cinder blocks
hazy sinew
I imagined they were much bigger than the rocks like bricks
well they arent on the grid
actually now that i think about it maybe all the stones are big
cuz i dont imagine ur gonna be tilting small rocks
and applying pressure
tranquil vapor
i was wondering why my puzzle input resulted in 0
hazy sinew
lmao i had that
tranquil vapor
brute force so slow
hazy sinew
fun timeskip(cycleLength: Int, startOffset: Int, totalCycles: Int) = (totalCycles - startOffset) % cycleLength
now i dont have to remember the formula anymore
steady fog
ventika so smart
tranquil vapor
today was easy
my solution takes ~10s
@real hinge
i wonder how much faster it would be if it didnt use ruby
steady fog
hop on crystal
tranquil vapor
i wonder how much faster it would be if it didnt use ruby
and tbh u can probably optimise these but im lazy
steady fog
ven do you know elixir
tranquil vapor
nop
real hinge
my solution takes ~10s
part1 or two
tranquil vapor
two
tranquil vapor
and tbh u can probably optimise these but im lazy
my solution:
|| rotate map in a loop and put the resulting array into a map with the loop index. after every rotation check if the array is already in the map. if yes, we have a cycle, so use index from map to calculate how many rotations until a cycle happens and break the loop
now subtract the amount of rotations done with the loop from 1000000000 and take modulo with the amount of rotations until cycle to get how many rotations are left to do, then do those rotations
||
hazy sinew
and tbh u can probably optimise these but im lazy
hey i did that too
hazy sinew
my solution:
|| rotate map in a loop and put the resulting array into a map with...
that takes 10 seconds?????
steady fog
fix
hazy sinew
it takes 71ms in kotlin with exactly the same method
real hinge
I was gonna optamize mine furthger
real hinge
but uhhhh ureaxjsal
steady fog
ruby so bad
steady fog
I was gonna optamize mine furthger
optamize
real hinge
is crystal and ruby same
steady fog
love
hazy sinew
you probably did something that is really inefficient like read the input again every iteration
tranquil vapor
no
hazy sinew
and didnt notice
tranquil vapor
its the same code bro
tranquil vapor
i dont think ruby has jit
ruby might not be much slower in every day apps that do IO and stuff
but this is pure CPU work
where compiled languages are a shit ton faster
steady fog
venny
tranquil vapor
you probably did something that is really inefficient like read the input again ...
❯ diff solution.cr solution.rb -w
0a1,2
> require "set"
>
87c89
< map = File.read_lines(file).map &.chars
---
> map = File.readlines(file, :chomp => true).map &:chars
94c96
< map = File.read_lines(file).map &.chars
---
> map = File.readlines(file, :chomp => true).map &:chars
96c98
< prev_maps = {} of Array(Array(Char)) => Int32
---
> prev_maps = {}
103c105
< dump = map.map &.dup
---
> dump = map.map &:dup
i dont speak weirdly formatted ruby
hazy sinew
wait are ruby and crystal the same thing
steady fog
no
tranquil vapor
i dont speak weirdly formatted ruby
||
--- solution.rb 2023-12-14 17:54:13.123423262 +0100
+++ solution.cr 2023-12-14 17:53:11.742596153 +0100
@@ -1,5 +1,3 @@
-require "set"
-
def tilt_north(map)
(0...map.first.size).each do |j|
loop do
@@ -86,23 +84,23 @@
end
def solve1(file)
- map = File.readlines(file, :chomp => true).map &:chars
+ map = File.read_lines(file).map &.chars
tilt_north map
puts calculate_load map
end
def solve2(file)
- map = File.readlines(file, :chomp => true).map &:chars
+ map = File.read_lines(file).map &.chars
- prev_maps = {}
+ prev_maps = {} of Array(Array(Char)) => Int32
rotations_done = 0
cycle_happens_after = 0
1000000000.times do |i|
rotate map
- dump = map.map &:dup
+ dump = map.map &.dup
if prev_maps.has_key? dump
rotations_done = i + 1
cycle_happens_after = i - prev_maps[dump]
||
this is the only difference between the files
hazy sinew
it has no semicolons and no braces so its python
tranquil vapor
wait are ruby and crystal the same thing
"Crystal’s syntax is heavily inspired by Ruby’s, so it feels natural to read and easy to write, and has the added benefit of a lower learning curve for experienced Ruby devs."
those are very minor similarities
hazy sinew
i dont care
tranquil vapor
that's like saying javascript is c++ cause they both have classes and {}
hazy sinew
no because c++ has semicolons and types and js doesnt
i see def, no semicolons, no types, prev_maps = {} which is python syntax
its python
stop coping
also theres no variable declaration keyword
tranquil vapor
the difference is that crystal is actually good
while python stinks
anyway
im surprised ruby is this slow
ofc its gonna be slower than llvm, but 62 times? O.O
hazy sinew
nah not even python2 takes 15 seconds
tranquil vapor
when it comes to performance, Ruby tends to be slower than Python
from random dev.to post
steady fog
elixir captures are funny
hazy sinew
i thought ruby was llvm?
neon ore
ruby is interpreted
real hinge
> when it comes to performance, Ruby tends to be slower than Python
why would u use ruby then
tranquil vapor
why would u use ruby then
cool language
tranquil vapor
i thought ruby was llvm?
crystal is
real hinge
python bette
steady fog
crystal de crystal de
ven what lang should my extensible modular de be in
go, v, crystal
tranquil vapor
i enabled ruby jit and it's still 14s
tranquil vapor
crystal
steady fog
can i hot load scripts into the same runtime
steady fog
lua 
tranquil vapor
steady fog
i am teh anti lua
tranquil vapor
linux users love lua for their configurations
proper totem
lua kinda cozy
steady fog
badbadbad
proper totem
not super cozy but a tiny bit cozy
steady fog
@proper totem hi
proper totem
i wrote my first programming language in lua
hewo @steady fog @steady fog @steady fog
steady fog
hewwo
steady fog
yop
proper totem
i want to try crystal it seems neat
will do once motivation comes back from negatives
I've done ruby for work but it was Bad ruby in a project that I just needed to patch enough to get barely working so I don't think it's a good gauge of how writing it feels
tranquil vapor
crystal would be so amazing if its language server was better
but its terrible 
the fate of most lesser used languages
ruined by bad LSP
tranquil vapor
but its terrible <:pensive_cat:937179384866046002>
well it could be worse ig
but its really not it
proper totem
ruined by bad lsp or strangely bigoted leadership (or both [hello nim])
steady fog
nim makes me scream
real hinge
can you run ruby code with crystal
steady fog
momtika
tranquil vapor
can you run ruby code with crystal
nop
theyre a bit different in some things
[1, 2].map &:to_s # ruby
[1, 2].map &.to_s # crystal
arr = [] # ruby
arr = [] of String # crystal
crystal requires you to specify type of arrays, maps, sets, etc
native pewter
bit weird that those annotations are required, could probably infer
esp on returns, you already know the type silly
tranquil vapor
bit weird that those annotations are required, could probably infer
yeah like it already does for other vars
ig they wanted to be explicit about arrays
steady fog
why does this comparison fail
Enum.count(~c"###", &(&1 === ~c"#"))
returns 0
???
unhusk
tranquil vapor
insane code
native pewter
mentally ale
tranquil vapor
zeet did u solve part 1 yet
steady fog
no
steady fog
wrong
real hinge
@tranquil vapor how is yours so fast
mine takes 0.98s minimum
do you check arrays in all directions
maybe its because I rotate the array and check again with same function all the time
tranquil vapor
<@343383572805058560> how is yours so fast
i wrote good code
tbh my code isnt even vastly optimised
i could optimise it better
real hinge
whats ur check
I just take O's to top
rotate and do same thing
ig this horribly written function takes most of the cpu time
checking normally would be way faster but I dont wanna do that
steady fog
@real hinge figure out why mine doesn't detect O correctly
real hinge
<@287555395151593473> figure out why mine doesn't detect O correctly
virus
steady fog
how
real hinge
I put virus in your computr
steady fog
whyyy
real hinge
money
tranquil vapor
ig this horribly written function takes most of the cpu time
oh
you create new array every time
that's probably why
real hinge
yop
daring marlinBOT
|| https://codeberg.org/Ven/AdventOfCode/src/branch/main/2023/day14/solution.cr#...
**solution.cr: **Lines 6-8
||```cr
if map[i][j] == 'O' && map[i - 1][j] == '.'
map[i][j] = '.'
map[i - 1][j] = 'O'
rust analyzer is the best
real hinge
**solution.cr: **Lines 6-8
||```cr
if map[i][j] == 'O' && map[i - 1][j] == '.'
...
yop very good
I do kinda same too
just rotating creates new array
steady fog
what adds those errors in the line @real hinge
real hinge
whats the difference
tranquil vapor
just rotating creates new array
why tho
real hinge
isnt rust analyzer just language server
tranquil vapor
do you mean you physically rotate the array by 90°
solemn python
Rustc gives the error
Rust analyzer puts it in your editor
real hinge
yes
solemn python
do you mean you physically rotate the array by 90°
I did that too