Finsished!

I had it four fricking hours ago but made an off-by-one mistake somewhere

(To be fair the four hours were spent elsewhere doing other things, not working on AoC)

Anyway, time to wrap presents and watch the best Christmas movie with my family

Which is of course Die Hard

@TheFifthMarshal yeah i'm pretty sure i have that covered...

gist.github.com/hyper-neutrino/88898c28e46fc508229b394a725023f9

I was so worried p2 was going to modify how the instructions worked, like changing the mul ? 0s to mul ? 1s, but no it's just a trivial modification lol

Well, I still have a small chance of getting top 100 for the year, but I'd need like top 40 on d25p1

except for whatever reason my trivial modification didn't make it work lol even though my p1 ran properly...

Shouldn't for w in range(9, 0, -1): be for w in range(1 9) for part 2?

no, because a) you want smaller values to come later so they overwrite and b) it does a check to only replace n[nz] if the new value is smaller

for p1 i did range(1, 10)

Then I have no idea

also you know how your p1 ans started with 4

91297395919993 was mine, lmao

nice lmao

wow

mine was 49917929934999

tbf neither is within unintelligent brute-forceability

I suspect the puzzle inputs were chosen without considering their effect on the answer

i'm half trying to remember what i thought i figured out before i crashed and half getting torn to bits over last minute christmas eve plans

probably

do you know whether your attempted part 2 answer was too high or too low?

high

ah it worked this time

i am stupid

rip

the code i sent that said doesn't work

hadn't finished running yet

What was the problem?

nice

mb

i found a scala solver which just bruteforces with a cache

5029/5640

woo

mine is really not very smart... it just brute forces each block and collapses state by equal z values in between

This makes the effort I spent manually disentangling the problem and cutting its size down by a factor of 7 seem wasted

cuz i noticed x and y are reset to 0 before being used in each block

... i am sad now cuz if i just made that one observation, i think this still would've been enough to get a few points

granted, it wouldn't matter; #4 is too far ahead

i just need to make sure i don't lose #5

i have decided i will work on a proper library/system for next year and try to not care as much about leaderboard

although that might not be possible

i realized it ages ago and look where i am now lmao

rip

I suspect you could get a near-instant part 1 if you were to combine your and my approaches somehow

i think the furthest i got was realizing i worked through it wrong and need to specifically take the "B" values into account when putting something in to try to hit something congruent to 1..9 mod 26

some people just threw the input in gcc lmao

gcc?

lmao

That is first do the logic to determine the needed values for the seven stages with negative x that I did, and then share state in the way you did.

@TheFifthMarshal yeah all the instructions are basic C stuff so gcc will optimize them to a much more simplified form

Oh

bruh

lmfao

That doesn't change the need to solve an equation in 14 variables

i'm still not sure i even have 7 negatives

wait no i do

fuck

what am i doing

My approach would still work if there were a different number of negatives, as long as the number of div x 26 is equal to the number of positives

@TheFifthMarshal the thing actually simplifies to something a bit different from an equations in 14 variables

to get a feel for what i need to think through i've set up an interactive prompt that doesn't actually constrain w to be within the valid range of digits so i can just match the x value whenever the fuck i want and it is neat that it doesn't especially seem like i have to specifically match it on the div 26 stages because it's still a matter of adding 0 digits instead of 1

Mine was pretty fast, which I'm proud of

It takes like 10s to start up, since it builds a big huge array and sorts it, but then it uses some recursion and binary searching so the actual answer finding is instant

10m

Wow

My code did part 1 very slowly, but it could well have been near-instant if I had an input whose max digit started with 9

It did part 2 near instantly

Mine would've taken 140 iterations worst case

It could go digit-by-digit backwards

My approach was using some fancy logic I explained earlier to nail down 7 of the 14 digits, and then brute force the other 7. (so worst case 4,782,969 ops)

5m

There's a lot to criticize about d24, but it does have a lot of unique approaches possible

3m

Day 24 is the first part 1 that took be more than about an hour to do

i don't like that d24 has you solve the input and not the problem description but i will say that it wasn't a bad problem overall

just felt a bit less like a coding problem

AOC lore is always the same "oh no something went wrong with Santa or the sleigh get 50 stars to fix the completely unrealistic situation". Why can't we have an AOC where Santa committed tax evasion and you have to collect 50 stars to bribe the IRS to drop their investigation into Santa's financials?

lol

@hyper-neutrino I wrote a total of 98 lines of code to solve day 24. That's definitely a coding problem

@lyxal Last year we went to a tropical island instead of saving Christmas

1m, glhf

1m

well ofc it required coding but like the majority of the work felt like it went into manually handling the input

not that that's objectively bad.

anyway. last day, glhf everyone

15s

5s

done 79

i missed a fucking 1-off

(and wow, today was much easier than anything in the past week)

93/75

i am a fucking idiot

both for missing the oneoff and because i wasn't actually following the spec correctly for one very tiny reason

also wow how did i get 18 ranks from p1 to p2 lmfao

anyway, my #5 was secured today

Probably a lot of people were unable to do one or more of the hard problems

oh fuck i forgot about that

but still did day 25

yeah, right, probably d24 got a lot of people that couldn't get the 50th

I'm one of them

disappointed in my last few days but not in my overall performance. GG

i set my sights on top 10 and got top 5 so, i can't be upset with that overall :)

Now, if only I can do Day 19 and get the last two stars

oh that one...

rip

ah. yeah i've done this before so i knew what it'd be, but yeah i imagine some people spent a while looking at p2 before realizing they should just click the button

Somewhat surprisingly, Day 19 did not lose more people than Day 18

huh, interesting

also yeah this year's ascii art was a bit not as great as the ones from previous years at least IMO

¯\_(ツ)_/¯

Finished!

I had it working 5m ago but was running it wrong while debugging

So I thought I had an issue I didn't have

Well, AoC is over, so I can finally have a healthy sleep schedule

9/- woot

First <=10 rank

oh hey, nice

my issue was i didn't reset the hitbox check after the east cells moved

anyway, literally just loaded the cucumbers into two sets, lol

I used roughly the same code twice, and just transposed the array before and after

ah

that'd honestly probably be equally equivalent cuz checking for collision on an array is O(1) bc random access and in a set is also O(1) bc set

I hate JS's modulo operator, I lost a few minutes due to (i - 1) % c.size being negative sometimes

I know it's that way in C/Java, but Python's is just so much more convenient

ah :/

wait, -1?

For changing > to .

oh right

I'd share my code but it's ugly enough it'd be negligent to curse other people with its cursedness

I was really not ready for tonight to be so easy

@tjjfvi nice

yeah, I'm glad tonight wasn't too hard

386/1382 because I actually just picked up a pencil and did yesterday after tonight’s part 1 lmao

Feels good to finally do all 50

I wonder if mod 26 is just a coincidence, or if there's some sort of hidden message

Probably a red herring tbh

Guys, I found it! The secret message was kmkmlkkpnnszwm :p

pog

also the best part of my rank today is i did try to finish yesterday before midnight but i was just too dead tired until i finished part 1

Come to think of it now I need to look at y’all’s code for yesterday because I literally did not use a program for anything but checking my work

lol

Oh yeah I almost forgot how long I ate shit on today because I forgot about operator precedence in Python for like the fourth time

I was doing x+1%w

There should be both a % and a mod operator, with mod having looser precedence

And took like two minutes to add enough prints to tell that my coordinates were just ducking exploding

*fucking

thank you for the clarification

very cool

Until then I thought my problem was with the nothing moved check so I completely replaced it

@TheFifthMarshal oh yeah you know how i still have a brute forcer running

My solution started with 2

just got an aoc shirt lmao

Today looks quite easty but I've kinda given up

so it's finally over huh

gratz to all the insane mfs who got all 50 stars