« first day (3613 days earlier)      last day (1232 days later) » 
00:00 - 15:0015:00 - 20:00

12:11 AM
Hi everyone! Could anyone help me with this question that I've proposed in the Sandbox? I wanted to post it, but since it's a popularity challenge and has no upvotes, I wanted to get more feedback (and perhaps tag it code-golf instead).
 
 
4 hours later…
4:38 AM
@user popularity challenges are discouraged because most people don't have any good scoring criteria other than votes
also people aren't really motivated to post answer to your question after it's timeline ends
 
if it's a parsing aoc again i am going to cry
 
but you finish it in like 5 mins anyway
 
sounds about right
 
Why is there some snake thing in the ASCII art
8
Is today's task to battle a sea serpent?
 
That's hard to do while on a train.
Or at the north pole, which is the other place a challenge could be set.
 
4:53 AM
Not if the sea serpent is in front of the train :p
 
Trains normally go on land.
 
2
 
5:02 AM
it's a good thing this doesn't care about overall orientation
 
p2 probably will though
 
i think there's a language called snail that's perfect for part 2
no idea how to use it or if it even exists tho
 
PMA/Snails
it has some docs
 
Forgot to get the actual tile number for part 1 and used jndex instead :/
And mistyped the index again :(
KNEW it was about sea monsters!
 
...wait
it flips as well as rotating
:(
i just spent a minute making it handle rotation cleanly
and i still don't even know how i'm going to assemble the image
i assume brute forcing permutations would... not be fast enough
that's a lot of people with part 1 done
and none with part 2
oh boy
...I cannot believe that solution worked
that was WAY too stupid to work
lmao no wonder nobody's done part 2 yet
 
5:27 AM
@UnrelatedString 144 factorial, basically. so no
 
you'll never guess how i solved part 1
except you probably already did
but holy fuck I wish I had tried that earlier
 
5:38 AM
@UnrelatedString finding just the four tiles which have only 2 common borders?
 
Part 2 is so hard
 
woah rank 861
 
yep
partly because APL is great with arrays
 
5:41 AM
Okay good, I was hoping 800 people hadn't finished p2 yet
 
IF I can figure out how to assemble this thing, p2 is mad easy
because APL
 
The if is the big part :p
 
yeah IF
 
i'm probably not gonna rank
 
Me neither
 
5:47 AM
i fucking hate myself this problem shouldn't be this hard for me
 
I'm using one letter variables and feeling the pain
It looks like I sat on my keyboard and compiled it, and I can't remember what half my arrays' formats are
 
wait...
i missed aoc...
:(
 
This one's really hard
 
lies
can't be harder than yesterday
 
Part 2 is
 
5:52 AM
part 2 definitely is
 
you stlil have time to rank for p2
only 27 users done almost an hour in
 
Why did they have to make this one so difficult
 
YES
okay. i temporarily don't hate myself
14/54. man this problem was soooo hard but actually quite satisfying to finish
 
Also apparently you need to find edges first, then center, because I've got three blocks matching my corner on one edge
 
6:02 AM
yeah, i wasted about 10m on that
 
interesting
 
could've ranked up about 20pts maybe...
 
I don't remember how to determine which ones are edges/corners :/
 
my input seems to just not have any extraneous adjacencies lol
 
I might just take the 5kth place and do this tomorrow morning
 
6:03 AM
i think if you do this tomorrow morning you'll still get better than 5000th lmao
 
lmao true.
 
im kinda bad
 
I need to write this using actually good programming practices, instead of giant tuples of arbitrary values whose purpose I don't remember
 
lel cant even do part 1 again
@RedwolfPrograms same
only diff is i remember exactly what everything is
and i still can't do it .-.
 
6:04 AM
I mean in theory it's not that hard, but when you put the pressure of time on me I just focus super hard on one part of the challenge and then don't remember how to use it later
 
i'm surprised i didn't confuse myself with naming
 
i got infinity as an answer
i have 1 letter names
and i know exactly what they stand for
 
Yeah, definitely doing this tomorrow morning.
No way more than 2k people finish this at all lol
 
for(const[n,u,d,l,r]of S)
except for S
tf is S
 
If anyone can tell me what the heck aa = aa.m(x => [...x, image.fi(i => i.s(n => n != -1 && tedges[n].s(e => x[0].i(e)))).fi(n => n != -1 && tedges[n].s(e => x[0].i(e)))]).m(x => [...x, [image.fin(i => i.s(n => n == x[2])), image.fin(i => i.s(n => n == x[2])).fin(n => n == x[2])]]); is supposed to do I'll pay them a million dollars lol
 
6:05 AM
what
 
Cause I sure can't
And I wrote it ten minutes ago
 
my code excluding template was 2777 bytes
 
@RedwolfPrograms beautiful
 
a lot of copy paste and debug though
 
Yeah, see y'all tomorrow :p
 
6:07 AM
3677: {d: 1321, l: 1321}
3739: {d: 1627, l: 1321}
3779: {d: 1321, l: 1321}
3793: {d: 1321, l: 1321}
3797: {d: 1321, l: 1321}
3821: {d: 1321, l: 1321}
3877: {d: 1321, l: 1321}
hey
i don't think this is right for some reason
 
well uh...
 
At least your output looks like something a machine would write and not a squirrel
 
i wonder: on how many AOCs did the first person get their 2* before the 100th person got their 1*
 
almost all?
 
answer: day 2, 3, 5, 6, 7, 9, 10, 11, 12, 14, 15, 17, 18, 19
so yeah all but 1, 4, 8, 13, 16
i'm getting closer to rank 14
and the 100th user finally completed 2* lol. i wonder if this is the longest ever 2* leaderboard completion
 
6:17 AM
i'm having a hell of a time debugging and i'm not even to the point of actually finding sea monsters
 
lel im so bad finally did part 1
took me 1h26m
is part2 actually hard tho
 
it is... very tedious
wouldn't necessarily say algorithmically hard. it's harder in implementation than concept but it's not particularly easy in either respect ig
 
is the stitching the hardest part?
 
i tried a dumb band aid fix and now it uses 64 unique tiles lmao
 
6:32 AM
yeah stitching is the hardest part. so, it wasn't too difficult for me it just took me a long time because i messed something up the first time and it wasn't really salvageable
finding sea monsters is easy
 
as i thought
rip me
should just make it a builtin in charcoal
time to do some shitty bfs stitching yey
 
somehow it's just given up on making the bottom left corner a corner
 
:D nice
 
i cannot get myself to understand my own rotation/reflection logic at this point
i'm...
i'm going to brute force a state transition lookup table
>>> 8**8
16777216
maybe not
kind of unfortunate that i actually do not have a pen and paper atm
 
6:48 AM
@UnrelatedString do it
 
7:10 AM
clearly it'd be faster manually counting :p
 
one of my configurations is creating alternating rows of 12 tiles and 1 tile
and i'm trying to do something with that directly
and it's not even close to working
 
7:35 AM
1667/684
 
Once I made a "extend-to-right" function, it was straightforward implementation
 
I'm building out each row in order from the left and it's... not going super well
the rows themselves seem to work fine
but moving from one to the next... yeah
 
Oh nice, I started today's problem like 1 hour late, yet I'm #3 to solve part 2 on CGCC
 
7:38 AM
Hint: extending to right ~ extending downwards
Apparently solver count drastically dropped on day 19, and it'll get even worse on 20
 
:o
ok so
i got distracted by youtube...
 
i made something to brute force the rotations and reflectednesses on the column and it's just coming up with zero valid grids lmao
 
ngn
can anyone share inputs&outputs? i'm stuck on part2 with a wrong answer, though it works for the sample. i need more tests.
 
@ngn Does your input start with "Tile 3187"?
 
ngn
@Bubbler no, but i have a tile with id 3187 further down
 
7:50 AM
what about 3593
 
ngn
@UnrelatedString i have that too, somewhere in the middle
 
ngn
@Bubbler thanks
 
My correct answer was something close to 2300
yes (you should delete it just in case)
Very weird if it works on my input but not yours
 
ngn
7:59 AM
@Bubbler do you mind running your solution on my input? the wrong answer i keep getting is 2759
 
i have an answer from my sol and i'm pretty sure mine is right
 
ngn
@HyperNeutrino different from mine?
 
ngn
ok, back to the drawing board.. thanks
success! finally
of course now it doesn't work for the sample but who cares
 
8:20 AM
i need to get myself some food before i can continue bashing my head into this now
because it's already been three hours and twenty minutes
 
lol
queue keeps growing? this isn't good...
 
mine doesn't work for the sample cuz the 12x12 grid size is hardcoded lol. can't be bothered to write good code, i just wanted to get it done and hopefully rank and get out of there lol
 
@HyperNeutrino same minus the hope of possibly ranking
 
what even is ranking bro
 
i may or may not justify time/effort i spend / validate myself with fake internet points
 
8:37 AM
i'm just hoping to get more points than i did last year
...not looking entirely likely
 
i need to remember next year to start it on time and i want to get top 10 next time lol. started writing my template too late and i missed a few days too rip
 
8:54 AM
btw
my stitched image
is very nice square
line lengths: [109, 109, 109, 109, 109, 109, 109, 109, 109, 109, 80, 80, 80, 80, 80, 80, 80, 80, 109, 80, 80, 80, 80, 80, 80, 80, 80, 118, 96, 96, 96, 96, 96, 96, 96, 96, 118, 96, 96, 96, 96, 96, 96, 96, 96, 118, 111, 111, 111, 111, 111, 111, 111, 111, 118, 111, 111, 111, 111, 111, 111, 111, 111, 118, 111, 111, 111, 111, 111, 111, 111, 111, 111, 109, 109, 109, 109, 109, 109, 109, 109, 109, 106, 106, 106, 106, 106, 106, 106, 106, 109, 109, 109, 109, 109, 109, 109, 109, 109, 109, …]
.-.
 
lolwat. rip
 
i have no clue how this can even happen fr
 
that makes no sense whatsoever
 
indeed
just goes to show my incompetence knows no bounds xd
this really is taking longer than yesterday...
 
9:10 AM
i think i've managed to actually get the stitching done through sheer force of will
 
@UnrelatedString is it correct?
 
@Razetime the sea monster detection comes next
 
oh yeah crap
 
sea monster detection appears to be offline
which would imply that perhaps my stitching is not entirely accurate
 
9:26 AM
@UnrelatedString tell me the secrets
@UnrelatedString and/or your whole-image flipping/rotation
 
my strong suspicion is that nothing works except that i am nominally creating some grid from the 144 pieces
 
@UnrelatedString hey same :D
 
also it's trying all 8 possible configurations once the image is stitched, but no sea monsters
...the grid should be 96 by 96, correct?
 
hyper said it was 12 x 12 tiles?
so 109 x 109? but what would i know
 
yeah t has to be 96 by 96
because each tile is 10 by 10 and then you remove the borders
so at least that part is correct
 
9:29 AM
remove the borders?
... oh
 
yeah
...well it turns out that my monster remover also doesn't work so that's cool
 
don't need to remove
just count
 
true
regardless it's not finding any
time to actually use the example input
so it's producing about 40 too many rows on the example
oh yeah i hardcoded the length in more places than that
 
9:46 AM
in my case i just cant reflect
 
...so if I don't ignore the sea monsters in the example, I should get 303, right?
I'm getting 299
and just to make sure I'm not going insane, it isn't any orientation of the correct example grid
but the different sum would indicate that, even worse...
it's using tiles multiple times
 
your probs messing up reflections like i am
or rotations.
 
why didn't i keep checking that
 
:o ono
 
i can't imagine how i'd mess up reflections in python, and then for rotations i'm just using c = [*zip(*c[::-1])]
 
9:52 AM
Instead of programming the reflections and rotations I'm going to lookup table
 
it's using five out of the nine tiles
HOW
and on the real input it's 133 out of 144
in both cases it's off by 1 less than the side length, so
I swear I checked against that what the hell
 
it's been what
5 hours now
 
yep
I need more food at this point
 
i need more braincells, if anyone has a source for those lmk
 
have you checked if your stitching is reusing tiles
 
10:06 AM
can you game the Coming Up Roses hat by upvoting all of the answers on your question?
 
I FOUND A SEA MONSTER
 
the answer is not correct
but I did find a sea monster
four of them actually
...............
IT'S STILL
133 TILES
HOW
oh that's how
it's what i originally thought it was originally and it wasn't
generator time
well I have 144 unique tiles now
but the monsters are gone
 
guess what
i have 129 tiles :D
 
:D
I finally hit a sea monster with 144 tiles but it's not the right answer
it's correct on the example
but i need food again
you know, i had plans for tonight
but i am not giving up
 
10:17 AM
this is probably the only problem which would be easier done manually
 
honestly yes
 
literal 96-piece jigsaw puzzle
 
except at least in my case all of the adjacencies are unique
so it shouldn't be hard
at all
 
would've been way faster too...
 
i have been reduced to eating almonds in ketchup and hot sauce
 
10:30 AM
o i fxied it
i was overthinking
now to find teh dragons
> That's not the right answer; your answer is too high.
thanks
 
@UnrelatedString what kind of war crime is this
 
@Razetime a surprisingly good one
 
im getting 292
neat
299 now
 
@UnrelatedString You cannot convince me that that is a good idea
 
@Razetime it's like, local artisanal hot sauce
 
10:45 AM
still ain't convinced
 
local artisan = you?
 
@ASCII-only nope lol
 
well
i was close
 
anyhow i'm somehow getting alternately 104 and 102 on the example now
after having gotten the right answer on the example
why couldn't the sane solution have worked
i'm just going to try 30 less than my current wrong answer in 5 minutes because... sea monsters can't overlap, can they?
wait no not 30
15
okay, that's not it either
@ngn tried mine on your input to see if somehow i'm having the same issue as you were and... my code looks for an unrotated top left corner, which your input doesn't have
 
ngn
11:01 AM
@UnrelatedString my issue was that i wasn't orienting the tiles properly. in some cases every other row came out flipped.
 
mmm
i think i had that issue too actually
but then i scrapped the entire thing that was capable of having that issue
 
ok so
i did it
i'm not even happy tbh
 
just relieved that that's over with
 
11:12 AM
Day       Time  Rank  Score       Time  Rank  Score
 20   01:26:52  2023      0   06:11:43  1798      0
 19   00:26:15   434      0   03:57:57  2956      0
this though
this is hilarious
 
is that H:M:S of runtime? O_O
ah, coding time.
 
i'm now randomizing my solution
10 minute timer hours
2134 is out, 2154 is out, I don't know how the hell it removed numbers of tiles that weren't a multiple of 15 starting from 2199...
 
@UnrelatedString send input?
i wont spoil
ill grab popcorn instead
 
wait does my sol break on that
lol
jk ofc. turns out it's trailing newline
@UnrelatedString so i got a solution, all i can say is... see you next decade
 
11:20 AM
oh boy
random now gave me 2124, which i already tried
wait why did i even try 2154 when 2134 is too large
 
whats the range of the random
 
i'm just making random choices between i don't even know what they are anymore
 
also you probs could have binary searched via high/low
 
you know what
I'm just gonna pass on this one
 
guess i'm trying 2079 next, as planned
or 2064
it looks like that's the lowest it'll give me
nope 2049 is also on there
2019...
uhhh
 
11:28 AM
@UnrelatedString randomizing, meaning?
 
yeah this isn't telling me anything more than "the answer is some multiple of 15 less than the number of #s if you ignore accounting for sea monsters
@Razetime if i knew what i was randomizing i wouldn't have to randomize it
update: 2049 is not it
 
Blade runner
 
the smallest it's given me is 1989
1989 ain't it
 
11:54 AM
I'm wondering how many minutes you're forced to wait between each submission at this point
 
@Bubbler last I read, it was 10
 
Dec 9 at 5:41, by Bubbler
I'm waiting for the moment the difficulty skyrockets, so I have a better chance of higher rank
(meanwhile this actually worked out, though the problem wasn't really mathy)
 
12:13 PM
@Razetime guess who just hit 15
 
@UnrelatedString update: it really does look like you'll be stuck till next decade trying solutions
 
is the correct answer, in fact, different from the ones i've been trying by some multiple of 15
no would mean one of two things:
either there's something beyond fixing, or sea monsters can overlap
 
@UnrelatedString doubt it
 
to which part
 
overlap
spoiler:
 
12:21 PM
that's a relief i suppose
 
answer is slightly outside the 1989-2100 range
for some definitions of slightly
 
hmm
so even my row-growing logic is completely broken
if it's lower than the smallest value i've gotten
yeah i'm just going to call it a night after guessing 1929
rewrite after some sleep
surprise surprise 1929 ain't it
 
@UnrelatedString yes please get some sleep
 
12:58 PM
blindly guessed 1629 because why not and apparently that's someone else's answer so maybe i;m kind of in the ballpark there
 
1:19 PM
please get some sleep
 
1:45 PM
Good morning!
 
2:19 PM
morning!
 
Working on p2 again, how many of y'all finished it?
 
ASCII-Only did
Neutrino I think
I'm gonna take my time with this
and maybe look at some stuff on the reddit
 
I think I've got it figured out, I actually made useful variable names and helper functions for this so bugs probably won't be a problem
 
good luck!
 
Thanks!
 
2:59 PM
This is so stupidly hard.
It would have been a challenge even if they weren't rotated or flipped.
 
00:00 - 15:0015:00 - 20:00

« first day (3613 days earlier)      last day (1232 days later) »