The Nineteenth Byte

4:50 AM

10m to AoC!

7

6

5

4

3

You were 15 seconds late!

4:57 AM

Yeah :p

2

intcode 2.0 here i come

1

@ASCII-only Why do you think that?

was just kidding

4:59 AM

0.5

0.25

probs another painful input parsing day judging by my luck

bruh

i forgot about regex total match i'm dumb

well now i have a 1m wait timer

part 2 is ...

i am going to cry

2/14. ehhhh, not bad. wish i had a better p2 but i don't think i made any critical flaws i'm just slow

5:18 AM

I can't do part 1 :(

once again

python cheese

you could probably do it via regex in js too

How?

It's not regex-y sort of stuff

lel i cant do part 1 either

@RedwolfPrograms how so?

part 1's basically regex tho

5:24 AM

I see sort of how it could be regex, but not really

i literally converted the rule into a regex pattern and then just matched each string

lel

local (global*) scoping killed me

also regex would be easy but too late

F

5:41 AM

star 2 is beating my ass to the point that i almost want to just do a regex translation

How would you even do it in regex?

might be smarter than trying to make my own engine with fucking generators

part 2 looks harder with regex

Definitely. First replace is easy, second is technically impossible.

i just did a really inefficient match function thing lmao, kinda troll

5:43 AM

oh true

it's the fuckin

but infinite loop is not really good for regex to handle.

literally the textbooxk example not-a-regular-language

@HyperNeutrino the rule 8 substitution would just be adding a +

but then 11 is... not good

oh true

wait. maybe you could if you add a bit of an abstraction layer and do 11 via just like matching 42 at the front and 31 behind and then repeat until empty or irreducible

Thank you so freaking much for that comment about adding a +, I was using * and would never have found it otherwise

that might be how 13 people beat me. idk

oof

5:46 AM

lol

I just cheated the second replace by trying every possible number of matching ones on either side from 1 to 40

I had a 60kb regex at the end of parsing

...lmao

wait yeah that's actually smart. i forgot about that, it's not like there can be infinite matches with short inp strings

hm. wish i didn't immediately assume regex wouldn't work

I'm actually working on something similar to this exact thing, basically regexes with variables

ooh. sounds like it'd be very useful for parsers lol

im still stuck on part 2 lmao

5:50 AM

the main thing that's tripping me up right now is i'm getting an answer, just not the right one

That way you can do recursive stuff like p:(p)|.+ to match anything in matching nested parentheses

im actually so bad

...I should try one of the sample inputs

tempted to just antlr this shit

or earley

mmm, so it matches one more on the sample than it should

...one more?

5:51 AM

dude

how the heck to do rule 11

erroneously matching aaaabbaaaabbaaa on the example

wish the sample was small and digestible enough to actually follow mentally

My horribly awful disgusting hack of a solution

...one interesting thing I just noticed is that rule 0 is just 8 and 11

Would've helped to know that before lol

wait

huh

5:59 AM

> That's not the right answer. Curiously, it's the right answer for someone else; you might be logged in to the wrong account or just unlucky.

facepalm

same happened to me too :D

ngn

me too

6:16 AM

SAME

had the idea of a greedy sort of algorithm that hardcoded the changes and it actually got the sample right but it got someone else's answer on my real input

hooray for 5 minute timers

i wonder if maybe i'm supposed to find some kind of special relationship between 42 and 31

why

...i forgot something big

:(

i would have gotten this like 10 minutes ago if I hadn't forgotten that line

and/or used smaller toy inputs

that... would have been the main thing

Day       Time  Rank  Score       Time  Rank  Score
 19   00:12:50    75     26   01:25:09  1136      0

maybe it would have been smart to actually follow the challenge's advice and look at the input

6:40 AM

ok part 2 is just adding ?R lol

:(

yesterday, by Bubbler

Frick

Why is this on the starboard

@Bubbler deserves to be there

i really need help wth

i match 6 messages on part 2 sample

which is wth

hmm, ?R isn't right

6:55 AM

@ASCII-only send?

@HyperNeutrino link

nvm i'm not sure i want to read that

>:(

i swear it should work

I'm heavily overfitting on the input and trying to finish it with regex

trying longer and longer regex until the answer is accepted

so I'm trying to make a useless capturing group to use it later

7:03 AM

it's easy with perl regex

and/or java/.net

(and also taking longer and longer time)

use .net regex and balancing groups

Is (abab){0}\1 correct?

{0}????

Factor regex is even more basic than JS regex, so meh

7:04 AM

;_; i'm actually stuck

it's been 2 hours

@ASCII-only i don't want it to match initially, I'm just storing it for use later

@Razetime that matches it 0 times

aka the empty string

l mao

yeah

yes

i need \1 to reference a previous capture group

7:07 AM

since it never matches, \1 is nothing

oh well

how do I store a regex pattern smh

aha

named capturing groups

????

(?<name>(<regex stuff>)\k{name}(even more regex stuff))

wth kinda lang has named captures but not recursion

Idk how to recurse within a capture group

it's PCRE flavour

7:20 AM

i used regex but only to match rule references versus characters

DOMAIN ERROR: Too many subpatterns

lol

loooool

i'm honestly beginning to think regex for p2 isn't the best idea xd

it is tho

it very likely is

7:25 AM

it's a lot easier if you look at the input

then again i didn't try it so maybe i should try to do it in pure regex before i judge it

how many of you guys are done

can i just give up

i literally don't know which part of it is wrong

really wish the provided example covered more edge cases lol

try uh

Tried various things but returned back to regex construction

0: 8 11
8: 42
11: 42 31
42: "a"
31: "b"

ab
aab
aabb
aaabb
abbbbb
abaab

7:35 AM

Let's see if it gives me an answer eventually

@UnrelatedString whats this

why is this returning 2 for me

a really simple example input you could try for diagnostic purposes

Isn't 2 the correct answer?

i think it should give 2

aab and aaabb

7:36 AM

yeah

ah ofc.

well so your problem is definitely not something obvious lol

Looks like a good code golf task

tried running this on my own solution:

0: 8 11
8: 42
11: 42 31
42: 31 1 | 1
31: "b"
1: "a"

aab
abbababa
bab
baaaaabab
aaaaaaaaaabbb

and somehow it just straight up errored

1 match here

7:41 AM

okay now it's 3

whats it supposed to return

it should definitely match aab, baaaaabab, and aaaaaaaaaaabbb

and the other two should not match

ok I'm just going to try recursion now

8:00 AM

Actually doing binary search on the answer box would have been much faster

3

wait.what?

wait how

(just tried it, it stops giving me low/high at 4th try and timeout increased to 5min)

I knew they're smart enough to stop that kind of cheating

oh wait you mean just brute forcing in the submission

i thought you found a way to divide-and-conquer programmatically lmfao

and my program just gave me an answer and the timeout is still 3min left

8:04 AM

F

ngn

my part2 worked with same code as my stupid recursive part1. i've no idea why. it should have crashed.

and the answer was indeed right

ngn

congrats

but horrible, super inefficient, and super overfitted to the input

@ngn my part2 should have worked because of this

but it doesn't

ugh

8:08 AM

using plain regex, no recursive regex or anything

@UnrelatedString i'm failing on this input but idk why

8:23 AM

Looks like the patterns 42 and 31 match 8~10 chars each - could it be more efficient to list all the strings that match them and work on that?

8:46 AM

@Bubbler I tried that approach then gave up

but that was mostly because I wanted to try a dumber/greedier version of the same chop-stuff-off-in-a-while-loop approach

and then that version happened to work

@ASCII-only what is the single line that it matches

aab lel

lol

what if you added uhhhhhh

baab

see if baab matches

the last observation was not quite correct, 42 and 31 matches 128 distinct strings of length 8 each, and the two are disjoint

that...makes things definitely easier

whatever

i did it

so the problem is i did it wrong

oh well

Day       Time  Rank  Score       Time  Rank  Score
 19   00:26:15   434      0   03:57:57  2956      0

4 hours

i wasted 4 hours of my life for this

f

and you still placed under 3000 lol

that made me check the stats page, and there are now more than 10000 people who have completed star 1 on day 4 but not star 2

it's not even hard

it's just

mildly tedious

@Bubbler ...that explains a lot

9:14 AM

Using chunks of size 8 indeed gives the correct answer in an instant

@UnrelatedString In terms of (part2 failed/part2 solved) ratio, days 10 and 13 seem to be the highest two, which do require some knowledge in algorithms and/or maths

I'm gonna take my time and solve this later

@Bubbler I can definitely see why 13 would be up there, but 10 is a bit of a surprise tbh

although I do also remember watching a friend of mine try day 10 star 2 and...

okay yeah 10 isn't a surprise

9:35 AM

10 is dp, 13 is crt so understandable

10 is so infamous that it's memed on the discord

db, dynamic 🅱️rogramming

wait theres an aoc discord?

probably not official but yeah what

it's probably run by some people off reddit

Wonder if there's some youtube channel who explains how to solve each AOC

wait i could do that that's actually an interesting idea

9:44 AM

yeah

i'd need to figure out how to properly solve things tho

cuz i completely memed yesterday's problem because python funny

lol

does python have some way to fuck with precedence or did you just make an int class where addition and multiplication themselves are swapped

(and then also swap addition and multiplication in the input)

the latter one

nice lol

idk why i thought shunting yard algorithm would get me there faster lmao

actually now that i think about it almost all of my approaches use some aspect of python specifically so coming up with a general solution might not even be too helpful

idk should i even bother, cuz i have a decently high rank so idk if people would even be interested if i made videos explaining each solution and how i got to them quickly enough to rank. /shrug/

10:01 AM

@HyperNeutrino I agree with memes

memes are always best solution

I must see it now

i don't think i can get it anymore but if i do write up explanations for my solutions i'll send it

I just wanna see the messy meme solution

no need of explanation

@Razetime tio.run/…

tio.run/##hc/NasMwDAfwc/…

part 2 is very similar. my actual part 1 solution was a sketchy parser cuz i didn't realize i could do this which is why i only ranked 105th

Lyxal

10:28 AM

@HyperNeutrino dang. I wrote an entire parser from scratch

Tokens and everything

I overengineered part 2 for that

With three functions: tokenise, bind and evaluate

10:58 AM

@HyperNeutrino ok this is hilarious

@Lyxal i did too

@HyperNeutrino the method does not affect ranking, it's 99% just the experience + good problem solving

@Razetime link

11:25 AM

@Bubbler oh, so everyone gets a different input?

11:37 AM

not everyone

but there's enough different inputs to discourage copying answer

wizzwizz4

But you could just copy someone else's code.

copying someone else's code is a whole other

thing

11:54 AM

@NewMainPosts in case anyone thinking of writing a Charcoal answer, best I could do by porting an existing answer is adapting xash's answer, which takes 20 bytes, but it can be done in 16 bytes, I just don't know whether it's a port because I don't understand some of the answers well enough

12:11 PM

@Neil What about retina

1:02 PM

@Razetime 48 bytes for Retina 0.8.2

1:25 PM

:( i'm not smart enough so my current charcoal solution is 18 lol

17 if i cheat by taking input as python array (so no cast needed)

wait

InclusiveRange is useless since i can just use Incremented???

1:46 PM

only for numbers; InclusiveRange also works on character ranges

also, my 16-byter uses InputNumber() so it doesn't matter if the input isn't in JSON format