Advent of Code

Redwolf Programs

04:40

@hyper-neutrino For heappush, is index supposed to not change? It looks like it's always just the length of the heap

emanresu A

04:51

9m

tjjfvi

8.25m

Unrelated String

closer to 8 smh

Redwolf Programs

Hoping today's not another algo-implementing one lol

hyper-neutrino

@RedwolfPrograms oops, that's a bug

just put index = half after the swap lol

Redwolf Programs

Learning about heaps rn, they seem cool

emanresu A

04:54

My code for day 14 only worked in firefox

hyper-neutrino

anyway, shameless self promo

Advent of Code 2021 Day 15 Walk-Through

►

cuz why not

Redwolf Programs

I'll watch that as soon as this AoC's over

emanresu A

Shameless djikstra library cuz why not

tjjfvi

5m

Redwolf Programs

I would say that's sort of cheating to use a djikstra library, but if the AoC challenge is literally as uncreative as "implement dijkstra", they're asking for it :p

hyper-neutrino

04:55

video's a... bit long :p and audio quality is pretty bad, sorry

tjjfvi

3.5m

3m

Redwolf Programs

3m

.at(-1)m

tjjfvi

2.25m

Redwolf Programs

Geometric series like yesterday?

tjjfvi

1.66666666666666m

lyxal

04:58

please oh please don't let there be anything that requires spatial thinking

emanresu A

^

lyxal

I do not spatial thinking

tjjfvi

1.49999999m

1m

Redwolf Programs

Space is cool, thinking is cool, so spatial thinking is cool

emanresu A

Today's goal: get ahead of ascii-only

Redwolf Programs

04:59

Mine's to not drop off the leaderboard

emanresu A

glagahf everyone

Redwolf Programs

yeah glhf!

tjjfvi

.5m

lyxal

mine's to get 2 stars lol

tjjfvi

.25m

5s

lyxal

05:00

gah frick what do they think this is? the web tech course I did this semester?

Unrelated String

what the fuck is this spec

lyxal

idfk

emanresu A

Probably doable with regex

emanresu A

05:14

When your input ends in 420

Lyxal, have you been hacking AoC again?

lyxal

@emanresuA m...maybe....

emanresu A

I'm somewhat glad that hyper isn't done yet

Because it reinforces how slow I am at this

lyxal

it reinforces how sh*tty the specs are

i'd have vtc'd this as lowkey unclear/needing clarity by now

have these guys never heard of the sandbox?

6

hyper-neutrino

28/17

it took me way too long to debug

emanresu A

I have a bug

hyper-neutrino

05:17

turns out i was dropping a leading 0 off of my real input :/

emanresu A

My one works for all but one testcase

and the real input

hyper-neutrino

because bin(int(input(), 16))[2:] only works if the input starts with 89ABCDEF

reading today's took me forever

emanresu A

Oh great trailing zeros

Unrelated String

@hyper-neutrino wait what

fuck

wait does that also apply to 0x

because i'm actually evalling with 0x

and SOMETHING is going horribly wrong tht i can't discern for the life of me

oh yeah of course the leading 0

just eats shit

fuck

emanresu A

WTF 85 p1

hyper-neutrino

05:20

dropped 26 points :/

emanresu A

I have points on global

Unrelated String

f

i literally made a typo in my hex digit list

hyper-neutrino

your what now

i ended up just doing "".join(bin(int(c, 16))[2:].zfill(0) for c in input()) lol

emanresu A

0x...n.toString(2).split('')

lyxal

HbvS∑:L:4%+$∆Zf⌊ in vyxal got me feeling like yeet

tjjfvi

05:26

My worst bug here was overflow :/

hyper-neutrino

rip :/

lyxal

@lyxal that's to just account for 0s

hyper-neutrino

functional lang for this seems super hard

i always do parsing problems by just mutating

leaderboard is closed; took 27:29

Redwolf Programs

I'm not sure how the trialing zeroes are meant to be handled

The Fifth Marshal

Done 168/127

My code just ignores them

that is, it will happily accept any trailing bits that aren't used by the code

Redwolf Programs

05:30

Yeah but how?

Wdym not used?

The Fifth Marshal

What is wdym?

Redwolf Programs

How are they not just parsed as the start of another operator?

hyper-neutrino

just ignore them

you only parse one thing

there is only one top-level packet in the transmission

Redwolf Programs

Ohhh

hyper-neutrino

> The BITS transmission contains a single packet at its outermost layer which itself contains many other packets. The hexadecimal representation of this packet might encode a few extra 0 bits at the end; these are not part of the transmission and should be ignored.

The Fifth Marshal

05:31

What was the point of specifying the puzzle input in hex, not binary?

Redwolf Programs

Oh no mine's really rboken

I'm not on the leaderboard anymore :(

The Fifth Marshal

(it's not to make it shorter, since there was a 30,000-character puzzle input one year)

Redwolf Programs

I've really screwed this one up

Unrelated String

holy fuck

hyper-neutrino

@TheFifthMarshal because what's an AOC challenge without unnecessary complications

emanresu A

05:35

I'm doing recursion

hyper-neutrino

smh, they should really sandbox these so we can iron out things like difficult I/O format

Unrelated String

i can't believe it took me that long to figure out i was literally skipping the last four bits of every literal

hyper-neutrino

oh

i see how one might do that

Unrelated String

183/250 :(

The Fifth Marshal

I had a more bizarre bug at one point: I was including the last 4 bits, but then treating that as part of the next subpacket as well

Redwolf Programs

05:36

Ojh, an off-by-one

o mine

Unrelated String

i very consciously decided not to parse the values on star 1 which led to looping in a way that kind of isn't built to parse the values

i also lost a ton of time on star 1 by completely skipping the total-length-given subpackets because i forgot that i still had to evaluate them for the version sum

Redwolf Programs

600/400 :(

I had a weird bug in p1 that took wayyy too long to debug

emanresu A

Great, infinite loop

Unrelated String

@hyper-neutrino the approach i used would translate pretty well to a functional language

emanresu A

Copy code and abandon ship

Unrelated String

05:40

mutated index variables and various accumulators but in a pretty straightforward way

Redwolf Programs

My sol

Still uses some library functions, but most of the meat of it is vanilla

I'm only 11 points off of the leaderboard, so I can easily get back on

Unrelated String

mmm

Redwolf Programs

Of all the things I didn't expect to do today, it was implement a whole programming language at 11 o'clock in half an hour lol

Unrelated String

lmao

Redwolf Programs

At least there's not much else they can add to this, so intcode 2.0 is unlikely

Well, making the operator ids 4 bits and adding some control flow could get interesting

The Fifth Marshal

05:43

And there's even a version field for back compat

Unrelated String

true

Redwolf Programs

This spec sucks

It's like it was intentionally written to be a game of hide-and-seek for the information you really need

Unrelated String

it's so fucking long and they don't tell you what you're looking for until the end

Redwolf Programs

At least the test cases were convenient

Unrelated String

yeah it was nice to actually have multiple

even if you can't parse them visually because they're in hex

emanresu A

05:45

TIL you put a limiter on a loop and it turns out it's recursion

hyper-neutrino

the thing is like

this is literally just a really long impl

there was, at least to me, nothing hard about this challenge, just really tedious to read the spec and implement

the only reason this challenge isn't just trivial "look at the task description and type" is because the description is unnecessarily convoluted and filled with useless flavortext

Unrelated String

^

Redwolf Programs

Yeah. And if you have a bug, it's a nightmare.

hyper-neutrino

yeah

Unrelated String

virtually impossible to debug without heavy guesswork

hyper-neutrino

05:47

i like challenges like day 14 where you have to actually think about a good approach and come up with optimizations, and there are multiple approaches

Unrelated String

and you can build an intuition for some part of it quickly

Redwolf Programs

Yeah I liked that one a lot

hyper-neutrino

yeah

easy to understand, challenging to solve

Unrelated String

here you can intuit broadly what's going on but there's so fucking much that you're going to mislead yurself trying

Redwolf Programs

Yesterday's was too textbook-algo-implementation, and today's was too cluttered

emanresu A

05:47

Speaking of debug...

Unrelated String

yeah

hyper-neutrino

^^^

emanresu A

(Don't run, infinite recursion)

Redwolf Programs

newd...pretty hot variable naming lol

hyper-neutrino

@emanresuA you are gonna run into issues with your parsing method if you're not careful

Unrelated String

05:48

imagine having variable names longer than two characters

emanresu A

Yeah I know, it's a bit of a mess

hyper-neutrino

because if your input doesn't start with 8/9/A/B/C/D/E/F, it will drop leading 0s

emanresu A

I know, that was the problem with p1

hyper-neutrino

ah

The Fifth Marshal

I have a variable named subpacket_bits in my code. That's 13 characters

emanresu A

05:49

I'm too lazy to zfill it :P

hyper-neutrino

wait i could've just prepended a 1 and dropped a bit lol

The Fifth Marshal

True

Redwolf Programs

@emanresuA next can be replaced with splice, for future reference

(is "for future reference" a thing that makes sense in english? too tired to tell)

@emanresuA Global variables

That's one of your issues

Recursion with global variables in functions does not work, you need to use var

emanresu A

var spammed everyywhere

Might ask for more help later

Redwolf Programs

Did it stop the recursion error?

emanresu A

05:53

Yep

Redwolf Programs

Yay!

emanresu A

6 * 9 = 15

Wait no that's just a typo

Now it's mostly working

Redwolf Programs

The test cases are really helpful

I've been super tired all day today from last night's AoC, so I'm going to go to sleep now o/

hyper-neutrino

goodnight o/

emanresu A

Yeah :p

Unrelated String

06:00

good idea

Razetime

@RedwolfPrograms this is pretty much exactly how networking do be

emanresu A

I have bugs

hyper-neutrino

homedepot.com/b/Outdoors-Garden-Center-Pest-Control/…

emanresu A

06:43

For some reason number literals are being added even though their parents have been parsed

Unrelated String

as in their parents have finished parsing?

emanresu A

This is for the last testcase

(value is when pushing a number literal)

And they're getting double-pushed

hyper-neutrino

probably a reference copy issue

emanresu A

Probably...

Code's still here

Unrelated String

07:31

it took me embarrassingly long to figure out that the mode setting thing was in a loop

when i saw that most of your code was in a function i assumed it was recursive

emanresu A

It is, mostly

Unrelated String

oh wait it is

emanresu A

I've made anotther version which seems to work but doesn't evaluate the final expression

Unrelated String

oh so you're using the modes as like
states
as you parse a single packet

emanresu A

Pretty much

Unrelated String

07:35

...and there aren't any breaks except in the switch so it doesn't terminate until it's actually exhausted the data?

that can't go well

and it actually relies on parsing multiple packets in one call

emanresu A

if (pc++ == packetCount) return packets

I kinda want to rewrite it from scratch

hyper-neutrino

yeah i think the easiest way is just mutating recursion

just have a parse function that takes a list

and returns one packet, consuming tokens as needed

Razetime

07:52

hi there

Unrelated String

i returned the number of bits parsed which was nice for the bit length of subpackets given type

emanresu A

I just parsed the next in bits

Razetime

yeah this looks like a really messed up recursive problem

Unrelated String

and then i guess with hyper's mutating approach you'd just remove that much from what you have then parse it until it's empty

emanresu A

It's really annoying how there are two ways to determine the number of packets

Ugh this is pain

Razetime

08:09

gotta be much easier to do this manually after converting to bits

emanresu A

08:39

Great, today's gonna end up like day 4

(In that I completely fail p2)

emanresu A

08:54

Ok, time to compile it into a js expression

emanresu A

09:25

This is proving a little harder than I thought

e(s(1,3)p(2,2)undefined(

Razetime

haha gottemmmm

emanresu A

Didyou finish

emanresu A

10:04

It's probably simpler to compile to lisp

Of course, my code has a meltdown as soon as I run it on full input.

> CE00C43D881120 -> M(7,8,9) => 9

> my input -> parenthetical syntactically invalid mess

gist.github.com/chunkybanana/e51df26491d2b995eb23305691369312

emanresu A

10:37

all of my code is failing horribly.

I've been working on this for five and a half hours and I'm pretty much back to where I started.

Unrelated String

f

emanresu A

I have three different versions of the code, all with different bugs... sighs

emanresu A

11:08

Time to make #4

emanresu A

11:50

Aight, I gave up and used Redwolf's because after seven hours of debugging, rewriting etc, all four attempts still refused to work.

And now, I'm going to sleep because it's 1am.

85/8126, that's gotta be a record

Unrelated String

damn

Transcript for