AoC 2022 | Day 10 | Solutions & Spoilers

[sweet mango]

J looks like ```
..##.
...#.
...#.
...#.
#..#.
.##..

[quasi minnow]

I, X, V, W, and Y all don't have proper forms AFAIK

K and H are fair though

[frank kayak]

You have a moving pointer and a theoretically infinite strip (if you make the width more than 40) though.

[nimble shuttle]

doesn't make it turing complete

[quasi minnow]

Yes but you can't read or loop

[nimble shuttle]

exactly

[frank kayak]

True I guess

[nimble shuttle]

[frank kayak]

If you could read though I don't think you'd have to have a dedicated loop here for it to be turing complete

As it resets its position at the end of the screen each time anyway

[nimble shuttle]

as long as you can't create a finite program with an infinite runtime, it's not turing complete

[quasi minnow]

There's no looping or branching, it's not turing complete

[light vortex]

impossible, as you can not disable the first pixel, no matter what you do. X is 1 at the start, and it takes at least 2 ops to change that, in that time, this pixel is already drawn.

my conclusion, might be faulty

@hollow void

[hollow void]

yes, youa re right

[frank kayak]

I'd argue there is looping, just no branching

[nimble shuttle]

what looping lmao

no there isn't

[hollow void]

but if you ignore the frst 4 pixels?

[nimble shuttle]

can't exceed 40 columns, so still not every image

[light vortex]

good question. will play with that tomorrow 👋

[hollow void]

every image as in any combination of pixels on the screen

[frank kayak]

The counter essentially sets itself to 0 every 40 cycles, that sounds like a form of looping to me.

[nimble shuttle]

uh that's not looping

[quasi minnow]

But the PC doesn't loop

[sweet mango]

why would the counter reset?

[nimble shuttle]

fr just read the wiki page on turing machines

[frank kayak]

To draw the next row

[nimble shuttle]

we can't discuss turing completeness if you don't know what it means

[sweet mango]

but part 1 relies on the counter not resetting

[nimble shuttle]

in that case, yes

it's possible

i could even write a program that generates the instructions for any given image (excluding starting pixels)

[hollow void]

proof

[nimble shuttle]

set the sprite to somewhere away from the beam or at the beam 2 cycles ahead

not that hard

[hollow void]

I'd prove it in this way:
I can freely update x every 2 pixels.

This way, I need to be able to draw all 4 combinations every 2 cycles

[nimble shuttle]

yeah, and you can position the sprite for any combo of the 2 pixels

[hollow void]

#., .., .# and ##

[hollow void]

yep

[quasi minnow]

But is it possible if width isn't divisible by 2 though

[nimble shuttle]

yes

you can just use a noop

wait

actually maybe not

[quasi minnow]

I don't think it is

(some rows, so you can choose the value of X)
# # # # #
# # # # #

I don't think it's possible to render that pattern

Without a dedicated blanking period anyway

[nimble shuttle]

yeah i think you're right

[vernal urchin]

I tried writing a thing to generate the code for a given output string, and well my brain hurts i am giving up

[nimble shuttle]

rip

[half mirage]

Just caught up on all my AoC - did all Day 9 and Day 10 today

Day 10 probably my fave day, though the off-by-one errors are not fun

QQ - in the Python match-case syntax, is there a way to cast a matched value to a different type from within the case statement? Or do you have to cast afterwards

[quasi minnow]

I think you have to cast after binding

[naive sun]

me too

[nimble shuttle]

me3

[half mirage]

right, I did wonder - good to know :)

Also only 4 classes for today's solution

[nimble shuttle]

my solution had 0

[hollow void]

mine had 1 function

[hollow void]

java user spotted

[hollow void]

cast afterwards

[half mirage]

Actually, true story, the first time I learned Java, I was lying on my bed reading the first chapter of a Java book, and fell asleep reading it

[quasi minnow]

Lmao

[naive sun]

What the cat??
Look at the tests for trying to cast in a case -
I would have expected it to be a syntax error - But it's actually changing the results...
https://github.com/shenanigansd/scratchpad/commit/0c20735eba3fc0031de1b7b773290e640e271a8c
https://github.com/shenanigansd/scratchpad/actions/runs/3665655477/jobs/6196870308

=================================== FAILURES ===================================
_______________________________ test_run_cycles ________________________________

    def test_run_cycles() -> None:
        cycles = run_cycles(EXAMPLE_INPUT)
    
>       assert cycles[20] * 20 == 420
E       assert (1 * 20) == 420

2022/10/python/test_day10.py:156: AssertionError
=========================== short test summary info ============================
FAILED 2022/10/python/test_day10.py::test_run_cycles - assert (1 * 20) == 420
========================= 1 failed, 36 passed in 0.17s =========================

[hollow void]

amazon: "often bought together: Data Structures in Java - rope - stepladder"

[naive sun]

Python 3.10.8 locally and CPython (3.11.0) in CI

[half mirage]

Every time I tried to use a cast in a case statement, Pytest freaked the fuck out

it did NOT like me doing that

[naive sun]

How so?

[half mirage]

ofc I can't recreate it now 🙃

it was this 50 line error

[naive sun]

Huh
For me all it's doing is making it to where x is always 1, and stopping after 35 cycles for some reason

[chilly plank]

I would have guessed that int was a constraint, not a cast

[hollow void]

def day10(inp: str) -> tuple[int, str]:
    xhist = [1]
    for i in inp.strip().split("\n"):
        match i.split(" "):
            case "noop",:
                xhist.append(xhist[-1])
            case "addx", val:
                xhist.append(xhist[-1])
                xhist.append(xhist[-1] + int(val))
    gsm = 0
    for i in [20, 60, 100, 140, 180, 220]:
        gsm += i * xhist[i - 1]
    data = []
    for i, j in enumerate(xhist):
        data.append((i % 40 - j) in (0, 1, -1))
    for j, i in enumerate(data):
        print("\u2588\u2588" if i else "  ", end = "")
        if j % 40 == 39:
            print()
    deeta = input("\b\bWhat can you read?\n> ").strip().upper()
    if deeta.isalpha() and deeta.isascii() and len(deeta) == 8:
        return (gsm, deeta)
    print("Misinput, aborting...")
    raise ValueError("Invalid input")```

[chilly plank]

In [5]: a = ["addx", "3"]

In [6]: match a:
   ...:     case ["addx", int(x)]:
   ...:         print('int')
   ...:     case ["addx", x]:
   ...:         print('other')
   ...: 
other

seems to be a constraint, yeah

[nimble shuttle]

makes sense

[naive sun]

OH

ic

[hollow void]

!e py match 4: case int(a): print(a) case _: pass

yep

[violet pond]

@hollow void :white_check_mark: Your 3.11 eval job has completed with return code 0.

4

[hot panther]

Anyone else's solution missing seemingly one pixel?

I mean clearly this is legible but that first Z is killing me

Ok figured it out. x at this point is -1, so my x % 40 check is yielding 39 which causes the pixel not to be drawn

[quasi minnow]

This is perhaps too high-resolution

(it's not even done encoding yet)

190M lines, 1.4G

Yikes

[bright quarry]

>>> match 4:
...   case int():
...     print("it's an int.")
...   case _:
...     print("it's not.")

it's an int.

It can be both

[chilly plank]

typod cast into case

I don't think case statements ever does a cast

[half mirage]

had this exact same issue

my robust unit testing caught this for me though :)

[coral sorrel]

I have poked at mine a bit and I can't get the little errors to go away

[half mirage]

today is really easy for off by one errors

I had to really think about exactly the error I was getting and what I needed to do

[coral sorrel]

I just can't see it 😅

[half mirage]

If you'd like, I'd be happy to have a look at your code once I finish polishing my unit tests :)

[coral sorrel]

Sure

It's only four pixels that are wrong as far as I can see

(spoilered for answer)

[sweet mango]

Everyone has a different answer

[pseudo crane]

guys what is tricky about this task because i think im missing out on sth that is important

[sweet mango]

An understanding of crt beams is important

[pseudo crane]

ehmm i dont know what that means

im talking about part 1 btw

[half mirage]

For me, it helped just being very regimented in coding part 1

[sweet mango]

Do you know how ticks work?

[coral sorrel]

My understanding is that the puzzle inputs are not unique to an individual, there's just an assortment of them

[pseudo crane]

not sure what ure talking about haha

but i can say how i understand it

[half mirage]

you really have to read part 1, IMO. It's not difficult, but getting the precise details just right can be tricky

[pseudo crane]

everytime there is addx i add the value to the "queue"

that has to wait 2 cycles to be added to the x

[coral sorrel]

The 'tricky' bit in part 1 is that the updating register operation takes two cycles

[sweet mango]

I just realized our "communication device" is a pager

[pseudo crane]

then it goes onto the next value in the queue

[coral sorrel]

so you might have to read the answer in the middle of it

[wispy topaz]

I made this mistake too. There is no queue. It simply takes 2 cycles before you go to the next input

[pseudo crane]

so if there is something waiting to be added i dont read the values from addx?

[sweet mango]

When do we get the challenge that becomes O(n^x) if we do it wrong?

[wispy topaz]

read input
cycle
if input is noop, start over
if input is addx
cycle
update x

[sweet mango]

I implemented mine using lots of generators.

[pseudo crane]

what do you mean by start over

[wispy topaz]

get next input

[pseudo crane]

so forget about what was before?

[wispy topaz]

noop is no op aka no operation.

[sweet mango]

basically pass

I did mine kind of like this. ```py
registers = {"X": 1}

def noop():
yield

def addx(value):
yield
yield
registers["X"] += value

opcodes = {
"noop": noop,
"addx": addx,
}

def run(data):
counter = 0
for line in data.splitlines():
oc, *args = line.split()
args = [*map(int, args)]
for _ in opcodesoc:
counter += 1

[pseudo crane]

i feel so dumb XD

i have no clue what u guys mean with this noop

[wispy topaz]

I just did a for loop.
for cycles in range(2)
if noop break
else
update x

[sweet mango]

noop is short for no operation

[pseudo crane]

i dont understand what does it change

[sweet mango]

noop does nothing.

[wispy topaz]

it's 1 of the 2 things you can get in the puzzle input.

[pseudo crane]

ye i know

but i get correct answer for the example

up to 20 cycles

then i get the wrong value of x on other cycles

[wispy topaz]

it's from the puzzle:

noop takes one cycle to complete. It has no other effect.

[pseudo crane]

    x = 1
    cycles_to_register = [20+(x * 40) for x in range(6)]
    addx = []
    num_of_cycles = 0
    for cycle in file:
        num_of_cycles += 1 
        cycle = cycle.strip().split(" ")
        if cycle[0].startswith("a"):
            v = int(cycle[1])
            addx.append([v,2])
            addx[0][1] -= 1

        if addx[0][1] == 0:
            x += addx.pop(0)[0]
    
            
        if num_of_cycles in cycles_to_register:
            print(x*num_of_cycles)

maybe you can take a look

[wispy topaz]

that seems odd. are you using the test input to verify your code?
I made the mistake of using the actual input

[pseudo crane]

on cycle 20 x is 21 for me

and its correct

[wispy topaz]

why do you have a list addx ?

[wispy topaz]

that's good, is 40 correct?

[pseudo crane]

addx is my so called queue

which some1 said doesnt have a place here

i store every value from addx there

[wispy topaz]

there is no queue. it's 2 cycles and then add.

no need to queue anything, your queue will never have more than 1 item

well maybe if you never take anything off of the queue it will act as a history.

[pseudo crane]

so for example

if i have
addx2
noop
noop
addx 3
addx 5
noop

what should the x be

4?

[wispy topaz]

but it will be:

queue is empty
cycle
read input
if addx: add to queue
cycle
remove from queue
queue is empty

[wispy topaz]

1 +2 +3 +5 is 10, and you used 2 + 1 +1 + 2 +2 +1 = 9 cycles

[pseudo crane]

ohhhh

i think i get it

so just correct me if im wrong now

when i read addx 2

i add 2 to cycles count

and add 2 to the x

and then move on to the next line

[wispy topaz]

sort of

[pseudo crane]

haha

[wispy topaz]

it's impoartnt that you understand that it's:

if addx:
+1cycle
+1cycle

and not:

if addx:
+2cycle

[coral sorrel]

⬆️

[wispy topaz]

there is stuff happening during each cycle.

and 1 thing happening after both cycles, but not during the cycle. 🙂

[chilly plank]

there is the option of replacing the 2 cycle addx with a noop + 1 cycle addx

[coral sorrel]

Like the issue I suspect is that at some point one of the points you have to read at happens in the middle of an addx

and the register is only changed on the second cycle of the addx

[wispy topaz]

on after

[coral sorrel]

yes, sorry

[pseudo crane]

so whats the difrence in

            for _ in range(2):
                num_of_cycles += 1

from

if cycle[0].startswith("a"):
            num_of_cycles += 2

if after reading addx 3

i have to wait 2 cycles for the 3 to be addex to x

what can happen during this 2 cycles that addx 3 takes to complete

[wispy topaz]

what is your objective? What number are you trying to turn in for your puzzle? think about how that number is calculated and when.

[pseudo crane]

ok i think i got it

thanks a lot ;d

i know why it mattered

at least i think so

because while addx is happening it may be cycle 20 that i have to register

and if i just added 2 cycles i may never be on cycle 20

[stark quest]

Dumb question, but im trying to figure out my loop condition, we need to read through every command in the list, but we also need to keep going until we read through all the cycles right?

[pseudo crane]

how the hell do i get correct signal strengths on all cycles except 220

[bright quarry]

I used a queue, when it's noop I just add a noop to the queue, when it's addx I add both a noop and an addx

[bright quarry]

oof, that hurts

[pseudo crane]

i got it

thank you to every1 who helped me

hopefully i get part2 without problems XD

[half mirage]

Just finished my code and test refactoring for today

Tests have 673 lines 👀

My code for today:

import collections
from collections.abc import Iterator, Container
import dataclasses
import enum
import itertools


class CycleState(enum.Enum):
    START = enum.auto()
    DURING = enum.auto()
    AFTER = enum.auto()


@dataclasses.dataclass(frozen=True)
class Cycle:
    no: int
    state: CycleState


@dataclasses.dataclass
class Instr:
    cycle_no: int
    add_val: int

    def decrement_cycle(self) -> None:
        self.cycle_no -= 1


class CathodeRayTube:
    def __init__(self, program: str, watch_cycles: Container[Cycle] = ()) -> None:
        self.program = collections.deque(program.splitlines())

        self.register = 1

        self._cycle_gen = self._cycle_gen_fn()
        self.cycle: Cycle

        self.instr: Instr | None = None

        self._watch_cycles = watch_cycles
        self.watched_cycles: dict[Cycle, int] = {}

        self.screen = ""

    @classmethod
    def read_file(cls, watch_cycles: Container[Cycle] = ()) -> "CathodeRayTube":
        with open("input.txt") as f:
            return cls(f.read(), watch_cycles=watch_cycles)

    @staticmethod
    def _cycle_gen_fn() -> Iterator[Cycle]:
        for x in itertools.count(1):
            for y in CycleState:
                yield Cycle(x, y)

    def _increment_cycle(self) -> None:
        self.cycle = next(self._cycle_gen)

    def _parse_instr(self, instr_text: str) -> None:
        match instr_text.split():
            case ["noop"]:
                instr = Instr(cycle_no=1, add_val=0)
            case ["addx", add_val]:
                instr = Instr(cycle_no=2, add_val=int(add_val))
            case _:
                raise Exception("Unexpected instruction")
        self.instr = instr

    def _execute_pre_cycle(self) -> None:
        if self.instr is None:
            try:
                instr_txt = self.program.popleft()
            except IndexError as exc:
                raise StopIteration from exc
            self._parse_instr(instr_txt)

    def _execute_mid_cycle(self) -> None:
        if ((self.cycle.no - 1) % 40) in (
            self.register - 1,
            self.register,
            self.register + 1,
        ):
            self.screen += "#"
        else:
            self.screen += "."

        if self.cycle.no % 40 == 0:
            self.screen += "\n"

    def _execute_post_cycle(self) -> None:
        assert self.instr
        self.instr.decrement_cycle()
        if self.instr.cycle_no == 0:
            self.register += self.instr.add_val
            self.instr = None

    @property
    def signal_strength(self) -> int:
        return self.cycle.no * self.register

    def __iter__(self) -> "CathodeRayTube":
        return self

    def __next__(self) -> int:
        self._increment_cycle()

        cycle_fns = {
            CycleState.START: self._execute_pre_cycle,
            CycleState.DURING: self._execute_mid_cycle,
            CycleState.AFTER: self._execute_post_cycle,
        }

        cycle_fns[self.cycle.state]()

        if self.cycle in self._watch_cycles:
            self.watched_cycles[self.cycle] = self.signal_strength

        return self.signal_strength


def main() -> None:
    key_cycles = [
        Cycle(no=20, state=CycleState.DURING),
        Cycle(no=60, state=CycleState.DURING),
        Cycle(no=100, state=CycleState.DURING),
        Cycle(no=140, state=CycleState.DURING),
        Cycle(no=180, state=CycleState.DURING),
        Cycle(no=220, state=CycleState.DURING),
    ]

    crt = CathodeRayTube.read_file(watch_cycles=key_cycles)

    for _ in crt:
        pass

    print(
        "Sum of interesting signal strengths:",
        sum(crt.watched_cycles.values()),
        end="\n\n",
    )
    print("Screen:\n")
    print(crt.screen)


if __name__ == "__main__":
    import timeit

    print(timeit.timeit(main, number=1))

And my unit tests, which is multiple times larger than that:
https://github.com/Fiddle-N/advent-of-code-2022/blob/master/day_10/test.py

[chilly plank]

😭

[naive sun]

Hello

Why?

[sweet mango]

when your tests are longer than your actual program.

[mossy dock]

I think I'm missing the first column but well it worked

[half mirage]

For all the test cases

You’ll see if you have a look

[mossy dock]

I think they mean why did you have to make it that long lol

[half mirage]

I had to make it that long to fit all the test cases

I want to be able to assert every detail given in the problem description about the test inputs

Part 2 has a crap ton of test cases though

20 cycles * 3 cycle stages (before, during, after) leads to around 60 cases to assert just for part 2

[naive sun]

I still don't even understand how it works

I'm opening up the docs just to figure out what a Container is

[chilly plank]

I had a weird solution for dealing with part 1, but otherwise my dist code is straightforward

fn solve(input: &[Command]) -> AocResult<(i32, String)> {
  use Command::*;

  let mut res = 0;

  let mut cycle = 0;
  let mut x = 1;

  let mut display = vec![];
  let mut draw = |pos, x| {
    display.push(pos - 1 <= x && x <= pos + 1);
  };

  for cmd in input {
    let prev_cycle = cycle;
    let prev_x = x;

    match cmd {
      Noop => {
        draw(cycle % 40, x);
        cycle += 1;
      }
      Addx(arg) => {
        draw(cycle % 40, x);
        cycle += 1;
        draw(cycle % 40, x);
        cycle += 1;
        x += arg;
      }
    }

    if (20 + prev_cycle) % 40 > (20 + cycle) % 40 {
      let aligned = cycle - cycle % 20;
      res += aligned * prev_x;
    }
  }

  let display = display
    .chunks_exact(40)
    .map(|row| {
      row
        .iter()
        .map(|b| if *b { '#' } else { '.' })
        .collect::<String>()
    })
    .join("\n");

  Ok((res, format!("\n{display}")))
}

parsing code left out, it's not too interesting

[wispy topaz]

Jeebs.

[chilly plank]

just parses the thing into commands, rust enums are neat

pub enum Command {
  Addx(i32),
  Noop,
}

[half mirage]

So that’s based off the fact that all I’m doing with that parameter is doing an “in” check on it

I pass in a list in one place, and dict.keys() in another, both of which support doing an “in” check

That means they both use __contains__, so they are both a Container

[chilly plank]

why use a deque for the instructions?

[half mirage]

Just so I can pop left . It’s a bit gratuitous I admit

[naive sun]

https://github.com/Fiddle-N/advent-of-code-2022/blob/master/day_10/process.py#L136-L142
Why do you get two blank lines at the end in two different ways?

https://github.com/Fiddle-N/advent-of-code-2022/blob/master/day_10/test.py#L5
https://github.com/Fiddle-N/advent-of-code-2022/blob/master/day_10/test.py#L157
Why is one a constant and one is not?

That's all I got because I don't understand the code

[violet pond]

day_10/process.py lines 136 to 142

print(
    "Sum of interesting signal strengths:",
    sum(crt.watched_cycles.values()),
    end="\n\n",
)
print("Screen:\n")
print(crt.screen)```
`day_10/test.py` line 5
```py
LONG_PROGRAM = """\```
`day_10/test.py` line 157
```py
program = """\```

[half mirage]

I didn’t have to use end for the first print, that much is true

And the constant is just because I have that long program in two test functions and didn’t want a copy paste of it

And hey, my code is understandable! XD

More than the golf courses that’s for sure

[naive sun]

It's very nice and clean
I'm just not a very good dev

[chilly plank]

||and verbose||

[half mirage]

I am verbose I admit

[naive sun]

That's not bad though

[chilly plank]

it can be

[naive sun]

necessarily

yeah

[half mirage]

I could stand to turn it down a notch I admit

Mostly with the variable names though

[naive sun]

nah, this is fine

[stark quest]

using the test input, for some reason mine gets the first two but thats it

cycle, val, cycle*x
20 21 420
60 19 1140
100 17 1700
140 30 4200
180 16 2880
220 18 3960

[naive sun]

Well, actually, no
Instr is bad
Don't shorten variable names
Not verbose enough

[stark quest]

the vals are wrong as well

[half mirage]

Naming things is unironically difficult

[naive sun]

There's only two hard things in computer science, cache invalidation, naming things, and off-by-one errors

[half mirage]

Yup that’s the joke I had in mind 😂

Had to deal with two of those today

[naive sun]

You too!?

[half mirage]

With all the testing I did, I still got an off by one error afterwards 😂

But I quickly found it

[stark quest]

and this was an off by one error lol

to interject

[half mirage]

Have you worked out the error or still trying to figure it out?

[stark quest]

i got it

i init'ed my cycle count at 1 for some reason

def check_cycle(cycle, x):
    if cycle % 40 == 20: 
        return cycle * x
    return 0

def part_one():
    with open("day10/instructions.txt", "r") as f:
        instructions = f.read().split('\n')

    x = 1
    cycle = 0
    strength = 0

    for i in instructions:
        if 'addx' in i.split():
            _, val = i.split()

            cycle += 1
            strength += check_cycle(cycle, x)

            cycle += 1
            strength += check_cycle(cycle, x)

            x += int(val)
        else:
            cycle += 1
            strength += check_cycle(cycle, x)
    print(strength)

[chilly plank]

to give some criticism, I see a bunch of methods that are used once and do trivial work

e.g. _increment_cycle should just be inlined in the one place it's called imo

read_file is a paper thin wrapper around the constructor with a file hard coded, why not just read the file and pass the string to the constructor in main?

[wispy topaz]

make sure you check during every cycle, and not skipping cycles, or after cycles.

[stark quest]

yeah i fixed it

accidently defined cycle = 1

[quasi minnow]

Lmao that'll do it

[half mirage]

So, the increment_cycle I have as a separate function just because it reads nicer. self._increment_cycle is IMO clearer than next(self._cycle_gen()) as to what it does

And yes, read_file absolutely doesn’t have to be a class method, it could be a separate one

But I like alternative initialisers :)

[pseudo crane]

@wispy topaz thanks a lot at part 2 i understood even more why making it 2 seperate cycles is that important

but you made me really understand this problem

[chilly plank]

alternative initializers are fine, but having one that just reads from a hard coded file is very weird from a usability standpoint

[wispy topaz]

Wow, your code is the antithesis of mine. it's impressive in it's level of detail, but I would get lost in it.

[wispy topaz]

yw! glad to help!

[chilly plank]

something like this would make more sense

def make_crt_from file(fname):
  ...

(though I would call also that overkill, but that's more in opinion land)

[half mirage]

It’s a fair point. For sure if I had the need to read in from different files I’d do that

I just have it hardcoded cos I know I’ll only read from one file ever in this case

[chilly plank]

you don't run the samples from file?

[wispy topaz]

What, you don't read your testdata from different input files and test the testdata as well? 🙂

[half mirage]

I get lost in the golf courses 😂

[quasi minnow]

If you're not careful, you might get some business logic in with those tests!

[wispy topaz]

Yeah, I think the golf is a fun experiment, but I don't think it's how people should be coding typically.
but there's quite a difference between "minimal readable code" and "unreadable golf"

[stark quest]

Could also make the center part a for loop to clean it up

[half mirage]

Oh yeah for sure! Totally agree

[wispy topaz]

You have the same couple of lines duplicated in 3 places. There might be a logical way to remove the duplicates.

[chilly plank]

I tend to opt to keep things as simple as possible, no extravagant structures unless they are obviously good ideas from the beginning

During writing patterns tend to emerge and you can factor pieces out

[quasi minnow]

^

People spout DRY but they need to consider WET

('Write Everything Twice', the principle that your code is actually fairly likely to diverge if you've only had to write it twice)

[half mirage]

I am definitely overengineering my code for these puzzles, that’s for sure

[chilly plank]

Oh, I do try to follow DRY, but things that repeat tend to become very obvious as you write code

[half mirage]

But the reason that the structures are there are so the unit tests work nicely

[quasi minnow]

I follow DRY as/where it makes sense
I.E. if the code has the same structure and purpose

[wispy topaz]

There's also the concept of reasonable numbers. The only reasonable numbers in computer science are 0, 1 and inf. 🙂

[quasi minnow]

If the code is for two different things that just look similar, that's where WET applies

[half mirage]

Well, that’s part of it. The other part is that I like named attributes for things

[wispy topaz]

me too.

[chilly plank]

I do find the fact that iter steps partway through steps a bit weird as well, but that could just be me. There could still be a split internally for unit testing purposes "execute_pre_cycle/execute_mid_cycle/execute_post_cycle", but a lot of the cruft seems to be to try to accommodate the iterator part

[tidal forum]

a

[chilly pumice]

with open("Inputs/10.txt") as f:
    ops = f.readlines()

# Cycle: value
graph: dict[int, int] = {}

cycle = 0
x = 1

for op in ops:
    op = op.strip()
    if op == "noop":
        cycle += 1
        graph[cycle] = x
    else:
        to_add = int(op.split(" ")[-1])
        graph[cycle] = x
        graph[cycle+1] = x
        x += to_add
        graph[cycle+2] = x
        cycle+=2

from pprint import pprint
pprint(graph)

and yes I did call that variable graph....

I couldnt think of a better name lol

[tidal forum]

i think you did your cycles wrong

[chilly pumice]

alright lemme try it again then

[chilly pumice]

I don't see how tho

[wispy topaz]

but I can't say my attributes always have great names during aoc, see my day8, lot's of them:

    # create an empty class
    class Tree(namedtuple('Tree', 'x y h')): v=0;s=1

    #read the data into a dictionary
    data = open(filename).read().splitlines()
    forest = {(x,y):Tree(x,y,h) for y,row in enumerate(data) for x,h in enumerate(row)}
    H,W = max(forest)

    # loop over each tree
    for t in forest.values():
        # loop over the 4 directions
        for z,d in [
                (list(zip( range(0,t.x)[::-1], repeat(t.y))),         t.x),
                (list(zip(   range(t.x+1,W+1), repeat(t.y))),         W-t.x),
                (list(zip(        repeat(t.x), range(0,t.y)[::-1])),  t.y),
                (list(zip(        repeat(t.x), range(t.y+1,H+1))),    H-t.y)]:
            t.v |= all(t.h>forest[x,y].h for x,y in z)
            t.s *= next((c for c,e in enumerate(z,1) if t.h<=forest[e].h),d)

    print('part1', sum(t.v for t in forest.values()))
    print('part2', max(forest.values(),key=lambda t:t.s).s)

[wispy topaz]

day7 at least had better names:

class File(namedtuple('File', 'name fsize folders files parent')):
    @property
    def size(self):
        return self.fsize + sum(d.size for d in [*self.folders.values(),*self.files])

[half mirage]

As you say, it’s really for the unit testing. If I didn’t have to care about unit testing then I’d definitely make it be one iteration per full cycle. But testing the intermediary steps is really the key thing IMO, as it’s the place one is most likely to slip up on

[tidal forum]

what you're doing looks really weird, are you starting your cycle count at 0 or 1

[chilly pumice]

0

[half mirage]

Any reason to use namedtuples over dataclasses btw?

[tidal forum]

if the first instruction is noop then your graph should look like {0: 1} no?

[wispy topaz]

1 liner class definition. That's about it really.

[half mirage]

Fair

[chilly pumice]

no

[tidal forum]

why

[chilly pumice]

its cycle:value, not the other way around

[wispy topaz]

also useful if you want attributes whose value cannot be changed.

[tidal forum]

yeah, i know

[chilly pumice]

so 1 cycle, 0 value

[chilly plank]

I'm also curious why _parse_instr isn't a free function that returns an Instr, it doesn't really have any business being part of the class. It doesn't depend on anything in the class at all

def parse_instr(self, instr_text: str) -> Instr:
    match instr_text.split():
        case ["noop"]:
            instr = Instr(cycle_no=1, add_val=0)
        case ["addx", add_val]:
            instr = Instr(cycle_no=2, add_val=int(add_val))
        case _:
            raise Exception("Unexpected instruction")
        return instr

class ...:
    # ...

    def _execute_pre_cycle(self) -> None:
        if self.instr is None:
            try:
                instr_txt = self.program.popleft()
            except IndexError as exc:
                raise StopIteration from exc
            self.instr = parse_instr(instr_txt)

[half mirage]

Dataclasses has a frozen parameter

[chilly pumice]

oh wait reg starts at 1

[tidal forum]

on the end of the 0th cycle (which is your first cycle) . . .

[chilly pumice]

so it would be 1:1

[half mirage]

Since dataclasses came out, namedtuples have basically been dead to me

[wispy topaz]

That's like having milk on your poptarts, you might as well cook at that point. 🙂

[tidal forum]

if you're doing 1:1 then why does your dict look like {0:1, 1:1, 2:2} if there's an addx 1

[wispy topaz]

For actual non-aoc coding I do really enjoy dataclasses a lot.

[half mirage]

Friendship ended with namedtuple
Now dataclasses is my best friend

[half mirage]

Dataclasses and enums are my favourite for my over engineered AOC code :)

[chilly plank]

I don't really like python enums

compared to enums in other langs

[chilly pumice]

um, 0:1 is 0th cycle and value has 1. 1:1 is first cycle, but value wont increase yet. and in 2nd cycle value will finally be increased. I'm sorry what am I missing

[chilly plank]

they are very much not first class citizens in python

[half mirage]

This could be a free function definitely. My unofficial rule of thumb is that if it’s a helper function only used in the one class, I keep it in that class. But if it were to be used across classes I’d bump it out of the class

[chilly plank]

if it doesn't depends on the state of the class, it for sure shouldn't be a member function

I guess you could make a case for a staticmethod or classmethod

[half mirage]

Definitely staticmethod

[chilly pumice]

okay got it!

[half mirage]

Definitely not classmethod

[chilly pumice]

my starting cycle was wrong @tidal forum. Thank you

[chilly plank]

(I keep forgetting the exact difference, I know staticmethod is the one you usually want)

[stark quest]

I dont really understand what we are doing in part 2. X is the position of the middle of our sprite? And that somehow relates to what gets drawn?

[chilly plank]

ah classmethod takes class as an arg

[half mirage]

Class method is really mostly for alternative initialisers

[calm shard]

If I ever feel the need to type @staticmethod. I stop and write a free function instead.

If it doesn't need an instance and doesn't need the class, it has nothing to do with the class and shouldn't be there.

[half mirage]

I don't mind the odd static method. Though at the moment I'm standing at 2 static methods having edited my code, which is definitely the upper limit I'd be comfortable with

If it's a function only used by a class, even if it doesn't need self, I'm ok with it as long as it's sparingly used

[calm shard]

tbh if it's basically a helper to another method I have no issue with it just being a method. regardless if you use self or not.

Otherwise, there's no room at the inn!

[stark quest]

is the gist of part 2?

if cycle in range(x-1,x+3): return '#'
return '.'

[calm shard]

I have a similar line of code in my solution... so it's not a terrible idea

[chilly plank]

wrong range, but basically

[stark quest]

oh yeah x+2

because we still need inclusive

and then like print newline at cycle % 40

[calm shard]

tbh I just went with in (x-1, x, x+1)

[chilly plank]

I just dumped true/false into a list and built the actual string afterwards

[stark quest]

like a list of bools and #/. based on the truth value?

[chilly plank]

I would do x-1 <= cycle <= x+1

feels like it describes the intent better

[calm shard]

Dunno, is the beam in one of these 3 positions feels better to me.

I'll see what my senior says at code review 😛

[half mirage]

I did enjoy the code review of my code just now. Usually I'm the one rigorously reviewing people's code, not the other way around XD

[chilly plank]

hopefully it comes off as constructive criticism rather than just ranting 😛

[half mirage]

Nah it's all good :)

[naive sun]

And a slots parameter too!

[stark quest]

wait, how are you guys printing the drawing as a contiguous block? I put everything into a list and was just going to iterate over it but that just prints down the terminal lol

and i couldnt just print the list because of the newlines

nvm i can see the data is wrong anyways

oh i had the wrong dataset loaded

long day

nah either way this is wrong lol

but the test input matches what is shown

[vernal urchin]

i just build up the solution as a string

[stark quest]

wtf

oh actually its just the first line thats printing

then its . always

[rancid cave]

i print with end=""

[chilly pumice]

day 10 part 2 done!

[stark quest]

isnt it actually cycle % 40 in [x-1,x,x+1]? because my cycle counter keeps going up to 240 and the x obviously never goes that high

[chilly pumice]

I really don't like my code tho

its garbage

[modest wren]

1-based indeximg go brrr

[stark quest]

cycle % 41

[potent wadi]

(cycle + 1) % 40?

[stark quest]

or that

[potent wadi]

I was pretty happy with my solution today

I didnt think the problem itself was that hard, working out what you really had to do was quite challenging though

[gray olive]

it was definitely easier than day 9

[chilly pumice]

I feel day 9 was easier

[gray olive]

the difference between part 2 and part 1 today is a couple variables and an if

[potent wadi]

Day 9 wasnt too tough either - I just got stuck on a silly bug for ages

[chilly pumice]

day 9 I at least knew what I was doing. day 10 I was literally changing the order of the lines to make it work

noop was causing issues

[gray olive]

i couldn't figure out how part 2 worked in day 9

[chilly pumice]

its part 1 but with a loop

pretty simple imo

[potent wadi]

If you kept a history of the register for each cycle then pt 2 was incredibly easy

[gray olive]

how the loop worked idk

[chilly pumice]

this is the loop you had to add https://github.com/Chiggy-Playz/AdventOfCode/blob/master/2022/Python/day9.py#L42

[violet pond]

2022/Python/day9.py line 42

for i in range(1, len(rope)):```

[chilly pumice]

well

[potent wadi]

Oh this is the rope one

[chilly pumice]

not very helpful

[potent wadi]

not the forrest trees one

[chilly pumice]

yeah

[potent wadi]

I liked that one actually

Pt 2 felt like a good extension as well

[chilly pumice]

me too

[potent wadi]

basically "take pt 1, make it work for n dimensions"

[chilly pumice]

I couldnt believe when my code ran first try and was correct

[potent wadi]

Yeah, the answer somewhat fell out on me too

[chilly pumice]

agreed

[gray olive]

um
#1050641552524976129 message

where to add?

[chilly pumice]

you expect me to read that? :kekw:

[potent wadi]

You really need to layout your code in a more human friendly way 😂

[chilly pumice]

I hate code golfing

like why

[gray olive]

that's literally the first thing that comes to my mind

what do you expect me to do in a rush

[potent wadi]

I always find that if you rush and dont take the time to make it readable you will never fix the bugs you come up against

[modest wren]

that's only an issue if you have bugs

[chilly pumice]

why rush

[modest wren]

to be faster

[chilly pumice]

eh

[gray olive]

i can read it all right

but the only thing i can't figure out is the bug

how does part 2 work

[chilly pumice]

what you're doing for head and tail, you gotta apply for current knot and previous knot

[calm shard]

Formatting usually comes second... sometimes third... because when I see what Part 2 is I usually need/want to refactor Part 1 to make it nicer code. Then I will format it. 😄

[chilly pumice]

and loop over all knots

[chilly pumice]

I do follow this too. Part 1 is just "throw anything at it" 😂

[gray olive]

the U 8 in the test case gives me ```
H
1
2
3
4
98765

[gray olive]

i did

it worked for part 1

[chilly pumice]

rewrite it with better names and I might just look at it 🙃

[calm shard]

I do get lucky sometimes and I correctly predict Part 2 well enough that my P1 code is mostly still suitable 😄

[modest wren]

wrong day ?

[chilly pumice]

or you can just see how my code works and compare. You say you are doing this, so it should be comparable

[chilly pumice]

yep happened to me on treehouse

[gray olive]

idk we're just talking about the difficulty of day 9

[potent wadi]

perhaps this is linked to your message above

[nocturne aurora]

I don't like that this one is nearly impossible to complete programmatically

I thought this was Advent of Code, not Advent of Reading

[gray olive]

my part 1 worked something like ```py
if abs(head - tail) > sqrt(2):
tail = previous_head

[chilly pumice]

yeah lmao

[nocturne aurora]

reminds me a bit of the cursed picoCTF 2022 challenge Wizardlike where it was nearly impossible to read the (ASCII art) flag and actually complete the challenge

[chilly pumice]

when I read What eight capital letters appear on your CRT? I was like wtf

[modest wren]

things like this have shown up before, to be fair

[chilly pumice]

that is EXACTLY what I did for part 1

but as you can see from my github, I ditched that

[gray olive]

come on

[nocturne aurora]

is this spoilers for day 9

[vernal urchin]

you all are not talking about day 10 right?

[nocturne aurora]

if so why

[chilly pumice]

maybe we should move

[gray olive]

probably

[vernal urchin]

I was just confused more than anything, but yeah

[stark quest]

did you use Euclidian distance or something?

[nocturne aurora]

I'm getting spoiled for day 9 lol I should wait to go to this channel

[gray olive]

what type of distance is the thing where you don't square because that's exactly what i did

[chilly pumice]

how are you on day 10 without 9 ignore, me stupid

[calm shard]

Please move Day 9 chat to #1050641552524976129

You can skip days

[chilly pumice]

riiiiiiiiiight

[stark quest]

https://iq.opengenus.org/content/images/2018/12/distance.jpg

[calm shard]

Please move Day 9 chat to #1050641552524976129

[nocturne aurora]

there's no need to do all of them

[chilly pumice]

yeah

[nocturne aurora]

Manhattan?

[chilly pumice]

I forget only I try to lol

[gray olive]

nvm it does

[nocturne aurora]

okk

[calm shard]

https://tenor.com/view/no-samueljackson-stare-gif-10537769

@nocturne aurora @gray olive

[gray olive]

see this is the reason i don't leave this chat
i keep getting pinged

[nocturne aurora]

😂

for Day 10 Part 2 I just add dummy array item before every addx, this is easiest way

[junior coyote]

assuming register start point is 1

and i have addx 15

cycle 1: register = 1

cycle 2: register = 16

[vernal urchin]

no

you update X after cycle 2

but before cycle 3

[junior coyote]

yes so while my cycle var is 2

register is 16

i mean

[vernal urchin]

any updates during cycle 2 must use X=1

you will only get X=16 effectively in 3rd cycle

[half mirage]

during cycle 2: x = 1
after cycle 2: x = 16

part 1 only cares about during values

[junior coyote]

im part 2

[vernal urchin]

to be fair even in part 2, the drawing to screen is based on during-value

[half mirage]

well, part 2 as well, if we are talking about setting pixels on the screen

[vernal urchin]

if line.startswith("addx"):
  tick()
  tick()
  X += int(line[4:])

essentially

[stark quest]

lets go

[junior coyote]

[(1, 1), (2, 1), (3, 16), (4, 16), (5, 5), (6, 5), (7, 11), (8, 11), (9, 8), (10, 8), (11, 13), (12, 13), (13, 12), (14, 12), (15, 4), (16, 4), (17, 17), (18, 17), (19, 21), (20, 21), (21, 21), (22, 20), (23, 20), (24, 25), (25, 25), (26, 24), (27, 24), (28, 29), (29, 29), (30, 28), (31, 28), (32, 33), (33, 33), (34, 32), (35, 32), (36, 37), (37, 37), (38, 36), (39, 36), (40, 1), (41, 1), (42, 2), (43, 2), (44, 26), (45, 26), (46, 7), (47, 7), (48, 8), (49, 8), (50, 24), (51, 24), (52, 13), (53, 13), (54, 13), (55, 13), (56, 34), (57, 34), (58, 19), (59, 19), (60, 19), (61, 19), (62, 16), (63, 16), (64, 25), (65, 25), (66, 26), (67, 26), (68, 23), (69, 23), (70, 31), (71, 31), (72, 32), (73, 32), (74, 37), (75, 37), (76, 37), (77, 37), (78, 37), (79, 37), (80, 37), (81, 1), (82, 1), (83, 1), (84, 2), (85, 2), (86, 9), (87, 9), (88, 9), (89, 9), (90, 9), (91, 11), (92, 11), (93, 17), (94, 17), (95, 17), (96, 17), (97, 17), (98, 17), (99, 17), (100, 18), (101, 18), (102, 18), (103, 18), (104, 25), (105, 25), (106, 26), (107, 26), (108, 26), (109, 13), (110, 13), (111, 26), (112, 26), (113, 33), (114, 33), (115, 33), (116, 34), (117, 34), (118, 1), (119, 1), (120, 1), (121, 1), (122, 1), (123, 3), (124, 3), (125, 3), ]

[stark quest]

back on track

[junior coyote]

so i have this

[nocturne aurora]

why do you have this

[vernal urchin]

ayy nice

[junior coyote]

in cycle one my sprite is on which pixel

?

i cant understand the next

[vernal urchin]

cycle 1/2 sprite is at X=1

[junior coyote]

but why is the sprite 3 twitter symbols

[stark quest]

I will be honest I couldnt figure out how to get the newlines working so i just used a library

print(textwrap.fill(''.join(drawing), 40))

lol

[vernal urchin]

because it's 3 characters wide

the leftmost character is at X=1

but it covers columns 1,2,3

[junior coyote]

ok so my first tuple should be drawn

a #welcome

on my render correct?

[vernal urchin]

Yeah first 2 should be ## because you can't move the sprite until 3rd cycle

[junior coyote]

at cycle 40

38

it will be

.....................................###```

?

[vernal urchin]

at cycle 38 it seems like the sprite is at X=36, so it covers 36,37,38, so you should get a #

[nocturne aurora]

str(map(lambda x:"\n"+x[1]if x[0]%40 else x[1], drawing))[1:]
``` should work

[calm shard]

seriously? 😄

"twitter symbols"

[naive sun]

That better be a language barrier

[vernal urchin]

octothorpe

[calm shard]

pound!

[naive sun]

WTF is that?

[junior coyote]

and on cycle 42 will it be

...................................###..```
or will it be

.......................................#```

[vernal urchin]

o c t o t h o r p e

[junior coyote]

which one of these is correct

[half mirage]

hash

[vernal urchin]

is octothorpe

[half mirage]

tag

[calm shard]

octothorpe is overly fancy name for it lol

[naive sun]

It's a hashtag you dummy

[vernal urchin]

it is, I like being that guy

[junior coyote]

i forgot its hashtag

[calm shard]

no just hash

[vernal urchin]

but it's not a hashable value

[junior coyote]

pls anyone

xd

[naive sun]

Why not?
It's a str

[echo wolf]

is the hash. A hashtag is the hash plus some text, like #python

[vernal urchin]

i don't want to count the dots lol

[vernal urchin]

stop bringing logic into this

[junior coyote]

its not about counting dots

its about will the thing go into the edge

or will it return

the moment the 3 hash touch the right side

*reverse

[vernal urchin]

on cycle 42 you are on the second row

and X=2

[nocturne aurora]

is an octothorpe

[junior coyote]

wth

[vernal urchin]

so it will be

..XXX....

[junior coyote]

i thought it moves like

[vernal urchin]

first 40 cycles = 1st row

[nocturne aurora]

is not a hash, a hash is 334359b90efed75da5f0ada1d5e6b256f4a6bd0aee7eb39c0f90182a021ffc8b

[vernal urchin]

second 40 cycles = 2nd row

[junior coyote]

https://tenor.com/view/cylon-battlestar-galactica-looking-around-scan-gif-9153290

i thought it moves like this

=(((((

[vernal urchin]

you get the "column" by cycle % 40

[naive sun]

You're right, I'm sorry, I forgot we're on Twitter
I'll now proceed with normal Twitter posting -
@vernal urchin is fired for wanting to use Python instead of Rust BTW

[junior coyote]

rust language is garbage

c++ is better it has more plus signs

[vernal urchin]

no, rust has no garbage collection

[gray olive]

C is better it's much simpler

[vernal urchin]

Java is garbage

collected

[junior coyote]

haha

[vernal urchin]

best programming language is aecor

[nocturne aurora]

boehm ftw

[gray olive]

and dumped into the pile of idk

[junior coyote]

only reason java is not dead is because of my minecraft channel

[naive sun]

Does it have more than one user yet?

[junior coyote]

https://tenor.com/view/minecraft-memes-pig-pigs-minecraft-gif-25398350

[vernal urchin]

100% of the people who have used it have loved it

[naive sun]

Can't argue with that

[vernal urchin]

i made rope simulation with it how bad can it be

[junior coyote]

2 + "2" = 22

or will it be string

i forgot

which makes more "sense"

[vernal urchin]

string

everything gets implicitly converted to string IIRC if js doesn't know what to do

[nocturne aurora]

string

[nocturne aurora]

a bunch of operators need to behave as if it converts to string even if it's not implemented that way (for performance reasons)

actually some security exploits came like this

[wispy topaz]

The only reason java is not dead is because Oracle spends more money on lawyers than programmers.

(I kid, java is widely distributed and will remain relevant for many years to come in some form or another)

[junior coyote]

i would like to learn it for android

but i also dont want to learn it

[naive sun]

Unfortunately

[nocturne aurora]

I almost put laughing emoji, but actually this is unironically true

[junior coyote]

i want to learn rust or c++

[naive sun]

Learn Rust BTW!!
Join us in the future of the developed world!!

[junior coyote]

i cant do second part from today

yesterday was easier

or is it me

[naive sun]

Yesterday took me the entire day
Today took me an hour

[nocturne aurora]

learn Rust! 🦀🚀

[nocturne aurora]

I did today's first than yesterday, because today looks easy

should I do yesterday's

this one second part is almost identical to first part

[junior coyote]

what does entire day mean

how fast did you do today?

[naive sun]

Like 7 hours of working on the problem

[naive sun]

Just under an hour and a half

[junior coyote]

i did the first part in 20min

and now i sit on second part

1 hour

because i cannot understand the task

[junior coyote]

it feels frightening(?) because its syntax is like c++s so much more difficult than python imo

[naive sun]

Yeah I'll admit I had to look at this channel for a second before I completely understood the problem

[kind egret]

Today is a bit like reading a datasheet and building up a simple circuit based on it. It's all about precision, timing, and reading comprehension

[vernal urchin]

Day 9 included a food break in between because i kinda got annoyed after seeing part 2 lmao

[echo wolf]

i remember it being midnight and I was about to go to bed because I was expecting this one to be harder like day 9, but then I quickly read through it and went "man I could do that pretty quickly"

[naive sun]

"08:44"

Takes me longer than that just to understand the problem

[junior coyote]

day9 was nice for brain imo

some other moderator/helper had a animation that looked cool

[naive sun]

salt-die

[junior coyote]

yes

[vernal urchin]

that was me

oh salt-die too I guess

[junior coyote]

show yours

too

id like to see

[naive sun]

here

[vernal urchin]

mine's more of an inspiration to continuous rather than discrete blocks

[junior coyote]

thats cool

[vernal urchin]

I also did a discrete version in python like salt-die but that wasn't as cool as this

[junior coyote]

why is at the end when you have let go of it already

some movement still happening?

[vernal urchin]

because you can move more than 1 pixel in every frame, so I make everything go closer over time

rather than just snapping instantly

basically each node goes closer to its parent at a speed proportional to how far away it is from the expected position

takes a few frames for it to converge

it's not completely rigid, basically, similar to how in the 2d version there was some "leeway" before the child snaps

[vernal urchin]

i meant for part 2, part 1 was pretty straightforward

Day 4 is my proudest achievement

[junior coyote]

###...##..#..#..##..####.###..####.####.
#..#.#..#.#.#..#..#.#....#..#.#.......#.
###..#....##...#..#.###..#..#.###....#..
#..#.#.##.#.#..####.#....###..#.....#...
#..#.#..#.#.#..#..#.#....#.#..#....#....
###...###.#..#.#..#.####.#..#.####.####.```

i get this

am i done finaly?

[vernal urchin]

nice!

yep

[junior coyote]

😭

[warm edge]

Here is my dreadful looking code for part 2

with open('day10\input.txt','r') as f:
    data = f.read().split('\n')


cycle = 0
x = 1
screenline = ''
screen = []
pixel = -1

def add_line(cycle):
    global pixel
    global screenline
    if cycle in [40, 80, 120, 160, 200, 240]:
        screen.append(screenline)
        screenline = ''
        pixel = -1

def draw_pixel(x):
    global pixel
    global screenline
    pixel += 1
    add_line(cycle)

    if pixel in (x+1,x-1,x):
        print(x, pixel, '#')
        screenline += '#'
    else:
        print(x, pixel)
        screenline += '.'
    

for line in data:
    if 'noop' in line:
        cycle += 1
        draw_pixel(x)
    else:
        line = line.strip('addx ')
        cycle += 1
        draw_pixel(x)
        cycle += 1
        draw_pixel(x)
        x += int(line)


for i in screen:
    print(i)```

As always, lmk if you see any obvious improvements

[naive sun]

Doesn't help that I go to sleep like an hour or two before they're released

[vernal urchin]

if you replace . with a space it's easier to read the characters @junior coyote

[echo wolf]

If you think about it, there could be a possibility that someone's random letters spelled out some word

[echo wolf]

just don't sleep

[vernal urchin]

i spent a few hours today morning trying to reverse engineer an input for a custom word, but I gave up lol

[half mirage]

global 👀

[vernal urchin]

here's my cleaned up implementation:

res = 0
screen = ""
X = 1
cycles = 0

def tick():
    global cycles,res,screen
    cycles += 1
    if cycles % 40 == 20:
        res += X * cycles
    
    if cycles % 40 == 0:
        screen += "\n"
    else:
        screen += " #"[X <= cycles % 40 < X+3]

for line in lines:
    if line == "noop":
        tick()
    elif line.startswith("addx"):
        tick()
        tick()
        X += int(line[4:])

print("Part 1:", res)
print("Part 2:")
print(screen)

[junior coyote]

hardcoded

[junior coyote]

lol your advice with " " instead of "."

finally can read

[vernal urchin]

I'm not against hardcoding but it's just... more work in this case

[vernal urchin]

ay

[sterile pagoda]

Try running this program. You'll need to increase the number of rows to 24 though: https://paste.pythondiscord.com/elutisiwex

[vernal urchin]

i don't have a number of rows harcoded, it just stops when the program stops 😛

[sterile pagoda]

Ah nice. Robust!

[gray olive]

why does the input go negative in some places

it completely ruins my VM

[vernal urchin]

lmao this is exactly what i was trying to do

with the rick roll

how did you generate this?

[sterile pagoda]

Erm, I'll send you the function, one sec...

[vernal urchin]

if you want to see the golfed solution we have all been optimizing in the other channel:

R=c=X=0
S=n="\n"
T='c+=1;W=c%40;R-=~X*c*(W==20);S+=n[W:]or" #"[X<W<X+4];'
for L in open(0):exec(T+(T>L)*f"{T}X+=int(L[4:])")
print(R,S)

it's so horrible I love it

[junior coyote]

what does

=~ supposed to do

i have never seen it

[vernal urchin]

R-=(~X)*...

bitwise not

[junior coyote]

i go out of my way to say today was harder than all other ones

my english reading comprehension is horrible ig

🥴

[sterile pagoda]

https://paste.pythondiscord.com/quficorape.py

I didn't do anything fancy to generate the ascii art itself 😄

I looked up pixel fonts but got bored.

[sterile pagoda]

It just looks two pixels ahead at each step, and always uses addx. There are probably more interesting/optimal ways to do it.

[vernal urchin]

that's basically what I was trying to do but i couldn't get it working, my brain was being dumb

then i just said screw it and started working on the rope thing

[junior coyote]

what did you use for rope thing

which gui

[vernal urchin]

custom graphics library in my own programming language lmao

[sterile pagoda]

Plot/puzzle predictions for day 11?

[vernal urchin]

saturday night, prolly gonna be a doozy

[sterile pagoda]

We're still in the ravine right?

[vernal urchin]

I'm guessing some sort of graph traversal

[restive path]

something like day 15 of last year I'd think

[stark quest]

i just used ⬜ and ⬛

[junior coyote]

how

emojis?

[stark quest]

its unicode

but yeah

[junior coyote]

idk what U+2B1C mean

[stark quest]

unicode

def check_draw(cycle, x):
    #Replaced '#' and '.' because they are difficult to read
    if (cycle-1) % 40 in [x-1,x,x+1]: return '⬜'
    return '⬛'

no idea

[mint tangle]

it's the unicode codepoint for that character

[junior coyote]

i think my console didnt like it ig?

it work in other one

oh no

i forgot smth in other=)

nvm

[nocturne aurora]

I used same

[junior coyote]

is that js?

[junior coyote]

how do you read in a file in js if its in the browser :S

[nocturne aurora]

yes

[nocturne aurora]

I did this on the challenge input which opens as a web page when you click the link (because it doesn't have a Content-Disposition that tells your browser to force you to download it)

[junior coyote]

ooh

i see because its somewhere in the HTML document

you can just use it

from browser?

[junior coyote]

its in this pre style="word-wrap" thing

why document.body.textContent

can fetch it

[nocturne aurora]

textContent only gets the text, and strips any HTML tags

it basically takes all the text nodes from every element and puts them together

[junior coyote]

oh so every <asdf> get excluded

cool

ty

+1 knowledge

[nocturne aurora]

:D

[vernal urchin]

@sterile pagoda https://paste.pythondiscord.com/micawisaye

[junior coyote]

can i pm you rq?

[vernal urchin]

threw it together with some of salt-die's code to put a utility together

[nocturne aurora]

sure

[sterile pagoda]

Ah nice

Did you make the letters of the font yourself?

[vernal urchin]

nope, stolen from salt-die which he crowdsourced I think

[gray olive]

i finally made the VM work

[gleaming kelp]

most fun challenge so far ngl

https://github.com/R1c4rdCo5t4/AdventOfCode2022

[gray olive]

now i wanna generate letters

[gray olive]

this one somehow doesn't work

nvm it does

[sweet mango]

First, you must define a font

[gray olive]

salt-die's seems to work pretty well ```

.\get_art.py "There! An entire sentence" > test_art_vm.txt & py vm.py < test_art_vm.txt

[rocky charm]

i made all the lowercase and digits!

and the missing uppers

[gray olive]

where is .

[rocky charm]

i only made one punctuation

[gray olive]

why

[vernal urchin]

I see, your contributions are much appreciated thank you

[rocky charm]

i'll probably do all printables later

[rocky charm]

because my time isn't unlimited

[sweet mango]

You could make a text file containing pre-rendered images of all the ascii characters

[gray olive]

also support a column argument to the program pls

[rocky charm]

it's not a program, it's a text file

[gray olive]

ok

[vernal urchin]

Why not just pipe instead of saving to a file and then redirecting

[gray olive]

does that work in cmd?

[vernal urchin]

I would hope so?

[sweet mango]

The redirects work, but not the &

[gray olive]

it doesn't

[sweet mango]

& is for running something in the background in bash

[vernal urchin]

Ah I guess I tried it on PowerShell at some point

I don't use windows often

Or at all, really

[sweet mango]

in powershell, & is to run a program by file path

pwsh (run program): & ./path/to/file.exe
bash (run program in background): ./path/to/file.sh &

[vernal urchin]

I'm talking about piping with |

Idk what you're referring to

[sweet mango]

<> is also a pipe

[vernal urchin]

🤷‍♂️

[gray olive]

i made the programs work with lengths greater than 40 now

[nocturne aurora]

epic

[gray olive]

turns out that's possible with windows cmd

[vernal urchin]

Yeah I was surprised

That's like, one of the fundamental ways to use a shell

[vernal urchin]

yeah haha i stuck with 40 so nobody would have to change their aoc code 😅

[gray olive]

>.\get_art.py "Have you ever thought of something as   great as this" 200 | .\vm.py 200

works pretty well
now all i want is word wrap instead of character wrap

[sweet mango]

Now do famous song lyrics

[abstract python]

never g-

[sweet mango]

do a url

you know which one

[gray olive]

:, ., and / aren't supported

[hollow herald]

-onna get it never gonna get it

[sweet mango]

Sounds like a bug

[vernal urchin]

add them!

[gray olive]

idk how

[sweet mango]

You need to support the full printable ascii table

0-19 can be skipped

[gray olive]

yeah well idk how salt-die did it

[vernal urchin]

print out the uncompressed data from load_font(), add whatever you want, re-compress it

salt-die just read it from a text file the compression shenanigans are mine

for easy sharing

lemme refactor it

[gray olive]

how do you recompress

[vernal urchin]

https://paste.pythondiscord.com/eluhejukus

here, should be much easier to change the font here

[gray olive]

ok

[nocturne aurora]

print(repr(zlib.compress(font.encode('utf-8'))))

then replace the FONT = line starting with b'

[gray olive]

inverse is actually pretty cool

[tidal forum]

did someone make a program to generate instructions to create anything for a monitor

.-.

[vernal urchin]

yeah look at the link a few messages above

[tidal forum]

nice

[nimble shuttle]

it is almost the hour of choking yet again

[vernal urchin]

indeed, wactch me choke so bad

[quasi minnow]

Lmao

[restive path]

see yall in like 3 hours

[gray olive]

i got word wrap done ```

.\get_art.py "Never gonna tell a lie and hurt you..." 50 | py vm.py 50 : 5

>.\get_art.py 200 < lyrics.txt | py vm.py 200 "#" 5

[rocky charm]

you think this is a better s:

.##..
#....
..#..
##...

[gray olive]

why only 4 lines

[rocky charm]

implied empty lines above it

[gray olive]

ok

probably

[rocky charm]

the 5 is exactly the same as a S too, trying to change it

maybe just add a # at the top-left corner so it isn't rounded

[violet pond]

Hey @gray olive!

It looks like you tried to attach a Python file - please use a code-pasting service such as https://paste.pythondiscord.com

[gray olive]

updated file btw https://paste.pythondiscord.com/bahoceraqi.py

[fading hull]

Had to change my code, after part 1, because it is not any good for part 2.

now with the rework part 1 doesn't work anymore 😦

I seem to be off one cycle, but I don't see where.
In case anyone wants to take a look: https://paste.pythondiscord.com/muvimeruso

[weak dirge]

thanks mustafa 🙂