long fulcrum Jan 25, 2023, 5:07 AM

#

?

meager zinc Jan 25, 2023, 5:07 AM

#

that's how it's encoded in hex

long fulcrum Jan 25, 2023, 5:08 AM

#

do you have to add more

#

i just run it through this print('CODE'.encode().decode('u16'))

#

if it doesnt work just add a space at the end of the code and it works

meager zinc Jan 25, 2023, 5:08 AM

#

That's because it has to be in chunks of 16

#

(two 8-bytes)

meager zinc Jan 25, 2023, 5:09 AM

#

long fulcrum i just run it through this `print('CODE'.encode().decode('u16'))`

This adds the extra stuff behind the scenes

#

to change encodings

long fulcrum Jan 25, 2023, 5:09 AM

#

oh right

meager zinc Jan 25, 2023, 5:12 AM

#

this is the version for windows:

c = __import__('ctypes.util')
c_lib = c.CDLL(c.util.find_library('msvcrt'))
c_lib.puts(
  c.cast(
    b'Hello msvcrt!',
    c.POINTER(c.c_char),
  )
)

versed eagle Jan 25, 2023, 5:13 AM

#

yep
windows doesn't have libc, so you have to use the microsoft c runtime

meager zinc Jan 25, 2023, 5:13 AM

#

Microsoft Visual C Runtime is the equivalent

versed eagle Jan 25, 2023, 5:13 AM

#

the v also stands for something but i forgot since i dont use windows

versed eagle Jan 25, 2023, 5:13 AM

#

meager zinc Microsoft Visual C Runtime is the equivalent

ah
visual

#

thats it

meager zinc Jan 25, 2023, 5:13 AM

#

visual c# actually maybe

#

lemme google it

#

no

#

just

#

Microsoft Visual C Run Time

#

idk why it has "Visual"

#

kinda random

eager ore Jan 25, 2023, 5:14 AM

#

that's why you count in bytes

meager zinc Jan 25, 2023, 5:16 AM

#

I think mac also has libc

#

but I'm not very familiar with mac anyways

meager zinc Jan 25, 2023, 5:17 AM

#

meager zinc idk why it has "Visual"

Apparently this is because the Microsoft Windows ™️ Visual C++ (MSVC) compiler has Visual in it

#

Actually... this means you can write a direct C to Python transpiler... I'm scared about that possibility.

#

how would you implement # pragma though

#

any ideas?

gleaming linden Jan 25, 2023, 5:24 AM

#

import sys
R=range
Q=R(-1,2)
G,E='#.'
A=[E+i+E for i in sys.argv[-1].split('\n')]
W=len(A[0])
*A,=E*W,*A,E*W
for Y in R(1,len(A)-1):print(''.join([E,G][2-(A[Y][X]==G)<sum(G==A[Y+y][X+x]for y in Q for x in Q if y|x)<4]for X in R(1,W-1)))
``` 236

long fulcrum Jan 25, 2023, 5:24 AM

#

jesus

meager zinc Jan 25, 2023, 5:25 AM

#

we've nearly cut it in half

long fulcrum Jan 25, 2023, 5:25 AM

#

when I started earlier it was 1300

#

like way before I posted it here

meager zinc Jan 25, 2023, 5:25 AM

#

oh we've def cut it in half

#

!e print(236 / 1300)

night quarryBOT Jan 25, 2023, 5:25 AM

#

@meager zinc :white_check_mark: Your 3.11 eval job has completed with return code 0.

0.18153846153846154

meager zinc Jan 25, 2023, 5:26 AM

#

it's 18% of original size

gleaming linden Jan 25, 2023, 5:26 AM

#

import sys;R=range;Q=R(-1,2);G,E='#.';A=[E+i+E for i in sys.argv[-1].split('\n')];W=len(A[0]);*A,=E*W,*A,E*W;[print(''.join([E,G][2-(A[Y][X]==G)<sum(G==A[Y+y][X+x]for y in Q for x in Q if y|x)<4]for X in R(1,W-1)))for Y in R(1,len(A)-1)]
``` and oneline version (+1c)

meager zinc Jan 25, 2023, 5:26 AM

#

Best compression algorithm?

long fulcrum Jan 25, 2023, 5:34 AM

#

gleaming linden ```py import sys R=range Q=R(-1,2) G,E='#.' A=[E+i+E for i in sys.argv[-1].split...

how does this skip all the padding?

gleaming linden Jan 25, 2023, 5:34 AM

#

range(1,len(A)-1) and range(1,W-1)

long fulcrum Jan 25, 2023, 5:35 AM

#

it can be range(1,len(A))

gleaming linden Jan 25, 2023, 5:35 AM

#

can it?

long fulcrum Jan 25, 2023, 5:35 AM

#

yep

gleaming linden Jan 25, 2023, 5:36 AM

#

wtf

#

why does that work

#

ah I see

#

it just errors and exits

#

does that count though

long fulcrum Jan 25, 2023, 5:37 AM

#

yep

gleaming linden Jan 25, 2023, 5:37 AM

#

I mean

#

we can just ignore stderr

#

and just look at stdout

#

ig that works

long fulcrum Jan 25, 2023, 5:59 AM

#

ok im having some issues ungolfing droidcry

#

i want to save a nice looking version too

#

import time
import os

dead = "⬜"
alive = "⬛"

os.system("cls")
inp = [
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬛⬛",
    "⬜⬜⬛⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬛⬛",
    "⬜⬜⬜⬛⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬛⬛⬛⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
]

for n, m in enumerate(inp):
    inp[n] = inp[n].center(len(inp[n]) + 2, dead)
inp.insert(0, dead * len(inp[0]))
inp.append(dead * len(inp[0]))


def nextgen(inp):
    nxt = []
    for y, i in enumerate(inp):
        temp = ""
        for x, j in enumerate(i):
            tmp = []
            if {y} & {0, len(inp) - 1} or {x} & {0, len(inp[y]) - 1}:
                count = 0
            else:
                for k in range(-1, 2):
                    for l in range(-1, 2):
                        if k == 0 and l == 0:
                            continue
                        tmp.append(inp[y + k][x + l])
            count = tmp.count(alive)
            if {y} & {0, len(inp) - 1} or {x} & {0, len(inp[y]) - 1}:
                continue
            else:
                if inp[y][x] == alive:
                    if 2 <= count <= 3:
                        temp += alive
                    else:
                        temp += dead
                else:
                    if count == 3:
                        temp += alive
                    else:
                        temp += dead
        if y == 0:
            continue
        nxt.append(temp)

    for i in nxt:
        print(i)
    return nxt


nxt = nextgen(inp)
while True:
    nxt = nextgen(nxt)

#

what the fuck is wrong with this

#

getting this error Traceback (most recent call last): File "c:\Users\\Documents\Code\Python\lifemulitplegen.py", line 85, in <module> nxt = nextgen(nxt) File "c:\Users\\Documents\Code\Python\lifemulitplegen.py", line 59, in nextgen tmp.append(inp[y + k][x + l]) IndexError: string index out of range

eager ore Jan 25, 2023, 6:01 AM

#

you should write a clean version first

#

then golf it

long fulcrum Jan 25, 2023, 6:02 AM

#

I did but ive lost it

eager ore Jan 25, 2023, 6:04 AM

#

you could use git

#

in the future

gleaming linden Jan 25, 2023, 6:40 AM

#

long fulcrum ```py import time import os dead = "⬜" alive = "⬛" os.system("cls") inp = [ ...

just use the set + complex number approach if you're gonna run multiple frames

#

or at least numpy arrays

long fulcrum Jan 25, 2023, 6:52 AM

#

i have no idea what any of those are

#

id just like to get my own version working

gleaming linden Jan 25, 2023, 6:55 AM

#

long fulcrum ```py import time import os dead = "⬜" alive = "⬛" os.system("cls") inp = [ ...

again, just use range instead of enumerate and get rid of the edges there

#

instead of dealing with edge checks

#

but this isn't esoteric anymore

long fulcrum Jan 25, 2023, 6:56 AM

#

ive got this now:

import time
import os

# os.system("cls")
inp = [
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬛⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬛⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬛⬛⬛⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
    "⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜⬜",
]


def nextgen(inp):
    nxt = []
    alive = "⬛"
    dead = "⬜"
    for n, m in enumerate(inp):
        inp[n] = inp[n].center(len(inp[n]) + 2, dead)
    inp.insert(0, dead * len(inp[0]))
    inp.append(dead * len(inp[0]))
    for y, i in enumerate(inp):
        temp = ""
        for x, j in enumerate(i):
            count = 0
            tmp=[]
            if y == 0 or y == len(inp)-1 or x == 0 or x == len(inp[y])-1: continue
            for n in range(-1,2):
                for m in range(-1,2):
                    if n==0 and m==0:
                        continue
                    tmp.append(inp[y+n][x+m])
            count = tmp.count(alive)

            if inp[y][x] == alive:
                if 2 <= count <= 3:
                    temp += alive
                else:
                    temp += dead
            else:
                if count == 3:
                    temp += alive
                else:
                    temp += dead
            if y == 0 or y == len(inp)-1 or x == 0 or x == len(inp[y])-1: continue                   
        nxt.append(temp)
    nxt=nxt[1:-1]
    for n in nxt:
        print(n)
    print()
    return nxt


nxt = nextgen(inp)
while True:
    nxt=nextgen(inp)

#

all working except its growing wider every cycle

long fulcrum Jan 25, 2023, 6:56 AM

#

gleaming linden but this isn't esoteric anymore

oh yeah true

quartz wave Jan 25, 2023, 10:28 AM

#

gleaming linden ```py import sys;R=range;Q=R(-1,2);G,E='#.';A=[E+i+E for i in sys.argv[-1].split...

is there a test input for this

#

here's an untested 235 ```py
import sys
R=range
Q=-1,0,1
G,E='#.'
A=[E+i+E for i in sys.argv[-1].split('\n')]
W=len(A[0])
A,=EW,A,EW
for Y in R(1,len(A)-1):print(''.join([E,G][2-(A[Y][X]==G)<sum(G==A[Y+y][X+x]for y in Q for x in Q if y|x)<4]for X in R(1,W-1)))

sick hound Jan 25, 2023, 10:39 AM

#

!e```py
exec(bytes('牰湩⡴漢⁮湯漊⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠੮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漊⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠੮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漊⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠੮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漊⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠੮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漊⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠੮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漠⁮湯漊⁮湯⤢','u16')[2:])

night quarryBOT Jan 25, 2023, 10:39 AM

#

@sick hound :x: Your 3.11 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 1, in <module>
003 |   File "<string>", line 1
004 |     print("on on
005 |           ^
006 | SyntaxError: unterminated string literal (detected at line 1)

gleaming linden Jan 25, 2023, 10:45 AM

#

@quartz wave test input

quartz wave Jan 25, 2023, 10:48 AM

#

233 ```py
import sys
R=range
Q=-1,0,1
G,E='#.'
A=[E+i+E for i in sys.argv[-1].split('\n')]
W=len(A[0])
A=E*W,A,EW
for Y in R(1,len(A)-1):print(''.join([E,G][2-(A[Y][X]==G)<sum(G==A[Y+y][X+x]for y in Q for x in Q if y|x)<4]for X in R(1,W-1)))

#

231 ```py
import sys
R=range
Q=-1,0,1
E,G=S='.#'
A=[E+i+E for i in sys.argv[-1].split('\n')]
W=len(A[0])
A=E*W,A,EW
for Y in R(1,len(A)-1):print(''.join(S[2-(A[Y][X]==G)<sum(G==A[Y+y][X+x]for y in Q for x in Q if y|x)<4]for X in R(1,W-1)))

dense nova Jan 25, 2023, 11:12 AM

#

what does this do

#

lmao

#

is it a really complicated adder

quartz wave Jan 25, 2023, 11:17 AM

#

dense nova is it a really complicated adder

idk

dense nova Jan 25, 2023, 11:21 AM

#

btw

#

i see room for an improvement of 1 character

#

import sys
R=range
Q=-1,0,1
E,G=S='.#'
A=[i+E*2for i in sys.argv[-1].split('\n')] # changing E+i+E to i+E*2 so the for keyword can 'stick' to the 2
W=len(A[0])
A=E*W,*A,E*W
for Y in R(1,len(A)-1):print(''.join(S[2-(A[Y][X]==G)<sum(G==A[Y+y][X+x]for y in Q for x in Q if y|x)<4]for X in R(1,W-1)))

quartz wave Jan 25, 2023, 12:02 PM

#

dense nova i see room for an improvement of 1 character

and you're sure that doesn't change anything?

dense nova Jan 25, 2023, 12:03 PM

#

dont think so

quartz wave Jan 25, 2023, 12:04 PM

#

dense nova dont think so

it did

#

output failed paste-and-compare test

dense nova Jan 25, 2023, 12:04 PM

#

!e

for i in range(10000):
  for j in range(10000):
    if j + i + j != i+j*2:
      print(f"oopsie? {i} {j}")
      break

night quarryBOT Jan 25, 2023, 12:04 PM

#

@dense nova :warning: Your 3.11 eval job timed out or ran out of memory.

[No output]

quartz wave Jan 25, 2023, 12:04 PM

#

dense nova !e ```py for i in range(10000): for j in range(10000): if j + i + j != i+j...

you're testing with numbers

dense nova Jan 25, 2023, 12:04 PM

#

oh

quartz wave Jan 25, 2023, 12:04 PM

#

‫we're talking about strings where addition is not associative or commutative

dense nova Jan 25, 2023, 12:04 PM

#

i see

#

did not know strings were allowed lmao

sick hound Jan 25, 2023, 12:16 PM

#

@quartz wave what's the challenge now?

quartz wave Jan 25, 2023, 12:16 PM

#

sick hound <@310263589913100288> what's the challenge now?

wdym

sick hound Jan 25, 2023, 12:17 PM

#

quartz wave 231 ```py import sys R=range Q=-1,0,1 E,G=S='.#' A=[E+i+E for i in sys.argv[-1]....

what's this

quartz wave Jan 25, 2023, 12:17 PM

#

sick hound what's this

golf of golf

#

i literally don't know what the coding problem is

#

i just golf golfed solutions

sick hound Jan 25, 2023, 12:18 PM

#

ill try to golf ypurs

gleaming linden Jan 25, 2023, 12:18 PM

#

dense nova is it a really complicated adder

it's game of life

quartz wave Jan 25, 2023, 12:18 PM

#

oh

sick hound Jan 25, 2023, 12:20 PM

#

exec(bytes('浩潰瑲猠獹刊爽湡敧儊ⴽⰱⰰ਱ⱅ㵇㵓⸧✣䄊嬽⭅⭩⁅潦⁲⁩湩猠獹愮杲孶ㄭ⹝灳楬⡴尧❮崩圊氽湥䄨せ⥝䄊䔽圪⨬ⱁ⩅੗潦⁲⁙湩删ㄨ氬湥䄨⴩⤱瀺楲瑮:sparkles:⸧潪湩匨㉛⠭孁嵙塛㵝䜽㰩畳⡭㵇䄽奛礫孝⭘嵸潦⁲⁹湩儠映牯砠椠⁮⁑晩礠硼㰩崴潦⁲⁘湩删ㄨ圬ㄭ⤩ ','u16')[2:])```

dense nova Jan 25, 2023, 12:21 PM

#

why

sick hound Jan 25, 2023, 12:24 PM

#

cos it's shorter

quartz wave Jan 25, 2023, 12:24 PM

#

sick hound ```py exec(bytes('浩潰瑲猠獹刊爽湡敧儊ⴽⰱⰰ਱ⱅ㵇㵓⸧✣䄊嬽⭅⭩⁅潦⁲⁩湩猠獹愮杲孶ㄭ⹝灳楬⡴尧❮崩圊氽湥䄨せ⥝䄊䔽圪⨬ⱁ⩅੗潦⁲⁙湩删ㄨ氬湥...

degen AKA packed solutions are separate

#

they always derive from the latest golfed code

versed eagle Jan 25, 2023, 12:31 PM

#

quartz wave i literally don't know what the coding problem is

its game of life

quartz wave Jan 25, 2023, 12:32 PM

#

versed eagle its game of life

13 minutes too late

versed eagle Jan 25, 2023, 12:34 PM

#

oh

#

sorry

quartz wave Jan 25, 2023, 12:38 PM

#

215 ```py
import sys
R=range
Q=-1,0,1
E,G=S='.#'
A=sys.argv[-1].split('\n')
L,W=len(A),len(A[0])
for Y in R(L):print(''.join(S[2-(A[Y][X]==G)<sum(G==A[Y+y][X+x]for y in Q for x in Q if y|x if W>X+x>-1<Y+y<L)<4]for X in R(W)))

#

210 ```py
import sys
R=range
Q=-1,0,1
L=len(A:=sys.argv[-1].split('\n'))
W=len(A[0])
for Y in R(L):print(''.join('.#'[2-(A[Y][X]=='#')<sum('#'==A[Y+y][X+x]for y in Q for x in Q if y|x if W>X+x>-1<Y+y<L)<4]for X in R(W)))

#

208 ```py
import sys
R=range
Q=-1,0,1
L=len(A:=sys.argv[-1].split('\n'))
W=len(A[0])
for Y in R(L):print(''.join('.#'[2-(A[Y][X]<'.')<sum('.'>A[Y+y][X+x]for y in Q for x in Q if y|x if W>X+x>-1<Y+y<L)<4]for X in R(W)))

gleaming linden Jan 25, 2023, 1:45 PM

#

200 ```py
import sys
R=range
L=len(A:=sys.argv[-1].split('\n'))
W=len(A[0])
for Y in R(L):print(''.join('.#'[2<sum('.'>A[y][x]for y in R(Y-1,Y+2)for x in R(X-1,X+2)if W>x>-1<y<L)<4+(A[Y][X]<'.')]for X in R(W)))

#

193 ```py
import sys
D=-1,0,1
E=enumerate
A=sys.argv[-1].split('\n')
for Y,l in E(A):print(''.join('.#'[4+('.'>c)>sum('.'>A[Y+y][X+x]for y in D for x in D if len(l)>X+x>-1<Y+y<len(A))>2]for X,c in E(l)))

#

191 ```py
import sys
D=-1,0,1
E=enumerate
A=sys.argv[-1].split('\n')
for Y,l in E(A):print(''.join('.#'[4+('.'>c)>sum(len(l)>X+x>-1<Y+y<len(A)!='.'>A[Y+y][X+x]for y in D for x in D)>2]for X,c in E(l)))

finite blaze Jan 25, 2023, 3:03 PM

#

D=-1,0,1

#

What's going on here?

long fulcrum Jan 25, 2023, 3:11 PM

#

finite blaze What's going on here?

basically this but golfed

for n in range(-1, 2):
  for m in range(-1, 2):```

finite blaze Jan 25, 2023, 4:11 PM

#

oh interesting

#

cool stuff

gleaming linden Jan 25, 2023, 4:11 PM

#

finite blaze `D=-1,0,1`

It's used to get the surrounding cells (where x and y coords are -1/0/1 units away)

finite blaze Jan 25, 2023, 4:12 PM

#

i didnt know you can declare ranges like that

#

i need to remember it

gleaming linden Jan 25, 2023, 4:12 PM

#

It's not a range btw

#

It's just a tuple containing -1, 0 and 1

#

Which works fine because we're just looping over it

#

It's only shorter than range(-1,2) because it only has 3 items

arctic skiff Jan 25, 2023, 4:21 PM

#

7800 chars

placid fractal Jan 25, 2023, 8:06 PM

#

does anyone know a good way to basically do

if len(value) == 1:
  x = y = value
else:
  x, y* = value

but in a single liner? reason im asking here is because itll probably become code butchering lmao

versed eagle Jan 25, 2023, 8:16 PM

#

len(value)==1and(x:=(y:=value))or(x:=value[0],y:=value[1:])

#

that should do it

#

but I'm tired so it might also fail entirely

rugged sparrow Jan 25, 2023, 8:49 PM

#

placid fractal does anyone know a good way to basically do ```py if len(value) == 1: x = y = ...

!e ```py
value = [1,2,3]
x, y = (value, value) if len(value) == 1 else (value[0], value[1:])
print(x, y)

value = [1]
x, y = (value, value) if len(value) == 1 else (value[0], value[1:])
print(x, y)

night quarryBOT Jan 25, 2023, 8:49 PM

#

@rugged sparrow :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | 1 [2, 3]
002 | [1] [1]

rugged sparrow Jan 25, 2023, 8:50 PM

#

like that?

placid fractal Jan 25, 2023, 8:50 PM

#

i mean thats the 1 liner version of it

#

i was kindof hoping for a way to do it with unpacking

#

like without using an if else statement ig

twin reef Jan 25, 2023, 8:52 PM

#

Not technically a one-liner, and not quite equivalent, but the shortest so far

x, *y = value
y = y or x

placid fractal Jan 25, 2023, 8:52 PM

#

value is an int or a tuple of ints

#

so it would be

x, *y = tuple(value)
y = y or x

rugged sparrow Jan 25, 2023, 8:53 PM

#

!e py tuple(1) this raises an exception

night quarryBOT Jan 25, 2023, 8:53 PM

#

@rugged sparrow :x: Your 3.11 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 1, in <module>
003 | TypeError: 'int' object is not iterable

rugged sparrow Jan 25, 2023, 8:53 PM

#

placid fractal so it would be ```py x, *y = tuple(value) y = y or x ```

so this would break if value is an int

placid fractal Jan 25, 2023, 8:53 PM

#

rugged sparrow so this would break if `value` is an int

right

#

i just realized that

twin reef Jan 25, 2023, 8:54 PM

#

(*value, )?

#

or even [*value]

rugged sparrow Jan 25, 2023, 8:54 PM

#

twin reef `(*value, )`?

that would also fail since value is not iterable

twin reef Jan 25, 2023, 8:54 PM

#

oh right...

placid fractal Jan 25, 2023, 8:55 PM

#

twin reef or even `[*value]`

the unpacking would also fail

#

sry its such a random question

earnest wing Jan 25, 2023, 8:58 PM

#

wait, that doesn't work either

#

can't unpack at any point

placid fractal Jan 25, 2023, 8:58 PM

#

that fails cuz *value wont work lol

earnest wing Jan 25, 2023, 8:58 PM

#

wait, then why is the initial prompt "if len(value) == 1, how can I ..."

placid fractal Jan 25, 2023, 8:59 PM

#

thats my fault

#

i like blanked and did that lmao

#

ig it should be len(list(val))

earnest wing Jan 25, 2023, 8:59 PM

#

also doesn't work

placid fractal Jan 25, 2023, 9:00 PM

#

so basically then id have to do
isinstance(val, int)

twin reef Jan 25, 2023, 9:01 PM

#

or type(val)==int

placid fractal Jan 25, 2023, 9:01 PM

#

right

placid fractal Jan 25, 2023, 9:37 PM

#

this is the best I was able to figure out

x, *y = val if type(val) is tuple else [val]
y = val if type(val) is list else val or x

tysm for yalls help tho

echo viper Jan 25, 2023, 10:23 PM

#

I have written a python code that by using openAI API, automatically generates texts, images and videos, and generate youtube videos and publish them. Would appreciate if you could take a look and give me your feedback. my channel id on youtube is UCDUG32YQkF5jKKCRs6qz2nw

versed eagle Jan 25, 2023, 10:24 PM

#

echo viper I have written a python code that by using openAI API, automatically generates t...

wrong channel

echo viper Jan 25, 2023, 10:28 PM

#

versed eagle wrong channel

which channel should I use

echo viper Jan 25, 2023, 10:28 PM

#

versed eagle wrong channel

I am new

versed eagle Jan 25, 2023, 10:28 PM

#

!ot

night quarryBOT Jan 25, 2023, 10:28 PM

#

Off-topic channel: #ot2-never-nester’s-nightmare

Please read our off-topic etiquette before participating in conversations.

versed eagle Jan 25, 2023, 10:28 PM

#

or #python-discussion

echo viper Jan 25, 2023, 10:28 PM

#

versed eagle or <#267624335836053506>

Ok, thanks

long fulcrum Jan 25, 2023, 11:05 PM

#

any possible way to golf this further?

def s(m,n,t=0):
    if n>0:
        yield list(range(t,m+t))
        for w in zip(*list(s(n-1,m,m+t))[::-1]):yield w
for i in s(10,10):print(*(f'{i:>2}'for i in i))```

versed eagle Jan 25, 2023, 11:06 PM

#

def s(m,n,t=0):
 if n>0:
  yield[*range(t,m+t)]
  yield from zip(*[*s(n-1,m,m+t)][::-1])
for i in s(10,10):print(*(f'{i:>2}'for i in i))

#

what is it supposed to do

#

could probably golf it more if i knew what its meant to be doing

long fulcrum Jan 25, 2023, 11:08 PM

#

prints a spiral from 0-99

long fulcrum Jan 25, 2023, 11:09 PM

#

versed eagle ```py def s(m,n,t=0): if n>0: yield[*range(t,m+t)] yield from zip(*[*s(n-1,...

doesnt the function have to have a name?

#

that errors out

#

!e

def s(m,n,t=0):
    if n>0:
        yield list(range(t,m+t))
        for w in zip(*list(s(n-1,m,m+t))[::-1]):yield w
for i in s(10,10):print(*(f'{i:>2}'for i in i))```

night quarryBOT Jan 25, 2023, 11:09 PM

#

@long fulcrum :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 |  0  1  2  3  4  5  6  7  8  9
002 | 35 36 37 38 39 40 41 42 43 10
003 | 34 63 64 65 66 67 68 69 44 11
004 | 33 62 83 84 85 86 87 70 45 12
005 | 32 61 82 95 96 97 88 71 46 13
006 | 31 60 81 94 99 98 89 72 47 14
007 | 30 59 80 93 92 91 90 73 48 15
008 | 29 58 79 78 77 76 75 74 49 16
009 | 28 57 56 55 54 53 52 51 50 17
010 | 27 26 25 24 23 22 21 20 19 18

versed eagle Jan 25, 2023, 11:10 PM

#

long fulcrum doesnt the function have to have a name?

yeah
im just tired lol

#

should be fixed now

long fulcrum Jan 25, 2023, 11:10 PM

#

somethings wrong with the brackets

versed eagle Jan 25, 2023, 11:11 PM

#

fixed

long fulcrum Jan 25, 2023, 11:12 PM

#

needs another bracket after the m+t in 3rd line as well

#

oh lines 2 and 3 can be combined now

versed eagle Jan 25, 2023, 11:12 PM

#

fixed

long fulcrum Jan 25, 2023, 11:12 PM

#

oh 2-4 actually

versed eagle Jan 25, 2023, 11:13 PM

#

does it need to be a list?

#

why do you call list

long fulcrum Jan 25, 2023, 11:13 PM

#

night quarry <@236448694172516353> :white_check_mark: Your 3.11 eval job has completed with r...

it can be whatever as long as the output is this

long fulcrum Jan 25, 2023, 11:14 PM

#

versed eagle why do you call list

honestly cant even remember I havent looked at this in ages

versed eagle Jan 25, 2023, 11:14 PM

#

ah yeah it does

long fulcrum Jan 25, 2023, 11:17 PM

#

def s(m,n,t=0):
    if n>0:yield[*range(t,m+t)];yield from zip(*[*s(n-1,m,m+t)][::-1])
for i in s(10,10):print(*(f"{i:>2}"for i in i))```

glass stone Jan 26, 2023, 2:29 AM

#

!paste

#

I wrote this a little while ago, does it count as esoteric? https://paste.pythondiscord.com/cadixuwiji

#

it is not windows friendly if you run it

fluid canyon Jan 26, 2023, 2:32 AM

#

hey i was wondering if someone would like to help me with a simple coding problem, im new to coding and the discord and i wanted to try out the advice someone might give me.

versed eagle Jan 26, 2023, 2:33 AM

#

fluid canyon hey i was wondering if someone would like to help me with a simple coding proble...

this is the wrong channel. check out #❓｜how-to-get-help for information on how to receive help

fluid canyon Jan 26, 2023, 2:33 AM

#

thanks

versed eagle Jan 26, 2023, 2:33 AM

#

in this channel, the type of help you will receive is not the type of help you want

glass stone Jan 26, 2023, 2:33 AM

#

lol. come back later

versed eagle Jan 26, 2023, 2:34 AM

#

glass stone I wrote this a little while ago, does it count as esoteric? https://paste.python...

from time import sleep as sleep

#

why

fluid canyon Jan 26, 2023, 2:34 AM

#

herald your a man you solved my problem

versed eagle Jan 26, 2023, 2:35 AM

#

im not a man, but im glad i could help regardless.

glass stone Jan 26, 2023, 2:35 AM

#

it's esoteric, you probably wou;ldn't understand

#

ya I will help you out in the help channel if you @ me

fluid canyon Jan 26, 2023, 2:35 AM

#

no im good now but thanks

placid fractal Jan 26, 2023, 2:35 AM

#

glass stone I wrote this a little while ago, does it count as esoteric? https://paste.python...

i like

versed eagle Jan 26, 2023, 2:36 AM

#

versed eagle `from time import sleep as sleep`

from time import sleep does the same thing, unless you're using some weird unicode characters that look the same as normal ascii ones

glass stone Jan 26, 2023, 2:36 AM

#

😄 Awesome! It took long time to make

placid fractal Jan 26, 2023, 2:36 AM

#

use the greek version of ;

glass stone Jan 26, 2023, 2:36 AM

#

Herald I know that now, like I said it was written quite some time ago

versed eagle Jan 26, 2023, 2:36 AM

#

ah

#

i see

#

i remember when i first found out about from thing import whatever statements

placid fractal Jan 26, 2023, 2:37 AM

#

is it possible to do something along the lines of
(this is in C)

define True 0

define False 1

glass stone Jan 26, 2023, 2:37 AM

#

I thought I was on to something with that too.

versed eagle Jan 26, 2023, 2:37 AM

#

versed eagle i remember when i first found out about `from thing import whatever` statements

i used them instead of normal import for everything

low lynx Jan 26, 2023, 2:37 AM

#

@placid fractal

versed eagle Jan 26, 2023, 2:38 AM

#

placid fractal is it possible to do something along the lines of (this is in C) # define True ...

in C, yes

glass stone Jan 26, 2023, 2:38 AM

#

from this import *

placid fractal Jan 26, 2023, 2:38 AM

#

yeah that ^

placid fractal Jan 26, 2023, 2:38 AM

#

low lynx <@640460002066300938>

YOO tysm

#

i didnt see that lmao

low lynx Jan 26, 2023, 2:38 AM

#

thank ionite

placid fractal Jan 26, 2023, 2:38 AM

#

now we can make cursed code even more cursed

versed eagle Jan 26, 2023, 2:39 AM

#

ah

versed eagle Jan 26, 2023, 2:39 AM

#

placid fractal is it possible to do something along the lines of (this is in C) # define True ...

i thought you meant macros lol

#

(which also can be done)

placid fractal Jan 26, 2023, 2:40 AM

#

versed eagle i thought you meant macros lol

nono sry. i meant like redefining the actual True and False bools

versed eagle Jan 26, 2023, 2:40 AM

#

ah

placid fractal Jan 26, 2023, 2:40 AM

#

macros are much easier to make no?

versed eagle Jan 26, 2023, 2:40 AM

#

not really

#

ionite did the bool thing in less than 10 lines

placid fractal Jan 26, 2023, 2:41 AM

#

thats just insanely awesome

versed eagle Jan 26, 2023, 2:41 AM

#

macros are relatively much harder to make (in python)

placid fractal Jan 26, 2023, 2:41 AM

#

macropy maybe?

#

ah

versed eagle Jan 26, 2023, 2:41 AM

#

AnonymousDapper has a macro project going iirc

#

and RocketRace has a project that does something similar as well

#

Cereal has a macro thingy also

placid fractal Jan 26, 2023, 2:45 AM

#

interesting ok

versed eagle Jan 26, 2023, 2:46 AM

#

the main thing with their macros though is that they don't use valid Python syntax, they use codecs and whatnot
it'd probably be easier to make a pretty basic macro system using annotations

placid fractal Jan 26, 2023, 2:47 AM

#

I see

#

that makes sense tbh

#

basically a struct of macros really

versed eagle Jan 26, 2023, 2:49 AM

#

yeah

#

well

#

sort of

#

really it depends how in depth you're willing to go

placid fractal Jan 26, 2023, 2:56 AM

#

makes sense, tys tho

dry mirage Jan 26, 2023, 3:21 AM

#

placid fractal is it possible to do something along the lines of (this is in C) # define True ...

!e 👀

from einspect import view

view(False) << 1
view(True) << 0

print(True, False)

night quarryBOT Jan 26, 2023, 3:21 AM

#

@dry mirage :white_check_mark: Your 3.11 eval job has completed with return code 0.

0 1

placid fractal Jan 26, 2023, 3:22 AM

#

o dayum ok

dry mirage Jan 26, 2023, 3:23 AM

#

!e or even

from einspect import view

view(False) << ("hi",)

for x in False:
    print(x)

night quarryBOT Jan 26, 2023, 3:23 AM

#

@dry mirage :white_check_mark: Your 3.11 eval job has completed with return code 0.

hi

placid fractal Jan 26, 2023, 3:31 AM

#

interesting

#

i feel like a noob rn, but ive never heard of einspect

versed eagle Jan 26, 2023, 3:33 AM

#

its a really cool module made by ionite

placid fractal Jan 26, 2023, 3:33 AM

#

i finally found the documentation

#

https://docs.ionite.io/einspect/

#

i love this

#

thank you ionite

low lynx Jan 26, 2023, 4:11 AM

#

dry mirage !e 👀 ```py from einspect import view view(False) << 1 view(True) << 0 print(T...

i thought that would error because it was unsafe

quartz wave Jan 26, 2023, 4:26 AM

#

long fulcrum ```py def s(m,n,t=0): if n>0:yield[*range(t,m+t)];yield from zip(*[*s(n-1,m,...

117 ```py
s=lambda m,n,t:n>0and[[*range(t,m+t)],*zip(s(n-1,m,m+t)[::-1])]or[]
for i in s(10,10,0):print(('%2d'%i for i in i))

low lynx Jan 26, 2023, 5:15 AM

#

116

s=lambda m,n,t:n and[[*range(t,m+t)],*zip(*s(n-1,m,m+t)[::-1])]or[]
for i in s(10,10,0):print(*('%2d'%i for i in i))

sonic birch Jan 26, 2023, 5:23 AM

#

Hi guys, am I allowed to put a link to my github repo here if the repo is related to this channel topic? Or such 'self-promoting' is not allowed here? Thanks

versed eagle Jan 26, 2023, 5:23 AM

#

if its related to the esoteric python then sure

sonic birch Jan 26, 2023, 5:33 AM

#

It depends what is 'esoteric python'. In my repo I'm trying to abuse different Python features for diff purposes. For example here https://github.com/LubomyrIvanitskiy/CrazyPython/tree/main/typespace I tried to implement simething like Dependent Types using function names as annotations. Or here https://github.com/LubomyrIvanitskiy/CrazyPython/tree/main/alt_imports I wrote a code with which your function can be 'located' in multiple modules at the same time. In general this community looks like the right place and would like to here some criticism/questions from you

UDP: here is a link for the whole repo: https://github.com/LubomyrIvanitskiy/CrazyPython

GitHub

CrazyPython/typespace at main · LubomyrIvanitskiy/CrazyPython

Making crazy and, in general, useless stuff by abusing Python features - CrazyPython/typespace at main · LubomyrIvanitskiy/CrazyPython

GitHub

CrazyPython/alt_imports at main · LubomyrIvanitskiy/CrazyPython

Making crazy and, in general, useless stuff by abusing Python features - CrazyPython/alt_imports at main · LubomyrIvanitskiy/CrazyPython

GitHub

GitHub - LubomyrIvanitskiy/CrazyPython: Making crazy and, in genera...

Making crazy and, in general, useless stuff by abusing Python features - GitHub - LubomyrIvanitskiy/CrazyPython: Making crazy and, in general, useless stuff by abusing Python features

versed eagle Jan 26, 2023, 5:36 AM

#

cool idea with the functions in different modules thing

#

though i think they would have to get imported somewhere beforehand so the @tag will run

sonic birch Jan 26, 2023, 5:43 AM

#

versed eagle though i think they would have to get imported somewhere beforehand so the `@tag...

For the first time you launch your app the modules is dynamically creayed and some funxtion stubs are stored by their alt path. So your IDE autofill will work. Next time the program will reuse them from sys.modules

versed eagle Jan 26, 2023, 5:45 AM

#

so do you import the modules beforehand then?

#

to use your example in ReadMe.md
if you try to import tags.math, it would fail, at least until one of the other modules has run and called tag, no?

#

let me rephrase then

#

it would fail unless one of the other modules have been imported, right?

old socket Jan 26, 2023, 5:51 AM

#

Yea the decorator runs @ runtime so how does it work

#

well decorators should at least

versed eagle Jan 26, 2023, 5:52 AM

#

unless you're doing some sort of introspection to see what needs to be imported, but i dont see anything like that in the code so i dont think thats whats going on here

sonic birch Jan 26, 2023, 5:53 AM

#

Good point, let me check

#

You are right. When import tags is called it calles tags.init.py where I execute manifest.py file - this is a file where you should put import for all modules that use tag decorator. I forget to mention it in the readme file, and this is the power of comunnity that you've remind me to do that:) Thank you!

sonic birch Jan 26, 2023, 6:27 AM

#

sonic birch You are right. When import tags is called it calles tags.__init__.py where I exe...

I know it's a bit overhead to import them in single file, but anyway those modules would be imported later. Let me know if you have a better idea

dry mirage Jan 26, 2023, 10:43 AM

#

low lynx i thought that would error because it was unsafe

there is a calculation for if the memory move is safe now, in general it will only be unsafe if one of the following is true

source size exceeds target allocation
source is a non-GC type moved into a GC type
source has an instance dict offset in negative memory and the target doesn't

#

!e

from einspect import view
from einspect.errors import UnsafeError

x = ("hi", "hello")
view(x) << (1, 2, 3)
print(x)

try:
    view(x) << (1, 2, 3, 4)
except UnsafeError as e:
    print(e)

night quarryBOT Jan 26, 2023, 10:49 AM

#

@dry mirage :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | (1, 2, 3)
002 | memory move of 56 bytes into allocated space of 48 bytes is out of bounds. Enter an unsafe context to allow this.

low lynx Jan 26, 2023, 11:21 AM

#

offset in negative memory?

dry mirage Jan 26, 2023, 12:14 PM

#

low lynx offset in negative memory?

sometimes the instance dict is stored before the object in memory, due to the object being variable in size (like a tuple subclass)

low lynx Jan 26, 2023, 12:24 PM

#

ah I see

quartz wave Jan 26, 2023, 12:38 PM

#

dry mirage !e ```py from einspect import view from einspect.errors import UnsafeError x = ...

it aligns .__sizeof__() right?

dry mirage Jan 26, 2023, 1:33 PM

#

quartz wave it aligns `.__sizeof__()` right?

sort of, I calculate memory myself based on the struct

quartz wave Jan 26, 2023, 1:34 PM

#

and it's aligned by 16?

dry mirage Jan 26, 2023, 1:34 PM

#

like list().__sizeof__ includes the array for some reason

dry mirage Jan 26, 2023, 1:35 PM

#

quartz wave and it's aligned by 16?

yeah https://github.com/python/cpython/blob/3.11/Objects/obmalloc.c#L878-L884

night quarryBOT Jan 26, 2023, 1:35 PM

#

Objects/obmalloc.c lines 878 to 884

#if SIZEOF_VOID_P > 4
#define ALIGNMENT              16               /* must be 2^N */
#define ALIGNMENT_SHIFT         4
#else
#define ALIGNMENT               8               /* must be 2^N */
#define ALIGNMENT_SHIFT         3
#endif```

dry mirage Jan 26, 2023, 1:35 PM

#

https://github.com/ionite34/einspect/blob/main/src/einspect/api.py#L47-L52

night quarryBOT Jan 26, 2023, 1:35 PM

#

src/einspect/api.py lines 47 to 52

if sizeof(c_void_p) > 4:
    ALIGNMENT = 16
    ALIGNMENT_SHIFT = 4
else:  # pragma: no cover
    ALIGNMENT = 8
    ALIGNMENT_SHIFT = 3```

quartz wave Jan 26, 2023, 1:36 PM

#

branchless ```py
ALIGNMENT_SHIFT = 3 + (sizeof(c_void_p) > 4)
ALIGNMENT = 1 << ALIGNMENT_SHIFT

dry mirage Jan 26, 2023, 1:37 PM

#

trying to be same as cpython source tho 😔

quartz wave Jan 26, 2023, 1:37 PM

#

fair

dry mirage Jan 26, 2023, 1:48 PM

#

quartz wave fair

!e also the struct creation now calls the proper C-APIs (for tuples _PyObject_GC_NewVar) so the allocation is correct for variable size objects

from einspect import NULL
from einspect.structs import PyTupleObject, PyTypeObject

t = PyTupleObject(
    ob_refcnt=1,
    ob_type=PyTypeObject(tuple).as_ref(),
    ob_size=22,
    ob_item=[1, 2] + [NULL] * 20,
)

print(t.into_object())

night quarryBOT Jan 26, 2023, 1:48 PM

#

@dry mirage :white_check_mark: Your 3.11 eval job has completed with return code 0.

(1, 2, <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NULL>)

sonic birch Jan 26, 2023, 5:37 PM

#

Does anybody know how to change a read-only attribue in Python. For example I have a function foo. I want to change it's globals for test purpose. When I try it like foo.globals = new_globals I got a 'read only attribute' error. Any ideas? Maybe einspect can help?

versed eagle Jan 26, 2023, 5:39 PM

#

does it have to be the same function?

#

cause you could make a new function with a different __globals__

#

and replace foo with the new function

#

but yeah, einspect can most likely do that

#

if you know the offset for the .__globals__ you can replace the ptr with your own ptr

#

to a different dict

sonic birch Jan 26, 2023, 5:44 PM

#

should I copy paste the function body in order to create the new function? Or what do you mean?

versed eagle Jan 26, 2023, 5:45 PM

#

create a new function instance

sonic birch Jan 26, 2023, 5:45 PM

#

Ah, you mean dynamically? Sounds reasonable, will try. Thanks

versed eagle Jan 26, 2023, 5:46 PM

#

https://docs.python.org/3/library/types.html#types.FunctionType

Python documentation

types — Dynamic type creation and names for built-in types

Source code: Lib/types.py This module defines utility functions to assist in dynamic creation of new types. It also defines names for some object types that are used by the standard Python interpre...

rugged sparrow Jan 26, 2023, 6:19 PM

#

sonic birch Ah, you mean dynamically? Sounds reasonable, will try. Thanks

!e ```py
def foo():
return Global_Value

try:
foo()
except Exception as e:
print(e)

foo = type(foo)(foo.code, foo.globals | {'Global_Value':'injected'})
print(foo())```

night quarryBOT Jan 26, 2023, 6:19 PM

#

@rugged sparrow :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | name 'Global_Value' is not defined
002 | injected

rugged sparrow Jan 26, 2023, 6:19 PM

#

^ that example re-builds foo with the new dictionary

#

you can also pass in a dictionary subclass, but that can have partially unpredictable results afaik

sonic birch Jan 26, 2023, 6:21 PM

#

rugged sparrow ^ that example re-builds `foo` with the new dictionary

Seems, you have modified the module global as well

rugged sparrow Jan 26, 2023, 6:21 PM

#

that would not modify the module globals, it only changes what foo thinks the module globals is

sonic birch Jan 26, 2023, 6:22 PM

#

!e Global_Value

night quarryBOT Jan 26, 2023, 6:22 PM

#

@sonic birch :x: Your 3.11 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 1, in <module>
003 | NameError: name 'Global_Value' is not defined

rugged sparrow Jan 26, 2023, 6:22 PM

#

yea, the line where I rebuild foo adds Global_Value to the dictionary

sonic birch Jan 26, 2023, 6:23 PM

#

rugged sparrow yea, the line where I rebuild `foo` adds `Global_Value` to the dictionary

I see. Cool, thanks a lot!

rugged sparrow Jan 26, 2023, 6:27 PM

#

!e ```py
from ctypes import *
def foo():
return Global_Value

func_mem = (c_ssize_t * (foo.sizeof() // tuple.itemsize)).from_address(id(foo))
dict_offset = [*func_mem].index(id(foo.globals))
d = {'Global_Value': 'injected'}
py_object.from_address(id(foo) + (dict_offset * tuple.itemsize)).value = d
print(foo())``` @sonic birch you can also use ctypes to directly swap the dictionary without rebuilding foo but that can be unstable

night quarryBOT Jan 26, 2023, 6:27 PM

#

@rugged sparrow :white_check_mark: Your 3.11 eval job has completed with return code 0.

injected

sonic birch Jan 26, 2023, 8:48 PM

#

rugged sparrow !e ```py from ctypes import * def foo(): return Global_Value func_mem = (c_...

Wow, that one looks really esoteric 😎

dry mirage Jan 26, 2023, 9:31 PM

#

sonic birch Does anybody know how to change a read-only attribue in Python. For example I ha...

well not previously but I've just added FunctionView along with PyFunctionObject https://github.com/ionite34/einspect/blob/dev/src/einspect/views/view_function.py#L16-L17

night quarryBOT Jan 26, 2023, 9:31 PM

#

src/einspect/views/view_function.py lines 16 to 17

class FunctionView(View[FunctionType, None, None]):
    _pyobject: PyFunctionObject```

dry mirage Jan 26, 2023, 9:31 PM

#

so you can do

from einspect import view


def foo():
    return Global_Value

with view(foo).unsafe() as v:
    v.globals = {"Global_Value": 123}

print(foo())
>> 123

versed eagle Jan 26, 2023, 9:33 PM

#

dry mirage well not previously but I've just added `FunctionView` along with `PyFunctionObj...

im curious, do you plan on adding a view for code objects as well?

sonic birch Jan 26, 2023, 9:39 PM

#

dry mirage so you can do ```py from einspect import view def foo(): return Global_Val...

Amazing😱

subtle solar Jan 26, 2023, 10:04 PM

#

!e i propose to you, my newfound method of checking whether a container is non-empty 😎 ```py
@type.call
class anything:
def eq(self, __o: object) -> bool:
return True

x = [5, 6, 7, 8]
print(f"{anything in x = }")
y = []
print(f"{anything in y = }")

night quarryBOT Jan 26, 2023, 10:04 PM

#

@subtle solar :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | anything in x = True
002 | anything in y = False

dry mirage Jan 26, 2023, 10:13 PM

#

sonic birch Amazing😱

in https://github.com/ionite34/einspect/releases/tag/v0.5.3

languid hare Jan 26, 2023, 10:17 PM

#

subtle solar !e i propose to you, my newfound method of checking whether a container is non-e...

brilliant

versed eagle Jan 26, 2023, 10:18 PM

#

versed eagle im curious, do you plan on adding a view for code objects as well?

@dry mirage sorry for the ping but you didn't respond so i don't think you saw my question

dry mirage Jan 26, 2023, 10:18 PM

#

versed eagle <@233059161401720832> sorry for the ping but you didn't respond so i don't think...

oh, yeah might do that next

versed eagle Jan 26, 2023, 10:19 PM

#

alr
was asking cause if you weren't planning on doing it, i was gonna offer to do it

dry mirage Jan 26, 2023, 10:22 PM

#

versed eagle alr was asking cause if you weren't planning on doing it, i was gonna offer to d...

you can have a try if you want :p feel free to PR or anything

versed eagle Jan 26, 2023, 10:22 PM

#

alr! will do that

dry mirage Jan 26, 2023, 10:22 PM

#

https://github.com/python/cpython/blob/3.11/Include/cpython/code.h#L36-L100

#

just need to define this as a new structs/py_code.py or something

versed eagle Jan 26, 2023, 10:24 PM

#

I've gotten the offsets for most of the attributes already
was mapping it out anyway for a different project of mine so i have most of the information already

#

it's just a matter of formatting it and making it like the other ones

dry mirage Jan 26, 2023, 10:25 PM

#

not sure if you're aware already but with my @struct decorator you can define fields using type hints

#

if you need to define a simple type like c_short you can use Annotated, the decorator will ignore whatever the first argument is, and take the second one as the field type

#

so like with this, you'll get the correct python type of int, and the decorator will use the c_short type to generate the correct field

_co_linearray_entry_size: Annotated[int, c_short]

versed eagle Jan 26, 2023, 10:27 PM

#

ah
that's nice

dry mirage Jan 26, 2023, 10:28 PM

#

if you need to do bit widths it's a third argument, so

_fields_ = [
  ("var", c_uint, 3)
]

would be

var: Annotated[int, c_uint, 3]

sonic birch Jan 27, 2023, 3:56 AM

#

subtle solar !e i propose to you, my newfound method of checking whether a container is non-e...

What does the @type.call do and why is it necessary here?

gleaming linden Jan 27, 2023, 3:56 AM

#

that's a decotator

#

@some_decotator
def foo(x):
    ...

# is roughly equivalent to
def foo(x):
    ...
foo = some_decorator(foo)

#

and the same for classes

sonic birch Jan 27, 2023, 4:10 AM

#

gleaming linden that's a decotator

I know. The question was why is it needed here? I see that without this decorator you will need to create an instance of anything class like anything() in x in order to make it work. But I don't get why it work

old socket Jan 27, 2023, 4:16 AM

#

sonic birch I know. The question was why is it needed here? I see that without this decorato...

type.__call__ decorator makes it so that anything is already an instance

#

!e ```py
@type.call
class foo:
...

print(foo)

night quarryBOT Jan 27, 2023, 4:16 AM

#

@old socket :white_check_mark: Your 3.11 eval job has completed with return code 0.

<__main__.foo object at 0x7f078da5ffd0>

old socket Jan 27, 2023, 4:16 AM

#

You can imagine it like this as well

#

!e ```py
@lambda _: _()
class foo:
...

print(foo)

night quarryBOT Jan 27, 2023, 4:16 AM

#

@old socket :white_check_mark: Your 3.11 eval job has completed with return code 0.

<__main__.foo object at 0x7f7e75bf4190>

sonic birch Jan 27, 2023, 4:23 AM

#

old socket !e ```py @lambda _: _() class foo: ... print(foo) ```

Never saw this tweak before. Thanks it've helped. BTW I asked ChatGPT this question, and ChatGPT wasn't able to explain it to me. So +1 to humanity 🙂

rugged owl Jan 27, 2023, 4:25 AM

#

sonic birch Never saw this tweak before. Thanks it've helped. BTW I asked ChatGPT this quest...

Wasnt able to answer because its not a common practice which is one of its flaws

#

dumb_ishrion

sonic birch Jan 27, 2023, 4:26 AM

#

rugged owl Wasnt able to answer because its not a common practice which is one of its flaws

Exactly. ChatGPT is not esoteric at all It is a pop-star 🙂

old socket Jan 27, 2023, 4:30 AM

#

ChaptGPT to prototype fast then bug fix afterwards for 8 hrs

rugged owl Jan 27, 2023, 4:32 AM

#

import time
g=lambda:int(time.time()*1e3)<<22|1<<12

anyway to shorten this more

#

its a basic snowflake id generation if that helps

versed eagle Jan 27, 2023, 4:32 AM

#

wdym "snowflake id generation"

gleaming linden Jan 27, 2023, 4:32 AM

#

^

versed eagle Jan 27, 2023, 4:33 AM

#

if you're looking for a number that won't be repeated, you can just use time.time()

#

since you're doing more than that, i assume you want the number to have some specific properties?

gleaming linden Jan 27, 2023, 4:39 AM

#

import time
g=lambda:int(time.time()*1e3)<<22|8<<9
```-1c

sonic birch Jan 27, 2023, 5:08 AM

#

dry mirage in https://github.com/ionite34/einspect/releases/tag/v0.5.3

!e BTW, you approach works if I try to assign a new globals to the function. If I just add a new item to it - the item is also added to the module's globals. Is it expected?

from einspect import view


def foo():
    a = 12
    print(a)


with view(foo).unsafe() as v:
    v.globals |= {'boo': 4}

print(globals()['boo'])

night quarryBOT Jan 27, 2023, 5:08 AM

#

@sonic birch :x: Your 3.11 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 10, in <module>
003 | AttributeError: 'View' object has no attribute 'globals'

rugged sparrow Jan 27, 2023, 5:10 AM

#

!e ```py
a = {'a': 1}
b = a
b |= {'b':2}
print(a, b)
print(a is b)

night quarryBOT Jan 27, 2023, 5:10 AM

#

@rugged sparrow :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | {'a': 1, 'b': 2} {'a': 1, 'b': 2}
002 | True

rugged sparrow Jan 27, 2023, 5:11 AM

#

sonic birch !e BTW, you approach works if I try to assign a new globals to the function. If ...

thats because |= updates the original dict

#

!e just like += for lists ```py
l1 = [1]
l2 = l1
l2 += [2]
print(l1, l2)

night quarryBOT Jan 27, 2023, 5:11 AM

#

@rugged sparrow :white_check_mark: Your 3.11 eval job has completed with return code 0.

[1, 2] [1, 2]

sonic birch Jan 27, 2023, 5:14 AM

#

I see. It is like operator equivalen for updaye and append. Yes, it is fixed by `v.globals = v.globals | {'boo': 4}. Just wonder isn't it will be confused for other guys

rugged sparrow Jan 27, 2023, 5:16 AM

#

!e its a bit of a quirk of python that the standard for inplace operators is to modify the original. its the source of this odd error: ```py
x = ([1],)
try:x[0] += [2]
except Exception as e:print(e)
print(x)

night quarryBOT Jan 27, 2023, 5:16 AM

#

@rugged sparrow :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | 'tuple' object does not support item assignment
002 | ([1, 2],)

versed eagle Jan 27, 2023, 5:17 AM

#

i dont think thats intended behaviour

rugged sparrow Jan 27, 2023, 5:18 AM

#

it is, because it ends up calling the following: py x[0] = x[0].__iadd__([1]) # or tuple.__setitem__(x, 0, list.__iadd__(x[0], [1]))

versed eagle Jan 27, 2023, 5:18 AM

#

huh

#

i didnt know it'd call setitem

sonic birch Jan 27, 2023, 5:19 AM

#

rugged sparrow !e its a bit of a quirk of python that the standard for inplace operators is to ...

What? How?😱

rugged sparrow Jan 27, 2023, 5:19 AM

#

!e import dis;dis.dis('x[0] += y')

night quarryBOT Jan 27, 2023, 5:20 AM

#

@rugged sparrow :white_check_mark: Your 3.10 eval job has completed with return code 0.

001 |   1           0 LOAD_NAME                0 (x)
002 |               2 LOAD_CONST               0 (0)
003 |               4 DUP_TOP_TWO
004 |               6 BINARY_SUBSCR
005 |               8 LOAD_NAME                1 (y)
006 |              10 INPLACE_ADD
007 |              12 ROT_THREE
008 |              14 STORE_SUBSCR
009 |              16 LOAD_CONST               1 (None)
010 |              18 RETURN_VALUE

rugged sparrow Jan 27, 2023, 5:20 AM

#

3.10 bytecode is more concise, but same concept

#

you can see that it calls the STORE_SUBSCR opcode

versed eagle Jan 27, 2023, 5:20 AM

#

i would think that it would call tuple.__getitem__(([1], ), 0).__iadd__([2])

versed eagle Jan 27, 2023, 5:20 AM

#

rugged sparrow you can see that it calls the `STORE_SUBSCR` opcode

yep

versed eagle Jan 27, 2023, 5:20 AM

#

versed eagle i would think that it would call `tuple.__getitem__(([1], ), 0).__iadd__([2])`

at least, that'd make more sense

rugged sparrow Jan 27, 2023, 5:20 AM

#

not all objects define __iadd__

versed eagle Jan 27, 2023, 5:21 AM

#

hm

#

true

rugged sparrow Jan 27, 2023, 5:21 AM

#

for example, strings

#

!e py x = list(['a']) x[0] += 'b' print(x)

night quarryBOT Jan 27, 2023, 5:21 AM

#

@rugged sparrow :white_check_mark: Your 3.11 eval job has completed with return code 0.

['ab']

rugged sparrow Jan 27, 2023, 5:21 AM

#

(also integers, floats and tuples)

versed eagle Jan 27, 2023, 5:21 AM

#

right

#

but we're talking specifically about lists

#

which do

rugged sparrow Jan 27, 2023, 5:22 AM

#

but the bytecode can't know that its working with a list

versed eagle Jan 27, 2023, 5:22 AM

#

night quarry <@274715613115711488> :white_check_mark: Your 3.10 eval job has completed with r...

10 INPLACE_ADD
that's the bytecode for __iadd__, no?

meager zinc Jan 27, 2023, 5:23 AM

#

yes

rugged sparrow Jan 27, 2023, 5:23 AM

#

INPLACE_ADD either does return stack[0].__iadd__(stack[1]) or if stack[0] doesnt have __iadd__ it does return stack[0] + stack[1]

versed eagle Jan 27, 2023, 5:23 AM

#

alr

#

so it already handles the case where __iadd__ isnt defined

#

on the object

meager zinc Jan 27, 2023, 5:24 AM

#

rugged sparrow `INPLACE_ADD` either does `return stack[0].__iadd__(stack[1])` or if `stack[0]` ...

note that it will also have an assignment call afterwards if it doesn't have __iadd__

rugged sparrow Jan 27, 2023, 5:26 AM

#

i was refering to just the INPLACE_ADD, which eventually calls https://github.com/python/cpython/blob/81f7359f67a7166d57a10a3d5366406d9c85f1de/Objects/abstract.c#L1247-L1267

GitHub

cpython/abstract.c at 81f7359f67a7166d57a10a3d5366406d9c85f1de · py...

The Python programming language. Contribute to python/cpython development by creating an account on GitHub.

rugged sparrow Jan 27, 2023, 5:27 AM

#

meager zinc note that it will also have an assignment call afterwards if it doesn't have `__...

the assignment call happens in the next few bytecodes, not the INPLACE_ADD opcode

meager zinc Jan 27, 2023, 5:28 AM

#

yes, I know

#

also have an assignment call afterwards

rugged sparrow Jan 27, 2023, 5:28 AM

#

the assignment call always happens, regardless of the existence of __iadd__

quartz wave Jan 27, 2023, 5:29 AM

#

rugged sparrow the assignment call happens in the next few bytecodes, not the `INPLACE_ADD` opc...

it's BINARY_OP 13 in 3.11+ right?

rugged sparrow Jan 27, 2023, 5:29 AM

#

yea, but INPLACE_ADD reads better lol

quartz wave Jan 27, 2023, 5:29 AM

#

i wonder why they didn't just separate INPLACE_* from BINARY_*

rugged sparrow Jan 27, 2023, 5:30 AM

#

because it is still a BINARY_OP

rugged owl Jan 27, 2023, 5:37 AM

#

!e

from einspect import view

_ = view("true")
__ = view("false")

with _.unsafe() as un_:
  un_ <<= True

with __.unsafe() as un__:
  un__ <<= False

print("true", "false")

night quarryBOT Jan 27, 2023, 5:37 AM

#

@rugged owl :white_check_mark: Your 3.11 eval job has completed with return code 0.

False False

rugged owl Jan 27, 2023, 5:37 AM

#

What is happening here

quartz wave Jan 27, 2023, 5:45 AM

#

rugged owl What is happening here

it seems like they're 2 different instances of False

#

in the end

#

!e ```py
from einspect import view

_ = view("true")

with .unsafe() as un:
un_ <<= True

print("true")

night quarryBOT Jan 27, 2023, 5:46 AM

#

@quartz wave :white_check_mark: Your 3.11 eval job has completed with return code 0.

False

rugged sparrow Jan 27, 2023, 5:46 AM

#

i am pretty sure that the repr for bool does an identity check, basically py def repr(self): if self is True: return 'True' else: return 'False'

quartz wave Jan 27, 2023, 5:47 AM

#

quartz wave !e ```py from einspect import view _ = view("true") with _.unsafe() as un_: ...

@rugged owl

quartz wave Jan 27, 2023, 5:47 AM

#

rugged sparrow i am pretty sure that the `repr` for `bool` does an identity check, basically ``...

also this

rugged sparrow Jan 27, 2023, 5:47 AM

#

https://github.com/python/cpython/blob/main/Objects/boolobject.c#L9-L14

night quarryBOT Jan 27, 2023, 5:47 AM

#

Objects/boolobject.c lines 9 to 14

static PyObject *
bool_repr(PyObject *self)
{
    PyObject *res = self == Py_True ? &_Py_ID(True) : &_Py_ID(False);
    return Py_NewRef(res);
}```

dry mirage Jan 27, 2023, 10:12 AM

#

rugged owl !e ```py from einspect import view _ = view("true") __ = view("false") with _....

!e yeah you'd have to override bool repr since it defaults to checking identity

from einspect import view

view(bool)["__repr__"] = lambda s: "True" if int(s) else "False"

view("true") << True
view("false") << False

print("true", "false")
print(int("true"), int("false"))
print(type("true"))
print("false" == False)
print("true" == True)

night quarryBOT Jan 27, 2023, 10:12 AM

#

@dry mirage :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | True False
002 | 1 0
003 | <class 'bool'>
004 | True
005 | True

languid hare Jan 27, 2023, 11:49 AM

#

!e

from einspect import view

view(bool)["__repr__"] = lambda s: "True" if int(s) else "False"

view("true") << True
view("false") << False

print("true", "false")
print("tr" + "ue")

night quarryBOT Jan 27, 2023, 11:49 AM

#

@languid hare :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | True False
002 | True

languid hare Jan 27, 2023, 11:49 AM

#

neat

low lynx Jan 27, 2023, 11:50 AM

#

hmm i wonder

#

!e

from einspect import view

view(int)['__add__'] = lambda n, m: 3

with view('1').unsafe() as un1:
    un1 <<= 1

print('1' + '1')

night quarryBOT Jan 27, 2023, 11:57 AM

#

@low lynx :white_check_mark: Your 3.11 eval job has completed with return code 0.

low lynx Jan 27, 2023, 11:57 AM

#

eh

#

!e

from einspect import view

view(str)['__add__'] = lambda n, m: 3

with view('1').unsafe() as un1:
    un1 <<= 1

print('1' + '1')

night quarryBOT Jan 27, 2023, 11:59 AM

#

@low lynx :white_check_mark: Your 3.11 eval job has completed with return code 0.

low lynx Jan 27, 2023, 11:59 AM

#

what

dry mirage Jan 27, 2023, 12:07 PM

#

low lynx !e ```py from einspect import view view(str)['__add__'] = lambda n, m: 3 with ...

'1' + '1' is inlined during compilation

low lynx Jan 27, 2023, 12:09 PM

#

ah

#

ofc

#

is it still inlined?

#

ah needs to be str.__add__

#

wait what

#

2?

#

!e

from einspect import view

view(str)['__add__'] = lambda n, m: 3

with view('1').unsafe() as un1:
    un1 <<= 1

one = '1'
print(one + one)

night quarryBOT Jan 27, 2023, 12:11 PM

#

@low lynx :white_check_mark: Your 3.11 eval job has completed with return code 0.

low lynx Jan 27, 2023, 12:12 PM

#

what

dry mirage Jan 27, 2023, 12:12 PM

#

well it's not str add since you moved it to an int

low lynx Jan 27, 2023, 12:12 PM

#

!e

from einspect import view

view(int)['__add__'] = lambda n, m: 3

with view('1').unsafe() as un1:
    un1 <<= 1

one = '1'
print(one + one)

night quarryBOT Jan 27, 2023, 12:12 PM

#

@low lynx :white_check_mark: Your 3.11 eval job has completed with return code 0.

low lynx Jan 27, 2023, 12:12 PM

#

int add isn't it either

dry mirage Jan 27, 2023, 12:13 PM

#

low lynx int add isn't it either

uh

#

so I think you overriding int add like that broke the offset calculation for the move afterwards

low lynx Jan 27, 2023, 12:14 PM

#

ah

dry mirage Jan 27, 2023, 12:14 PM

#

!e

from einspect import view

view('1') << 1

view(int)['__add__'] = lambda n, m: 3

x = '1'
print(x + '1')

night quarryBOT Jan 27, 2023, 12:14 PM

#

@dry mirage :white_check_mark: Your 3.11 eval job has completed with return code 0.

low lynx Jan 27, 2023, 12:15 PM

#

what

#

ohh so the offset was why i had to do unsafe()

#

that kinda makes sense

dry mirage Jan 27, 2023, 12:15 PM

#

oh lol

#

yeah that's supposed to be safe

#

!e

from einspect import view

view(int)['__add__'] = lambda n, m: 3

try:
    view('1') << 1
except RuntimeError as e:
    print(e)

night quarryBOT Jan 27, 2023, 12:17 PM

#

@dry mirage :white_check_mark: Your 3.11 eval job has completed with return code 0.

memory move of 3 bytes into allocated space of 0 bytes is out of bounds. Enter an unsafe context to allow this.

dry mirage Jan 27, 2023, 12:17 PM

#

3 bytes into allocated space of 0 bytes

yeah 🥴 completely brok

low lynx Jan 27, 2023, 12:17 PM

#

wow something is being stored in 0 bytes?

#

we just broke information theory

dry mirage Jan 27, 2023, 12:20 PM

#

!e there's also a new singleton NULL now if you want to set a NULL PyObject pointer for some reason

from einspect import view, NULL

t = [1, 2, 3]

view(t).item[1] = NULL

print(t)

night quarryBOT Jan 27, 2023, 12:20 PM

#

@dry mirage :white_check_mark: Your 3.11 eval job has completed with return code 0.

[1, <NULL>, 3]

low lynx Jan 27, 2023, 12:48 PM

#

does that make the int 2 into NULL or just the list item?

gleaming linden Jan 27, 2023, 12:53 PM

#

just the item I think

dry mirage Jan 27, 2023, 2:30 PM

#

!e for some reason the built-in reprs support formatting null pointers, but if you try to access the value:

from einspect import view, NULL

t = [1, 2, 3]
view(t).item[1] = NULL

print(t[1])

night quarryBOT Jan 27, 2023, 2:30 PM

#

@dry mirage :warning: Your 3.11 eval job has completed with return code 139 (SIGSEGV).

[No output]

low lynx Jan 27, 2023, 2:32 PM

#

huh

#

wait so why are you using view().item instead of __setitem)__ or []

quartz wave Jan 27, 2023, 2:39 PM

#

!e ```py
from einspect import view, NULL

t = [1, 2, 3]
view(t).item[1] = NULL

print(repr(t[1]))

night quarryBOT Jan 27, 2023, 2:39 PM

#

@quartz wave :warning: Your 3.11 eval job has completed with return code 139 (SIGSEGV).

[No output]

quartz wave Jan 27, 2023, 2:39 PM

#

aw

dry mirage Jan 27, 2023, 2:47 PM

#

low lynx wait so why are you using ``view().item`` instead of ``__setitem)__`` or ``[]``

__setitem__ calls PyObject.from_object on the value

#

I'll need to add some override to recognize NULL

#

!e

from einspect import view, NULL

t = (1, 2, 3)
view(t)[1] = NULL

print(t)
print(t[1])

night quarryBOT Jan 27, 2023, 2:48 PM

#

@dry mirage :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | (1, <NULL ptr[PyObject] at 0x7f426207df00>, 3)
002 | <NULL ptr[PyObject] at 0x7f426207df00>

dry mirage Jan 27, 2023, 2:48 PM

#

otherwise it literally sets that PyObject pointer as an object

low lynx Jan 27, 2023, 2:52 PM

#

ah

cedar isle Jan 27, 2023, 2:56 PM

#

Heh, made this abomination can ya guess what it does?

C = lambda o, b = tuple(), m = MappingProxyType({
        "|" : "__or__",
        "&" : "__and__",
        "^" : "__xor__",
        "<<": "__lshift__",
        ">>": "__rshift__",
        "+" : "__add__",
        "-" : "__sub__",
        "*" : "__mul__",
        "/" : "__truediv__",
        "//": "__floordiv__",
        "%" : "__mod__",
        "**": "__pow__",
        "==": "__eq__",
        "!=": "__ne__",
        "<" : "__lt__",
        "<=": "__le__",
        ">" : "__gt__",
        ">=": "__ge__",
}): type(
        "P",
        (type,),
        f := {m.get(o, o): lambda self, other: self(other)},
)("C", b, f)

the dict should give a clue

night quarryBOT Jan 27, 2023, 3:12 PM

#

Rules

8. Do not help with ongoing exams. When helping with homework, help people learn how to do the assignment without doing it for them.

cedar isle Jan 27, 2023, 3:25 PM

#

Was this directed at me? ^

#

if so, not been a student in a few decades lol

#

this is evil of my own concoction

versed eagle Jan 27, 2023, 3:28 PM

#

my guess is that someone else sent a message

#

and they deleted it after the rule was brought up

cedar isle Jan 27, 2023, 3:31 PM

#

oh lol, derp

#

#

its evil lol

rugged owl Jan 27, 2023, 7:40 PM

#

pithink

cedar isle Jan 27, 2023, 7:41 PM

#

rugged owl <:pithink:652247559909277706>

OMG, that emoji is so cute

sonic birch Jan 27, 2023, 7:59 PM

#

Man, what are you doing? Are you using some kind of functional programming library for Python?

cedar isle Jan 27, 2023, 8:00 PM

#

cedar isle Heh, made this abomination can ya guess what it does? ``` C = lambda o, b = tupl...

Clue here 😛

#

Basically, a functional pipeline tool

#

I have "simplified" it

#

since that last post

#

and by simplified I mean removed the need for a seperate generator class

sonic birch Jan 27, 2023, 8:01 PM

#

Few months ago I tried to implement similar library. Not finished so far)

cedar isle Jan 27, 2023, 8:03 PM

#

Oh, this was just to show someone it can be done lol, its basically combinator logics, which I recently buiot a whole lang parser from, using similar overloads so you can define grammars in python kinda like EBNF

sonic birch Jan 27, 2023, 8:05 PM

#

Any links? Sounds interesting

night quarryBOT Jan 27, 2023, 8:05 PM

#

Hey @cedar isle!

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

cedar isle Jan 27, 2023, 8:06 PM

#

sonic birch Any links? Sounds interesting

Don't have links for it yet, as still a little bit unstable

cedar isle Jan 27, 2023, 8:06 PM

#

sonic birch Any links? Sounds interesting

here is link for todays stuff
https://paste.pythondiscord.com/qiyuqunozu

cedar isle Jan 27, 2023, 8:08 PM

#

sonic birch Any links? Sounds interesting

and here is the logics for my combinator parser
(its still WIP so bear with)
https://paste.pythondiscord.com/yugatupalu

#

and a snippet of how it works

Identifier = Regex(r"[a-zA-Z_][a-zA-Z0-9_]*")
Assignment = Regex(r"=")
Number = Regex(r"0x[0-9a-fA-F]+|[0-9]+(\.[0-9]+)?")
String = Regex(r'"[^"]*"')
Comment = Regex(r";.*")
Semicolon = Regex(r";")

variable = Identifier >> ~Whitespace >> Assignment >> ~Whitespace >> Number >> ~Whitespace >> Semicolon

sonic birch Jan 27, 2023, 8:19 PM

#

Why do you want to have it on Python. Why do not use the pure EBNF? Just for fun?

versed eagle Jan 27, 2023, 8:21 PM

#

cedar isle and a snippet of how it works ``` Identifier = Regex(r"[a-zA-Z_][a-zA-Z0-9_]*") ...

instead of calling Regex, you could make it so that it calls regex when you get the variable

#

it'd be closer to the ebnf syntax then

old socket Jan 27, 2023, 8:22 PM

#

Wasn't there a way to compile multiple lines in together

#

I forgot a lot about regex stuff

#

By lines I mean regular exprs

versed eagle Jan 27, 2023, 8:23 PM

#

you mean where it would have to match all of the ones passed to succeed?

old socket Jan 27, 2023, 8:24 PM

#

No I remember seeing someone compile a large string that had a bunch of regular exprs that match independently

#

But I must've remembered wrong cause that sounds not right

versed eagle Jan 27, 2023, 8:24 PM

#

that's possible

#

regex can match multiple exprs

old socket Jan 27, 2023, 8:25 PM

#

Oh yea possible, just use groups as meta data

#

So when you match, check group name to determine token kind

eager ore Jan 28, 2023, 3:06 AM

#

@old socket

sick hound Jan 28, 2023, 8:38 AM

#

Hello, does anyone know how to create a perceptron in python? help me please

tough willow Jan 28, 2023, 9:45 AM

#

sick hound Hello, does anyone know how to create a perceptron in python? help me please

#❓｜how-to-get-help

old socket Jan 28, 2023, 8:18 PM

#

@sick hound Sorry for ping, but I keep getting SyntaxError encoding problem when messing around codecs. Do you know what I'm doing wrong?

#

Using -*- coding: custom -*- and # coding: custom both yields same result

#

I need to import manually then it'll load the codec

#

*And I know custom_autload.pth is running because I added a print

old socket Jan 28, 2023, 8:53 PM

#

Oh I had to add an incrementaldecoder

sick hound Jan 28, 2023, 9:10 PM

#

no idea honestly, I just tell people to copy manually lol

#

you could try https://github.com/pyston/pyston/blob/de7c5c28d9fcb09b67c29d6c41dcc2427b955f48/pyston/pyston_lite/autoload/setup.py

GitHub

pyston/setup.py at de7c5c28d9fcb09b67c29d6c41dcc2427b955f48 · pysto...

A faster and highly-compatible implementation of the Python programming language. - pyston/setup.py at de7c5c28d9fcb09b67c29d6c41dcc2427b955f48 · pyston/pyston

old socket Jan 28, 2023, 9:10 PM

#

The main reason I wanted a function like this is because with poetry you can set the build script

#

So when building via poetry then uploading to pypi people will be able to skip that whole step

sick hound Jan 28, 2023, 9:11 PM

#

yea I don't think you can automate this with poetry

#

I also use that

#

also a tip I didn't notice until I was already done: beware of bytecode cache

old socket Jan 28, 2023, 9:12 PM

#

Did you do ```toml
[tools.poetry]
build = ...

sick hound Jan 28, 2023, 9:12 PM

#

you will still have old bytecode if you change only the encoder, and not update the script using it and try importing it

old socket Jan 28, 2023, 9:13 PM

#

How do I update the cache?

#

I was planning on taking in the bytes, turning it into a string then lexing & parsing

sick hound Jan 28, 2023, 9:13 PM

#

if you delete __pycache__ it'll get regenerated

old socket Jan 28, 2023, 9:13 PM

#

then after parsing use a visitor to spit out a huge ast.Module

#

then ast.unparse

sick hound Jan 28, 2023, 9:14 PM

#

depends on what you want to do

old socket Jan 28, 2023, 9:14 PM

#

sick hound if you delete `__pycache__` it'll get regenerated

Ohh, I see

sick hound Jan 28, 2023, 9:14 PM

#

https://github.com/dzshn/lispy this only uses tokenize.tokenize

old socket Jan 28, 2023, 9:14 PM

#

So If I made any updates to the compiler it'll use old bytecode unless pycache is deleted

sick hound Jan 28, 2023, 9:14 PM

#

wait no that doesnt use codecs yet

old socket Jan 28, 2023, 9:14 PM

#

sick hound depends on what you want to do

I mainly want to use ast module bc It's a lot cleaner imo and the import interface is easy

#

And as long as I add each of the ast nodes it'll be interop

sick hound Jan 28, 2023, 9:16 PM

#

sick hound https://github.com/dzshn/lispy this only uses `tokenize.tokenize`

anyways I gave this as example cuz by using tokenize it actually made parsing simpler

#

there's only like 20 lines of code doing parsing

old socket Jan 28, 2023, 9:17 PM

#

I was gonna use lark

#

So I can write some good "EBNF" type grammar

#

bytes -> str -> lark -> visitor -> ast.Module -> return ast.unparse

#

Do you need to do anything special to make the code execute? Because I manually put print("Hello World!") into the codec decoder and it didn't print at all when running the file

sick hound Jan 28, 2023, 9:19 PM

#

you just return valid code from your codec

old socket Jan 28, 2023, 9:20 PM

#

Oh

#

Do you know why it wasn't running then?

#

I basically had ```py
def decode(...) -> tuple[str, int]:
data = b"print('Hello World')"

return data.decode(), len(data)

#

Or am I looking at this too simply

sick hound Jan 28, 2023, 9:22 PM

#

return len(input)

#

not data

#

this code works for plugging in a function that transforms bytes to bytes: https://github.com/dzshn/pypp/blob/main/pypp/_codec.py

old socket Jan 28, 2023, 9:24 PM

#

The integer being returned is the amount of lines or?

#

I've been kinda looking at everyone elses codecs without doing much looking at the documentation

#

Or the read line count?

sick hound Jan 28, 2023, 9:24 PM

#

it's the amount of bytes consumed from input (if it's less than input's length the function will be called again)

sick hound Jan 28, 2023, 9:25 PM

#

old socket I've been kinda looking at everyone elses codecs without doing much looking at t...

tbh I had to try and guess a lot

old socket Jan 28, 2023, 9:25 PM

#

Oh, so you just get the length on the input since you can assume all the bytes had been read

#

I see

sick hound Jan 28, 2023, 9:25 PM

#

somehow there's even wrong stubs for this

old socket Jan 28, 2023, 9:25 PM

#

I guess you can also do BytesIO and read line by line

#

then use .tell

sick hound Jan 28, 2023, 9:26 PM

#

I recommend doing what that script does and just consume the entire input at once, and hook whatever code you want

#

otherwise it's kind of a headache

old socket Jan 28, 2023, 9:27 PM

#

Hmmm I still cannot get it to run

#

I'll try again

#

Yea, I cannot get the codec to replace invalid code with valid code

#

λ  ~/custom  poetry run python test.py
  b'\n\nlet foo: str = "Hello World"\n'
``` It does get the invalid syntax here when printing

old socket Jan 28, 2023, 9:53 PM

#

~~Okay... I guess I need to add \n\n then finish with \n in order to get the code to run~~

#

I'm super confused lmao

#

Looks like I just need to end with a new line other wise it won't actually execute

quartz wave Jan 28, 2023, 11:28 PM

#

sick hound https://github.com/dzshn/lispy this only uses `tokenize.tokenize`

how do you not get stuck in positioning hell

sick hound Jan 28, 2023, 11:42 PM

#

quartz wave how do you not get stuck in positioning hell

by simply ignoring everything but type and string lol

quartz wave Jan 28, 2023, 11:56 PM

#

sick hound by simply ignoring everything but `type` and `string` lol

so like doesn't that normalize whitespace

#

because in mine it's an objective to keep it as-is

sick hound Jan 28, 2023, 11:57 PM

#

with lispy it doesn't matter cuz it build's it's own thing and executes it

#

pypp though.. lol

#

the input for that

old socket Jan 29, 2023, 2:14 AM

#

Is input of decode for codec only memoryview when an error happens?

meager zinc Jan 29, 2023, 5:28 AM

#

At this point you can definitely write a C to Python transpiler (platform-dependent).

c = __import__('ctypes.util')
c_lib = c.CDLL(c.util.find_library('msvcrt'))

fptr = c_lib._get_osfhandle(1)
realf = c_lib._open_osfhandle(fptr, 8)

c_lib._fdopen.restype = c.POINTER(c.c_char)
fd = c_lib._fdopen(realf, 'w')

c_lib.fputs(b'Hello msvcrt!', fd)

c_lib.fflush(fd)
c_lib.fclose(fd)
c_lib._exit(0)

Output:

Hello msvcrt!

(keep in mind that this actually bypasses things like contextlib.redirect_stdout)

languid hare Jan 29, 2023, 5:31 AM

#

sick hound pypp though.. lol

what even

#

macro hell

meager zinc Jan 29, 2023, 5:33 AM

#

The implementation details are pretty smart

#

it uses ##encoding pypp and preprocesses it by manipulating the encoding

sick hound Jan 29, 2023, 6:02 AM

#

Hi I’m back

old socket Jan 29, 2023, 6:08 AM

#

old socket bytes -> str -> lark -> visitor -> ast.Module -> return ast.unparse

Just realised after thinking about it more that it isn't this simple 😔

#

I'll need to make a node visitor class then check column and line numbers to manually indent

dry mirage Jan 29, 2023, 1:03 PM

#

meager zinc At this point you can definitely write a C to Python transpiler (platform-depend...

this isn't really a transpiler though..? it just calls C from Python

#

sort of equivalent to a python to C transpiler calling the python subprocess from C

meager zinc Jan 29, 2023, 4:59 PM

#

dry mirage this isn't really a transpiler though..? it just calls C from Python

Not sure if there would be another way though? How would you simulate malloc, calloc and pointers well otherwise?

dry mirage Jan 29, 2023, 5:00 PM

#

it only needs to be functionally equivalent, not implementation details though?

meager zinc Jan 29, 2023, 5:00 PM

#

I suppose but also a lot of the C library is inaccessible

#

Like you can't catch floating-point errors with _clear87

#

If you are using python floats

dry mirage Jan 29, 2023, 5:01 PM

#

why would you have a floating point error with python floats pithink

meager zinc Jan 29, 2023, 5:01 PM

#

No, in C you can have a floating point error (underflow/overflow)

#

and then catch it

#

lemme find an example

dry mirage Jan 29, 2023, 5:02 PM

#

right, but surely you can do the same functional thing without messing with float errors?

meager zinc Jan 29, 2023, 5:02 PM

#

#

Usually yes, but you would have to know at compile-time when such issues would occur so that you can manually set the fp error

#

(technically transpile-time)

dry mirage Jan 29, 2023, 5:03 PM

#

meager zinc

!e

a = b = 1e-40

y = a
print(y)

b = y
print(b)

night quarryBOT Jan 29, 2023, 5:03 PM

#

@dry mirage :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | 1e-40
002 | 1e-40

dry mirage Jan 29, 2023, 5:03 PM

#

this all just works without errors

meager zinc Jan 29, 2023, 5:04 PM

#

Which is also technically a problem

meager zinc Jan 29, 2023, 5:04 PM

#

meager zinc

because sometimes C code abuses such behavior

dry mirage Jan 29, 2023, 5:04 PM

#

an implementation that relies on float errors can be refactored to not

meager zinc Jan 29, 2023, 5:04 PM

#

true

dry mirage Jan 29, 2023, 5:04 PM

#

like functionally you want to calculate Fibonacci sequences, it doesn't matter how you get there

meager zinc Jan 29, 2023, 5:04 PM

#

but refactoring on a scale like that is not typically possible in a transpiler

pure dew Jan 29, 2023, 8:15 PM

#

sick hound this code works for plugging in a function that transforms bytes to bytes: <http...

you wrote this? damn good job

sinful stirrup Jan 31, 2023, 2:02 AM

#

wrote this one liner to raise a matrix m to the power of n. think you can make it shorter? (no semicolons or using numpy)

[[m := [[sum([r[i] * c[i] for i in range(len(c))]) for c in zip(*p)] for r in m] for _ in range(n - 1)] for p in [m]][-1][-1]

past plank Jan 31, 2023, 2:04 AM

#

without actually thinking about the math, you've got a lot of unnecessary spaces

#

[[m:=[[sum([r[i]*c[i]for i in range(len(c))])for c in zip(*p)]for r in m]for _ in range(n - 1)]for p in [m]][-1][-1]

languid hare Jan 31, 2023, 2:18 AM

#

[[m:=[[sum(R*C for R,C in zip(r,c))for c in zip(*p)]for r in m]for _ in range(n-1)]for p in [m]][-1][-1]

other low hanging optims

versed eagle Jan 31, 2023, 2:19 AM

#

for p in [m] feels wrong

languid hare Jan 31, 2023, 2:20 AM

#

yea i wasnt too sure what that was for either

versed eagle Jan 31, 2023, 2:22 AM

#

versed eagle `for p in [m]` feels wrong

I think this does the same thing, but without that
not sure though

[m:=[[sum(R*C for R,C in zip(r,c))for c in zip(*m)]for r in m]for _ in range(n-1)][-1]

low lynx Jan 31, 2023, 2:23 AM

#

range(n-1) is shorter as [0]*~-n I think

meager zinc Jan 31, 2023, 2:23 AM

#

smart

#

[m:=[[sum(R*C for R,C in zip(r,c))for c in zip(*m)]for r in m]for _ in[0]*~-n][-1]

languid hare Jan 31, 2023, 2:24 AM

#

firEyes

low lynx Jan 31, 2023, 2:24 AM

#

does this even work

meager zinc Jan 31, 2023, 2:24 AM

#

Good question

versed eagle Jan 31, 2023, 2:25 AM

#

i don't even know what is supposed to do

#

in the first place

low lynx Jan 31, 2023, 2:25 AM

#

!e

m=[[1, 1], [1, 0]]
n = 5
[m:=[[sum(R*C for R,C in zip(r,c))for c in zip(*m)]for r in m]for _ in[0]*~-n][-1]
print(m)

night quarryBOT Jan 31, 2023, 2:25 AM

#

@low lynx :white_check_mark: Your 3.11 eval job has completed with return code 0.

[[1597, 987], [987, 610]]

low lynx Jan 31, 2023, 2:25 AM

#

no it doesn't

languid hare Jan 31, 2023, 2:25 AM

#

ive no idea if that is right

meager zinc Jan 31, 2023, 2:25 AM

#

!e ```py
m, n = [[1, 2, 3], [4, 5, 6], [7, 8, 9]], 2
print([[m := [[sum([r[i] * c[i] for i in range(len(c))]) for c in zip(p)] for r in m] for _ in range(n - 1)] for p in [m]][-1][-1])
m, n = [[1, 2, 3], [4, 5, 6], [7, 8, 9]], 2
print([m:=[[sum(RC for R,C in zip(r,c))for c in zip(m)]for r in m]for _ in[0]~-n][-1])

night quarryBOT Jan 31, 2023, 2:26 AM

#

@meager zinc :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | [[30, 36, 42], [66, 81, 96], [102, 126, 150]]
002 | [[30, 36, 42], [66, 81, 96], [102, 126, 150]]

meager zinc Jan 31, 2023, 2:26 AM

#

Nope, they are functionally equivalent

versed eagle Jan 31, 2023, 2:26 AM

#

(what is it supposed to be doing here)

low lynx Jan 31, 2023, 2:26 AM

#

did the original even work

meager zinc Jan 31, 2023, 2:26 AM

#

Matrix power

meager zinc Jan 31, 2023, 2:26 AM

#

low lynx did the original even work

Not correctly it seems

languid hare Jan 31, 2023, 2:26 AM

#

mm

versed eagle Jan 31, 2023, 2:26 AM

#

meager zinc Matrix power

idk what that is

languid hare Jan 31, 2023, 2:27 AM

#

is it updating m in place or having that expression be the power, with m modified along the way

sharp moon Jan 31, 2023, 2:27 AM

#

versed eagle idk what that is

@BugleBoy#1234 multing a matrix by itself n times

meager zinc Jan 31, 2023, 2:27 AM

#

!e ```py
import numpy as np
x = np.arange(1, 10, 1).reshape(3,3)
y = np.matrix(x)

a = y**3
print(a)

versed eagle Jan 31, 2023, 2:27 AM

#

sharp moon @BugleBoy#1234 multing a matrix by itself n times

dk what a matrix is ¯_(ツ)_/¯

night quarryBOT Jan 31, 2023, 2:27 AM

#

@meager zinc :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | [[ 468  576  684]
002 |  [1062 1305 1548]
003 |  [1656 2034 2412]]

low lynx Jan 31, 2023, 2:27 AM

#

!e

import numpy as np
m = [[1, 1],[1, 0]]
n = 10
print(np.array(m)**n)
print([[m := [[sum([r[i] * c[i] for i in range(len(c))]) for c in zip(*p)] for r in m] for _ in range(n - 1)] for p in [m]][-1][-1])

night quarryBOT Jan 31, 2023, 2:27 AM

#

@low lynx :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | [[89, 55], [55, 34]]
002 | [[-5711768217452919631   253295162119140625]
003 |  [  253295162119140625     2064377754059776]]

low lynx Jan 31, 2023, 2:28 AM

#

wait wtf

#

numpy?

meager zinc Jan 31, 2023, 2:28 AM

#

versed eagle dk what a matrix is ¯\_(ツ)_/¯

pretty much a 2d list

languid hare Jan 31, 2023, 2:28 AM

#

uhh

#

lmao

#

oh

versed eagle Jan 31, 2023, 2:28 AM

#

meager zinc pretty much a 2d list

oh ok

languid hare Jan 31, 2023, 2:28 AM

#

m got updated

meager zinc Jan 31, 2023, 2:28 AM

#

low lynx wait wtf

M is modified on the prev line

low lynx Jan 31, 2023, 2:28 AM

#

ah

versed eagle Jan 31, 2023, 2:28 AM

#

meager zinc pretty much a 2d list

then what would multiplying them do

languid hare Jan 31, 2023, 2:28 AM

#

!e

import numpy as np
m = [[1, 1],[1, 0]]
n = 10
print([[m := [[sum([r[i] * c[i] for i in range(len(c))]) for c in zip(*p)] for r in m] for _ in range(n - 1)] for p in [m]][-1][-1])

m = [[1, 1],[1, 0]]
print(np.array(m)**n)

night quarryBOT Jan 31, 2023, 2:28 AM

#

@languid hare :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | [[89, 55], [55, 34]]
002 | [[1 1]
003 |  [1 0]]

versed eagle Jan 31, 2023, 2:28 AM

#

you can't multiply lists

meager zinc Jan 31, 2023, 2:28 AM

#

Heh

sharp moon Jan 31, 2023, 2:28 AM

#

versed eagle dk what a matrix is ¯\_(ツ)_/¯

Its a grid of numbers used in maths with special rukes on how to treat them which is useful in fields such as 3d rotation and loads of stufff

low lynx Jan 31, 2023, 2:28 AM

#

!e

import numpy as np
m = [[1, 1],[1, 0]]
n = 10
print(np.array(m)**n)
print([[m := [[sum([r[i] * c[i] for i in range(len(c))]) for c in zip(*p)] for r in m] for _ in range(n - 1)] for p in [m]][-1][-1])

night quarryBOT Jan 31, 2023, 2:28 AM

#

@low lynx :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | [[1 1]
002 |  [1 0]]
003 | [[89, 55], [55, 34]]

low lynx Jan 31, 2023, 2:29 AM

#

...

meager zinc Jan 31, 2023, 2:29 AM

#

Interesting

languid hare Jan 31, 2023, 2:29 AM

#

oh piss

versed eagle Jan 31, 2023, 2:29 AM

#

sharp moon Its a grid of numbers used in maths with special rukes on how to treat them whic...

ah

low lynx Jan 31, 2023, 2:29 AM

#

am I bad at numpy

languid hare Jan 31, 2023, 2:29 AM

#

** is elementwise

meager zinc Jan 31, 2023, 2:29 AM

#

the original didn't even work

low lynx Jan 31, 2023, 2:29 AM

#

ahhh

#

ofc

versed eagle Jan 31, 2023, 2:29 AM

#

sharp moon Its a grid of numbers used in maths with special rukes on how to treat them whic...

it's a fancy math thing

#

alr

languid hare Jan 31, 2023, 2:29 AM

#

!e

import numpy as np
m = [[1, 1],[1, 0]]
n = 10
print(np.linalg.matrix_power(m, n))
print([[m := [[sum([r[i] * c[i] for i in range(len(c))]) for c in zip(*p)] for r in m] for _ in range(n - 1)] for p in [m]][-1][-1])

#

wait

#

..

night quarryBOT Jan 31, 2023, 2:30 AM

#

@languid hare :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | [[89 55]
002 |  [55 34]]
003 | [[89, 55], [55, 34]]

languid hare Jan 31, 2023, 2:30 AM

#

there

meager zinc Jan 31, 2023, 2:30 AM

#

import numpy as np
m = [[1, 1],[1, 0]]
n = 10
print(np.linalg.matrix_power(m, n))
m = [[1, 1],[1, 0]]
n = 10
print([[m := [[sum([r[i] * c[i] for i in range(len(c))]) for c in zip(*p)] for r in m] for _ in range(n - 1)] for p in [m]][-1][-1])

#

so it is working correctly

#

shortest:

[m:=[[sum(R*C for R,C in zip(r,c))for c in zip(*m)]for r in m]for _ in[0]*~-n][-1]

languid hare Jan 31, 2023, 2:31 AM

#

fyi original broke for n = 0 and 1

meager zinc Jan 31, 2023, 2:32 AM

#

meager zinc shortest: ```py [m:=[[sum(R*C for R,C in zip(r,c))for c in zip(*m)]for r in m]fo...

ok well assuming we're just working with the original then this is the best so far

versed eagle Jan 31, 2023, 2:32 AM

#

the goal is to function the same as the original, no?

languid hare Jan 31, 2023, 2:32 AM

#

0 is a little messy since you need to return the identity, 1 should be an easy fix
but we could just assume n > 1 :)

meager zinc Jan 31, 2023, 2:32 AM

#

I would suppose so

versed eagle Jan 31, 2023, 2:33 AM

#

doesn't matter if the original is flawed, as long as ours has the same flaw

meager zinc Jan 31, 2023, 2:33 AM

#

Man list comprehensions are great but also too long

#

[m:=[[sum(R*C for R,C in zip(r,c))for c in zip(*m)]for r in m]for _ in[0]*~-n][-1]
# PEP 123.456: Implicit comprehensions
[m:=[[sum(R*CR,Czip(r,c))czip(*m)]rm]_[0]*~-n][-1]

languid hare Jan 31, 2023, 2:39 AM

#

Thonk

simple sphinx Jan 31, 2023, 4:00 AM

#

I made a thing I think people here would like, here is a kata about minimalist python:
https://www.codewars.com/kata/63d4b700bce90f0024a9ca19

Codewars

Joyful Transpiler

Background
After being inspired by Marie Kondo's teachings on minimalism, you've decided to apply her principles to your computer science work. You've noticed that your keyboard has 104 keys and ...

versed eagle Jan 31, 2023, 4:03 AM

#

exciting

versed eagle Jan 31, 2023, 6:24 AM

#

!e

(
    (c := __import__("ctypes")),
    (gfp := (lambda f: id(f) + 16)),
    (cvp_fa := (lambda o: c.c_voidp.from_address(o))),
    (sw_f := (lambda f, F, *, s={}: (
        s.__setitem__(f, cvp_fa(gfp(f)).value),
        cvp_fa(gfp(f)).__setattr__("value", s.get(F, False) or cvp_fa(gfp(F)).value),
    None)[-1]))
)

sw_f(max, min)
sw_f(min, max)

print(min(1, 2, 3), max(1, 2, 3))

night quarryBOT Jan 31, 2023, 6:24 AM

#

@versed eagle :white_check_mark: Your 3.11 eval job has completed with return code 0.

3 1

versed eagle Jan 31, 2023, 6:25 AM

#

versed eagle !e ```py ( (c := __import__("ctypes")), (gfp := (lambda f: id(f) + 16)),...

this isn't super complicated, but its fun to play around with

versed eagle Jan 31, 2023, 7:16 AM

#

<@&831776746206265384> unrelated to channel topic, and sending the same message in multiple channels rapidly

#

also, is a scam

simple sphinx Jan 31, 2023, 7:17 AM

#

heck yea I love crypto bots 😄

versed eagle Jan 31, 2023, 7:18 AM

#

versed eagle <@&831776746206265384> unrelated to channel topic, and sending the same message ...

ty mods!

versed eagle Jan 31, 2023, 7:18 AM

#

simple sphinx heck yea I love crypto bots 😄

mhm.

simple sphinx Jan 31, 2023, 7:18 AM

#

lol

sinful stirrup Jan 31, 2023, 11:18 AM

#

meager zinc shortest: ```py [m:=[[sum(R*C for R,C in zip(r,c))for c in zip(*m)]for r in m]fo...

this doesn't seem to work

#

!e

from numpy.linalg import matrix_power

n = 3
m = [[1, 2, 3], [4, 5, 6], [7, 8, 9]]

print([m:=[[sum(R*C for R,C in zip(r,c))for c in zip(*m)]for r in m]for _ in[0]*~-n][-1])

m = [[1, 2, 3], [4, 5, 6], [7, 8, 9]]

print(matrix_power(m, n))

night quarryBOT Jan 31, 2023, 11:20 AM

#

@sinful stirrup :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | [[7560, 9288, 11016], [17118, 21033, 24948], [26676, 32778, 38880]]
002 | [[ 468  576  684]
003 |  [1062 1305 1548]
004 |  [1656 2034 2412]]

sinful stirrup Jan 31, 2023, 11:20 AM

#

🤔

#

i can't think of a way to do this without storing a copy of m to multiply with every time like what i do with for p in [m] (which btw feels really wrong but i think it's necessary)

sinful stirrup Jan 31, 2023, 11:24 AM

#

meager zinc shortest: ```py [m:=[[sum(R*C for R,C in zip(r,c))for c in zip(*m)]for r in m]fo...

so judging by my test this doesn't work but this:

[[m:=[[sum(R*C for R,C in zip(r,c))for c in zip(*m)]for r in p]for _ in[0]*~-n][-1]for p in[m]]

does

low lynx Jan 31, 2023, 11:25 AM

#

!e

from numpy.linalg import matrix_power
arr = [[1, 2, 3], [4, 5, 6], [7, 8, 9]]
print(matrix_power(arr, 8))

night quarryBOT Jan 31, 2023, 11:25 AM

#

@low lynx :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | [[ 510008400  626654232  743300064]
002 |  [1154967822 1419124617 1683281412]
003 |  [1799927244 2211595002 2623262760]]

low lynx Jan 31, 2023, 11:26 AM

#

hmm

#

not fixable

#

!e

from numpy.linalg import matrix_power
arr = [[1, 2, 3], [4, 5, 6], [7, 8, 9]]
print(matrix_power(arr, 4))

night quarryBOT Jan 31, 2023, 11:26 AM

#

@low lynx :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | [[ 7560  9288 11016]
002 |  [17118 21033 24948]
003 |  [26676 32778 38880]]

low lynx Jan 31, 2023, 11:26 AM

#

oh nvm

sinful stirrup Jan 31, 2023, 11:30 AM

#

sinful stirrup so judging by my test this doesn't work but this: ```py [[m:=[[sum(R*C for R,C i...

without using the p thing m keeps multiplying by itself and it also gets bigger every time so the numbers are gonna get exponentially bigger than they should be

arctic skiff Jan 31, 2023, 12:25 PM

#

night quarry <@395901889046249475> :white_check_mark: Your 3.11 eval job has completed with r...

where are commas?

earnest wing Jan 31, 2023, 12:36 PM

#

numpy doesn't repr with commas

#

they're a lot of noise when you have multidimensional arrays

low lynx Jan 31, 2023, 12:43 PM

#

numpy also uses ellipses when it gets big

#

!e

import numpy as np
print(np.ones((100, 100)))

night quarryBOT Jan 31, 2023, 12:44 PM

#

@low lynx :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | [[1. 1. 1. ... 1. 1. 1.]
002 |  [1. 1. 1. ... 1. 1. 1.]
003 |  [1. 1. 1. ... 1. 1. 1.]
004 |  ...
005 |  [1. 1. 1. ... 1. 1. 1.]
006 |  [1. 1. 1. ... 1. 1. 1.]
007 |  [1. 1. 1. ... 1. 1. 1.]]

old socket Jan 31, 2023, 12:45 PM

#

Is it just me or is anyone else slightly annoyed by the two brackets on top left corner and bottom right corner

#

Like, I'm sure it would be easy to get rid of those two

proper vault Jan 31, 2023, 12:51 PM

#

they are needed for consistency with 3d arrays afaik

vernal pivot Jan 31, 2023, 12:51 PM

#

old socket Is it just me or is anyone else slightly annoyed by the two brackets on top left...

think of how confusing it would be if multi dim arrays would be printed as a bunch of one dim arrays

#

you couldn't tell the difference between a 2d array and a bunch of 1d array printed on multiple lines

#

that would be a horrible design choice

#

!e like here:

import numpy as np

print(np.ones((5, 5)))
print(np.ones(5))

night quarryBOT Jan 31, 2023, 12:53 PM

#

@vernal pivot :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | [[1. 1. 1. 1. 1.]
002 |  [1. 1. 1. 1. 1.]
003 |  [1. 1. 1. 1. 1.]
004 |  [1. 1. 1. 1. 1.]
005 |  [1. 1. 1. 1. 1.]]
006 | [1. 1. 1. 1. 1.]

vernal pivot Jan 31, 2023, 12:53 PM

#

there is a clear distinction between the 2d and 1d array

#

without the double brackets it would look like the 1d array is part of the 2d array or the 2d array is a bunch of 1d arrays

quartz wave Jan 31, 2023, 12:54 PM

#

old socket Is it just me or is anyone else slightly annoyed by the two brackets on top left...

they're aligned quite nicely for me

old socket Jan 31, 2023, 12:55 PM

#

vernal pivot think of how confusing it would be if multi dim arrays would be printed as a bun...

Oh yea, I forgot to consider what it would look like with two arrays

low lynx Jan 31, 2023, 1:26 PM

#

you also need them for higher dims

#

!e

import numpy as np
print(np.zeros((3, 3, 3, 3, 3)))

night quarryBOT Jan 31, 2023, 1:28 PM

#

@low lynx :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | [[[[[0. 0. 0.]
002 |     [0. 0. 0.]
003 |     [0. 0. 0.]]
004 | 
005 |    [[0. 0. 0.]
006 |     [0. 0. 0.]
007 |     [0. 0. 0.]]
008 | 
009 |    [[0. 0. 0.]
010 |     [0. 0. 0.]
011 |     [0. 0. 0.]]]
... (truncated - too many lines)

Full output: https://paste.pythondiscord.com/arewajaqug.txt?noredirect

earnest wing Jan 31, 2023, 2:46 PM

#

old socket Like, I'm sure it would be easy to get rid of those two

!e import numpy as np; print(np.ones((4,4,4,4)))

night quarryBOT Jan 31, 2023, 2:46 PM

#

@earnest wing :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | [[[[1. 1. 1. 1.]
002 |    [1. 1. 1. 1.]
003 |    [1. 1. 1. 1.]
004 |    [1. 1. 1. 1.]]
005 | 
006 |   [[1. 1. 1. 1.]
007 |    [1. 1. 1. 1.]
008 |    [1. 1. 1. 1.]
009 |    [1. 1. 1. 1.]]
010 | 
011 |   [[1. 1. 1. 1.]
... (truncated - too many lines)

Full output: https://paste.pythondiscord.com/sifoteleya.txt?noredirect

earnest wing Jan 31, 2023, 2:47 PM

#

well looks like that was already answered above

dry mirage Jan 31, 2023, 8:45 PM

#

versed eagle !e ```py ( (c := __import__("ctypes")), (gfp := (lambda f: id(f) + 16)),...

added View.swap now

from einspect import view

view("dog").swap("cat")
print("dog", "cat")
# >> cat dog

view(max).swap(min)
print(min(1, 2, 3), max(1, 2, 3))
# >> 3 1

versed eagle Jan 31, 2023, 8:46 PM

#

cool

low lynx Jan 31, 2023, 8:55 PM

#

ooh nice

#

!e

from einspect import view

view(True).swap(False)
print(True, False)

while True:
    pass
else:
    if False:
         print('cursed')

night quarryBOT Jan 31, 2023, 8:58 PM

#

@low lynx :x: Your 3.11 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 3, in <module>
003 | AttributeError: 'BoolView' object has no attribute 'swap'

low lynx Jan 31, 2023, 8:58 PM

#

aww

rugged owl Jan 31, 2023, 8:58 PM

#

pithink

meager zinc Jan 31, 2023, 11:43 PM

#

low lynx !e ```py from einspect import view view(True).swap(False) print(True, False) w...

I've done this before -- it gets optimized out by the bytecode compiler sadly

low lynx Jan 31, 2023, 11:44 PM

#

PensiveCat

#

I figured it might

#

wonder if putting it in a function would change things

versed eagle Jan 31, 2023, 11:57 PM

#

low lynx wonder if putting it in a function would change things

probably not

#

since it'd still get const folded

#

it should work with bool() though

#

since that's not a constant

topaz haven Feb 1, 2023, 1:36 AM

#

what does this channel do?
something like


print.__call__((lambda x,y,z: "\150\151\041"+str(x*y) * z )(2838,493734,0))
i = (lambda: 1+1+1)()
eval("\160\162\151"+"\156\164("+str(print)[3]+"+"+str((1248+9384)%1)+")")

right

lavish dirge Feb 1, 2023, 2:17 AM

#

Google Docs

EdgeLore Language Updated Jan/23

POWER = "AMOUNT" STRENGTH = "LEVEL INTENSITY" ENGINE = "MOTOR IN WHICH AN OPERATOR USES TO POWER A SYSTEM" SCAN = "ANALYZE A SPECIFIC WORD OR FIELD AND OR GIVE DATA2 ON THE ASKED INFORMATION TO SEARCH FOR" ANALYZE = "READ AND LOOK OVER" IMMUNE = "DOES NOT AFFECT" DOMAIN = "AREA OWNED AND CONTROL...

Google Docs

Class Definitions

class CONNECTIONCLASSIFICATIONS: NETWORKBOUNADARIES = [0,1,2,3,4,5,6,7,8,9,10,11,12,13] NETWORKBOUNADARIES[0] = INPUT NETWORKBOUNADARIES[1] = OUTPUT NETWORKBOUNADARIES[2] = INCOMING NETWORKBOUNADARIES[3] = OUTCOMING NETWORKBOUNADARIES[4] = INGOING NETWORKBOUNADARIES[5...

Google Docs

Words With Plural Usage (Partial)

THE,WORD,CONTROLS,IS,PLURAL THE,WORD,CONTROLLERS,IS,PLURAL THE,WORD,GUILDS,IS,PLURAL THE,WORD,ROUNDTABLES,IS,PLURAL THE,WORD,MOVEMENTS,IS,PLURAL THE,WORD,MOVES,IS,PLURAL THE,WORD,ROTATES,IS,PLURAL THE,WORD,ROTATIONS,IS,PLURAL THE,WORD,EDITORS,IS,PLURAL THE,WORD,SKILLSETS,IS,PLURAL THE,WORD,SKILLS...

Google Docs

Words As Nouns And Verbs

THE,WORD,POWER,IS,A,NOUN THE,WORD,STRENGTH,IS,A,NOUN THE,WORD,ENGINE,IS,A,NOUN THE,WORD,SCAN,IS,A,VERB THE,WORD,ANALYZE,IS,A,VERB THE,WORD,IMMUNE,IS,A,NOUN THE,WORD,DOMAIN,IS,A,NOUN THE,WORD,VIRTUAL,IS,A,NOUN THE,WORD,SOUND,IS,A,NOUN THE,WORD,FREQUENCY,IS,A,NOUN THE,WORD,IMMUNITY,IS,A,VERB THE,W...

Google Docs

Words As Singular

THE,WORD,POWER,IS,SINGULAR THE,WORD,STRENGTH,IS,SINGULAR THE,WORD,ENGINE,IS,SINGULAR THE,WORD,SCAN,IS,SINGULAR THE,WORD,ANALYZE,IS,SINGULAR THE,WORD,IMMUNE,IS,SINGULAR THE,WORD,DOMAIN,IS,SINGULAR THE,WORD,VIRTUAL,IS,SINGULAR THE,WORD,SOUND,IS,SINGULAR THE,WORD,FREQUENCY,IS,SINGULAR THE,WORD,IMMUN...

pure dew Feb 1, 2023, 4:44 AM

#

lavish dirge Hi everyone! For those that were interested in keeping up with my work before, H...

respectfully, what the fuck is this

lavish dirge Feb 1, 2023, 4:50 AM

#

Building a language from scratch using constants as a base, building a set of information to be compiled all at once using an RNN.

Every word has to be defined and that includes the words on the definition to go further. Once that is made, the language is made to function as a set of metadata like parameters that it pulls information from within by using similar information related to the existing terminology.

The logic behind it is simple. It can only use what is available to it. Once the classes are added, it acts as a form of structure for the parameters to form structure.

pure dew Feb 1, 2023, 4:51 AM

#

oh damn ok

lavish dirge Feb 1, 2023, 4:52 AM

#

Sorry should have said that first

pure dew Feb 1, 2023, 4:52 AM

#

why python as a base tho?

#

why not a simpler custom parser?

lavish dirge Feb 1, 2023, 4:52 AM

#

Able to disable character length limiter.

versed eagle Feb 1, 2023, 4:52 AM

#

lavish dirge Building a language from scratch using constants as a base, building a set of in...

that would be infinitely recursive, no?

#

since each word would be continually defined by more words

#

which are also defined by words

#

and so on

lavish dirge Feb 1, 2023, 4:53 AM

#

Evolutionary algorithm base in more better terms

#

Infinite addition of words yes.

pure dew Feb 1, 2023, 4:54 AM

#

lavish dirge Able to disable character length limiter.

what now?

lavish dirge Feb 1, 2023, 4:57 AM

#

It's chosen to use with RNN because of the high complexity of frequency and patterns that are able to be used with it.

What I'm working with now is developing class and def functions and then attempting to make classmethod.

The way I want to teach it is using persistent memory to store the set of commands in ram as a form of recall. I'm making the structure and classes but need to find a way to put the classes first so that way I can use the words in the classes. I have an interface that compiles already. One compiles the definition of each word and the other compiles the word itself

versed eagle Feb 1, 2023, 4:59 AM

#

how is the definition compilation not infinitely recursive

pure dew Feb 1, 2023, 4:59 AM

#

but why python syntax? wouldn't it have been easier to build your own parser that processes the input into a super simple tree?

versed eagle Feb 1, 2023, 4:59 AM

#

unless you have some words which are given "constant" definitions, with no other words in them (i.e. stopping points for the recursion)

pure dew Feb 1, 2023, 4:59 AM

#

i really wanna see the code

versed eagle Feb 1, 2023, 5:00 AM

#

im quite curious as well lmao

lavish dirge Feb 1, 2023, 5:00 AM

#

Add me on discord. What time zone you both in

versed eagle Feb 1, 2023, 5:01 AM

#

why?

lavish dirge Feb 1, 2023, 5:01 AM

#

I'm in eastern

#

Respect to time difference

versed eagle Feb 1, 2023, 5:02 AM

#

ah

lavish dirge Feb 1, 2023, 5:02 AM

#

12am here

versed eagle Feb 1, 2023, 5:02 AM

#

its fine for you to message me whenever
i turn notifications off for when i sleep so it wont wake me up or anything

pure dew Feb 1, 2023, 5:03 AM

#

wdym add you? friends?

lavish dirge Feb 1, 2023, 5:03 AM

#

Yeah I don't mind

pure dew Feb 1, 2023, 5:03 AM

#

i'm eastern but DND is a lifesaver

lavish dirge Feb 1, 2023, 5:03 AM

#

I'm eastern time too

versed eagle Feb 1, 2023, 5:04 AM

#

pure dew i'm eastern but DND is a lifesaver

the game or Do Not Disturb?

lavish dirge Feb 1, 2023, 5:04 AM

#

By the way if used in cmd it is recursive

#

But not python. Do you know why

versed eagle Feb 1, 2023, 5:05 AM

#

versed eagle the game or Do Not Disturb?

im pretty sure you mean Do Not Disturb, but i feel a need to ask for clarification

lavish dirge Feb 1, 2023, 5:05 AM

#

It is because in Python, when you use the word and it has a definition that is within the word base, it will compile an error until the word within the definition is defined.

versed eagle Feb 1, 2023, 5:05 AM

#

hm

lavish dirge Feb 1, 2023, 5:06 AM

#

Python will not compile the definition unless the words within the definitions are defined as well

versed eagle Feb 1, 2023, 5:06 AM

#

so then how do you compile the first word definition

lavish dirge Feb 1, 2023, 5:06 AM

#

To compile the first word I had to define power

versed eagle Feb 1, 2023, 5:06 AM

#

if it wont compile until all the words in it are defined

lavish dirge Feb 1, 2023, 5:06 AM

#

Which is why it took over 4 years

#

It did have some ups and downs. But I was only able to compile everything with the help of codeblocks running a c++ version to find errors, and a Unity 3D compilation of errors as well as serializing the whole language manually each word one by one

pure dew Feb 1, 2023, 5:07 AM

#

versed eagle the game or Do Not Disturb?

both tbh, but the latter

versed eagle Feb 1, 2023, 5:07 AM

#

alr

pure dew Feb 1, 2023, 5:08 AM

#

honestly man I'm having trouble understanding your explanations

#

is there a paper on this?

lavish dirge Feb 1, 2023, 5:09 AM

#

Yes as a matter of fact and there is a new method other than binary for it to run

pure dew Feb 1, 2023, 5:10 AM

#

i would love to hear more, but this might not be the channel for extended discussion

#

mm no threads either

lavish dirge Feb 1, 2023, 5:12 AM

#

We could both start a group chat if that helps.

rough moat Feb 1, 2023, 6:59 AM

#

print(str("\n".join(["".join(["█"if 5<(((x-10)**2+(y-10)**2)**0.5)<7 or 5<(((x-40)**2+(y-10)**2)**0.5)<7 or(x==20 and 9<y<16)or(x==30 and 9<y<16)or(((x+y+3)%8>5)and(20<x<25)and(8<y<15))or(((8-(x-y+3)%8)>5)and(24<x<30)and(8<y<15))else" "for x in range(60)])for y in range(30)])))

low lynx Feb 1, 2023, 7:11 AM

#

!e

print(str("\n".join(["".join(["█"if 5<(((x-10)**2+(y-10)**2)**0.5)<7 or 5<(((x-40)**2+(y-10)**2)**0.5)<7 or(x==20 and 9<y<16)or(x==30 and 9<y<16)or(((x+y+3)%8>5)and(20<x<25)and(8<y<15))or(((8-(x-y+3)%8)>5)and(24<x<30)and(8<y<15))else" "for x in range(60)])for y in range(30)])))

night quarryBOT Feb 1, 2023, 7:11 AM

#

@low lynx :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 |                                                             
002 |                                                             
003 |                                                             
004 |                                                             
005 |        ███████                       ███████                
006 |       ████ ████                     ████ ████               
007 |      ██       ██                   ██       ██              
008 |     ██         ██                 ██         ██             
009 |     ██         ██                 ██         ██             
010 |     ██         ██                 ██         ██             
011 |     █           █   █    █    █   █           █             
... (truncated - too many lines)

Full output: https://paste.pythondiscord.com/kepopinaju.txt?noredirect

low lynx Feb 1, 2023, 7:12 AM

#

...

rough moat Feb 1, 2023, 7:12 AM

#

teehee

#

owo in one line of code

low lynx Feb 1, 2023, 7:12 AM

#

I shall golf once my internet returns

olive aspen Feb 1, 2023, 7:12 AM

#

i was looking through the discord randomly n see that pop up instantly

#

quite the treat

rough moat Feb 1, 2023, 7:13 AM

#

#

its not necessarily as short as possible, i just wrote it while bored

low lynx Feb 1, 2023, 7:23 AM

#

preliminary golf 243

print('\n'.join(''.join(' █'[25<(x-10)**2+(y-10)**2<49or 25<(x-40)**2+(y-10)**2<49or(x==20and 9<y<16)or(x==30and 9<y<16)or(((x+y+3)%8>5)and(20<x<25)and(8<y<15))or((((x-y+3)%8)<3)and(24<x<30)and(8<y<15))]for x in range(60))for y in range(30)))

versed eagle Feb 1, 2023, 7:50 AM

#

print('\n'.join(''.join(' █'[25<(x-10)**2+(y-10)**2<49or 25<(x-40)**2+(y-10)**2<49or(x==20and 9<y<16)|(x==30and 9<y<16)|(((x+y+3)%8>5)&(20<x<25)&(8<y<15))|(((x-y+3)%8<3)&(24<x<30)&(8<y<15))]for x in range(60))for y in range(30)))

#

small optimisations

#

228

quartz wave Feb 1, 2023, 10:22 AM

#

major optimizations, 194 ```py
print('\n'.join(''.join(' █'[25<(x-10)**2+(d:=y-10)d<49or 25<(x-40)**2+dd<49or 9<y<16<x in{20,30}or(8<y<15)==1in((x+y+3)%8>5<x-15<10,(x-y+3)%8<3<x-21<9)]for x in range(60))for y in range(30)))

low lynx Feb 1, 2023, 12:07 PM

#

minor optimizations, 187

print('\n'.join(''.join(' █'[5<abs((z:=x+1j*y-10j)-10)<7or 5<abs(z-40)<7or 9<y<16<x in{20,30}or(9<y<16)and((x+y+3)%8>5<x-15<10or(x-y+3)%8<3<x-21<9)]for x in range(60))for y in range(30)))

quartz wave Feb 1, 2023, 12:18 PM

#

and -> &, 185 ```py
print('\n'.join(''.join(' █'[5<abs((z:=x+1j*y-10j)-10)<7or 5<abs(z-40)<7or 9<y<16<x in{20,30}or(9<y<16)&((x+y+3)%8>5<x-15<10or(x-y+3)%8<3<x-21<9)]for x in range(60))for y in range(30)))

#

combined comparisons, 178 ```py
print('\n'.join(''.join(' █'[5<abs((z:=x+1j*y-10j)-10)<7or 5<abs(z-40)<7or-7<y-16<0<(x in{20,30}or(x+y+3)%8>5<x-15<10or(x-y+3)%8<3<x-21<9)]for x in range(60))for y in range(30)))

low lynx Feb 1, 2023, 12:28 PM

#

anything possible with switching to mod stuff and cutting out the second for loop?

gleaming linden Feb 1, 2023, 12:29 PM

#

168 turn outside generator into explicit for loop```py
for y in range(30):print(''.join(' █'[5<abs((z:=x+1j*y-10j)-10)<7or 5<abs(z-40)<7or-7<y-16<0<(x in{20,30}or(x+y+3)%8>5<x-15<10or(x-y+3)%8<3<x-21<9)]for x in range(60)))

low lynx Feb 1, 2023, 12:35 PM

#

167 combined the two Os into one comparison

for y in range(30):print(''.join(' █'[5<min(abs((z:=x+1j*y-10j)-10),abs(z-40))<7or-7<y-16<0<(x in{20,30}or(x+y+3)%8>5<x-15<10or(x-y+3)%8<3<x-21<9)]for x in range(60)))

quartz wave Feb 1, 2023, 12:43 PM

#

low lynx 167 combined the two Os into one comparison ```py for y in range(30):print(''.jo...

165 take advantage of the fact that x and y are already separated ```py
for y in range(30):print(''.join(' █'[5<min(abs(z:=x+1j*y-10j-10),abs(z-30))<7or-7<y-16<0<(x in{20,30}or(x+y+3)%8>5<x-15<10or(x-y+3)%8<3<x-21<9)]for x in range(60)))

gleaming linden Feb 1, 2023, 12:43 PM

#

158 combine the two Os even more ```py
for y in range(30):print(''.join(' █'[5<abs(x+1j*y-10j-[40,10][x<30])<7or-7<y-16<0<(x in{20,30}or(x+y+3)%8>5<x-15<10or(x-y+3)%8<3<x-21<9)]for x in range(60)))

quartz wave Feb 1, 2023, 12:43 PM

#

aw

gleaming linden Feb 1, 2023, 12:46 PM

#

157 ```py
for y in range(30):print(''.join(' █'[5<abs(x+1jy-10j-10-30(x>30))<7or-7<y-16<0<(x in{20,30}or(x+y+3)%8>5<x-15<10or(x-y+3)%8<3<x-21<9)]for x in range(60)))

low lynx Feb 1, 2023, 1:35 PM

#

156

for y in range(30):print(''.join(' █'[5<abs(x+1j*y-10j-10-x//30*30)<7or-7<y-16<0<(x in{20,30}or(x+y+3)%8>5<x-15<10or(x-y+3)%8<3<x-21<9)]for x in range(60)))

#

wait

low lynx Feb 1, 2023, 1:44 PM

#

low lynx minor optimizations, 187 ```py print('\n'.join(''.join(' █'[5<abs((z:=x+1j*y-10j...

this one actually has a different W

#

from the original

#

it doesn't matter ig

#

an OwO is an OwO

quartz wave Feb 1, 2023, 1:45 PM

#

still looks like a W

#

154 by Divisor * Quotient = Dividend - Remainder ```py
for y in range(30):print(''.join(' █'[5<abs(x+1j*y-10j-10-x+x%30)<7or-7<y-16<0<(x in{20,30}or(x+y+3)%8>5<x-15<10or(x-y+3)%8<3<x-21<9)]for x in range(60)))

low lynx Feb 1, 2023, 1:50 PM

#

here's an image of the current so as to not mess it up again

Screen_Shot_2023-02-02_at_12.50.38_am.png

#

152 by adjusting the modulo inequalities, first one to turn 10->7 and second to remove the +3

for y in range(30):print(''.join(' █'[5<abs(x+1j*y-10j-10-x+x%30)<7or-7<y-16<0<(x in{20,30}or(x+y-3)%8<2<x-18<7or(x-y)%8>4<x-20<10)]for x in range(60)))

#

148 if we're fine with extra whitespace

for y in(r:=range(60)):print(''.join(' █'[5<abs(x+1j*y-10j-10-x+x%30)<7or-7<y-16<0<(x in{20,30}or(x+y-3)%8<2<x-18<7or(x-y)%8>4<x-20<10)]for x in r))

tough willow Feb 1, 2023, 5:39 PM

#

exec(bytes('潦⁲⁹湩爨㴺慲杮⡥〶⤩瀺楲瑮✨⸧潪湩✨袖嬧㰵扡⡳⭸樱礪ㄭ樰ㄭⴰ⭸╸〳㰩漷⵲㰷⵹㘱〼⠼⁸湩㉻ⰰ〳潽⡲⭸⵹⤳㠥㈼砼ㄭ㰸漷⡲⵸⥹㠥㐾砼㈭㰰〱崩潦⁲⁸湩爠⤩','u16')[2:]) 100 characters if you dont count bytes

burnt pasture Feb 1, 2023, 6:26 PM

#

I wrote a blog post a decade ago about crashing Python that tries to prevent malicious code: https://nedbatchelder.com/blog/201206/eval_really_is_dangerous.html Today I got a question about whether it's possible if you forbid underscores in the malicious code. I figured this channel would be up to the challenge.

Eval really is dangerous

eval() is tempting, but it’s truly dangerous. Here’s how common safety attempts can be circumvented.

flint hollow Feb 1, 2023, 6:56 PM

#

x.__class__ is the same as type(x) at least

rugged sparrow Feb 1, 2023, 7:13 PM

#

burnt pasture I wrote a blog post a decade ago about crashing Python that tries to prevent mal...

!e this gets the string for any dunder, without using ord encoding tricks you could probably detect for ```py
dunder = {n[2:][:-2]: n for n in dir([c for c in getattr(object, [n for n in dir(type) if 'subclasses' in n][0])() if 'weakref.ProxyType' in repr(c)][0])}.get
print(dunder('init'))

night quarryBOT Feb 1, 2023, 7:13 PM

#

@rugged sparrow :white_check_mark: Your 3.11 eval job has completed with return code 0.

__init__

rugged sparrow Feb 1, 2023, 7:14 PM

#

combined with getattr and setattr you can bypass underscores being blacklisted

versed eagle Feb 1, 2023, 7:16 PM

#

burnt pasture I wrote a blog post a decade ago about crashing Python that tries to prevent mal...

i would say that it's possible but with a massive amount of work

rugged sparrow Feb 1, 2023, 7:17 PM

#

as long as your backing interpreter is cpython anytime in the next 5 years, it will always be possible to wiggle through edge cases

#

even without constructing code objects you can get to arbitrary code exec at least 6 ways off the top of my head

#

and then you could use ROP or JOP to do whatever you want

versed eagle Feb 1, 2023, 7:18 PM

#

what im thinking of is using something like ast.parse and then interpreting that

#

to check if the code does stuff that you don't want

rugged sparrow Feb 1, 2023, 7:19 PM

#

versed eagle what im thinking of is using something like ast.parse and then interpreting that

that would work, as long as you only allow a limited subset

versed eagle Feb 1, 2023, 7:19 PM

#

wdym "allow a limited subset"

rugged sparrow Feb 1, 2023, 7:20 PM

#

a limited subset of python code

#

you would have to disable large chunks of the language in order to keep it secure, to the point it would be hard to call it python

versed eagle Feb 1, 2023, 7:21 PM

#

rugged sparrow you would have to disable large chunks of the language in order to keep it secur...

i disagree

#

the primary tool would be recursion

#

any time a code object is constructed, recurse onto it to check the bytecode

rugged sparrow Feb 1, 2023, 7:22 PM

#

that doesnt take into account other methods of code exec

versed eagle Feb 1, 2023, 7:22 PM

#

and then do the same for other such events that allow things that you don't want to allow

versed eagle Feb 1, 2023, 7:22 PM

#

rugged sparrow that doesnt take into account other methods of code exec

such as?

rugged sparrow Feb 1, 2023, 7:22 PM

#

at the assembly level

#

memory corruption -> code exec

versed eagle Feb 1, 2023, 7:24 PM

#

how would you achieve memory corruption then?

rugged sparrow Feb 1, 2023, 7:25 PM

#

race conditions with the garbage collector, abuse assumptions made in some builtin objects, use one of the bugs that have been patched in bypassable ways

versed eagle Feb 1, 2023, 7:25 PM

#

can you provide an example

rugged sparrow Feb 1, 2023, 7:27 PM

#

!e ```py
bytearray_mem = memoryview(bytearray(bytearray.basicsize)).cast('P')
bytearray_mem[0] = 1 # refcount
bytearray_mem[1] = id(bytearray) # ob_type
bytearray_mem[2] = (2 ** (tuple.itemsize * 8) - 1) // 2 # ob_size
bytearray_mem = bytearray_mem.tobytes()

class Fake:
slots = ['value']
def del(self):
Fake.resurrect = self # this causes a leak at close, but without it we crash hard in gc
Fake.fake = self.value

Fake_mem = memoryview(bytearray(Fake.basicsize + (tuple.itemsize * 2))).cast('P') # add header otherwise GC complains
Fake_mem[2] = 1
Fake_mem[3] = id(Fake)
Fake_mem[4] = id(bytearray_mem) + bytes.basicsize - 1
Fake_mem = Fake_mem.tobytes()
bytes_mem = memoryview(bytearray(property.basicsize - (bytes.basicsize - 1))).cast('P')
for i in range(len(bytes_mem)):
bytes_mem[i] = id(Fake_mem) + bytes.basicsize - 1 + (tuple.itemsize * 2)

bytes_mem = bytes_mem.tobytes()

class Bug(property):
def new(self, *args, **kwargs):
return bytes_mem

p = property.new(Bug)
ns = p.getter(lambda:0)

mem = memoryview(Fake.fake)

print(mem, len(mem))``` this is one i sent a bit ago, it gets a memoryview over the entire process address space (0 -> PY_SSIZE_T_MAX)

night quarryBOT Feb 1, 2023, 7:27 PM

#

@rugged sparrow :white_check_mark: Your 3.11 eval job has completed with return code 0.

<memory at 0x7fda32538a00> 9223372036854775807

rugged sparrow Feb 1, 2023, 7:27 PM

#

@versed eagle ^ that one exploits property

#

so no imports, and no code objects

#

and nothing that looks inherently malicious

versed eagle Feb 1, 2023, 7:28 PM

#

a way to stop that would be

#

replace any pointer to a c function within builtin_function_or_method with a pointer to your own function that validates inputs first

#

lots of work, but possible, no?

rugged sparrow Feb 1, 2023, 7:30 PM

#

versed eagle replace any pointer to a c function within `builtin_function_or_method` with a p...

how would you validate

#

it would be difficult (near impossible) to validate successfully while also detecting every bug

#

and at that point just patch them in python, but more will always pop up

versed eagle Feb 1, 2023, 7:33 PM

#

rugged sparrow !e ```py bytearray_mem = memoryview(bytearray(bytearray.__basicsize__)).cast('P'...

at least for that bug specifically, could you not validate the addresses that memoryview reads/writes to before it reads/writes to them

rugged sparrow Feb 1, 2023, 7:33 PM

#

how do you validate that

#

what determines whether a memoryview should have access to a given range

rugged sparrow Feb 1, 2023, 7:35 PM

#

versed eagle at least for that bug specifically, could you not validate the addresses that `m...

also, you don't need to make the memoryview over the entire space. It could just be over 8 bytes that happen to be a function pointer, then you can replace that pointer with the entry point to some form of shell code (ROP, JOP, whatever)

versed eagle Feb 1, 2023, 7:37 PM

#

check the size of the object that it's meant to be viewing
if ptr > addr_of_object && ptr < addr_of_object + size_of_object then it's in the correct range and dereferencing can be allowed

rugged sparrow Feb 1, 2023, 7:38 PM

#

in that example, I construct the bytearray manually, with fully controlled arguments. I control ptr and size_of_object

versed eagle Feb 1, 2023, 7:39 PM

#

even if you can control size_of_object, how would you control ptr?

rugged sparrow Feb 1, 2023, 7:40 PM

#

versed eagle even if you can control `size_of_object`, how would you control `ptr`?

i build the bytearray, so i control the address of its buffer

#

bytearray_mem = memoryview(bytearray(bytearray.__basicsize__)).cast('P')
bytearray_mem[0] = 1 # refcount
bytearray_mem[1] = id(bytearray) # ob_type
bytearray_mem[2] = (2 ** (tuple.__itemsize__ * 8) - 1) // 2 # ob_size
bytearray_mem = bytearray_mem.tobytes()``` @versed eagle these lines build the backing memory of the bytearray

#

i never set the address because it starts at 0

versed eagle Feb 1, 2023, 7:43 PM

#

so, disallow interpreting a user modified memoryview as a PyObject*

#

char* only

rugged sparrow Feb 1, 2023, 7:44 PM

#

how could you do that? I never modified the memoryview directly

burnt pasture Feb 1, 2023, 7:46 PM

#

@rugged sparrow sorry, i missed all this. in the blog post, getattr is removed because __builtins__ is empty.

versed eagle Feb 1, 2023, 7:46 PM

#

rugged sparrow ```py bytearray_mem = memoryview(bytearray(bytearray.__basicsize__)).cast('P') b...

this is you modifying it, no?

rugged sparrow Feb 1, 2023, 7:46 PM

#

so you would fully disable memoryview's in your "sandboxed" python?

versed eagle Feb 1, 2023, 7:47 PM

#

no
you can build memoryviews, read from them, etc

#

just not write to them

versed eagle Feb 1, 2023, 7:47 PM

#

rugged sparrow so you would fully disable memoryview's in your "sandboxed" python?

i wouldnt call it sandboxed

rugged sparrow Feb 1, 2023, 7:47 PM

#

versed eagle just not write to them

i think that could be considered making them useless lol

rugged sparrow Feb 1, 2023, 7:48 PM

#

versed eagle no you can build memoryviews, read from them, etc

also all of that code can be done without memoryviews

#

i just did it for ease of coding

versed eagle Feb 1, 2023, 7:48 PM

#

this is more like shooting python in the leg and then chaining it down

#

rather than sandboxing

rugged sparrow Feb 1, 2023, 7:48 PM

#

i could have done it with bytearrays

#

or bytes

#

or strings

versed eagle Feb 1, 2023, 7:48 PM

#

strings aren't meant to be mutable anyway

rugged sparrow Feb 1, 2023, 7:48 PM

#

i don't have to write into the object at all

#

i would just construct it inplace

versed eagle Feb 1, 2023, 7:49 PM

#

can you provide an example of that please?

rugged sparrow Feb 1, 2023, 7:49 PM

#

hell, you could create memory with the correct shape using range_iterator

#

b''.join([
  (1).to_bytes(8, 'little'),
  id(bytearray).to_bytes(8, 'little'),
  ...
])```

versed eagle Feb 1, 2023, 7:51 PM

#

and then you would construct a memoryview of that to get access to the process memory?

rugged sparrow Feb 1, 2023, 7:51 PM

#

nope, i only do that as an example

#

you could leverage the bug to write a function pointer directly if you wanted to

#

the bug itself is the problem, the rest is just using various primitives to get it to do what you want

#

and python (like any language) has tons of primitives to build memory

#

directly or indirectly

#

more specifically, that bug allows for a free of an arbitrary address, provided the first 8 bytes of that address equals 1

versed eagle Feb 1, 2023, 7:56 PM

#

so, disallow any addresses where the first 8 bytes = 1

rugged sparrow Feb 1, 2023, 7:57 PM

#

lol

#

thats all PyObjects with a refcount of one

versed eagle Feb 1, 2023, 7:57 PM

#

that seems simple to fix

versed eagle Feb 1, 2023, 7:57 PM

#

rugged sparrow thats all PyObjects with a refcount of one

alright, so increment all refcounts

rugged sparrow Feb 1, 2023, 7:57 PM

#

so now you never free memory

versed eagle Feb 1, 2023, 7:57 PM

#

garbage collect when refcount = 1

#

instead of 0

rugged sparrow Feb 1, 2023, 7:58 PM

#

if you add that hook, then now the bug triggers fine with address where the first 8 bytes equal 2

versed eagle Feb 1, 2023, 7:58 PM

#

disallow self reference in __del__

rugged sparrow Feb 1, 2023, 7:58 PM

#

the address that gets freed could have a pointer to a fake type

#

with a fake tp_del

#

and at that point you have arbitary code exec

versed eagle Feb 1, 2023, 7:59 PM

#

rugged sparrow the address that gets freed could have a pointer to a fake type

check if the types are Real before getting attributes

rugged sparrow Feb 1, 2023, 8:00 PM

#

how can you check if a type is real

versed eagle Feb 1, 2023, 8:00 PM

#

check if it's valid

rugged sparrow Feb 1, 2023, 8:00 PM

#

what defines valid

versed eagle Feb 1, 2023, 8:00 PM

#

no arbitrary code exec

rugged sparrow Feb 1, 2023, 8:00 PM

#

(fyi, you are not gonna win this one lmao)

versed eagle Feb 1, 2023, 8:01 PM

#

rugged sparrow (fyi, you are not gonna win this one lmao)

i dont think an actual implementation is possible

rugged sparrow Feb 1, 2023, 8:01 PM

#

it could be a valid type, but the instance is fake and holds a pointer that is freed by the types tp_free

versed eagle Feb 1, 2023, 8:01 PM

#

i think that its theoretically possible

#

given infinite time and computing power

rugged sparrow Feb 1, 2023, 8:02 PM

#

oh well with that small caveat

versed eagle Feb 1, 2023, 8:02 PM

#

versed eagle i think that its *theoretically* possible

it's the difference between testing that something is true for every single possible input and proving that something is true

rugged sparrow Feb 1, 2023, 8:03 PM

#

if you were to build this system, you would have to be able to prove that it is safe for every possible input

versed eagle Feb 1, 2023, 8:03 PM

#

e.g. in math, pythag theorem can be proven without testing every single possible input

versed eagle Feb 1, 2023, 8:03 PM

#

rugged sparrow if you were to build this system, you would have to be able to prove that it is ...

if you were to actually build it*

#

which im not talking about doing

rugged sparrow Feb 1, 2023, 8:03 PM

#

that is a very simple equation compared to complex computing

#

and there are many relatively straight forward math problems that cannot be proven

#

https://en.wikipedia.org/wiki/Millennium_Prize_Problems

Millennium Prize Problems

The Millennium Prize Problems are seven well-known complex mathematical problems selected by the Clay Mathematics Institute in 2000. The Clay Institute has pledged a US$1 million prize for the first correct solution to each problem.
The Clay Mathematics Institute officially designated the title Millennium Problem for the seven unsolved mathemat...

versed eagle Feb 1, 2023, 8:04 PM

#

rugged sparrow that is a very simple equation compared to complex computing

infinite time and computing power

versed eagle Feb 1, 2023, 8:04 PM

#

rugged sparrow and there are many relatively straight forward math problems that cannot be prov...

yeah ik

rugged sparrow Feb 1, 2023, 8:04 PM

#

versed eagle infinite time and computing power

the moment you say this, this discussion has no purpose

versed eagle Feb 1, 2023, 8:05 PM

#

how so?

rugged sparrow Feb 1, 2023, 8:05 PM

#

once you bring in infinite time, the theoretical system is useless

versed eagle Feb 1, 2023, 8:06 PM

#

again, how so

#

you were saying that it's not possible at all
im saying that it's possible, but not practically possible

rugged sparrow Feb 1, 2023, 8:06 PM

#

its not possible

#

not while allowing a full set of python syntax and the current implementation + spec

versed eagle Feb 1, 2023, 8:07 PM

#

i mean the whole point was to disallow underscores

#esoteric-python

define True 0

define False 1