esoteric-python

[night quarry]

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

(<built-in function iter>, (<class 'int'>, 0))

[rugged sparrow]

!e @fleet bridge found something else weird ```py
corrupt = iter(lambda:0, 0)

class Cstr:
def hash(self):
return hash('iter')
def eq(self, other):
[*corrupt]
return other == 'iter'

builtins = builtins.dict if hasattr(builtins, 'dict') else builtins
oiter = builtins['iter']
del builtins['iter']
builtins[Cstr()] = oiter
print(corrupt.reduce())```

this code should not trigger a segfault, it should pass a NULL into PyBuild_Value inside of callable_iter.__reduce__ and trigger a SystemError exception. But for some reason the PyCLEAR of iterable->seq_callable is not actually clearing the pointer?

[night quarry]

@rugged sparrow :x: Your 3.11 eval job has completed with return code 139 (SIGSEGV).

(<built-in function iter>, (<function  at 0x7f5c07b404a0>, 0))

[rugged sparrow]

so it is freeing the lambda but not clearing the pointer (which is a weird + big issue if it is a problem with PyCLEAR)

the extra weird bit is that I cannot trigger it on my machine, only on the bot

[astral rover]

how do you find these things?

[rugged sparrow]

I was trying to find a way to execute python code when _PyEval_GetBuiltin is called

And use that to clear some attributes before they are read so that it would pass NULLs to to PyBuild_Value

[serene garnet]

hey guys

sombody can understand my code?

for x,i in enumerate(n:=list(input()[::-1])):
    if x!=len(n)-1:i=int(i);n[x]='0';n[x+1]=[n[x+1],str(int(n[x+1])+1)][i>4]
print(*n[::-1],sep='')

[gleaming linden]

Increment?

Oh wait

[quartz wave]

round?

round to max digit?

carry-round to max digit?

[gleaming linden]

Yeah that's probably right

[quartz wave]

yep ```pycon

for x,i in enumerate(n:=list(input()[::-1])):
... if x!=len(n)-1:i=int(i);n[x]='0';n[x+1]=[n[x+1],str(int(n[x+1])+1)][i>4]
...
1445
print(*n[::-1],sep='')
2000

it took a minute to inspect and a few seconds to figure out what it actually does

not bad imo

[gleaming linden]

I can't read code on mobile

[gleaming linden]

r=lambda x:r(x[1:])if'4'==x[0]else['4']<x
a,*b=input()
print(str(int(a)+r(b))+'0'*len(b))
``` golfed

[serene garnet]

wow

[serene garnet]

wait, but it doesn't work at 1-digit numbers.
it raises indexerror

[gleaming linden]

oops

[serene garnet]

how about this code:

n=int(input());d=1
while n>9:n+=5;n//=10;d*=10
print(n*d)

i found an algorithm

[gleaming linden]

oh that's pretty smart

[serene garnet]

yeah i googled it 🙂

[sly root]

I've computed the coefficients of the sin function by hand

This is more cursed than esoteric

The base formula:

Played a little with factorials of 5 and 7

Function:

sin = lambda x: (x-((x**3)/6)+((x**5)/118)-((x**7)/4482.4706564094))```

And its output is the same as output of math.sin

!e ```py
import math

sin = lambda x: (x-((x3)/6)+((x5)/118)-((x**7)/4482.4706564094))

print(math.sin(2))
print(sin(2))```

[night quarry]

@sly root :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | 0.9092974268256817
002 | 0.9092974268256817

[gleaming linden]

!rule english

[night quarry]

Rules

4. Use English to the best of your ability. Be polite if someone speaks English imperfectly.

[subtle herald]

!pastebin

[finite blaze]

How fast is it in comparison with math.sin() ?

[turbid dragon]

!e

import timeit

sin = lambda x: (x-((x**3)/6)+((x**5)/118)-((x**7)/4482.4706564094))

print(timeit.timeit("sin(100000)", setup="from __main__ import sin"))

print(timeit.timeit("sin(100000)", setup="from math import sin"))

[night quarry]

@turbid dragon :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | 0.6931726009934209
002 | 0.07194424699991941

[finite blaze]

any ideas why math.sin is faster?

[wide crow]

Because it's written in C.

[finite blaze]

ah fair

I wonder how it looks in C

[wide crow]

What's more it works for arguments other than 2.

[astral rover]

it basically just is a wrapper around math.h's sin iirc

[wide crow]

Try this:

#include<stdio.h>
#include<math.h>

double factorial(int);
void calc(float, float*);

int main()
{
    int x;
    float radian, result = 0;
    
    printf("Enter value of x in degrees\n");
    scanf("%d", &x);
    
    radian = x * (3.1415926 / 180.0);  // convert degrees to radians
    
    calc(radian, &result);
    
    printf("sin(%d) = %f\n", x, result);
    
    return 0;
}

void calc(float num, float *res)
{
    int count, n = 1, sign = 1;
    
    for(count = 1; (n <= 10); count += 2)
    {
        *res  +=  sign * (pow(num, count) / factorial(count));
        n += 1;
        sign *= -1;
    }
}

double factorial(int num)
{
    int count;
    double prod = 1;
    
    for(count = 1; count <= num; count++)
    {
        prod *= count;
    }
    return prod;
}

[astral rover]

but what if i go and pass 10pi radians?

[wide crow]

Well it does not allow for that. This was an answer to an earlier question about a an implementation posted even earlier. It answers that question. I just did this for a laugh because someone wanted to see it in C. It's a very long time since I played with all that unnecessarily complicated C stuff.

[low lynx]

typically sin and cos Taylor series approximations would shift it down to (-pi/2, pi/2) first I think

[vast basin]

[low lynx]

import numpy as np
def key(keÿ):
 kèÿ,kéy,kéÿ,kêy,*kèy=0,0,*keÿ.shape
 while kèÿ<kéÿ and kéy<kêy:
  try:kêÿ=kèÿ+keÿ[kèÿ:,kéy].nonzero()[0][0]
  except IndexError:kéy+=1;continue
  if kêÿ!=kèÿ:keÿ[[kèÿ,kêÿ]]=keÿ[[kêÿ,kèÿ]]
  keÿ[kèÿ]/=keÿ[kèÿ,kéy];keÿ[-~kèÿ:,kéy:]-=keÿ[-~kèÿ:,kéy:-~kéy]*keÿ[kèÿ,kéy:];kèy+=[(kèÿ,kéy)];kèÿ+=1;kéy+=1
 for këy,këÿ in kèy[::-1]:keÿ[:këy,këÿ:]-=keÿ[:këy,këÿ:-~këÿ]*keÿ[këy,këÿ:]
 return keÿ, kèy

key

[quartz wave]

what does this do exactly ```py
import numpy as np

def func(A):
a, b, c, d, *e = 0, 0, *A.shape
while a<c and b<d:
try:
B = a + A[a:, b].nonzero()[0][0]
except IndexError:
b += 1
continue
if B!=a:
A[[a,B]] = A[[B,a]]
A[a] /= A[a,b]
A[a+1:, b:] -= A[a+1:, b:b+1] * A[a, b:]
e += [(a,b)]
a += 1
b += 1
for x, y in e[::-1]:
A[:x, y:] -= A[:x, y:-~y] * A[x, y:]
return A, e

[low lynx]

wait i forgot i removed the copy bit

!e

import numpy as np
def key(keÿ):
 kèÿ,kéy,kéÿ,kêy,*kèy=0,0,*keÿ.shape
 while kèÿ<kéÿ and kéy<kêy:
  try:kêÿ=kèÿ+keÿ[kèÿ:,kéy].nonzero()[0][0]
  except IndexError:kéy+=1;continue
  if kêÿ!=kèÿ:keÿ[[kèÿ,kêÿ]]=keÿ[[kêÿ,kèÿ]]
  keÿ[kèÿ]/=keÿ[kèÿ,kéy];keÿ[-~kèÿ:,kéy:]-=keÿ[-~kèÿ:,kéy:-~kéy]*keÿ[kèÿ,kéy:];kèy+=[(kèÿ,kéy)];kèÿ+=1;kéy+=1
 for këy,këÿ in kèy[::-1]:keÿ[:këy,këÿ:]-=keÿ[:këy,këÿ:-~këÿ]*keÿ[këy,këÿ:]
 return keÿ, kèy

arr1 = np.arange(9, dtype=float).reshape((3, 3))
arr2 = np.arange(12, dtype=float).reshape((3, 4))
arr3 = arr1.copy()
arr3[0, 0] = -1
print(arr1)
print(key(arr1))
print(arr2)
print(key(arr2))
print(arr3)
print(key(arr3))

[night quarry]

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

001 | [[0. 1. 2.]
002 |  [3. 4. 5.]
003 |  [6. 7. 8.]]
004 | (array([[ 1.,  0., -1.],
005 |        [ 0.,  1.,  2.],
006 |        [ 0.,  0.,  0.]]), [(0, 0), (1, 1)])
007 | [[ 0.  1.  2.  3.]
008 |  [ 4.  5.  6.  7.]
009 |  [ 8.  9. 10. 11.]]
010 | (array([[ 1.,  0., -1., -2.],
011 |        [ 0.,  1.,  2.,  3.],
... (truncated - too many lines)

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

[vast basin]

I assume this is what you sent me in Discord earlier. I'll no longer be soliciting advice from individuals performing summoning of the old gods with their code.

[low lynx]

it's just linalg tho

[vast basin]

pandas does nested for loops to address a nested list directly by index: Blasphemy.
Stickie uses 20 different accents on the same word in one valid output: #esoteric-python

Maybe I'll just hide in here and claim that my musings are just deliberately esoteric.

[low lynx]

ooh you just gave me a good idea

[vast basin]

oh no

[low lynx]

!e

from einspect import impl
from itertools import product

@impl(int)
def __matmul__(self, other):
    return int(not(self&other)) # nand gate

def full_adder(è, é, ê):
  return è@é@(ê@(è@é@è@(è@é@é))),è@é@è@(è@é@é)@ê@(è@é@è@(è@é@é))@(ê@(ê@(è@é@è@(è@é@é))))

for i, j, k in product(range(2), repeat=3):
    print(i, j, k, full_adder(i, j, k))

[night quarry]

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

001 | 0 0 0 (0, 0)
002 | 0 0 1 (0, 1)
003 | 0 1 0 (0, 1)
004 | 0 1 1 (1, 0)
005 | 1 0 0 (0, 1)
006 | 1 0 1 (1, 0)
007 | 1 1 0 (1, 0)
008 | 1 1 1 (1, 1)

[low lynx]

it's now even more unreadable

[vast basin]

When Sir Bankston Binary the Third created Binary back in 927 AD on the sacred tablets, I think he would be horrified to see what humanity has done.

[sly root]

Found another method
sin(x) = e^(X*1j)

[night quarry]

@sly root :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | 0.2383187820087187
002 | 0.07195955800125375

[sly root]

!e ```py
import timeit

print(timeit.timeit("(2.718281828459045**(100000*1j)).imag"))

print(timeit.timeit("sin(100000)", setup="from math import sin"))

[night quarry]

@sly root :white_check_mark: Your 3.10 eval job has completed with return code 0.

001 | 0.06613369201659225
002 | 0.08648197798174806

[sly root]

.

[magic dirge]

Hey esoteric people!

You guys seem like the people to talk to about micro optimizations. Not that such a thing should ever be entertained

But out of curiosity

https://hastebin.com/share/penerufeta.py

What might I do so squeeze a little more out of this?

[quartz wave]

like 186 to 192 ```py
match self.observe():

            case '\n':
                yield from self.consume() # ignore

            case '\r':
                yield from self.consume()  # ignore

this can just bepy
match self.observe():

            case '\n' | '\r':
                yield from self.consume() # ignore

[magic dirge]

Will that improve performance?

I figured it would have desugared into the same two checks in the bytecode

[quartz wave]

it'll certainly avoid unnecessary code duplication

[magic dirge]

Derp. Totally true

[quartz wave]

but the duplicated code should be equivalent in performance to the code if not duplicated

numerals is redundant here in lines 250 to 257 ```py
numerals = {
'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', 'P',
'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', 'a', 'b', 'c', 'd', 'e', 'f',
'g', 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', 's', 't', 'u', 'v',
'w', 'x', 'y', 'z', '0', '1', '2', '3', '4', '5', '6', '7', '8', '9'
}

        return numerals, 'ε', 'ι'

other than that i have nothing else to suggest

[magic dirge]

I guess that's a good thing! I really have taken pains in the interest of performance

I'm hoping that once converted to Cython, and with the help of some clever inlining perhaps, it'll be lightning fast

Not that lexer's need to be all that fast. But it's good to at least try

My only real concern is the chaining of yield froms. I don't really understand how all that works under the hood, but I wouldn't be surprised if it comes with some non-negligible overhead

XD It's that damn newline tokenizer. It's the only one that yields more than one token in a single pass. On the other hand, there are a few that don't yield anything, so I suppose generator mechanics clean that mess up rather nicely

[fleet bridge]

!e ```py
k=6
s='E6 84 9B E3 81 97 E3 81 A6 E3 82 8B E3 82 88 E5 85 88 E8 BC A9'
i=[iter(s:=''.join(map(lambda x:x[2:],map(bin,bytes.fromhex(s)))))]*k
n=len(s)//k
r=[*map(int,map(''.join,map(map,[next]*n,[i]*n)),[2]*n)]
print(r)

[night quarry]

@fleet bridge :white_check_mark: Your 3.11 eval job has completed with return code 0.

[57, 40, 18, 27, 56, 56, 6, 23, 56, 56, 6, 38, 56, 56, 10, 11, 56, 56, 10, 8, 57, 24, 22, 8, 58, 11, 50, 41]

[fleet bridge]

i love this part: [*map(int,map(''.join,map(map,[next]*n,[i]*n)),[2]*n)]

[low lynx]

map(map, ...) is cursed lmao

[finite blaze]

what is going on here

[gleaming linden]

[*map(int,  # convert base 2 to ints
    map(''.join,  # join string 
        map(
            map,  # get an item from i
            [next]*n,
            [i]*n
        )
    ),
    [2]*n
)]

So basically just ```py
r = [int(''.join(j), 2) for j in i]

[long fulcrum]

nope

[versed eagle]

!e

*map(print,(1,2,3))

[night quarry]

@versed eagle :x: Your 3.11 eval job has completed with return code 1.

001 |   File "<string>", line 1
002 | SyntaxError: can't use starred expression here

[versed eagle]

¯_(ツ)_/¯

[split salmon]

!e ```py
x = ["SEGEMENTATION FAULT"]
for _ in range(100_000):
x = map(int, x)

list(x)```

[night quarry]

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

[No output]

[split salmon]

Clean way to exit

[twin reef]

This fees like a cpython bug

[split salmon]

It is

[earnest wing]

works on my machine

had to raise it up to 2^17 iterations

but yeah just seems like the implementation of map overflows its stack if recursively evaluating itself without checks

[split salmon]

I think it overflows into kernel space

[split salmon]

I discovered it

[split salmon]

#python-discussion message

[long fulcrum]

!e

*map(print, (1,2,3)),```

[night quarry]

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

001 | 1
002 | 2
003 | 3

[long fulcrum]

¯_(ツ)_/¯

[versed eagle]

i mean yeah
if you unpack it into a tuple

which is not the same thing as having it bare

[long fulcrum]

it's literally what you said wouldn't work

[versed eagle]

i said that *map(print,stuff) wouldn't work

[rugged owl]

[fleet bridge]

isnt it overflowing stack?
and it is happening in C so it cannot be checked?

>>> x = []
>>> for _ in range(100_000): x = zip(x)
...
>>> x
<zip object at 0x000001E37F6C5B40>
>>> list(x)
<crashes>

>>> x = []
>>> for _ in range(100_000): x = enumerate(x)
...
>>> list(x)
<crashes>

>>> x = []
>>> for _ in range(100_000): x = filter(bool, x)
...
>>> list(x)
<crashes>

>>> x = []
>>> for _ in range(100_000): x = reversed(x)
...
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: 'list_reverseiterator' object is not reversible
``` :-(

>>> x = []
>>> for _ in range(100_000): x = iter(x)
...
>>> list(x)
[]
``` this works because `iter(<iterator>)` returns iterator itself

[split salmon]

Yea

[split salmon]

Try a higer number (1M)

!e py x = ["wot?"] for _ in range(100_000): x = (_ for _ in x) list(x)

[night quarry]

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

[No output]

[split salmon]

Yea

Its for function that yield values

Like map, zip, enum, filter, genexp, etc

Apparently it is allocated when you call the list function

Hold on, multiprocessing uses processes, lets see what happens

from multiprocessing import Pool
def crash(a):
    print(a)
    for _ in range(100_000):
        a = (_ for _ in a) # func calls are expensive
    list(a)
with Pool(4) as p:
    p.map_unordered(crash, [0, 1, 2, 3])```

!e py from multiprocessing import Pool import time def crash(a): print(a) for _ in range(100_000): a = (_ for _ in a) # func calls are expensive time.sleep(0.5) list(a) with Pool(4) as p: p.map_unordered(crash, [0, 1, 2, 3])

[night quarry]

@split salmon :x: Your 3.11 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 7, in <module>
003 |   File "/usr/local/lib/python3.11/multiprocessing/context.py", line 119, in Pool
004 |     return Pool(processes, initializer, initargs, maxtasksperchild,
005 |            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
006 |   File "/usr/local/lib/python3.11/multiprocessing/pool.py", line 191, in __init__
007 |     self._setup_queues()
008 |   File "/usr/local/lib/python3.11/multiprocessing/pool.py", line 346, in _setup_queues
009 |     self._inqueue = self._ctx.SimpleQueue()
010 |                     ^^^^^^^^^^^^^^^^^^^^^^^
011 |   File "/usr/local/lib/python3.11/multiprocessing/context.py", line 113, in SimpleQueue
... (truncated - too many lines)

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

[split salmon]

Lol piping failed ig

[night quarry]

@split salmon :x: Your 3.11 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 9, in <module>
003 |   File "/usr/local/lib/python3.11/multiprocessing/context.py", line 119, in Pool
004 |     return Pool(processes, initializer, initargs, maxtasksperchild,
005 |            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
006 |   File "/usr/local/lib/python3.11/multiprocessing/pool.py", line 191, in __init__
007 |     self._setup_queues()
008 |   File "/usr/local/lib/python3.11/multiprocessing/pool.py", line 346, in _setup_queues
009 |     self._inqueue = self._ctx.SimpleQueue()
010 |                     ^^^^^^^^^^^^^^^^^^^^^^^
011 |   File "/usr/local/lib/python3.11/multiprocessing/context.py", line 113, in SimpleQueue
... (truncated - too many lines)

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

@split salmon :x: Your 3.11 eval job has completed with return code 1.

001 |   File "<string>", line 12
002 |     p.map_unordered(crash, [0, 1, 2, 3])
003 |                                         ^
004 | IndentationError: unindent does not match any outer indentation level

[split salmon]

!e py from multiprocessing import Pool import time def crash(a): print(a) for _ in range(100_000): a = (_ for _ in a) # func calls are expensive time.sleep(0.5) list(a) with Pool(4) as p: print("hi");p.map_unordered(crash, [0, 1, 2, 3])

[night quarry]

@split salmon :x: Your 3.11 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 9, in <module>
003 |   File "/usr/local/lib/python3.11/multiprocessing/context.py", line 119, in Pool
004 |     return Pool(processes, initializer, initargs, maxtasksperchild,
005 |            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
006 |   File "/usr/local/lib/python3.11/multiprocessing/pool.py", line 191, in __init__
007 |     self._setup_queues()
008 |   File "/usr/local/lib/python3.11/multiprocessing/pool.py", line 346, in _setup_queues
009 |     self._inqueue = self._ctx.SimpleQueue()
010 |                     ^^^^^^^^^^^^^^^^^^^^^^^
011 |   File "/usr/local/lib/python3.11/multiprocessing/context.py", line 113, in SimpleQueue
... (truncated - too many lines)

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

[split salmon]

!e py from multiprocessing import Pool import time def crash(a): print(a) for _ in range(100_000): a = (_ for _ in a) # func calls are expensive time.sleep(0.5) list(a) with Pool(4) as p: print("hi");p.map_unordered(crash, [0, 1, 2, 3])

[night quarry]

@split salmon :x: Your 3.11 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 9, in <module>
003 |   File "/usr/local/lib/python3.11/multiprocessing/context.py", line 119, in Pool
004 |     return Pool(processes, initializer, initargs, maxtasksperchild,
005 |            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
006 |   File "/usr/local/lib/python3.11/multiprocessing/pool.py", line 191, in __init__
007 |     self._setup_queues()
008 |   File "/usr/local/lib/python3.11/multiprocessing/pool.py", line 346, in _setup_queues
009 |     self._inqueue = self._ctx.SimpleQueue()
010 |                     ^^^^^^^^^^^^^^^^^^^^^^^
011 |   File "/usr/local/lib/python3.11/multiprocessing/context.py", line 113, in SimpleQueue
... (truncated - too many lines)

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

[split salmon]

!e py from multiprocessing import Pool import time def crash(a): print(a) for _ in range(100_000): a = (_ for _ in a) # func calls are expensive time.sleep(0.5) list(a) with Pool() as p: print("hi");p.map_unordered(crash, [0, 1, 2, 3])

[night quarry]

@split salmon :x: Your 3.11 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 9, in <module>
003 |   File "/usr/local/lib/python3.11/multiprocessing/context.py", line 119, in Pool
004 |     return Pool(processes, initializer, initargs, maxtasksperchild,
005 |            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
006 |   File "/usr/local/lib/python3.11/multiprocessing/pool.py", line 191, in __init__
007 |     self._setup_queues()
008 |   File "/usr/local/lib/python3.11/multiprocessing/pool.py", line 346, in _setup_queues
009 |     self._inqueue = self._ctx.SimpleQueue()
010 |                     ^^^^^^^^^^^^^^^^^^^^^^^
011 |   File "/usr/local/lib/python3.11/multiprocessing/context.py", line 113, in SimpleQueue
... (truncated - too many lines)

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

@split salmon :white_check_mark: Your 3.11 eval job has completed with return code 0.

['usr', 'snekbox', 'lib64', 'lib', 'etc']

[split salmon]

There is no dev shm

[night quarry]

@split salmon :x: Your 3.11 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 2, in <module>
003 | OSError: [Errno 30] Read-only file system: '/dev'

@split salmon :x: Your 3.11 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 2, in <module>
003 | OSError: [Errno 30] Read-only file system: '/tmp'

[sick hound]

exit = lambda:(l:=(p:=__import__("ctypes").cdll.LoadLibrary("libc.so.6")).malloc(16)) and p.free(l)

segfault in python speedrun

[quartz wave]

(linux-only)

[sick hound]

doesn't this work with mingw?

[quartz wave]

general windows python, no it doesn't

[sick hound]

sad

my approach works on Windows too:

exit = lambda:__import__("ctypes").c_char_p.from_address(0)

[quartz wave]

51 ```py
import ctypes;f=lambda:ctypes.c_int.from_address(0)

[fleet bridge]

import ctypes;exit=lambda:ctypes.c_int.from_address(0) # 54
import ctypes as _;exit=lambda:_.c_int.from_address(0) # 54
from ctypes import c_int as _;exit=_.from_address # 49
import ctypes as _;exit=_.c_int.from_address # 44

from os import _exit as exit # 28

[quartz wave]

exit needs a status though

and it doesn't "segfault" like the .from_address() stuff do

[fleet bridge]

piping is a funny word

[rugged sparrow]

!e ```py
import ctypes;f=lambda:ctypes.string_at(0)

f()

[night quarry]

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

[No output]

[dry mirage]

what about shortest segfault without importing ctypes 👀

[fleet bridge]

somthing like this

deep C-level recursion

!e x=[];[*[(x:=zip(x))for _ in [0]*100000][-1]]

[night quarry]

@fleet bridge :warning: Your 3.11 eval job has completed with return code 0.

[No output]

[fleet bridge]

🧐

it crashes on my pc

[sick hound]

!e 42 ```py
x=[];[*[(x:=zip(x))for _ in[0]*10**5][-1]]

[night quarry]

@sick hound :warning: Your 3.11 eval job has completed with return code 0.

[No output]

[versed eagle]

if you're going to do it that way, normal loop is shorter
36

x=[]
for _ in[0]*10**6:x=zip(x)
[*x]

though it doesn't segfault either way

so, a bit pointless

[earnest wing]

+3 using map(id, instead which should segfault

[long fulcrum]

can you do 10e6 there or does it have to be an int?

[sick hound]

there must be an int

[low lynx]

is 9^9 greater than 10^6?

[versed eagle]

yes

>>> 10^6
1000000
>>> 9^9
387420489

[low lynx]

35 then

[fleet bridge]

why 10e6 if you can do 1e9 ?

[versed eagle]

cant use that in range though

dw chilaxan

that segfaulted on my machine

its just the bot being weird

[rugged sparrow]

x=[];[[x:=zip(x)for()in[x]*6**7],*x]
``` weird

[sick hound]

Chatgpt made rock paper scissors and guess my number in as few characters as possible, what do you guys think?

(These are all in 1 line of code)

a,b=input().lower(),input().lower();print("D"if a==b else"1W"if a+b in("rockpaperscissors","scissorspaperrock","paperscissorsrock")else"2W")

import random as r;a,g=r.randint(1,100),0;exec("g=int(input());print('LHW'[a>g::2][a<g::-2])\nwhile a!=g:")

[last locust]

I mean, it's definitely not the shortest possible

[sick hound]

How come

I don’t even understand it that well tbh

[versed eagle]

also using exec is kinda cheating

since using a \n in an exec

gets around 1 lining

[versed eagle]

also also that doesn't even work

[sick hound]

😂

Asked chatgpt so many times if it was sure it would work too

[gleaming linden]

~~```py
a,b=('rockpaperscissors'.find(input().lower())for _ in(0,1));print('LWD'[(a+b==9)-(a>=b)+(a==b)])

```py
a,b=('rockpaperscissors'.find(input().lower())for _ in(0,1));print([f'{[a<b,a>b][a+b==9]+1}W','D'][a==b])
``` 106 vs 141, still shorter

[fleet bridge]

doesnt even work because of SyntaxError

[low lynx]

I think you could remove like scissors and do functionality with -1

[gleaming linden]

oh yeah

[low lynx]

75 chars

a,b=('rp'.find(input().lower()[0])for _ in'__');print('D12 WW'[(a-b)%3::3])

assuming a few things about the input

[quartz wave]

66 chars ```py
print('D12 WW'[eval("'rp'.find(input().lower()[0])-"*2+'0')%3::3])

[dry mirage]

!e no ctypes

import gc

def g():
    v = object()
    yield v
    next(x for x in gc.get_referrers(v) if isinstance(x, tuple))[-1]
    
tuple(g())

[night quarry]

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

[No output]

[sick hound]

!e ```py
import gc
def g():
v=object()
yield v
next(x for x in gc.get_referrers(v)if isinstance(x,tuple))[-1]
tuple(g())

[night quarry]

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

[No output]

[sick hound]

this is the shortest it can get (as done by me)

this is 114, including newlines

[astral rover]

i have a good way to do it, import a module compiled with mypyc :)

[arctic skiff]

def g():
 v=object()
 yield v
 next(x for x in __import__('gc').get_referrers(v)if isinstance(x,tuple))[-1]
tuple(g())

[sick hound]

length is 118, 4 longer than mine

[arctic skiff]

https://www.nas.nasa.gov/hecc/support/kb/common-causes-of-segmentation-faults-(segfaults)_524.html why does nasa have this type of support?

!epy (x:=[],[x:=map(int,x) for _ in range(100000)],list(x))

[night quarry]

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

[No output]

[arctic skiff]

!epy (x:=(),[x:=map(int,x) for _ in range(100000)],[*x])

[night quarry]

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

[No output]

[arctic skiff]

!e```py
print(len('(x:=(),[x:=map(int,x) for _ in range(100000)],[*x])'))

[night quarry]

@arctic skiff :white_check_mark: Your 3.11 eval job has completed with return code 0.

51

[arctic skiff]

its @split salmon's method anyways

i just made it in 1 line

[tranquil ermine]

esoteric Python? I'm ready to release my skill of using assembly language though .dll in my projects

            mov rax, offset point
point: lock jmp rax

[pure dew]

don't think you're in the right spot

[versed eagle]

dll. that's more of a #c-extensions type of thing

[rugged sparrow]

!e @dry mirage heres another weird way of raising SystemError ```py
def bug():
if bug.clear:
return 1
else:
bug.clear = True
list(bug.iterator)
return 0

bug.iterator = iter(bug, 1)
bug.clear = False
next(bug.iterator)

[night quarry]

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

001 | Traceback (most recent call last):
002 |   File "<string>", line 11, in <module>
003 | SystemError: Objects/object.c:722: bad argument to internal function

[meager zinc]

this can 100% be shorter

[dry mirage]

https://github.com/python/cpython/blob/3.11/Objects/object.c#L720-L724

[night quarry]

Objects/object.c lines 720 to 724

if (v == NULL || w == NULL) {
    if (!_PyErr_Occurred(tstate)) {
        PyErr_BadInternalCall();
    }
    return NULL;```

[rugged sparrow]

abuses this part of calliter_iternext ```c
static PyObject *
calliter_iternext(calliterobject *it)
{
PyObject *result;

if (it->it_callable == NULL) {
    return NULL;
}

result = _PyObject_CallNoArgs(it->it_callable); // clears the iterable inside this call
if (result != NULL) {
    int ok;

    ok = PyObject_RichCompareBool(it->it_sentinel, result, Py_EQ); // which means that `it->it_sentinel` is NULL here
    if (ok == 0) {
        return result; /* Common case, fast path */
    }

[dry mirage]

722 is rich compare?

[rugged sparrow]

for whatever version of 3.11 the bot is running

[dry mirage]

so it_sentinel got set to NULL here?

[rugged sparrow]

yea, because bug exhausts its own iterator

[dry mirage]

amazing

I guess this just needs another check of it->it_sentinel == NULL after the pyobject call?

[rugged sparrow]

store it->it_sentinel (dangerous, need to keep track of ref counts) or that ^

[dry mirage]

what should even happen here

StopIteration?

[rugged sparrow]

yea you can probably fall through to return NULL

*i think

[meager zinc]

!e -2 chars ```py
(x:=(),[x:=map(int,x)for _ in range(99999)],[*x])

[night quarry]

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

[No output]

[meager zinc]

!e ```py
def b():
if b.c:return 1
b.c=1
[*b.i]
b.i=iter(b,1)
b.c=0
next(b.i)

[night quarry]

@meager zinc :x: Your 3.11 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 7, in <module>
003 | SystemError: Objects/object.c:722: bad argument to internal function

[rugged sparrow]

yea that works too, I didn't want to golf mine so it was more clear where the bug comes from

[meager zinc]

unfortunately you cannot use walrus on attributes

or it would be shorter

is it possible to un-inherit from a class during runtime?

yeah

import ctypes
{ "NULL": ctypes.c_voidp.from_address(0) }

[earnest wing]

it's a null pointer wrapped in a regular python object pointer

it's perfectly fine

oops i was thinking of something else

[rugged sparrow]

You can change __class__ in some cases

[rugged sparrow]

You can make .i and .c globals instead

[earnest wing]

ctypes.POINTER(ctypes.c_int)() creates a null int pointer @half escarp

that's access checked by ctypes

a.k.a doesn't immediately halt

[meager zinc]

This is probably way too long, but I tried using global

def b():global i,c;return 1if c else((c:=1)&0)*[*i]
c=0
next(i:=iter(b,1))

[rugged sparrow]

thats clever

!e ```py
*(i:=iter(lambda x=[1]:x and[x.pop(),*i],[])),

[night quarry]

@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 | SystemError: Objects/object.c:722: bad argument to internal function

[earnest wing]

https://docs.python.org/3/library/ctypes.html#ctypes._CData.from_address

creates a void* from the data at address 0 from my understanding there

[meager zinc]

One doesn't have safety around it (creates data from the pointer) and gets dereferenced by print(), the other is technically a ctypes wrapper around it

pretty much if you do anything with c_voidp.from_address(0) you will segfault

[earnest wing]

because the void* gets used as a PyObject* ?

[meager zinc]

I believe so

that's from the wrapper

lemme find the source

Looks like <NULL> comes from object.c:
https://github.com/python/cpython/blob/3.11/Objects/object.c#L413-L414

[night quarry]

Objects/object.c lines 413 to 414

if (v == NULL)
    return PyUnicode_FromString("<NULL>");```

[meager zinc]

looks like it's not a PyObject* because that would be handled by that code

so I'm not quite sure what's happening

There's three different things:

import ctypes
# Creates a void pointer from 0, segfault upon dereference
c_voidp.from_address(0)  # segfault
# Creates a `ctypes` POINTER safety-checked wrapper
ctypes.POINTER(ctypes.c_int)()  # <NULL>
# Creates a PyObject* that is handled in `object.c`
ctypes.py_object()  # py_object(<NULL>)

Looks like the first two are PyCData_Type and PyCPointer_Type
https://github.com/python/cpython/blob/main/Modules/_ctypes/_ctypes.c

One stores the pointer (so it's not safe), the other stores a pointer to the pointer (so it's safer)

what's pythonapi?

oh

Looks like that returns a pointer to a PyObject*

hmm

Looks like it's a PyObject* -> NULL

*not really null

it's a pointer to a PyObject with empty memory

so like 00000000000 in memory

No, I don't think so

Because v is a pointer to valid memory, it's just all zeroes

(or it might be garbage? not sure)

Actually it seems like it's garbage

So v (the PyObject* containing the PyListObject*) has random stuff (0101101011110101...) and it will show up as py_object(MEMORY_ADDRESS) e.g. py_object(17192621)

[old socket]

Is 00000000000 equivalent to 0?

[gleaming linden]

as a number, yes

[meager zinc]

Appears to use malloc in the definition

Ah nevermind

it does use calloc

This is the line: op->ob_item = (PyObject **) PyMem_Calloc(size, sizeof(PyObject *));

[old socket]

I meant memory context wise

[meager zinc]

Idk what a 0000000000000000 PyObject would actually be...? I would presume it would either trigger some error or be something like None

[gleaming linden]

I think it's null

[old socket]

Like, would 00000000000 point to the same memory as 0

Is what I was asking

[dry mirage]

it is null?

null *PyObject pointer is still null

that creates a new list with a sized allocation

and returns the PyObject pointer

PyList_New(1) would create a list object with size 1 allocation and return a pointer to that object

!e

from ctypes import *


PyList_New = pythonapi.PyList_New
PyList_New.restype = py_object
PyList_New.argtypes = [c_int]

ls = PyList_New(3)
print(ls)

[night quarry]

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

[<NULL>, <NULL>, <NULL>]

[meager zinc]

well the memory gets allocated and it doesn't point to 0, but it does technically count as NULL still.

[dry mirage]

null pointers just don't point to valid memory, it doesn't have to be 0

[meager zinc]

Ah i see

[arctic skiff]

!e ```py
print(len('(x:=(),[x:=map(int,x)for _ in range(99999)],[*x]) '))

[night quarry]

@arctic skiff :white_check_mark: Your 3.11 eval job has completed with return code 0.

50

[arctic skiff]

Its only 1 char short

[arctic skiff]

100% so you making this in 0 chars?

[quartz wave]

at this point that's just hardware changes

[dense nova]

!e

(x:=(),[x:=map(int,x)for _ in range(99999)],[*x])
print(x == 1)

[night quarry]

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

[No output]

[dense nova]

huh

someone explain

[arctic skiff]

-1 bits of ram is enough

[arctic skiff]

[*x] causes seg fault which crashes before executing print

[sick hound]

Or because it does too much recursion on the C side?

[earnest wing]

yes that seems to be the consensus

[sick hound]

bruh you talk in this channel, respect bruh

[indigo harbor]

This place needs to be alive xD

[sly ibex]

i need to practice wolf:(

[meager zinc]

golfing?

[sly ibex]

i was doing a easy exercise in codewars
the kata is

The length of your code should be less or equals to 44.

Your code should be one line.

The length of the reference solution is 37.```

@test.it("Basic Tests")
    def tests():
        test.assert_equals(f("HelloWorld"), 4)
        test.assert_equals(f("Codewars"), 7)
        test.assert_equals(f("X"), 0)
        test.assert_equals(f("findX"), 0)```

my code

def f(s):
    s = s[::-1]
    m = 0
    for x in s:
        if x.isupper():
            m = s.index(x)
            break
    return m```

[low lynx]

you want a golf for this?

[sly ibex]

the cata is about golf

it doesnt take if the code is not in one single line

[low lynx]

it's about one lining

not golf

[fleet bridge]

About golfing too

[sly ibex]

but the name of the cata is code gofl

😦

[fleet bridge]

Onelining is easy

[versed eagle]

does this work?

f=lambda s:[*map(str.isupper,s)].index(1)

im not sure if it does

[low lynx]

return next(filter(str.isupper,string[::-1]))

this is code golf so my use of filter doesn't count

don't smite me pls

[versed eagle]

oh wait

does it have to be reversed

[low lynx]

it said reversed string so i think so?

[sly ibex]

yeah reversed

[fleet bridge]

f=lambda s:[*map(str.isupper,s[::-1])].index(1)
``` easy fix

[versed eagle]

if it does, then this works

f=lambda s:[*map(str.isupper,s[::-1])].index(1)

[versed eagle]

ah
denball got to it first

[low lynx]

f=lambda s:next(filter(str.isupper,s[::-1]))

is shorter

[versed eagle]

thats nice

and probably a good deal faster as well

[low lynx]

next and reduce are the go-tos when you want to one line a for loop imo

for the cases it does work

[sly ibex]

what the heck with codewars he said this code is too long

[flint hollow]

probably not golfed very well, but

def f(s): return [*map(lambda x:x.isupper(), s[::-1])].index(True)
``` works

[sly ibex]

ohh wait

[versed eagle]

ah
you used the same approach we did previously

[low lynx]

lambda x:x.isupper() is just str.isupper tho

[versed eagle]

up here

[flint hollow]

me am blind apparently 🤦

[versed eagle]

its ok lol

[low lynx]

since you can call a class' method with just the class and manually pass in self

[versed eagle]

i didnt read half the problem

before i started

writing stuff

so i missed the reversed bit

[versed eagle]

im pretty sure this is the shortest it can be

[sly ibex]

however you can try it in
https://www.codewars.com/kata/62f8d0ac2b7cd50029dd834c/train/python if you want

[flint hollow]

is this even correct? it returns the letter, you're supposed to return the index

[versed eagle]

oh true

[versed eagle]

in which case we're back to here

[low lynx]

oh index

then yeah

[sly ibex]

there's one single error
47 should be less or equal to 44.

in reference to len

[versed eagle]

huh

[low lynx]

oh duh

wait

hmm

still not trivial nvm

oh wait we want index from back

i have this so far

f=lambda s:['@'<c<'['for c in s[::-1]].index(1)

but still 47

[versed eagle]

ooh

ordinals

[low lynx]

oh ofc

then yeah it's solved for 44 c

[flint hollow]

44? that's 43

[low lynx]

oh i had extra whitespace

also damn way better ideas in the solutions

[sly ibex]

unlock the solutions?

[low lynx]

yep i got a 34

[sly ibex]

what

what

[low lynx]

||```py
f=lambda s:s[-1]>'Z'and-~f(s[:-1])

[versed eagle]

wow

[sly ibex]

booom yeahh baby

[versed eagle]

that is
a good one

[flint hollow]

this returns False for X and findX instead of 0

[low lynx]

0==False so it passes the test

oh that one is the first one on the solutions lmao

[flint hollow]

ah it doesn't test strictly to 0

[low lynx]

i just glanced at it and saw recursion and was like "OHHHH"

well good thing to note for the future

when in doubt try recursion

[verbal spear]

Is regex allowed?

[low lynx]

regex where?

the problem above?

[verbal spear]

Finding first upper

[low lynx]

sure, but importing it is way more characters than it is worth

and it's probably longer even still

[quartz wave]

eh ```py
import re;f=lambda s:re.search('[A-Z]',s).start()
f=lambda s:next(i for i in range(len(s))if'@'<s<'[')

longer than what i thought of

but shorter than what you have

[low lynx]

you need to start searching from the back

also that's not the most optimized one

[quartz wave]

ok got it

yeah regex would suck here

[low lynx]

it'd be comparable if you didn't have to import

but you do

so not worth

[fresh helm]

how can i do this in one like, trying to capitalize the first letter without disturbing the others

a = ' '.join(input()).split()
a[0] = a[0].upper()
print(''.join(a))```

[quartz wave]

.capitalize()?

!e ```py
print("hi there".capitalize())

[night quarry]

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

Hi there

[quartz wave]

nvm

it disturbs the others

[gleaming linden]

.title() does a similar thing but that counts ' as separate words

There's a method in string somewhere that does that

!d string.capwords

[night quarry]

string.capwords

string.capwords(s, sep=None)```
Split the argument into words using [`str.split()`](https://docs.python.org/3/library/stdtypes.html#str.split "str.split"), capitalize each word using [`str.capitalize()`](https://docs.python.org/3/library/stdtypes.html#str.capitalize "str.capitalize"), and join the capitalized words using [`str.join()`](https://docs.python.org/3/library/stdtypes.html#str.join "str.join"). If the optional second argument *sep* is absent or `None`, runs of whitespace characters are replaced by a single space and leading and trailing whitespace are removed, otherwise *sep* is used to split and join the words.

[fresh helm]

its one word not multiple

so it disturbs la letters

[gleaming linden]

Oh wait I completely misread that

[quartz wave]

this does the job exactly how you want it to ```py
a=a[0].upper()+a[1:]

[fresh helm]

i need to take input tho

can i club all in one line>

[quartz wave]

full code ```py
a=input()
a=a[0].upper()+a[1:]

one line would look like ```py
a=(a:=input())[0].upper()+a[1:]

[fresh helm]

walrus isnt supported

keeps giving error

ty tho

[quartz wave]

get python 3.8+

[fresh helm]

3.11.1

idle still gobs me error

[quartz wave]

it shouldn't error then

it will run fine

[fresh helm]

wait reinstalling

👍

works

[paper shell]

can someone help me upload a text file in python and read it?

i just don’t know how to know the location and input it in code, the file is saved on my desktop

[fresh helm]

go to windows make it so that indexing searches all folders, after that search for ur file copy path paste it in ur input(if ur using windows)

[paper shell]

using a mac

and jupyter notebook

[fresh helm]

jus do ctrl n paste the contents of the text file and save as text file and input ur relative path

[paper shell]

import os

base_path = "/path/to/directory/holding/file/"
filename = "text.txt"
path_to_file = os.path.join(base_path, filename)
fd = open(path_to_file , 'r')

is that wrong

[fresh helm]

do we have an error?
we could do r"/path/to/..." instead of current one

[paper shell]

not a directory error

[fresh helm]

single / could cause an error

[paper shell]

also tried os.chdir('file:///Users/myname /Desktop/text.txt')

[fresh helm]

///?

[paper shell]

that gave a file not found error but the location was copy and pasted from the url that opened when i opened the file on google chrome

[pure dew]

yea that's a url not a file path

[paper shell]

how do i get the file path

i’m very new to coding so my questions are silly sorry

[fresh helm]

if u have a file explorer right click copy path?

[paper shell]

don’t have a file explorer?

it’s just a text file saved on my desktop

[fresh helm]

u can move the text file to current working dir or simply google how to copy path of a file in mac

[paper shell]

kk

[pure dew]

the path here would just be /Users/myname/Desktop/text.txt

[paper shell]

not a directory error

[pure dew]

so something like

from pathlib import Path
with Path("/Users/myname/desktop/text.txt").open() as f:
    # do stuff
    ...

[paper shell]

YES!

[pure dew]

also, this really isn't the channel for this question

[paper shell]

what channel

is better

[last locust]

#1035199133436354600 (see #❓｜how-to-get-help)

[tough willow]

No

[verbal spear]

U need r"<filename>" ?

[fresh helm]

yesh

[tough willow]

its only a problem with \ paths, there is no issues with /

[fresh helm]

ok sorry did / instead of \

[unique heath]

ihave returned
anyway what are we doing

[night quarry]

:incoming_envelope: :ok_hand: applied mute to @verbal spear until <t:1676556886:f> (10 minutes) (reason: discord_emojis rule: sent 225 emojis in 10s).

The <@&831776746206265384> have been alerted for review.

[verbal spear]

It was done with a oneliner and 77 characters (including spaces)

[verbal spear]

!e for z in"啕⪪䗑⯪坵⺺嵝㨮嵝⺺坵⯪䗑⪪啕":print(f'{ord(z):015b}'.translate({49:"⬜",48:"⬛"}))

[night quarry]

@verbal spear :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/yewijimama.txt?noredirect

[past plank]

this is why we typically count bytes instead of characters for codegolf

[restive void]

!e I tried this a while back and couldn't figure it out, don't know why..

class MatchMeta(type):
    def __instancecheck__(cls, inst):
        return cls.isinstance(inst)

def matching(fn):
    class cls(metaclass=MatchMeta):
        isinstance = fn
    return cls
# ---------
@matching
def even(number):
    return number % 2 == 0

for i in [3, 4, "wat%s"]:
    match i:
        case even():
            print(i, "is even")
        case _:
            print(i, "is not")

[night quarry]

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

001 | 3 is not
002 | 4 is even
003 | wat%s is not

[sonic birch]

Did you choose each symbol manually?

UDP: is it true that I potentially can draw any image with this method?

[verbal spear]

Nope... it was reverse
I had the pattern and translated to 0/1 to int to char

[dry mirage]

!e

from einspect import impl

@impl(int | str)
def is_even(self):
    return isinstance(self, int) and not self % 2

for x in (3, 4, "wat"):
    print(x, x.is_even())

[night quarry]

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

001 | 3 False
002 | 4 True
003 | wat False

[low lynx]

that wouldn't match against a type in a match stmt tho

[restive void]

Nice, I haven't really had time to look at einspect yet, but the | is a nice touch. But yeah, I specifically wanted to customize match-case behaviour.

[rugged sparrow]

thats a cool way of making match case utilities

[unique heath]

!e print(sum(i for i in range(1000)if i%3<1or i%5<1))

how do i golf this even more lmfao

[night quarry]

@unique heath :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | <string>:1: SyntaxWarning: invalid decimal literal
002 | 233168

[dry mirage]

!e

print(sum({*(r:=range(1000))[::3],*r[::5]}))

[night quarry]

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

233168

[unique heath]

awht

confosed

big fusion

[dry mirage]

you wanted every number divisible by either 3 or 5 from 0 to 1000

so that's also the union of the set of every third number from 0 to 1000 and every 5th number

[unique heath]

um ok

HOLY FUCK I JUST REALIZED AND I FEEL SO STUPID NOW

[sonic birch]

Your approach reminds me this lib https://github.com/LubomyrIvanitskiy/CrazyPython/tree/main/typespace

[fresh helm]

just a question why is this program faster than the other
slower:

s = input()
a =len(''.join(i for i in s if i.isupper()))
print(s.upper() if not a==len(s)*0.5 and not a<len(s)*0.5 else s.lower())

faster:

s = input()
print(s.upper() if not len(''.join(i for i in s if i.isupper()))==len(s)*0.5 and not len(''.join(i for i in s if i.isupper()))<len(s)*0.5 else s.lower())```

[sudden osprey]

What makes you say the second one is faster?

[fresh helm]

i used the time module

and it suggests so

[versed eagle]

have you tried using timeit?

also, if you're using input(), it'll be variant
unless you've set stdin to be something
but even then, time for system calls can vary quite a bit depending on what else is happening

[arctic skiff]

!ti py s = 'example' a =len(''.join(i for i in s if i.isupper())) print(s.upper() if not a==len(s)*0.5 and not a<len(s)*0.5 else s.lower())

[night quarry]

@arctic skiff :white_check_mark: Your 3.11 timeit job has completed with return code 0.

200000 loops, best of 5: 1.49 usec per loop

[arctic skiff]

!tipy s = 'example' print(s.upper() if not len(''.join(i for i in s if i.isupper()))==len(s)*0.5 and not len(''.join(i for i in s if i.isupper()))<len(s)*0.5 else s.lower())

@night quarry forgot to respond?

!ti py s = 'example' print(s.upper() if not len(''.join(i for i in s if i.isupper()))==len(s)*0.5 and not len(''.join(i for i in s if i.isupper()))<len(s)*0.5 else s.lower())

[night quarry]

@arctic skiff :white_check_mark: Your 3.11 timeit job has completed with return code 0.

100000 loops, best of 5: 2.22 usec per loop

[last locust]

You probably forgot the space between the !ti and the ` of the codeblock

[arctic skiff]

yeah

because !e don't require it

!epy ...

[night quarry]

@arctic skiff :warning: Your 3.11 eval job has completed with return code 0.

[No output]

[arctic skiff]

¯_(ツ)_/¯

[last locust]

!epy print("hi")

huh, cool

I guess it's added as an alias

[night quarry]

@last locust :white_check_mark: Your 3.11 eval job has completed with return code 0.

hi

[arctic skiff]

anyways 2nd one is slower

[last locust]

!src e

[night quarry]

Command: eval

Run Python code and get the results.

Source Code

Go to GitHub

bot/exts/utils/snekbox.py:490

[arctic skiff]

ti is also alias

!src ti

[night quarry]

Command: timeit

Profile Python Code to find execution time.

Source Code

Go to GitHub

bot/exts/utils/snekbox.py:526

[sudden osprey]

It's special cased in the error handler for !e, probably could for timeit too...

[versed eagle]

if you want to profile short snippets of python, timeit is generally a reliable way to go

rather than trying to make a profiler yourself

[last locust]

https://github.com/python-discord/bot/blob/main/bot/exts/backend/error_handler.py#L66-L71

[night quarry]

bot/exts/backend/error_handler.py lines 66 to 71

if isinstance(e, errors.CommandNotFound) and not getattr(ctx, "invoked_from_error_handler", False):
    if await self.try_silence(ctx):
        return
    if await self.try_run_eval(ctx):
        return
    await self.try_get_tag(ctx)  # Try to look for a tag with the command's name```

[last locust]

There's special handling for !eval (try_run_eval), but not !timeit

[arctic skiff]

oh

timeit needs justice...

[jaunty flame]

is there any one who can help me to begin python please

i'm new in this "world"

and i'm french

[arctic skiff]

#❓｜how-to-get-help

[fresh helm]

no just trying it, also its a module from standard library right?,can u help im getting some error

[arctic skiff]

globals().update({'add_ints':lambda *_:(__:=__import__('time').time(),[__import__('time').sleep(___)for ___ in _],int(__import__('time').time()-__))[-1]})```

[fresh helm]

ran it

same

error even after running

ok reinstalled, works now

[meager zinc]

any ideas for golfing this?

for n in range(101):print(n,all(n%i!=0for i in range(2,n)))

It prints the numbers from 0 to 100 and indicates whether each is prime (it fails for 1, but I couldn't find a short fix)

[low lynx]

n%i>0 is a quick char save

hmm

actually in fact n%i is enough by itself

!e

def is_prime(n):
    return n>1 and all(n%p for p in range(2, n))
num = ''.join(str(int(is_prime(n))) for n in range(101))
print(f'{int(num, 2):x}')

ok definitely not worth it lmao

[night quarry]

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

6a28a20a08a20828228220808

[low lynx]

still not worth

[meager zinc]

!e ```py
for n in range(101):print(n,all(n%i for i in range(2,n)))

[night quarry]

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

001 | 0 True
002 | 1 True
003 | 2 True
004 | 3 True
005 | 4 False
006 | 5 True
007 | 6 False
008 | 7 True
009 | 8 False
010 | 9 False
011 | 10 False
... (truncated - too many lines)

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

[quartz wave]

for n in range(101):print(n,n-1>0<all(n%i for i in range(2,n)))
``` is the shortest fix i can find

[low lynx]

!e

for n in range(101):print(n,bool(0x6a28a20a08a20828228220808>>n&1))

[night quarry]

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

001 | 0 False
002 | 1 False
003 | 2 False
004 | 3 True
005 | 4 False
006 | 5 False
007 | 6 False
008 | 7 False
009 | 8 False
010 | 9 False
011 | 10 False
... (truncated - too many lines)

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

[low lynx]

wait

oh I stored it backwards I'm bad

[meager zinc]

Smart but too long

[low lynx]

yeah

honestly if the bool call didn't exist

and it actually worked

[meager zinc]

I think you encoded it wrong perhaps

[low lynx]

yeah I encoded it backwards

[meager zinc]

It would be a 101 bit number which the max would be x20000000000000000000000000 which is unfortunate

unless we can store it in 10e10 format somehow

but it seems unlikely

[gleaming linden]

can't we do ```py
n=0x...
while n:print(bool(n%2));n//=2

[low lynx]

!e

n = 0x6a28a20a08a20828228220808
r = f'{n:b}'[::-1]
print(f'{int(r, 2):x}', int(r, 2))

[night quarry]

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

808220a20a08228828228a2b 39771395718504928067455191595

[meager zinc]

sure but the question is how long is it ?

[low lynx]

that's the actual one I think

[meager zinc]

it's shorter

actually it might be the same

[low lynx]

!e

for n in range(101):print(n,bool(0x808220a20a08228828228a2b>>n&1))

[night quarry]

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

001 | 0 True
002 | 1 True
003 | 2 False
004 | 3 True
005 | 4 False
006 | 5 True
007 | 6 False
008 | 7 False
009 | 8 False
010 | 9 True
011 | 10 False
... (truncated - too many lines)

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

[low lynx]

wait

[meager zinc]

0x6a28a20a08a20828228220808
0x808220a20a08228828228a2b

one shorter

[low lynx]

it's still wrong tho

[meager zinc]

hm

[versed eagle]

here's one that doesn't count 0 and 1 as prime

for n in range(101):print(n,len([i for i in range(1,n)if n%i<1])==1)

since they arent

[versed eagle]

its only a few chars longer than this

[low lynx]

!e

def is_prime(n):
    return n>1 and all(n%p for p in range(2, n))
num = ''.join(str(int(is_prime(n))) for n in range(100, -1, -1))
print(f'{int(num, 2):x}')

[night quarry]

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

20208828828208a20a08a28ac

[meager zinc]

lemme try something

[low lynx]

we need a counter

[gleaming linden]

ah ofc

[low lynx]

!e

for n in range(101):print(n,bool(0x20208828828208a20a08a28ac>>n&1))

[night quarry]

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

001 | 0 False
002 | 1 False
003 | 2 True
004 | 3 True
005 | 4 False
006 | 5 True
007 | 6 False
008 | 7 True
009 | 8 False
010 | 9 False
011 | 10 False
... (truncated - too many lines)

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

[meager zinc]

!e ```py
def is_prime(n):
return n>1 and all(n%p for p in range(2, n))
num = ''.join(str(int(is_prime(n))) for n in range(100, -1, -1))
new = ''.join('1' if 0x20208828828208a20a08a28ac >> n & 1 else '0' for n in range(100, -1, -1))
print(f'{num}\n{new}\n{new[::-1]}')

[night quarry]

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

001 | 00010000000100000100010000010100010000010100000100000100010100010000010100000100010100010100010101100
002 | 00010000000100000100010000010100010000010100000100000100010100010000010100000100010100010100010101100
003 | 00110101000101000101000100000101000001000101000100000100000101000001000101000001000100000100000001000

[meager zinc]

Yeah it provides the same numerical output

so idk

[low lynx]

actually wait

!e

for n in range(101):print(n,0x20208828828208a20a08a28ac>>n&1>0)

[night quarry]

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

001 | 0 False
002 | 1 False
003 | 2 True
004 | 3 True
005 | 4 False
006 | 5 True
007 | 6 False
008 | 7 True
009 | 8 False
010 | 9 False
011 | 10 False
... (truncated - too many lines)

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

[low lynx]

is that shorter than the current one?

[meager zinc]

Probably

[versed eagle]

yes

by 5 chars

[low lynx]

neat

[versed eagle]

well

its shorter than mine

idk if its shorter than

cereal's

[low lynx]

I'm on mobile someone else do the counting

[meager zinc]

for n in range(101):print(n,0x20208828828208a20a08a28ac>>n&1>0) #ties
for n in range(101):print(n,n-1>0<all(n%i for i in range(2,n))) #ties
for n in range(101):print(n,len([i for i in range(1,n)if n%i<1])==1)
for n in range(101):print(n,all(n%i for i in range(2,n))) # technically fails

[versed eagle]

it is the same

as cereals

and shorter than mine by 5

[low lynx]

oh damn

[versed eagle]

though, cereals also gets 1 and 0 wrong

iirc

[meager zinc]

63 chars

No he said that it was the shortest fix

[low lynx]

no cereal's the fix

[versed eagle]

nvm

checked it

[meager zinc]

see this

[low lynx]

maybe we could look at expressing the huge number as like x**y + c

or something

I'll have a look at it once I get home

[meager zinc]

I can take a crack at the number with the smallest difference

[low lynx]

go ahead

[meager zinc]

it'll take me a couple minutes brb

[versed eagle]

there's lots of repeating chunks

its really close to 2**97

if that helps at all

[low lynx]

!e

print(f'{0x20208828828208a20a08a28ac-2**97:x}')

[night quarry]

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

208828828208a20a08a28ac

[gleaming linden]

2**97 is just the msb

[low lynx]

oh wait yeah

that won't work

[versed eagle]

i was just
taking the number
and dividing it by stuff

[low lynx]

we're looking for like (smallish number)*(smallish number)**(smallish number) +- (smol remainder)

the remainder is really the main thing we need to minimize I think

everything else should be pretty small

[gleaming linden]

I don't think that's possible

the bits don't repeat

[versed eagle]

i have an idea

[gleaming linden]

what if you get rid of all the evens and hardcode 2 in

[meager zinc]

I'm figuring it out rn

bruteforce style

but the number is really big so I might have to figure out a threaded solution

hmm

I need to find a better method

I wonder if we could use a sort-of binary search

where we take 0 to NUM and get the number in the middle and check if a smaller number would be closer or a larger, then cut it

I can't quite figure it out but it might work

[pure dew]

is this where we rediscover IEEE754

[vague cairn]

!e ```py
from functools import reduce as _reduce

def zip_dicts(*dicts, sentinal=None):
"yields a superset of keys from all dicts"
MY_SENTINAL = object()
def my_get(dic,k):
ret = dic.get(k, MY_SENTINAL)
if ret==MY_SENTINAL:return sentinal
return ret

keys = _reduce(lambda a,b: a|set(b), dicts, set())
for k in keys:
    yield k, *(my_get(d,k) for d in dicts)

test1={1:2}
test2={3:4}
test3={5:6}
print(list(zip_dicts(test1,test2,test3)))```

[night quarry]

@vague cairn :white_check_mark: Your 3.11 eval job has completed with return code 0.

[(1, 2, None, None), (3, None, 4, None), (5, None, None, 6)]

[vague cairn]

Is there a more canonical / pythonic way to do this?

for use like this:py for k,v1,v2,v3 in zip_dicts(d1,d2,d3): ...

[haughty nova]

Looks good

[vague cairn]

I hate the my_get() in there, but I'm paranoid that somewhere in the library there's a mapping object whose .get() raises errors when default=None and I specifically want a default of None to be possible.)

Oh, that argument should probably be renamed to default.

[flint hollow]

it's longer and less esoteric, but this works:

!e

d1 = {'a': 1, 'b': 2, 'c': 3}
d2 = {'x': 7, 'y': 8, 'z': 9}

for kv1, kv2 in zip(d1.items(), d2.items()):
    k1, v1 = kv1
    k2, v2 = kv2
    print(k1, v1, k2, v2)

[night quarry]

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

001 | a 1 x 7
002 | b 2 y 8
003 | c 3 z 9

[haughty nova]

Use a simple if instead of get, then

[vague cairn]

... that also could be useful, but not what I needed.

[haughty nova]

Yea that's a different thing

Thurisatic's code

That's what you replied to I assume?

[vague cairn]

yes

[haughty nova]

!e ```py
from functools import reduce as _reduce

def zip_dicts(*dicts, default=None):
"""yields a superset of keys from all dicts"""
keys = _reduce(lambda a,b: a|set(b), dicts, set())
for k in keys:
yield k, *(d.get(k, default) for d in dicts)

test1={1:2}
test2={3:4}
test3={5:6}
print(list(zip_dicts(test1,test2,test3)))```

Ayo

[night quarry]

@haughty nova :white_check_mark: Your 3.11 eval job has completed with return code 0.

[(1, 2, None, None), (3, None, 4, None), (5, None, None, 6)]

[haughty nova]

Ok right?

[vague cairn]

Yes, that's exactly what I meant by, my_get() shouldn't be needed.

[versed eagle]

i think what grass is doing is, for each key in the set of all the keys, put all the values that correspond with that key from each dict into the tuple
and use the default value when a dict doesn't contain a value for that key

[haughty nova]

Correctus

!e
Better example I suppose ```py
from functools import reduce as _reduce

def zip_dicts(*dicts, default=None):
"""Yields a superset of keys from all dicts"""

keys = _reduce(lambda a,b: a|set(b), dicts, set())

for k in keys:
    yield k, *(d.get(k, default) for d in dicts)

test1={1:2, 3:8}
test2={3:4, 9:25}
test3={5:6, 1:100}

print(list(zip_dicts(test1,test2,test3)))```

[night quarry]

@haughty nova :white_check_mark: Your 3.11 eval job has completed with return code 0.

[(1, 2, None, 100), (3, 8, 4, None), (5, None, None, 6), (9, None, 25, None)]

[haughty nova]

I think you don't have to worry about custom mappers having an odd get method which would raise errors, but if you want to foolproof for user defined stuff```py

def ensure_get(d, key):
return d[key] if key in d else default

for ...
yield k, *(ensure_get(d, k) ...) ```

Don't think you can or should foolproof beyond that

[vague cairn]

fair enough.

[meager zinc]

If they have a strange get implementation that seems like more of a problem with that custom mapper than something here.

Also, maybe I'm missing something but you might consider moving this to a help channel - see #❓｜how-to-get-help

[vague cairn]

I wish I remembered which Mapping was the problem. It may have been fixed since 3.5 when I started...

[vague cairn]

But I don't want help, I already closed my library and moved on, I just enjoyed the general-ness of the solution I found and wanted to A) share, and B) seek creative inspiration how to make it better.

[low lynx]

i mean the Mapping type from collections.abc has get as one of its methods so if it doesn't then that's their issue i think

[haughty nova]

I think the name of the function should be changed to

[versed eagle]

def zip_dicts(D*, default=None):
    for k in {k for k in d.keys() for d in D}:
        yield k, *(d.get(d,default) for d in D)

test1={1:2}
test2={3:4}
test3={5:6}
print(list(zip_dicts(test1,test2,test3)))

[haughty nova]

Because while it zips

[versed eagle]

you dont need functools

[low lynx]

[haughty nova]

Zip yields values as an intersection, as a shortest of all iterables

[vague cairn]

Nice!

[haughty nova]

Here it is unioning, equal or longer than all iterables

[low lynx]

this is really more dict_zip_longest yeah

[low lynx]

you have the asterisk the wrong way around

oop sorry about ping

[versed eagle]

its ok lol

i sent the wrong

version

[haughty nova]

D* algorithm

[versed eagle]

!e

def zip_dicts(*D, default=None):
    for k in {k for d in D for k in d.keys()}:
        yield k, *(d.get(k, default) for d in D)

test1={1:2}
test2={3:4}
test3={5:6}
print(list(zip_dicts(test1,test2,test3)))

[night quarry]

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

[(1, 2, None, None), (3, None, 4, None), (5, None, None, 6)]

[versed eagle]

that is

whats intended

the other one is just

entirely wrong

[low lynx]

for k in d should work the same as for k in d.keys() i think

[haughty nova]

Yes

[versed eagle]

ah true

[low lynx]

on this, best i've found so far is

(206*6936**7+0xeb81c62d490d4a28ac) # new
0x20208828828208a20a08a28ac # old

which is significantly longer still
i'm about to switch to using numpy so i can actually run the code im using for larger values since i think you'll need much larger values to get that remainder down

from a quick search doesn't seem like having it of the form a*b**p-c yields anything better

[meager zinc]

I can't create code that's efficient enough

[low lynx]

i'm just doing this so far

num = 0x20208828828208a20a08a28ac
results = []
for b in range(3, 10000):
    for a in range(1, 5000):
        if foo(x := log(num, b) - log(a, b)) < 1e-7:
            print(a, b, x)
            results.append((a, b, x))

where foo is just x - int(x)

i have a feeling p-adic numbers could be helpful for this

but idk enough about them to utilize anything important

[meager zinc]

I feel like solving for the smallest a and b that provide the smallest x difference for any number n of the form a**b-x=n could be a really useful problem in general

We could also try encoding the number using modular arithmetic

But it would likely be similar length

[low lynx]

just a**b isn't very restrictive tbh

[meager zinc]

You can solve for a with any b and solve for b with any a (logs + sqrt), and solve for x with either of those. n is given.

If you could figure out one then you would be good

also if we could figure out a guess-improvement method like Shor's algo or Newton's method that would be interesting. Also, a modified Shor's could be helpful. If we could get it O(N) that would be impressive. Currently it looks for prime factors and is O(N^2).

[low lynx]

is our number prime or not?

nope divisible by 4

that's obvious since 0 and 1 are composite ig

[low lynx]

here's the prime factorization btw

73442231024708479 * 5058929 * 271 * 79 * 5 * 2**2

took my computer ages to check that last number for primality

really should add wheel factorization

[vapid finch]

I can't get 1 to work right, but otherwise it's fine.

x=1
y=1
while x<101:print(x,y%x+1==x);y*=x;x+=1

[sick hound]

couldnt u do x:=1

[vapid finch]

Needs to be x:=x+1 if I understand you right

[sick hound]

im confused

esoteric hurts my head

[vapid finch]

I'm using https://en.wikipedia.org/wiki/Wilson's_theorem.

[night quarry]

@verbal spear :x: Your 3.11 eval job has completed with return code 1.

001 |   File "<string>", line 1
002 |     while x<101:print(x,y%x+1==x) or y,x =y*x,x+1
003 |                                        ^^^^^^
004 | SyntaxError: invalid syntax. Maybe you meant '==' or ':=' instead of '='?

@verbal spear :x: Your 3.11 eval job has completed with return code 1.

001 |   File "<string>", line 1
002 |     while x<101:print(x,y%x+1==x) or (y,x =y*x,x+1)
003 |                                         ^^^^^^
004 | SyntaxError: invalid syntax. Maybe you meant '==' or ':=' instead of '='?

[verbal spear]

!e x=1
y=1
while x<101:
print(x,y%x+1==x) or exec("x,y = x+1,y*x")

[night quarry]

@verbal spear :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | 1 True
002 | 2 True
003 | 3 True
004 | 4 False
005 | 5 True
006 | 6 False
007 | 7 True
008 | 8 False
009 | 9 False
010 | 10 False
011 | 11 True
... (truncated - too many lines)

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

[sick hound]

wait, another way for a sieve? also, 1 is neither a prime or a composite number

[manic prairie]

hello

[low lynx]

wilson's theorem is pretty slow for a primality test

even naive trial division is faster

fine for this tho

[low lynx]

!e

x=y=1
while x<101:print(x,y%x>1<x);y*=x;x+=1

[night quarry]

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

001 | 1 False
002 | 2 False
003 | 3 True
004 | 4 True
005 | 5 True
006 | 6 False
007 | 7 True
008 | 8 False
009 | 9 False
010 | 10 False
011 | 11 True
... (truncated - too many lines)

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

[low lynx]

wait

doesn't work for 2 or 4

because 1! is 1 and 4! is that one case where the residue is nonzero

that's incredibly annoying

[sly ibex]

my mouth drop when i see stickie and n0oO0oOob codes

[pure dew]

🤔

[vapid finch]

!e I can fix 1 with

x=y=1
while x<101:print(x,y%x+1==x>1);y*=x;x+=1

48 bytes.

[night quarry]

@vapid finch :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | 1 False
002 | 2 True
003 | 3 True
004 | 4 False
005 | 5 True
006 | 6 False
007 | 7 True
008 | 8 False
009 | 9 False
010 | 10 False
011 | 11 True
... (truncated - too many lines)

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

[vapid finch]

Actually it's even worse than that. I compute the factorial over the integers, so pretty soon you're doing lots of big integer arithmetic, and not long after, it eats all your memory.

(Well, I guess "not long" is relative. It'll go for a long time. But way less long than if you computed the factorial modulo x.)

[coarse void]

my dude vibin...

[muted cipher]

!e

import struct, math

def fxor(a, b):
  rtrn = []
  a = struct.pack('d', a)
  b = struct.pack('d', b)
  for ba, bb in zip(a, b):
    rtrn.append(ba ^ bb)

  return struct.unpack('d', bytes(rtrn))[0]

print(fxor(float("inf"),1))

[night quarry]

@muted cipher :white_check_mark: Your 3.11 eval job has completed with return code 0.

2.0

[muted cipher]

yes

how to get 2 without 2

ok wtf

!e

import struct, math
def fxor(a, b):
  rt = []
  a = struct.pack('d', a)
  b = struct.pack('d', b)
  for ba, bb in zip(a, b):
    rt.append(ba ^ bb)
  return struct.unpack('d', bytes(rt))[0]
print(fxor(float("inf"),3))

[night quarry]

@muted cipher :white_check_mark: Your 3.11 eval job has completed with return code 0.

1.5

[muted cipher]

hmmm

ieee 754 go brrr

[quartz wave]

1+1

[versed eagle]

~-3

[low lynx]

0b10

[quartz wave]

1<<1

3^1

7&~5

len(' ')

[versed eagle]

7//3

[quartz wave]

int(chr(50))

[versed eagle]

len((len,len))

[muted cipher]

len((len((len,len)),len((len,len))))

[pure dew]

8 ** (1/3)

[muted cipher]

4**.5

[quartz wave]

__name__.__len__().__bool__().__add__(__name__.__len__().__bool__())

[low lynx]

sum(-~a for a, b in zip(iter(int, 1),'xx'))

[muted cipher]

how long can we make this

[low lynx]

i mean you could always just add as many .__call__s as you want

[muted cipher]

ah so my method

go to __call__ hell

[karmic pumice]

!e

print((("а" > "a") + ("a" < "а")))

[night quarry]

@karmic pumice :white_check_mark: Your 3.11 eval job has completed with return code 0.

2

[quartz wave]

pulled a sneaky there

[low lynx]

!e

print((("а" > "a") + ("a" < "а")))
print((("a" > "а") + ("a" < "а")))
print((("a" > "а") + ("а" < "a")))

[night quarry]

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

001 | 2
002 | 1
003 | 0

[karmic pumice]

xd

imagine actually using it to troll people

[tawny helm]

ok

ok

ok

[meager zinc]

!e ```py
from inspect import*
print(getframeinfo(currentframe()).lineno)

[night quarry]

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

2

[meager zinc]

!e print(len(((),())))

[night quarry]

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

2

[meager zinc]

!e print(True+True)

[night quarry]

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

2

[arctic skiff]

!e print(2)

[night quarry]

@arctic skiff :white_check_mark: Your 3.11 eval job has completed with return code 0.

2

[quartz wave]

!e ```py
import sys
print(sys._getframe().f_lineno)

[night quarry]

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

2

[wheat river]

!e

print((lambda x:(a:=x/2,[a:=1+x/a**2for _ in range(x*5)])[1][-1].__int__())(4))

[night quarry]

@wheat river :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | <string>:1: SyntaxWarning: invalid decimal literal
002 | 2

[wheat river]

!e

print((lambda x:(a:=x/2,[a:=(1).__add__(x.__truediv__(a.__pow__(2)))for _ in range(x*5)])[1][-1].__int__())(4))

[night quarry]

@wheat river :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, in <lambda>
004 |   File "<string>", line 1, in <listcomp>
005 | AttributeError: 'NotImplementedType' object has no attribute '__pow__'. Did you mean: '__doc__'?

[wheat river]

rip

!e

from collections import deque

@print
@deque.popleft
@deque
@int.__str__
@int.__sizeof__
@int
@lambda _:_.__doc__
class _:
    '1'

[night quarry]

@wheat river :white_check_mark: Your 3.11 eval job has completed with return code 0.

2

[wheat river]

!e

@print
@int
@next
@str.__iter__
@int.__str__
@int.__sizeof__
@int
@lambda _:_.__doc__
class _:
    '1'

[night quarry]

@wheat river :white_check_mark: Your 3.11 eval job has completed with return code 0.

2

[near siren]

!E

import re, types
from math import e, pi, tau

print(types.LambdaType(compile(f"{dir(types)[round(pi ** e - pi * e)][:int(tau)].casefold()} λ :--λ--λ", "sadness", "eval"), globals())()(re.T))

[night quarry]

@near siren :white_check_mark: Your 3.11 eval job has completed with return code 0.

2

[verbal spear]

!e import os
print("hi")
print('<style>body{background-color:"red"}</style>')
os.system('touch pic.png')

[night quarry]

@verbal spear :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | hi
002 | <style>body{background-color:"red"}</style>

[verbal spear]

!e

[night quarry]

Missing required argument

code

Command Help

!eval [python_version] <code, ...>
Can also use: e

Run Python code and get the results.

This command supports multiple lines of code, including code wrapped inside a formatted code block. Code can be re-evaluated by editing the original message within 10 seconds and clicking the reaction that subsequently appears.

If multiple codeblocks are in a message, all of them will be joined and evaluated, ignoring the text outside of them.

By default your code is run on Python's 3.11 beta release, to assist with testing. If you run into issues related to this Python version, you can request the bot to use Python 3.10 by specifying the python_version arg and setting it to 3.10.

We've done our best to make this sandboxed, but do let us know if you manage to find an issue with it!

[verbal spear]

!e import os
os.system("dir")

!e import os
os.system(' ls - ali')

[night quarry]

@verbal spear :warning: Your 3.11 eval job has completed with return code 0.

[No output]

[verbal spear]

!e import os
print(os.system(' ls - ali'))

[night quarry]

@verbal spear :white_check_mark: Your 3.11 eval job has completed with return code 0.

32512

[arctic skiff]

!e ```py
from random import randint
print(randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, 99999))))))))))))))))))))))))))))))))))))))))))))))))))))))))

[night quarry]

@arctic skiff :white_check_mark: Your 3.11 eval job has completed with return code 0.

0

[tough willow]

!e
from random import randint
print(randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, 99999))))))))))))))))))))))))))))))))))))))))))))))))))))))))

[night quarry]

@tough willow :white_check_mark: Your 3.11 eval job has completed with return code 0.

0

[tough willow]

!e
from random import randint
print(randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, 999999999999999))))))))))))))))))))))))))))))))))))))))))))))))))))))))

[night quarry]

@tough willow :white_check_mark: Your 3.11 eval job has completed with return code 0.

0

[tough willow]

Ok

[fleet bridge]

!e ```py

from random import randint
print(randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, randint(0, 2**100))))))))))))))))))))))))))))))))))))))))))))))))))))))))

[night quarry]

@fleet bridge :white_check_mark: Your 3.11 eval job has completed with return code 0.

177459

[fleet bridge]

it has very weird distribution

[meager zinc]

!e ```py
print(
len(
((()
,( ),
() ,()
,(),
(),
()
,
((),)),())
)
)

[night quarry]

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

2

[low lynx]

neat

[night quarry]

Hey @sick hound!

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

[quartz wave]

so um i made code to see how long it takes to reach the end and

still not done

after hours

so latest took 1h 44m 49s 835 ms to reach that point

[versed eagle]

fun

[meager zinc]

You can probably compute the probability

but I'm not particularly familiar with it

you could probably speed it up with C and threading and SIMD vector operations

(or Rust)

also who is mango time

[pure dew]

not bad, but you can do the entire calculation inside the transformer class and speed it ut

[meager zinc]

!e assert "​" == ""

[night quarry]

@meager zinc :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 | AssertionError

[meager zinc]

anyone here want to fork black with me and make it remove all the spaces

or any other formatter

[dry mirage]

!raw 1077068322354507908

[night quarry]

== Raw message ==

!e `​assert "​" == ""`​

[smoky scarab]

!e print(-42.0 .is_integer())

[night quarry]

@smoky scarab :white_check_mark: Your 3.11 eval job has completed with return code 0.

-1

[rugged owl]

[meager zinc]

alr do we want to do black or something like autopep8 instead

[rugged owl]

Honestly would have to looking at the repositorys first to see

[quartz wave]

‫only 2 remaining after 5h 38m 36s 258ms

unfortunately i have to stop now

experiment stopped at 8h 31m 38s 454 ms 646 us 400+ ns

iteration 2687300449

[magic wraith]

I can create one for you if you still need it, seems like a simple function

[karmic pumice]

!pypi LoneDruid
made it myself after asking this

[night quarry]

lonedruid v0.5.2

Package to convert stuff to esoteric stuff, maybe something else later.

[magic wraith]

seems cool, have you thought of creating a repository for it? I'm sure a lot of people would have fun ideas for contributing

[wheat river]

what is this

[quartz wave]

@wheat river

j is the amount of randint()s

[pure dew]

the code would also be a lot nicer

for big project, yea

[wheat river]

ah k

[summer orchid]

https://youtu.be/lYOKkoVHWTU guess this fits into here

[magic wraith]

heya everyone, any reverse engineers here? i have a simple obfuscated script and i'd like to know how good is the protection and how long it will take someone to break it (i'm testing a obfuscator i made) if you want to give it a try here is the files, dm me with any feedback you have
https://cdn.discordapp.com/attachments/879026043846017044/1077314731469176902/files.zip

[night quarry]

Hey @magic wraith!

It looks like you tried to attach file type(s) that we do not allow (.zip). We currently allow the following file types: .gif, .jpg, .jpeg, .mov, .mp4, .mpg, .png, .mp3, .wav, .ogg, .webm, .webp, .flac, .m4a, .csv, .json.

Feel free to ask in #community-meta if you think this is a mistake.

[next flame]

unzipped files here
obf.py: https://paste.pythondiscord.com/aqinobojav.py
main.py:

from obf import string_hash

# task: find the algorithm for string_hash function
print(string_hash('Hello World!'))
print(string_hash('1234567890'))
print(string_hash('⫊ṣՖᔲᗶ'))

[verbal spear]

thats really cool! if you begin with 2**20, when will it in average reach 0?
I thoght after 20 iterations, but it is less.

[low lynx]

the expected number of times starting at 2**20 for it to reach 0 should be around 15.44

[arctic skiff]

Min 1 is enough reach 0 max 20**20 = 4000 ig

[low lynx]

the expected number of iterations needed for a number n would be around 1+H_n where H_n is the nth harmonic number

[arctic skiff]

(if the given number is 20

[rugged sparrow]

@magic wraith at first glance it looks like a virtual machine?

B.a is the instruction pointer

global A is the stack? (or one of them)

[earnest wing]

Let me take a look

yeah, a vm with a constant program and a few stacks

A for data, F and G as a form of call stack

the globals manipulation and the large number of instructions makes it kind of annoying to read though

[rugged owl]

Anyone know a way to make this sliding window algo shorter

81 bytes / chars

import sys;a,w=sys.argv[1:];print([a[i:i+int(w)]for i in range(len(a)-int(w)+1)])

[magic wraith]

that's true

i've been trying to make the shortest bytecode possible though

[rugged sparrow]

how are you making the bytecode?

did you write your own compiler?

[magic wraith]

well yes

but it doesn't work on every python code

for example async, yield or recursive functions

[drowsy anvil]

class x:
  class y:
    pass

[magic wraith]

!e py def deco(func): def inner(obj, *args, **kwargs): return obj.__class__(*args, **kwargs) return inner class A: @deco def __init__(self): ... x = A()

[night quarry]

@magic wraith :x: Your 3.11 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 9, in <module>
003 |   File "<string>", line 3, in inner
004 |   File "<string>", line 3, in inner
005 |   File "<string>", line 3, in inner
006 |   [Previous line repeated 496 more times]
007 | RecursionError: maximum recursion depth exceeded

[sly ibex]

ruby golfing vs python golfing

i've been seeing ruby can more shorter the codes than python does

[proper vault]

Ye, ruby is generally better for code golf

[fleet bridge]

Can you show some examples of golfed code in ruby? I think it is suitable for this channel

[sick hound]

puts"Hello world!"
Ruby does not require parentheses in method calls

[quartz wave]

fun thing about lua is that it also doesn't require parentheses for table literals and strings

[sick hound]

Also, one should remember that in Python, a newline is only required after each decorator and before block declarations. So, this is completely valid:

def three():r=range(10);return[1for i in r if i==2][0]
for i in range(4):print(i)

[magic wraith]

also notice that : has higher precedence than ;
so py if thing:first();second() is equivalent to py if thing: first() second() rather than py if thing: first() second()

[odd light]

!e

import gc
import ctypes

gc.get_referents(list.__dict__)[0]["map"] = lambda _, fn: list(map(fn, _))
ctypes.pythonapi.PyType_Modified(ctypes.py_object(list))

print([1, 2, 3, 4].map(lambda x: x+1))

[night quarry]

@odd light :white_check_mark: Your 3.11 eval job has completed with return code 0.

[2, 3, 4, 5]

[odd light]

kek

[rugged sparrow]

!e ```py
from fishhook import hook
@hook(list, name='map')
def list_map(self, fn):
return type(self)(map(fn, self))

print([1, 2, 3, 4].map(lambda x: x+1))

[night quarry]

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

[2, 3, 4, 5]

[rough moat]

print("".join(chr(ord(i)-0b100111000000000) for i in "么乥乬乬乯丠乷乯乲乬乤丠乵乷乵"))

[sturdy rune]

anyone got a code here that doesn't use variable or things like print()? Basically 100% esoteric

[low lynx]

don't think its possible to display stuff without print

[rugged sparrow]

!e ```py
open(0, 'w+').write('hello world')

[night quarry]

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

[No output]

[rugged sparrow]

huh that should have worked

wack

[low lynx]

I phrased that wrong

I don't think it's possible to display stuff without any letters

[rough moat]

!e
print("".join(chr(ord(i)-0b100111000000000) for i in "么乥乬乬乯丠乷乯乲乬乤丠乵乷乵"))

[night quarry]

@rough moat :white_check_mark: Your 3.11 eval job has completed with return code 0.

Hello world uwu

[versed eagle]

!e

open(0, 'w').write('hello world')

[night quarry]

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

[No output]

[versed eagle]

that works on my machine^

my guess is that it wont work on this bc it only captures stdin/stderr

[versed eagle]

but, the 'w+' won't work

[meager zinc]

stdin is likely overwritten and your code prints to the terminal running python which isn't captured

[versed eagle]

your code prints to the terminal running python which isn't captured
that's exactly what i just said??

also wdym "overwritten"

[meager zinc]

1. stdin = my_stdin # my_stdin is a custom class from snekbox
2. similar but I added the fact that it prints to the terminal running the script

I need to look at the snekbox source

[versed eagle]

wdym you "added the fact that it prints to the terminal"
i had already said that.

im confused as to what you're saying

since you're repeating what i've already said

[meager zinc]

my guess is that it wont work on this bc it only captures stdin/stderr
not sure where this mentions the fact it's printing to the original terminal rather than another stdout
I would rather not argue about this, this is a dumb thing to argue over. would you prefer that I delete all the messages?

My guess is that it's printing to the Snekbox thread instead of the NsJail isolated thread.

[versed eagle]

not sure where this mentions the fact it's printing to the original terminal rather than another stdout
it says right there, it's not captured. and, as a side note, it's not printing to stdout at all. not sure where that came from

[versed eagle]

to my knowledge we aren't arguing?

im asking for clarification since i genuinely didnt understand what you're saying

[meager zinc]

okay

here's what I'm trying to clarify

[versed eagle]

sorry if it seemed like i was arguing with you

that wasnt my intention

i'll try to phrase my questions better in the future

[meager zinc]

open(0) is referring to the stdout of the parent process Snekbox runs
stdout in Python is referring to the stdout of the child process NsJail executes (and isolates)

stdin is not captured ever
stdout/stderr is only captured from the child process
the parent process is never captured from.

[versed eagle]

open(0) is referring to the stdout of the parent process Snekbox runs
0 is the file descriptor for stdin

not stdout

[meager zinc]

oh

yeah

I missed that

!e ```py
open(1, 'w').write('hello world')

[night quarry]

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

hello world

[meager zinc]

that's what it was supposed to be