esoteric-python

[versed eagle]

so by definition you cant allow full python syntax in the problem

.

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.

by definition, if the goal is succeeded, the "full python syntax" cannot be allowed

because underscores would have to be disallowed

preventing arbitrary code exec was part of preventing underscores

[rugged sparrow]

i said a limit subset of python code right here

you disagreed

[old socket]

What is going on rn

[versed eagle]

i disagreed, aside from the restrictions imposed by the goal

which i thought was implicit

sorry if i didnt make that clear.

[rugged sparrow]

now you are just backtracking

[versed eagle]

no, i am not

im saying the same thing that i have been saying

[rugged sparrow]

plenty of the code i sent after that message contained dunders and underscores and you never clarified

[versed eagle]

because none of those need the underscores to function

if you know the sizes of things beforehand, which presumably you would, you can use constants

and if you dont know the sizes, you can brute force it

given that there aren't really that many likely possibilities

[rugged sparrow]

#esoteric-python message this entire bug revolves around __new__ which contains underscores

[versed eagle]

so then the bug is irrelevant

because to use it, you first have to get something that would already be disallowed

[rugged sparrow]

if this has been your argument from the beginning, then we agree, you would have to limit it to such a small subset that you could not call it python

[versed eagle]

the point was to disallow underscores and underscored names

that's part of the goal

__new__ is an underscored name, and so would be disallowed

[rugged sparrow]

dunders are a core language feature, without them, I would not consider it python

[earnest wing]

have fun without any of the stdlib btw

[versed eagle]

then we agree, you would have to limit it to such a small subset that you could not call it python
we do not agree, actually

[rugged sparrow]

then what point are you trying to make

[versed eagle]

you would have to limit it to such a small subset that you could not call it python
i never said that?

[versed eagle]

that it is theoretically possible, in the python language (specifically, with cpython) to do AT LEAST one of the following given a piece of code:
determine whether it grants user access to underscore names statically OR prevent such access during runtime

[rugged owl]

[versed eagle]

@old socket
no im just distracted and trying to do other stuff at the same time, so it takes a long time to type stuff lol

no large paragraph here

[rugged sparrow]

that is not at all what you were arguing lol

[versed eagle]

that is exactly what i have been arguing

this entire time

[old socket]

I'm super confused right now but none the less, I am enjoying the show

[rugged sparrow]

right here you disagreed when i said limited subset

[rugged sparrow]

this is a limited subset

[versed eagle]

i wasn't disagreeing with limited subset

i was replying to this

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

of course it's limited, underscore names aren't allowed!

we've been through this

the goal is to prevent access to underscore names

by definition, achieving the goal creates a subset of python

[rugged sparrow]

what are you arguing then?

[versed eagle]

.

[rugged sparrow]

the OR prevent such access during runtime is where it gets fuzzy

[versed eagle]

also, please dont call me "dude"
not mad, ik you didn't mean anything by it
but i use she/her

[rugged sparrow]

sorry, my bad

[split stirrup]

I have been reliably informed by my nieces that "dude" is non-gendered

[rugged sparrow]

im still gonna respect @versed eagle 's wishes

[still charm]

this thread's name scares me

[rugged sparrow]

!e py print('{0.\u005F\u005Fclass\u005F\u005F.\u005F\u005Fbase\u005F\u005F.\u005F\u005Fsubclasses\u005F\u005F}'.format(1)) @versed eagle you can at least read __ values by abusing .format

[night quarry]

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

<built-in method __subclasses__ of type object at 0x7f121a099580>

[rugged sparrow]

*but you only get the resulting value as a string

[versed eagle]

from a strictly security standpoint, that wouldn't matter as much as writing to it, but for this challenge that would have to be prevented
could probably be prevented by hooking str.format to not work if it would format into a dunder

[rugged sparrow]

afaik, @burnt pasture 's challenge was to bypass a theoretical system that evals input in a call that looks like this ```py
def safe(inp):
if '_' in inp:raise Exception('invalid')
exec(inp, {'builtins':{}})

[old socket]

I don't think you'd be hooking format or __format__ cause that just get's the string representation of whatever object

I think so at least, I only remember __format__ takes in a specification that you can use to determine what to do

[rugged owl]

[old socket]

Playing with codecs has been very fun, but some annoying stuff I've come across when developing in poetry is that

Sometimes if the package you are working on isn't installed, and you add your .pth file

It'll corrupt poetry stopping you from installing new packages

You need to delete the poetry cache

Somehow adding the .pth before installing your codec package will make poetry error with a json decoder error

[rugged sparrow]

poetry is probably reading the .pth's and failing in there

[old socket]

Should I make an issue on the GitHub? I feel like it might not be to relevant though since not a lot of people will be doing stuff like that

[rugged sparrow]

depends on how its failing (like if you were building something with a "normal" .pth will if fail there)

[old socket]

I think it's failing whenever you have add a .pth file that references the current module (the one you are developing) before you do poetry install to add the current module to site-packages

So I think it'll error in perhaps "normal" cases but only if someone messes with .pth files

[rugged sparrow]

ah, so poetry install loads site which loads .pth files, and yours fails because the package it references isnt installed yet

[rugged sparrow]

i wouldn't bother reporting it as a bug then

[old socket]

Yeop sounds about right, it's just tedious to fix

Just deleting the cache, re-installing env and done

But until you do that, you cannot do poetry install ... and some other commands

I'm pretty sure poetry debug info is also gonna report that you have no current directory

[rugged sparrow]

actually, i would report that. Poetry probably shouldn't completely fail like that

[old socket]

I'll see if I can reproduce right now

I'll send the error as well

Oh yeop looks like it's reproduce-able, but I forgot the exact steps to get it there under "norma" circumstances

So I forced it to error

[old socket]

Here's what it looks like

And yea, after this happens poetry reports that there is no cache directory

Or it errors when trying to get a cache directory? *

[rugged sparrow]

thats a weird failure

[old socket]

Yea

I want to take a nap, so I'll probably write an issue later

But I want to first figure out the exact steps I did to get it normally, cause I reproduced just barely by manually causing the issue to happen

[rugged owl]

!e

def b(c,d=0,data=[0]*30000,o=""):
    c,_=list(c),"".join(c);b={};s=[]
    for i,char in enumerate(c):
        if char=="[":s.append(i)
        elif char=="]":b[s.pop()]=i;b[i]=b[-~s[0]]
    while d<len(c):
        char=c[d];d+=(char==">")-(char=="<");d=0 if d<0 else d
        data[d]=255&(data[d]+(char=="+")-(char=="-")),
        o+=chr(data[d])if char=="."else o;data[d]=ord(input()[0])if char==","else data[d]
        d=b[d]if char in"[]"and data[d]^48else d+1
    return o
print(b("++++++++++[>+++++++>++++++++--->+++>+<<<<-]>++.>+.+++++++..+++.>++.<<+++++++++++++++.>.+++.------.--------.>+.>."))

[night quarry]

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

001 | <string>:10: SyntaxWarning: invalid decimal literal
002 | Traceback (most recent call last):
003 |   File "<string>", line 12, in <module>
004 |   File "<string>", line 5, in b
005 | IndexError: list index out of range

[rugged owl]

what am i doing wrong here

[pure dew]

access after pop

[sonic birch]

Hi, not sure it's the right channel, but.. is it possible having a regex1 check whether it implies regex2? For example regex1=.+.gmail.com and regex2=.+.com so regex1 -> regex2

[topaz haven]

How do I make statements inside linear lambda?

For example,

x = 1
print("hi")
x += 1
print(x)

Or sth like this

[cloud fossil]

Lambdas can't have statements in them

But they can have a function call which runs statements

[low lynx]

!e

f=lambda:(x:=1)and(print('hi'))or(x:=x+1)and(print(x))
f()

[night quarry]

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

001 | hi
002 | 2

[quartz wave]

f=lambda:(x:=1,print('hi'),x:=x+1,print(x)) condensed

[arctic skiff]

!e```py
(lambda:(x:=1,print('hi'),x:=x+1,print(x)))()

[night quarry]

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

001 | hi
002 | 2

[finite blaze]

why x:=1 works in lambdas but x=1 doesn't?

[earnest wing]

expression vs statement

[versed eagle]

!e

(lambda x=1:(print('hi'),x:=x+1,print(x)))()

[night quarry]

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

001 | hi
002 | 2

[earnest wing]

in a nutshell, you can think of it this way

• expressions can be put inside parentheses ()
• statements cannot, because that causes a syntax error

[finite blaze]

Alright, thanks

[versed eagle]

nope, it saves a char @arctic skiff

[arctic skiff]

yeah i see thats why i del

[versed eagle]

>>> len("""(lambda:(x:=1,print('hi'),x:=x+1,print(x)))()""")
45
>>> len("""(lambda x=1:(print('hi'),x:=x+1,print(x)))()""")
44 

[arctic skiff]

!e```py
(x:=1,print('hi'),x:=x+1,print(x))

[night quarry]

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

001 | hi
002 | 2

[arctic skiff]

!epy print(len("""(x:=1,print('hi'),x:=x+1,print(x))"""))

[night quarry]

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

34

[versed eagle]

How do I make statements inside linear lambda?

[arctic skiff]

you can't

[versed eagle]

the point is to do it inside a lambda

[versed eagle]

yes ik that. you have to use expressions.

but the point is still the same

[arctic skiff]

oh didn't see >

[versed eagle]

it's meant to be inside a lambda

[finite blaze]

how could i do this in a for loop in one line?

[arctic skiff]

!e```py
(lambda x:(print('hi'),x:=x+1,print(x)))(1)

[night quarry]

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

001 | hi
002 | 2

[arctic skiff]

!epy print(len("""(lambda x:(print('hi'),x:=x+1,print(x)))(1)"""))

[night quarry]

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

43

[arctic skiff]

but still x=1 is a arg so its still a part of statement

[versed eagle]

!e

(lambda x:(print('hi'),print(x:=x+1)))(1)

[night quarry]

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

001 | hi
002 | 2

[versed eagle]

41

[arctic skiff]

!epy (lambda x:(print('hi'),print(x+1)))(1)

[versed eagle]

oh wait

[night quarry]

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

001 | hi
002 | 2

[versed eagle]

now the prints

can be combined

[arctic skiff]

wait yeah

!epy (lambda x:print('hi\n',x:=x+1))(1)

[versed eagle]

!e

(lambda x:print(f'hi\n{x+1}'))(1)

[night quarry]

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

001 | hi
002 | 2

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

001 | hi
002 |  2

[arctic skiff]

since you are using x+1

[versed eagle]

i used it because you used it here

was gonna keep using walrus until then lol

[arctic skiff]

then i realized we were using lambda and expressions

!epy print('hi\n2')is shortest

[night quarry]

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

001 | hi
002 | 2

[versed eagle]

not a lambda though

[arctic skiff]

lambda went with lama

[versed eagle]

what?

[calm loom]

Llama

🦙

[versed eagle]

what llama

there's no llama

[calm loom]

!e

(lambda :print("hi\n2"))()

[night quarry]

@calm loom :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | hi
002 | 2

[calm loom]

🦙

[rugged owl]

!e

(lambda:print("hi\n2"))()

[night quarry]

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

001 | hi
002 | 2

[rugged owl]

1 less char

[rough moat]

print(__import__("os").system("")or"","\n".join(["\n"]+[f"\x1b[38;5;{[80,181,231,181,80][y]}m█"*15 for y in range(5)]),"\x1b[0m")

wont work in !e or whatever

it uses ansi color

[restive void]

What's the system("") for?

[earnest wing]

Force color in windows command prompt

[dry mirage]

!e

from einspect import view

view(1).swap(0)

print(0 > 1)
print(bool(0), bool(1))

print(0 * 0 == 1.0)

[night quarry]

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

001 | True
002 | True False
003 | True

[dry mirage]

🥴 slightly cursed

[versed eagle]

do the 0 > 1 and 0 * 0 == 1.0 not get const folded?

[dry mirage]

0 > 1 is not folded

0 * 0 is folded to 0 == 1.0

and since 0 is now 1, it's true

[meager zinc]

The smallest that I could come up with (slightly different format, dynamically allocated stack) but here it is (this could probably be better):

def b(p):
    s=[i:=0];c=''
    for o in p:c+=' '*i+[f'i+=1\n{" "*i}if i>=len(s):s+=[0]','i-=1','s[i]+=1','s[i]-=1','print(end=chr(s[i]))','s[i]=ord(__import__("sys").stdin.read(1))','while s[i]:','s','']['><+-.,['.find(o)]+'\n';i+=(92-ord(o))*(o in'][')
    exec(c)

[meager zinc]

anyone got any ideas to golf this further?

[quartz wave]

comparisons do not get constant folded

[meager zinc]

if True and if __debug__ do get const folded unfortunately

[split salmon]

!e I made this on phone py __import__('pickle').loads(bytearray(map((lambda x:x-4),__import__('gzip').decompress(bytearray(map((lambda x:x-1 if x!=0xff else x),[1,1,1,92,39,114,106,209,1,145,131,49,206,133,104,59,52,171,26,6,102,198,214,43,198,6,6,154,63,3,10,139,161,11,163,221,179,147,139,161,115,163,221,139,131,221,203,181,159,3,153,94,77,200,12,70,188,135,126,50,160,41,206,79,41,45,225,77,104,51,12,244,212,115,76,148,185,20,2,3,195,26,153,226,108,255,3,100,191,154,235,1,9,140,32][::-1]))))))

[night quarry]

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

Hello world!

[split salmon]

"No module found named 'einspect'"?

[dry mirage]

hm?

[split salmon]

Python 3.9.7

[dry mirage]

!pypi einspect is not a builtin 👀

[night quarry]

einspect v0.5.8

Extended Inspect - view and modify memory structs of runtime objects.

[split salmon]

Oh damn

Sh-t

[wheat river]

https://github.com/ionite34/einspect/blob/807b70c16410898e2676438c864c9b1aa23137fd/src/einspect/views/view_cfunction.py#L61

What is vectorcall??

[night quarry]

src/einspect/views/view_cfunction.py line 61

def vectorcall(self) -> vectorcallfunc:```

[acoustic quail]

for i in range(len(l.split())):
    try: eval(f'f2r{int(i//rows)}')
    except: exec(f'f2r{int(i//rows)} = Frame(f2)')
    exec(f'f2r{int(i//rows)}c{i%rows} = Button(f2r{int(i//rows)}, font = f,\
           text = l.split()[i], command = lambda: e1.insert(END, l.split()[{i}][1]))')
    exec(f'f2r{int(i//rows)}c{i%rows}.pack(side = LEFT)')
    exec(f'f2r{int(i//rows)}.pack()')```a snippet of a thing i'm doing

[potent flare]

what the fuck

what in the actual fuck

why

[acoustic quail]

because i don't want to make 143 buttons arranged in 10 rows

manually

that do almost the same thing

also this is #esoteric-python isnt it

[low lynx]

why not use enumerate?

[acoustic quail]

what

[low lynx]

if you're doing this to be esoteric then sure go ahead and do whatever but you could use a list instead of exec statements

[low lynx]

for i, v in enumerate(l.split()):
    ...

and replace all your l.split()[i] with just v

[acoustic quail]

i want the value to be saved

exec(f' ··· lambda: ··· l.split()[{i}][1])) works as needed

··· lambda: ··· l.split()[ i ][1]) gives
··· lambda: ··· l.split()[142][1]), which works as if all buttons were the last one

[quartz wave]

condensed version/base for golfing (271) ```py
p=0;i=min(rows,n:=len(L:=l.split()))
while z:=i-p:j=p//rows;p=i;a=f'f2r{j}';g=rows*j;exec(a+'=Frame(f2);'+f'(f2r{j}c%d:=Button({a},font=f,text=L[g+%d],command=lambda:e1.insert(END,L[{g}+%d][1]))).pack(side=LEFT);'z%sum(zip([range(z)]*3),())+a+'.pack()');i=min(n,i+rows)

does exactly the same thing

[finite blaze]

for i in range(1000):
    if i%3==0 or i%5==0: s+=i

how can I oneline something like this?

[fleet bridge]

sum(a for b in c if d)

[finite blaze]

I just found a pretty interesting page

with a lot of problems to golf

https://projecteuler.net

[last locust]

Or if you want the same approach then for i in range(1000): s+= i * (i % 3 == 0 or i % 5 == 0)

[gleaming linden]

or sum(a*(condition)for a in b)

[low lynx]

3*((999//3)*(999//3+1)//2)+5*((999//5)*(999//5+1)//2)-15*((999//15)*(999//15+1)//2)

for a pure mathematical solution

[dry mirage]

https://docs.python.org/3/c-api/call.html#the-vectorcall-protocol

it's only used when internal python C calls target the function iirc

but faster call without making tuples and dicts

[rugged sparrow]

tp_vectorcall is preferred any time that a call occurs

!e thats why ```py
from fishhook import *

@hook(type(lambda:0))
def call(self, *args, **kwargs):
print(self, args, kwargs)
return orig(self, *args, **kwargs)

def foo(a):pass

foo(1)

[night quarry]

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

[No output]

[rugged sparrow]

(for better or for worse)

[dry mirage]

[rugged sparrow]

if you null out PyFuncType->tp_vectorcall to switch to using tp_call then you crash due to recursion

[dry mirage]

python defined funcs also have a separate tp_call and tp_vectorcall?

also why does that crash from recursion

[rugged sparrow]

it does if you null tp_vectorcall then hook it

because that overwrites with the default tp_call which looks up the py_object in the class mapping and then uses PyObject_Call

[dry mirage]

isn't tp_vectorcall optional?

[rugged sparrow]

the recursion issue will only trigger if you use fishhook to hook PyFuncType.__call__ and then null PyFuncType.tp_vectorcall

at least it did in 3.8 when i was first developing fishhook

[dry mirage]

would it work if you cast __call__ to a PYFUNCTYPE first

or does that still recurse

[rugged sparrow]

i believe it still recurses as ctypes uses PyObject_Call or some derivative internally

[dry mirage]

does PyFunctionType even have tpcall?

oh the type

the object instance at least only has vectorcall?

[rugged sparrow]

one sec ill test

@dry mirage how do i modify tp_vectorcall with einspect? setting view(obj)._pyobject.tp_vectorcall did not work

[dry mirage]

does that not?

[rugged sparrow]

>>> from fishhook import getmem
>>> funcmem = getmem(type(lambda:0))
>>> from einspect import view
>>> PyFuncType = view(type(lambda:0))._pyobject
>>> from ctypes import cast
>>> s1 = funcmem.tolist()
>>> PyFuncType.tp_vectorcall = cast(0, type(PyFuncType.tp_vectorcall))
>>> s2 = funcmem.tolist()
>>> s1 == s2
True

getmem just takes an object and returns a memoryview from id(obj) -> id(obj) + obj.__basicsize__

[sick hound]

thoughts?

[dry mirage]

strange

https://github.com/ionite34/einspect/blob/main/src/einspect/structs/py_type.py#L121

[night quarry]

src/einspect/structs/py_type.py line 121

tp_vectorcall: vectorcallfunc```

[dry mirage]

it's just a Structure field so

[dry mirage]

ah I think that might be the issue

[rugged sparrow]

lol if that was an issue then fishhook would not work at all

[dry mirage]

!e

from einspect import view

fnt = type(lambda: 0)

print(view(fnt).mem_size)
print(fnt.__basicsize__)

[night quarry]

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

001 | 416
002 | 136

[dry mirage]

PyTypeObject is a PyVarObject so you can't just use basicsize

[rugged sparrow]

*it gets the basic size of the type

https://github.com/chilaxan/fishhook/blob/master/fishhook/fishhook.py#L15-L28

[night quarry]

fishhook/fishhook.py lines 15 to 28

def getmem(obj_or_addr, size=None, fmt='P'):
    if size is None:
        if isinstance(obj_or_addr, type):
            # type(cls).__sizeof__(cls) calls a function with a heaptype member
            # if cls is currently unlocked and member happens to overlap with other data
            # crash occurs
            # we avoid this by hardcoding TYPE_BASICSIZE
            size = TYPE_BASICSIZE
        else:
            size = sizeof(obj_or_addr)
        addr = id(obj_or_addr)
    else:
        addr = obj_or_addr
    return memoryview((c_char*size).from_address(addr)).cast('c').cast(fmt)```

[dry mirage]

!e

from einspect.structs import PyTypeObject

print(PyTypeObject.tp_vectorcall.offset)

[night quarry]

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

400

[rugged sparrow]

https://github.com/chilaxan/fishhook/blob/master/fishhook/fishhook.py#L10-L11

[night quarry]

fishhook/fishhook.py lines 10 to 11

def sizeof(obj):
    return type(obj).__sizeof__(obj)```

[rugged sparrow]

@dry mirage fishhook.getmem gets the object memory properly for types, that is not the issue

[dry mirage]

yeah it doesn't seem to be being set in any case

[rugged sparrow]

if it did not then it would not be able allocate the structs + insert the structs that it needs to in order for it to actually work

[dry mirage]

maybe it's to do with the PYFUNCTYPE I've assigned to it

[rugged sparrow]

maybe

I think I remember running into an issue with ctypes where struct fields that hold a C/PyFuncType pointer don't set properly

[dry mirage]

wait what

!e

from ctypes import *
from einspect import view

fnt = type(lambda: 0)
obj = view(fnt)._pyobject

print(cast(obj.tp_vectorcall, c_void_p))

[night quarry]

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

c_void_p(None)

[dry mirage]

it's already null? 🥴

[rugged sparrow]

huh

maybe it is special cased now inside of the call.c

[dry mirage]

isn't tp_vectorcall on the function type for the actual function type and not instances?

so function objects can still have vectorcall

https://github.com/python/cpython/blob/433fb3ef08c71b97a0d08e522df56e0afaf3747a/Objects/funcobject.c#L817

[night quarry]

Objects/funcobject.c line 817

PyTypeObject PyFunction_Type = {```

[dry mirage]

definition doesn't seem to include tp_vectorcall

tp_vectorcall_offset is for the instance (I think?)

[rugged sparrow]

maybe

tbh i havent looked at how vectorcall works too much

[dry mirage]

https://github.com/python/cpython/blob/433fb3ef08c71b97a0d08e522df56e0afaf3747a/Objects/funcobject.c#L832

[night quarry]

Objects/funcobject.c line 832

PyVectorcall_Call,                          /* tp_call */```

[dry mirage]

or I guess tp_call is just always PyVectorcall_Call

[rugged sparrow]

• would need to remove has_vectorcall from tp_flags

then it would break

[dry mirage]

then function calls would actually reach FunctionType->tp_call yeah

[royal sun]

https://sadh.life/post/cursed-for/

[rugged sparrow]

cool use of codec

[finite blaze]

My solution for another projecteuler problem

"""
A palindromic number reads the same both ways. The largest palindrome made from the product of two 2-digit numbers is 9009 = 91 × 99.

Find the largest palindrome made from the product of two 3-digit numbers.
"""
r=0;[r:=x*y for x in range(1000)for y in range(x+1,1000)if str(x*y)==str(x*y)[::-1]and x*y>r];print(r)

[sly ibex]

    elif month == 12 and day >=22 or month == 1 and day <= 20:

possible to shorten this?

[tough willow]

 elif any(all(month==12,day>=22),all(month==1,day<=20))
#elif month == 12 and day >=22 or month == 1 and day <= 20:

@sly ibex

[sudden osprey]

Assuming month is between 1 and 12 I think this would work: not(1,20)<(month,day)<(12,22)

[rugged sparrow]

@dry mirage ```py

from fishhook import hook, orig, find_offset, getmem
PyFuncType = getmem(type(lambda:0))
flag_offset = find_offset(PyFuncType, type(lambda:0).flags)
PyFuncType[flag_offset] &= ~(1 << 11) # Py_TPFLAGS_HAVE_VECTORCALL

@hook(type(lambda:0))
... def call(self, *a, **k):
... print(self, a, k)
... return orig(self, *a, **k)
...
Traceback (most recent call last):
File "<stdin>", line 2, in <module>
File "/usr/local/lib/python3.10/site-packages/fishhook/fishhook.py", line 302, in wrapper
File "/usr/local/lib/python3.10/site-packages/fishhook/fishhook.py", line 147, in force_setattr
RecursionError: maximum recursion depth exceeded while calling a Python object``` this is what i figured would happen

[dry mirage]

does that just set __call__?

I wonder which slot it actually sets

PyFuncType->tp_call?

[rugged sparrow]

it sets tp_call

[dry mirage]

is there even an original implementation of tp_call?

[rugged sparrow]

yea

PyObject_Vectorcall i think is set there

[dry mirage]

I guess there's no way around it then

since it needs to call back into itself

wonder if you can hook it for a function instance though

[rugged sparrow]

probably

lemme check

[dry mirage]

!e

from einspect import view

fn = lambda: 0
v = view(fn)

def new_call(func, args, n_args, kw_names):
     return "hi"

with v.unsafe():
    v.vectorcall = new_call

print(fn())

[night quarry]

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

0

[dry mirage]

hm

[rugged sparrow]

weird

[dry mirage]

seems it's a 3.11 optimization or something?

PyFunctionObjects don't seem to actually get their vectorcall called

!e

from einspect import view

fn = lambda: 0

obj = view(fn)._pyobject
obj.vectorcall = type(obj.vectorcall)()

print(fn())

[night quarry]

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

0

[dry mirage]

In 3.10 you get a

TypeError: 'function' object does not support vectorcall

due to https://github.com/python/cpython/blob/3.11/Objects/call.c#L283-L288

[night quarry]

Objects/call.c lines 283 to 288

if (func == NULL) {
    _PyErr_Format(tstate, PyExc_TypeError,
                  "'%.200s' object does not support vectorcall",
                  Py_TYPE(callable)->tp_name);
    return NULL;
}```

[rugged sparrow]

interesting. seems like i will have to find some other way to hook calls globally

[quartz wave]

i am still not done with my codec project 🥲

[dry mirage]

key=reversed

[dry mirage]

hm

yeah it's only iterable

[low lynx]

use lambda x: tuple(reversed(x))

[night quarry]

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

001 |   1
002 | H 1
003 | b 1
004 | e 1
005 | W 3
006 | d 3
007 | r 3
008 | o 4
009 | l 5
010 | a 7

[solar tulip]

Thanks guys.

[wheat river]

So basically instead of passing args as py tuple it passes it as an array?

[dry mirage]

essentially yeah

kwargs as well

usually those are a dict

[wheat river]

Sweet

[pearl socket]

!e

import re
class _int_t:
    def __init__(self, ln, s):
        self.ln = ln // 8
        self.s = s
    def __gt__(self, other):
        return int.from_bytes(other[0:self.ln], 'little', signed=self.s)
class _reinterpret_cast:
    def __lt__(self, other):
        return True
def create(name):
    unsigned = name[0] == 'u'
    name = name.lstrip('u').replace('_t', '')
    types = ['byte', 'short', 'int', 'long']
    if name.startswith('int'):
        length = re.findall(r'\d+', name)
        if len(length) > 0:
            length = int(length[0])
        else:
            length = 2 ** types.index(name)
    else:
        length = 2 ** types.index(name)
    return _int_t(length, unsigned)
for t in [
    'uint8_t', 'uint16_t', 'uint32_t', 'uint64_t',
    'int8_t', 'int16_t', 'int32_t', 'int64_t',
    'ubyte', 'ushort', 'uint', 'ulong',
    'byte', 'short', 'long'
]:
    globals()[t] = create(t)
reinterpret_cast = _reinterpret_cast()

if __name__ == '__main__':
    print(reinterpret_cast<uint32_t>(b'pytn'))
    print(reinterpret_cast<int32_t>(b'cast'))
    print(reinterpret_cast<uint16_t>(b'to'))
    print(reinterpret_cast<int16_t>(b'in'))
    print(reinterpret_cast<uint8_t>(b't'))
    print(reinterpret_cast<int8_t>(b's'))

[night quarry]

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

001 | 1853127024
002 | 1953718627
003 | 28532
004 | 28265
005 | 116
006 | 115

[pearl socket]

python casting v2

[rugged owl]

Neat use of gt and lt

[quartz wave]

!e ```py
print(sorted([[1, 2], [2], [0, 1]], key=lambda x: (x[:], x.setitem(0, -1))[0] if len(x) == 1 else x))

[night quarry]

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

[[0, 1], [1, 2], [-1]]

[quartz wave]

this is cursed

[fathom hemlock]

guys is this worth of esoteric python?

def phone_number(n):
    return f"({''.join(map(str, n[:3]))}) {''.join(map(str, n[3:6]))}-{''.join(map(str, n[6:]))}"

!e

def phone_number(n):
    return f"({''.join(map(str, n[:3]))}) {''.join(map(str, n[3:6]))}-{''.join(map(str, n[6:]))}"
print(phone_number([1,2,3,4,5,6,7,8,9,0]))

[night quarry]

@fathom hemlock :white_check_mark: Your 3.11 eval job has completed with return code 0.

[phone]

[restive void]

!e

def us_phone_number(n):
    return "({}{}{}) {}{}{}-{}{}{}{}".format(*n)
print(us_phone_number([1,2,3,4,5,6,7,8,9,0]))

[night quarry]

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

[phone]

[earnest wing]

!e

def us_phone_number(n):
    return "(%d%d%d) %d%d%d-%d%d%d%d"%(*n,)
print(us_phone_number([1,2,3,4,5,6,7,8,9,0]))

[night quarry]

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

[phone]

[restive void]

!e

def us_phone_number(n):
    n = int("".join(map(str, n)))
    return f"({n//10**7}) {n%10**7//10**4}-{n%10**4}"
print(us_phone_number([1,2,3,4,5,6,7,8,9,0]))

[night quarry]

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

[phone]

[quartz wave]

!e ```py
a='{}{}{}'.format
f=lambda n:'(%s) %%s-'%a(next(d:=zip([n:=iter(n)]*3)))%a(*next(d))+''.join(map(str,n))
print(f([*range(1,10)]+[0]))

[night quarry]

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

[phone]

[quartz wave]

@fathom hemlock it's just the basics
here's a better one

[fathom hemlock]

that looks so good

[low lynx]

!e inspired by @maiden blade

from einspect import impl

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

def full_adder(a, b, c):
    return a@b@(c@(a@b@a@(a@b@b))),a@b@a@(a@b@b)@c@(a@b@a@(a@b@b))@(c@(c@(a@b@a@(a@b@b))))

def add(a, b):
    c, *n = 0,
    while a|b:
        c, i = full_adder(a & 1, b & 1, c)
        n += [i]
        a >>= 1
        b >>= 1
    return [c, *n[::-1]]

print(add(1, 2), f'{3:b}')
print(add(7, 54), f'{61:b}')
print(add(100, 100), f'{200:b}')

[night quarry]

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

001 | [0, 1, 1] 11
002 | [0, 1, 1, 1, 1, 0, 1] 111101
003 | [1, 1, 0, 0, 1, 0, 0, 0] 11001000

[maiden blade]

this is my first time getting pinged in esoteric area

i feel sad and happy at this time rn

[low lynx]

appreciate the juxtaposition between the somewhat pythonic add and the utter monstrosity that is full_adder

[low lynx]

!e version with less brackets

from einspect import impl
from itertools import product

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

def full_adder(a: int, b: int, c: int):
    return a@b@(a@b@a@(a@b@b)@c),a@b@a@(a@b@b)@c@(a@b@a@(a@b@b))@(a@b@a@(a@b@b)@c@c)

for n in product(range(2), repeat=3):
    print(*n, full_adder(*n))

[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)

[split salmon]

!e ```py
@(lambda a:(lambda:(print("How"),a(),print("does this work?"))[0]))
def func():
print("dafuq")

func()```

[night quarry]

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

001 | How
002 | dafuq
003 | does this work?

[fleet bridge]

isnt it an adder implemented using sheffer stroke?

[calm loom]

Lmfao

[arctic skiff]

15875 chars long

[karmic pumice]

nice

[calm loom]

💀

[acoustic quail]

apparently i did this today but forgorpy a, b, c = int(input()), int(input()), int(input()) print([(q + c + 1 if q < 0 and q + c >= 0 else q + c) for q in [(a - b - 1 if a > 0 and a - b <= 0 else a - b)]][0])

[acoustic quail]

what the [[Hyperlink Blocked]]

[gentle plover]

nand my beloved

[quartz wave]

that red book about CPU internals will surely be an interesting read rn 😔

[meager zinc]

!e ```py
from einspect import impl
from itertools import*
@impl(int)
def matmul(a,b):return a&b^1
for n in product([0,1],repeat=3):print(*n,(lambda a,b,c:(a@b@(a@b@a@(a@b@b)@c),a@b@a@(a@b@b)@c@(a@b@a@(a@b@b))@(a@b@a@(a@b@b)@c@c)))(*n))

[night quarry]

@meager zinc :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)

[meager zinc]

you should implement matmul as a&b^1 for maximum esoteric

[finite blaze]

Anyone up for a golfing challange?

What is the greatest product of four adjacent numbers in the same direction (up, down, left, right, or diagonally) in the 20×20 grid?

[versed eagle]

does it loop around the borders or no

[quartz wave]

what 20x20 grid?

[versed eagle]

like, would grid[18][5] -> grid[19][5] -> grid[0][5] -> grid[1][5] be in a vertical line, for example

[finite blaze]

random one

or i can provide one if u want to compare your results

[quartz wave]

yeah let's go with that

[quartz wave]

ok

[quartz wave]

also ^

[finite blaze]

g = [[8, 2, 22, 97, 38, 15, 00, 40, 00, 75, 4, 5, 7, 78, 52, 12, 50, 77, 91, 8],
[49, 49, 99, 40, 17, 81, 18, 57, 60, 87, 17, 40, 98, 43, 69, 48, 4, 56, 62, 00],
[81, 49, 31, 73, 55, 79, 14, 29, 93, 71, 40, 67, 53, 88, 30, 3, 49, 13, 36, 65],
[52, 70, 95, 23, 4, 60, 11, 42, 69, 24, 68, 56, 1, 32, 56, 71, 37, 2, 36, 91],
[22, 31, 16, 71, 51, 67, 63, 89, 41, 92, 36, 54, 22, 40, 40, 28, 66, 33, 13, 80],
[24, 47, 32, 60, 99, 3, 45, 2, 44, 75, 33, 53, 78, 36, 84, 20, 35, 17, 12, 50],
[32, 98, 81, 28, 64, 23, 67, 10, 26, 38, 40, 67, 59, 54, 70, 66, 18, 38, 64, 70],
[67, 26, 20, 68, 2, 62, 12, 20, 95, 63, 94, 39, 63, 8, 40, 91, 66, 49, 94, 21],
[24, 55, 58, 5, 66, 73, 99, 26, 97, 17, 78, 78, 96, 83, 14, 88, 34, 89, 63, 72],
[21, 36, 23, 9, 75, 00, 76, 44, 20, 45, 35, 14, 00, 61, 33, 97, 34, 31, 33, 95],
[78, 17, 53, 28, 22, 75, 31, 67, 15, 94, 3, 80, 4, 62, 16, 14, 9, 53, 56, 92],
[16, 39, 5, 42, 96, 35, 31, 47, 55, 58, 88, 24, 00, 17, 54, 24, 36, 29, 85, 57],
[86, 56, 00, 48, 35, 71, 89, 7, 5, 44, 44, 37, 44, 60, 21, 58, 51, 54, 17, 58],
[19, 80, 81, 68, 5, 94, 47, 69, 28, 73, 92, 13, 86, 52, 17, 77, 4, 89, 55, 40],
[4, 52, 8, 83, 97, 35, 99, 16, 7, 97, 57, 32, 16, 26, 26, 79, 33, 27, 98, 66],
[88, 36, 68, 87, 57, 62, 20, 72, 3, 46, 33, 67, 46, 55, 12, 32, 63, 93, 53, 69],
[4, 42, 16, 73, 38, 25, 39, 11, 24, 94, 72, 18, 8, 46, 29, 32, 40, 62, 76, 36],
[20, 69, 36, 41, 72, 30, 23, 88, 34, 62, 99, 69, 82, 67, 59, 85, 74, 4, 36, 16],
[20, 73, 35, 29, 78, 31, 90, 1, 74, 31, 49, 71, 48, 86, 81, 16, 23, 57, 5, 54],
[1, 70, 54, 71, 83, 51, 54, 69, 16, 92, 33, 48, 61, 43, 52, 1, 89, 19, 67, 48]]

[versed eagle]

@finite blaze

[finite blaze]

nope

[versed eagle]

alr

[meager zinc]

what would it be then?

also are there any negative numbers?

[finite blaze]

only positive numbers

[versed eagle]

are they all integers

[finite blaze]

yes

[quartz wave]

what's the range?

[meager zinc]

ok so we can ignore rows that are less than 4 big entirely

[quartz wave]

random.randint(1, 100)?

[finite blaze]

0-100

[quartz wave]

ok

inclusive?

[finite blaze]

exclusive

[quartz wave]

ok

are we gonna go with brute force first

[meager zinc]

ok im dumb

or something

[finite blaze]

i can provide a bruteforce base for golfing if u want

[quartz wave]

!e ```py
import random as r
print([[r.randrange(100) for _ in range(20)] for _ in range(20)])

[night quarry]

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

[[83, 32, 84, 36, 65, 99, 80, 26, 87, 22, 50, 20, 7, 23, 80, 82, 92, 42, 29, 45], [16, 73, 61, 60, 32, 22, 96, 9, 16, 52, 28, 46, 6, 51, 87, 52, 4, 36, 29, 77], [29, 79, 45, 61, 18, 95, 82, 93, 52, 83, 23, 4, 35, 21, 85, 66, 17, 76, 58, 71], [16, 86, 77, 50, 7, 52, 73, 30, 56, 77, 13, 65, 75, 84, 27, 96, 64, 66, 31, 52], [10, 9, 87, 95, 48, 87, 43, 95, 88, 20, 70, 95, 15, 79, 38, 32, 41, 65, 26, 31], [0, 47, 83, 18, 14, 24, 97, 83, 13, 13, 10, 69, 56, 39, 46, 54, 26, 6, 45, 49], [4, 64, 61, 25, 18, 64, 5, 95, 15, 32, 31, 55, 34, 8, 93, 17, 90, 64, 19, 26], [16, 15, 84, 13, 71, 80, 82, 20, 62, 36, 97, 24, 5, 37, 5, 91, 84, 79, 30, 0], [13, 67, 57, 11, 9, 94, 53, 63, 72, 14, 30, 45, 43, 9, 86, 33, 76, 2, 36, 87], [4, 4, 9, 85, 70, 44, 46, 67, 89, 68, 32, 60, 64, 89, 48, 72, 45, 14, 99, 72], [96, 60, 10, 38, 0, 76, 47, 56, 40, 77, 34, 96, 17, 66, 6, 97, 37, 86, 49, 24], [85, 28, 72, 92, 17, 70, 86, 35, 21, 57, 46, 92, 8, 83, 8, 96, 61, 32, 0, 89], [42, 21, 41, 13, 17, 16, 27, 69, 91, 7, 7
... (truncated - too long)

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

[meager zinc]

yes

this is probably (maybe) dynamic programming and likely sliding window

for optimal

but of course we aren't looking for optimal

[quartz wave]

sliding window is a good idea

[meager zinc]

this is also embarrassingly parallel

harold are you making a base or do I need to

[versed eagle]

im making something

but you should probably make your own since im gonna be a bit

im busy and distracted so it'll be a while before i have something working

[meager zinc]

ok so its just the maximum 4x1 column?

[versed eagle]

@finite blaze do you have an input with a known correct output

[meager zinc]

oh no all directions

I see

[finite blaze]

yes

[versed eagle]

so that we can test whatever we have

once we get something

[finite blaze]

^ input, answer:

||70600674||

[quartz wave]

base, 302 ```py
import math
p=math.prod;R=range(17)
def f(x):
r=0
for A in x+[*zip(*x)]:r=max(r,*map(p,zip(A,A[1:],A[2:],A[3:])))
for i in R:
a,b,c,d=x[i:i+4]
for j in R:n=r=max(r,a[j]*b[j+1]*c[j+2]*d[j+3])
for i in R:
a,b,c,d=x[i:i+4]
for j in range(19,2,-1):n=r=max(r,a[j]*b[j-1]*c[j-2]*d[j-3])
return r

[meager zinc]

ok i was too slow

[quartz wave]

don't assign math.prod() to a variable, 298 ```py
import math
R=range(17)
def f(x):
r=0
for A in x+[*zip(*x)]:r=max(r,*map(math.prod,zip(A,A[1:],A[2:],A[3:])))
for i in R:
a,b,c,d=x[i:i+4]
for j in R:n=r=max(r,a[j]*b[j+1]*c[j+2]*d[j+3])
for i in R:
a,b,c,d=x[i:i+4]
for j in range(19,2,-1):n=r=max(r,a[j]*b[j-1]*c[j-2]*d[j-3])
return r

inline second for, 262 ```py
import math
R=range(17)
def f(x):
r=0
for A in x+[*zip(*x)]:r=max(r,*map(math.prod,zip(A,A[1:],A[2:],A[3:])))
for i in R:
a,b,c,d=x[i:i+4]
for j in R:r=max(r,a[j]*b[j+1]*c[j+2]*d[j+3])
for j in range(19,2,-1):r=max(r,a[j]*b[j-1]*c[j-2]*d[j-3])
return r

[meager zinc]

is a[j]*b[j+1]*c[j+2]*d[j+3] really smaller than math.prod(x[i:i+4])

[quartz wave]

hmm good one

[meager zinc]

what I thought might be easier is to transpose the list and repeat the same stuff as the row calcluations

[quartz wave]

various optimizations, 224 ```py
import math
p=math.prod;R=range(17)
def f(x):
r=max(sum([[*map(p,zip(A,A[1:],A[2:],A[3:]))]for A in x+[zip(x)]],[]))
for i in R:X=x[i:i+4];r=max(r,[p(X[j:j+4])for j in R],[p(X[j-4:j])for j in range(20,3,-1)])
return r

[meager zinc]

because then it would have a lot of duplicates

[quartz wave]

‫nvm wait

[meager zinc]

the multiplication doesn't work my bad

[quartz wave]

revert, 258 ```py
import math
p=math.prod;R=range(17)
def f(x):
r=max(sum([[*map(p,zip(A,A[1:],A[2:],A[3:]))]for A in x+[*zip(x)]],[]))
for i in R:a,b,c,d=x[i:i+4];r=max(r,[a[j]*b[j+1]*c[j+2]d[j+3]for j in R],[a[j]*b[j-1]*c[j-2]*d[j-3]for j in range(19,2,-1)])
return r

[finite blaze]

import math
r=0
for x in range(0,20):
 for y in range(0,17):
  d=a=0
  v=math.prod(g[x][y:y+3])
  h=g[y][x]*g[y+1][x]*g[y+2][x]*g[y+3][x]
  if x<17:d=g[x][y]*g[x+1][y+1]*g[x+2][y+2]*g[x+3][y+3]
  if x>2:a=g[x][y]*g[x-1][y+1]*g[x-2][y+2]*g[x-3][y+3]
  r=max([v,h,d,a,r])
print(r)

and thats my base with barely any optimisations, 278 chars

[meager zinc]

interesting

using r=max([v,h,d,a,r])

[quartz wave]

a little optimization, 256 ```py
import math
p=math.prod;R=range(17);t=0,1,2,3
def f(x):
r=max(sum([[*map(p,zip(A,A[1:],A[2:],A[3:]))]for A in x+[*zip(*x)]],[]))
for i in R:X=x[i:i+4];r=max([r]+[p(X[k][j+k]for k in t)for j in R]+[p(X[k][j-k]for k in t)for j in range(19,2,-1)])
return r

wait why do we have a function

[meager zinc]

are we not recursing?

[quartz wave]

no we aren't

[meager zinc]

no wearent

ok we can remove it

[quartz wave]

so now we manually take input

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

reformat, 285 ```py
import math
p=math.prod;R=range(17);t=0,1,2,3;x=[[*map(int,x.split())]for x in open(0)]
r=max(sum([[*map(p,zip(A,A[1:],A[2:],A[3:]))]for A in x+[*zip(*x)]],[]))
for i in R:X=x[i:i+4];r=max([r]+[p(X[k][j+k]for k in t)for j in R]+[p(X[k][j-k]for k in t)for j in range(19,2,-1)])
print(r)

[meager zinc]

this assumes input as variable g

[finite blaze]

yep, I guess we could ignore the array part

and just count the algorithm

[quartz wave]

optimized, 277 ```py
import math
r=0;g=[[*map(int,x.split())]for x in open(0)]
for x in range(20):
for y in range(17):r=max(r,math.prod(g[x][y:y+3]),g[y][x]*g[y+1][x]*g[y+2][x]*g[y+3][x],x>16or g[x][y]*g[x+1][y+1]*g[x+2][y+2]*g[x+3][y+3],x<3or g[x][y]*g[x-1][y+1]*g[x-2][y+2]*g[x-3][y+3])
print(r)

[meager zinc]

there's bound to be some way ot compress all these:
g[y][x]*g[y+1][x]*g[y+2][x]*g[y+3][x]
g[x][y]*g[x+1][y+1]*g[x+2][y+2]*g[x+3][y+3]
g[x][y]*g[x-1][y+1]*g[x-2][y+2]*g[x-3][y+3]

they are very similar but ever so slightly differnet

[quartz wave]

yep

[meager zinc]

I tried using a loop with (-1,1) for the last two but it didn't work for some reason

[quartz wave]

compressed these, 238 ```py
import math
r=0;g=[[*map(int,x.split())]for x in open(0)];t=0,1,2,3;p=math.prod
for x in range(20):
for y in range(17):r=max(r,p(g[x][y:y+3]),p(g[y+k][x]for k in t),x>16or p(g[x+k][y+k]for k in t),x<3or p(g[x-k][y+k]for k in t))
print(r)

[meager zinc]

nicee

[meager zinc]

assuming input like this, 210:

import math
r=0;exec(open(0));t=0,1,2,3;p=math.prod
for x in range(20):
 for y in range(17):r=max(r,p(g[x][y:y+3]),p(g[y+k][x]for k in t),x>16or p(g[x+k][y+k]for k in t),x<3or p(g[x-k][y+k]for k in t))
print(r)

or, similar but as a list (212):

import math
r=0;g=eval(open(0));t=0,1,2,3;p=math.prod
for x in range(20):
 for y in range(17):r=max(r,p(g[x][y:y+3]),p(g[y+k][x]for k in t),x>16or p(g[x+k][y+k]for k in t),x<3or p(g[x-k][y+k]for k in t))
print(r)

[finite blaze]

import math
r=0;exec(open(0));t=0,1,2,3;p=math.prod
for x,y in zip(range(20),range(17):r=max(r,p(g[x][y:y+3]),p(g[y+k][x]for k in t),x>16or p(g[x+k][y+k]for k in t),x<3or p(g[x-k][y+k]for k in t))
print(r)

maybe something like this?

[quartz wave]

nope

zip() isn't equivalent to nesting

[finite blaze]

👍

[quartz wave]

merge all the for k in ts, 226 ```py
import math
r=0;g=[[*map(int,x.split())]for x in open(0)];p=math.prod
for x in range(20):
for y in range(17):r=max(r,p(g[x][y:y+3]),map(p,zip([(g[y+k][x],x<17and g[x+k][y+k],x>2and g[x-k][y+k])for k in(0,1,2,3)])))
print(r)

walrus, 225 ```py
import math
r=0;g=[[*map(int,x.split())]for x in open(0)];p=math.prod
for x in range(20):
for y in range(17):r=max(r,p(g[x][y:y+3]),map(p,zip([(g[d:=y+k][x],x<17and g[x+k][d],x>2and g[x-k][d])for k in(0,1,2,3)])))
print(r)

[low lynx]

I think sheffer strokes are nand gates

[quartz wave]

oh i found it

‫it just describes how to get the carry from the addition at the ith bit

[glass stone]

wow ya'll crazy

when you say y:y+3 is that a lambda?

[quartz wave]

sum is (not a)&(not b)&c | (not a)&b&(not c) | a&(not b)&(not c) | a&b&c and carry is b&c | x&c | x&y

[quartz wave]

no it's a slice

lambdas are prefixed with lambda

[glass stone]

neat. I can tell r is important but that is all I can make sense of

thought so, i am scacred of lambdas so i don't use them ever but trying to make sense of that nested for loop

is that a list comprehension or more slicing

it hurts my brain to look at is it safe to run for my ubuntu?

[vapid finch]

Don't be scared. lambda is nearly the same as def.

[glass stone]

silence of the lambdas

[meager zinc]

mwhahaha (this only works on windows cpython, btw)

(lambda: (lambda _:(lambda __:(lambda ___:(lambda ____:[(lambda _____:__.fputs(f'Hello {(x:=(lambda:_).__name__):.8}'.encode(),_____)&__.fflush(_____)&__.fclose(_____)&__._exit(o:=0))(__._fdopen(____,'w'))for __._fdopen.restype in(_.POINTER(_.c_char),)][o])(__._open_osfhandle(___,8)))(__._get_osfhandle(-~0)))(_.CDLL(_.util.find_library('msvcrt'))))(__import__('ctypes.util')))()

Output:

Hello <lambda>!

[old socket]

LInux users when msvcrt

[meager zinc]

this only works on windows cpython
me - 3 seconds ago

[old socket]

I know

[rain herald]

💀

[glass stone]

only works on windows im safe using solaris

ya'll are amazing thanks for being here ❤️

[finite blaze]

Alright, i guess thats the final version

[low lynx]

no

221, assigned range to a variable

import math
r=0;g=[[*map(int,x.split())]for x in open(0)];p=math.prod;R=range
for x in R(20):
 for y in R(17):r=max(r,p(g[x][y:y+3]),*map(p,zip(*[(g[d:=y+k][x],x<17and g[x+k][d],x>2and g[x-k][d])for k in R(4)])))
print(r)

[night quarry]

@olive bramble Per Rule 6, your invite link has been removed. If you believe this was a mistake, please let staff know!

Our server rules can be found here: https://pythondiscord.com/pages/rules

[crisp fog]

Today I was rummaging through other people's sources and found just such a thing.
The guy invented strptime

[fleet bridge]

my brute-force solution for simple school problem (find 5 numbers that sum up to 2017 and have no common digits): ```py
t = 2017
for n1 in range(1000, t):
s1 = set(t1 := str(n1))
if len(s1) != len(t1): continue
for n2 in range(n1):
s2 = set(t2 := str(n2))
if len(s2) != len(t2): continue
if not s1.isdisjoint(s2): continue
for n3 in range(n2):
s3 = set(t3 := str(n3))
if len(s3) != len(t3): continue
if not s1.isdisjoint(s3): continue
if not s2.isdisjoint(s3): continue
for n4 in range(n3):
s4 = set(t4 := str(n4))
if len(s4) != len(t4): continue
if not s1.isdisjoint(s4): continue
if not s2.isdisjoint(s4): continue
if not s3.isdisjoint(s4): continue
for n5 in range(n4):
s5 = set(t5 := str(n5))
if len(s5) != len(t5): continue
if not s1.isdisjoint(s5): continue
if not s2.isdisjoint(s5): continue
if not s3.isdisjoint(s5): continue
if not s4.isdisjoint(s5): continue
if n1 + n2 + n3 + n4 + n5 == t:
print(n1, n2, n3, n4, n5)

all lines here (except first and two last) have some optimisations or are optimisations themselves

[finite blaze]

t=2017;l=len;s=set;R=range;S=str
for n1 in R(1000,t):
 s1=s(t1:=S(n1))
 if l(s1)!=l(t1):continue
 for n2 in R(n1):
  s2=s(t2:=S(n2))
  if l(s2)!=l(t2):continue
  if not s1.isdisjoint(s2):continue
  for n3 in R(n2):
   s3=s(t3:=S(n3))
   if not s1.isdisjoint(s3) or not s2.isdisjoint(s3)or l(s3)!=l(t3):continue
   for n4 in R(n3):
    s4=s(t4:=S(n4))
    if not s1.isdisjoint(s4)or not s2.isdisjoint(s4)or not s3.isdisjoint(s4)or l(s4)!=l(t4):continue
    for n5 in R(n4):
     s5=s(t5:=S(n5))
     if not s1.isdisjoint(s5)or not s2.isdisjoint(s5)or not s3.isdisjoint(s5)or not s4.isdisjoint(s5)or l(s5)!=l(t5):continue
     if n1+n2+n3+n4+n5==t:print(n1,n2,n3,n4,n5) 

^ some really basic optimisations. without changing how the algorithm works
1170 to 668 chars

[gleaming linden]

You could probably demorgan some of those conditions

And assign isdisjoint to a variable

[finite blaze]

I tried doing that

but it doesn't work

[gleaming linden]

d=set.isdisjoint and d(s1,s2) should work

[finite blaze]

oh you can do things like this

good to know

[gleaming linden]

Yes, instance.method(param) is just class_of_instance.method(instance, param)

[finite blaze]

noted

t=2017;l=len;s=set;R=range;S=str;d=set.isdisjoint
for n1 in R(1000,t):
 s1=s(t1:=S(n1))
 if l(s1)!=l(t1):continue
 for n2 in R(n1):
  s2=s(t2:=S(n2))
  if l(s2)!=l(t2) or not s1.d(s2):continue
  for n3 in R(n2):
   s3=s(t3:=S(n3))
   if not s1.d(s3) or not s2.d(s3)or l(s3)!=l(t3):continue
   for n4 in R(n3):
    s4=s(t4:=S(n4))
    if not s1.d(s4)or not s2.d(s4)or not s3.d(s4)or l(s4)!=l(t4):continue
    for n5 in R(n4):
     s5=s(t5:=S(n5))
     if not s1.d(s5)or not s2.d(s5)or not s3.d(s5)or not s4.d(s5)or l(s5)!=l(t5):continue
     if n1+n2+n3+n4+n5==t:print(n1,n2,n3,n4,n5) 

down to 584 chars

[gleaming linden]

Demorgan

not a or not b is not(a and b)

[finite blaze]

what does "demorgan" mean?

[gleaming linden]

https://en.m.wikipedia.org/wiki/De_Morgan's_laws

Basically just a and b is the same as not(not a or not b) and a or b is the same as not(not a and not b)

[finite blaze]

will it work for more conditions?

[gleaming linden]

Yeah

[finite blaze]

like not a or not b or not c or not d

okay, cool

down to 567

t=2017;l=len;s=set;R=range;S=str;d=set.isdisjoint
for n1 in R(1000,t):
 s1=s(t1:=S(n1))
 if l(s1)!=l(t1):continue
 for n2 in R(n1):
  s2=s(t2:=S(n2))
  if l(s2)!=l(t2)or not s1.d(s2):continue
  for n3 in R(n2):
   s3=s(t3:=S(n3))
   if not(s1.d(s3)and s2.d(s3))or l(s3)!=l(t3):continue
   for n4 in R(n3):
    s4=s(t4:=S(n4))
    if not(s1.d(s4)and s2.d(s4)and s3.d(s4))or l(s4)!=l(t4):continue
    for n5 in R(n4):
     s5=s(t5:=S(n5))
     if not(s1.d(s5)and s2.d(s5)and s3.d(s5)and s4.d(s5))or l(s5)!=l(t5):continue
     if n1+n2+n3+n4+n5==t:print(n1,n2,n3,n4,n5) 

[gleaming linden]

I feel like this can be a recursive function instead of nested loops

[finite blaze]

yeah, I didn't want to change the initial algorithm that much

i am leaving this for someone else ;p

[gleaming linden]

all the x.d(y) should be d(x,y) btw

[gleaming linden]

from itertools import*
for N in range(11):
 for C in combinations(range(10),N):
  for P in product(range(4),repeat=N):
   c=P.count
   if not 5==c(0)>=c(1)>=c(2)>=c(3):continue
   if sum(a:=[d*10**v for v,d in zip(P,C)])==2017:
    R=[0]*5
    for d,v in zip(a,P):
     R[[*map(lambda x:0==x//10**v%10,R)].index(1)]+=d
    print(*R);e

basically goes through every possible combination of digits, then goes through every permutation of powers of 10 for each digit (for place values) and checks if the sum is 2017

the result includes a 0 though, not sure if that counts

[finite blaze]

Won't declaring range as a variable save some chars?

[gleaming linden]

yes, this is just a base golf for now

[finite blaze]

👍

[meager zinc]

you should assign range to a variable:

from itertools import*
r=range
for N in r(11):
 for C in combinations(r(10),N):
  for P in product(r(4),repeat=N):
   c=P.count
   if not 5==c(0)>=c(1)>=c(2)>=c(3):continue
   if sum(a:=[d*10**v for v,d in zip(P,C)])==2017:
    R=[0]*5
    for d,v in zip(a,P):
     R[[*map(lambda x:0==x//10**v%10,R)].index(1)]+=d
    print(*R);e

also, instead of using not + continue can we just use elif?

or use and not?

[gleaming linden]

you can combine the ifs

if c(3)<=c(2)<=c(1)<=c(0)==5<2017==sum(V:=[j*10**i for i,j in zip(P,C)]):

[meager zinc]

from itertools import*
r=range
for N in r(11):
 for C in combinations(r(10),N):
  for P in product(r(4),repeat=N):
   c=P.count
   if sum(a:=[d*10**v for v,d in zip(P,C)])==2017and not 5==c(0)>=c(1)>=c(2)>=c(3):
    R=[0]*5
    for d,v in zip(a,P):
     R[[*map(lambda x:0==x//10**v%10,R)].index(1)]+=d
    print(*R);e

we can also remove a bunch of indentation

from itertools import*
r=range
for N in r(11):
 for C in combinations(r(10),N):
  for P in product(r(4),repeat=N):
   c=P.count
   if c[3]<=c[2]<=c[1]<=c[0]==5<2017==sum(V:=[j*10**i for i,j in zip(P,C)]):
    R=[0]*5
    for d,v in zip(a,P):R[[*map(lambda x:0==x//10**v%10,R)].index(1)]+=d
    print(*R);e

what does the last ;e do?

[gleaming linden]

it causes an error

so the code terminates

[meager zinc]

oh to short circuit?

lol I see

[gleaming linden]

stderr doesnt matter so why not

[meager zinc]

can we just do print(*R)e or is that a syntax error

[gleaming linden]

I think you can just rename R to N

and it will error on the next iteration because it changes N

[meager zinc]

smart

from itertools import*
r=range
for N in r(11):
 for C in combinations(r(10),N):
  for P in product(r(4),repeat=N):
   c=P.count
   if c[3]<=c[2]<=c[1]<=c[0]==5<2017==sum(V:=[j*10**i for i,j in zip(P,C)]):
    N=[0]*5
    for d,v in zip(a,P):N[[*map(lambda x:0==x//10**v%10,N)].index(1)]+=d
    print(*N)

[gleaming linden]

has to be c(x) instead of c[x]

[meager zinc]

no difference in characters, but more esoteric

from itertools import*
r=range
for N in r(11):
 for C in combinations(r(t:=10),N):
  for P in product(r(4),repeat=N):
   c=P.count
   if c(3)<=c(2)<=c(1)<=c(0)==5<2017==sum([j*t**i for i,j in zip(P,C)]):
    N=[0]*5
    for d,v in zip(a,P):N[[*map(lambda x:0==x//t**v%t,N)].index(1)]+=d
    print(*N)

[gleaming linden]

and get rid of the V:= walrus

[meager zinc]

I can't think of anything else

we would probably have to change how we are doing it to get much better

[gleaming linden]

I'm rewriting the end conversion bit

to use groupby

[meager zinc]

ok

[finite blaze]

@fleet bridge look what they did to your code 😿

[meager zinc]

:)

[finite blaze]

well, i touched it too ;p

[meager zinc]

wait what is variable a?

oh I see

we removed the walrus

at some point

fixed```py
from itertools import*
r=range
for N in r(11):
for C in combinations(r(t:=10),N):
for P in product(r(4),repeat=N):
c=P.count
if c(3)<=c(2)<=c(1)<=c(0)==5<2017==sum(a:=[j*ti for i,j in zip(P,C)]):
N=[0]*5
for d,v in zip(a,P):N[[*map(lambda x:0==x//tv%t,N)].index(1)]+=d
print(*N)

note that it outputs:

1942 63 5 7 0
1932 74 5 6 0
Traceback (most recent call last):
    for P in product(r(4),repeat=N):
TypeError: 'list' object cannot be interpreted as an integer

[gleaming linden]

oh, interesting

I can't get groupby to work because you need to sort somewhere for groupby to work

I'll go sleep now, gn everyone

[finite blaze]

o/

[meager zinc]

gn

[finite blaze]

sum = 0
for y in range(len(g)-1,0,-1): 
    for x in range(0,len(g[y-1])):
        g[y-1][x]+=max(g[y][x], g[y][x+1])

^ trying to oneline this

[g[y-1][x]+=max(g[y][x], g[y][x+1])for x in range(0,len(g[y-1])) for y in range(len(g)-1,0,-1)]

but it doesnt work, += invalid syntax

any ideas what i might be doing wrong here?

[fleet bridge]

i love you guys

you are amazing

[wispy ivy]

Hello, can any of the smart brain here golf the code below ? I really don't mind about readability I want to see how short this code can become

def balance(s):
    stack = []
    for c in s:
        if c in "([":
            stack.append(c)
        elif c in ")]":
            if not stack:
                return 0
            if c == ")" and stack[-1] != "(":
                return 0
            if c == "]" and stack[-1] != "[":
                return 0
            stack.pop()
    return 1 if not stack else 0

s = input()
print(balance(s))

Ty, if you ever take the time to work on it

[finite blaze]

what is this supposed to do?

[rugged owl]

I mean this is the most I can do without knowing what the program does

def b(s):
 return 1-(len([c for c in s if c in "(["]) != len([c for c in s if c in ")]"]))

[proper vault]

I don't think that one works, the program checks if the parentheses are balanced and matching.

[wispy ivy]

Sorry the program is given a string of parentheses of two different types is balanced, meaning that the opening and corresponding closing parentheses are properly nested
And print 1 if it is and 0 if it's not

[proper vault]

i=[];s=input()
for c in s:i+=')]'['(['.index(c)]if c in'(['else(0*i.pop()if i[-1:] ==[c]else c)if c in')]'else[]
print('01'[not i])
```this seems to work

you could probably do better using and and or instead of .. if .. else ..

but I CBA to figure that out

[earnest wing]

import re
S=re.sub
try:compile(S("\[","[0",S("[^\[\]\(\)]","",input())),"","exec")
except:exit("0")
print(1)

if ()[] are matching, then they are valid python syntax as long as [] is replaced with [something]

(because you can't index without args)

[proper vault]

non-parentheses are also allowed in the input

[earnest wing]

that's what the first S() does

[proper vault]

ah, I C now

[earnest wing]

at least in theory

haven't tested this

[wispy ivy]

You guys are cracked, super impressive

[earnest wing]

ah i forgot compile requires other args

this should work tm

it'll throw syntax warnings and clobber the return code fyi

so I recommend python3 -Wignore

xd

[plush halo]

!e```py
a = []
a.append(a)
print(a)
while True:
a = a[0]

[night quarry]

@plush halo :x: Your 3.11 eval job timed out or ran out of memory.

[[...]]

[static marsh]

This is like the first time I've tried this stuff in python but how is it lol ```py
import sys;

class String():
init: None = lambda self, _str: setattr(self, '_str', _str);
string: list[int] = property(lambda self: [ord(self._str[i]) for i in range(len(self._str))], lambda self, v: setattr(self, '_str', ''.join(chr(i) for i in v)));
getitem: str = lambda self, i: [''.join(map(lambda *args, **kwargs: chr(*args, **kwargs), [self.string[i]]))][0];
iter: list[any] = lambda self: [self.string].iter();
str: str = lambda self: ''.join([''.join(i) for i in [self.getitem(i) for i in range(self.string.len())]]);
add = lambda self, other: ['teehee', self][exec(('String'!=type(other).name)"raise Exception('type must be String')")==self.setattr('string', ([v for v in self.string]+other.string))];

out: String = String("Hello, ") + String("World!") + String(" Hello2");
sys.stdout.write(f"{out}\n");

I got some help from cereal

it was fun to make

ngl

[low lynx]

I don't think list[any] as a type hint does what you think it does

[static marsh]

I know 😩 idk how to specify return type on lambda's

😭

[low lynx]

you'd type hint it as a Callable not a list

[static marsh]

yeah I didn't really want to type hint the lambda but the return type of the lambda

[quartz wave]

!e ```py
import sys;

class String:
(init := lambda self, _str: setattr(self, '_str', _str)).annotations|={'return':None};
(string := property(lambda self: [ord(self._str[i]) for i in range(len(self._str))], lambda self, v: setattr(self, '_str', ''.join(chr(i) for i in v)))).annotations|={'return':list[int]};
(getitem := lambda self, i: [''.join(map(lambda *args, **kwargs: chr(*args, **kwargs), [self.string[i]]))][0]).annotations|={'return':str};
(iter := lambda self: [self.string].iter()).annotations|={'return':list[any]};
(str := lambda self: ''.join([''.join(i) for i in [self.getitem(i) for i in range(self.string.len())]])).annotations|={'return':str};
add = lambda self, other: ['teehee', self][exec(('String'!=type(other).name)"raise Exception('type must be String')")==self.setattr('string', ([v for v in self.string]+other.string))];

out: String = String("Hello, ") + String("World!") + String(" Hello2");
sys.stdout.write(f"{out}\n");

[night quarry]

@quartz wave :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 |   File "<string>", line 5, in String
004 | AttributeError: 'property' object has no attribute '__annotations__'

[quartz wave]

aww

[static marsh]

oh ahh

I see

what does the walrus thingy do

I have no clue

[low lynx]

assignment expression

[sick hound]

!e you forgot .getter ```py
import sys;

class String:
(init := lambda self, _str: setattr(self, '_str', _str)).annotations|={'return':None};
(string := property(lambda self: [ord(self._str[i]) for i in range(len(self._str))], lambda self, v: setattr(self, '_str', ''.join(chr(i) for i in v)))).getter().annotations|={'return':list[int]};
(getitem := lambda self, i: [''.join(map(lambda *args, **kwargs: chr(*args, **kwargs), [self.string[i]]))][0]).annotations|={'return':str};
(iter := lambda self: [self.string].iter()).annotations|={'return':list[any]};
(str := lambda self: ''.join([''.join(i) for i in [self.getitem(i) for i in range(self.string.len())]])).annotations|={'return':str};
add = lambda self, other: ['teehee', self][exec(('String'!=type(other).name)"raise Exception('type must be String')")==self.setattr('string', ([v for v in self.string]+other.string))];

out: String = String("Hello, ") + String("World!") + String(" Hello2");
sys.stdout.write(f"{out}\n");

[night quarry]

@sick hound :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 |   File "<string>", line 5, in String
004 | TypeError: property.getter() takes exactly one argument (0 given)

[sick hound]

eh still doesn't work

[versed eagle]

!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,0)or cvp_fa(gfp(F)).value),None)[~0]))
sw_f(sum,print)
sw_f(print,repr)
sum(1, print(1), repr(1))

[night quarry]

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

1 <I object at 0x7fc73e770090> 1

[versed eagle]

can someone explain this behaviour?

(why does print(1) return something different than repr(1))

[sick hound]

print() produces output to a file (usually sys.stdout) and returns None, repr() calls the object's __repr__ dunder and returns a string

[versed eagle]

i know what the functions do

im asking because i've swapped the function pointer for print

to be the function pointer for repr

so now they should be the same

!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,0)or cvp_fa(gfp(F)).value),None)[~0]))
sw_f(sum,print)
sw_f(max,min)
sw_f(min,max)
sum(min(1, 2, 3), max(1, 2, 3))

[night quarry]

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

3 1

[versed eagle]

see here, I've swapped min and max

as well as using sum for print

[versed eagle]

the question here is "why is the behaviour different when the function pointers are the same"

[dry mirage]

you swapped sum and print first right?

[versed eagle]

i dont swap them
its 1 way

[dry mirage]

so swapping print and repr later should make repr work as sum?

oh

[dry mirage]

!e

from einspect import view

view(sum) << print
view(print) << repr

sum(1, print(1), repr(1))

[night quarry]

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

1 1 1

[dry mirage]

seems fine here

[versed eagle]

do you switch more than just the pointer?

all im doing is switching the pointer for the functions

[dry mirage]

what pointer?

[versed eagle]

the pointer to the c function

[dry mirage]

where?

I mean in the struct

[versed eagle]

uh
the offset i used is 16

dont remember how i got to that number

[dry mirage]

16 is like... ob_type?

[versed eagle]

8 is type iirc

[dry mirage]

https://github.com/ionite34/einspect/blob/main/src/einspect/structs/py_cfunction.py#L15-L21

[night quarry]

src/einspect/structs/py_cfunction.py lines 15 to 21

@struct
class PyCFunctionObject(PyObject[_T, None, None]):
    m_ml: ptr[PyMethodDef]  # Description of the C function to call
    m_self: ptr[PyObject]  # Passed as 'self' arg to the C func, can be NULL
    m_module: ptr[PyObject]  # The __module__ attribute, can be anything
    m_weakreflist: ptr[PyObject]  # List of weak references
    vectorcall: vectorcallfunc```

[dry mirage]

oh it's the m_ml then

yeah you have to change vectorcall as well

sum and print both take *args

but repr only takes 1 arg

so you need to change its vectorcall as well

[versed eagle]

ah

[dry mirage]

!e

from einspect import view

view(sum).ml.ml_meth = view(print).ml.ml_meth
view(print).ml.ml_meth = view(repr).ml.ml_meth
view(print)._pyobject.vectorcall = view(repr)._pyobject.vectorcall

sum(1, print(1), repr(1))

[night quarry]

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

1 1 1

[versed eagle]

ty

[dry mirage]

the << move will move the entire object minus ob_refcnt

[wispy ivy]

Sorry for the late reply, but there is something wrong no ?
For example if you double the imbalanced input it will output 1 even if it shouldn't
Example : [])[])

[proper vault]

ah, yeah, seems broken

mb

[rugged owl]

You dont have to use unsafe here?

[dry mirage]

the objects are the same size

[rugged owl]

ah,

[dry mirage]

well, also neither have an instance dict, and they're all gc types

!e so a normal func to a builtin func would be unsafe

from einspect import view

view(repr) << (lambda x: int(x))

[night quarry]

@dry mirage :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 |   File "/snekbox/user_base/lib/python3.11/site-packages/einspect/views/view_base.py", line 392, in __lshift__
004 |     return self.move_from(other)
005 |            ^^^^^^^^^^^^^^^^^^^^^
006 |   File "/snekbox/user_base/lib/python3.11/site-packages/einspect/views/view_base.py", line 302, in move_from
007 |     check_move(self, other)
008 |   File "/snekbox/user_base/lib/python3.11/site-packages/einspect/views/_moves.py", line 56, in check_move
009 |     raise UnsafeError(
010 | einspect.errors.UnsafeError: memory move of instance dict at offset 88 from 'function' to 'builtin_function_or_method' is out of bounds. Enter an unsafe context to allow this.

[proper vault]

i=[];s=input()
for c in s:i+=')]'['(['.index(c)]if c in'(['else(0*i.pop()if i[-1:]==[c]else'_')if c in')]'else[]
print('01'[not i])
```this one should work at least a bit better.

[wispy ivy]

Works wonders ty !

[long fulcrum]

[print(i)for i in range(51)if~i.bit_count()&1]```
can I golf this further?

[earnest wing]

!e [print(i)for i in range(51)if~i.bit_count()&1]

[night quarry]

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

001 | 0
002 | 3
003 | 5
004 | 6
005 | 9
006 | 10
007 | 12
008 | 15
009 | 17
010 | 18
011 | 20
... (truncated - too many lines)

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

[long fulcrum]

prints numbers that have an even amount of 1s in its binary version

[twin reef]

A slightly different idea, probably can be optimized further:

s=set();[s.add(x)for x in range(51)if~-x&x not in s];print(*s)

[long fulcrum]

thats 25 chars longer though

i had to add ,sep="\n"

[twin reef]

I know😢

[long fulcrum]

the print can be changed to *map(print,s),

[versed eagle]

would have to be inside brackets, parens, or braces

[rugged sparrow]

@dry mirage want to help me work on making fishhook.asm work on windows later tonight?

I also want to add in a patchfinder to let it apply assembly hooks into the middle of functions

[dry mirage]

👍 sure

[glacial lantern]

import __hello__

No longer prints "Hello world!" in 3.11
You have to do:

import __hello__
__hello__.main()

This ruined my day.

[dry mirage]

!e import __hello__

[night quarry]

@dry mirage :white_check_mark: Your 3.10 eval job has completed with return code 0.

Hello world!

[dry mirage]

hm

[glacial lantern]

3.11

[last locust]

!e __import__('__hello__').main()

[night quarry]

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

Hello world!

[last locust]

Thankfully you can at least still one-line it

[glacial lantern]

That's longer than just

print("Hello world!")

now though.
I wonder why this change happened.

[glad belfry]

I am not sure where to place this issue

I only just started working with python about 2 months ago

I've created a small save backup tool

and I am trying to use pyinstaller to create an exe for it

but it keeps missing one particular module over and over

so the exe never works

and despite some heavy googling the past hour

I don't feel I am any closer to grasping what is going on exactly

the module in question is checksumdir

this is how I am making the exe currently, although I have tried quite a few different variations

such as one file

I'm not really sure what else info wise to add

to help out

I just haven't had pyinstaller give me a fuss over packages not being there that are before

[meager zinc]

This is definitely the wrong channel - see #❓｜how-to-get-help

[glad belfry]

Thanks 👍

[limpid spruce]

this is a good way to figure out if a number is even or odd

[past plank]

yeah especially because it stops working for numbers larger than 15

[versed eagle]

to check if a number is even or odd you only need to check the least significant bit

[fleet bridge]

Too boring

!e print(len(str(__import__('math').pi)))

[night quarry]

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

17

[limpid spruce]

!e bin(__import__('math').__getattribute__('pi').__getattribute__('__str__')().__getattribute__('__len__')()).__getattribute__("__len__")()+__import__("math").__getattribute__("e").__getattribute__('__str__')().__getattribute__('__len__')().__getattribute__('__and__')(1)-7

[night quarry]

@limpid spruce :warning: Your 3.11 eval job has completed with return code 0.

[No output]

[limpid spruce]

xd

!e print(bin(__import__('math').__getattribute__('pi').__getattribute__('__str__')().__getattribute__('__len__')()).__getattribute__("__len__")()+__import__("math").__getattribute__("e").__getattribute__('__str__')().__getattribute__('__len__')().__getattribute__('__and__')(1)-7)

[night quarry]

@limpid spruce :white_check_mark: Your 3.11 eval job has completed with return code 0.

1

[past plank]

just using __getattribute__ instead of directly dot-accessing a member isn't really esoteric

just longer

same for .__str__() and .__len__() instead of just str() and len()

[versed eagle]

if we're not using num%2 or num&1, i like not not int(bin(num)[~0])

in most cases it won't actually matter whether you have an int or a bool, so you can most likely drop the not not if you want to

[past plank]

n//2==n/2

[fleet bridge]

I believe struct.Struct.pack() requires bools for bool fields

[past plank]

ez boolean

[fleet bridge]

Won't work for huge numbers

[versed eagle]

thats why i say "in most cases"

because there are certainly cases where a bool specifically is needed

[past plank]

since when has that been a concern for esoteric code

[versed eagle]

but most times an int will work just fine

[past plank]

also it works for any integer less than 18014398509481985 which is pretty "huge" anyway

[fleet bridge]

!e

def is_even(n): return n/2==n//2

assert is_even(0)
assert not is_even(1)
assert is_even(2)
assert not is_even(3**1000)
assert is_even(3**1000+1)

[night quarry]

@fleet bridge :x: Your 3.11 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 6, in <module>
003 |   File "<string>", line 1, in is_even
004 | OverflowError: integer division result too large for a float

[fleet bridge]

Oh

3**1000 is overkill

[past plank]

yeah

!e ```py
def is_even(n): return n/2==n//2

max = 18014398509481984

print(is_even(max))
print(is_even(max-1))
print(is_even(max-2))
print(is_even(max-3))

print(is_even(max+1)) # no work
print(is_even(max+2)) # no work
print(is_even(max+3)) # no work```

[night quarry]

@past plank :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | True
002 | False
003 | True
004 | True
005 | True
006 | False
007 | False

[past plank]

hm why -3 doesn't work

[fleet bridge]

weird

[past plank]

ah I messed up my calculations on the max

[fleet bridge]

Maybe your max is not the actual limit

[past plank]

!e ```py
def is_even(n): return n/2==n//2

max = 18014398509481984 // 2

print(is_even(max))
print(is_even(max-1))
print(is_even(max-2))
print(is_even(max-3))

print(is_even(max+1)) # no work
print(is_even(max+2)) # no work
print(is_even(max+3)) # no work```

[night quarry]

@past plank :white_check_mark: Your 3.11 eval job has completed with return code 0.

001 | True
002 | False
003 | True
004 | False
005 | True
006 | True
007 | False

[past plank]

there

I doubled it accidentally

regardless 9 quadrillion is pretty large

will work in most applications lol

[fleet bridge]

It is in range of int64

[past plank]

this gets you the same rangepy is_even = range(-9007199254740992, 9007199254740992, 2).__contains__

[past plank]

not quite

it's 53 bits of double precision

[fleet bridge]

53<64, isn't it?

[past plank]

because x / 2 needs to be able to represent that number at a 0.5 resolution

and python uses IEEE754 doubles

you get 0.5 resolution at 54 bits, but x/2 so x needs to be <=53 bits

[fleet bridge]

I mean it is technically possible to have large int64, use this algorithm and have problems with it

[past plank]

yeah

any number larger than 2**53 but less than 2**64 is going to have issues

and so will numbers larger than 2^64 obviously

[fleet bridge]

This is approximately 99.95% of all int64 numbers 😄

[past plank]

yeah

and approximately 100% of all numbers

but regardless, the majority of computations occur with numbers less than 9 quadrillion

you can always just do py is_even = range(-ARBITRARILY_LARGE_INTEGER*2, ARBITRARILY_LARGE_INTEGER*2, 2).__contains__

[fleet bridge]

is_even=range(-int(float('inf'))*2,int(float('inf'))*2).__contains__

[past plank]

int(float('inf')) lol

[sick hound]

@lambda a:(u:=(i:=a.__code__).co_code,setattr(a,"__code__",i.replace(co_code=b"".join({u[e+e]}<{*b"et"}!=u[e-~e]==i.co_names.index(a.__name__)and b"q\0"or u[e+e:2+e+e]for e in range(len(u)//2)))),a)[2]

handy decorator for "tail" call "optimisation"

[sly ibex]

question
i have this
coffe1 = 20
coffe2 = 50
coffe3 = 20
there's any way to say
while coffe1 != 0 or coffe2 != 0 or coffe3 !=0? this part

[last locust]

If you're trying to shorten it then maybe while (0, 0, 0) != (coffe1, coffe2, coffe3)?

[versed eagle]

while any((coffe1, coffe2, coffe3)) is shorter

wait actually

while coffe1|coffe2|coffe3:

is even shorter

[finite blaze]

could u explain how this works?

[versed eagle]

it's bitwise or

each output bit will be set to 1 if either of the input bits are 1

if both of the input bits are 0, then the output bit is 0

if all of the bits in each of the inputs are 0, then the bitwise or will return 0

0 == False

if all the numbers are 0, then it will be False, and the while loop will end

[finite blaze]

Thanks 👍

[pure dew]

@fleet bridge @past plank
is_even = lambda x: x % 2 == 0

[quartz wave]

f=lambda x:x%2<1 done

[versed eagle]

what about is_even = lambda x: x & 1 ^ 1

[sick hound]

you can do this??

[versed eagle]

please dont ping me for no reason

[sick hound]

why mad over a little red dot

[versed eagle]

i literally just asked you to not ping me for no reason.

please stop.

[low lynx]

is_even=lambda s:~s&1

[meager zinc]

def is_even(number):
  if number == 0:
    return "idk"
  if number == 1:
    return False
  if number == 2:
    return True
  if number == 3:
    return False
  if number == 4:
    return True
  raise OverflowError

[sick hound]

best

[tough willow]

!e ```py
def is_even(number):
if number == 0:
return "idk"
if number == 1:
return False
if number == 2:
return True
if number == 3:
return False
if number == 4:
return True
return is_even(number-4)

print(is_even(123))

[night quarry]

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

False

[low lynx]

why are you returning idk for 0

[sick hound]

thats smart

[low lynx]

0 is definitely even

[tough willow]

Yes

[violet pollen]

You guys really got me curious about esoteric programming in Python--are there any good resources I can use to learn about it?

[low lynx]

i kinda just started golfing stuff and then learning stuff by looking at other people's golfs

if you have an interesting problem to golf send it here and you can see how other people golf

there's a bunch of golfs online as well

[versed eagle]

IFcoltransG made a guide for it. It's not finished, and I think it's been shelved, but it's out there if you want to read what exists so far.

https://ifcoltransg.github.io/esoteric-python-guide/

[low lynx]

code golfing aoc was like golf boot camp so if you do that next year can recommend

[versed eagle]

other than that, there aren't any resources that i know of for "learning esoteric python"

mainly it's just exploring the boundaries of the language on your own and seeing what other people do with it

[violet pollen]

ahh, i see

thank you :)

[past plank]

Are you aware of what "esoteric" means

[versed eagle]

that's a bit rude

and dapper is fairly active in this channel
it's safe to say that they know what esoteric means
as such, questioning them in a relatively antagonistic way like that is rude
please dont do that <3

[arctic skiff]

@versed eagle

[arctic skiff]

The correct use of recursion

[lean warren]

can anyone know, why there is a need of lambda function in python

[sick hound]

you can sort dicts in a list by a specific key in those dicts

key=lambda x: x["timestamp”] for example

[lean warren]

oh, interesting!!

[sick hound]

theres many other uses but thats one im certain on

[lean warren]

got it, thanks for letting me know...

[sick hound]

[lean warren]

awesome😂

[sick hound]

im sure someone here can educate you more on lambda, im not an expert or anything

[restive void]

There's no strict need for lambdas, you could always just use a named function. It just makes some simple throwaway functions (e.g. for use as key functions as shown above) easier to write.

[versed eagle]

please don't
i've asked twice already for people to stop doing that

now im having to ask a third time.

:/

[arctic skiff]

Ok i won't

[versed eagle]

You've done it again.

Please stop.

Please.

[arctic skiff]

Ok lol

[calm loom]

hm

[pure dew]

oh, forgot which channel I was in when I sent this

[tough willow]

Hi @versed eagle

[arctic skiff]

You should definitely turn off pings

[versed eagle]

I have asked 4 times to not be pinged. Every single time, my request has been subsequently ignored. can you please just stop. this harassment doesn't gain you anything. there's no point. please stop

[arctic skiff]

Is there any reason you don't put yourself on dnd or turn of this server pings

[tough willow]

Sorry i will Stop

[versed eagle]

then i would miss pings that have an actual purpose, and neither of those would fix the actual issue here.
people are refusing to stop, after multiple instances of me asking them to stop. instead, they are going out of their way to harass me and disrespect my requests for them to stop.

[versed eagle]

how many times am i going to have to ask before you actually stop

:/

[tough willow]

I stopped

[versed eagle]

no, you haven't. you're continuing to ping me.

[tough willow]

Im not

[arctic skiff]

@zinc aurora is here

[rugged owl]

The reply automatically pings them, think thats what they are meaning

[tough willow]

okay sorry @versed eagle i didnt know that

i will stop

[arctic skiff]

you can click this so it won't ping on reply

[versed eagle]

i know. im trying to refrain from involving others, because that shouldn't be necessary
people should be mature enough to realise that what they're doing is a problem, and stop on their own

[tough willow]

okay

[tough willow]

i turned it off

[versed eagle]

you just pinged me again.

[arctic skiff]

you have to turn it off on every reply it don't toggle for every reply

[arctic skiff]

lol

[tough willow]

okay

[rugged owl]

[violet pollen]

🧍

[tough willow]

hi

[finite blaze]

Hey, I have a big txt file with names like this: "TOM","TIMMY","MAX".....
what would be the shortest way to store these names in a list?

[versed eagle]

the names are separated by commas, and are all inside strings?

[versed eagle]

eval(f"[{open("yourfile").read()}]")

[finite blaze]

hmm interesting

ah i understand

smart smart, thanks

but it doesn't work ;p

[versed eagle]

ah

switch the quotes

i accidentally used the wrong type

of quotes

[finite blaze]

yea 👍

[versed eagle]

eval(f"[{open('yourfile').read()}]")

[arctic skiff]

.close left the chat

[versed eagle]

the file gets closed in its destructor

which gets called bc its ob_refcnt will be 0

[finite blaze]

hmm i have this little challange

text file containing over five-thousand first names, begin by sorting it into alphabetical order. Then working out the alphabetical value for each name, multiply this value by its alphabetical position in the list to obtain a name score.

For example, when the list is sorted into alphabetical order, COLIN, which is worth 3 + 15 + 12 + 9 + 14 = 53, is the 938th name in the list. So, COLIN would obtain a score of 938 × 53 = 49714.

What is the total of all the name scores in the file?

names = sorted(eval(f"[{open('022.txt').read()}]"))
print(sum([sum(ord(letter) - 64 for letter in name) * names.index(name) for name in names]))

thats what i am doing right now

870873746 is my result but page won't accept it

any ideas what i might be missing?

[serene stratus]

#❓｜how-to-get-help

[rugged owl]

He is trying to do it via esoteric practices

they are*

[serene stratus]

not so sure since his solution doesn't look esoteric and he is asking what's wrong with his code

[violet pollen]

i think they are trying to make a one liner? idk

[low lynx]

I don't see why you don't just enumerate instead of indexing

[pure dew]

alphabetical value, not character value

[versed eagle]

off by one error

[frail forum]

names = ['AVRAM','JACK','BOB']
print(sum([sum([ord(letter)-64 for letter in name])*(names.index(name)+1)for name in names]))``` I dont have you're file of names so I just used 3 for testing purposes

python indexes start at 0 instead of 1

it was multiplying the score of the first name by 0 causing the score itself to be 0 which was probably your problem

upon 2 more minutes of research i have gotten the names file

the correct score should be 871198282 👍🏾

[quartz wave]

does it look like 'A', 'B', 'C', ...?

[frail forum]

yes, it is comma separated

[versed eagle]

it's this

[quartz wave]

we don't know that

[versed eagle]

cereal

[quartz wave]

ok

[frail forum]

yes, yes you do

[quartz wave]

about to golf this

[versed eagle]

so far what we have is

N=sorted(eval(f"[{open('f').read()}]"))
print(sum(sum(ord(l)-64for l in n)*-~N.index(n)for n in N))

[frail forum]

let him cook...

[frail forum]

did this get longer?

[versed eagle]

yeah i edited it
since i accidentally pasted an older version that didn't account for the off by 1 error

[frail forum]

ah

[quartz wave]

ok so we'll read from stdin now

so open(0)

90 ```py
N=sorted(eval(open(0).read()))
print(sum(sum(ord(l)-64for l in n)*-~N.index(n)for n in N))

[versed eagle]

that would require changing the input

[quartz wave]

why?

[frail forum]

because

[versed eagle]

wait no nvm

[frail forum]

are u sure?

[quartz wave]

ok we're good then

[versed eagle]

yeah

[quartz wave]

wait a min

[frail forum]

whats the input for this?

because it has to turn

comma separated into a table

eval(f"[{open('f').read()}]")

[versed eagle]

it will default to a tuple

[frail forum]

kk

[versed eagle]

if there aren't enclosing braces/brackets

[quartz wave]

89 assuming all unique ```py
N=sorted(eval(open(t:=0).read()))
print(sum(sum(ord(l)-64for l in n)*(t:=t+1)for n in N))

[versed eagle]

ooh thats a good one

[frail forum]

will in that case

print(sum(sum(ord(l)-64for l in n)*(t:=t+1)for n in sorted(eval(open(t:=0).read()))))```

85

[quartz wave]

no walrus in comprehension iterable

[frail forum]

sec

[frail forum]

good catch

[quartz wave]

although 86 ```py
t=0;print(sum(len(n)(t:=t+1)+tsum(map(ord,n))for n in sorted(eval(open(0).read()))))

nvm

wait

[frail forum]

l is not defined

[frail forum]

i dont this this works

[quartz wave]

ye it doesn't work

[frail forum]

6010368747 is the value i got out of that

[quartz wave]

86 again ```py
t=0;print(sum(sum(ord(l)-64for l in n)*(t:=t+1)for n in sorted(eval(open(0).read()))))

[frail forum]

trying to think of ways to golf

[versed eagle]

found a shortening, i think

instead of eval(open(0).read()), eval(*open(0))
if we assume the input contains no newlines

which i think is a fair assumption

[frail forum]

there are no newlines

you are correct

this is a problem from project euler#022

[night quarry]

Hey @frail forum!

You either uploaded a .txt file or entered a message that was too long. Please use our paste bin instead.

[frail forum]

https://paste.pythondiscord.com/maremomezu

this is the names file

[quartz wave]

that's actually a good one

that reduces it to a good 80 ```py
t=0;print(sum(sum(ord(l)-64for l in n)*(t:=t+1)for n in sorted(eval(*open(0)))))

but can eval() handle a 46447 length file

[frail forum]

yea

[quartz wave]

it can't for me

[frail forum]

hmmm

[quartz wave]

nvm it can

the warnings were causing the problem

[night quarry]

Hey @rough moat!

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

[rough moat]

lol

its a megabyte

[night quarry]

Hey @rough moat!

You either uploaded a .txt file or entered a message that was too long. Please use our paste bin instead.

[rough moat]

bro

i dont think it even lets me paste lmao

i'll just put it on github

https://github.com/Sea-Pickle/gloop_scripts/blob/main/milan_onefile.py

yeah hastebin doesn't even work

[last locust]

It works fine for me: https://paste.pythondiscord.com/voqepavigi

!paste has instructions on how to use it

[night quarry]

Pasting large amounts of code

If your code is too long to fit in a codeblock in Discord, you can paste your code here:
https://paste.pythondiscord.com/

After pasting your code, save it by clicking the floppy disk icon in the top right, or by typing ctrl + S. After doing that, the URL should change. Copy the URL and post it here so others can see it.

[rough moat]

it didnt work for me

idk

[finite blaze]

ah thanks, stupid mistake, knew it ;p

[night quarry]

:incoming_envelope: :ok_hand: applied mute to @nimble crescent until <t:1675958919:f> (10 minutes) (reason: newlines rule: sent 106 newlines in 10s).

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

[river idol]

!unmute @nimble crescent

[night quarry]

:incoming_envelope: :ok_hand: pardoned infraction mute for @nimble crescent.

[river idol]

Sorry about that. Use this for longer blocks of code: https://paste.pythondiscord.com/

[nimble crescent]

ok, thanks

[versed eagle]

i think this is the wrong channel for what you want

[nimble crescent]

Im sorry, what would be a better channel?

[versed eagle]

my guess would be #networks
or you could do one of those new thread things they have

[rugged sparrow]

I have a challenge for yall. try to find the most creative way of getting python to raise a SystemError (raise/throw directly are not allowed lol)

[fleet bridge]

!e open(4)

[night quarry]

@fleet bridge :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 | OSError: [Errno 9] Bad file descriptor

[fleet bridge]

When SystemErrors can occur?

[rugged sparrow]

!e ```py
exec('''
corrupt = iter(int, 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 should create a system error

weird that it isn't