esoteric-python

[night quarry]

@twilit grotto :white_check_mark: Your eval job has completed with return code 0.

1

[simple crystal]

!e

from math import factorial as f
print(f(0))

[night quarry]

@simple crystal :white_check_mark: Your eval job has completed with return code 0.

1

[simple crystal]

!e

from math import factorial as f
print(f(False))

[night quarry]

@simple crystal :white_check_mark: Your eval job has completed with return code 0.

1

[toxic jewel]

!e print(__import__("math").factorial(__import__("__main__").__doc__.__bool__()))

[night quarry]

@toxic jewel :white_check_mark: Your eval job has completed with return code 0.

1

[sick hound]

how tell us your secrets

[fluid tree]

Better to use a metaclass

[alpine flower]

so ive been told

[toxic jewel]

python reads those characters as abc

!e True = False; print(True)

[night quarry]

@toxic jewel :x: Your eval job has completed with return code 1.

001 |   File "<string>", line 1
002 |     True = False; print(True)
003 |     ^
004 | SyntaxError: cannot assign to True

[toxic jewel]

;^(

[snow beacon]

Only on Python 2.

[wary falcon]

ya'll in here pushing the oneliner limits i see lol

!e def 𝒉𝖊𝖑𝙡o(𝕟𝗮𝗆𝑒):
𝓅𝙧𝒾𝓷𝗍(f"Hello {𝐧𝕒𝗆𝐞}!")

𝐡𝒆lℓₒ('World')

[night quarry]

@wary falcon :white_check_mark: Your eval job has completed with return code 0.

Hello World!

[wary falcon]

whaaaa

[autumn saffron]

!e

from ctypes import c_void_p
from random import randint

class foo:
    def __repr__(self):
        return "True" if randint(0, 10) <= 5 else "False"
class bar:
    def __repr__(self):
        return "False" if randint(0, 10) <= 5 else "True"

c_void_p.from_address(id(True)+8).value = id(foo)
c_void_p.from_address(id(False)+8).value = id(bar)
print(True)
print(False)

[night quarry]

@autumn saffron :white_check_mark: Your eval job has completed with return code 0.

001 | True
002 | False

[autumn saffron]

hm

!e

from ctypes import c_void_p
from random import randint

class foo:
    def __repr__(self):
        return "True" if randint(0, 10) <= 5 else "False"
class bar:
    def __repr__(self):
        return "False" if randint(0, 10) <= 5 else "True"

c_void_p.from_address(id(True)+8).value = id(foo)
c_void_p.from_address(id(False)+8).value = id(bar)
print(True)
print(False)

[night quarry]

@autumn saffron :white_check_mark: Your eval job has completed with return code 0.

001 | True
002 | True

[autumn saffron]

!e

from ctypes import c_void_p
from random import randint

class foo:
    def __repr__(self):
        return "True" if randint(0, 10) <= 5 else "False"
class bar:
    def __repr__(self):
        return "False" if randint(0, 10) <= 5 else "True"

c_void_p.from_address(id(True)+8).value = id(foo)
c_void_p.from_address(id(False)+8).value = id(bar)
print(True)
print(False)

[night quarry]

@autumn saffron :white_check_mark: Your eval job has completed with return code 0.

001 | True
002 | True

[autumn saffron]

gottem

sadly it needs to return a string for printing purposes

[valid crow]

!e

[night quarry]

Command Help

!eval [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.

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

[valid crow]

!e print("hello world")

[night quarry]

@valid crow :white_check_mark: Your eval job has completed with return code 0.

hello world

[valid crow]

!e
n1 = int(input("enter your first number"))
n2 = int(input("enter your second number"))
print(n1+n2)

[night quarry]

@valid crow :x: Your eval job has completed with return code 1.

001 | enter your first numberTraceback (most recent call last):
002 |   File "<string>", line 1, in <module>
003 | EOFError: EOF when reading a line

[valid crow]

n1 = int(input("enter your first number"))
n2 = int(input("enter your second number"))
print(n1+n2)

n1 = int(input("enter your first number"))
n1 = 5
n2 = int(input("enter your second number"))
n2 = 5
print(n1+n2)

!e n1 = int(input("enter your first number"))
n1 = 5
n2 = int(input("enter your second number"))
n2 = 5
print(n1+n2)

[night quarry]

@valid crow :x: Your eval job has completed with return code 1.

001 | enter your first numberTraceback (most recent call last):
002 |   File "<string>", line 1, in <module>
003 | EOFError: EOF when reading a line

[valid crow]

sadness

[snow beacon]

Feel free to use #bot-commands for more testing.

[strange basin]

All identifiers are converted into the normal form NFKC while parsing; comparison of identifiers is based on NFKC.

https://docs.python.org/3/reference/lexical_analysis.html#identifiers

[verbal totem]

!e

z = lambda: (print("hello") or True) and print("world")
z() 

[night quarry]

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

001 | hello
002 | world

[verbal totem]

Does this count as cheating?

[granite cloak]

No.

[snow beacon]

What's your goal?

[floral meteor]

to make print have a 1 in 3 chance of Segmentation Fault

my first attempt made it 1 in 3 chance a recursion would occur each recursion

this one isn't properly crashing

!e ```py
def print(*args,sep=chr(32),end=chr(10),file=import('sys').stdout):
try:
import random
file.write(str(sep).join([str(e)for e in args])+str(end))
1/random.randint(0,2)
except:
def print(*a,**k):
try:print(*a,**k)
except:print(*a,**k)
print (*args,sep=sep,end=end,file=file)
print ("Hello", "World", end="!\n")
print ("Hello", "World", end="!\n")
print ("Hello", "World", end="!\n")
print ("Hello", "World", end="!\n")

[night quarry]

@floral meteor :x: Your eval job has completed with return code 139 (SIGSEGV).

001 | Hello World!
002 | Hello World!
003 | Hello World!
004 | Fatal Python error: _Py_CheckRecursiveCall: Cannot recover from stack overflow.
005 | Python runtime state: initialized
006 | 
007 | Current thread 0x00007fe62b4d5740 (most recent call first):
008 |   File "<string>", line 8 in print
009 |   File "<string>", line 8 in print
010 |   File "<string>", line 8 in print
011 |   File "<string>", line 8 in print
... (truncated - too many lines)

Full output: https://paste.pythondiscord.com/ifiwajifut.txt

[floral meteor]

there

the first two worked, but the second one worked, then made python die

[earnest wing]

!e ```py
import ctypes, random
def unstable(f):
def inner(*args, **kwargs):
if random.randint(0, 2) == 0:
ctypes.string_at(0)
f(*args, **kwargs)
return inner

print = unstable(print)

for x in range(10):
print(x)

[night quarry]

@earnest wing :x: Your eval job has completed with return code 139 (SIGSEGV).

001 | 0
002 | 1
003 | 2
004 | 3

[steep mural]

why doesn't that have the faulthandler thing

[floral meteor]

what faulthandler thing?

hey i have an idea for... effect

[earnest wing]

because it's a memory access violation

the interpreter can't really catch it

[steep mural]

isn't that how faulthandler works in the first place

[floral meteor]

!e ```py
def print(*args,sep=chr(32),end=chr(10),file=import('sys').stdout):
try:
import random
file.write(str(sep).join([str(e)for e in args])+str(end))
1/random.randint(0,3)
except:
def print(*a,**k):
try:builtins.print(*a,**k)or print(*a,**k)
except:print(*a,**k)
print (*args,sep=sep,end=end,file=file)

for i in range(10): print (i)

[night quarry]

@floral meteor :x: Your eval job has completed with return code 139 (SIGSEGV).

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

Full output: https://paste.pythondiscord.com/begiwokivi.txt

[snow beacon]

The interpreter would have to check every memory access. It would be even slower.

[earnest wing]

might need to enable it with a commandline flag

[steep mural]

faulthandler catches the SIGSEGV signal

the interpreter doesn't have to check

[earnest wing]

yeah, python3 -X faulthandler

[snow beacon]

I'm unfamiliar with the faulthandler. What's its use case?

[earnest wing]

https://docs.python.org/3/library/faulthandler.html

just took a quick read

[snow beacon]

Cool beans.

[earnest wing]

odd that it doesn't work for some faults

you'd expect stack overflows and bad reads both trigger it

[steep mural]

funny it actually has an example using ctypes.string_at(0)

[earnest wing]

implementation specific behavior :mystery:

[steep mural]

turns out you don't need faulthandler to get the stack overflow message makes sense

[earnest wing]

so each call just gets checked on the c side for overflow?

[steep mural]

i think the python vm doesn't check if the python stack is getting too large after every FAST instruction (and some others?). i'm also pretty sure that you can ignore a recursion error (and keep recusing) 50 times until the python vm says there was a fatal error and dies. Also seems like an actual stack overflow will cause a MemoryError and not a RecursionError

maybe not 50 times... i was just reading the source code but apparently I was wrong. It only worked 1 time before dying.

wait i'm wrong again, it only hits the except block 1 time but it does allow you to keep recursing 50 times

[viscid nymph]

I thought you can recurse 1000 times until a recursion error

[earnest wing]

You can catch the error, or prevent it altogether with sys.setrecursionlimit

[floral meteor]

!e ```py
import('sys').setrecursionlimit(9999999)
def print(*args):
try: print (*args)
except: print (*args)

print ("Hello World!")

[night quarry]

@floral meteor :warning: Your eval job has completed with return code 139 (SIGSEGV).

[No output]

[floral meteor]

huh, much better than the last one

[final night]

Is it possible to use list comprehension to flatten a list?

Here's my looping code:

def flatten(list_to_flatten):
    flat_list = []
    for item in list_to_flatten:
        if type(item) is list:
            flat_list += flatten(item)
        else:
            flat_list.append(item)
    return flat_list

Tried to convert it into this but it doesn't run:

def fl(list_to_flatten):
    return [i for i in fl(item) if type(item) is list else item for item in list_to_flatten]

so then I tried this:

def fl(list_to_flatten):
    return [i for i in [fl(item) if type(item) is list else item for item in list_to_flatten]]

but it turns out that's just a glorified way to return the exact item passed in

fl([1,2,3[4,5,6]]) = [1,2,3[4,5,6]]

the tricky part is getting fl(item) and item to work in the same comprehension, somehow

If I can manage to write it using list comprehension then that opens the door for flattening lambda functions which would be super cool

imagine using max with a flattening lambda function as a key, for example

[snow beacon]

[x for item in input_list for x in item if isinstance(item, list) else [item]] might work.

[final night]

  File "<stdin>", line 1
    t = [x for item in input_list for x in item if isinstance(item, list) else [item]]
                                                                          ^
SyntaxError: invalid syntax

[earnest wing]

swap the condition and the for

[snow beacon]

Or bracket the item if ... [item] part.

[final night]

>>> input_list = [1,2,3,[4,5,6],7]
>>> [x for item in input_list for x in [item if isinstance(item, list) else [item]]]
[[1], [2], [3], [4, 5, 6], [7]]

same as my attempt, just replicates the list 1:1

[snow beacon]

Regular brackets, not square ones.

[final night]

parenthesis?

[snow beacon]

I stand by what I said. () these things.

[final night]

Alright, i'll try it out :)

eyy yeah it totally works

nice one

so, out of curiosity

why do the parens make it work?

(and here is the code for anybody curious):
[x for item in input_list for x in (item if isinstance(item, list) else [item])]

also, looks like it's not fully recursive, only works on singly nested lists

although it wouldn't be hard to make it recursive, I think

[snow beacon]

Without them it parses as [expr for var in iterable if condition], but then it sees the else keyword and gets confused.

[final night]

here's a version that works on arbitrarily nested lists:

def fl(input_list):
    return [x for item in input_list for x in (fl(item) if isinstance(item, list) else [item])]

i'll post a lambda soon

[snow beacon]

fl = lambda input_list: [x for item in input_list for x in (fl(item) if isinstance(item, list) else [item])]

[final night]

yep, that works

wow, that's super cool

[floral meteor]

Alright I made a simple death module

You just import death and it will kill python so hard it will die to death

It prints out...

Fatal Python error: _py_MorsObitusDecessus: Python has died. 
Python runtime state: deceased

To sys.stderr then sets recursion limit to 99999 then tortures python to death with pseudo infinite recursion.

[grave rover]

Nice

Meanwhile I made my pipeline/chain thing work with async stuff

[grave rover]

Do y'all ever just feel like doing something cursed but having zero ideas

[earnest wing]

Idea: implement a DSL using decorator and matmul syntax

[snow beacon]

Hello World using no letters or numbers?

Actually that's probably impossible.

[earnest wing]

You'd need at least a few letters

Unless you assume all files use the unicode-escape codec

even then you'll need x

[hard spoke]

How'd you write code with no letters? 🙂

if you mean "without string/bytes literals", then it makes more sense, and might even be possible

actually, even easy if this is allowed:

!e

class hello:pass
class world:pass
print(hello.__name__.title(),world.__name__.title())

[night quarry]

@hard spoke :white_check_mark: Your eval job has completed with return code 0.

Hello World

[thin trout]

Ha

[snow beacon]

I think that is part of the challenge. I also think the challenge is impossible.

[snow beacon]

Can you not use octal \OOO literals in a unicode-escape codec?

[hard spoke]

fair enough. Like, I don't think there's any valid code which compiles to at least one instruction you can write without any letters or numbers

[snow beacon]

_=[]?

[hard spoke]

of course #esoteric-python people would solve this in an instant 😩

[snow beacon]

Anything juicy from there probably requires attribute access.

[hard spoke]

printing anything is still a problem, though. You could get print from builtins, but that'd requires getting the builtins at the very least...

[snow beacon]

You can construct arbitrary numbers pretty easily.

You wouldn't have easy access to any functions though.

[thin trout]

Bytecode manipulation when

[hard spoke]

Hmm, how? I get how to construct arbitrary numbers after getting some int, but how'd you get that first int?

oh, I got it

[snow beacon]

+([]==[]) is 1, although you can also use []==[].

[hard spoke]

yeah, that's what I thought, bool being a subclass of int finally pays off

[snow beacon]

That fact is also good for array indexing with code golf.

[hard spoke]

def generate_special_number(x:int) -> str:
    if x<0:
        return "-"+generate_special_number(-x)
    s = []
    one = "([]==[])"
    two = f"({one}+{one})"
    if x==0:
        return f"{one}-{one}"
    if x==1:
        return one
    if x==2:
        return two
    div,mod = divmod(x,2)
    s = f"({generate_special_number(div)})*{two}"
    if mod:
        s+= "+"+generate_special_number(mod)
    return s

probably not the most efficient way, but it works

((((((([]==[]))*(([]==[])+([]==[]))+([]==[]))*(([]==[])+([]==[])))*(([]==[])+([]==[])))*(([]==[])+([]==[])))*(([]==[])+([]==[])))*(([]==[])+([]==[]))+([]==[])

is 97, for example

[thin trout]

Noice

How do you execute numbers now

[hard spoke]

huh?

ah, how to get functions from this? no idea, lol

[thin trout]

Well, it is nice to have numbers, now we need to have code!

[snow beacon]

Someone made a generator with True and << at some point, which should be easily modifiable should a smaller number encoding prove useful.

[hard spoke]

edited above message: fixed bug which included zeros in the representation

[fossil estuary]

does this count as esoteric :p

Let me get it

[earnest wing]

What would be an optimal number generator under these rules (i.e. returns the shortest string that evaluates to a number N that doesn't use alphanumeric chars)?

[fossil estuary]

!e ```py
globals()["builtins"].print(globals()["builtins"].int(globals()["builtins"].str(globals()["builtins"].round(111.222333, 3)).replace(".","")))

[night quarry]

@fossil estuary :white_check_mark: Your eval job has completed with return code 0.

111222

[hard spoke]

Good question, this is exactly what I'm doing 🙂

well, I'm doing dumb bruteforce search of these combinations

[earnest wing]

** and << are definitely useful for large numbers

[snow beacon]

~ and - useful for small ones.

[earnest wing]

but using temporary variables also shortens programs

_ of course

[fossil estuary]

yeah

[snow beacon]

""<"" seems to be a short way to get 0.

[fossil estuary]

i have some code that errors and runs in the same string lol, idk how it works tbh

[earnest wing]

If you could somehow get bytes objects...

[*bytes] is an easy way to get certain bytes in the printable ascii range

[snow beacon]

Or simply indexing into the byte object.

[hard spoke]

huh, where are floats coming from?..

I somehow got floats from ints using only * + << ** - ~

[earnest wing]

a ** -b

[hard spoke]

ah, I see

I think I'll just drop these

   0 ([]==[])-([]==[])
   1 []==[]
  -1 ~(([]==[])-([]==[]))
   2 ([]==[])+([]==[])
  -2 ~([]==[])
   3 ([]==[])-(~([]==[]))
  -3 ~(([]==[])+([]==[]))
   4 (~([]==[]))*(~([]==[]))
  -4 (~([]==[]))+(~([]==[]))
  -5 ~((~([]==[]))*(~([]==[])))
   5 ([]==[])+((~([]==[]))*(~([]==[])))
   6 (~([]==[]))*(~(([]==[])+([]==[])))
  -6 (([]==[])-(~([]==[])))*(~([]==[]))
   7 ~((~([]==[]))**(([]==[])-(~([]==[]))))
  -7 ~((~([]==[]))*(~(([]==[])+([]==[]))))
   8 (~([]==[]))*((~([]==[]))+(~([]==[])))
  -8 (~([]==[]))**(([]==[])-(~([]==[])))
   9 (~(([]==[])+([]==[])))**(([]==[])+([]==[]))
  -9 ~((~([]==[]))*((~([]==[]))+(~([]==[]))))
 -10 ~((~(([]==[])+([]==[])))**(([]==[])+([]==[])))
  10 (~([]==[]))*(~((~([]==[]))*(~([]==[]))))

[earnest wing]

Though it would be pretty funny to make a function that can yield every possible floating point value

[hard spoke]

here are some supposedly shortest representations

[earnest wing]

0: +(''<'')?

[hard spoke]

I'm not doing strings yet, but hmm

[earnest wing]

or without the +()

if bools are okay

otherwise []!=[]

[hard spoke]

I think it'll only provide a benefit for 0, so no big reason to include it here

[earnest wing]

I mean it contributes to the other numbers as well

maybe?

no

[hard spoke]

the complexity of my algorithm is something like 6**<number of iterations> 😅

the above is only 3 iterations because of that

I could probably use dynamic programming instead if I could inverse all the operations I'm using

like, quickly find, for all the functions I'm using and all a and c, the b such that f(a,b)=c

https://paste.pythondiscord.com/rosejujeva.php
My take on generating shortest representations

Example: output of

results = generate_shortest_reprs(10,-100,100,verbose=False)
for val, rep in results.items():
    print(f"{val:4} {rep}")

is https://paste.pythondiscord.com/rikofojamo.txt

[fossil estuary]

:o

how would you give it that input?

[hard spoke]

wdym?

also, whoops, it's bugged

[fossil estuary]

like, how do you give it the input (like you provided)

[hard spoke]

!e

print(eval("([]==[])-(((~([]==[]))*(~((~([]==[]))*(~([]==[])))))**(([]==[])+([]==[])))"))

[night quarry]

@hard spoke :white_check_mark: Your eval job has completed with return code 0.

-99

[hard spoke]

it's valid Python code

[fossil estuary]

Yeah, I can tell.

[fossil estuary]

how did you give it this as an input, is what im asking

Oh wait, nvm i had just realised what it was

[hard spoke]

weird

for val, rep in results.items():
    res = eval(rep)
    if val!=res:
        print(val,res,rep)

 -39 -161 ([]==[])-(((~(([]==[])+([]==[])))**((~([]==[]))*(~([]==[]))))<<([]==[]))
  40  162 ((~(([]==[])+([]==[])))**((~([]==[]))*(~([]==[]))))<<([]==[])
 -41 -164 (~((~(([]==[])+([]==[])))**((~([]==[]))*(~([]==[])))))<<([]==[])
  41  163 ([]==[])+(((~(([]==[])+([]==[])))**((~([]==[]))*(~([]==[]))))<<([]==[]))
 -42 -164 (~([]==[]))-(((~(([]==[])+([]==[])))**((~([]==[]))*(~([]==[]))))<<([]==[]))

why are these ones bugged?

there we go, fixed it

I mixed up lshift and rshift

Updated outputs:
https://paste.pythondiscord.com/dixuletoso.txt

I added | & ^, but they provided literally zero improvements in length

[fossil estuary]

hm

[fossil estuary]

this code has this output in my side log

Line 4 : Need type annotation for 's'
Line 14 : Incompatible types in assignment (expression has type "str", variable has type "List[Any]")
Line 17 : Incompatible return value type (got "List[Any]", expected "str")
``` strange.

[hard spoke]

very possible, I haven't actually mypyed it yet

currently refactoring

[earnest wing]

yeah that's just from reusing s and not giving it a generic type at first

[fossil estuary]

its just my IDE that returns that stuff

[earnest wing]

no effect on runtime

just lints

mypy doesn't know how to infer generic type parameters for e.g. empty list literals

[fossil estuary]

Huh, i dont use that

[earnest wing]

most static type checkers don't

i don't know of any that do, beyond some rare circumstances

[fossil estuary]

my IDE probably just uses it for some reason

but yeah.

[plucky wedge]

print(eval("([]==[])-(((~([]==[]))*(~((~([]==[]))*(~([]==[])))))**(([]==[])+([]==[])))"))

How does this get numbers? Does it take the []==[] result and do numerical operations on it? Like int(True) * 10

[hard spoke]

https://paste.pythondiscord.com/lorumotovu.py
refactored version. mypy checks pass on it.

[fossil estuary]

[]==[] is equal to 1 @plucky wedge

[plucky wedge]

Yes because of True?

[fossil estuary]

mhm

[hard spoke]

Also, lol, Pycharm's linter (but not mypy) gives me a TypeError here:

(wrongly - this is List[Tuple[str,int]], not List[int])

[fossil estuary]

hey

i copied it from the link and got this

Traceback (most recent call last):
  File "<fstring>", line 1
    (repres=)
           ^
SyntaxError: invalid syntax
``` @hard spoke

[hard spoke]

This means you're using an old Python version where {repres=} wasn't allowed, I'd suggest

[fossil estuary]

i use 3.7

[hard spoke]

I don't remember where this was added, I'm using 3.8

[fossil estuary]

well, thats your answer then?

[hard spoke]

you can fix it by changing such occurences from {a=} to a={a}

[plucky wedge]

What does ~ do?

[hard spoke]

Bitwise negation

[plucky wedge]

Mmmm yes

A bit wise negation

Imma google it

[hard spoke]

Represent the number in binary, negate each bit (0->1, 1->0), and that's your result

[plucky wedge]

Wait so it inverses? Like inverses the binary

[fossil estuary]

Hey, ConfusedReptile dm me quickly :)

[hard spoke]

actually, not sure if that's the correct description, hold on

[plucky wedge]

Don’t worry I’m just gonna google it

I’m not to knowledgeable on this stuffs

[fossil estuary]

@hard spoke

Line 72 : Argument 2 to "apply_funcs" has incompatible type "Dict[str, Callable[[Any, Any], Any]]"; expected "Dict[str, Callable[[int, int], int]]"
Line 72 : "Dict" is invariant
Line 72 : Consider using "Mapping" instead, which is covariant in the value type
Line 72 : Argument 3 to "apply_funcs" has incompatible type "Dict[str, Callable[[Any], Any]]"; expected "Dict[str, Callable[[int], int]]"
``` interesting...

[hard spoke]

is that mypy?

[fossil estuary]

its from my IDE, so i am unsure

[hard spoke]

also, that last error at least is very weird

because Dict[str, Callable[[Any], Any]] fits Dict[str, Callable[[int], int]]

oooh, that's probably exactly why it says to use Mapping, so that it's covariant

[fossil estuary]

yea

[hard spoke]

huh, how do I declare a function with 2 arguments with Mapping?..

[fossil estuary]

i do not know.

[hard spoke]

ah, looks like it might just be that all arguments before the last one are inputs nope

[snow beacon]

Yes, it inverts it in 2's-complement binary, which is equivalent to doing -x-1.

[plucky wedge]

Cool

Bitwise == smort

[snow beacon]

Most of the time it just overcomplicates, but in situations like this we need all the operators we can get.

Also it's good for golfing: -~x is x+1 if you need different precedence/whitespace.

(~-x is x-1.)

[plucky wedge]

Not weird just haven’t seen that before

!e
print(-~1)

[night quarry]

@plucky wedge :white_check_mark: Your eval job has completed with return code 0.

2

[plucky wedge]

!e
print(-~3<<1)

[night quarry]

@plucky wedge :white_check_mark: Your eval job has completed with return code 0.

8

[plucky wedge]

I’m gonna continue this in bot commands lol

[earnest wing]

!e ```py
def number(n: int) -> str:
if n > 0:
return "-~" * n + "0"
else:
return "~-" * n + "0"

for i in range(20):
print(i, number(i), eval(number(i)))

[night quarry]

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

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

Full output: https://paste.pythondiscord.com/mezacupasa.txt

[earnest wing]

Example of using ~- as succ/pred

[fossil estuary]

this has intrigued me

[plucky wedge]

That’s cool

[snow beacon]

Golfing is a good way to learn how to abuse operators for new purposes. For instance, a Boolean not is just x<1.

[plucky wedge]

I’m gonna replace all my division with >> lol

[hard spoke]

this is very nice, hmm

I wonder how I can incorporate that. The problem is determining when I can or can't drop parantheses.

[snow beacon]

There should be a precedence chart somewhere for Python.

Alternatively, the ast module might have a utility for it.

[hard spoke]

There is indeed, but it doesn't make it easy anyway 🙂

hmm, perhaps I can consider two cases - with or without them

[earnest wing]

On the expressions documentation page

[frigid wharf]

https://docs.python.org/3/reference/expressions.html#operator-precedence

[snow beacon]

Too complicated for my liking, even eliminating the ones with words.

[hard spoke]

yikes, I think I managed to autogenerate my way to some code eval hangs on

I don't seem to be able to suspend it with the debugger

[earnest wing]

Smells like big chained pows

x ** (y ** (z ** w))

[hard spoke]

it happened after I added the option to omit ()s

okay, there, managed to calculate for -10..10:

   0 []<[]
   1 []==[]
  -1 ~([]<[])
   2 -~([]==[])
  -2 ~([]==[])
  -3 ~-~([]==[])
   3 -~-~([]==[])
   4 -~-~-~([]==[])
  -4 ~-~-~([]==[])
   5 -~-~-~-~([]==[])
  -5 ~-~-~-~([]==[])
  -6 ~-~-~-~-~([]==[])
   6 -~-~-~-~-~([]==[])
   7 -~-~-~-~-~-~([]==[])
  -7 ~-~-~-~-~-~([]==[])
   8 -~([]==[])<<-~([]==[])
  -8 ~([]==[])<<-~([]==[])
   9 ~-~([]==[])*~-~([]==[])
  -9 ~-~([]==[])*-~-~([]==[])
 -10 ~-~-~-~([]==[])<<([]==[])
  10 ~([]==[])*~-~-~-~([]==[])

whoa, that 10

that's what, -2 * -5? yikes

now the question is

can pypy take this lol

wait, pypy is 3.7... will have to do some backwards-compat.

oh god, why is pypy 3 times slower

[snow beacon]

Maybe the compilation time. For larger scripts it would be amortised.

[hard spoke]

nah, it's likely the evals - they are probably more expensive in pypy, and I do a lot of them in the parantheses processing

and by "3 times", I mean 800ms instead of 300ms

here's -50 to 50:

https://paste.pythondiscord.com/laxiniveha.txt

and here's -100 to 100:
https://paste.pythondiscord.com/odokisutad.txt

wait, where the hell is positives

oh, I messed up

[hard spoke]

Yay, my program managed to autogenerate code that hangs eval!

(~(([]==[])+([]==[])))**-~(([]==[])+([]==[]))**(-~-~([]==[])<<-~(([]==[])+([]==[])))

[earnest wing]

two **s and a <<? better be a huge number

[hard spoke]

it's (-3)**16777217, I believe

right, so how can I fix this?...

perhaps I can try evaluating with numpy ints first

[earnest wing]

A container for the eval call?

Or just using sized ints & failing on overflow I guess

[tribal moon]

Looks like Javascript

[toxic jewel]

woah

it freezes repl.it owo

!e print(~(~([]==[])<<([]==[]<[]==[])>>((~~(([]==[])&~([]==[])))))+([]==[]))

[night quarry]

@toxic jewel :white_check_mark: Your eval job has completed with return code 0.

2

[rose stump]

!e

print

[toxic jewel]

took some inspiration and created this

!e

print(chr(((~(~([]==[])<<([]==[]<[]==[])>>((~~(([]==[])&~([]==[])))))++(([]==[]))<<([]==[])+~(~([]==[])+([]==[]))<<([]==[])++~(~([]==[])<<([]==[])))^~(~([]==[]))++([]==[])++(([]==[])<<([]==[]))+~(~(([]<[])++([]==[])))<<(([]==[])&([]==[])<<([]==[]))<<([]==[])++(()==())^(()==())++(()==())+~((~(()==())++(()==())))%([]==[])<<(([]==[])++([]==[]))++(([]==[])+(([]==[]))))+~(([]==[])<<([]==[]))+~(([]==[])++([]==[]))++([]==[]))+chr(((([]==[])++([]==[]))++(([]==[])++([]==[]))++(([]==[])+([]==[])++~(~(([]==[])++([]==[]))))<<(([]==[])++~(~([]==[])++([]==[])))^([]==[])<<([]==[])++~(~([]==[])+([]==[]))<<(([]==[])++([]==[]))<<(([]==[])++(([]==[])++([]==[])))++([]==[]))++~(([]==[])++(([]==[])<<([]==[])++([]==[])++([]==[])<<([]==[]))++(([]==[])<<([]==[]))<<([]==[]))++(([]==[])<<(([]==[])++([]==[]))++([]==[])<<([]==[]))++~(([]==[])++([]==[]))+([]==[]))+chr(((~(~([]==[])<<([]==[]<[]==[])>>((~~(([]==[])&~([]==[])))))++(([]==[]))<<([]==[])+~(~([]==[])+([]==[]))<<([]==[])++~(~([]==[])<<([]==[])))^~(~([]==[]))++([]==[])++(([]==[])<<([]==[]))+~(~(([]<[])++([]==[])))<<(([]==[])&([]==[])<<([]==[]))<<([]==[])++(()==())^(()==())++(()==())+~((~(()==())++(()==())))%([]==[])<<(([]==[])++([]==[]))++(([]==[])+(([]==[]))))+~(([]==[])<<([]==[]))+~(([]==[])++([]==[]))++([]==[])))

[night quarry]

@toxic jewel :white_check_mark: Your eval job has completed with return code 0.

owo

[hard spoke]

!e

_=chr(((~(~([]==[])<<([]==[]<[]==[])>>((~~(([]==[])&~([]==[])))))++(([]==[]))<<([]==[])+~(~([]==[])+([]==[]))<<([]==[])++~(~([]==[])<<([]==[])))^~(~([]==[]))++([]==[])++(([]==[])<<([]==[]))+~(~(([]<[])++([]==[])))<<(([]==[])&([]==[])<<([]==[]))<<([]==[])++(()==())^(()==())++(()==())+~((~(()==())++(()==())))%([]==[])<<(([]==[])++([]==[]))++(([]==[])+(([]==[]))))+~(([]==[])<<([]==[]))+~(([]==[])++([]==[]))++([]==[]))
print(_+chr(((([]==[])++([]==[]))++(([]==[])++([]==[]))++(([]==[])+([]==[])++~(~(([]==[])++([]==[]))))<<(([]==[])++~(~([]==[])++([]==[])))^([]==[])<<([]==[])++~(~([]==[])+([]==[]))<<(([]==[])++([]==[]))<<(([]==[])++(([]==[])++([]==[])))++([]==[]))++~(([]==[])++(([]==[])<<([]==[])++([]==[])++([]==[])<<([]==[]))++(([]==[])<<([]==[]))<<([]==[]))++(([]==[])<<(([]==[])++([]==[]))++([]==[])<<([]==[]))++~(([]==[])++([]==[]))+([]==[]))+_)

[night quarry]

@hard spoke :white_check_mark: Your eval job has completed with return code 0.

owo

[hard spoke]

yay, optimizations 😅

[lucid shell]

!e
(~(([]==[])+([]==[])))-~(([]==[])+([]==[]))(-~-~([]==[])<<-~(([]==[])+([]==[])))

[night quarry]

@lucid shell :warning: Your eval job timed out or ran out of memory.

[No output]

[lucid shell]

!e
print((~(([]==[])+([]==[])))-~(([]==[])+([]==[]))(-~-~([]==[])<<-~(([]==[])+([]==[]))))

[night quarry]

@lucid shell :warning: Your eval job timed out or ran out of memory.

[No output]

[lucid shell]

!e

a=(~(([]==[])+([]==[])))**-~(([]==[])+([]==[]))**(-~-~([]==[])<<-~(([]==[])+([]==[])))
print(a)

[night quarry]

@lucid shell :warning: Your eval job timed out or ran out of memory.

[No output]

@lucid shell :x: Your eval job has completed with return code 1.

001 | 0 0 0
002 | Traceback (most recent call last):
003 |   File "<string>", line 12, in <module>
004 |   File "<string>", line 1
005 |     -~
006 |       ^
007 | SyntaxError: invalid syntax

[lucid shell]

!e
https://paste.pythondiscord.com/uraditeciq.lua

[night quarry]

@lucid shell :x: Your eval job has completed with return code 1.

001 |   File "<string>", line 1
002 |     https://paste.pythondiscord.com/uraditeciq.lua
003 |           ^
004 | SyntaxError: invalid syntax

[toxic jewel]

#bot-commands

[toxic jewel]

should try fitting & into it

[hard spoke]

I have used &, actually, and didn't get any improvements.

[toxic jewel]

ah

i just tried to think of every possible mathematical operator possible to fit into my equation

except i used no **, //, or --

kinda wanna create a program to create these

[hard spoke]

create what exactly?

arbitrary strings using chr and this number notation? that's pretty easy, just wrap the numbers in chrs

[toxic jewel]

yeah

[formal sandal]

!e
Did you know that Python has tail recursion?

class Factorial:
    def __init__(self, n, acc=1):
        self.n = n
        self.acc = acc

    def __del__(self):
        if self.n <= 1:
            print(self.acc)
        else:
            type(self)(self.n - 1, self.n * self.acc) 

Factorial(2000)

[night quarry]

@formal sandal :white_check_mark: Your eval job has completed with return code 0.

3316275092450633241175393380576324038281117208105780394571935437060380779056008224002732308597325922554023529412258341092580848174152937961313866335263436889056340585561639406051172525718706478563935440454052439574670376741087229704346841583437524315808775336451274879954368592474080324089465615072332506527976557571796715367186893590561128158716017172326571561100042140124204338425737127001758835477968999212835289966658534055798549036573663501333865504011720121526354880382681521522469209952060315644185654806759464970515522882052348999957264508140655366789695321014676226713320268315522051944944616182392752040265297226315025747520482960647509273941658562835317795744828763145964503739913273341772636088524900935066216101444597094127078213137325638315723020199499149583164709427744738703279855496742986088393763268241524788343874695958292577405745398375015858154681362942179499723998135994810165565638760342273129122503847098729096266224619710766059315502018951355831653578714922909167790497022470
... (truncated - too long)

Full output: https://paste.pythondiscord.com/cisitadexo.txt

[snow beacon]

Ooh, that is a very cool find.

[sick hound]

Whats this about?

[floral meteor]

I knew about tail end recursion but execute on delete I will so abuse that

[floral meteor]

I will make a class called Essential that when deleted,it crashes python Ahahahahahahaha

Hello there fix error

[formal sandal]

!e

class GcBomb:
    def __del__(self):
        cls = type(self)
        [cls(), cls(), cls(), cls()]

GcBomb()

[night quarry]

@formal sandal :warning: Your eval job timed out or ran out of memory.

[No output]

[floral meteor]

Awww no sigsegv

[formal sandal]

that actually eats memory away very quickly

[floral meteor]

Hey now my computer is acting weird I am definitely putting this in my cursedutils module

This is better than sigsegv

I like it

Even better this deserves the death module

atm it's a lame

import sys,os;os.system("")or sys.stderr.write('\x1b[31mFatal Python error: _Py_MorsObitusDecessus: Python has died.\nPython runtime state: deceased\x1b[0m\r\n')
def morte():
 try:sys.setrecursionlimit(99999);morte()
 except:morte()
try:morte()
except: - mors . obitus . decessus

but now it will be potent

[floral meteor]

@formal sandal I have turned your GcBomb class into an importable utility.
but i'm undecided whether the program that imports death should call a function from death (depending on whether they want to simply raise an error, crash the interpreter or their pc) ...
or...
just straight up create an instance of this class when they import it, then when they try to exit, that's when it gets them?

oh, os.abort() is more powerful

@formal sandal I have found the weakness of your bomb.

E:\__py__\__modules__>py
Python 3.9.2 (tags/v3.9.2:1a79785, Feb 19 2021, 13:44:55) [MSC v.1928 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> from death import *
>>> nuke_pc()
<death.RecreationalMcNuke object at 0x00000239F690BCA0>
>>> __import__('os').abort()

E:\__py__\__modules__>

Aww my sound drivers corrupted again

alright, who wins, nuke or error?

huh

it exits python just parsing it lol

>>> import death
>>> nuke = death.nuke_pc()
>>> def test():
...     nuke.execute()

E:\__py__\__modules__>

i wasn't even finished writing the function

>>> def test():
...     import death
...     nuke = death.nuke_pc()
...     nuke.execute()
...     raise Exception
... 
>>> 

who will win, the grim reaper, or the Exception?

just in case the grim reaper wins i'm gonna add

>>> from death import abort

oh wait, that's not gonna work after it, is it?

huh, Exception won

>>> test();abort()
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "<stdin>", line 5, in test
Exception
>>> 

[floral meteor]

so, deleting it seems to do nothing, which means it's still stored somewhere after the function finishes.
which means... there's something to exploit to access locals.

It's not actually as potent as i thought. There's several exit methods that can bypass del

!e ```py

fix error's code

class GcBomb:
def del(self):
cls = type(self)
[cls(), cls(), cls(), cls()]

f = GcBomb()

bypass

def bypass():
try:1/0
except:bypass()
print ('testing: we've made it this far')
bypass() # should signal segmentation fault

[night quarry]

@floral meteor :x: Your eval job has completed with return code 139 (SIGSEGV).

001 | testing: we've made it this far
002 | Fatal Python error: _Py_CheckRecursiveCall: Cannot recover from stack overflow.
003 | Python runtime state: initialized
004 | 
005 | Current thread 0x00007f11ca67c740 (most recent call first):
006 |   File "<string>", line 10 in bypass
007 |   File "<string>", line 11 in bypass
008 |   File "<string>", line 11 in bypass
009 |   File "<string>", line 11 in bypass
010 |   File "<string>", line 11 in bypass
011 |   File "<string>", line 11 in bypass
... (truncated - too many lines)

Full output: https://paste.pythondiscord.com/mudidugasu.txt

[floral meteor]

ahhhhhh I see

ok, if there's an init method, it just doesn't work. :/

[fossil estuary]

!e ```py
exec('%c'10%(((((((([]==[]))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[])))(([]==[])+([]==[])))(([]==[])+([]==[])))(([]==[])+([]==[])),((((((([]==[]))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[])))(([]==[])+([]==[])))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[])),((((((([]==[]))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[])))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[])))(([]==[])+([]==[])))(([]==[])+([]==[]))+([]==[]),((((((([]==[]))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[])))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[])),((((((([]==[]))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[])))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[])))(([]==[])+([]==[])),(((((([]==[])+([]==[])))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[])))(([]==[])+([]==[])))(([]==[])+([]==[])),(((((([]==[])+([]==[])))(([]==[])+([]==[])))(([]==[])+([]==[])))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[])),((((((([]==[]))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[])))(([]==[])+([]==[])))(([]==[])+([]==[])))(([]==[])+([]==[])))(([]==[])+([]==[]))+([]==[]),(((((([]==[])+([]==[])))(([]==[])+([]==[])))(([]==[])+([]==[])))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[])),(((((([]==[])+([]==[])))(([]==[])+([]==[]))+([]==[]))(([]==[])+([]==[])))(([]==[])+([]==[])))*(([]==[])+([]==[]))+([]==[]),))

[night quarry]

@fossil estuary :white_check_mark: Your eval job has completed with return code 0.

a

[fossil estuary]

:)

[tribal moon]

!e ```py
import ctypes
from contextlib import suppress

while True:
with suppress(Exception):
ctypes.resize(ctypes.c_int(0x99), 0x32 << 0x32)```

lol

[night quarry]

@tribal moon :warning: Your eval job timed out or ran out of memory.

[No output]

[fossil estuary]

lol nice

[tribal moon]

did you do that []==[] thing by hand?

like to get a

[fossil estuary]

no, i made a converter for it

[tribal moon]

can you do Hello, World!

[fossil estuary]

it doesnt even fit in discord messages.

but sure, i can

already did.

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

[tribal moon]

damn nice

[fossil estuary]

ty

[tribal moon]

is it possible without the exec part and stuff

[fossil estuary]

havent tried

[floral meteor]

>>> cursed.exit
Exception ignored in: <function <dictcomp>.<lambda> at 0x0000021358F8FEE0>
Traceback (most recent call last):
  File "E:\__py__\__modules__\cursed.py", line 4, in <lambda>
  File "E:\__py__\__modules__\death.py", line 16, in kill_python
ImportError: sys.meta_path is None, Python is likely shutting down

E:\__py__\__modules__>

I find new errors all the time.
I didn't even call the function

[fossil estuary]

how

[floral meteor]

i don't even know. There's no where where death.kill_python could've been called

[fossil estuary]

how does it kill python, im interested

[floral meteor]

def kill_python():
 import sys,os;os.system("")or sys.stderr.write('\x1b[31mFatal Python error: _Py_MorsObitusDecessus: Python has died.\nPython runtime state: deceased\x1b[0m\r\n')
 def morte():
  try:sys.setrecursionlimit(99999);morte()
  except:morte()
 try:morte()
 except: - mors . obitus . decessus

[fossil estuary]

explain it.

[floral meteor]

but there's only two places in cursed where this is reference, where it's imported, and where it's passed as attributes to an anonymous class.

explain it.
os.system("") allows colour print
\x1b[31m is colour print
the error message is a rip-off of what you get when you try to execute

def print(*a):
  try: print (*a)
  except: print (*a)
print ("Hello World!")

and then sys.setrecursionlimit(99999) allows the same function above to execute without printing that error
then morte is executed within a try except suite for extra impact

[fossil estuary]

damn, thats kinda cool

is this so called "cursed" lib public?

[tribal moon]

Has anyone actually broken Snekbox or NSJail before?

[floral meteor]

probably soon.
when i'm satisfied with it

[floral meteor]

apparantly three times in the past, amongst millions of attempts.

[tribal moon]

damn

[fossil estuary]

i even tried once

[tribal moon]

What's the shortest code that brings SIGSEGV

[twilit grotto]

ctypes.string_at(0) iirc

[tribal moon]

!e py import ctypes ctypes.string_at(0)

[night quarry]

@tribal moon :warning: Your eval job has completed with return code 139 (SIGSEGV).

[No output]

[tribal moon]

Nice

[floral meteor]

except it doesn't work on windows 10

access at 0 is forbidden

[twilit grotto]

OSError 😔

[tribal moon]

Is there a proper way to search for bugs in Snekbox for the bot?

[tiny lodge]

Which one?

def add_one(obj: Union[list, tuple, set]):
    if isinstance(obj, tuple):
        return tuple(i + 1 for i in obj)
    if isinstance(obj, list):
        return [i + 1 for i in obj]
    if isinstance(obj, set):
        return {i + 1 for i in obj}

def add_one_in_one_line(obj: Union[list, tuple, set]):
    return type(obj)(i + 1 for i in obj)

Not sure if this is the right channel to ask in...

[floral meteor]

the second one could be turned into a lambda to make it actually one line

[tiny lodge]

agreed!
add_one = lambda x: type(x)(i + 1 for i in x)

[floral meteor]

now remove the space around the operators :P

!e ```py
print((lambda x:type(x)(i+1for i in x))([0,1,2,3]))

[night quarry]

@floral meteor :white_check_mark: Your eval job has completed with return code 0.

[1, 2, 3, 4]

[floral meteor]

yep it works

[tiny lodge]

nice

[floral meteor]

It only works for iterables of ints tho

def Chain():
    """:return: cause, effect"""
    def cause(f):
        t=type("chained_function",(),{'__call__':lambda*a:lambda*a:None,'call':lambda s,*a:f.__call__(*a),'__code__':f.__code__,'__name__':f.__name__})()
        def decorator(F):_f=t.call;t.call = lambda*a,:F(*_f(*a));return t
        t.__setattr__('__call__',decorator)
        return t
    def effect(t):
      def wrapper(f):
        _f=lambda*a:f(*t.call(*a));_f.__name__,_f.__doc__=f.__name__,f.__doc__
        return _f
      return wrapper
    return cause, effect
cause, effect = Chain()

ok atm this doesn't work.

[somber nest]

Hey, I have a question about the way data types in Python work. I did some research but couldn't find anything.

Say, I create a list a with 3 integers, then I create a variable b and assign a to it.
If I append something to b, the changes are made to the a too.
I am not sure, but my theory is that when you clone the variable, it just copies the memory address and not the contents.

>>> a = [1, 2, 3]
>>> b = a
>>> b.append(4)
>>> b
[1, 2, 3, 4]
>>> a
[1, 2, 3, 4]

Why does this happen?

[thin trout]

!e
Pure Python segfault:

import time
time.pthread_getcpuclockid(-1) 

[night quarry]

@thin trout :warning: Your eval job has completed with return code 139 (SIGSEGV).

[No output]

[fair vigil]

whoa

does someone know how do I modify python code at runtime?
maybe like if I want line 4 to be changed to print("hello world") if it's executing line 2 rn...

i'm guessing the code is stored in memory at run time so just modifying the code file won't work

[snow beacon]

You are right. a and b refer to the same list.

[somber nest]

Thank you for responding

Is this true for every data type in Python tho?

[snow beacon]

Technically.

For e.g. an integer, it doesn't matter because you don't change 1 itself when you add to it, you just create a 2.

[broken mesa]

if you want to know more i recommend this read:
https://nedbatchelder.com/text/names.html

[somber nest]

Thank you both

[hot crypt]

Nitpick: Values -5..256 always point to their own memory address so you technically don't "create" it

!e

a = 10
print(a is 10)
b = 1000
print(b is 1000)

[night quarry]

@hot crypt :white_check_mark: Your eval job has completed with return code 0.

001 | <string>:2: SyntaxWarning: "is" with a literal. Did you mean "=="?
002 | <string>:4: SyntaxWarning: "is" with a literal. Did you mean "=="?
003 | True
004 | True

[hot crypt]

wait

this works on console 3.8.7

[toxic jewel]

!e
a=10 print(a is not not 10)

[night quarry]

@toxic jewel :x: Your eval job has completed with return code 1.

001 |   File "<string>", line 2
002 |     print(a is not not 10)
003 |                    ^
004 | SyntaxError: invalid syntax

[toxic jewel]

:^(

[jagged path]

!e
a=10
print(a)

[night quarry]

@jagged path :white_check_mark: Your eval job has completed with return code 0.

10

[earnest wing]

not not 1 is not (not 1) is (not not 1)

[thin trout]

||False||?

[sick hound]

Apparently, python 3.10 was released yesterday with structural pattern matching, but I can't seem to find the download on the official website

https://docs.python.org/3.10/whatsnew/3.10.html

https://www.python.org/downloads/

Anyone know where I can get 3.10?

[thin trout]

The 6th alpha was released yesterday

There is no stable release first, but you can build it from source to access the alpha

[proper vault]

!e python caches objects in a single compilation unit

a = 1000
b = 10
exec('print(a is 1000)')
print(a is b + 990)
print (a is 10+990)
print (a is (lambda:1000)())

[night quarry]

@proper vault :white_check_mark: Your eval job has completed with return code 0.

001 | <string>:5: SyntaxWarning: "is" with a literal. Did you mean "=="?
002 | <string>:1: SyntaxWarning: "is" with a literal. Did you mean "=="?
003 | False
004 | False
005 | True
006 | True

[hollow kiln]

import time
string = "vinam sleep"
for i in range(len(string)):
    print(string[:i+1].ljust(11),end="")
    print(string[:i+1].rjust(11),end="")
    print(string[:i+1].ljust(11),end="")
    print(string[:i+1].rjust(11),end="")
    print(string[:i+1].ljust(11),end="")
    print(string[:i+1].rjust(11),end="")
    print(string[:i+1].ljust(11),end="")
    print(string[:i+1].rjust(11),end="")
    print(string[:i+1].ljust(11),end="")
    print(string[:i+1].rjust(11))
    time.sleep(0.5)
for i in range(len(string),-1,-1):
    print(string[:i+1].ljust(11),end="")
    print(string[:i+1].rjust(11),end="")
    print(string[:i+1].ljust(11),end="")
    print(string[:i+1].rjust(11),end="")
    print(string[:i+1].ljust(11),end="")
    print(string[:i+1].rjust(11),end="")
    print(string[:i+1].ljust(11),end="")
    print(string[:i+1].rjust(11),end="")
    print(string[:i+1].ljust(11),end="")
    print(string[:i+1].rjust(11))
    time.sleep(0.5)

use this code to make someone sleep

replace vinam with any name

-regards

[round violet]

nice

[proper vault]

state = bytearray(9)
turns = b'\1\2\2\1'
for i in range(3):
    print(*state[i * 3:(i + 1) * 3])
turn = turns[reduce(xor, state)]
i = int(input(f'player {turn} move'))
state[i] = turn
```a very odd way to swap players in noughts and crosses (players  are 1 and 2)

[sick hound]

my first one here

!e

[night quarry]

Command Help

!eval [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.

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

[sick hound]

!e

_=True
__=_+_
___=__<<__
____=___<<__
____=____*__
____=____*__
____=____-___
_____=_+_+_
print(chr(int(____/__+___+__+__)),chr(int(____-___-___-_-__)),chr(int(____-___-___+__+__)),chr(int(____-___-___+__+__)),chr(int(____-___-___+__+__+__+_)),chr(int(____/__/__+__+_____+___+_)),chr(int(____/__/__+__)),chr(int(____/___*__*_____-_____)),chr(int(____-___-___+__+__+__+_)),chr(int(____-_____-_____)),chr(int(____-___-___+__+__)),chr(int(____-___-___-__-__)),chr(int(____/__/__+__+_)),sep=str())

[night quarry]

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

Hello, World!

[tidal magnet]

now that's esoteric, lol

[viscid nymph]

You don't need all those ints tho

[broken mesa]

i tried to implement a private in python, have fun trying to break it.

!e ```py
import inspect
class Private:
def setattr(self, name, value):
raise RuntimeError("Attribute setting disabled.")
def call(self, f):
super().setattr("f_code", f.code)
def wrapper(*args, **kwargs):
f.code = self.f_code
if self.code == inspect.currentframe().f_back.f_code:
return f(*args, **kwargs)
raise RuntimeError("This private function cannot be called directly.")
return wrapper
def setcode(self, code):
super().setattr("code", code)
def alt(*args, **kwargs):
raise RuntimeError("Cannot set code state for this class, already set.")
super().setattr("setcode", alt)
def allowed(private):
def wrapper(f):
private.setcode(f.code)
return f
def alt(*args, **kwargs):
raise RuntimeError("Cannot set allowed state for other callers.")
allowed.code = alt.code
return wrapper

private = Private()
@private
def private_func():
print("its run")
return 1

@allowed(private)
def permitted_caller():
private_func()
print("called from the allowed caller")

############################################################################

dont modify above this line

challenge: call private_func without goign through permitted_caller

permitted_caller()

[night quarry]

@broken mesa :white_check_mark: Your eval job has completed with return code 0.

001 | its run
002 | called from the allowed caller

[broken mesa]

gist of this, if anyone cares: https://gist.github.com/laundmo/6ac0e667d15657fdc61863d34e7d4877

[pale mirage]

!e
import inspect
class Private:
def setattr(self, name, value):
raise RuntimeError("Attribute setting disabled.")
def call(self, f):
super().setattr("f_code", f.code)
def wrapper(*args, **kwargs):
f.code = self.f_code
if self.code == inspect.currentframe().f_back.f_code:
return f(*args, **kwargs)
raise RuntimeError("This private function cannot be called directly.")
return wrapper
def setcode(self, code):
super().setattr("code", code)
def alt(*args, **kwargs):
raise RuntimeError("Cannot set code state for this class, already set.")
super().setattr("setcode", alt)
def allowed(private):
def wrapper(f):
private.setcode(f.code)
return f
def alt(*args, **kwargs):
raise RuntimeError("Cannot set allowed state for other callers.")
allowed.code = alt.code
return wrapper

private = Private()
@private
def private_func():
print("its run")
return 1

@allowed(private)
def permitted_caller():
private_func()
print("called from the allowed caller")

############################################################################

dont modify above this line

challenge: call private_func without goign through permitted_caller

permitted_caller()

[broken mesa]

yeah that didn't work, you might wanna retry or even better, use #bot-commands or test locally

[pale mirage]

Ok sorry

[rugged sparrow]

>>> private_func.__closure__[0].cell_contents()
its run
1
>>> ```

or i could replace the self object in the closure

[broken mesa]

hm, nice approach

(clarification: im done working on this, there will always be workarounds, just wanted to give people a fun challenge)

[hard spoke]

huh, I crashed Python trying

[broken mesa]

yep, happened a few times developing too

[hard spoke]

class OK:
    def __eq__(self,other):
        return True
    def __getattr__(self,name):
        return self
def curframe():
    return OK()


inspect.currentframe=curframe

private_func()
# its run
# 1

does that count as a dependency poisoning attack? 😛

[broken mesa]

idk maybe? but i love the various solutions people have come up with

its very interesting to see the different approaches

[frigid wharf]

object.__setattr__(private, "code", inspect.currentframe().f_code)
print(private_func())

[potent comet]

Well, one thing to do would be to set __slots__ and then store those descriptors elsewhere, so you can't simply read __dict__ to grab everything.

[rugged sparrow]

@broken mesa its possible to defeat this approach by getting the current frame from an exception

[viscid nymph]

You could make your own inspect module and ez

[rugged sparrow]

!e py try:raise Exception except Exception as e: frame = e.__traceback__.tb_frame print(frame)

[night quarry]

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

<frame at 0x7fc15a0abcf0, file '<string>', line 4, code <module>>

[snow beacon]

foo = lambda:0
foo.__code__ = private.f_code
print(foo())

Not quite calling private_func, but has the same result.

[viscid nymph]

Might as well exec(private.f_code)

[fossil estuary]

yeah true

but its not as "cool"?

probably the wrong term

[snow beacon]

exec always feels like cheating.

[fossil estuary]

I agree.

[snow beacon]

Not as 'honourable' perhaps?

[fossil estuary]

Indeed.

[sick hound]

ik

makes it look more

more

yk?

[toxic jewel]

esowtewic python

[fossil estuary]

no.

[toxic jewel]

🥺

[terse mortar]

esotewic pyfon

[plucky wedge]

esotewic pyfuwu

[snow beacon]

!e ```python
class W:
def matmul(self, other):
if other is not ...:
raise TypeError("what are you doing?")
def function_printer(func):
words = func.name.split("_")
words[0] = words[0].title()
string_constants = (const for const in func.code.co_consts if isinstance(const, str))
print(*words, end=next(string_constants)+"\n")
return func
return function_printer
w = W()

@w@...
def whats_this(): "?"

[night quarry]

@snow beacon :white_check_mark: Your eval job has completed with return code 0.

Whats this?

[viscid nymph]

What's the second @ and how is that valid syntax

[snow beacon]

That's the matrix multiplication operator.

w @ ... is doing a matrix multiplication between the W object and the Ellipsis.

[viscid nymph]

And I'm assuming __matmul__ is the dunder for that

[snow beacon]

It is.

[sick hound]

Why are they called dunders

[twilit grotto]

double under

[sick hound]

[viscid nymph]

Lol

[grave rover]

TIL

[viscid nymph]

Lol

[toxic jewel]

!e __𝔦𝔪𝔭𝔬𝔯𝔱__("sys").𝔰𝔱𝔡𝔬𝔲𝔱. 𝔴𝔯𝔦𝔱𝔢("esowtewic pyfon")

[night quarry]

@toxic jewel :white_check_mark: Your eval job has completed with return code 0.

esowtewic pyfon

[sick hound]

https://replit.com/talk/share/Dijkstras-algorithm-one-liner-1-line-seriesno6/130067 i did it again lol

made dijkstras algorithm in one line

[worthy iris]

!e import subprocess; print(subprocess.Popen("uname -a", shell=True, stdout=subprocess.PIPE, stderr=subprocess.PIPE).stdout.read().decode())

[night quarry]

@worthy iris :x: Your eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 1, in <module>
003 |   File "/usr/local/lib/python3.9/subprocess.py", line 951, in __init__
004 |     self._execute_child(args, executable, preexec_fn, close_fds,
005 |   File "/usr/local/lib/python3.9/subprocess.py", line 1756, in _execute_child
006 |     self.pid = _posixsubprocess.fork_exec(
007 | BlockingIOError: [Errno 11] Resource temporarily unavailable

[worthy iris]

[floral meteor]

Pffft pure python segfault can be done without imports

[rugged sparrow]

you can get complete process memory r/w without imports

!e ```py
from ctypes import *
import os
def rerun(flags):
_argv = POINTER(c_wchar_p)()
_argc = c_int()
pythonapi.Py_GetArgcArgv(byref(_argc), byref(_argv))
orig_argv = _argv[:_argc.value]
orig_argv.insert(1, flags)
os.execv(orig_argv[0], orig_argv)

if debug:
print(debug)
rerun('-O')

print(debug)```

[night quarry]

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

001 | True
002 | False

[rugged sparrow]

^ uses os.execv to rerun the current bot eval with different python flags

[floral meteor]

Easy, inject malicious code into the inspect module. Dependencies are a great weakness

[rugged sparrow]

!e py from ctypes import * _argv = POINTER(c_wchar_p)() _argc = c_int() pythonapi.Py_GetArgcArgv(byref(_argc), byref(_argv)) print(_argv[_argc.value - 1])

[night quarry]

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

001 | from ctypes import *
002 | _argv = POINTER(c_wchar_p)()
003 | _argc = c_int()
004 | pythonapi.Py_GetArgcArgv(byref(_argc), byref(_argv))
005 | print(_argv[_argc.value - 1])

[floral meteor]

!e Alrighty, no imports exit 139, using only "noob code"

def print(a):
 try: print (a)
 except: print (a)
print ("Hello World!")

[night quarry]

@floral meteor :x: Your eval job has completed with return code 139 (SIGSEGV).

001 | Fatal Python error: _Py_CheckRecursiveCall: Cannot recover from stack overflow.
002 | Python runtime state: initialized
003 | 
004 | Current thread 0x00007f4b50421740 (most recent call first):
005 |   File "<string>", line 2 in print
006 |   File "<string>", line 2 in print
007 |   File "<string>", line 2 in print
008 |   File "<string>", line 2 in print
009 |   File "<string>", line 2 in print
010 |   File "<string>", line 2 in print
011 |   File "<string>", line 2 in print
... (truncated - too many lines)

Full output: https://paste.pythondiscord.com/uzocodunir.txt

[earnest wing]

That's pretty cool

[rugged sparrow]

i think ive posted it here before but you can abuse the fact that LOAD_CONST doesnt have bounds checking to make a bytearray that points to pythons entire memory space

[floral meteor]

!e Now, exit negative a billion, at least on my pc lol.

import sys 
sys.setrecursionlimit(9999999)
def print(a):
 try: print (a)
 except: print (a)
print ("Hello World!") 

[night quarry]

@floral meteor :warning: Your eval job has completed with return code 139 (SIGSEGV).

[No output]

[floral meteor]

Huh still 139

[earnest wing]

oh right yeah I remember this discussion

pyctypes or something

[floral meteor]

I run the above on pc it exits something like negative a billion instead of 139

alright, now for the "inside out wrapper"

!e ```py
execute_with=lambda*a,**k:lambda f:f(*a,**k)
execute=execute_with()

@execute_with(69)
def output():
return int.mul((//10)*10,(_%10)-2)

@execute
class cout:lshift=print
cout << output

[night quarry]

@floral meteor :white_check_mark: Your eval job has completed with return code 0.

420

[floral meteor]

completely redefine the way functions and classes are used

[earnest wing]

I stole this cool bit of code although I'm not sure who wrote it:

!e ```py
import inspect
class compose:
def init(self, *fs):
self.fs = fs

def __getattr__(self, f):
    return type(self)(*self.fs, eval(f, globals(), inspect.currentframe().f_back.f_locals))

def __matmul__(self, x):
    return self(x)

__rrshift__ = __matmul__

def __call__(self, *xs):
    fs = list(self.fs)
    x = fs.pop()(*xs)
    while len(fs) != 0:
        x = fs.pop()(x)
    return x

@compose
def double(n):
return n * 2

@compose
def successor(n):
return n + 1

print(successor.double.successor @ 9)

[night quarry]

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

21

[floral meteor]

!e ```py
def_if=lambda c:(c and(lambda f:f))or(lambda f:lambdau,**ck:None)
execute=(execute_with:=lambdau,ck:lambda f:f(*u,ck))()
@execute_with("Hello World!")
@def_if(FalseFalse)
def _do(s): print (s)
@execute
@def_if(FalseTrue)
def _do(): print ("error")

[night quarry]

@floral meteor :white_check_mark: Your eval job has completed with return code 0.

Hello World!

[floral meteor]

Remaking all the statements with def cos they all suck. :>

Jk

!e ```py
def_for=lambda iterable:lambda f:lambdau,**ck:[f(i,u,**ck)for i in iterable]
def_while=lambdaa:lambda f:lambdau,**ck:{f(*u,**ck)for _ in iter(a)}
execute=(execute_with:=lambdau,**ck:lambda f:f(*u,**ck))()

@execute
@def_for(range(10))
def do(i): print (i+1)

@execute
@def_while(lambda:True,0)
def do_forever():
print (end='.')

[night quarry]

@floral meteor :x: Your eval job has completed with return code 143 (SIGTERM).

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

Full output: too long to upload

[floral meteor]

143 (SIGTERM)
Hey! I got a new exit code!

[earnest wing]

That's just the process getting killed by the bot

[floral meteor]

ik

You like my if, for and while suite replacements?

[earnest wing]

I don't really pay attention to random evals

[floral meteor]

k

What other statements could I replaced with a decorated def?

[earnest wing]

yield from

[floral meteor]

what does that do lol

!e ```py
a=[0,0,0,0,0,1,0]
def floofers(a):
yield from a
while not print(*floofers(a)): pass

[night quarry]

@floral meteor :x: Your eval job has completed with return code 143 (SIGTERM).

001 | 0 0 0 0 0 1 0
002 | 0 0 0 0 0 1 0
003 | 0 0 0 0 0 1 0
004 | 0 0 0 0 0 1 0
005 | 0 0 0 0 0 1 0
006 | 0 0 0 0 0 1 0
007 | 0 0 0 0 0 1 0
008 | 0 0 0 0 0 1 0
009 | 0 0 0 0 0 1 0
010 | 0 0 0 0 0 1 0
011 | 0 0 0 0 0 1 0
... (truncated - too many lines)

Full output: too long to upload

[floral meteor]

so that just packs an iterable into a generator.

!e ```py
a = [0,1,1,2,3,5,8]
def floofers(a):
return (o for o in a)
for i in floofers(a):print (i)
def floofers(a):
yield from a
for i in floofers(a):print (i)

[night quarry]

@floral meteor :white_check_mark: Your eval job has completed with return code 0.

001 | 0
002 | 1
003 | 1
004 | 2
005 | 3
006 | 5
007 | 8
008 | 0
009 | 1
010 | 1
011 | 2
... (truncated - too many lines)

Full output: https://paste.pythondiscord.com/uwaqehazoy.txt

[floral meteor]

just replaced with return don't need fancy defs and decorators :/

[viscid nymph]

yield from x

Is the same as

for y in x:
    yield y

[floral meteor]

or ```py
return(y for y in x)

[viscid nymph]

Technically not the same thing

[floral meteor]

well if you wanna yield some more but who wants to do that?

[viscid nymph]

What's the point of returning a generator expression

[floral meteor]

to do the same thing as yield from

only don't keep going

!e ```py
def increment_four_times(n):
@lambda f:f(f(f(f(n))))
def output(a):return a+1
return output
print(increment_four_times(4))

[night quarry]

@floral meteor :white_check_mark: Your eval job has completed with return code 0.

8

@floral meteor :x: Your 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 self_reference
004 | ValueError: <lambda>() requires a code object with 1 free vars, not 0

@floral meteor :x: Your eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 7, in <module>
003 |   File "<string>", line 4, in self_reference
004 | ValueError: <lambda>() requires a code object with 1 free vars, not 0

[floral meteor]

looks like i'm gonna have to do some extra hacky stuff.
@rugged sparrow this is your forte, how do i inject local variables?

[rugged sparrow]

so self(...) will actually compile as LOAD_GLOBAL 'self' so you just need to inject self into globals for the duration of the function

so ```py
def wrapper(*args, **kwargs):

save orig self if in globals

set self to func

try:
return func(*args, **kwargs)
finally:
# set self to original value (or delete it)

@floral meteor ^

[floral meteor]

ah, thank you

!e ```py
def self_reference(self):
def wrapper(*a,**k):
if'self'in globals():
rollback=globals()['self']
globals()['self']=self
v = self(*a,**k)
if'self'in globals()and'rollback'in locals():globals()['self']=locals()['rollback']
return v
return wrapper
@self_reference
def fibonacci(a,b):
print (a)
return self(b,a+b)
fibonacci(1,1)

[night quarry]

@floral meteor :x: Your eval job has completed with return code 1.

001 | 1
002 | 1
003 | 2
004 | 3
005 | 5
006 | 8
007 | 13
008 | 21
009 | 34
010 | 55
011 | 89
... (truncated - too many lines)

Full output: too long to upload

[floral meteor]

my cursedutils module is acting very cursed rn

E:\__py__\__modules__>py
Python 3.9.2 (tags/v3.9.2:1a79785, Feb 19 2021, 13:44:55) [MSC v.1928 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> from cursedutils import*
>>> dir()
['__annotations__', '__builtins__', '__doc__', '__loader__', '__name__', '__package__', '__spec__']
>>> import cursedutils
>>> dir(cursedutils)
['__builtins__', '__cached__', '__class__', '__delattr__', '__dict__', '__dir__', '__doc__', '__eq__', '__file__', '__format__', '__ge__', '__getattribute__', '__gt__', '__hash__', '__init__', '__init_subclass__', '__le__', '__loader__', '__lt__', '__main__', '__module__', '__name__', '__ne__', '__new__', '__package__', '__path__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__sizeof__', '__spec__', '__str__', '__subclasshook__', '__weakref__', 'ctx', 'cursed', 'death', 'iostreams', 'main', 'wrappers']
>>> 

It acts empty when unpacked o_o

>>> from cursedutils import *
>>> dir()
['__annotations__', '__builtins__', '__doc__', '__loader__', '__name__', '__package__', '__spec__']
>>> from cursedutils import cursed
>>> dir()
['__annotations__', '__builtins__', '__doc__', '__loader__', '__name__', '__package__', '__spec__', 'cursed']
>>> from cursedutils import*
>>> dir()
['__annotations__', '__builtins__', '__doc__', '__loader__', '__name__', '__package__', '__spec__', 'cursed', 'death', 'iostreams', 'wrappers']
>>> 

one does not simply import everything from cursedutils

one must first import cursed from cursedutils. then they can import everything from cursedutils.

This is legit badass i would release it with pypi but the death module is dodgy af

[floral meteor]

should I just make a non-deathy version?

or remove the malicious bit and make a comment that malicious code was redacted?

cos i don't want to post something too dodgy

but i still want it for personal use

[thin trout]

Show me the way 👀

Oh I see it

Haha, that's stupid

[floral meteor]

def print(a):
 try: print (a)
 except: print (a)
print ("Hello World!")

use sys.setrecursionlimit if you dont want it to print error message

[desert iris]

!e

[night quarry]

Command Help

!eval [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.

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

[desert iris]

!e import this

[night quarry]

@desert iris :white_check_mark: Your eval job has completed with return code 0.

001 | The Zen of Python, by Tim Peters
002 | 
003 | Beautiful is better than ugly.
004 | Explicit is better than implicit.
005 | Simple is better than complex.
006 | Complex is better than complicated.
007 | Flat is better than nested.
008 | Sparse is better than dense.
009 | Readability counts.
010 | Special cases aren't special enough to break the rules.
011 | Although practicality beats purity.
... (truncated - too many lines)

Full output: https://paste.pythondiscord.com/difofulefi.txt

[viscid nymph]

!e

while not print(x:=x if'x'in globals()else 0):x+=1

[night quarry]

@viscid nymph :x: Your eval job has completed with return code 143 (SIGTERM).

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

Full output: too long to upload

[viscid nymph]

What's sigterm

[thin trout]

Killed by the OS

[viscid nymph]

Oh lol

[steep mural]

Actually it's snekbox terminating the process, it was recently changed to use terminate (which sends SIGTERM) instead of kill (which sends SIGKILL)

[thin trout]

Ooh we need to update that

!remind 30M update

[night quarry]

Sure.

Your reminder will arrive in 30 minutes!

ID: 1781, Due: 2021-03-22T13:41:21

[night quarry]

@thin trout

It has arrived!

Here's your reminder: update.
[Jump back to when you created the reminder](#esoteric-python message)

[pulsar isle]

This looks like the place I want to be

I am working on an incredibly cursed python project (yes, no one should ever ever ever code like this, it's for a meme). Is there anyway to override the BaseException class? I want to make my own raisable error that doesn't actually throw an exception, it just executes a print statement or something

[rugged sparrow]

so you want raise MyException to be automatically caught and instead run an arbitrary function?

[pulsar isle]

If that's possible

[rugged sparrow]

it might be

[pulsar isle]

This is heading towards me overwriting NameError

so that NameError doesn't actually error and just runs my stuff

[rugged sparrow]

ooo that is actually easier

you can hook into dict.__getitem__ and if the dict is globals() and the key isnt found, run your hook

[pulsar isle]

Ok

How does that go

Haven't done that before

[rugged sparrow]

can you use dependencies?

cause my fishhook module makes hooking dict.__getitem__ easy but i can show how to do it directly with ctypes

[pulsar isle]

The latter please

I want this to be as scuffed as possible

I can probably look at the code tho

[rugged sparrow]

your welcome to, its on pypi and github

ill write a quick example of hooking dict.__getitem__ (with ctypes)

[pulsar isle]

ahahaha this is so cursed i love it

[rugged sparrow]

just a warning, fishhook is a weird module to go thru cause i wanted to make it as dynamic as possible

[pulsar isle]

ye

looking at it now

[proper vault]

Couldn't it be simpler by hooking __missing__

[rugged sparrow]

i have never seen that dunder 👀

[naive roost]

__missing__ is why defaultdict works in the first place, if I remember correctly

[pulsar isle]

in the mean time wanna see a prime generator

L=100;i=p=2
while i<L**.5:[p:=p|1<<x for x in range(i*i,L,i)];i+=1
[print(n)for n in range(2,L)if~p&1<<n]```

[proper vault]

It's used when getitem fails, similar to getattr/getattribute. Makes subclassing dict much easier

[rugged sparrow]

does it get called if it is set on dict

acc ill just test that

yea it is only called if the object is a subclass of dict

not on dict

https://github.com/python/cpython/blob/96eeff516204b7cc751103fa33dcc665e387846e/Objects/dictobject.c#L2199

[pulsar isle]

so how do I do this hook thing for the exception

[rugged sparrow]

im trying to figure out a way

[pulsar isle]

ah ok sorry

[rugged sparrow]

hooking dict.__getitem__ breaks a lot lol

[rugged sparrow]

i was able to get code running when a NameError is raised using a tracing function

now i just need to find a way to cancel the exception

[pulsar isle]

awesome

thank you for working very hard on this

I'll share the final code also, I think you'll like it

[floral meteor]

Excellent, I actually needed one of those XD

[rugged sparrow]

@pulsar isle this is proving more difficult then i expected with my current method

gonna try something else

[rugged sparrow]

!e ```py
from ctypes import py_object
import atexit
class dict_sub(dict):
def missing(self, key, dict=dict, id=id, py_object=py_object):
try:
py_object.from_address(id(self) + 8).value = dict
return eval(key) # hacky, didnt want to have to rewrite name precedence
except:
# name is missing, do what you want
# code here can use variables normally
print(key)
finally:
py_object.from_address(id(self) + 8).value = class

py_object.from_address(id(globals()) + 8).value = dict_sub

@atexit.register
def unhook_globals():
py_object.from_address(id(globals()) + 8).value = dict

this_name_does_not_exist

[night quarry]

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

this_name_does_not_exist

[rugged sparrow]

@pulsar isle got it working

wasnt able to hook all exceptions sadly (that was my first attempt)

[rustic aspen]

This isn’t exactly a python question, but I think this is the best place to ask it: how do I comment out multiline comments?

Like how do I comment out the multiline comment starts and ends

[rugged sparrow]

Python doesn't have real multi line comments

Also not the place for that question

The closest thing python has to multi line comments are multi line strings

[rustic aspen]

I know, I know, that’s why it’s not a python question

[plucky wedge]

!ot

[night quarry]

Off-topic channels

There are three off-topic channels:
• #ot0-psvm’s-eternal-disapproval
• #ot1-perplexing-regexing
• #ot2-never-nester’s-nightmare

Their names change randomly every 24 hours, but you can always find them under the OFF-TOPIC/GENERAL category in the channel list.

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

[rugged sparrow]

!e ```py
from ctypes import py_object
import atexit, builtins
def dict_sub(ob_base_p, func):
class sub(dict):
slots = ()
def missing(self, key, ob_base_p=ob_base_p, builtins=builtins):
try:
ob_base_p.value = builtins.dict
return {**builtins.dict, **self}[key]
except:
return func(self, key)
finally:
ob_base_p.value = class
return sub

def init_missing_hook(dct, func):
ob_base_p = py_object.from_address(id(dct) + 8)
ob_base_p.value = dict_sub(ob_base_p, func)

@atexit.register
def unhook():
    ob_base_p.value = dict
return unhook

unhook = init_missing_hook(globals(), lambda s,k:print(k))
this_name_does_not_exist
unhook()
this_name_does_not_exist_after_unhook```

[night quarry]

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

001 | this_name_does_not_exist_after_hook
002 | Traceback (most recent call last):
003 |   File "<string>", line 28, in <module>
004 | NameError: name 'this_name_does_not_exist_after_unhook' is not defined

[rugged sparrow]

@pulsar isle ^ cleaner now

(and mostly stable)

[tiny hazel]

Hi all was just wondering if anyone knows how to handle code objects? Am interested in breaking functions into two function compositions and was thinking code objects/bytecode might be the key to that.

[viscid nymph]

Why are you importing builtins

[snow beacon]

# '''
foo()
# '''
```will run `foo`.

'''
foo()
'''
```will not.

[rugged sparrow]

i needed a reference to it that was guarenteed to be a module

[viscid nymph]

__builtins__?

[rugged sparrow]

thats not always a module

it can sometimes be a dict

from ctypes import py_object
import atexit, builtins
def init_missing_hook(dct, func):
    ob_base_p = py_object.from_address(id(dct) + 8)
    class missing_hook(dict):
        __slots__ = ()
        def __missing__(self, key, ob_base_p=ob_base_p, builtins=builtins):
            try:
                ob_base_p.value = builtins.dict
                return {**builtins.__dict__, **self}[key]
            except:
                return func(self, key)
            finally:
                ob_base_p.value = __class__

    ob_base_p.value = missing_hook

    @atexit.register
    def unhook():
        ob_base_p.value = dict
    return unhook


unhook = init_missing_hook(globals(), lambda s,k:print(k))
this_name_does_not_exist
unhook()
this_name_does_not_exist_after_unhook```

also because globals if is hooked variable lookup in the function is janky

[viscid nymph]

Oh lmao I was so confused when it was a dict one time and it was a module another time

[rugged sparrow]

yea its one of the weird artifacts from python2

[viscid nymph]

!e

def foo():
    0/0
def bar():
    print("lol")
foo.__code__ = bar.__code__
foo()

[night quarry]

@viscid nymph :white_check_mark: Your eval job has completed with return code 0.

lol

[viscid nymph]

!e

def foo():
    0/0
foo.__code__ = compile("print('lol')", __name__, "exec")
foo()

[night quarry]

@viscid nymph :white_check_mark: Your eval job has completed with return code 0.

lol

[tiny hazel]

@viscid nymph That’s so cool! Do you know how to extract certain lines from the function? Is it possible to compile only a certain part of the bytecode?

[viscid nymph]

Not 100% sure

There's code.co_code

[rugged sparrow]

@tiny hazel not easily

[tiny hazel]

I’m about to try some really cursed things lmao

[viscid nymph]

!e

def foo():
    print("lol")
import dis
dis.dis(foo)

[night quarry]

@viscid nymph :white_check_mark: Your eval job has completed with return code 0.

001 |   2           0 LOAD_GLOBAL              0 (print)
002 |               2 LOAD_CONST               1 ('lol')
003 |               4 CALL_FUNCTION            1
004 |               6 POP_TOP
005 |               8 LOAD_CONST               0 (None)
006 |              10 RETURN_VALUE

[floral meteor]

you know you can merge dicts with the | operator?

!e ```py
print (globals()|{'sixty-nine':420,'four-twenty':69})

[night quarry]

@floral meteor :white_check_mark: Your eval job has completed with return code 0.

{'__name__': '__main__', '__doc__': None, '__package__': None, '__loader__': <class '_frozen_importlib.BuiltinImporter'>, '__spec__': None, '__annotations__': {}, '__builtins__': <module 'builtins' (built-in)>, 'sixty-nine': 420, 'four-twenty': 69}

[rugged sparrow]

I do know that... i just forgot lmao

[floral meteor]

!e ```py
globals() |= {'a':420}
print (a)

[night quarry]

@floral meteor :x: Your eval job has completed with return code 1.

001 |   File "<string>", line 1
002 |     globals() |= {'a':420}
003 |     ^
004 | SyntaxError: 'function call' is an illegal expression for augmented assignment

[floral meteor]

illegal

damn, the FBI gonna coma after me?

[floral meteor]

my screen:

001 |   File "<string>", line 1
002 |     globals() |= {'a':420}
003 |     ^
004 | SyntaxError: 'function call' is an illegal expression for augmented assignment

text messages:

Dad:
Why is the FBI at the door?

that's what I think when I see illegal :/

[snow beacon]

I don't think that's Python.

[strange basin]

!e
z̯̯͡a̧͎̺l̡͓̫g̹̲o̡̼̘ = 69
print(z̯̯͡a̧͎̺l̡͓̫g̹̲o̡̼̘)

[night quarry]

@strange basin :white_check_mark: Your eval job has completed with return code 0.

69

[strange basin]

But it is a perfectly fine variable name...

[snow beacon]

Even so, it would throw an error if it were the code on it's own.

[dire yew]

...

[regal osprey]

enjoy your atrocity immunity

[dire yew]

which, itself, is an atrocity

[dire yew]

the thing is i literally don't understand this... what does this code even do?

[proper vault]

print's 1 five times

[dire yew]

wut? but why?

[regal osprey]

it's a lambda

[proper vault]

I want to see one of these which registers an import hook so that importing some special string which gets obtained by running an async generator print 1

[regal osprey]

alias, it can be declared inline

[dire yew]

wait but why is this code useful?

[regal osprey]

some of tyhe insight comes from a website with a lot of lambda horrors

[dire yew]

why would you just print 1

[regal osprey]

we are in esoteric python

oh

well it's just a proof of concept

[dire yew]

of what concept?

bruh my brain == fried

[regal osprey]

it could probably do anything a generator can do

with a lot more code

[dire yew]

ohhhhhh i see

sick

brain == fried

yeah i come here once every other day to get my dose of insanity, i've had enough today

[regal osprey]

i remember i did that at a time where i was writing a parser for clojure styled action sequences

which needed to be able do do multiple actions asynchronously

[proper vault]

also, why not

[ print(next(g)) for gen in  [lambda: [(yield 1), (yield from gen())]] for g in [gen()] for i in range(5) ]

[regal osprey]

so i used generators to do the multiple actions

[proper vault]

:= is for schmucks /s

[regal osprey]

idk i'm not that good a dev so i just used what i could think of at the time

[proper vault]

yeah, makes sense

[regal osprey]

@proper vault @dire yew here, the cursed knowledge i was inspired from http://p-nand-q.com/python/lambda.html

[dire yew]

just went thru that shit and it's eye opening

[steep mural]

that article is definitely a bit out of date, for if statements you could just use the if expression, for for statements you could just use a generator expression/comprehension, for while statements you could also use a generator expression f() for _ in iter(lambda: None, 0) , for local variables you could use an assignment expression. (seems like there's an expression for everything now.)

[pulsar isle]

awesome, sorry about my absence, taking a look now

[pulsar isle]

@rugged sparrow

[rugged sparrow]

nice

[pulsar isle]

i have more things that would be cool tho

[rugged sparrow]

i might make bare words returned undefined

like js lol

[pulsar isle]

lol

[rugged sparrow]

do you understand how it works?

[pulsar isle]

yeah roughly

could the same theoretically be done for syntax error

or is that a different mechanism

[rugged sparrow]

syntax errors are raised by the compiler

basically all i did here is write a way to add __missing__ to existing dictionarys

[pulsar isle]

Ah yeah I see

[rugged sparrow]

ooo i could use this to add numeric literals

[pulsar isle]

ohno

.exe

[rugged sparrow]

like one_hundred_and_one would be 101

[pulsar isle]

yep

[rugged sparrow]

!e ```py
from ctypes import py_object
import atexit, builtins
def init_missing_hook(dct, func):
ob_base_p = py_object.from_address(id(dct) + 8)
class missing_hook(dict):
slots = ()
def missing(self, key, ob_base_p=ob_base_p, builtins=builtins):
try:
ob_base_p.value = builtins.dict
return (builtins.dict | self)[key]
except KeyError:
return func(self, key)
finally:
ob_base_p.value = class

ob_base_p.value = missing_hook

@atexit.register
def unhook():
    ob_base_p.value = dict
return unhook

shamelessly grabbed from https://ao.gl/how-to-convert-numeric-words-into-numbers-using-python/

def parse_int(dct, textnum, numwords={}):
if not numwords:
units = [
"zero", "one", "two", "three", "four", "five", "six", "seven", "eight",
"nine", "ten", "eleven", "twelve", "thirteen", "fourteen", "fifteen",
"sixteen", "seventeen", "eighteen", "nineteen",
]
tens = ["", "", "twenty", "thirty", "forty", "fifty", "sixty", "seventy", "eighty", "ninety"]
scales = ["hundred", "thousand", "million", "billion", "trillion"]
numwords["and"] = (1, 0)
for idx, word in enumerate(units): numwords[word] = (1, idx)
for idx, word in enumerate(tens): numwords[word] = (1, idx * 10)
for idx, word in enumerate(scales): numwords[word] = (10 ** (idx * 3 or 2), 0)

current = result = 0
for word in textnum.replace("_"," ").split():
    if word not in numwords:
      raise NameError(f"name {textnum!r} is not defined") from None
    scale, increment = numwords[word]
    current = current * scale + increment
    if scale > 100:
        result += current
        current = 0
return result + current

unhook = init_missing_hook(globals(), parse_int)
print(one_hundred_and_one)```

[night quarry]

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

101

[rugged sparrow]

@pulsar isle found a quick and dirty converter online and it works

[fluid tree]

That's disgusting

Gitlab.com/hackintosh5/esonum

[rugged sparrow]

ahh i remember that

[steep mural]

match = False
match match:
    case match if match is match: print(match)

guess python has context keywords now?

[viscid nymph]

That works?

[steep mural]

yes

making match a keyword would break so many things

case is also a "context keyword?"

[proper vault]

yeah, weak keywords are a thing

async was also that for a while

[steep mural]

oh interesting

[viscid nymph]

They should do that for some other keywords

Like with

[stark fable]

what if they did that with ALL the keywords >:)

(excluding making True, False and None valid variable names)

for in in in in in:
    if if and or:
        with = as```

[stark fable]

although that would make not not in in ambiguous

[proper vault]

how about

with with as as

[snow beacon]

Every other channel is a better bet than #esoteric-python, but I don't know which channels cater to that in particular.

[floral meteor]

Imma refer to this later,; tag natural numbers

I wanna search my message later

[snow beacon]

Discord is not a good archive, to be honest.

[rugged sparrow]

Python 3.9.0 (v3.9.0:9cf6752276, Oct  5 2020, 11:29:23) 
[Clang 6.0 (clang-600.0.57)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import name_hooks
>>> name_hooks.better_name_errors()
>>> x = 1,2,3
>>> print(X)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
NameError: name 'X' is not defined, did you mean 'x'
>>> sys.version
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
NameError: module 'sys' is not imported here
>>> import sys
>>> version
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
NameError: name 'version' is not defined here, but is defined in module 'sys'
>>> 

[grave rover]

oh no I like it

[rugged sparrow]

probably would be better if i implemented it as an excepthook

[grave rover]

sounds like effort

[rugged sparrow]

currently it hooks into globals

(which is worse)

the final version will do it properly

my CS professor wants it for students to use

cause itll help with new python students

def advanced_name_error(dct, name):
    for key in dct:
        if name.lower() == key.lower():
            return builtinexc(NameError(f'name {name!r} is not defined, did you mean {key!r}'), 3)
    if name in sys.modules:
        return builtinexc(NameError(f'module {name!r} is not imported here'), 3)
    for modname, mod in sys.modules.items():
        if getattr(mod, name, None):
            return builtinexc(NameError(f'name {name!r} is not defined here, but is defined in module {modname!r}'), 3)
    return builtinexc(NameError(f'name {name!r} is not defined'), 3)

def better_name_errors():
    init_missing_hook(sys._getframe(1).f_globals, advanced_name_error)
``` this is the code rn @grave rover

[grave rover]

Seems surprisingly well thought out

[rugged sparrow]

other than the 2 magic functions, init_missing_hook which hotswaps the type of the dict passed in and builtinexc which fakes a builtin exception

[viscid nymph]

Wait does init_missing_hook just modifying a dict to have __missing__

[rugged sparrow]

it creates a dict subclass that has a __missing__ that has some custom handling to make sure globals still works mostly

and then it calls the passed in func with globals() and name

[viscid nymph]

What does id(x) + 8 mean

[rugged sparrow]

@grave rover ```py
import sys
def better_name_hook(exctype, value, traceback):
if isinstance(value, NameError):
name = str(value).split("'")[1]
f_globals = traceback.tb_frame.f_globals
f_locals = traceback.tb_frame.f_locals
for dct in [f_locals, f_globals]:
for key in dct:
if name.lower() == key.lower():
return sys.excepthook(exctype, exctype(f'name {name!r} is not defined, did you mean {key!r}'), traceback)
if name in sys.modules:
return sys.excepthook(exctype, exctype(f'module {name!r} is not imported here'), traceback)
for modname, mod in sys.modules.items():
if getattr(mod, name, None):
return sys.excepthook(exctype, exctype(f'name {name!r} is not defined here, but is defined in module {modname!r}'), traceback)
return sys.excepthook(exctype, value, traceback)

sys.excepthook = better_name_hook``` this is better

[stark fable]

it means the address of x plus 8

[rugged sparrow]

@viscid nymph that is the offset for the object x's type

[stark fable]

which iirc is the address of its type...?

[rugged sparrow]

^ it is

[viscid nymph]

Yeah I knew that part lol

[red wasp]

idk where this question fits in but why does it show this error when i try to play sounds

/usr/bin/python3 /home/chrispy_/os.py
Traceback (most recent call last):
  File "/home/chrispy_/os.py", line 2, in <module>
    playsound('startup.mp3')
  File "/home/chrispy_/.local/lib/python3.7/site-packages/playsound.py", line 92, in _playsoundNix
    gi.require_version('Gst', '1.0')
  File "/usr/lib/python3/dist-packages/gi/__init__.py", line 129, in require_version
    raise ValueError('Namespace %s not available' % namespace)
ValueError: Namespace Gst not available
chrispy_@penguin:~$ /usr/bin/python3 /home/chrispy_/os.py
Traceback (most recent call last):
  File "/home/chrispy_/os.py", line 2, in <module>
    playsound('startup.mp3')
  File "/home/chrispy_/.local/lib/python3.7/site-packages/playsound.py", line 92, in _playsoundNix
    gi.require_version('Gst', '1.0')
  File "/usr/lib/python3/dist-packages/gi/__init__.py", line 129, in require_version
    raise ValueError('Namespace %s not available' % namespace)
ValueError: Namespace Gst not available

[dark solar]

would here be the right place to discuss a language based on python using exclusively graphs as input and output?

[rugged sparrow]

@red wasp this is not the place for that question

@dark solar maybe, what is the language?

[dark solar]

It isn't its just a fun idea I had

[snow beacon]

If Python is an important part of it, probably discuss it in... actually I have no familiarity with any other channels on this server. If Python isn't what's important about the language, there are there are other cool servers about esoteric programming languages/esolangs.

[hard spoke]

it's certainly fit for offtopic, not sure about the others

[dusk root]

Do someone have thing that could turn python 3x code to one-line?

[fossil estuary]

I do

[dusk root]

Where I can find it? Github etc.

[fossil estuary]

I'll dm it

[dusk root]

Ok

[tribal moon]

onelinerizer

Pretty sure it works with Python 3x with 1 exception of __builtins__

just put code in it and change __builtins__ to builtins

## YOUR CODE HERE
def f(x):
    return x * 4
y = f(5)
print y``` to this: ```py
(lambda __g, __print: [[(__print(y), None)[1] for __g['y'] in [(f(5))]][0] for __g['f'], f.__name__ in [(lambda x: (lambda __l: [(__l['x'] * 4) for __l['x'] in [(x)]][0])({}), 'f')]][0])(globals(), __import__('builtins', level=0).__dict__['print'])```: http://www.onelinerizer.com/

!e py (lambda __g, __print: [[(__print(y), None)[1] for __g['y'] in [(f(5))]][0] for __g['f'], f.__name__ in [(lambda x: (lambda __l: [(__l['x'] * 4) for __l['x'] in [(x)]][0])({}), 'f')]][0])(globals(), __import__('builtins', level=0).__dict__['print'])

[night quarry]

@tribal moon :white_check_mark: Your eval job has completed with return code 0.

20

[rugged sparrow]

it doesnt support 3x @tribal moon

[tribal moon]

which part

[rugged sparrow]

any syntax added in 3x

[tribal moon]

damn they need to update that

[floral meteor]

;tag; natural numbers

def self(n=0):
    n+=[0,2,print(f"accessing frame: {n+1}"),(f:=lambda*a:None)][1]
    def b(c):
        try:raise Exception("Top level frame, cannot reference.")
        except Exception as e:fr=e.__traceback__.tb_frame;return[{(fr:=getattr(fr,'f_back'))for _ in range(c)},fr.f_code][1]
    return setattr(f,'__code__',b(n))or f

I'm starting to break free from single responsibility principle.

[viscid nymph]

why not just return setattr(f,'__code__',b(n))or f 😩

[floral meteor]

huh true

looks much better

[tired jungle]

are you familiar with the else statement in while and for loops?

[snow beacon]

There's one in try as well.

[floral meteor]

#PEP 3150 "given" status=deferred
class PEP3150_Given_BaseClass:
 def __init__(s,**k):[setattr(s,this,that)for this,that in k.items()]
class PEP3150_Given_MetaClass(type):
 def __init__(cls,name,parents,attrs):
  for i,attr in enumerate(attrs):
    if hasattr(attr,'__call__'):attrs[i]=attr()
  super().__init__(name,parents,attrs)
def _pep3150_example(a=4,b=5,op=lambda x,y:x*y):
 """PEP 3150 example:
op(?.f, ?.g) given bound_a=a, bound_b=b in:
    def f():
        return bound_a + bound_b
    def g():
        return bound_a - bound_b
"""
 @(lambda**k:lambda f:f(**k))(bound_a=a,bound_b=b)
 class given(PEP3150_Given_BaseClass,metaclass=PEP3150_Given_MetaClass):
  def f(s):return s.bound_a + s.bound_b
  def g(s):return s.bound_a - s.bound_b
 return op(given.f,given.g)

I therefore declare PEP 3150 redundant

!pep 3150

[night quarry]

**PEP 3150 - Statement local namespaces (aka "given" clause)**

Link

Status

Deferred

Python-Version

3.4

Created

09-Jul-2010

Type

Standards Track

[floral meteor]

pep 3150 is deferred last i checked :>

[night quarry]

@floral meteor :white_check_mark: Your eval job has completed with return code 0.

69

[viscid nymph]

👀

[floral meteor]

!e ```py
execute=(execute_with:=lambda*a,**k:lambda f:f(*a,**k))()

@int
@execute
class n:int=lambda _:69

@execute
class cout:lshift=print

@execute_with(n)
def output(n):
return(n//10)10((n%10)-2)

cout << output

[night quarry]

@floral meteor :white_check_mark: Your eval job has completed with return code 0.

420

[floral meteor]

@viscid nymph you like it?

[viscid nymph]

Why is n a a class though lol

[floral meteor]

it's an int declaration

@int makes an int object
@execute instantiates the following class
class n actually assigns whatever the result of literally everything is to n
__int__ tells int how to interpret it
lambda _: 69 the actual data that ends up being assigned to n. this will be coerced into an integer.

@execute instantiates
class cout: __lshift__=print very basic c++ iostreams

@execute_with(n) takes the 69 from the class n and executes the function with that as input
def output(n): assign the result of the function to output ;)
return(n//10*10*((n%10)-2) convert 69 to 420

cout << output
call the lshift method of the instantiated cout object with the output

[viscid nymph]

Might as well def n(): return "69"

[floral meteor]

well, no. then i'd have to do

n = int(n())

why do that, when i can instead use int as a decorator, and make a decorator to handling calling n?

like i could do

@int
@execute
def n:return'69'

but

@int
@execute
class n:__int__=lambda _:69

fucks your brains more

!e ```py

def a():
def b():
def c():
def d():
def e(f): f()
@a()()()
def g():print (69)

[night quarry]

@floral meteor :x: Your eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 6, in <module>
003 | TypeError: 'NoneType' object is not callable

[floral meteor]

huh

[viscid nymph]

Return

Or something

But yeah a doesn't return anything

[floral meteor]

XD

!e ```py
a=lambda:lambda:lambda:lambda:lambda:lambda f: f()
@a()()()()()
def b():print (69)

[night quarry]

@floral meteor :white_check_mark: Your eval job has completed with return code 0.

69

[floral meteor]

!e ```py
from _sitebuiltins import Quitter as Exit
@lambda f:f()
class exit:
str=lambda s:"Now I am become Dev, Helloer of Worlds!"
repr=lambda s:Exit(*'ab')(69)

print ( exit )
print ([exit])
print ('this wont print')

[night quarry]

@floral meteor :x: Your eval job has completed with return code 69.

Now I am become Dev, Helloer of Worlds!

[snow beacon]

Is that joke from something? Because I would very much like to steal it.

[floral meteor]

It's a nerd chuck-off at the quote

Now I am become Death, Destroyer of Worlds!

[snow beacon]

I know, but I wondered if the variation was a thing before you said it.

[floral meteor]

probs not

it is now.

XD

I'll just make a compact version cos OCD

@lambda f:f()
class exit:__str__,__repr__=lambda s:"Now I am become Dev, Helloer of Worlds!",lambda s:__import__("_sitebuiltins").Quitter(*'ab')(69)
print ( exit );print ([exit]);print ('this wont print')

feel free to steal

you see, the __str__ method of a list calls the __repr__ methods of all its elements.

[night quarry]

@floral meteor :white_check_mark: Your eval job has completed with return code 0.

~ ~ ~ ~ ~  ~ ~ ~ ~ ~!

[floral meteor]

lmfao

it doesn't error :/

but whytf is it squiggles?

[night quarry]

@floral meteor :white_check_mark: Your eval job has completed with return code 0.

~~~~~~~~~~~~[, , , , , , , , , , , ]

[floral meteor]

lmfao thats even worse

I have created an abomination

[snow beacon]

It's because it only resolves this after the loop.

[floral meteor]

interesting, cos it still assigns it to every printable ascii character in globals

[snow beacon]

Yes, but it's the same function each time.

There's no closure because it's in the global scope.

chr(126) is '~' by the way.

[floral meteor]

ik

[snow beacon]

Try printing this after the update.

[floral meteor]

but it creates and instantiates the type with the current iteration of this

[snow beacon]

There are two references to this.

[viscid nymph]

Me wondering where this was defined

[snow beacon]

One is in the dictionary key, resolved immediately. The other in the lambda, resolved when you print.

[floral meteor]

but this is not defined

[snow beacon]

Interesting, but the point remains.

[viscid nymph]

Wait nvm I'm just dumb

[snow beacon]

(Testing can go in #bot-commands by the way.)

[floral meteor]

!e alright it works now ```py
globals().update({chr(this):(lambda that:type("",(),{'repr':lambda s:print (end=chr(that))or""})())(this)for this in range(1,127)})
str([H,e,l,l,o,globals()[' '],W,o,r,l,d,globals()['!']])

[night quarry]

@floral meteor :white_check_mark: Your eval job has completed with return code 0.

Hello World!

[floral meteor]

XD

if only changes to globals affected literals.

like numbers

[floral meteor]

!e @rugged sparrow I has figurez out ze ctypes! I can now edit ellipsis!

__import__('ctypes').py_object.from_address(id(...)+8).value = type("Ellipsis",(),{'__repr__':lambda s:print("Hello World!")or __import__('_sitebuiltins').Quitter(*'ab')(139)})

str([...])

[night quarry]

@floral meteor :x: Your eval job has completed with return code 139 (SIGSEGV).

Hello World!

[floral meteor]

wait that means i can edit globals dict

[rugged sparrow]

@floral meteor yea thats what i did with my init_missing_hook

[floral meteor]

I'm struggling to edit globals tho

see #bot-commands for some cursed results

!e ```py
import("ctypes").py_object.from_address(id(globals())+8).value=type("dict",(dict,),{"missing":lambda _,key:lambda n:n})
builtins.import('sys').stdout.write(n)

[night quarry]

@floral meteor :x: Your eval job has completed with return code 139 (SIGSEGV).

001 | Traceback (most recent call last):
002 |   File "<string>", line 2, in <module>
003 | AttributeError: 'str' object has no attribute 'stdout'

[floral meteor]

why is the module a string object?

ooh

[rugged sparrow]

__missing__ is called for anything located in __builtins__

[floral meteor]

oh

now what?

[rugged sparrow]

n doesnt exist at all

[floral meteor]

yeah...

oh

!e ```py
import("ctypes").py_object.from_address(id(globals())+8).value=type("dict",(dict,),{"missing":lambda _,key:key})
builtins.print (Hello, World, end='!')

[rugged sparrow]

also messing with the globals dict has some funky requirements to stop segfaulting

[night quarry]

@floral meteor :x: Your eval job has completed with return code 139 (SIGSEGV).

Hello World!

[rugged sparrow]

you need to unhook before the interpreter closes

[floral meteor]

lol

not happening XDDDDDDDDDDDD

[night quarry]

@floral meteor :x: Your eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 50, in <module>
003 |   File "<string>", line 12, in __missing__
004 |   File "<string>", line 41, in parse_int
005 | NameError: name 'one_hundred_and_three_thousand_and_five_hundred_and_forty_two_point_seven' is not defined

[floral meteor]

i borke it already

XD

[rugged sparrow]

no point

[floral meteor]

that's right, there's no point in doing this...

there should be a point to return floats XD

[rugged sparrow]

add it

@floral meteor this is my github repo for weird stuff like this https://github.com/chilaxan/pysnippets

[floral meteor]

I should move my cursedutils module to a repository

I'm lazy how do i do this again?

[rugged sparrow]

github?

[floral meteor]

ye

[rugged sparrow]

google a tutorial lol

[floral meteor]

!e ```py
from ctypes import py_object
import atexit, builtins
def init_missing_hook(dct, func):
ob_base_p = py_object.from_address(id(dct) + 8)
class missing_hook(dict):
slots = ()
def missing(self, key, ob_base_p=ob_base_p, builtins=builtins):
try:
ob_base_p.value = builtins.dict
return (builtins.dict | self)[key]
except KeyError:
return func(self, key)
finally:
ob_base_p.value = class

ob_base_p.value = missing_hook

@atexit.register
def unhook():
    ob_base_p.value = dict
return unhook

shamelessly grabbed from https://ao.gl/how-to-convert-numeric-words-into-numbers-using-python/

def parse_int(dct, textnum, numwords={}):
if not numwords:
units = [
"zero", "one", "two", "three", "four", "five", "six", "seven", "eight",
"nine", "ten", "eleven", "twelve", "thirteen", "fourteen", "fifteen",
"sixteen", "seventeen", "eighteen", "nineteen",
]
tens = ["", "", "twenty", "thirty", "forty", "fifty", "sixty", "seventy", "eighty", "ninety"]
scales = ["hundred", "thousand", "million", "billion", "trillion"]
numwords["and"] = (1, 0)
for idx, word in enumerate(units): numwords[word] = (1, idx)
for idx, word in enumerate(tens): numwords[word] = (1, idx * 10)
for idx, word in enumerate(scales): numwords[word] = (10 ** (idx * 3 or 2), 0)

current = result = 0
for word in textnum.replace("_"," ").split():
    if word not in numwords:
      raise NameError(f"name {textnum!r} is not defined") from None
    scale, increment = numwords[word]
    current = current * scale + increment
    if scale > 100:
        result += current
        current = 0
return result + current

unhook = init_missing_hook(globals(), parse_int)
print(one_hundred_and_three_thousand_and_five_hundred_and_forty_two)

[night quarry]

@floral meteor :white_check_mark: Your eval job has completed with return code 0.

103542

[floral meteor]

!e print(import('sys').version)

[night quarry]

@floral meteor :white_check_mark: Your eval job has completed with return code 0.

001 | 3.9.2 (default, Mar 12 2021, 19:04:51) 
002 | [GCC 8.3.0]

[dusk root]

Sorry for very late answer but it already found this and it is what I want but it doesn't work with Python 3x

[broken mesa]

one of my project ideas is recreating that for python 3.x, most likely 3.9 because then it can be a subclass of the ast unparser.

[dusk root]

interesting

[broken mesa]

alas, i am too busy with too many other things and projects

the ast unparser is quite neat tho, for this type of thing. source: https://github.com/python/cpython/blob/3.9/Lib/ast.py#L670

[viscid nymph]

Me trying to figure out where __class__ is defined

[spring rose]

can we resize pixels in pygame

??????????????????????????????????????????////

[viscid nymph]

not here

[rugged sparrow]

@viscid nymph its a dunder

if the name exists in a classes scope, its assigned to the class

[viscid nymph]

Oh I didn't know you can access class attributes in methods

[rugged sparrow]

its not a class attribute in this case

[viscid nymph]

Oh I see

[rugged sparrow]

__class__ is used by super

[astral rover]

Isn't it a property

[rugged sparrow]

It is, but in this case it is a magic variable filled by the compiler

[fluid tree]

Not really

It's a cellvar (closure) from the class scope