#esoteric-python | Python | Page 76

grave rover Oct 14, 2019, 2:54 PM

#

if function.__new__ is what creates and returns a function
what would happen if I were to replace it

grizzled cloak Oct 14, 2019, 2:54 PM

#

curse(FunctionType, "__new__", my_func)

grave rover Oct 14, 2019, 2:54 PM

#

FunctionType is just a utility for instance checks or creating functions, not the actual type

#

try type(lambda: 0)

grizzled cloak Oct 14, 2019, 2:56 PM

#

On mobile

grave rover Oct 14, 2019, 2:56 PM

#

actually

#

it wont even work I think

sick hound Oct 14, 2019, 2:58 PM

#

uh, no, it is the actual type

#

>>> import types
>>> types.FunctionType is type(lambda: 0)
True```

grave rover Oct 14, 2019, 2:58 PM

#

huh

#

too bad we need to actually curse type to modify instance behavior

grizzled cloak Oct 14, 2019, 2:59 PM

#

But thats ok right?

grave rover Oct 14, 2019, 3:00 PM

#

you can't modify type because those functions are handled in C

grizzled cloak Oct 14, 2019, 3:01 PM

#

Whenever I curse I do something like this:

i_add = int.__add__
def new_add(...):
    if (...):
        return ...
    return i_add(...)
curse...```

grave rover Oct 14, 2019, 3:01 PM

#

I guess we could try using import hooks

#

get all attributes

#

find all instances of functions

#

and patch them

#

problems will happen with nested functions but I'll figure it out

rugged sparrow Oct 14, 2019, 3:16 PM

#

@grave rover its possible to change it with ctypes. it just takes some effort

marsh void Oct 14, 2019, 3:28 PM

#

*and a bit of segfaults on failing

rugged sparrow Oct 14, 2019, 3:30 PM

#

well yea

grave rover Oct 14, 2019, 3:33 PM

#

Ctypes won't fix it because you'd override the default wouldn't you

#

Also, there's no way to access a class after bytecode was generated but before the creation of the class object, right

rugged sparrow Oct 14, 2019, 3:35 PM

#

well ik juanita wrote a way to mod int.__add__ by modifying the func pointer

formal sandal Oct 14, 2019, 3:46 PM

#

@pure dew
Actually

#

This is valid python syntax, right?
Pair<int, int>(5, 6)

sick hound Oct 14, 2019, 3:57 PM

#

yes, although you might get a bit of a problem with int>(5, 6)

rugged sparrow Oct 14, 2019, 4:03 PM

#

>>> int = type('',(int,),{'__gt__':lambda self, other:print(other) if type(other) == tuple else super().__gt__(self, other)})()
>>> int > (5,6)
(5, 6)
>>>``` @sick hound could do this

#

or something like that

pure dew Oct 14, 2019, 4:10 PM

#

@formal sandal yea its a valid Ast, but a little messy

grave rover Oct 14, 2019, 4:32 PM

#

@rugged sparrow yeah but now try (int, int) > (5, 6)

rugged sparrow Oct 14, 2019, 4:35 PM

#

had to mod it a bit

#

still passes tho

grave rover Oct 14, 2019, 4:36 PM

#

so what does MyClass<(int, int)>(5, 6) return

#

because we need to call something on MyClass with int, int and then call the result of that with 5, 6

rugged sparrow Oct 14, 2019, 4:38 PM

#

so itll evaluate ret = MyClass.__gt__(self, (int, int)) then ret.__gt__(self, (5,6))

grave rover Oct 14, 2019, 4:38 PM

#

except it wont

#

a < b < c is shorthand for a < b && b < c

rugged sparrow Oct 14, 2019, 4:40 PM

#

youre doing a < b > c

grave rover Oct 14, 2019, 4:41 PM

#

same thing

#

a < b && b > c

rugged sparrow Oct 14, 2019, 4:42 PM

#

>>> class MyClass():
...     def __lt__(self , other):
...             return (self, other)
...
>>> MyClass = MyClass()
>>> MyClass<(int, int)>(5, 6)
False``` i at least got it to run

#

oh wait i had int overwritten

#

hmm

grizzled cloak Oct 14, 2019, 4:44 PM

#

@grave rover what about writing a __main__ wrapper

grave rover Oct 14, 2019, 4:44 PM

#

wdym

grizzled cloak Oct 14, 2019, 4:45 PM

#

So you call your file like this:
python mywraper file.py

#

And you do some find all def and add your decorator

grave rover Oct 14, 2019, 4:46 PM

#

it'd be impossible to trigger file.py:__main__

grizzled cloak Oct 14, 2019, 4:46 PM

#

?

pure dew Oct 14, 2019, 4:47 PM

#

set __name__ in your wrapper then

grave rover Oct 14, 2019, 4:47 PM

#

how are you gonna make if __name__ == '__main__': trigger

#

not possible

pure dew Oct 14, 2019, 4:47 PM

#

not?

grizzled cloak Oct 14, 2019, 4:47 PM

#

You set name

pure dew Oct 14, 2019, 4:47 PM

#

even by messing with runpy or w/e?

grizzled cloak Oct 14, 2019, 4:47 PM

#

exec takes a env dict

#

You can set name etc in that

grave rover Oct 14, 2019, 4:48 PM

#

__name__ == '__main__'
import file.py

pure dew Oct 14, 2019, 4:48 PM

#

nah nah not like that

grave rover Oct 14, 2019, 4:48 PM

#

wait you're planning to actually just exec it

#

I'm not touching that shit

grizzled cloak Oct 14, 2019, 4:48 PM

#

No

#

That how things like cprofile etc works

#

Exec isnt bad if you know what you are doing

#

Or dont care hahah

formal sandal Oct 14, 2019, 4:57 PM

#

@pure dew I though #esoteric-python was all about being messy!

pure dew Oct 14, 2019, 4:58 PM

#

well yea

#

its just more work to parse properly compared to Pair<(int, int)>(5, 6)

formal sandal Oct 14, 2019, 5:00 PM

#

Hm, this is better and still valid

pure dew Oct 14, 2019, 5:32 PM

#

lets do it

snow beacon Oct 14, 2019, 7:04 PM

#

You can pass exec a globals dict

rugged sparrow Oct 15, 2019, 1:00 AM

#

i just installed 3.8

#

we can now set local vars in oneliners easily

#

with the walrus operator

pure dew Oct 15, 2019, 1:19 AM

#

ive got 3.9 installed already

rugged sparrow Oct 15, 2019, 1:23 AM

#

any major stuff?

pure dew Oct 15, 2019, 1:27 AM

#

havent messed with it lol

#

just happened to see it was available on pyenv

rugged sparrow Oct 15, 2019, 1:28 AM

#

ah nice

#

i was installing on a chromebook in dev mode

#

so i may have bugs

#

¯_(ツ)_/¯

pine wagon Oct 15, 2019, 5:45 AM

#

d = {}
d[1] = d

print(d)

>>> {1: {...}}

#

Had no idea you could do this

#

idk if this is esoteric but it is interesting

crystal mica Oct 15, 2019, 5:47 AM

#

That's simply assigning value

pine wagon Oct 15, 2019, 5:47 AM

#

Yeah, I had no idea you could assign a dictionary value to itself

sick hound Oct 15, 2019, 5:51 AM

#

well you can also make a tuple that contains itself

#

>>> t
((...),)
>>> t[0] is t
True```

crystal mica Oct 15, 2019, 5:52 AM

#

You are not assigning value, you are assigning references

#

so yes it can contain references to itself

pine wagon Oct 15, 2019, 5:55 AM

#

Cool

gilded orchid Oct 15, 2019, 7:20 AM

#

You can now do this terrible thing instead of for loops


n=10

while(n:=n-1):
 print(n)

#

Might be useful for golfing, since while(n:=n-1) is shorter than for n in range(1,10,-1)

#


n=0

while (n:=n+1)%10:
 print(n)

You can also do this

marsh void Oct 15, 2019, 9:00 AM

#

Actually epic

#

I gotta install 3.8 today

snow beacon Oct 15, 2019, 9:09 AM

#

I'm a bit of a stranger to Ubuntu, but I couldn't seem to install it with apt

grave rover Oct 15, 2019, 9:09 AM

#

Ubuntu only updates repositories like twice a year

#

Hence why I use Arch

#

Though I think you could get it working with pyenv

snow beacon Oct 15, 2019, 9:10 AM

#

Ah. I guess I'll go the tarball route.

grave rover Oct 15, 2019, 9:13 AM

#

Oof

#

Meanwhile I just do pacman -Sy python or pyenv install 3.9.0

grizzled cloak Oct 15, 2019, 9:44 AM

#

@snow beacon you can add the repo manually

#

i think its called deadsneks or something

#

yeah
sudo add-apt-repository ppa:deadsnakes/ppa

#

sudo apt-get install python3.8

zealous widget Oct 15, 2019, 9:50 AM

#

i was trying to install 3.8 some weeks ago but it wouldn't detect openssl

whole kiln Oct 15, 2019, 9:51 AM

#

I think it should automatically build with it if you just install openssl

zealous widget Oct 15, 2019, 9:52 AM

#

you'd think

whole kiln Oct 15, 2019, 9:52 AM

#

Unless you're on macos which requires some build env vars iirc

zealous widget Oct 15, 2019, 9:52 AM

#

ubuntu

grave rover Oct 15, 2019, 10:16 AM

#

I wonder if 3.8 and 3.9 have stuff that we could use to make shit even more broken

zealous widget Oct 15, 2019, 10:17 AM

#

might just go back to arch, i only moved to ubuntu because the wireless drivers arch was using were wonky --- but it was the better distro

grave rover Oct 15, 2019, 10:19 AM

#

No wireless problems on Arch for me, just install iw/iwd and it works somehow, make sure to use NetworkManager tho

zealous widget Oct 15, 2019, 10:20 AM

#

i think it's because of the age of my machine, the drivers it used when booting from usb were fine

grave rover Oct 15, 2019, 10:20 AM

#

If it works from USB it should work from the device itself

zealous widget Oct 15, 2019, 10:20 AM

#

you'd think

grave rover Oct 15, 2019, 10:21 AM

#

Try listing all packages installed on the USB

grizzled cloak Oct 15, 2019, 10:32 AM

#

How old is you pc?

zealous widget Oct 15, 2019, 10:36 AM

#

almost old enough to have its own discord account

grizzled cloak Oct 15, 2019, 10:39 AM

#

Haha

#

U ever think of upgrading

#

a lot of our projects seem to be vizualizing stuff so id imagine some amount of power is needed? Or was it powerfull when you bought it?

zealous widget Oct 15, 2019, 11:05 AM

#

i'll upgrade when it breaks

rugged sparrow Oct 15, 2019, 1:52 PM

#

@marsh void 3.8 breaks our new method implementation

marsh void Oct 15, 2019, 2:23 PM

#

omo

#

@rugged sparrow this is pretty sad actually

rugged sparrow Oct 15, 2019, 2:23 PM

#

Yeah

marsh void Oct 15, 2019, 2:24 PM

#

Maybe there is another way actually?

rugged sparrow Oct 15, 2019, 2:24 PM

#

There def is. Also it could be my installation specifically, I had to compile on a chrome book

#

but offset of 16 isnt the dict anymore

📎 rP2e6hpVGWgAAAAASUVORK5CYII.png

marsh void Oct 15, 2019, 2:51 PM

#

🤔

grizzled cloak Oct 15, 2019, 2:51 PM

#

salt, grats on helper!

marsh void Oct 15, 2019, 2:52 PM

#

Hm

#

I am gonna get py3.8 today and will have a look

zealous widget Oct 15, 2019, 2:54 PM

#

thanks!

rugged sparrow Oct 15, 2019, 3:33 PM

#

@marsh void the assignment expressions mean stuff like ```py

(lambda:(x := int(input()),x**x,)[1])()
10
10000000000

#

hurrah for setting locals

marsh void Oct 15, 2019, 3:33 PM

#

haha

#

epic

rugged sparrow Oct 15, 2019, 3:34 PM

#

>>> (lambda y=[0]:(y[0] := 10,y))()
  File "<stdin>", line 1
SyntaxError: cannot use named assignment with subscript
>>> ``` sadly sub-scripting doesn't work :/

marsh void Oct 15, 2019, 3:34 PM

#

Could you do some for loop for a check?

#

on the new __dict__ address

rugged sparrow Oct 15, 2019, 3:34 PM

#

it segfaults

#

soooo

marsh void Oct 15, 2019, 3:34 PM

#

by hand

rugged sparrow Oct 15, 2019, 3:35 PM

#

oh

marsh void Oct 15, 2019, 3:35 PM

#

haha

rugged sparrow Oct 15, 2019, 3:35 PM

#

acc i could prob do a bash script

#

gotta finish my class work first tho gimme 15

marsh void Oct 15, 2019, 3:35 PM

#

k

#

try:
    exit()
except SystemExit:
    print(13)``` haha

rugged sparrow Oct 15, 2019, 3:37 PM

#

Lol

gilded orchid Oct 15, 2019, 4:18 PM

#

n=0
while n:=-~n%11:print(n)

can this be golfed anymore? for context this is equivelent to (but is one byte shorter than):

for n in range(1,11):print(n)

marsh void Oct 15, 2019, 4:21 PM

#

print(*range(1,11),sep='\n')

#

🤔

gilded orchid Oct 15, 2019, 4:21 PM

#

oh yeah

#

wow I'm bad at this

marsh void Oct 15, 2019, 4:21 PM

#

haha

#

No one is

#

@rugged sparrow have you figured the number though

#

For me +16 returns NULL

rugged sparrow Oct 15, 2019, 4:35 PM

#

Yeah I think that python handles the methods declared in c differently now

#

So idk

marsh void Oct 15, 2019, 5:04 PM

#

@juanita

formal sandal Oct 15, 2019, 6:56 PM

#

Maybe there is a more concise version?

#

!e

def curry(func, n=None):
    if n is None:
        n = func.__code__.co_argcount
    if n <= 1:
        return func
    else:
        return lambda x: curry(
                            lambda *args: func(x, *args),
                            n - 1)

addc = curry(lambda a, b: a + b)
inc = addc(1)
print(inc(6), inc(7))

night quarryBOT Oct 15, 2019, 6:58 PM

#

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

7 8

tame crow Oct 15, 2019, 7:06 PM

#

@formal sandal Well, you could remove the else 🙂

formal sandal Oct 15, 2019, 7:08 PM

#

I'm not talking about golfing...

#

I mean, conceptually more concise.

#

Maybe something like

def len(x):
    if x is a function:
        if x has __len in its default kwargs => return its value
        else: return number of arguments
    else:
        return the old len

#

Or just make a separate flen function.

tame crow Oct 15, 2019, 8:16 PM

#

Ah sorry, I misunderstood you 🙂 And perhaps I misunderstand that curry function, but isn't it what functools.partial does?

formal sandal Oct 15, 2019, 9:04 PM

#

curry is supposed to transform a function like f(a, b, c) into f(a)(b)(c)

rugged sparrow Oct 16, 2019, 1:33 AM

#

(f:=lambda n:1 if n == 1 else n * f(n-1))

print(f(6)) #720
``` recursive factorial lambda using assignment expressions

snow beacon Oct 16, 2019, 2:45 AM

#

You could have done that with a regular assignment statement

rugged sparrow Oct 16, 2019, 2:55 AM

#

true

#

but that can be contained within another lambda

#

and you can use them in map

#

*for example

snow beacon Oct 16, 2019, 3:48 AM

#

Or you could use a y-combinator like the rest of us peasants and groundlings

opal totem Oct 16, 2019, 4:06 AM

#

https://discordapp.com/channels/267624335836053506/470884583684964352/633542449981030410 😖 What?

#

All I can find are ctypes hacks

snow beacon Oct 16, 2019, 4:09 AM

#

That's all I'd heard of.

grave rover Oct 16, 2019, 1:43 PM

#

@opal totem yup, welcome to #esoteric-python

gilded orchid Oct 16, 2019, 2:26 PM

#

What are some other cool ways to golf some builtins (aside from these)

list(...)
[*...]

math.sqrt(9)
9**.5

grave rover Oct 16, 2019, 2:42 PM

#

I can offer the reverse```py
[f(x) for x in y]
[*map(f, y)]

gilded orchid Oct 16, 2019, 2:44 PM

#

math.floor(n)
n//1

grave rover Oct 16, 2019, 2:45 PM

#

Note: // does not cast to int

gilded orchid Oct 16, 2019, 2:45 PM

#

Oh yeah oops

#

Although they're pretty similar

#

It'd still be shorter even if you put an int() around it due to the import math

sick hound Oct 16, 2019, 3:21 PM

#

just int(n) would work

rugged sparrow Oct 16, 2019, 3:21 PM

#

^

gilded orchid Oct 16, 2019, 3:21 PM

#

oh yeah good point

brisk zenith Oct 16, 2019, 3:53 PM

#

only for positive floats though

#

it rounds up for negatives because it actually truncated the decimal.

sick hound Oct 16, 2019, 4:02 PM

#

oh

#

>>> int(-0.5)
0
>>> int(-3.5)
-3```

gilded orchid Oct 16, 2019, 4:08 PM

#

>>> 1j
1j
>>> 2j
2j

what is j?

brisk zenith Oct 16, 2019, 4:09 PM

#

an imaginary unit

edgy kelp Oct 16, 2019, 4:09 PM

#

imaginary unit

brisk zenith Oct 16, 2019, 4:09 PM

#

wow even the same wording

#

:D

#

basically used to create complex numbers

edgy kelp Oct 16, 2019, 4:11 PM

#

strange that it doesn't accept i which I think is more used by the people that need them in python

brisk zenith Oct 16, 2019, 4:12 PM

#

i believe it's because i is commonly used as an index variable, as in for i in range (...)

fallen heath Oct 16, 2019, 4:13 PM

#

It's very common in those old style for loops.

gilded orchid Oct 16, 2019, 4:16 PM

#

oh ok

#

I was just wondering because I made a thing to test every two character combination as python code and see what it evaluates to

#

and I got a bit confused when I saw 5j

brisk zenith Oct 16, 2019, 4:18 PM

#

:D

gilded orchid Oct 16, 2019, 4:19 PM

#

>>> -1j
(-0-1j)

can someone explain this

rugged sparrow Oct 16, 2019, 4:19 PM

#

>>> 1_2_3_4_5_6_7_8_9_0
1234567890``` or this?

marsh void Oct 16, 2019, 4:19 PM

#

Lol

#

chilaxan's one: this is a new syntax in python ~3.5 I believe, it is used to make your numbers more readable

#

100_000 instead of 100000

edgy kelp Oct 16, 2019, 4:20 PM

#

complex numbers are a + bj so I guess when you do -bj, it also shows the a to be clearer

marsh void Oct 16, 2019, 4:21 PM

#

maybe

gilded orchid Oct 16, 2019, 4:22 PM

#

>>> 0o1750
1000

what?

fallen heath Oct 16, 2019, 4:22 PM

#

Octal

edgy kelp Oct 16, 2019, 4:22 PM

#

octal I think

fallen heath Oct 16, 2019, 4:22 PM

#

base 8 numbers

gilded orchid Oct 16, 2019, 4:23 PM

#

oh right

rugged sparrow Oct 16, 2019, 4:23 PM

#

>>> 0x
  File "<stdin>", line 1
SyntaxError: invalid hexadecimal literal
>>> 0o
  File "<stdin>", line 1
SyntaxError: invalid octal literal
>>> 0_
  File "<stdin>", line 1
SyntaxError: invalid decimal literal``` these are new in 3.8 right?

#

the specific errors

edgy kelp Oct 16, 2019, 4:23 PM

#

I'm getting invalid token in 3.7

rugged sparrow Oct 16, 2019, 4:23 PM

#

cool

#

also

#

>>> lambda:1()
<stdin>:1: SyntaxWarning: 'int' object is not callable; perhaps you missed a comma?
<function <lambda> at 0x7ca952f78430>
>>>``` this is neat

fallen heath Oct 16, 2019, 4:24 PM

#

There's also a warning for using is with literals now.

rugged sparrow Oct 16, 2019, 4:25 PM

#

>>> () is ()
<stdin>:1: SyntaxWarning: "is" with a literal. Did you mean "=="?
True
>>> ``` nice

marsh void Oct 16, 2019, 4:28 PM

#

Very epic

#

Now python tells if you're being dumb lol

rugged sparrow Oct 16, 2019, 4:29 PM

#

im curious if the syntax warnings are silenceable

marsh void Oct 16, 2019, 4:31 PM

#

__import__('sys').excepthook = lambda *_: None

#

edgy kelp Oct 16, 2019, 4:32 PM

#

why the kwargs

rugged sparrow Oct 16, 2019, 4:33 PM

#

also the warnings arent handled as exceptions

marsh void Oct 16, 2019, 4:34 PM

#

Oh this is sad

#

xd

#

Ok no kwargs anyway

grizzled cloak Oct 16, 2019, 8:18 PM

#

Anyone wanna golf pythagorean triplets?

sick hound Oct 16, 2019, 8:29 PM

#

Sounds like ~~fun~~ torture on mobile

snow beacon Oct 16, 2019, 8:30 PM

#

Here's a program that completes a Pythagorean triple: lambda*v:sum(a**2for a in v)**.5

#

(I'm not sure this is the objective)

sick hound Oct 16, 2019, 8:31 PM

#

Btw there's an equation for the triplets

snow beacon Oct 16, 2019, 8:32 PM

#

lambda a,b:(a**2+b**2)**.5 is even shorter, actually

#

It calculates c for a and b

brazen geyser Oct 16, 2019, 8:34 PM

#

a*a and b*b to cut off a couple more

marsh void Oct 16, 2019, 8:38 PM

#

Did you mean the triple creator or calculator?

sick hound Oct 16, 2019, 8:41 PM

#

It generates the third term of a Pythagorean triple

#

I was thinking creator

#

I have a non-golfed creator script

#

But then again it's really simple

tame crow Oct 16, 2019, 9:15 PM

#

f=lambda n,m:(m*m-n*n,2*m*n,m*m+n*n)``` Give it two different positive integers in ascending order, generates a triplet

tame crow Oct 16, 2019, 9:39 PM

#

Shorter, but less general:

g=lambda a:(a,c:=a*a//2+1,c-1)```  It doesn't find all triplets, only those where there's a difference of 1 between the longest sides (still infinitely many). Call with an odd number >= 3.

snow beacon Oct 16, 2019, 9:42 PM

#

The brackets aren't necessary

tame crow Oct 16, 2019, 9:43 PM

#

Oh? Sweet, two chars saved 🙂

vestal solstice Oct 16, 2019, 10:05 PM

#

a*a is shorter than a**2

tame crow Oct 16, 2019, 10:06 PM

#

Indeed it is. Edit time.

vestal solstice Oct 16, 2019, 10:08 PM

#

there's no reason for double /

tame crow Oct 16, 2019, 10:08 PM

#

But ew, floats

vestal solstice Oct 16, 2019, 10:08 PM

#

it won;t go to floats

brisk zenith Oct 16, 2019, 10:08 PM

#

it will

vestal solstice Oct 16, 2019, 10:08 PM

#

it's wise

#

ah ok

brisk zenith Oct 16, 2019, 10:08 PM

#

/ always gives a float

#

although

vestal solstice Oct 16, 2019, 10:09 PM

#

then a*a//2+1

brisk zenith Oct 16, 2019, 10:09 PM

#

a*a+1>>1 is perfectly valid

vestal solstice Oct 16, 2019, 10:09 PM

#

less brackets

brisk zenith Oct 16, 2019, 10:09 PM

#

yeah both of those work

vestal solstice Oct 16, 2019, 10:09 PM

#

oh wait

#

no I meant (a*a)//2+1

brisk zenith Oct 16, 2019, 10:09 PM

#

i mean

#

a*a//2+1 would still work too

vestal solstice Oct 16, 2019, 10:10 PM

#

no

brisk zenith Oct 16, 2019, 10:10 PM

#

it would.

vestal solstice Oct 16, 2019, 10:10 PM

#

oh I get it

edgy kelp Oct 16, 2019, 10:10 PM

#

why wouldn't it work?

brisk zenith Oct 16, 2019, 10:10 PM

#

:D

vestal solstice Oct 16, 2019, 10:10 PM

#

I'm retarded

#

I was thinking right to left

#

for no reason

tame crow Oct 16, 2019, 10:11 PM

#

Getting pretty short!

rugged sparrow Oct 17, 2019, 3:20 AM

#

chilaxan@iBP-PC:~$ python3
Python 3.8.0 (default, Oct 16 2019, 20:51:08)
[GCC 7.4.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> x
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
NameError: name 'x' is not defined
>>> x[0] = (x := [0])
>>> x
[[...]]
>>>``` uh this works in 3.8

#

i understand why it works. but it feels broken

pure dew Oct 17, 2019, 3:31 AM

#

hell no

#

ill stick with 3.7 for now

rugged sparrow Oct 17, 2019, 3:33 AM

#

it makes sense why it works

#

¯_(ツ)_/¯

#

x[0] = (globals().update({'x':[0]}),x)[1]``` its the same as this in 3.7

grave rover Oct 17, 2019, 10:28 AM

#

ohno

marsh void Oct 17, 2019, 3:08 PM

#

is try-except one-lineable?

rugged sparrow Oct 17, 2019, 3:09 PM

#

It already was @marsh void, but it might be easier now

marsh void Oct 17, 2019, 3:09 PM

#

I am just wondering how to do it

#

In a non-hacky way

rugged sparrow Oct 17, 2019, 3:10 PM

#

I mean, using a context decorator isn't too hacky

marsh void Oct 17, 2019, 5:02 PM

#

or an old tryExceptRet

grave rover Oct 17, 2019, 11:55 PM

#

with suppress(Exception): ... lul

rugged sparrow Oct 18, 2019, 1:52 AM

#

@marsh void my tryExceptRet uses a context decorator to suppress and gather what exceptions happened

marsh void Oct 18, 2019, 3:28 AM

#

oh didn't see it

#

nice

rugged sparrow Oct 18, 2019, 3:34 AM

#

Yeah. I'm probably gonna rewrite it for 3.8 to use walrus tho

rugged sparrow Oct 18, 2019, 4:22 PM

#

>>> (x:=[]).append(x)
>>> y[0]=(y:=[0])``` these both have the same result of initializing the var, then storing itself within it

sick hound Oct 18, 2019, 4:23 PM

#

that makes sense

rugged sparrow Oct 18, 2019, 4:23 PM

#

yea

#

the second feels like it shouldnt work tho

sick hound Oct 18, 2019, 4:24 PM

#

why not

#

it evaluates (y:=[0])

#

then puts that into y[0]

rugged sparrow Oct 18, 2019, 4:24 PM

#

it makes sense that it works

#

but id feel like python should check that y is subscriptable before getting the value to store

sick hound Oct 18, 2019, 4:25 PM

#

>>> 'a'[0] = print('hello')
hello
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: 'str' object does not support item assignment```

rugged sparrow Oct 18, 2019, 4:25 PM

#

¯_(ツ)_/¯

#

i guess i just assumed it would eval assignments left to right

#

at least for subscriptable items

sick hound Oct 18, 2019, 4:26 PM

#

>>> x[0]=x=[0]
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
NameError: name 'x' is not defined
>>> y=y[0]=[0]
>>> ```

rugged sparrow Oct 18, 2019, 4:26 PM

#

i mean i get why it works

#

guess i thought it translated x[0] = 1 to x.__setitem__(0,1) internally

sick hound Oct 18, 2019, 4:31 PM

#

>>> dis.dis('x[0] = 1')
  1           0 LOAD_CONST               0 (1)
              2 LOAD_NAME                0 (x)
              4 LOAD_CONST               1 (0)
              6 STORE_SUBSCR
              8 LOAD_CONST               2 (None)
             10 RETURN_VALUE```

rugged sparrow Oct 18, 2019, 4:31 PM

#

huh

#

cool

sick hound Oct 18, 2019, 4:33 PM

#

also x.__setitem__(0,1) wouldn't make sense

#

>>> test[0] = 1
Real setitem 0 1
>>> test.__setitem__(0, 1)
Fake setitem 0 1```

rugged sparrow Oct 18, 2019, 4:36 PM

#

never really thought to test

#

¯_(ツ)_/¯

hollow patrol Oct 19, 2019, 1:56 PM

#

Here's a little challenge idea

#

Find a way to change the __class__ attribute of an object, if you can

marsh void Oct 19, 2019, 2:36 PM

#

We need to find __dict__ of a class which is non-mappingproxy first

#

>>> s = __import__('ctypes').py_object.from_address(id(object)+16).value
>>> s['__class__'] = None
>>> s
segmentation fault``` @hollow patrol ![FlamesLUL](https://cdn.discordapp.com/emojis/409236578725134348.webp?size=128 "FlamesLUL")

hollow patrol Oct 19, 2019, 2:46 PM

#

@marsh void Let's go one farther

#

Actually no

#

Okay, you did that

#

Somehow you actually did that

#

Good job

marsh void Oct 19, 2019, 2:46 PM

#

lol

#

but it segfaulted so I guess I didn't

hollow patrol Oct 19, 2019, 2:47 PM

#

Well

marsh void Oct 19, 2019, 2:47 PM

#

And it is an old method, not going to work on 3.8

hollow patrol Oct 19, 2019, 2:47 PM

#

oh

rugged sparrow Oct 19, 2019, 2:47 PM

#

@hollow patrol you can in <3.8

hollow patrol Oct 19, 2019, 2:47 PM

#

but not 3.8?

marsh void Oct 19, 2019, 2:47 PM

#

btw chilaxan have you figured out the 3.8 one?

#

because they changed the address

hollow patrol Oct 19, 2019, 2:47 PM

#

alright, find out how to do it in 3.8 🙂

sick hound Oct 19, 2019, 2:47 PM

#

>>> __import__('ctypes').py_object.from_address(id(object)+8).value = int
>>> object
0```

rugged sparrow Oct 19, 2019, 2:47 PM

#

Offset of sizeof(c_size_t)

marsh void Oct 19, 2019, 2:48 PM

#

oh

sick hound Oct 19, 2019, 2:48 PM

#

our method works in 3.8 too

rugged sparrow Oct 19, 2019, 2:48 PM

#

Oh yeah

marsh void Oct 19, 2019, 2:48 PM

#

oh nice

hollow patrol Oct 19, 2019, 2:48 PM

#

huh

marsh void Oct 19, 2019, 2:48 PM

#

epic

hollow patrol Oct 19, 2019, 2:48 PM

#

neat

rugged sparrow Oct 19, 2019, 2:48 PM

#

Just not for mappingproxies

#

@sick hound

marsh void Oct 19, 2019, 2:49 PM

#

haha

sick hound Oct 19, 2019, 2:49 PM

#

what

rugged sparrow Oct 19, 2019, 2:49 PM

#

Mappingproxies are different in memory now

#

Cant add methods the same way yet

hollow patrol Oct 19, 2019, 2:49 PM

#

What module is the sizeof from

rugged sparrow Oct 19, 2019, 2:49 PM

#

Ctypes

sick hound Oct 19, 2019, 2:50 PM

#

huh

#

ok apparently it doesn't work for mappingproxies

hollow patrol Oct 19, 2019, 2:50 PM

#

Alright

rugged sparrow Oct 19, 2019, 2:50 PM

#

I think they were optimized in 3.8

hollow patrol Oct 19, 2019, 2:51 PM

#

Also chilaxan I rewrote anony

rugged sparrow Oct 19, 2019, 2:51 PM

#

Nice

hollow patrol Oct 19, 2019, 2:51 PM

#

This time now it's much better

#

And much more flexible

#

And also much more readable

#

You can create custom statements by overriding anony's statement class

#

So next

#

Create some sort of metaclass or something

#

Which allows any instance of any class of this metaclass to have an overridable class method

sick hound Oct 19, 2019, 2:55 PM

#

hang on what ```py

cast(id(object.dict), py_object)
py_object(<method-wrapper 'repr' of py_object object at 0x000001B9F8602CC0>)```

rugged sparrow Oct 19, 2019, 2:57 PM

#

Yup

#

Something funky is going on behind the scenes

#

@sick hound the weird thing is, I saw the same behavior on a rpi3 with python 3.5 yesterday

wind maple Oct 19, 2019, 3:45 PM

#

I feel like the repr just happens to be at the same location as the dict start

#

by chance

rugged sparrow Oct 19, 2019, 3:45 PM

#

its not actually the repr

#

if you try to store a ref to it the interpreter segfaults

marsh void Oct 19, 2019, 3:51 PM

#

so uhm

#

I have a solution but it is meh

#

[obj for obj in gc.get_objects() if obj == dict(cls.__dict__)][0]

#

@rugged sparrow @wind maple here is some terrible solution which doesn’t include ctypes at all

rugged sparrow Oct 19, 2019, 3:55 PM

#

hmm

#

that works

marsh void Oct 19, 2019, 3:55 PM

#

Let’s improve this one then

rugged sparrow Oct 19, 2019, 3:55 PM

#

should use is

wind maple Oct 19, 2019, 3:55 PM

#

solution for what

sick hound Oct 19, 2019, 3:55 PM

#

no

#

if you use is it won't work

marsh void Oct 19, 2019, 3:56 PM

#

Yeah because

sick hound Oct 19, 2019, 3:56 PM

#

dict(...) creates a new dict

marsh void Oct 19, 2019, 3:56 PM

#

Different ID

rugged sparrow Oct 19, 2019, 3:56 PM

#

oh right

wind maple Oct 19, 2019, 3:56 PM

#

what are you even trying to do

marsh void Oct 19, 2019, 3:56 PM

#

cls.__dict__ retrieving

#

so it is possible to add new attributes to classes and stuff

wind maple Oct 19, 2019, 3:57 PM

#

oh cpython again

#

this channel should just be renamed to cpython-abuse

#

or a new channel be created

#

😤

marsh void Oct 19, 2019, 3:57 PM

#

@wind maple I did not use ctypes

#

but hm

#

does PyPy and others have GC?

wind maple Oct 19, 2019, 3:58 PM

#

most do

marsh void Oct 19, 2019, 3:59 PM

#

then epic

limber thunder Oct 19, 2019, 4:00 PM

#

im gonna fail the US ones

#

fuck

#

o they filtered

marsh void Oct 19, 2019, 4:01 PM

#

you ok?

rugged sparrow Oct 19, 2019, 4:05 PM

#

getDict = lambda c:next(o for o in __import__('gc').get_objects() if o=={**c.__dict__})```

marsh void Oct 19, 2019, 4:05 PM

#

yeah this one is cool

#

The problem is though

#

Doesn’t our dict creating add to the newer objects?

rugged sparrow Oct 19, 2019, 4:06 PM

#

that dict has one ref

#

which is discarded after the func is ran

sick hound Oct 19, 2019, 4:06 PM

#

the dict is only created after get_objects is called

rugged sparrow Oct 19, 2019, 4:06 PM

#

^ that too

marsh void Oct 19, 2019, 4:08 PM

#

Oh yep

#

It is not adding another one

sick hound Oct 19, 2019, 4:12 PM

#

although there is a problem

#

>>> import gc
>>> class Trick:
...     def __eq__(self, other):
...             return True
...
>>> trick = Trick()
>>> getDict = lambda c:next(o for o in __import__('gc').get_objects() if o=={**c.__dict__})
>>> getDict(object)
<__main__.Trick object at 0x000002196DBC3F98>```

marsh void Oct 19, 2019, 4:32 PM

#

o == {**c.__dict__} and type(o) is dict

#

🤔

#

not using isinstance on purpose I think

sick hound Oct 19, 2019, 4:36 PM

#

>>> import gc
>>> trick = dict(object.__dict__)
>>> getDict = lambda c:next(o for o in __import__('gc').get_objects() if o=={**c.__dict__} and type(o) is dict)
>>> getDict(object)
{'__repr__': <slot wrapper '__repr__' of 'object' objects>, '__hash__': <slot wrapper '__hash__' of 'object' objects>, '__str__': <slot wrapper '__str__' of 'object' objects>, '__getattribute__': <slot wrapper '__getattribute__' of 'object' objects>, '__setattr__': <slot wrapper '__setattr__' of 'object' objects>, '__delattr__': <slot wrapper '__delattr__' of 'object' objects>, '__lt__': <slot wrapper '__lt__' of 'object' objects>, '__le__': <slot wrapper '__le__' of 'object' objects>, '__eq__': <slot wrapper '__eq__' of 'object' objects>, '__ne__': <slot wrapper '__ne__' of 'object' objects>, '__gt__': <slot wrapper '__gt__' of 'object' objects>, '__ge__': <slot wrapper '__ge__' of 'object' objects>, '__init__': <slot wrapper '__init__' of 'object' objects>, '__new__': <built-in method __new__ of type object at 0x000000006DF7C5C0>, '__reduce_ex__': <method '__reduce_ex__' of 'object' objects>, '__reduce__': <method '__reduce__' of 'object' objects>, '__subclasshook__': <method '__subclasshook__' of 'object' objects>, '__init_subclass__': <method '__init_subclass__' of 'object' objects>, '__format__': <method '__format__' of 'object' objects>, '__sizeof__': <method '__sizeof__' of 'object' objects>, '__dir__': <method '__dir__' of 'object' objects>, '__class__': <attribute '__class__' of 'object' objects>, '__doc__': 'The most base type'}
>>> _ is trick
True```

#

nope

marsh void Oct 19, 2019, 4:37 PM

#

hm

#

Not sure if we can fix that

#

Actually wait

#

oh no xd

marsh void Oct 19, 2019, 5:19 PM

#

def get_dict(cls):
    getDict(cls)['__test__'] = None
    actual = getDict(cls)
    actual.pop('__test__', None)
    return actual``` @sick hound

#

two calls but no tricks

gilded orchid Oct 19, 2019, 5:54 PM

#

>>> 0xfor 0
15

🤔

#

oh wow syntax highlighting ruins that immediately

#

nice

marsh void Oct 19, 2019, 6:07 PM

#

lol

sick hound Oct 19, 2019, 7:28 PM

#

>>> 0xfor 0
15```

brisk zenith Oct 19, 2019, 8:22 PM

#

took me a moment but it makes sense

grizzled cloak Oct 19, 2019, 8:25 PM

#

the syntax highlighting did ruin the surprise for me though. still didnt understand why it picked 15 over 0 until i remembered 0 is false...

snow beacon Oct 19, 2019, 8:26 PM

#

The 0 shouldn't matter

#

The one at the end, I mean.

grizzled cloak Oct 19, 2019, 8:27 PM

#

oh i guess it always picks the first one if they both are True

snow beacon Oct 19, 2019, 8:29 PM

#

[0xfor i in range(10)]

formal sandal Oct 19, 2019, 8:55 PM

#

I have made an atrocious thing

#

!e

class run_on_in:
    def __init__(self, func, *args):
        self.func = func
        self.range = range(*args)
        
    def __contains__(self, x):
        for i in self.range:
            self.func(i)

def Range(*args):
    def inner(func):
        return run_on_in(func, *args)
    return inner
    

_ = object()

(0xfor _) in Range(5)(
    print
)```

night quarryBOT Oct 19, 2019, 8:56 PM

#

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

formal sandal Oct 19, 2019, 9:06 PM

#

By the way... Did you know that there is a Category Theory module in sympy?

#

spends the rest of his life in front of a computer with the latest version of sympy

grave rover Oct 20, 2019, 8:10 PM

#

do we know of a way to replace a method-wrapper function with our own function? (typically in a mappingproxy)

sick hound Oct 20, 2019, 8:55 PM

#

well you can make ints callable ```py

from ctypes import *
def new_call(self, args, kwargs):
... print('hello')
... return self
...
(c_void_p*52).from_address(id(int))[16] = cast(CFUNCTYPE(c_void_p, c_void_p, c_void_p, c_void_p)(new_call), c_void_p)
5()
<stdin>:1: SyntaxWarning: 'int' object is not callable; perhaps you missed a comma?
hello
5```

pure dew Oct 20, 2019, 8:59 PM

#

syntaxwarnings are gonna be the bane of #esoteric-python

#

yes i know this is horrific and traumatizing, let me do it anyway

sick hound Oct 20, 2019, 9:04 PM

#

well in that case, #esoteric-python's next project is to find out how to disable syntax warnings

pure dew Oct 20, 2019, 9:05 PM

#

use 3.7

#

easy

edgy kelp Oct 20, 2019, 9:05 PM

#

just hide all warnings and exceptions

#

if it doesn't crash you can push it farther

sick hound Oct 20, 2019, 9:09 PM

#

-W ignore

#

works

marsh void Oct 20, 2019, 10:08 PM

#

lol

#

very epic

#

How terrible was my truck patch in the new get_dict function though?

#

Does someone know how to override any of the dunder methods though?

gilded orchid Oct 21, 2019, 7:15 AM

#

You can do a stop warnings from printing using warnings.filterwarnings I think

#

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

#

Wait nvm, you can't disable syntax warnings with that

distant wave Oct 21, 2019, 8:05 AM

#

!e ```python

instr = "Mary had a little lamb"

output = list(instr)
output[::2], output[1::2] = instr[::2], instr[1::2][::-1]
print(''.join(output))

night quarryBOT Oct 21, 2019, 8:05 AM

#

@distant wave Your eval job has completed with return code 0.

Mbra  al t liatdehlyma

distant wave Oct 21, 2019, 8:08 AM

#

I wonder

#

!e ```python
a=lambda x,y=[]:(y:=list(x),(y[::2],y[1::2]:=x[::2],x[1::2][::-1]),''.join(y))[2]
print(a("Mary had a little lamb"))

night quarryBOT Oct 21, 2019, 8:08 AM

#

@distant wave Your eval job has completed with return code 1.

001 | File "<string>", line 1
002 |     a=lambda x,y=[]:(y:=list(x),(y[::2],y[1::2]:=x[::2],x[1::2][::-1]),''.join(y))[2]
003 |                       ^
004 | SyntaxError: invalid syntax

distant wave Oct 21, 2019, 8:09 AM

#

No walrus support

#

F

whole kiln Oct 21, 2019, 8:26 AM

#

I'll try to get on that soon

#

But touching the ci monstrosity I wrote gives me a headache

edgy kelp Oct 21, 2019, 8:33 AM

#

Open an issue and get someone else to do it :P

whole kiln Oct 21, 2019, 8:36 AM

#

That sounds like a worse idea lol

#

Well

#

Just bumping the version is simple

#

But

crystal mica Oct 21, 2019, 8:36 AM

#

It's snekbox rite?

whole kiln Oct 21, 2019, 8:36 AM

#

The ci is a little broken, sort of not really

#

So I wanna fix it

#

But want to perhaps redo it...

crystal mica Oct 21, 2019, 8:37 AM

#

I wanna watch, it sounds fun

whole kiln Oct 21, 2019, 8:37 AM

#

Yeah its senkbox

#

Tbh you could upgrade and it will probably all work

#

I don't think there's a need to revamp the in conjunction

#

Can be done separately

gilded orchid Oct 21, 2019, 2:27 PM

#

Just discovered that {'a', 'b'}.pop() randomly picks 'a' or 'b'

#

But it doesn't work with ints for some reason

formal sandal Oct 21, 2019, 2:40 PM

#

Does it work with large ints?

#

Like 25000

#

Hm, that's not very large

#

I guess it's just that ints have constant hash values.

rugged sparrow Oct 21, 2019, 3:01 PM

#

>>> {'a','b'}.pop()
'b'
>>> {'a','b'}.pop()
'b'
>>> {'a','b'}.pop()
'b'
>>> {'a','b'}.pop()
'b'
>>> {'a','b'}.pop()
'b'
>>> {'a','b'}.pop()
'b'
>>> {'a','b'}.pop()
'b'
>>> {'a','b'}.pop()
'b'
>>> {'a','b'}.pop()
'b'
>>> ``` @gilded orchid seems to be fixed in 3.8

gilded orchid Oct 21, 2019, 3:02 PM

#

Oh wow

#

Weird

rugged sparrow Oct 21, 2019, 3:02 PM

#

was prob a bug with how sets handle .pop

gilded orchid Oct 21, 2019, 3:02 PM

#

Yeah, since sets are meant to be unordered, but pop pops the last item

rugged sparrow Oct 21, 2019, 3:03 PM

#

yea just ran it 10k times

#

its always b in python 3.8

formal sandal Oct 21, 2019, 3:04 PM

#

Even if you restart the interpreter?

rugged sparrow Oct 21, 2019, 3:05 PM

#

yea

#

and there doesnt seem to be a pattern if you use other chars

gilded orchid Oct 21, 2019, 3:08 PM

#

Oh wow

#

I really need to update to 3.8

rugged sparrow Oct 21, 2019, 3:10 PM

#

its nice

#

:= walrus operator makes lambdas god-tier

gilded orchid Oct 21, 2019, 3:13 PM

#

(though the warnings are annoying)

rugged sparrow Oct 21, 2019, 3:13 PM

#

alias python3="python3 -W ignore"

gilded orchid Oct 21, 2019, 3:14 PM

#

wow

rugged sparrow Oct 21, 2019, 3:14 PM

#

lmao

marsh void Oct 21, 2019, 3:25 PM

#

true that

#

alias windows=".bat is meh"

formal sandal Oct 21, 2019, 5:17 PM

#

@rugged sparrow
alias python3="python3 -W ignore"

#

Shouldn't it make it infinitely recursive?..

rugged sparrow Oct 21, 2019, 5:17 PM

#

Oh yeah

formal sandal Oct 21, 2019, 5:17 PM

#

Or does it only allow one level?

rugged sparrow Oct 21, 2019, 5:18 PM

#

Honestly idk. I have mine set to alias python

#

Not python3

#

It only allows one level

#

Just tested

formal sandal Oct 21, 2019, 5:23 PM

#

📎 unknown.png

#

Maybe it is actually smart and it detects recursion.

marsh void Oct 21, 2019, 6:02 PM

#

smart lol

rugged sparrow Oct 21, 2019, 6:03 PM

#

probably

marsh void Oct 21, 2019, 6:03 PM

#

class hm:
    def __matmul__(self, other):
        print('why')
sent, everyone = hm(), hm()
sent @everyone```

snow beacon Oct 21, 2019, 6:40 PM

#

Before 3.8, I think sets were meant to pop in orders based on hash values. Strings have nondeterministic hashes.

#

Afterwards, they probably put the and dicts in line with each other.

pure dew Oct 21, 2019, 11:23 PM

#

did you know you can do "enhanced" expansion in loops

#

for (arg, (param_type, _)) in zip(args, self.sig_map):

#

where sig_map is a list of tuples

potent comet Oct 21, 2019, 11:31 PM

#

According to the reference, it's the same gammar type as regular assignments, so you can use anything you could there:

for [obj.attribute(blah)[45], *rest] in myiter:

snow beacon Oct 21, 2019, 11:33 PM

#

Can you do that with parameters?

pure dew Oct 21, 2019, 11:38 PM

#

function parameters?

snow beacon Oct 21, 2019, 11:39 PM

#

Yeah.

#

I think I remember a PEP removing it

pure dew Oct 21, 2019, 11:40 PM

#

do what tho

#

@potent comet thats valid syntax, sure, but it wont run

snow beacon Oct 21, 2019, 11:42 PM

#

Looks like you can't do this: ```python

def foo(a, (b, c)):
File "<stdin>", line 1
def foo(a, (b, c)):
^
SyntaxError: invalid syntax```

pure dew Oct 21, 2019, 11:42 PM

#

oh that?

#

nah

#

its not quite as smart as haskell

snow beacon Oct 21, 2019, 11:42 PM

#

It's a little inconsistent with other forms of assignment, but I understand the reasons.

#

It's for introspection, from memory.

potent comet Oct 21, 2019, 11:44 PM

#

Python 2 allowed unpacking there, it was removed in Python 3 (since it means signatures are really weird).

#

And that code snippet would run, if you have an obj with the method defined.

distant wave Oct 22, 2019, 12:46 AM

#

ooOOOoo

#

Feast your eyes

#

!e ```python
print((lambda x, y=[]:[y.clear(), y.extend(x), list(None for y[::2], y[1::2] in [(x[::2].lower(), x[1::2].upper())]), ''.join(y)][-1])("Mary had a little lamb"))

night quarryBOT Oct 22, 2019, 12:55 AM

#

@distant wave Your eval job has completed with return code 0.

mArY HaD A LiTtLe lAmB

distant wave Oct 22, 2019, 1:04 AM

#

Needs more brackets

rugged sparrow Oct 22, 2019, 1:05 AM

#

weirdCase = lambda s:''.join(map(lambda c:''.join(c[0].lower()+c[1].upper()),zip(s[::2],s[1::2])))

distant wave Oct 22, 2019, 1:05 AM

#

print(lambda x,y=[]:[y.clear(),y.extend(x),[0 for y[::2],y[1::2] in [(x[::2].lower(),x[1::2].upper())]],''.join(y)][-1])("Mary had a little lamb"))

#

!e ```python
sarcasm = lambda p: sum(map(lambda c:c[0]+c[1],zip(p[::2].upper(),p[1::2].lower())),'')
print(sarcasm("who's fleece was white as snow"))

night quarryBOT Oct 22, 2019, 1:08 AM

#

@distant wave Your eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 2, in <module>
003 |   File "<string>", line 1, in <lambda>
004 | TypeError: sum() can't sum strings [use ''.join(seq) instead]

distant wave Oct 22, 2019, 1:08 AM

#

-.-

pure dew Oct 22, 2019, 1:09 AM

#

!e ```py
(lambda s,a=(str.lower,str.upper):''.join(map(lambda p:ap[0]%2,enumerate(s))))("Mary had a little lamb and its code was shorter than all the others")

night quarryBOT Oct 22, 2019, 1:09 AM

#

Sorry, but you may only use this command within #bot-commands.

pure dew Oct 22, 2019, 1:09 AM

#

WHY

#

oh ur a helper

rugged sparrow Oct 22, 2019, 1:12 AM

#

weirdCase = lambda s:''.join(map(''.join,zip(s[::2].lower(),s[1::2].upper())))

distant wave Oct 22, 2019, 1:12 AM

#

lambda s,t=''.join:t(map(t,zip(s[::2].lower(),s[1::2].upper())))

rugged sparrow Oct 22, 2019, 1:13 AM

#

Lemme get to an actual computer and ill try to golf more

distant wave Oct 22, 2019, 1:15 AM

#

hm

#

" "*len(s) feels important

vestal solstice Oct 22, 2019, 1:23 AM

#

from functools import reduce
print(reduce(lambda x, y: x.swapcase()+y,"mary had a little lamb"))

pure dew Oct 22, 2019, 1:32 AM

#

what've i done

>>> Settings.set_command
<bound method ? of <class 'pyskell.repl.commands.Settings'>>

#

how does this hapen

rugged sparrow Oct 22, 2019, 1:33 AM

#

?

#

@pure dew what did you do?

pure dew Oct 22, 2019, 1:34 AM

#

mixing classmethod and custom decorators

#

what have i done..

#

def settings_classmethod_hack(func):
    func_signature = inspect.signature(func)

    def inner_hack(*args):
        return func(Settings, *args)

    inner_hack.__signature__ = func_signature

    return inner_hack


class Settings:
    settings = {}

    @classmethod
    @CommandHandler.command(name="set")
    @settings_classmethod_hack
    def set_command(cls, opt: str, val: bool):
        print(cls)
        cls.settings[opt] = val
        print(cls.settings)

#

and it works

zealous widget Oct 22, 2019, 1:45 AM

#

print(*map(lambda a:(str.upper, str.lower)[a[0]&1](a[1]), enumerate("Mary had a little lamb.")),sep="")

really tried to find a way to unpack this tuple without using partial

distant wave Oct 22, 2019, 2:00 AM

#

@pure dew isn't that what functools.wraps is for?

pure dew Oct 22, 2019, 2:01 AM

#

idk, ive never used it

#

oh

#

i guess so

distant wave Oct 22, 2019, 2:03 AM

#

Not sure if it clones the signature, but it does do name and docs

pure dew Oct 22, 2019, 2:04 AM

#

it does clone the signature

#

but i already handle the name elsewhere

distant wave Oct 22, 2019, 2:04 AM

#

!e ```python
from functools import wraps

def d(a):
@wraps(a)
def b(*args):
return a(*args)
return b

@d
def test(a=1, b=2): pass

print(test.signature)

pure dew Oct 22, 2019, 2:04 AM

#

still have that horrid return func(Settings, *args) tho

distant wave Oct 22, 2019, 2:05 AM

#

Also not handling kwargs

pure dew Oct 22, 2019, 2:05 AM

#

not gonna have kwargs

#

the command function aren't even gonna have vargs

crystal mica Oct 22, 2019, 2:05 AM

#

s='Mary had a little lamb.';print(*map(lambda s:''.join(s), zip(s[::2].upper(),s[1::2].lower())),sep='')```

#

Hmm, shorter, but rip on odd length

pure dew Oct 22, 2019, 2:05 AM

#

its a command dispatcher for a repl

distant wave Oct 22, 2019, 2:05 AM

#

You could do func(*((Settings,)+args))

#

but that looks worse

pure dew Oct 22, 2019, 2:06 AM

#

i could

#

but yea, yikes

distant wave Oct 22, 2019, 2:06 AM

#

I feel like args.insert(0, Settings) makes more sense

#

but if it's a tuple that doesn't work

#

Honestly what you've got seems fine

pure dew Oct 22, 2019, 2:06 AM

#

all that seems like extra work tbh

#

this is another command

@CommandHandler.command(name="quit")
def quit_command():
    print(paint("\nExiting..", "red"))
    exit()

#

that earlier one only needs the classmethod decorator because its a classmethod and my decorator doesn't play nicely with that

distant wave Oct 22, 2019, 2:07 AM

#

Here's mine, which is probably ugly as sin and indev

#


        if event in EVENT_MAPPING:
            if EVENT_MAPPING[event] == "bot":
                preargs = (self.bot,)
            elif EVENT_MAPPING[event] == "context":
                # BAD BAD BAST
                preargs = (models.Context(args[0], args[0].channel, args[0].channel.guild),)
            else:
                preargs = (self.bot,)
        else:
            preargs = (self.bot,)
        [
            asyncio.create_task(coroutine(*preargs, *args, **kwargs))
            for coroutine in self.event_registry[event]
        ]

pure dew Oct 22, 2019, 2:08 AM

#

oh wew

#

mines nowhere near that complex

distant wave Oct 22, 2019, 2:09 AM

#

It's really not that bad, it's just async, and doing the same really

#

coroutine(*preargs, *args, **kwargs) is the same as your func(Settings, *args)

pure dew Oct 22, 2019, 2:09 AM

#

but thats only for commands inside the settings class

distant wave Oct 22, 2019, 2:09 AM

#

I just build preargs dynamically based on whether the handler is a command specific one or not

pure dew Oct 22, 2019, 2:09 AM

#

all the others are just result = command(*new_args)

#

I don't know whether to be proud or sad

#

I don't have any particularly nasty bits of code in here

#

yet

#

its mostly clean

#

I did make a discovery tho

#

if source[0] == ":":
    name, *args = source[1:].split()

    print(handler.dispatch(name, *args))
``` this is allowable syntax, and actually works

distant wave Oct 22, 2019, 2:12 AM

#

yeah, it's pretty nice

pure dew Oct 22, 2019, 2:52 PM

#

@formal sandal Here's my syntax demo so far for the pyskell "compiler"

class ListManipulation:
    getFourth    : str.str
    getFourth(s) | s @ 3

    head        : list(a).a
    head(a, *_) | a

    tail           : list(a).list(a)
    tail(_, *tail) | tail

#

equivalent haskell code:

module ListManipulation where

getFourth :: [Char] -> [Char]
getFourth s = s !! 3

head :: [a] -> a
head x:xs = x

tail :: [a] -> [a]
tail x:xs = xs

#

@grave rover do you prefer this or the monstrosity that was pycss?

pure dew Oct 22, 2019, 3:19 PM

#

and boom, this is a totally valid ast

#

class ListManipulation:
    getFourth    : str.str
    getFourth(s) | s @ 3

    head        : list(a).a
    head(a, *_) | a

    tail           : list(a).list(a)
    tail(_, *tail) | tail


class Main(main):
    from ListManipulation import getFourth, head

    main : IO()
    main | (arr := [0, 1, 2, 3, 4])
    print, head, arr
    print, getFourth, arr

pure dew Oct 22, 2019, 11:30 PM

#

more spam, but here's some custom ast progress https://paste.a-sketchy.site/nucagiwapo.py

snow beacon Oct 22, 2019, 11:59 PM

#

This channels is too quiet for spam be a concern.

pure dew Oct 22, 2019, 11:59 PM

#

true

grizzled cloak Oct 23, 2019, 5:53 AM

#

@pure dew is this just you redifining a buch of dunder methods or is there something really weird going on?

sick hound Oct 23, 2019, 8:13 AM

#

looks like something weird going on

#

it is valid ast, but not valid much else

pure dew Oct 23, 2019, 11:31 AM

#

^

#

im just taking the ast this makes

pure dew Oct 23, 2019, 10:12 PM

#

To make debugging easier, I wrote a little graphing function using pydot

#

Here's the result of graphing the ast from the code in the paste: https://paste.a-sketchy.site/nucagiwapo.py

📎 test.png

#

because i can, ya know

pure dew Oct 24, 2019, 2:39 AM

#

Somebody stop me before I go too far
test : Dict(Char, List(Pair(_, b))).Char.Int.b

📎 test.png

sick hound Oct 24, 2019, 9:02 AM

#

            if res[1] and res[4].startswith(res[5][0]) and not res[6]:
                new_current.append((res[0], res[1], res[2], res[5][0], res[4][len(res[5][0]):], [], res[5][0]))``` is this code esoteric or just bad?

distant wave Oct 24, 2019, 9:04 AM

#

It brings the words "Why not use objects" to mind

#

Dataclasses make it a lot easier

snow beacon Oct 24, 2019, 9:57 AM

#

Oi! Helpful advice belongs in other channels.

#

This is the channel for golfing code until it's unreadable, but maybe that code is leaning too heavily on the 'unreadable' side.

distant wave Oct 24, 2019, 10:00 AM

#

b=bool
nc += [r.m,r.t,r.x,r.s.l,r.i[r.s.l.l:],[],r.s.l] if b(r.t)&b(r.sw(r.s.l))&~b(r.v) else []

#

b=bool
nc += ([r.m,r.t,r.x,r.s.l,r.i[r.s.l.l:],[],r.s.l],[])[b(r.t)&b(r.sw(r.s.l))&~b(r.v)]

#

Hm, it's missing something

#

b=bool
# do not touch
nc += ([r.m,r.t,r.x,r.s.l,r.i[r.s.l.l:],[],r.s.l],[])[b(r.t)&b(r.sw(r.s.l))&~b(r.v)]

#

There

#

!e ```python
nc = []
nc += []
print(nc)
b = bool
print((1, 2)[b(1)&b(0)])

night quarryBOT Oct 24, 2019, 10:05 AM

#

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

001 | []
002 | 1

distant wave Oct 24, 2019, 10:08 AM

#

b=bool
# do not touch
m,t,x,s,i,v=*r.it()
nc+=([m,t,x,s.l,i[s.l.l:],[],s.l],[])[b(t)&b(r.sw(s.l))&~b(v)]

vestal solstice Oct 24, 2019, 1:59 PM

#

a.e.s.t.h.e.t.i.c

wide cypress Oct 25, 2019, 1:37 PM

#

yeah fuck it lets go: longest code to print "hello"

#

most meaningful lines

#

alpha = range(255)
count = (0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
         0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
         0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
         0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
         0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0,
         0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
         0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 2, 0, 0, 1,
         0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
         0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
         0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
         0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
         0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
         0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
         0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
         0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
         0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0)

alpha_dict = {}
alpha_list = []

for letter, count in zip(alpha, count):
    alpha_dict[letter] = count

for letter, count in alpha_dict.items():
    if count:
        alpha_list.append(chr(letter) * count)

print("".join(alpha_list))```

#

@plucky bane that'll be £200k please.

#

okay, £98k if we ignore pep8

sudden osprey Oct 25, 2019, 2:07 PM

#

Nice

plucky bane Oct 25, 2019, 2:24 PM

#

Hello

#

@wide cypress I counter offer you 76p

wide cypress Oct 25, 2019, 2:36 PM

#

fine

#

please send it via cheque

plucky bane Oct 25, 2019, 2:41 PM

#

Sorry I broke my cheque

pure dew Oct 25, 2019, 3:28 PM

#

@wide cypress thats horrific

wide cypress Oct 25, 2019, 3:32 PM

#

thank you, thank you.

marsh void Oct 25, 2019, 3:34 PM

#

lol

marsh void Oct 25, 2019, 7:42 PM

#

>>> @new_method(int)
... @property
... def length(self):
...     return len(str(self))
... 
>>> a = 13
>>> a.length
2
>>> a = 241
>>> a.length
3``` haha, seems fun

gilded orchid Oct 25, 2019, 7:51 PM

#

cool decorator

marsh void Oct 25, 2019, 7:55 PM

#

It takes a while to overwrite with a property though, which is weird

#

Using the new GC method though

grave rover Oct 25, 2019, 11:37 PM

#

new gc method? 👀

rugged sparrow Oct 26, 2019, 4:25 AM

#

@grave rover I assume he means the method of getting a classes dictionary using the garbage collector instead of ctypes

chrome mauve Oct 26, 2019, 5:41 AM

#

😮 i heard ppl golf here

#

me and a few others have been golfing a q from my exam a few days, where we had to convert a string of length n even >= 16 to a string such that the the first n/2 - 5 characters are flipped, the next 10 characters are flipped, and then the last n/2 - 5 characters are flipped

#

our current best is 54 characters

#

e=lambda a:(a[(x:=len(a)//2+5):]+a[-x:x]+a[:-x])[::-1]

vestal solstice Oct 26, 2019, 7:28 AM

#

e=lambda a:(a+a[(x:=len(a)//2-5):-x]+a[:x])[:x+9:-1]

vestal solstice Oct 26, 2019, 8:03 AM

#

e=lambda a:(a+a[-(x:=len(a)//2+5):x]+a)[~x:x-1:-1]

marsh void Oct 26, 2019, 8:28 AM

#

Yeah I meant the method which uses garbage collector to get the class.__dict__ which I can edit. It is slower than ctypes' one, but in theory works on more python implementations

#

Gotta install some python implementations to test tho

chrome mauve Oct 26, 2019, 8:35 AM

#

oh nice job @vestal solstice

#

50 chars

#

we're also doing the same for other algorithm in the paper, where we had to convert a 2 x n array of players and scores to an a x 11 array, where the initial 2 x n array is structured as follows:

Player   Score
123456   12345
123456   13452
112345   12331
...

and the finally array had to list at most 10 scores achieved by the same player

123456   12345   13452   ...
112345   12331   ...
...

The initial input is just a list of tuples

[(123456, 12345), (123456, 13452), (112345, 12331), ...]

and our output is just larger tuples.
Currently, our best golfed solution is

s=lambda a:[*{(i,)+(*(r for j,r in a if i==j),)[:10]for i,_ in a}]

grizzled cloak Oct 26, 2019, 8:49 AM

#

so you are matching scores to players?

chrome mauve Oct 26, 2019, 8:49 AM

#

yes

grizzled cloak Oct 26, 2019, 8:49 AM

#

but only the top 10?

#

or the first 10

chrome mauve Oct 26, 2019, 8:49 AM

#

the question never specified either so we just went with any 10

grizzled cloak Oct 26, 2019, 8:49 AM

#

k

chrome mauve Oct 26, 2019, 8:49 AM

#

like, some players may still score less than 10 times

grizzled cloak Oct 26, 2019, 8:50 AM

#

k

#

but no more than 10?

#

could you post a more complete input?

chrome mauve Oct 26, 2019, 8:51 AM

#

yeah no more than 10

#

[(123456, 12345), (123456, 13452), (112345, 12331), (3, 2), (3, 4), (3, 1), (3, 2), (3, 123), (3, 213), (3, 123), (3, 1212), (3, 124321), (3, 12), (3, 212)]

grizzled cloak Oct 26, 2019, 8:52 AM

#

thx

chrome mauve Oct 26, 2019, 8:52 AM

#

any 10 of the 11 scores by player 3 could be in it

grizzled cloak Oct 26, 2019, 8:52 AM

#

and the output should be in order?

chrome mauve Oct 26, 2019, 8:52 AM

#

nope

#

the q in the exam itself was to do it in pseudocode (classic high school exams)

grizzled cloak Oct 26, 2019, 8:53 AM

#

yours outputs this:
[(123456, 12345), (123456, 13452), (112345, 12331)]
so i dont need to know what scores belong to what player?

#

oh wait

chrome mauve Oct 26, 2019, 8:53 AM

#

uh, idt thats the output?

grizzled cloak Oct 26, 2019, 8:54 AM

#

hmm for me it just outputs the same as the input

chrome mauve Oct 26, 2019, 8:54 AM

#

[(797, 53), (797, 86), (204, 82), (204, 54), (204, 62), (204, 87), (204, 90), (204, 19), (313, 96), (313, 71), (313, 74), (313, 66), (313, 89), (313, 33), (313, 57), (313, 37), (313, 66), (313, 100), (313, 68), (313, 82), (313, 46), (313, 30), (553, 49), (553, 49), (553, 28), (553, 47), (553, 33), (553, 39), (553, 83), (553, 23), (553, 92), (553, 34), (553, 94), (553, 69), (553, 50), (553, 70), (553, 85), (469, 28), (469, 76), (469, 86), (469, 45), (469, 88), (469, 70), (469, 52), (469, 58), (469, 51), (469, 83), (634, 70), (634, 39), (634, 80), (634, 58), (634, 34), (634, 19), (634, 59), (634, 20), (133, 49), (133, 55), (133, 83), (133, 98), (133, 68), (133, 90), (133, 66), (133, 89), (133, 12), (133, 97), (133, 11), (133, 31), (133, 64), (756, 89), (756, 64), (756, 42), (756, 84), (880, 35), (880, 75), (880, 28), (880, 73), (382, 93), (382, 61), (382, 77), (382, 49), (382, 11), (382, 34), (382, 48), (382, 69), (382, 77), (382, 33), (382, 17), (382, 19), (382, 72), (382, 87), (828, 35), (828, 72), (828, 99), (828, 17), (828, 12), (828, 54), (828, 75), (828, 78), (828, 26), (828, 82), (143, 33), (969, 25), (730, 75), (730, 75), (730, 73), (730, 58), (730, 76), (591, 100), (591, 92), (591, 65), (591, 21), (591, 20), (591, 21), (591, 71), (591, 91), (591, 29), (591, 43), (591, 32), (591, 95), (591, 96), (273, 49), (101, 55), (293, 62), (293, 17), (293, 40), (293, 97), (885, 38), (885, 21), (885, 32), (885, 63), (885, 19), (885, 34), (885, 99), (461, 94), (461, 65), (461, 10), (461, 13), (461, 30), (461, 48), (461, 66), (461, 85), (461, 19), (461, 25), (461, 99), (461, 13)]

grizzled cloak Oct 26, 2019, 8:54 AM

#

inp = [(123456, 12345), (123456, 13452), (112345, 12331)]
s(inp)
[(112345, 12331), (123456, 12345, 13452)]```

chrome mauve Oct 26, 2019, 8:54 AM

#

thats a randomly generated series of (player, score) tuples

#

i got

[(880, 35, 75, 28, 73), (133, 49, 55, 83, 98, 68, 90, 66, 89, 12, 97), (469, 28, 76, 86, 45, 88, 70, 52, 58, 51, 83), (969, 25), (730, 75, 75, 73, 58, 76), (313, 96, 71, 74, 66, 89, 33, 57, 37, 66, 100), (756, 89, 64, 42, 84), (461, 94, 65, 10, 13, 30, 48, 66, 85, 19, 25), (382, 93, 61, 77, 49, 11, 34, 48, 69, 77, 33), (553, 49, 49, 28, 47, 33, 39, 83, 23, 92, 34), (273, 49), (634, 70, 39, 80, 58, 34, 19, 59, 20), (591, 100, 92, 65, 21, 20, 21, 71, 91, 29, 43), (204, 82, 54, 62, 87, 90, 19), (828, 35, 72, 99, 17, 12, 54, 75, 78, 26, 82), (797, 53, 86), (293, 62, 17, 40, 97), (143, 33), (101, 55), (885, 38, 21, 32, 63, 19, 34, 99)]

as output

grizzled cloak Oct 26, 2019, 8:55 AM

#

ah i must have redefined s at some point

#

oops

vestal solstice Oct 26, 2019, 9:28 AM

#

you can unpack instead of concatenating tuples```py
s=lambda a:[{(i,[r for j,r in a if i==j][:10])for i,_ in a}]

#

saves 4

chrome mauve Oct 26, 2019, 9:32 AM

#

lol, thats 3 ppl who've found it independently now

#

wait no

#

thats slightly different

#

s=lambda a:[*{(i,*(r for j,r in a if i==j))[:11]for i,_ in a}]

distant wave Oct 27, 2019, 9:29 AM

#

!e ```python
from typing import List

oracle_none = object()

def oracle(path: str) -> str:
return ".".join(_oracle(path))

def _oracle_iterrepr(obj: iter) -> str:
MAX = 40
res = ""
for item in obj:
n = repr(item)
if len(res + n) > MAX:
res += ", ..."
break

    if res:
        res += ", "
    res += n

if not res:
    res = "empty"
return res

def _oracle_repr(obj: object) -> str:
if isinstance(obj, dict):
return f"{{{_oracle_iterrepr(obj.keys())}}}"

if isinstance(obj, list):
    return f"[{_oracle_iterrepr(obj)}]"

return repr(obj)

def _oracle(path: str) -> List[str]:
path = path.split(".")
res = []
parent = globals()
for item in path:
next_parent = getattr(parent, item, oracle_none)

    if next_parent is not oracle_none:
        parent = next_parent
        res.append(_oracle_repr(parent))
        continue

    try:
        parent = parent[item]
    except KeyError:
        return res + [f"(Not Found: {item})"]
    except TypeError:
        try:
            parent = parent[int(item)]
        except (ValueError, IndexError):
            return res + [f"(Not Found: {item})"]

    res.append(_oracle_repr(parent))

return res

testlist = []
testlist2 = [1, 2, 3, "a", "b", oracle]

print(oracle("oracle.globals.testlist2.-1.globals.testlist.0"))

night quarryBOT Oct 27, 2019, 9:29 AM

#

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

<function oracle at 0x7f42321b5830>.{'__name__', '__doc__', '__package__', ...}.[1, 2, 3, 'a', 'b', ...].<function oracle at 0x7f42321b5830>.{'__name__', '__doc__', '__package__', ...}.[empty].(Not Found: 0)

brazen geyser Oct 27, 2019, 10:47 AM

#

@distant wave i did some more benchmarks and this seems to be very slightly faster than the other resolve:

@lru_cache(128)
def compiled(code):
    return compile(code, '<cached>', 'eval', optimize=2)
    
def resolve(name):
    return eval(compiled(name))

distant wave Oct 27, 2019, 10:48 AM

#

odd

brazen geyser Oct 27, 2019, 10:48 AM

#

plus the added benefit of automatically resolving attributes

#

oh wow, changing that eval call to eval(compiled(name), globals()) actually makes it faster too

#

not sure why

#

guess there's a check that gets bypassed if globals param isnt None

zealous widget Oct 27, 2019, 12:04 PM

#

In [26]: f = lambda:lambda x:x

In [27]: a = defaultdict(f)

In [28]: a[1]
Out[28]: <function __main__.<lambda>.<locals>.<lambda>(x)>

i feel like this is useful for something

brazen geyser Oct 27, 2019, 6:26 PM

#

continuing from the messages between
https://discordapp.com/channels/267624335836053506/303906556754395136/638073584157327391
and
https://discordapp.com/channels/267624335836053506/303906556754395136/638080360260829205
what exactly does LOAD_CLOSURE do to cause this behaviour?

#

bytecode gurus hear my call

sick hound Oct 27, 2019, 6:29 PM

#

>>> def test():
...     x = []
...     def inner():
...             x.append(1)
...             return x
...     return inner
...
>>> something = test()
>>> something()
[1]
>>> something()
[1, 1]
>>> something()
[1, 1, 1]
>>> something.__closure__
(<cell at 0x000002077C84E708: list object at 0x000002077C9D24C8>,)
>>> something.__closure__[0]
<cell at 0x000002077C84E708: list object at 0x000002077C9D24C8>
>>> something.__closure__[0].cell_contents
[1, 1, 1]
>>> something.__closure__[0].cell_contents.append(4)
>>> something()
[1, 1, 1, 4, 1]```

#

@brazen geyser it seems to be something to do with __closure__

brazen geyser Oct 27, 2019, 6:30 PM

#

right, according to the docs

MAKE_FUNCTION(argc)¶
Pushes a new function object on the stack. From bottom to top, the consumed stack must consist of values if the argument carries a specified flag value

0x01 a tuple of default values for positional-only and positional-or-keyword parameters in positional order

0x02 a dictionary of keyword-only parameters’ default values

0x04 an annotation dictionary

0x08 a tuple containing cells for free variables, making a closure

the code associated with the function (at TOS1)

the qualified name of the function (at TOS)

#

and the MAKE_FUNCTION for wrapped has 0x08 passed as its argument so ye

#

so then the question becomes wtf is a closure in this context

#

ive heard of the term used to refer to anonymous functions but nothing quite like this

snow beacon Oct 27, 2019, 6:33 PM

#

Doesn't the closure just allow inner to store the value of x even though it's outside the function?

sick hound Oct 27, 2019, 6:33 PM

#

yeah something like that

brazen geyser Oct 27, 2019, 6:34 PM

#

yep seems so

#

looking for some solid docs to lay it all out

sick hound Oct 27, 2019, 6:34 PM

#

usually x would be deleted after test finished, but inner keeps a reference to it and there's a reference to inner so it gets kept

#

and it's in the closure

brazen geyser Oct 27, 2019, 6:39 PM

#

okay now it's making sense.

#

though id still like to see an exact definition of what cell and closure mean

pure dew Oct 27, 2019, 6:39 PM

#

closure is just a fancy term for upvalue capture in this context i think

snow beacon Oct 27, 2019, 6:40 PM

#

A cell is an object with a reference contained in it. Used exclusively for function closures iirc

pure dew Oct 27, 2019, 6:41 PM

#

cell == rust Box<>?

#

or Ref maybe

#

probably Ref tbh

snow beacon Oct 27, 2019, 6:41 PM

#

You don't need a box in a dynamically typed language.

pure dew Oct 27, 2019, 6:41 PM

#

yea ik

#

I'm just trying to relate the term to something I know

snow beacon Oct 27, 2019, 6:42 PM

#

Although honesty I don't know why you need a cell.

brazen geyser Oct 27, 2019, 6:42 PM

#

so it's like a temporary storage location for a reference used during compilation?

#

storage location with an index

snow beacon Oct 27, 2019, 6:42 PM

#

Not temporary, it's used when the function runs, I think.

#

Not certain of that, actually.

formal sandal Oct 27, 2019, 9:01 PM

#

!e

import re
re.compile(r"\bHello[0-9]+\s+world(?=123)", re.DEBUG)

night quarryBOT Oct 27, 2019, 9:01 PM

#

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

001 | AT AT_BOUNDARY
002 | LITERAL 72
003 | LITERAL 101
004 | LITERAL 108
005 | LITERAL 108
006 | LITERAL 111
007 | MAX_REPEAT 1 MAXREPEAT
008 |   IN
009 |     RANGE (48, 57)
010 | MAX_REPEAT 1 MAXREPEAT
... (truncated - too many lines)

Full output: https://paste.pythondiscord.com/ozukojuyar

marsh void Oct 27, 2019, 9:17 PM

#

def load_cls_dict(cls):
    """Gets 'cls.__dict__' that can be edited."""
    for obj in gc.get_objects():

        try:
            if obj == dict(cls.__dict__) and type(obj) is dict:
                return obj

        except:
            continue

    raise ValueError(
        'Failed to find editable __dict__ for {}.'.format(cls)
    )

def get_class_dict(cls):
    """Slower than load_cls_dict. Used to prevent cls.__dict__ copy trick."""
    maybe_actual = load_cls_dict(cls)
    maybe_actual[13] = 13
    actual = load_cls_dict(cls)
    actual.pop(13, None)
    return actual``` Ok here are the versions I am using, feel free to golf and other stuff

formal sandal Oct 27, 2019, 9:21 PM

#

Maybe except: instead of except Exception:?

marsh void Oct 27, 2019, 9:22 PM

#

Not sure if it is possible to get BaseException direct subclass, haha

#

But ok, let's have it

formal sandal Oct 27, 2019, 9:23 PM

#

is it possible that dict doesn't return a dict?

marsh void Oct 27, 2019, 9:23 PM

#

it is not imo

#

I mean, I think my checks are enough

formal sandal Oct 27, 2019, 9:25 PM

#

My point was that type(obj) is dict: is redundant.

#

Or is it not?

marsh void Oct 27, 2019, 9:26 PM

#

Oh, no, lemme explain

#

class tricker:
    def __eq__(self, other):
        return True

trick = tricker()```

#

@formal sandal here

formal sandal Oct 27, 2019, 9:31 PM

#

oh

brazen geyser Oct 27, 2019, 10:10 PM

#

is it possible to make a function run with a specific globals and locals dict?

#

without exec

worthy pollen Oct 27, 2019, 10:27 PM

#

eval?

marsh void Oct 27, 2019, 10:34 PM

#

@brazen geyser both eval and exec take globals and locals arguments

brazen geyser Oct 27, 2019, 10:35 PM

#

without exec or eval

gilded orchid Oct 27, 2019, 10:59 PM

#

could maybe be possible with a decorator or something

#

idk tho

brazen geyser Oct 27, 2019, 11:07 PM

#

>>> import inspect
>>> inspect.currentframe().f_locals = {}
Traceback (most recent call last):
  File "<pyshell#3>", line 1, in <module>
    inspect.currentframe().f_locals = {}
AttributeError: attribute 'f_locals' of 'frame' objects is not writable
>>>

#

rip

brazen geyser Oct 27, 2019, 11:25 PM

#

alternatively, can you make name resolution in a dynamically created module work like it would in an ordinary module?
e.g. imagine you had this file: module.py

class A:
    pass

def new_a():
    return A()

doing this in another file would be no problem:

import module
a = module.new_a()

v.s. using a dynamically generated module

import sys
from types import ModuleType

module = ModuleType('module')
sys.modules['module'] = module

class A:
    pass

def new_a():
    return A()

module.A = A
module.new_a = new_a

del new_a
del A

a = module.new_a()

which results in the following error:

Traceback (most recent call last):
  File "C:/Users/admin/Documents/Projects/shm_server/scrap.py", line 19, in <module>
    a = module.new_a()
  File "C:/Users/admin/Documents/Projects/shm_server/scrap.py", line 11, in new_a
    return A()
NameError: name 'A' is not defined

#

im guessing it involves making a spec first then using importlib to create the module based on the spec instead

pure dew Oct 28, 2019, 3:39 AM

#

Repost from python-discussion, but I'm especially proud of this, and it is pretty much just esoteric python https://paste.a-sketchy.site/mujopepuye.py

#

File imports work

snow beacon Oct 28, 2019, 6:44 AM

#

That file doesn't actually run, right?

#

It doesn't start with an 'import PyskellParser' or anything, so it'll error on include being an undefined identifier.

pure dew Oct 28, 2019, 11:45 AM

#

uh no

#

you still have to python -m pyskell pyskell/examples/import.py

#

I'll make an import based entry point after a while

grave rover Oct 28, 2019, 3:34 PM

#

@pure dew actually, imports dont work like that in C

pure dew Oct 28, 2019, 3:34 PM

#

yea I know

grave rover Oct 28, 2019, 3:34 PM

#

#import "..." just searches local first, then global, and the other way around for <>

pure dew Oct 28, 2019, 3:35 PM

#

i said similar

#

i wonder if i oughta just use one style

formal sandal Oct 28, 2019, 4:38 PM

#

Actually, this works in HTML (at least in firefox):

<b> aaaaaaaaaaa <i> bbbbbbbbbbb </b> ccccccccccc </i>

Despite the fact that you can't really turn it into a nested structure.

#

So firefox transforms that HTML:

📎 unknown.png

#

Maybe it's possible to do this in programming as well?

#

fizzbuzz(n) =
    from  [   ( n%3==0  {            ) n%5==0 } ]    
    mapto [ n (  "Fizz" { "FizzBuzz" ) "Buzz" } ]

#

See, it's like a Venn diagram.

#

There are two circles, "divisible by 3" and "divisible by 5". Everything outside the circles will result in n being return, the intersection will give "FizzBuzz" and so on.

formal sandal Oct 28, 2019, 5:18 PM

#

Insertion sort:

isort list =
    let m = min list
    split  [(== m)       ]
    apply  [( id ) isort ]
    as     [( a  ) b     ]
    return a ++ b

#

And, of course,

qsort x:xs =
    split  [( < x )    ( > x )]
    apply  [(qsort) id (qsort)]
    as     [( lt  ) eq ( gt  )]
    return lt ++ eq ++ gt

#

A bit verbose, but

binsearch list x = 
    if list == []
        return Nothing
    else
        let mid = list.middle

        if mid == x
            return Just x
        list
        split      [(< mid) (> mid)]
        as         [( lt  ) ( gt  )]
        elif mid < x
            return (binsearch gt x)
        elif mid > x  
            return (binsearch lt x)
        else
            impossible

#

The idea is that you can intertwine not only data structures, but control flow.

#

This might be useful if you have something like:

if green:
    color = CO_GREEN
    GreenManagerSingleton.Initialize()
else:
    print(123)

robot.abort_exhumation()

if green:
    color = CO_RED
    GreenManagerSingleton.deinitialize()
    return 0
else:
    return 1
```where the control flow is "interrupted" by some common action.

formal sandal Oct 28, 2019, 6:03 PM

#

Check whether a user (if the username and password are correct) is an admin:

check_adminness uname pwd = 
    from  [(uname in UNAMES {match uname pwd})              ]
    mapto [(Error "pwd"     { User uname    }) Error "uname"]
    apply [( id             { isadmin       })  id          ]

zealous widget Oct 28, 2019, 6:22 PM

#

this isn't right, but it's a start --- it should check for complete containment and probably other things too:

from itertools import combinations

class Venn:
    opens = "<({"
    closes = ">)}"

    def __init__(self, venn_string):
        self.venn_string = venn_string

    def __repr__(self):
        sets = [(character,index)
                for index, character in enumerate(self.venn_string)
                if character not in self.opens + self.closes]

        closures = [(index, self.venn_string.find(self.closes[self.opens.find(character)], index))
                    for index, character in enumerate(self.venn_string)
                    if character in self.opens]

        for (s1, e1), (s2, e2) in sorted([*combinations(closures, 2)]):
            if e1 > s2:
                if (s1, e1) in closures:
                    closures.remove((s1, e1))
                closures.extend([(s1, s2), (s1, e2)])
        closures.sort()

        return ", ".join("&".join(set_ for set_, index in sets if start < index < end)
                         for start, end in closures)

"""
Venn("<a(b>)") -> a, a&b, b
"""

formal sandal Oct 28, 2019, 6:43 PM

#

thinkmon

#

Actually... I don't know how to specify a three-circle Venn diagram.

#

Like A, B, C, A&B\C, A&C\B, B&C\A, A&B&C.

zealous widget Oct 28, 2019, 7:00 PM

#

yeah, i tried it with 3 and got:

In [159]: print(Venn("<a(b{c)}>"))
a, a&b, a&b&c, a&b&c, b, b&c, c

#

which isn't right

#

well, b & c are completely contained in a

formal sandal Oct 28, 2019, 7:17 PM

#

Well,

<  a (  b    {  c  )      } >
<  a  a&b\c  a&b&c   a&c\b

#

I don't think you can communicate this picture in 1D:

📎 3-circles-png-venn-diagram-transparent-background-free-venn-diagram-transparent-820_816.png

#

Oh, wait

#

< a ( b     { c    >      )   }
< a ( a&b\c {a&b&c > b&c\a) c }

#

But you still can't split it into 7 distinct cases.

edgy kelp Oct 28, 2019, 7:21 PM

#

I just peeked over here and I'm very confused and scared from these symbols you're throwing here 😄

formal sandal Oct 28, 2019, 7:21 PM

#

Maybe it will become more clear if you scroll up to the

fizzbuzz(n) =
    from  [   ( n%3==0  {            ) n%5==0 } ]    
    mapto [ n (  "Fizz" { "FizzBuzz" ) "Buzz" } ]
```message

thin trout Oct 28, 2019, 7:22 PM

#

They are doing some voodoo invocations @edgy kelp no worries

formal sandal Oct 28, 2019, 7:22 PM

#

Actually I'm just procrastinating because I have to do a huge homework task.

edgy kelp Oct 28, 2019, 7:22 PM

#

automate it

formal sandal Oct 28, 2019, 7:22 PM

#

I can't...

#

That's a physics assignment, and while I can automate the computation, I will still have to write all the steps on a piece of paper

thin trout Oct 28, 2019, 7:23 PM

#

Automate the printing

formal sandal Oct 28, 2019, 7:24 PM

#

I would have to buy a plotter since i can't print it.

thin trout Oct 28, 2019, 7:25 PM

#

Automate the buying

grizzled cloak Oct 28, 2019, 7:26 PM

#

I love doing physics calculations with python

zealous widget Oct 28, 2019, 7:26 PM

#

<a{b(c>})

i think this is as close as you can get, it doesn't include a&c by themselves

grizzled cloak Oct 28, 2019, 7:26 PM

#

Especially if im not quite getting them, teaching a computer just does it for me

formal sandal Oct 28, 2019, 7:38 PM

#

@zealous widget
a{b(c>}<)

#

brainmon

zealous widget Oct 28, 2019, 7:40 PM

#

is there any a&b without c

formal sandal Oct 28, 2019, 7:40 PM

#

Yes.

#

The only region missing is B \ (A | C)

zealous widget Oct 28, 2019, 7:42 PM

#

you could do non contiguous sets

formal sandal Oct 28, 2019, 7:42 PM

#

a   {   b   (   c   >       }       <     )
a      ab\c    abc    bc\a   c\(a|b)  ac\b

#

Actually, the correct way to write this would be {b(c>}<a)

zealous widget Oct 28, 2019, 7:47 PM

#

this is a tiny one-dimensional torus

formal sandal Oct 28, 2019, 7:47 PM

#

boring mode on Maybe <a{b(c>}){b\(a|c)}?

zealous widget Oct 28, 2019, 7:48 PM

#

class Venn:
    opens = "<({"
    closes = ">)}"

    def __init__(self, venn_string):
        self.venn_string = venn_string

    def __repr__(self):
        sets = [character for character in self.venn_string
                if character not in self.opens + self.closes]

        closures = [(index, self.venn_string.find(self.closes[self.opens.find(character)], index))
                    for index, character in enumerate(self.venn_string)
                    if character in self.opens]

        closure_set = {set_:set(range(start, end+1)) for (start, end), set_ in zip(closures, sets)}

        return str(closure_set)



"""
In [170]: print(Venn("<a(b{c>)}"))
{'a': {0, 1, 2, 3, 4, 5, 6}, 'b': {2, 3, 4, 5, 6, 7}, 'c': {4, 5, 6, 7, 8}}
"""

just comparing relations between these sets of numbers would get a lot of the way

#

but these sets don't have to be contiguous, if you find a nice way to break them up

#

hmm, you could instead not use the brackets and braces and just use the first appearance of a set symbol to open and a 2nd to close

a b c a b c
| | | | | |
a | | | | |
 a&b| | | |
  a&b&c | |
     b&c| |
        c |
          ∅

#

seems easier to parse

#

then you could do:

abcabcaccabb

i think gets all the combinations?

#

class Venn:
    def __init__(self, venn_string):
        self.venn_string = venn_string

    def __repr__(self):
        relations = set()
        s = ""
        for symbol in self.venn_string:
            if symbol in s:
                s = "".join(char for char in s if char != symbol)
            else:
                s += symbol
            relations.add("&".join(s))

        return ", ".join(relations)

In [177]: print(Venn("abcabc"))
c, , a&b, a, a&b&c, b&c

In [178]: print(Venn("abcabcaccabb"))
c, , a&b, a&c, a, b, a&b&c, b&c

#

hmm, it uses an empty string for the empty set

#

i guess that makes sense

#

class substring(str):
    def __sub__(self, other):
        return substring("".join(char for char in self if char not in other))

    def __add__(self, other):
        return substring("".join((self, other)))

class Venn:
    def __init__(self, venn_string):
        self.venn_string = venn_string

    def __repr__(self):
        relations = set()
        s = substring("")
        for symbol in self.venn_string:
            s = s - symbol if symbol in s else s + symbol
            relations.add("&".join(s) if s else "∅")
        return ", ".join(relations)

now we got a nice symbol -- looks nicer if i can subtract with strings

zealous widget Oct 28, 2019, 9:57 PM

#

i was able to get all combinations with this string:

In [199]: Venn("abcbabcb")
Out[199]: c, a&b, a&c, a, b, ∅, c&b, a&b&c

snow beacon Oct 28, 2019, 10:09 PM

#

It might be cleaner to subclass frozenset

zealous widget Oct 28, 2019, 10:17 PM

#

i used a set instead, but it's awkward cause sets dont have __add__

class Venn:
    def __init__(self, venn):
        self.__make__(venn)

    def __make__(self, venn):
        relations = set()
        s = set()
        for symbol in venn:
             (set.remove if symbol in s else set.add)(s, symbol)
             relations.add("&".join(s) if s else "∅")
        self.__rep__ = ", ".join(relations)

    def __repr__(self):
        return self.__rep__

#

i mean, instead of subclassing anything

#

it seems odd that __add__ isn't used for union and __iadd__ isn't update

#

maybe there's some ambiguity somewhere

snow beacon Oct 28, 2019, 11:10 PM

#

It would get confusing with .add, for one.

zealous widget Oct 28, 2019, 11:37 PM

#

maybe, but i wouldn't ever call the dunder method with .__add__

snow beacon Oct 28, 2019, 11:42 PM

#

If you're subclassing it, it'd come up.

zealous widget Oct 28, 2019, 11:46 PM

#

fair

formal sandal Oct 29, 2019, 10:33 AM

#

Well, I don't think this will be very useful with conditions. You'd need to make sure they don't overlap and write all the individual cases.

rugged sparrow Oct 29, 2019, 2:12 PM

#

in 3.8 globals().clear() doesnt wipe globals

#

however it will remove variables created by the script

#

it basically resets globals instead of clearing it

#

__builtins__.clear() still breaks everything tho

sick hound Oct 29, 2019, 5:15 PM

#

>>> __name__
'__main__'
>>> globals().clear()
>>> __name__
'builtins'``` (3.8)

brazen geyser Oct 29, 2019, 5:19 PM

#

lol

#

how does that happen

#

>>> globals().clear()
>>> globals()['__builtins__'].clear()

=============================== RESTART: Shell ===============================
>>>

nice

#

oh right that should be __builtins__.__dict__.clear()

sick hound Oct 29, 2019, 5:25 PM

#

>>> globals().keys()
dict_keys(['__name__', '__doc__', '__package__', '__loader__', '__spec__', '__annotations__', '__builtins__'])
>>> type(__builtins__)
<class 'module'>
>>> globals().clear()
>>> globals().keys()
dict_keys(['__builtins__'])
>>> type(__builtins__)
<class 'dict'>```

#

it gets cleared, then __builtins__ gets put back as a dictionary

#

nothing else gets put back

#

so __name__ goes to __builtins__ and you get 'builtins' ```py

builtins['name']
'builtins'```

brazen geyser Oct 29, 2019, 5:26 PM

#

thats.. weird. to say the least.

sick hound Oct 29, 2019, 5:27 PM

#

maybe you shouldn't have done globals().clear() then? :P

#

this is #esoteric-python, everything is weird

brazen geyser Oct 29, 2019, 5:27 PM

#

it gets cleared, then __builtins__ gets put back as a dictionary why does this happen in particular lmao

sick hound Oct 29, 2019, 5:28 PM

#

well putting it back is probably so that you don't do something silly like del __builtins__ and mess everything up

#

we have no idea why it becomes a dictionary though

snow beacon Oct 29, 2019, 7:51 PM

#

It probably needs to be reconstructed, and dicts are easier.

sick hound Oct 29, 2019, 7:53 PM

#

>>> globals().clear()
>>> import builtins
>>> builtins
<module 'builtins' (built-in)>```

snow beacon Oct 29, 2019, 7:58 PM

#

Does it reimport the module, or is that just assignment?

sick hound Oct 29, 2019, 8:04 PM

#

we don't know what you're asking but does this answer your question? ```py

globals().clear()
import sys
sys.modules['builtins']
<module 'builtins' (built-in)>```

snow beacon Oct 29, 2019, 8:39 PM

#

Yep.

crystal mica Oct 30, 2019, 2:16 AM

#

👀

distant wave Oct 30, 2019, 2:16 AM

#

!e ```python
import sys
builtins.dict.clear()
sys.modules.clear()
import builtins
builtins.print(sys.modules)

night quarryBOT Oct 30, 2019, 2:16 AM

#

@distant wave :x: Your eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 4, in <module>
003 | ImportError: __import__ not found

distant wave Oct 30, 2019, 2:16 AM

#

I wonder if that's even recoverable

crystal mica Oct 30, 2019, 2:17 AM

#

I dont think so

#

That sounds pretty dangerous

rugged sparrow Oct 30, 2019, 2:19 AM

#

@distant wave if object and True and False exist then its recoverable

#

Oh no

#

They dont

snow beacon Oct 30, 2019, 2:31 AM

#

I thought we figured out it was recoverable

distant wave Oct 30, 2019, 2:36 AM

#

If you already had an object remaining, likely

#

!e ```python
import sys
builtins.dict.clear()
sys.modules.clear()
sys.dict

night quarryBOT Oct 30, 2019, 2:36 AM

#

@distant wave :warning: Your eval job has completed with return code 0.

[No output]

distant wave Oct 30, 2019, 2:36 AM

#

!e ```python
import sys
builtins.dict.clear()
sys.modules.clear()
sys.dict.class.globals

night quarryBOT Oct 30, 2019, 2:36 AM

#

@distant wave :x: Your eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 4, in <module>
003 | AttributeError: type object 'dict' has no attribute '__globals__'

distant wave Oct 30, 2019, 2:37 AM

#

!e ```python
import sys
builtins.dict.clear()
sys.modules.clear()
sys.dict['stdout'].write(sys.dict.class.mro)

night quarryBOT Oct 30, 2019, 2:37 AM

#

@distant wave :x: Your eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 4, in <module>
003 | TypeError: write() argument must be str, not tuple

distant wave Oct 30, 2019, 2:38 AM

#

!e ```python
import sys
builtins.dict.clear()
sys.modules.clear()
sys.dict['stdout'].write(sys.dict.class.mro.str())

night quarryBOT Oct 30, 2019, 2:38 AM

#

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

(<class 'dict'>, <class 'object'>)

distant wave Oct 30, 2019, 2:38 AM

#

Looks recoverable

#

!e ```python
import sys
builtins.dict.clear()
sys.stdout.write(sys.dict"globals".str())

night quarryBOT Oct 30, 2019, 2:39 AM

#

@distant wave :x: Your eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 3, in <module>
003 | KeyError: 'globals'

distant wave Oct 30, 2019, 2:40 AM

#

!e ```python
import sys
builtins.dict.clear()
sys.modules.clear()
sys.stdout.write(sys.dict"import".str())

night quarryBOT Oct 30, 2019, 2:40 AM

#

@distant wave :x: Your eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 4, in <module>
003 | KeyError: '__import__'

rugged sparrow Oct 30, 2019, 2:54 AM

#

It is recoverable even if you don't have sys. Just takes some effort

#

Cause you still have integer literals

distant wave Oct 30, 2019, 2:59 AM

#

How do you get to importing from integer literals?

#

I mean, we can get to object and maybe traverse the subclasses tree maybe?

rugged sparrow Oct 30, 2019, 3:00 AM

#

Yea you use the subclasses tree

distant wave Oct 30, 2019, 3:00 AM

#

I wonder

#

oh

rugged sparrow Oct 30, 2019, 3:01 AM

#

(1).

distant wave Oct 30, 2019, 3:01 AM

#

!e ```python
import sys
builtins.dict.clear()
sys.modules.clear()
(1).class.mro[-1].subclasses = 1

night quarryBOT Oct 30, 2019, 3:01 AM

#

@distant wave :x: Your eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 4, in <module>
003 | TypeError: can't set attributes of built-in/extension type 'object'

rugged sparrow Oct 30, 2019, 3:03 AM

#

Also (1).__class__.__base__ == object

distant wave Oct 30, 2019, 3:04 AM

#

!e ```python
print((1).class.mro)

night quarryBOT Oct 30, 2019, 3:04 AM

#

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

(<class 'int'>, <class 'object'>)

distant wave Oct 30, 2019, 3:04 AM

#

You'd need to ctypes object to break it further imo

snow beacon Oct 30, 2019, 3:44 AM

#

You don't need the 1, the brackets suffice

marsh void Oct 30, 2019, 11:23 AM

#

Admins should create #ctypes-abuse since this is only related to CPython implementation, and not fully belongs to esoteric-python >.>

brisk zenith Oct 30, 2019, 11:33 AM

#

nah, ctypes fuckery still falls under weird things to do with python, it doesn't matter if its only for one implementation :D

sick hound Oct 30, 2019, 1:24 PM

#

~~I think that was sarcasm~~

marsh void Oct 30, 2019, 2:16 PM

#

And yes and no

formal sandal Oct 30, 2019, 2:35 PM

#

Maybe we should just have 8 or so different eso channels.

#

#eso-languages
#eso-ctypes
#eso-obfuscation
#eso-golfing
...

#

That wouldn't make any sense, though, because this channel is not that popular.

sullen rock Oct 30, 2019, 9:39 PM

#

I'd definitely join the esoteric-python discord instance, just sayin'

gilded orchid Oct 30, 2019, 10:19 PM

#

Idea: could we start weekly or monthly #esoteric-python golfing competitions

#

maybe there could be a simple task and whoever is able to golf it the best wins I guess

snow beacon Oct 30, 2019, 10:20 PM

#

Wasn't there already a competition each week/month that died out?

gilded orchid Oct 30, 2019, 10:20 PM

#

oh was there?

#

I've only been in this server like a year so I wouldn't know

brisk zenith Oct 30, 2019, 10:59 PM

#

i mean, the esoteric python challenges aren't strictly about golfing, but there's no reason why people can't submit golfed solutions or work with others to golf it together.

#

also, i'm trying not to let it die out but interesting ideas are difficult to come up with and i ~~don't have the time~~ forget to spend time every couple weeks or so trying to think of a new challenge.

zealous widget Oct 30, 2019, 11:18 PM

#

no one told me this was a thing : for i in range(10):{print(i)}

brisk zenith Oct 30, 2019, 11:19 PM

#

i mean

#

yes

edgy kelp Oct 30, 2019, 11:21 PM

#

what are the braces for?

brisk zenith Oct 30, 2019, 11:24 PM

#

the braces do nothing really, except make it look slightly like a language that uses braces

#

technically the braces are just syntax for creating sets

zealous widget Oct 30, 2019, 11:28 PM

#

    from __future__ import braces
                                 ^
SyntaxError: not a chance

sick hound Oct 30, 2019, 11:29 PM

#

yert

zealous widget Oct 30, 2019, 11:29 PM

#

pretty funny easter egg

hollow patrol Oct 31, 2019, 12:51 AM

#

I created a Context Manager with the power of a decorator

#

It's a class

#

You just subclass it, and implement a extend method (you can also do an __init__)

#

It converts everything in the context manager to a function

#

And it also blocks the natural call of the Context Manager

#

So you can implement your own custom calls for the context manager

#

With a context manager like this:

#

class iterate(ContextManager):
  def __init__(self, name, iterable):
    self.name = name
    self.iterable = iterable
  def extend(self, func, glob, loc):
    for item in self.iterable:
      loc[self.name] = item
      func()```

#

You can now do

#

with iterate('x', range(3)):
  print(x)
>>> 0
>>> 1
>>> 2```

#

If anyone finds it useful, I'll publish it

snow beacon Oct 31, 2019, 2:31 AM

#

I would be interested, although my interest is academic rather than practical. I tried to do something similar, and decided it was impossible.

frozen mica Oct 31, 2019, 8:38 AM

#

how do you bring the variables to the scope tho? Is it some magical stack-hack or something simple that I am not seeing?

gilded orchid Oct 31, 2019, 10:05 AM

#

(lambda x:(lambda l: print('print(\'\'.join([*'+(''.join([*map(l,x)]))+']))'))(lambda x:f'(lambda x:chr(x))({ord(x)}),'))```
I just made this terrible thing for the string obfuscation challenge

#

I probably won't submit it though since it's a bit basic

grave rover Oct 31, 2019, 2:07 PM

#

mfw I had to write shitcode today

#

instead of calling main I had to call main.callback.__closure__[0].cell_contents

gilded orchid Oct 31, 2019, 2:45 PM

#

oh god why

thin trout Oct 31, 2019, 3:13 PM

#

I got a challenge for you guys at #esoteric-python (you always came up with crazy thing) can you make this to work? (with some C or something)

False, True = True, False```

marsh void Oct 31, 2019, 3:17 PM

#

oh, it is here lol

grizzled cloak Oct 31, 2019, 3:20 PM

#

@thin trout you mean catch the syntaxError from assigning to a keyword or actually change the behavior

#

like make False evaluate to True after that

thin trout Oct 31, 2019, 3:21 PM

#

Actually change it

#

Yep

marsh void Oct 31, 2019, 3:23 PM

#

True is not even in globals(), haha

rugged sparrow Oct 31, 2019, 3:24 PM

#

I think it's doable

marsh void Oct 31, 2019, 3:24 PM

#

Same but uhm

#

Edit __debug__ for me pls if we manage to change False to True

rugged sparrow Oct 31, 2019, 3:26 PM

#

Haha

marsh void Oct 31, 2019, 3:27 PM

#

segmentation fault``` there we go

#

lol

rugged sparrow Oct 31, 2019, 3:27 PM

#

First segfault of the challenge lol

marsh void Oct 31, 2019, 3:27 PM

#

Yeah

#

>>> True = False
SyntaxError: why would you do this?```

grizzled cloak Oct 31, 2019, 3:37 PM

#

this does seem to break everything because python uses those internally.

#

i just get a bunch of errors because something evaluates to False instead of True etc

rugged sparrow Oct 31, 2019, 3:38 PM

#

what did you do to swap them?

grizzled cloak Oct 31, 2019, 3:38 PM

#

some weird trikery overwriting the bool method

#

cant do that at all on the boolean type

#

but you can for ints.

#

and then everything breaks

rugged sparrow Oct 31, 2019, 3:39 PM

#

>>> bool.__bool__
<slot wrapper '__bool__' of 'int' objects>``` hmm 🤔

grizzled cloak Oct 31, 2019, 3:44 PM

#

yeah

wind maple Oct 31, 2019, 7:29 PM

#

bool is a subclass of int afterall

marsh void Oct 31, 2019, 7:34 PM

#

^