esoteric-python

[fluid tree]

Gitlab.com/onelinerize

[rugged sparrow]

@fluid tree __class__ doesnt exist in the upper scope in a typical way. Its added here if it is explicitly requested (by super or manually) https://github.com/python/cpython/blob/a81fca6ec8e0f748f8eafa12fb12cf9e12df465c/Python/compile.c#L588

the closure is only added if __class__ is requested inside of a class method

!e py class A: print(__class__)

[night quarry]

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

001 | Traceback (most recent call last):
002 |   File "<string>", line 1, in <module>
003 |   File "<string>", line 2, in A
004 | NameError: name '__class__' is not defined

[rugged sparrow]

^ this fails because of that

!e py class A: def __init__(self): print(__class__) A()

[night quarry]

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

<class '__main__.A'>

[rugged sparrow]

due to this being done by the compiler like this, it means that classes constructed using metaclass(name, bases, namespace) cannot use the "magic" __class__ variable without needing extra work

[astral rover]

Wtf

So you don't need to self.class if you're in a method?

[rugged sparrow]

self.__class__ can be a subclass of the class you are defining

__class__ will always be the class you are defining

[fluid tree]

Afaik that constucts a synthetic reference to it. It doesn't actually implement the class at the class level

[fluid tree]

I know

[rugged sparrow]

it tells python to add a closure there. the actual __class__ closure is filled inside __built_class__ i believe

[astral rover]

Anyone here actually know how to use build_class?

I couldn't figure it out

[frigid wharf]

look at the channel description

Golfing, Python VM languages, obfuscation, code gore and other general Python weirdness

[stark fable]

s=lambda a:lambda b:lambda c:a(c)(b(c));k=lambda a:lambda b:a;factorial=lambda a:s(s(s(s(k)(k))(k(s(s(k(s(s(k(s))(s(k(k))(s(k(s))(s(k(s(s(k)(k))))(k)))))(k(k))))(s(s(k(s(s(k(s))(s(s(k(s))(k))(k(s(s(k)(k))(k(s(s(k(s))(s(k(s(k(s))))(s(k(k)))))(k(s(k)(k)))))))))(k(k(k(s(k)(k)))))))(s(s(k)(k))(k(k))))(s(s(k)(k))(k(k(s(k)(k)))))))(s(k(s(s(k(s))(s(k(s(k(s))))(s(k(k)))))(k(s(k)(k)))))(s(s(k)(k))(k(k(s(k)(k)))))))))(k(s(s(s(k)(k))(k(s(k)(k))))(k(s(k)(k))))))(k(k))((lambda a:a(a))(lambda a:lambda b:(lambda c:lambda d:a(a)(b-1)(c)(c(d)))if b else lambda a:lambda a:a)(a))(lambda a:a+1)(0)```

aside from failing on anything >= 6, this is a completely valid and working definition of factorial

[fluid tree]

"Failing" as in slow

[acoustic gust]

I joined chat and saw this....

wtf is this......

[shadow quest]

[loud flicker]

HOW DO YOU SAY HELLO WORLD

[floral meteor]

last time i tried that I got this instead:

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

[loud flicker]

FACTS

Print "hello world" doesnt work

[snow beacon]

That's because you didn't make a closure that time.

[floral meteor]

ik

So I was doing my usual esoterically making cool stuffs, but then this random line is giving me IndexError for no reason so I decided to use that as an extra random factor to change the colour instead of trying to debug it.

import os,sys;os.system('')
green_resmon=__import__('hexspam').print_green
def _init():
 w = os.get_terminal_size()[0]
 from random import randint as r
 a = [r(0,1)*r(0,1)*r(0,1)*r(0,1)for _ in range(w)]
 from functools import reduce
 return locals()
@lambda f:lambda:f(_init())
def matrix(__l):
 print(__l)
 a,r,reduce=__l['a'],__l['r'],__l['reduce']
 while w:=os.get_terminal_size()[0]:
  w>len(a)and a.extend([r(0,1)*r(0,1)*r(0,1)*r(0,1)for _ in range(w-len(a)+1)])
  w<len(a)and(a:=a[:w])
  try:
    if reduce(int.__mul__,[r(0,3)for i in range(w//16)]):a[r(0,w)]=r(0,1)*r(0,1)*r(0,1)*r(0,1)
  except:print(f'\x1b[{r(30,37)}')
  @print
  @lambda f:''.join([chr(f(a,i))for i in range(w)])
  def s(a,i):return r(33,127)if a[i]else 32
if __name__=='__main__':
 try:matrix()
 finally:print('\x1b[0m')

[snow beacon]

Which line?

[floral meteor]

The one where reduce is used

But it hasn't changed colour for a long time now

Also that script occasionally has a line that's just... out of place

also, when i make the window smaller, it goes faster

oh cool it turned purple now

[snow beacon]

It because random.randint is inclusive.

When random.randint(0, w) chooses w, that's outside of a, which has length w.

[floral meteor]

Thank you it works beautifully now

[olive stag]

can someone help me with this program

[floral meteor]

since you asked in the esoteric python channel...

[floral meteor]

!e ```py
{print(f"the cube root of {a} is {a**3}")for c in iter(lambda a=int(input('Enter a positive number, or a negative to quit')):a<0,True)}

that should work @olive stag

[olive stag]

was i supposed to ask in a different channel i don’t understand this or discord 😭

[earnest wing]

#❓｜how-to-get-help

[stark fable]

no, "failing" as in you get a recursionerror

[fluid tree]

That counts

[stark fable]

no it doesn't

[sick hound]

there is python syntax \ which indicates that a statement is continued on the next line, but is there syntax to indicate that a statement should be broken up to two lines?

print(1) | print(2)

_______outputs_______
>1
>2 

[earnest wing]

;

[fluid tree]

If you increase the recursion limit

[fluid tree]

There's ; but it never played well with indent for me. If you need to put a lot of stuff on one line, use esoteric python.

[sick hound]

What is the least used keyword in python?

[proper vault]

likely lambda

or maybe __peg_parser__/__debug__

also nonlocal

[sick hound]

I use and like lambda

[proper vault]

there are very few cases for lambda

[sick hound]

But yeah, nonlocal is unusual

[viscid nymph]

I would say nonlocal is used much less than lambda

[fluid tree]

Not keywords

Wait

They are?

[proper vault]

__peg_parser__ is, __debug__ is a bit more questionable, but since it isn't a variable name, I would call it a keyword

[tribal moon]

!e py import keyword print(keyword.kwlist)

[night quarry]

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

['False', 'None', 'True', '__peg_parser__', 'and', 'as', 'assert', 'async', 'await', 'break', 'class', 'continue', 'def', 'del', 'elif', 'else', 'except', 'finally', 'for', 'from', 'global', 'if', 'import', 'in', 'is', 'lambda', 'nonlocal', 'not', 'or', 'pass', 'raise', 'return', 'try', 'while', 'with', 'yield']

[frigid wharf]

does __peg_parser__ do anything except raising a SyntaxError?

[astral rover]

Don't think so

[astral rover]

Surely if False None and True are keywords so is debug

[proper vault]

well, obj.__debug__ is possible, but obj.None isn't

[astral rover]

Hmm

Ig so then

[proper vault]

it also isn't in keywords.kwlist

[viscid nymph]

Well we can't judge keywords that way 👀

[proper vault]

those are soft keywords

they are possible since 3.9 and exist since 3.10

[cosmic night]

idk if right place for this, but anyone have experience with monad libraries in python? most fleshed out one i could find is returns https://github.com/dry-python/returns/blob/master/README.md. Particularly interested in Final Tagless or Effect Handler approaches, only thing ive found in that vein is python-effect

[astral rover]

Looks very rust like

[bitter iris]

Counter({'def': 2469801,
         'None': 2079423,
         'if': 1952145,
         'return': 1535357,
         'import': 1264292,
         'from': 900304,
         'in': 900118,
         'True': 734511,
         'not': 714855,
         'assert': 566377,
         'False': 560890,
         'for': 559342,
         'is': 539126,
         'else': 493987,
         'class': 473044,
         'raise': 328813,
         'as': 273045,
         'and': 262468,
         'with': 217100,
         'elif': 205600,
         'try': 201381,
         'except': 198135,
         'or': 166449,
         'pass': 123573,
         'lambda': 86972,
         'await': 83763,
         'async': 73891,
         'yield': 51696,
         'while': 39673,
         'continue': 39332,
         'break': 32802,
         'del': 31962,
         'finally': 19407,
         'global': 7115,
         'nonlocal': 775})

aggregated results from a couple thousand python projects

In case of anyone curious, this is the script: https://github.com/isidentical/snippets/blob/master/most_used_keyword.py

[twilit grotto]

wow, didn't expect None so high

and how are there so many more defs than returns

[earnest wing]

side effects

lots of side effects

[hard spoke]

'async': 73891,
I wonder if it's disproportionately discord.py projects 😅

[viscid nymph]

Of course lol

[viscid nymph]

Yay I was right about nonlocal

Surprised that they're more awaits than async

[twilit grotto]

how so

you probably use the function more than once

[earnest wing]

plus, async iterators and context managers are far less common

[stark fable]

why would there be more asyncs than awaits

you would need to have a large number of async functions with no awaits in them

enough to balance out the many, many async functions that will contain multiple awaits

(well, empty async functions and also async for/async with since i think those exist, but they're not going to be ridiculously common)

[hot crypt]

apis

[plucky wedge]

What does nonlocal do?

[astral rover]

Allow for closures

https://youtu.be/E9wS6LdXM8Y I thought explained it quite well when

[hard spoke]

closures that assign variables in their environment, specifically - I think closures in general are any functions that can capture their environment, and in Python any function can do that

[floral meteor]

Edit: I used default kwarg to eliminate the need to init a

[floral meteor]

In this channel we use lambda more often than we drink water

[simple crystal]

!e

import *

[night quarry]

@simple crystal :x: Your eval job has completed with return code 1.

001 |   File "<string>", line 1
002 |     import *
003 |            ^
004 | SyntaxError: invalid syntax

[simple crystal]

what if I wanna

import everything

????

hmmm

[snow beacon]

That can be a project.

[simple crystal]

lol yes

someone in another server gave a good solution

import pkgutil
for module in [mod[1] for mod in pkgutil.iter_modules()][1:]:
    __import__(module)

[simple crystal]

!e

import pkgutil
for module in [mod[1] for mod in pkgutil.iter_modules()][1:]:
    __import__(module)

[night quarry]

@simple crystal :warning: Your eval job timed out or ran out of memory.

[No output]

[fossil estuary]

not sure if it's, technically esoteric but

!e ```py
x=123456789
print(f"{x:,}")

[night quarry]

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

123,456,789

[simple crystal]

...what does comma do to format specifier?

ohhhh thousands seperator

!e

x=123456789
print(f"{x:_}")

[night quarry]

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

123_456_789

[simple crystal]

!e

x=123456789
print(f"{x:_x}")

[night quarry]

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

75b_cd15

[fossil estuary]

what

oh, _x means hex@simple crystal

!e ```py
x=123456789
print(f"{x:x}")

[night quarry]

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

75bcd15

[fossil estuary]

Well, that's a good feature.

[snow beacon]

I started work on a guide for all of the stuff we talk about in this channel. Please feel free to write things for it. https://github.com/IFcoltransG/esoteric-python-guide/wiki

[viscid nymph]

oops

!d object.format

[night quarry]

`object.__format__`

object.__format__(self, format_spec)```
Called by the [`format()`](../library/functions.html#format "format") built-in function, and by extension, evaluation of [formatted string literals](lexical_analysis.html#f-strings) and the [`str.format()`](../library/stdtypes.html#str.format "str.format") method, to produce a “formatted” string representation of an object. The *format\_spec* argument is a string that contains a description of the formatting options desired. The interpretation of the *format\_spec* argument is up to the type implementing [`__format__()`](#object.__format__ "object.__format__"), however most classes will either delegate formatting to one of the built-in types, or use a similar formatting option syntax.

See [Format Specification Mini-Language](../library/string.html#formatspec) for a description of the standard formatting syntax.

The return value must be a string object.... [read more](https://docs.python.org/3/reference/datamodel.html#object.__format__)

[fossil estuary]

who pinged me :/

was it you?

@viscid nymph

[viscid nymph]

Yep

[fossil estuary]

oh, why

[simple crystal]

been learning javascript past couple days and damn if it doesn't allow for some esoteric ridiculous stuff except some of it actually get used in production code apparently lol

[west cipher]

being pythonic is for nerds

[floral meteor]

Why you have so many lines?

```py
Post it in a code block and I'll show you unpythonic
```

[west cipher]

async def sendAll(s):
    global MODE, STRIP_COLOR, SPEED, SEG_COUNT
    s = s.replace(' ', '').lower().replace('staticrgb', 'staticRGB').replace('rgb', 'RGB').replace('setspeed', 'setSpeed')
    DNIDH = {i[1]['name'].lower(): i[1]["websocket"] for i in clients.items() if "name" in i[1]}
    keyList = list(DNIDH.keys())
    for i in s.split(';'):
        j = [t.split(',') for t in i.split('>')]
        if len(j) == 1: j = [keyList + ["room"], j[0]] #Set to all devices if none are given
        for a in j[1]:
            if a.startswith("#"): a = "RGB|"+'|'.join(str(int(a[h+1:h+3], 16)) for h in range(0, 6, 2)) #Hex to RGB
            for d in j[0]:
                if d in keyList:
                    await DNIDH[d].send(a)
                elif d == "room": #process room LEDs
                    if a in ["white", "rainbow", "staticRGB", "off"]:
                        MODE = a
                    elif a.startswith("RGB"):
                        MODE = "RGB"
                        STRIP_COLOR = [int(c) for c in a.split('|')[1:]]
                    elif a.startswith("setSpeed"):
                        SPEED = float(a.split('|')[1])

i mean yea i could technically make this like, 3 lines (excluding the multiline ; trick)

but this is actual code im using and wrote not as a meme

@floral meteor

[floral meteor]

optimisation in progress...

why are you using global variables?

[west cipher]

well

i have a bunch of ESP32 LED drivers for various LED devices around my room

and i have a strip directly connected to the PI

being ran in another thread

i could make that strip connect to the websocket server and make a conversion layer and make it cleaner but i can't be bothered

[floral meteor]

why do you have space around your operators and after brackets?

[west cipher]

?

i mean, i didn't design this code to be compressed or anything

i made it to work, just my style is super stream of consiouness so it ended up looking like this

[floral meteor]

__69__ = {'MODE':MODE,'STRIP_COLOR':STRIP_COLOR,'SPEED':SPEED,'SEG_COUNT':SEG_COUNT}
async def sendAll(s):
    global __69__;s=s.replace(' ',str()).lower().replace('staticrgb', 'staticRGB').replace('rgb', 'RGB').replace('setspeed', 'setSpeed');__={'DNIDH':{i[1]['name'].lower():i[1]["websocket"]for i in clients.items()if"name"in i[1]},'keyList':list(__['DNIDH'].keys())}
    for i in s.split(';'):
        j=[t.split(',')for t in i.split('>')]
        if len(j)==1:j=[__['keyList']+["room"],j[0]]
        for a in j[1]:
            if a[0]=="#":a="RGB|"+'|'.join(str(int(a[h+1:h+3],16))for h in range(0,6,2))
            for d in j[0]:
                if d in __['keyList']:await __['DNIDH'][d].send(a)
                elif d=="room":
                    if a in["white","rainbow","staticRGB","off"]:__69__|={'MODE':a}
                    elif a[:3]="RGB":__69__|={'MODE':"RGB",'STRIP_COLOR':[int(c)for c in a.split('|')[1:]]}
                    else:a.startswith("setSpeed")and __69__.update({'SPEED':float(a.split('|')[1])})

[simple crystal]

beautiful

[west cipher]

y tho lol

[floral meteor]

i didn't even use walrii

[simple crystal]

also

...nice

[west cipher]

yea i would use walrus

[simple crystal]

NEEDS MORE WALRII 69

69ing walrii

[floral meteor]

alright i'll edit it

[simple crystal]

:=

=:

[west cipher]

i once did a giant command parser in 1 line

lemme try and find it

[floral meteor]

I used dunder 69 as a variable, are you proud?

[simple crystal]

proud isn't the word

[west cipher]

args = [a for group in args.split('|') if len(a:=[b for command in group.split(',') if len(b:=(((([g[0], g[1] if par[g[0]][0] == S else (float(('-' if g[1][0] == '-' else '') + d) if len(str(d:=(re.sub(r"([^0-9.])", '', g[1]).replace('.', '', max(0, (g[1]).count('.') - 1))))) > 0 else 1)]) if len(c) > 1 else [g[0], 'true' if par[g[0]][0] == S else 1.0])) if len(c) > 0 and (par[g[0]][0] == S or g[-1] != "false") else []) if len(g:=(c if ((c:=[(r.strip()) for r in (command.split('=', 1) if '=' in command else command.split(' ', 1))])[0] in par) else (([bind[c[0]], c[1]] if len(c) > 1 else [bind[c[0]]]) if (c[0] in bind) else []))) > 0 else []) > 0]) > 0][:3]```

here it is

heres the note that went with it:

#splts by '|', splits by ',', splits by '=' if '=' is in it, otherwise split by ' ', strip those, exit those args if any of those args are 'false', make sure args[0] (the filter name) is in the filter list, if that filter value is a float, turn it into a float by removing all but the last '.'. If any of the created lists have a length of 0, don't add it (applys to all layers)

[floral meteor]

! ! CEO — Today at 11:19 AM
__69__ = {'MODE':MODE,'STRIP_COLOR':STRIP_COLOR,'SPEED':SPEED,'SEG_COUNT':SEG_COUNT}
async def sendAll(s):
    global __69__;s=s.replace(' ',str()).lower().replace('staticrgb', 'staticRGB').replace('rgb', 'RGB').replace('setspeed', 'setSpeed');__={'DNIDH':{i[1]['name'].lower():i[1]["websocket"]for i in clients.items()if"name"in i[1]},'keyList':list(__['DNIDH'].keys())}
    for i in s.split(';'):
        len((j:=[t.split(',')for t in i.split('>')]))==1and(j:=[__['keyList']+["room"],j[0]])
        for a in j[1]:
            a[0]=="#"and(a:="RGB|".__add__('|'.join(str(int(a[h+1:h+3],16))for h in range(0,6,2))))
            for d in j[0]:
                d in __['keyList']and await __['DNIDH'][d].send(a)
                elif d=="room":
                    if a in["white","rainbow","staticRGB","off"]:__69__|={'MODE':a}
                    elif a[:3]="RGB":__69__|={'MODE':"RGB",'STRIP_COLOR':[int(c)for c in a.split('|')[1:]]}
                    else:a.startswith("setSpeed")and __69__.update({'SPEED':float(a.split('|')[1])})

[west cipher]

using ; to seperate lines is cheating imo

[floral meteor]

k, you've forced my hand, i will have to traumatise you now

[west cipher]

you can just
use a big ass print statment to define and run like everything

print(var1 := "yea", var2 := var1.strip(), await asdfasdf)

something like that

also you can prob avoid global by using globals()

[floral meteor]

globals()['.']={'MODE':MODE,'STRIP_COLOR':STRIP_COLOR,'SPEED':SPEED,'SEG_COUNT':SEG_COUNT}
async def sendAll(s):
 s,__=s.replace(' ',str()).lower().replace('staticrgb', 'staticRGB').replace('rgb', 'RGB').replace('setspeed', 'setSpeed'),{'DNIDH':{i[1]['name'].lower():i[1]["websocket"]for i in clients.items()if"name"in i[1]},'keyList':list(__['DNIDH'].keys())}
 for i in s.split(';'):
  len((j:=[t.split(',')for t in i.split('>')]))==1and(j:=[__['keyList']+["room"],j[0]])
  for a in j[1]:
   a[0]=="#"and(a:="RGB|".__add__('|'.join(str(int(a[h+1:h+3],16))for h in range(0,6,2))))
   for d in j[0]:
    d in __['keyList']and await __['DNIDH'][d].send(a)
    if d not in __['keylist']and d=="room":
     if a in["white","rainbow","staticRGB","off"]:globals()['.']|={'MODE':a}
     elif a[:3]="RGB":globals()['.']|={'MODE':"RGB",'STRIP_COLOR':[int(c)for c in a.split('|')[1:]]}
     else:a.startswith("setSpeed")and globals()['.'].update({'SPEED':float(a.split('|')[1])})
    globals().update(globals()['.'])

don't make me import ctypes and hook the ellipsis

[west cipher]

why would you need either of those

[simple crystal]

this channel isn't about needing

it's about wanting

[floral meteor]

there, massive indents can go die in a hole

[west cipher]

don't make me import ctypes and hook the ellipsis
how would that lower line count

[floral meteor]

the line count is the target

insanity is the target

[west cipher]

ok how is importing some random stuff insanity tho

[floral meteor]

you wanna find out?

[west cipher]

ok?

also if you want 1 line count you can just exec a string but thats boring

[floral meteor]

__import__('ctypes').py_object.from_address(id(...)+8).value=type(str(),(object,),dict(__call__=globals))
...()['.']={'MODE':MODE,'STRIP_COLOR':STRIP_COLOR,'SPEED':SPEED,'SEG_COUNT':SEG_COUNT}
@lambda f:...().update({'sendAll':f})
async def yeet(s):
 s,__=s.replace(' ',str()).lower().replace('staticrgb', 'staticRGB').replace('rgb', 'RGB').replace('setspeed', 'setSpeed'),{'DNIDH':{i[1]['name'].lower():i[1]["websocket"]for i in clients.items()if"name"in i[1]},'keyList':list(__['DNIDH'].keys())}
 for i in s.split(';'):
  len((j:=[t.split(',')for t in i.split('>')]))==1and(j:=[__['keyList']+["room"],j[0]])
  for a in j[1]:
   a[0]=="#"and(a:="RGB|".__add__('|'.join(str(int(a[h+1:h+3],16))for h in range(0,6,2))))
   for d in j[0]:
    d in __['keyList']and await __['DNIDH'][d].send(a)
    if d not in __['keylist']and d=="room":
     if a in["white","rainbow","staticRGB","off"]:...()['.']|={'MODE':a}
     elif a[:3]="RGB":...()['.']|={'MODE':"RGB",'STRIP_COLOR':[int(c)for c in a.split('|')[1:]]}
     else:a.startswith("setSpeed")and...()['.'].update({'SPEED':float(a.split('|')[1])})
 ...().update(...()['.'])

[west cipher]

wait

explain line 1

like

py_object.from_address(id(...)+8).value=type(str(),(object,),dict(__call__=globals))

that part

or

[floral meteor]

chilaxin seems to be missing :/

they made a module called fishhook

that's a quick implementation of that module

essentially, I'm telling EllipsisType to go die in a hole, and replacing it with an similar object that has a __call__ attribute, hence making ...() call globals instead of ellipsis

[west cipher]

oh

[floral meteor]

id(yay)+8 is the address of yay's class

now you know where EllipsisType lives, you can go torture them >XD

[west cipher]

… = globals

does this work?

[floral meteor]

no

you can't assign to constants, and builtin methods are read-only

hence why you need to fŭck around with addresses

[west cipher]

pain

i mean

you can do like

t = globals

t()['h'] = 5

h
>>> 5

[twilit grotto]

globals is a function

[west cipher]

yea but like

you can bind t to it

[floral meteor]

!e you could do

t=globals()
t.update({"2":3})
print(globals()['2']+2)

[night quarry]

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

5

[floral meteor]

!e but you can't do

t=globals()
t|={'2':3}
print(globals()['2']+2)

[night quarry]

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

5

[floral meteor]

wait wat

we should make an esoteric team for pyweek

and do the whole thing esoterically XD

[rugged sparrow]

@floral meteor the dict returned by globals is mutable

[floral meteor]

ik, but i thought augmented operator still instantiated a new instance from a merge.
apparantly it calls update instead

[rugged sparrow]

Makes sense that it calls update

Like list += does the same thing

[floral meteor]

!e ```py
f=type("",(),{'iadd':print})()
f += "Hello World!"
f += ...

[night quarry]

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

001 | Hello World!
002 | Traceback (most recent call last):
003 |   File "<string>", line 3, in <module>
004 | TypeError: unsupported operand type(s) for +=: 'NoneType' and 'ellipsis'

[floral meteor]

wait it still assigns the return?

O_O

[rugged sparrow]

Yes

[floral meteor]

@rugged sparrow how do i reset its recursion thingy?
nvm

[night quarry]

@floral meteor :warning: Your eval job timed out or ran out of memory.

[No output]

[floral meteor]

;-;

that'll turn it into a memory bomb won't it?

looks like I'll have to print as I go then

or set a limit

wow it's nomming my cpu only a little bit

!e ```py
a=[0]
@lambda cls:cls(1)
class fibonacci(int):
def repr(s):
a.append(s.class(int.add((lambda i:print(int(i))or i)(a[-2]),a[-1])))
return int.repr(s)

a+=[fibonacci]
print (a) # in console: >>> a

[night quarry]

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

001 | 0
002 | 1
003 | 1
004 | 2
005 | 3
006 | 5
007 | 8
008 | 13
009 | 21
010 | 34
011 | 55
... (truncated - too many lines)

Full output: too long to upload

[boreal slate]

import sys
p=print;s=sum;e=len
l=[s([int,ord][y.isalpha()](y)for y in x)for x in[[x for x in y]for y in sys.argv[1].split(",")]]
if e(l)==2:p(s(l))
elif e(l)==3:p(s(l[:2])*l[2])
else:
 _=e(l);p((s(l[:_-2])*l[_-2])//l[_-1])

How can i shorten this down

[viscid nymph]

I mean

You can shorten it by 2 characters

Because sum(gencomp) works

[boreal slate]

right, updated it

ok, made it shorter

import sys
p,s=print,sum
l=[s([int,ord][y.isalpha()](y)for y in x)for x in sys.argv[1].split(",")]
_=len(l)
p([0,1,s(l),s(l[:-1])*l[-1]][_]if _<4else(s(l[:_-2])*l[_-2])//l[_-1])

[simple crystal]

you can do p, s = print, sum cuz semicolons are cheating

[earnest wing]

yes!

[boreal slate]

there is nothing like "cheating" in #esoteric-python

[simple crystal]

semicolons aren't esoteric they are some nerd shit

unless you're stuffing the entire program onto one line there's no excuse, this ain't C++

[strange basin]

!e

assert "I" in "VOWELS"

[night quarry]

@strange basin :x: Your eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 1, in <module>
003 | AssertionError

[strange basin]

!e

assert "I" in "TEAM"

[night quarry]

@strange basin :x: Your eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 1, in <module>
003 | AssertionError

[strange basin]

Maybe time to go look for a new job ... 😦

[boreal slate]

it reduces your char count, \n is two chars, while ; is just one.

[simple crystal]

\n is one char 😤

p,s=print,sum is the same char count though

[thin trout]

It depends on the platform

Well

I guess it depends how you count

[simple crystal]

it's literally a character tho

when you aren't escaping it

calling it "\n" is like calling \ \\

but we call it \n because saying
is a character and
is only one character is unclear since we can't see

[tawny arch]

I think windows uses two characters for a new line

[simple crystal]

do you need to escape two characters for a new line in windows?

[tawny arch]

not when you're writing python

[proper vault]

it is autoconverted, but ye, \r\n is the correct sequence

[tawny arch]

maybe using a semicolon saves disk space?

[simple crystal]

exactly

but yeah it's been \r\n on windows

though that's changing https://devblogs.microsoft.com/commandline/extended-eol-in-notepad/

as far as I can tell because it's stupid

[proper vault]

I mean, literally everything other than unix and mac text files is \r\n

HTTP, most other network protocols

gzipped text files

[simple crystal]

lol [-6] eh

carriage return makes no sense in most formats imo

I use it to reprint to same console line but \n implies carriage return

why would cursor continue printing where previous line ended?

[proper vault]

\n means one column down and \r means return to start of line

[simple crystal]

on a typewriter, sure

[proper vault]

teletypes and terminals also do this

[simple crystal]

but there's no "word wrap" character

[proper vault]

though now there are flags you can set to change how this is handled

one thing unix definitely does right is trailing newline

[simple crystal]

and those flags are by default set to return to beginning of new line

[proper vault]

yeah, since that is how unix decided to do text files

[simple crystal]

because it makes sense

you don't need to emulate a typewriter

[proper vault]

emulating typewriters is all teletypes did

[simple crystal]

yes and that is now defunct

[proper vault]

ye, but now the characters have 2 distinct meanings

which is imo quite silly

[simple crystal]

and carriage return remaisn useful

for returnign to origin ont he same line

I use \r all the time

[proper vault]

or a newline on macs

its not about being useful, its about newlines working differently because unix decided they need to be special

[simple crystal]

it should be about being useful

[proper vault]

as in the problem isn't some ascii characters becomes useless

[simple crystal]

if something isn't useful it's just art

[proper vault]

most of the <32 codes are useless

the problem is that there isn't a way to do a newline sequence in ascii that will work on all computers implementing ascii rendering

[simple crystal]

what computers do we need to worry about that need to produce human readable text that can't handle a newline?

if \n is just vertical to any computer it's a "how quaint" moment

I get that ascii is based on typing stuff but it's far more complex to handle dates correctly and simple computers generally get programmed to do that fine

[proper vault]

backwards compatability is a always a concern. And it doesn't seem like an issue until something requires the networking \r\n form, and your editor is really convinced it needs to save \n. Or the fact you now have to keep track of which files are text and which are not text on linux in order to not produce non standard zip archives or incorrect HTTP requests.

I did only run into issues about 3 times, but that 3 more than really should have happened considering there is a standard on this

[simple crystal]

\n being the standard, right?

[proper vault]

\r\n. HTTP, zip, .. all use CRLF

[simple crystal]

but if backwards compatibility and standards really mattered we wouldn't have two different configurations of friggin usb-c

[proper vault]

its really just text files being silly

the networking world agrees on \r\n pretty consistently

[simple crystal]

so they add an unnecessary character in case an ancient computer has to process text data

[proper vault]

yes, you can't change standards

at least not for something as silly as a newline sequence

[simple crystal]

poor computer would just move the cursor vertically and htere'd be nothing behind the cursor

but it'd be suspended horizontally

what a conundrum

[proper vault]

yeah, that is how the character worked when ascii was defined

[simple crystal]

despite this, we have new ports every second tuesday we're supposed to be able to plug a quarter inch into somehow

standards do change

[proper vault]

not networking ones, unless its something really severe

[simple crystal]

cuz the internet is a gross behemoth but yeah I'm just complaining

but back on topic python files generally can be assumed to just use \n

and that's a visual representation of an escaped single character

[proper vault]

ye, python does handle all newline sequences regardless of OS

so yeah, it is definitely possible to codegolf with \n even on windows

[simple crystal]

the only time I've used semicolons in python is when I try to exec a whole python file and insert ;*code* on lines that can handle that to run that code a bunch

[proper vault]

they are useful for golfing since they let you avoid indentation

while 1:print('a');print('b')

[simple crystal]

!e

while 1:print('a');print('b')

[night quarry]

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

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

Full output: too long to upload

[simple crystal]

hehe

trufax

[thin trout]

You can also use a tuple, but that’s longer

[simple crystal]

I am more impressed by code that avoids the semis, personally

unless it's one line

if it's one line I like semis

more than one line, why bother?

[thin trout]

!e

while 1:(print("a"),print("b"))```

[proper vault]

onelines without semis are best imo

[night quarry]

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

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

Full output: too long to upload

[thin trout]

No semi colon required

[simple crystal]

yeha agreed

usin tuples like that is good

and you can use ternaries and comprehensions and so on

[proper vault]

comprehensions can also be used for local variables

[simple crystal]

with a walrus?

[proper vault]

without walrus

def fib(n):
    return max(
        d[1]
        for d in [{}]
        for d[0], d[1] in [(0, 1)]
        for _ in range(n)
        for d[0], d[1] in [[d[1], d[0] + d[1]]]
    )

source: nedbat twitter

[simple crystal]

oh wow what does for d in [{}] do

oh just iterates once

as a dictionary

I can't tell if this is a good job interview fib solution or it'd make them mad

neat tho

[sick hound]

hi

[fossil estuary]

hi

[boreal slate]

can i get this any shorter?

[fossil estuary]

use p,s=print,sum

@boreal slate

[boreal slate]

won't matter, count will remain the same

[frigid wharf]

p([0,1,s(l),s(l[:-1])*l[-1]][_]if _<4else(s(l[:_-2])*l[_-2])//l[_-1])
``` should make it shorter

[viscid nymph]

The 4else part probably raises an error

[proper vault]

it doesn't

the highligher just doesn't like it

[snow beacon]

y.isalpha() might work as y>"9" depending on your goal.

[thin trout]

Fun fact it would on micropython

[boreal slate]

yes that would work!

[viscid nymph]

!e

while not print("a"):print("b")

[night quarry]

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

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

Full output: too long to upload

[fossil estuary]

lol

too long

[viscid nymph]

!e

print(len('while not print("a"):print("b")'))
print(len('while 1:(print("a"),print("b"))'))

[night quarry]

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

001 | 31
002 | 31

[fossil estuary]

how so

oh wait

nvm

that's the length of the string 😅

[viscid nymph]

I would say while1

but why does this not work

!e

while1:print('lol')
print('huh')

[night quarry]

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

001 | lol
002 | huh

[viscid nymph]

wait nvm

[viscid nymph]

it's just a type hint

[vague hearth]

class DunderOverload(metaclass=TypeOverload):

    def __init__(self, value: int):
        self.value = value

    def __add__(self, other: int):
        print("int")
        return self.value + other

    def __add__(self, other: str):
        print("str")
        return self.value + int(other)

    def __add__(self, other):
        print("Any")
        return NotImplemented


inst = DunderOverload(10)
print(inst + 2)
print(inst + "3")
print(inst + [])

int
12
str
13
Any
TypeError: unsupported operand type(s) for +: '__getattribute__' and 'list'```a clean way to overload methods https://paste.pythondiscord.com/epupesecih.py

[twilit mountain]

_=()==()
__=(_+_)+(_+_)
___=((_+_)+(_+_)+(_+_)+(_+_)*(_+_)+(_+_)+(_+_)+(_+_))
____="%c"*(_+_+_+_+_)%(((_+_)*((___*__+(__+__))//(_+_))),((_)*((___*__+(__+__)-(_+_+_)))),((_+_)*((___*__+(__+__)+(__))//(_+_))),((_+_)*((___*__+(__+__)+(__))//(_+_))),((_)*((___*__+(__+__)+(__+_+_+_)))))+" "+"%c"*(_+_+_+_+_)%(((_)*((___*__+(__+__)+(__+__+__+_+_+_)))),((_)*((___*__+(__+__)+(__)+(_+_+_)))),((_+_)*((___*__+(__+__)+(__)+(_+_+_)+(_+_+_))//(_+_))),((_+_)*((___*__+(__+__)+(__))//(_+_))),((_+_)*((___*__+(__+__)-(__))//(_+_))))
print(____) 

Anyone here did anything like this?

[fossil estuary]

yea

me

wait a fukn

that prints hello world

doesnt it

I made it, I should know

!e

_=()==()
__=(_+_)+(_+_)
___=((_+_)+(_+_)+(_+_)+(_+_)*(_+_)+(_+_)+(_+_)+(_+_))
____="%c"*(_+_+_+_+_)%(((_+_)*((___*__+(__+__))//(_+_))),((_)*((___*__+(__+__)-(_+_+_)))),((_+_)*((___*__+(__+__)+(__))//(_+_))),((_+_)*((___*__+(__+__)+(__))//(_+_))),((_)*((___*__+(__+__)+(__+_+_+_)))))+" "+"%c"*(_+_+_+_+_)%(((_)*((___*__+(__+__)+(__+__+__+_+_+_)))),((_)*((___*__+(__+__)+(__)+(_+_+_)))),((_+_)*((___*__+(__+__)+(__)+(_+_+_)+(_+_+_))//(_+_))),((_+_)*((___*__+(__+__)+(__))//(_+_))),((_+_)*((___*__+(__+__)-(__))//(_+_))))
print(____)

[night quarry]

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

HELLO WORLD

[fossil estuary]

MF

@twilit mountain

#esoteric-python message

[twilit mountain]

Oh

[fossil estuary]

🙃 lol

[twilit mountain]

😂u only made it

I copied it from yr DM only

[fossil estuary]

loool

you've come a long way.

[terse mortar]

Lol ik

[ornate dew]

My physics test is in 5 hours, I know nothing, and this is what I'd do rather than study:

digits = {'0':'1', '1':'2', '2':'3', '3':'4', '4':'5', '5':'6', '6':'7', '7':'8', '8':'9', '9':'0'}
num = '0'
while num != ['1', '0', '0']:
    num = list(num)
    lastdigit = num.pop()
    increment = digits[lastdigit]
    if lastdigit == '9':
        try:
            nextlastdigit = num.pop()
            if nextlastdigit == '9':
                num = ['1', '0', '0']
            else:
                nextincrement = digits[nextlastdigit]
                num.append(nextincrement)
                num.append(increment)
        except IndexError:
            num = ['1', '0']
    else:
        num.append(increment)
    
    for digit in num:
        print(digit, end='')
    
    print()

Maybe one day I'll be masochistic enough to make it recursive so it works for any upper limit but that day is not today xD

[alpine flower]

!e

def increment(num: list) -> list:
  if num[-1] == 9:
    if len(num) > 1:
      num = increment(num[:-1]) + [0]
  else:
    num[-1] += 1
  return num

num = [1, 9, 9, 9]
print(increment(num))```

[night quarry]

@alpine flower :white_check_mark: Your eval job has completed with return code 0.

[2, 0, 0, 0]

[alpine flower]

@ornate dew not really the same but there's no limit

!e

def increment(num: list) -> list:
  if num[-1] == 9:
    if len(num) > 1:
      num = increment(num[:-1]) + [0]
  else:
    num[-1] += 1
  return num
print(increment([9 for i in range(10 ** 60)]))```

[night quarry]

@alpine flower :warning: Your eval job timed out or ran out of memory.

[No output]

[alpine flower]

haven't seen that one before

[ornate dew]

Well the idea was to not to use any integers at all lmao

[alpine flower]

finnne

[ornate dew]

Also what happened to my syntax highlighting?

[viscid nymph]

!code

[night quarry]

Here's how to format Python code on Discord:

```py
print('Hello world!')
```

These are backticks, not quotes. Check this out if you can't find the backtick key.

[ornate dew]

Oh xD

[alpine flower]

!e

digits = {'0':'1', '1':'2', '2':'3', '3':'4', '4':'5', '5':'6', '6':'7', '7':'8', '8':'9', '9':'0'}
def increment(num: list) -> list:
  if num[-1] == 9:
    if len(num) > 1:
      num = increment(num[:-1]) + ['0']
  else:
    num[-1] = digits[num[-1]]
  return num
print(increment(['9', '9', '9']))
```\

[night quarry]

@alpine flower :white_check_mark: Your eval job has completed with return code 0.

['9', '9', '0']

[ornate dew]

Don't mean to nitpick, but this still uses -1 to index lol

[alpine flower]

that's easy to fix tho eval("-1")

[ornate dew]

Or int('-1') lmao

But that would be cheating

[alpine flower]

recursion hurts my head tho idk why it isn't working

[ornate dew]

No variables of datatype int and nothing that returns int

[ornate dew]

No cap, same here

[simple crystal]

....can you assign exec with a walrus in an f string with this method?

hmmm

you'd need the f string syntax nvm

[floral meteor]

!e ```py
def increment(num: list[int]) -> list:
(num[-1]==9)and len(num)>1 and(num:=increment(num[:-1])+[0])or num.setitem(-1,num[-1]+1)
return num

num = [1, 9, 9, 9]
print(increment(num))

[night quarry]

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

[2, 0, 0, 0]

[snow beacon]

You forgot the quote marks around the 9.

[floral meteor]

I forgot all the quote marks

[molten wadi]

been codiing with python for 5 years and never knew you could do increment(num: list[int]) -> list

within a function name

crazy

[wide laurel]

!w

!e

[night quarry]

Command Help

!eval [code]
Can also use: e

*Run Python code and get the results.

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

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

[wide laurel]

!eval

def main():
  exit()
main()

[night quarry]

@wide laurel :x: Your eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 3, in <module>
003 |   File "<string>", line 2, in main
004 | NameError: name 'exit' is not defined

[wide laurel]

oh

idk then

[floral meteor]

@wide laurel #bot-commands if your trying to break the bot XD

[wide laurel]

ok lol

[molten wadi]

lmao

[floral meteor]

!e my code after the coffee: ```py
print(420)

[night quarry]

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

420

[floral meteor]

does the same thing

!e post moved: "my code before the coffee"```py
import('ctypes').py_object.from_address(id(...)+8).value=type("wrapper",(object,),{'call':lambda s,f:f()})
@int
@...
class n:int=lambda s:69
@...
class cout:lshift=print
@...
class main:init=lambda s: cout << int.mul((n//10)*10,(n%10)-(True<<10**False))

[night quarry]

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

420

[floral meteor]

hey chilaxin

[rugged sparrow]

!e ```py
from ctypes import (
pythonapi, POINTER, byref,
c_int, c_wchar_p
)
import os

def rerun(*new_flags, keep_old=True):
_argv = POINTER(c_wchar_p)()
_argc = c_int()
pythonapi.Py_GetArgcArgv(byref(_argc), byref(_argv))
orig_argv = _argv[:_argc.value]
if keep_old:
orig_argv[1:1] = new_flags
orig_argv[-1] = orig_argv[-1].replace('x = 0', 'x = 0x11')
os.execv(orig_argv[0], orig_argv)
else:
os.execv(orig_argv[0], (orig_argv[0],) + new_flags)
x = 0
if x == 0x11:
print('here')
else:
rerun()```

[night quarry]

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

here

[rugged sparrow]

magic ^

@floral meteor i live in this channel lol

[floral meteor]

same XD

imma see if i can use this cursed class syntax to make brainfuck

[boreal slate]

import sys
s=sum
l=[s([int,ord][y>"9"](y)for y in x)for x in sys.argv[1].split(",")]
_=len(l)
print([0,1,s(l),s(l[:-1])*l[-1]][_]if _<4else(s(l[:_-2])*l[_-2])//l[_-1])
``` anyway i can get rid of the if statement, probably that should make this shorter

i can share the annotated version of this if you want

[floral meteor]

!e ```py
import('ctypes').py_object.from_address(id(...)+8).value=type("wrapper",(object,),{'call':lambda s,a,**k:lambda f:f(a,**k),'repr':lambda s:'...'})
@...("++++++++++[>++++++++++<-]>--.-.+++.")
class output:
@str
@...()
class input(str):''
def init(s,c):
if not c:return''
@list
@...()
class a:list=lambda list:[0]
@...(a)
class a(list):
"""This is a list, only infinite."""
str=lambda s:list.str([...]+s+[...])
@int
@...()
class offset:int=lambda int:0
def getitem(self,item):
if(item:=item+self.offset)<=0:
self.offset-=item
@int
@...(0)
class item(int):int=lambda int:0
try:return list.getitem(self,item)
except IndexError:self.extend([0](item-len(self)+1))
return list.getitem(self,item)
def setitem(self,item):
if(item:=item+self.offset)<=0:
self.offset-=item
@int
@...(0)
class item(int):int=lambda int:0
try:list.getitem(self,item)
except IndexError:self.extend([0](item-len(self)+1))
return list.setitem(self,item)
@int
@...(0)
class i(int):0
@int
@...(0)
class p(int):-1
while(p:=p+1)<len(c):
if p<0:raise SyntaxError
@str
@...()
class t():str=lambda s:c[p]
if t in['+','-']:
@lambda n:a.setitem[i,n]
@int
@...()
class _:int=lambda s:(a[i]+1-2*(t=='-'))%256
elif t in['<','>']:
@int
@...(i+1-2*(this=='>'))
class i(int):...
elif t=='.':
@lambda b:print(end=chr(int(b)))
@...()
class cout:int=a.pop()
elif(t=='['and a[i]==0)or(t==']'and a[i]!=0):
@int
@...()
class :
def int(s):
nonlocal p;p=p+1-2*(t==']')
@int
@...(0)
class (int):0
while >=0:
if 0>(p:=p+1-2*(t==']')) or p>len(c):raise SyntaxError
c[p]in['[',']']and(:=+1-2*(c[p]!=t))
return p
@int
@...()
class p(int):...

[night quarry]

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

001 | <string>:2: SyntaxWarning: 'ellipsis' object is not callable; perhaps you missed a comma?
002 | <string>:5: SyntaxWarning: 'ellipsis' object is not callable; perhaps you missed a comma?
003 | <string>:10: SyntaxWarning: 'ellipsis' object is not callable; perhaps you missed a comma?
004 | <string>:12: SyntaxWarning: 'ellipsis' object is not callable; perhaps you missed a comma?
005 | <string>:17: SyntaxWarning: 'ellipsis' object is not callable; perhaps you missed a comma?
006 | <string>:23: SyntaxWarning: 'ellipsis' object is not callable; perhaps you missed a comma?
007 | <string>:32: SyntaxWarning: 'ellipsis' object is not callable; perhaps you missed a comma?
008 | <string>:38: SyntaxWarning: 'ellipsis' object is not callable; perhaps you missed a comma?
009 | <string>:41: SyntaxWarning: 'ellipsis' object is not callable; perhaps you missed a comma?
010 | <string>:46: SyntaxWarning: 'ellipsis' object is not callable; perhaps you missed a comma?
011 | <string>:51: SyntaxWarning
... (truncated - too long, too many lines)

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

[floral meteor]

bugger it, i'll work on my cursedutils module

[ornate dew]

Boys I was wrong, today was the day

digits = {'0':'1', '1':'2', '2':'3', '3':'4', '4':'5', '5':'6', '6':'7', '7':'8', '8':'9', '9':'0'}

def increment(num):
    try:
        digit = num.pop()
        if digit == '9':
            increment(num)
        num.append(digits[digit])
    except IndexError:
        num.append('1')
    return num

num = ['0']
while num != ['1', '0', '0', '0']:
    num = increment(num)
    for i in num:
        print(i, end='')
    print()

This was much shorter, and dare I say, simpler lmao

It's finally short enough to fit on vsinder

[plain imp]

wtf is this lmao

[vestal solstice]

you treat increment as "void" inconsistently

digits = {'0':'1', '1':'2', '2':'3', '3':'4', '4':'5', '5':'6', '6':'7', '7':'8', '8':'9', '9':'0'}

def increment(num):
    try:
        digit = num.pop()
        if digit == '9':
            increment(num)
        num.append(digits[digit])
    except IndexError:
        num.append('1')


num = ['0']
while num != ['1', '0', '0', '0']:
    increment(num)
    for i in num:
        print(i, end='')
    print()

[alpine flower]

that's partially my fault I have a habit of naming global + local vars the same thing and treating them as local vars

is there a way to use contextlib.suppress without a context manager?

[simple crystal]

!e

import gc

f = gc.get_referents(list.__dict__)[0]
f.pop("append")
a = list()
print(a)
a.append()
print(a)
print(2)

[night quarry]

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

[]

[simple crystal]

segmentation fault

when I try to debug to see what happens the debugger breaks

and then it will raise a "list has no attribute append" once

then it goes back to segmentation fault

I want an exception to get raised if I try to append...

oh fishhook works

I will figure out why

[ornate dew]

Not quite, the return is outside the try-catch

So it should return every time

I think 😂

!e

digits = {'0':'1', '1':'2', '2':'3', '3':'4', '4':'5', '5':'6', '6':'7', '7':'8', '8':'9', '9':'0'}

def increment(num):
    try:
        digit = num.pop()
        if digit == '9':
            increment(num)
        num.append(digits[digit])
    except IndexError:
        num.append('1')
    return num

num = ['0']
while num != ['1', '0', '0', '0']:
    num = increment(num)
    for i in num:
        print(i, end='')
    print() 

[night quarry]

@ornate dew :white_check_mark: Your eval job has completed with return code 0.

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

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

[ornate dew]

Seems to work lmao

[ornate dew]

That shouldn't matter, unless you explicitly declare a variable as global like global var within a context, it should be specific to they context

I just happened to forget that python passes by object reference, so both variables point to the same object in memory, meaning I don't really have to return it lmao

[viscid nymph]

Maybe that should be a list

[alpine flower]

!e

digits = [str(ord(i) - ord('a')) for i in "abcdefghij"]

def increment(num: list[str]) -> None:
    try:
        digit = num.pop()
        if digit == str(ord('j') - ord('a')):
            increment(num)
        num.append(digits[int(digit) + ord('a') if int(digit) < ord('j') - ord('a') else ord('a') - ord('a')])
    except IndexError:
        num.insert(ord('a') - ord('a'), str(ord('b') - ord('a')))
    return num

print(increment(["1", "9", "9"]))```

[night quarry]

@alpine flower :white_check_mark: Your eval job has completed with return code 0.

['1', '0', '0']

[alpine flower]

my point is it's really hard to do it without a dict + not having any sort of numbers in the code

I can't figure out how to get around the indices tho

[ornate dew]

This is like the opposite problem, you can use int datatype, but no literal numbers in the code

I was doing, you can't use anything that returns an int, but you can use numbers as strings

I'm pretty sure it's impossible to do both :P

[alpine flower]

ya there's probably some backend way to do it, I just have absolutely no idea how to do it

[sick hound]

!e

!e
print("Hello world!")

[night quarry]

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

Hello world!

[sick hound]

hell yea

!e

num2 = input("Enter a number: ")
result = float(num1) + float(num2)

print(result)```

[night quarry]

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

001 | Enter a number: Traceback (most recent call last):
002 |   File "<string>", line 1, in <module>
003 | EOFError: EOF when reading a line

[naive roost]

as you can see (even though you should probably play on #bot-commands), input doesn't work in this sandboxed interpreter

[runic ruin]

_B='108.0';_A='34.0';String=str;Integer=int;Float=float;Run_Code=exec;Character=chr;Character_Number_1=Character(Integer(Float(String('112.0'))));Character_Number_2=Character(Integer(Float(String('114.0'))));Character_Number_3=Character(Integer(Float(String('105.0'))));Character_Number_4=Character(Integer(Float(String('110.0'))));Character_Number_5=Character(Integer(Float(String('116.0'))));Character_Number_6=Character(Integer(Float(String('40.0'))));Character_Number_7=Character(Integer(Float(String(_A))));Character_Number_8=Character(Integer(Float(String('72.0'))));Character_Number_9=Character(Integer(Float(String('101.0'))));Character_Number_10=Character(Integer(Float(String(_B))));Character_Number_11=Character(Integer(Float(String(_B))));Character_Number_12=Character(Integer(Float(String('111.0'))));Character_Number_13=Character(Integer(Float(String(_A))));Character_Number_14=Character(Integer(Float(String('41.0'))));Final_String=String(Character_Number_1+Character_Number_2+Character_Number_3+Character_Number_4+Character_Number_5+Character_Number_6+Character_Number_7+Character_Number_8+Character_Number_9+Character_Number_10+Character_Number_11+Character_Number_12+Character_Number_13+Character_Number_14);Run_Code(Final_String)
```This took me seventeen years to code, and it is finally perfect.

[night quarry]

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

Hello

[sick hound]

it's beautiful

!e

_=((()==())+(()==()))
__=(((_<<_)<<_)*_)
__import__(('c%'[::(([]!=[])-(()==()))])*((_<<_)+(()==()))%((__+(((_<<_)*_)+((_<<_)+((_*_)+(_+(()==())))))),(__+(((_<<_)*_)+((_<<_)+((_*_)+(_+(()==())))))),(__+(((_<<_)<<_)+(_<<_))),(__+(((_<<_)<<_)+((_*_)+(()==())))),(__+(((_<<_)<<_)+((_<<_)+(_*_)))),(__+(((_<<_)<<_)+((_<<_)+(_*_)))),(__+(((_<<_)<<_)+((_<<_)+((_*_)+(_+(()==())))))),(__+(((_<<_)*_)+((_<<_)+((_*_)+(_+(()==())))))),(__+(((_<<_)*_)+((_<<_)+((_*_)+(_+(()==())))))))) 

[night quarry]

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

Hello world!

[sick hound]

this is one of mine

[late talon]

just curious what is you favorite VM type

[late talon]

also what type of VM software

[sick hound]

qemu/kvm

I use manjaro linux

passthrough a 6800 XT into windows

[twilit grotto]

only pyhton

[edgy fiber]

oh

I thought I wouldn't get an answer, but thanks.

[late talon]

sounds like Muthar anas(someordinarygamers)

[sick hound]

i know lol

[late talon]

Just curious do you use Archlinux and Do you Use oracle virtual box or VMWARE

[sick hound]

I use an arch distro, manjaro seemed good for me since I do a lot of coding

[ornate dew]

You're a psychopath

[grave rover]

!e ```py
f=(lambda n:(o:=n//2)>n or''.join(0*(m:=x/n3)(l:=y/n3)or" #"[(mm+ll-1)**3-mm*l**3<0]+"\n"[x<o-1:]for y in range(o,-o,-1)for x in range(-o,o)))

print(f(20))```

[night quarry]

@grave rover :white_check_mark: Your eval job has completed with return code 0.

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

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

[grave rover]

(lambda n:(lambda o=int(.5*n//1):[([(lambda m=(x/n*3),l=(y/n*3):print([" ","#"][(m**2+l**2-1)**3-m**2*l**3<0],end=""))()for x in range(-o,o)],print())for y in range(o,-o,-1)])())

(lambda n:(lambda o:"".join(" #"[(m**2+l**2-1)**3-m**2*l**3<0]+"\n"*(o-x==1)for y in range(o,-o,-1)for x in range(-o,o)for m,l in[[(x/n*3),(y/n*3)]]))(n//2))(10)

(lambda n:(o:=int(n/2))>n or''.join((m:=x/n*3)*(l:=y/n*3)*0 or" #"[(m**2+l**2-1)**3-m**2*l**3<0]+"\n"[x<o-1:]for y in range(o,-o,-1)for x in range(-o,o)))

(lambda n:(o:=n//2)>n or''.join(0*(m:=x/n*3)*(l:=y/n*3)or" #"[(m*m+l*l-1)**3-m*m*l**3<0]+"\n"[x<o-1:]for y in range(o,-o,-1)for x in range(-o,o)))

``` kinda satisfying to look at how this one evolved over time

[fossil estuary]

!e
f=(lambda n:(o:=n//2)>n or''.join(0*(m:=x/n3)(l:=y/n3)or" #"[(mm+ll-1)**3-mm*l**3<0]+"\n"[x<o-1:]for y in range(o,-o,-1)for x in range(-o,o)))

print(f(12))

[night quarry]

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

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

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

[plucky wedge]

CODEBLOCK REEEEEEEEEEEEEE

like dis

[fossil estuary]

I know

I like not using them because it's funny

@plucky wedge

[dusk root]

Someone have idea how to run http://www.onelinerizer.com/ in python3?
I got:

Traceback (most recent call last):
  File ".../oneline.py", line 22, in <module>
    onelinerize(code)
  File "...\lib\site-packages\onelinerizer\onelinerizer.py", line 804, in onelinerize
    return get_init_code(t, table)
  File "...\lib\site-packages\onelinerizer\onelinerizer.py", line 70, in get_init_code
    output = Namespace(table).many_to_one(tree.body)
  File "...\lib\site-packages\onelinerizer\onelinerizer.py", line 214, in many_to_one
    return reduce(
NameError: name 'reduce' is not defined

[grave rover]

oh I was writing something like that for py3.8

[last locust]

Fwiw that specific error can be fixed by using return __import__('functools').reduce(...)

[dusk root]

did you finished it?

[fossil estuary]

mine works. ;)

[grave rover]

been a while since I really worked on it lol

not 100% working yet sadly

but it can do most things iirc

[dusk root]

I just wanted it to do exec

[grave rover]

I mean

# Generated by Mart Obfuscator

a=lambda b:b<([]==[])<<(()>=())and()==()or a(b-([]==[]))+a(b-((()>=())<<(()==())))
``` here's fibonacci

>>> x = [6, 4, 3, 2, 9, 1]
>>> sort.quick_sort(x, 0, len(x)-1)
(None, (None, -1, (None, (None, -1, -1), (None, -1, -1))), (None, -1, (None, -1, -1)))
>>> x
[1, 2, 3, 4, 6, 9]
```quicksort and selectionsort

tic tac toe is still a WIP sadly, the number obfuscator seems to actually decide whether the loop runs or exits, and even then it seems there's an issue somewhere causing an indexerror

[eternal prawn]

Not sure if this belongs here but I managed to achieve something I'm really proud of, but it has no real use other than making my code a bit prettier with type hinting.
I made a type so that I can make numpy arrays with a specified length. Kinda like a generic, so I can type hint Vector[6] if I'm expecting a (6,) shape numpy array.

This is what the syntax looks like

x = Vector[3]([1, 2, 3])

And this will drop an error, because of the shape mismatch:

y = Vector[4]([1, 2, 3])

Now I can type hint:

def foo(x: Vector[2]) -> Vector[2]:
    return x+2

If someone is interested I would be happy to receive criticism: https://gist.github.com/Speterius/69026b1376a671f9c1715fb2eb813b29

[rugged sparrow]

@eternal prawn noice. If you want to make it a bit simpler, python has __class_getitem__

!e ```py
class a:
def class_getitem(cls, arg):print(arg)

a[0]```

[night quarry]

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

0

[dull anvil]

factors = lambda num: [int(j) for i in map(lambda x: [x] if x == num else [x, num / x], list(filter(lambda x: num % x == 0, range(1, int(num**0.5))))) for j in i]```

nice way of calculating factors

aint it

[late talon]

what is lamada

[stable hornet]

read https://realpython.com/python-lambda/

[viscid nymph]

and use a metaclass instead of a decorator

[astral rover]

You don't need a meta class here

You could use init subclass

And then not make it a function that returns a class

[viscid nymph]

I know but the decorator was interesting

[floral meteor]

alright, going into python from matlab, i really missed the struct object.
Well with some globals hooking and a cursed struct class I will make struct work in python.

!e ```py
import ctypes
#hackery

class struct:
def init(self,**k):[setattr(self,this,that)for this,that in k.items()]
def getattr(self,item):
if not hasattr(self,item):setattr(self,item,self.class())
return object.getattr(self,item)
def call(self,f):setattr(self,f.name,f);return f
class _hooked(type):
@lambda c:c()
class __sentinal:...
gulag={}
def del(self,other=__sentinal):
if other is __sentinal
for victim, orig in gulag.items():ctypes.py_object.from_address(id(victim)+8).value=orig
elif other in gulag:ctypes.py_object.from_address(id(other)+8).value=gulag[other]
class curse(metaclass=_hooked):
"""example usages:

@curse(...)
def call(self,*a,**k):
def wrapper(f):
return f(*a,**k)

curse(None,'call',lambda s,*a:s)
"""
def new(self,other,name='',f=None):
if f is None:return lambda f:self(other,name or f.name,f)
else:return self(other,name or f.name,f)
def init(self,other,name='',f=None):
if other not in self.class.gulag:self.class.gulag|={other,type(other)}
hack=ctypes.py_object.from_address(id(other)+8)
try:hack.value=type(hack.value.name,(hack.value,),{name:f})
except:hack.value=type(hack.value.name,(),{this:getattr(hack.value,this)for this in dir(hack.value)}|{name:f})
@curse(globals())
def missing(self, key):
try:return getattr(builtins,key)
except:self[key]=struct(name=key,doc=None);return self[key]
cout.lshift=print
cout << "Hello World!"

oh whoops i wasnt finished lol

!e ok first i perfectses the class

class struct:
 def __init__(self,**k):[setattr(self,this,that)for this,that in k.items()]
 def __getattr__(self,item):
  if not hasattr(self,item):setattr(self,item,self.__class__())
  return object.__getattr__(self,item)
 def __call__(self,f):setattr(self,f.__name__,f);return f
cout, output = struct(__lshift__=print), struct()
output.data.x = 3
output.data.y = 4
@cout << f(output.data)
def _(data):
  return data.x * data.y

I have no clue, it shouldn't be recursionings

it should be workings

[floral meteor]

AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA

[alpine flower]

"".join(['A' for i in range(43)])* smh

[vestal solstice]

I believe you meant f"{11478986304257167078711585130701756989053283163632459947822157877600493794161163746264989691490697448857580612687974391859056618193877774016661073255152478229773002314437327253923119931013370082604311376279371017487563790178827046139928382510691977723156789183826289319143052391855134253864827181651765249703991141830847575261485010803545942957275674094024174250:X}"

[dull anvil]

'A'*43

[unique glacier]

🥳

[unique glacier]

what is this

[floral meteor]

i was trying to make a matlab object

[unique glacier]

wow,It's cool👍

I trying to make an AI project. But it's not working, Please help guys to make it.

If yo want a make money ? try this link https://www.fiverr.com/

[snow beacon]

This channel tends to solve problems in what can only be described as the worst way possible. Maybe try #❓｜how-to-get-help if you want to get good code.

[grave rover]

if __name__ == "__main__":
    x = 10

    with open("/proc/self/mem", "r+b") as mem:
        mem.seek(id(x))
        obj = PyLongObject()
        obj.set_bytes(mem.read(obj.get_size()))
        print(obj.get_named_values())
        # => PyLongObject(ob_refcnt=74, ob_type=140678365892192, ob_size=1, ob_digit=10)

``` repurposing some old code

[fluid tree]

... why

[grave rover]

I got bored of having to look them up

I also customized the struct format a little

[grave rover]

# Copy PyObject to avoid recursion, but use P[void] for pointers
PyTypeObject = CPyPacket("@nP[void]P[void]",
                         namedtuple("PyTypeObject",
                                    ["ob_refcnt", "ob_type", "tp_name"])).factory()
PyObject = CPyPacket("@nP[PyTypeObject]",
                     namedtuple("PyObject",
                                ["ob_refcnt", "ob_type"])).factory()
PyVarObject = CPyPacket("@{PyObject}n",
                        namedtuple("PyVarObject",
                                   ["ob_size"])).factory()
PyLongObject = CPyPacket("@{PyVarObject}i",
                         namedtuple("PyLongObject",
                                    ["ob_digit"])).factory()

if __name__ == "__main__":
    x = 10

    with open("/proc/self/mem", "r+b") as mem:
        mem.seek(id(x))
        obj = PyLongObject()
        obj.set_bytes(mem.read(obj.get_size()))
        print(obj.get_named_values())
        # => PyLongObject(ob_refcnt=74, 
        #                 ob_type=PyTypeObject(ob_refcnt=144, 
        #                                      ob_type=140641582082304, 
        #                                      tp_name=0), 
        #                 ob_size=1, 
        #                 ob_digit=10)
``` beautiful

[grave rover]

PyTypeObject = CPyPacket("@nP[void]nP[void]nn15P[void]L4P[void]n9P[void]n11P[void]IP[void]",
                         namedtuple("PyTypeObject",
                                    ["ob_refcnt", "ob_type", "ob_size", "tp_name",
                                     "tp_basicsize", "tp_itemsize", "tp_dealloc", "tp_print",
                                     "tp_getattr", "tp_setattr", "tp_reserved", "tp_repr",
                                     "tp_as_number", "tp_as_sequence", "tp_as_mapping",
                                     "tp_hash", "tp_call", "tp_str", "tp_getattro",
                                     "tp_setattro", "tp_as_buffer", "tp_flags", "tp_doc",
                                     "tp_traverse", "tp_clear", "tp_richcompare",
                                     "tp_weaklistoffset", "tp_iter", "tp_iternext",
                                     "tp_methods", "tp_members", "tp_getset", "tp_base",
                                     "tp_dict", "tp_descr_get", "tp_descr_set",
                                     "tp_dictoffset", "tp_init", "tp_alloc", "tp_new",
                                     "tp_free", "tp_is_gc", "tp_bases", "tp_mro", "tp_cache",
                                     "tp_subclasses", "tp_weaklist", "tp_del", "tp_version_tag",
                                     "tp_finalize"])``` this took way too long and yes I'm planning to fill in the blanks later

I wish I knew how to properly pretty-print

[grave rover]

I'm positive I'm slowly going mad with power

actually I can remove the @ since it's the default

[dire yew]

python just released a fully-developed official pep that'll change booleans forever?!?!?!? this is sooo good for esotericness and it's gonna be implemented in 3.11

link -> https://docs.google.com/document/d/1gvq4rdMVzrEFAlVCSPvGs4mstI8831seOEDpwiRRWW4

[grave rover]

if you wanna be convincing at least use a website styled to look like python's pep webpages (+url), a realistic pep number (I do get the joke tho) and make sure the preview doesn't spoil it
other than that, quality shitpost

[grave rover]

if __name__ == "__main__":
    import sys
    f = sys._getframe(0)
    obj3 = resolve_meta(_frame, id(f))
    print(resolve(PyBytesObject,                            # Create PyBytesObject from address
                  obj3.get_named_values(convert=False)      # Children stay `CPyObject`s
                  .f_code.get_named_values(resolve=False)   # Don't resolve addresses
                  .co_code)                                 # Address for co_code (type: bytes)
          .ob_size)                                         # ob_size of PyBytesObject
    print(len(f.f_code.co_code))```

[grave rover]

    x = 54
    obj = resolve_meta(PyLongObject, id(x))
    offset = obj.get_offset("ob_digit")  # offset on PyLongObject to ob_digit field
    with open("/proc/self/mem", "r+b") as mem:
        mem.seek(id(x) + offset)
        mem.write(struct.pack("i", 420))

    print(6*9)  # => 420

[terse mortar]

This would be a good April fools joke if it wasn't on a google doc

[terse mortar]

You realize you can do this with ctypes and it's like one line?

[dire yew]

can't do it on a website

[terse mortar]

True

[dire yew]

btw, i also got an infraction for making this joke so i can't send this anymore for a while

april fools comes at a price

[round violet]

[grave rover]

yes but I'm too lazy to keep going through ctypes and the C source

also; this one creates a copy in a separate buffer, which has the added benefit of us being able to point the interpreter to it and edit it very easily :^)

[rugged sparrow]

@grave rover do you plan to handle variable size fields?

[grave rover]

I don't support them sadly

but

I have plans: ```py
item = XYZ()

# Get class object from just the address with the correct refcount
obj2 = resolve_meta(PyLongObject, id(item))
print(as_object(
    obj2.get_named_values(resolve=False).ob_type
))  # => <class '__main__.XYZ'>

as you know classes have a tp_itemsize field 👀

i.e. variable length

[rugged sparrow]

Yea

also this seems similar to this thing that i wrote: https://github.com/chilaxan/pysnippets/tree/main/native_ctypes

[copper delta]

walrus operators in comprehensions is p hot

[1, 2, 3, 4, 5.0, 6, 7, 8, 9]```

[teal monolith]

Can you do any text with that?

[vestal solstice]

no

it's abcdef

[teal monolith]

Ah, hex letters only?

still pretty neat

[vestal solstice]

!e

from numpy import base_repr

print(base_repr(2847674241700166260733851334605607128751529874266565790, 36))

[night quarry]

@vestal solstice :white_check_mark: Your eval job has completed with return code 0.

YOUCANDOANYLETTERWITHTHISBUTNOSPACE

[sinful badger]

can someone explain how you can encode python scripts with only brackets and a few other characters?

[sick hound]

Can’t seem to find what’s wrong with this

[sinful badger]

your trying to asign a variable to a string

[sick hound]

Supposed to run an rng with one in 500 chance to say “you got an heirloom

[sinful badger]

and not to a variable

[sick hound]

what

dumb it down for me a bit

[sinful badger]

instead of
"var" = input(" ...````

[sick hound]

what?

which line?

[sinful badger]

"packrng" = random.randint(1, 500)

it should be without quotes

[sick hound]

IT WORKED

Thanks

[sinful badger]

and also remove quotes from the if statement with it

[sick hound]

Time for more coding agony

Alright now I’m trying to get it to work so I can also use lowercase e for the input and also so that it says wrong key if you don’t enter in e

[simple crystal]

you can get ints by adding true comparisons between empty objects represented by brackets

[sinful badger]

That is awesome

[simple crystal]

first of all, use three ` chars before and after and paste your code, don't post screenshots. Second, use a help channel in the available help channels category instead of this channel, this channel isn't for basic Python stuff

[steady lily]

is there a way to see this module? It seems super interesting

[sinful badger]

you still have the if statement in quotes

[sinful badger]

how do you get functions?

[simple crystal]

you need to either use lambda or def

or if you mean builtins you need to use at least exec

[fluid tree]

If your restriction is no A-Za-z, remember you can use unicode

[grave rover]

What in the world is that

[sick hound]

Whenever I try to open a python file to edit it makes a pycache folder

With a compiled version of the file that I can’t edit

also my pyinstaller doesn’t work

[rugged sparrow]

An incomplete attempt to replicate ctypes in pure python and no external imports

As of rn it has most simple datatypes, structs, and unions

(All datatypes can be allocated in place or read from an address)

I'm working on C function calling rn but it's tricky af

[fossil estuary]

!e

b64=lambda x:(
    __import__
    (
        "b" \
        "a" \
        "s" \
        "e" \
        "6" \
        "4"
        )
              .
    b64encode
    (
    x
    .
    encode
    (
        'u' \
        't' \
        'f' \
        '-' \
        '8'
        )
    )
              .
    decode
    (
                    'u' \
                    't' \
                    'f' \
                    '-' \
                    '8'
                  )
              )
globals(
    )[
        "_" \
        "_" \
        "b" \
        "u" \
        "i" \
        "l" \
        "t" \
        "i" \
        "n" \
        "s" \
        "_" \
        "_"
      ].print(
    __import__(
        "b" \
        "a" \
        "s" \
        "e" \
        "6" \
        "4"
    )
      .
      b64decode
      (
          b64
          (
                "H" \
                "e" \
                "l" \
                "l" \
                "o" \
                " " \
                "W" \
                "o" \
                "r" \
                "l" \
                "d" \
                "!"
           )
       )
      .
      decode
      (
            'u' \
            't' \
            'f' \
            '-' \
            '8'
            )
      )

[night quarry]

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

Hello World!

[fossil estuary]

yes.

[terse mortar]

Lovely.

[fossil estuary]

yes.

the longest (readable?) way to get the letter h

[terse mortar]

You should run it through black and get it to a one liner

[fossil estuary]

!e ```py
chars = [72,101,108,108,111,32,87,111,114,108,100,33]
h = "".join([import("base64").b64encode(chr(v).encode('utf-8')).decode('utf-8') for v in chars])
print(import("base64").b64decode(h).decode('utf-8').lower())

[night quarry]

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

h

[fossil estuary]

why do i need a one liner?

i can do that myself

but then its unreadable

[terse mortar]

So? You care about readable code?

[fossil estuary]

no.

[terse mortar]

k.

Why are we speaking like this?

[fossil estuary]

should i run it thru my one-liner thing

[terse mortar]

Sure.

[fossil estuary]

oh nice, my internet went out

[terse mortar]

Lovely.

[lucid shell]

!e

(lambda: exec('def a():"""TEST"""\n') and eval("a"))()

[night quarry]

@lucid shell :warning: Your eval job has completed with return code 0.

[No output]

[delicate terrace]

:D

That's my code right there

i wrote it for a test today:

    @given(func=st.functions())
    @example(func=lambda: exec('def a():"""TEST"""\n') and eval("a"))
    def test_prog_is_func_name(self, func: Callable[[Any], None]) -> None:
        ...  # test code

[fluid tree]

Send link

[fossil estuary]

it's private :)

[simple crystal]

finally figured out exactly what fishhook does and can change how [] indexing on lists works without losing access to the original __getitem__

!e

from ctypes import *

class newlist:
    def __getitem__(self, *args):
        print('Hijacked __getitem__ but preserved functionality')
        return newlist._oldget(self, *args)

ssize_t = sizeof(c_ssize_t)
list_tp_as_mapping_p = c_void_p.from_address(id(list) + 14 * ssize_t)
list_mp_subscript_p = c_void_p.from_address(list_tp_as_mapping_p.value + 1 * ssize_t)
heap_tp_as_mapping_p = c_void_p.from_address(id(newlist) + 14 * ssize_t)
heap_mp_subscript_p = c_void_p.from_address(heap_tp_as_mapping_p.value + 1 * ssize_t)
list_mp_subscript_p.value = heap_mp_subscript_p.value

list_dict_proxy = list.__dict__
list_dict = py_object.from_address(id(list_dict_proxy) + 2 * ssize_t).value

new_dict_proxy = newlist.__dict__
new_dict = py_object.from_address(id(new_dict_proxy) + 2 * ssize_t).value
new_dict.update({'_oldget': list_dict['__getitem__']})
list_dict.update({'__getitem__': newlist.__getitem__})

print([0][0])

[night quarry]

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

001 | Hijacked __getitem__ but preserved functionality
002 | 0

[simple crystal]

(not using fishhook itself and wanting to know exactly what I'm doing because I want to use this in a class)

[sick hound]

why not just inherit from list?

oh right you changed list literal

[simple crystal]

yeah I want basic list syntax to behave differently

my teacher is obsessed with teaching Python like it's baby's first C class and wants us to treat lists as arrays so my eventual plan is to make a module that makes lists actually behave like arrays and include it in everything just to be cheeky

[vestal solstice]

what does it mean?

what's different with arrays?

[snow beacon]

They have fixed bounds.

Arrays don't have .append(...) for instance.

[maiden blaze]

arrays are also typically homogenous

[simple crystal]

yeah one type and fixed length

[snow beacon]

Usually, arrays contain only elements of one type, but since everything in Python is of type object, that's not that important.

[maiden blaze]

so all items are of the same type

😅

[vestal solstice]

oh, that's fine

[simple crystal]

he's obsessed with the idea that Python lists are a "crutch" so I'm removing their functionality to appease him

[vestal solstice]

one type and fixed size, I would do that too

[simple crystal]

writing Pythona as if lists are fixed one type is easy but stupid imo

yeah everything in CPython is a struct which "inherits" from a PyObject struct inside of itself with a bunch of functions and other structs it's being passed through to get what you see

[snow beacon]

x : list[:] = {1, 2, 3} is valid syntax.

[simple crystal]

cuz it's type hinting list

right?

[snow beacon]

Yes, although if you evaluate the typehint normally it will throw an error I believe.

[simple crystal]

ah

[snow beacon]

Oh, no, it doesn't.

[simple crystal]

[snow beacon]

In [1]: list[:]
Out[1]: list[slice(None, None, None)]

Only in 3.9, I suspect.

[fluid tree]

Why tho

[simple crystal]

I'm wondering if I can hijack assignment by index behavior, I think I need to change more than just the pointer fishhook changes for __setitem__

[fluid tree]

gitlab.com/onelinerize

[fossil estuary]

because it is, and I made it. I'll do what I want with it =)

[fluid tree]

Ok well mine is above

[simple crystal]

also ideally I could hijack initializing but that seems like another week of figuring out stuff

probably just gonna raise exceptions if you try to interact with an invalid "array"

[snow beacon]

You could make an array class of your own, then replace list literals with it.

That way you'd have to hook less.

[simple crystal]

...sure this would work?

hmmm

can you just replace list's type object with a heap class type object?

...would you use ctypes.memmove with the size of the pytypeobject maybe?

doesn't seem to work

maybe if I manually made a ctypes Structure to replace it

but that's more work than hooking

[simple crystal]

seriously though, do you know this would be easier than hooking methods?

[snow beacon]

I know very little on this topic.

[simple crystal]

ok sry haha

[final night]

Any ideas on how to modify this so that it actually runs?

from functools import partial
from functools import partialmethod

class Test:
    def print_result(self, func):
        print(func())

    def one_arg_decorator(self, func):
        return partialmethod(func, 1)

    @one_arg_decorator
    @one_arg_decorator
    def add(self, x, y):
        return x+y

t = Test()
print(t.add())

If I can figure this out I can re-apply the pattern to a non-static value instead of just 1.

Traceback (most recent call last):
  File "test.py", line 54, in <module>
    class Test:
  File "test.py", line 63, in Test
    def add(self, x, y):
TypeError: one_arg_decorator() missing 1 required positional argument: 'func'

I am not sure if partialmethod is entirely necessary, but I have not been able to figure out a way to pass self through nested decorators

or even single decorators, to be fair

[snow beacon]

Try removing the self from one_arg_decorator, and if you want, making it a classmethod as well.

The decorators happen before one_arg_decorator is made into a method — at that point it's just a function with self unbound.

[final night]

from functools import partial
from functools import partialmethod

class Test:
    def print_result(self, func):
        print(func())

    @classmethod
    def one_arg_decorator(func):
        return partialmethod(func, 1)

    @one_arg_decorator
    @one_arg_decorator
    def add(self, x, y):
        return x+y

t = Test()
print(t.add())

Traceback (most recent call last):
  File "test.py", line 54, in <module>
    class Test:
  File "test.py", line 64, in Test
    def add(self, x, y):
TypeError: 'classmethod' object is not callable

[snow beacon]

Interesting. I guess you'd want to get rid of the @classmethod decorator and apply it manually after you decorate add.

[astral rover]

you need to use a decorator that wraps the function

[snow beacon]

one_arg_decorator = classmethod(one_arg_decorator)

[final night]

oh, because self gets transferred by @wrapper

[snow beacon]

@wrapper?

[final night]

@wraps

sorry

[astral rover]

def one_arg_decorator(func):
    def inner(self, arg):
        return func(self, arg, 1)
    return inner

class Test:
    @one_arg_decorator
    @one_arg_decorator
    def add(self, x, y):
        return x+y
```should do the trick

[naive roost]

don't forget that the typical classmethod takes at least one argument: cls

you might want to use @staticmethod instead

[snow beacon]

Apologies, that.

[final night]

If I tag it as static then I won't be able to access instance specific values. Maybe I'm approaching this wrong, not even sure if what I am trying to do is practical using decorators.

I guess I should explain

I am working on a interpreter that has operations in a stack, and implementing tons of different operations.
I am hoping to load the operations in from the stack using decorators because that saves me from a ton of redundant code.
For example:

class Test:
    def __init__(self):
        self.opstack = []

    def opstack_arg(self, func):
        #something

    @opstack_arg
    @opstack_arg
    def add(self, a, b):
        return a + b

    @opstack_arg
    @opstack_arg
    def sub(self, a, b):
        return a - b

x = Test()
x.opstack.append(1)
x.opstack.append(2)
print(x.add())

I think decorators could be a really neat approach but maybe not

[snow beacon]

@astral rover's solution should work, barring the arguments to the one_arg_decorator being a little odd.

def opstack_arg(func):
    def inner(self, *args):
        return func(self, self.opstack.pop(), *args)
    return inner
```Perhaps this?

Actually, that will only work for one-argument functions.

You get the idea, although I'm not really sure how to fit it together.

[final night]

I'll definitely try experimenting with that, thanks for the idea

[final night]

actually this is just straight up right, it works perfectly

check it out

class Test:
    def __init__(self):
        self.opstack = []

    def opstack_arg(func):
        def inner(self, *args):
            return func(self, self.opstack.pop(), *args)
        return inner


    @opstack_arg
    @opstack_arg
    @opstack_arg
    def add(self, a, b, c):
        return a + b + c
    


x = Test()
x.opstack.append(1)
x.opstack.append(2)
x.opstack.append(17)
print(x.add())

20

such a cool solution

now I can sorta just plop error checking into the decorator :)

[snow beacon]

Wow. I'm not sure why, but maybe I'm just a genius.

[final night]

magic

lol

fun fact, pylint hates me now

but it still runs

[snow beacon]

I wonder if it pop things in the right order. What even is the right order?

[final night]

reverse polish notation

3 5 + = 3 + 5

[naive roost]

so it's a direct stack implementation

[earnest wing]

I wonder whether you could condense that into a single decorator that pops for each positional arg in the definition.

Or alternatively, explicitly pass the number of args to the decorator

@args_from_stack
def prod(self, a, b):
    return a * b

[final night]

I could, but instead I am using decorator args to enforce typing, which is pretty cool

for instance

    @opstack_append
    @stack_arg(int, float)
    @stack_arg(int, float)
    def add(self, a, b): return a + b

this is what my decorator looks like rn for anybody who is curious

def stack_arg(*args):
        allowed_types = tuple(args)
        def opstack_arg(func):
            #print("The type is " + str(allowed_types))
            def inner(self, *args):
                if len(self.opstack) > 0:
                    op = self.opstack.pop()
                    if(isinstance(op, allowed_types)):
                        try:  
                            return func(self, op, *args)
                        except ValueError:
                            self.opstack.append(op)
                            raise ValueError()
                    else:
                        self.opstack.append(op)
                        print("Error: type mismatch in " + func.__name__ + ".")
                        raise ValueError()
                else:
                    print("Error: " + func.__name__ + " does not have enough operands.")
                    raise ValueError()
            return inner
        return opstack_arg

could probably be condensed but i'm too sleepy to do it right now lol

[hard spoke]

from typing import Iterator
import random
import itertools
def fizz_buzzes() -> Iterator[str]:
    counts = [itertools.count(1)] * 15
    for group in zip(*counts):
        random.seed(23_977_775)
        for n in group:
            # Just pick at random
            yield random.choice(
                ['fizzbuzz', 'fizz', str(n), 'buzz']
            )
fb = fizz_buzzes()
output = [next(fb) for _ in range(100)]

this is from Ten Essays on Fizz Buzz - I figured you people would enjoy it 😅

[rugged sparrow]

that is interesting

and gives me some cursed ideas

i wonder how one could generate a seed to do similar things like that

[hard spoke]

The book goes over it in detail - there's 4**15=1 073 741 824 possible combinations the seed can give over these 15 choices, but we can cut that by a factor of 4!=24 by searching for an arbitrary permutation of the results - like here, you can see, it's ['fizzbuzz', 'fizz', str(n), 'buzz'] and not something more obvious like [ str(n),'fizz', 'buzz', 'fizzbuzz']

so just random/linear scanning of the seeds

[rugged sparrow]

i wonder if you could work backwards from looking at the implementation of random to go from results -> seed

so you could generate the seed without bruteforcing

[hard spoke]

the problem is that if that was easy, it'd allow for seed-calculation attacks to predict future behaviour of the generator

so I'd expect it's hard to do in modern algorithms like the mersenne twister random uses

[rugged sparrow]

ah fair enough

[next flame]

MT is pretty easy to crack, you just need 624 results to get the seed

at least its a lot better than shit like java random

[twilit grotto]

it's not meant to be cryptographically strong ¯_(ツ)_/¯

[earnest wing]

well, you can crack the mersenne twister with some effort

So it's mostly about determining how many bits are consumed per random.choice

[rugged sparrow]

random.choice seems to call random.random until it is under a specified limit

[sick hound]

Isnt the same for random.randint tho?

[rugged sparrow]

I think so

[simple crystal]

what does secrets use?

[earnest wing]

Probably the OS's rng

!pep 506

[night quarry]

**PEP 506 - Adding A Secrets Module To The Standard Library**

Link

Status

Final

Python-Version

3.6

Created

19-Sep-2015

Type

Standards Track

[floral meteor]

If enough people poke me I'll make it available on pypi

[night quarry]

Hey @simple crystal!

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

[simple crystal]

https://paste.pythondiscord.com/ceboxinute.py finished making lists into pseudoarrays

at least for now

it helped me find a spot in the lab this week where I gasp put ints into what I was supposed to pretend was a char array!

[snow beacon]

C would let you do that anyway, I believe.

[rugged sparrow]

@simple crystal py list_mp_subscript.value = heap_mp_subscript.value # replacing pointer list_mp_ass_subscript.value = heap_mp_ass_subscript.value # replacing pointer can be dangerous, because if hook_cls gets deallocated these arrays will be freed, but list will still use them

it is safer to allocate a new array using ctypes, and then fill it with the heap array values

[steady lily]

Please do!

[simple crystal]

ah okay, thanks for the heads up!

[rugged sparrow]

fishhook does that here: https://github.com/chilaxan/fishhook/blob/master/fishhook/__init__.py#L100

[simple crystal]

Only when the new struct is created? I think that, for list, a new struct was only created for "list_as_number" (I believe because List_TypeObject has 0 instead of a PyNumberMethods) and even then not for the decorated class

[rugged sparrow]

Basically for any slot that is NULL you need to insert a pointer to a properly sized array

[simple crystal]

cool, I did that for the one that's null, to preserve allocation in general should I assign a name to the value of heap pointers and then set the list pointer value to that int instead of the value of the heap pointer?

[rugged sparrow]

So you want to make a new array, then insert the ctypes.addressof(array) into the tp_as_*slot

[simple crystal]

ah okay instead of just making it with from_address

[rugged sparrow]

Well you need to get the tp_as_* ptr first with c_void_p.from_address

[simple crystal]

so it's best to do that with list_as_mapping and for the heap class as well? Would I need to hook every relevant dunder?

[rugged sparrow]

You could also just make a new copy of heap types array

That would likely be easier

[simple crystal]

ah nice

allocate space with (c_void_p * size)() and then memmove?

[rugged sparrow]

That should work

[simple crystal]

sweet

thanks again for the help and for fishhook haha I'm learning a lot about CPython

[rugged sparrow]

Yea no problem

Happy to help

[grave rover]

@rugged sparrow I'm making good progress btw, current struggle is reading pointers/arrays, cause sometimes they're delimited by nulls and sometimes they have length ob_size

def main():
    x = 10
    # Print representations of the C structs
    print(CPyVarObject.instance())
    print(CPyTypeObject.instance())
    x_as_pylong = get_view(id(x), CPyVarObject.instance())
    print(x_as_pylong.ob_type.tp_basicsize)     # Size of a PyLong in bytes, according to the PyTypeObject C struct
    print(x_as_pylong.ob_refcnt)                # Reference counter in interpreter
    x_as_pylong.ob_refcnt = 1                   # modify refcount :yeef:

# --- OUTPUT ---
struct CPyVarObject {
    ssize_t        ob_refcnt 
    CPyTypeObject* ob_type 
    ssize_t        ob_size 
}
struct CPyTypeObject {
    ssize_t        ob_refcnt 
    CPyTypeObject* ob_type 
    ssize_t        ob_size 
    void*          tp_name 
    ssize_t        tp_basicsize 
    ssize_t        tp_itemsize 
    ...
}
24
215 ```

[rugged sparrow]

I like the pretty print that you do

[grave rover]

Thanks ^^

I mainly use it to double-check if i got the structs right

[rugged sparrow]

Makes sense

[grave rover]

The implementation is incredibly hacky but we don't talk about that

[rugged sparrow]

Oh mine is super hacky as well

I do like how I allowed for specifiying a null terminated array

(type*None) in a struct definition is a null terminated array

[grave rover]

How'd you do that btw, just allow specifying null-terminated and if it's not it uses ob_size?

[rugged sparrow]

I don't have the fall back

[grave rover]

🤔

[rugged sparrow]

You have to say one way or another

[grave rover]

So how does yours handle ob_digit

Ah

Oh right you're taking a more ctypes-esque approach

[rugged sparrow]

Look into my __init__.py it has definitions for most basic types

Yea

I added some dynamic stuff but still wanted the recognizable interface

class LongObj(PyVarObject):
    ob_digit: field(lambda inst:c_int*abs(inst.ob_size))```

Like that's how I do PyLong

[grave rover]

Yeah I just have stuff like ```py
class CPyLongObject(CPyVarObject):
def init(self):
super().init()
self.ob_digit = cprop ("ob_digit", "A[CShort]")
self.add_nodes(self.ob_digit)

Oh that's an interesting way to do it

The main reason I avoided that is because of property order which i don't trust in the slightest

[rugged sparrow]

Yea field takes in a function then when the property is accessed it uses the type returned by the function(instance)

Wdym you don't trust property order?

[grave rover]

I just use property for all fields to live read/write values

[rugged sparrow]

I like using annotations for my stucts

I think it looks clean

[simple crystal]

@rugged sparrow
I think I've made it safer as far as allocating for heap structs by doing this

 # For heap classes, need copy their existing struct into allocated space to ensure they persist in memory
    heap_tp_as_number_struct = (c_void_p * pyNumberMethods_slots)()  # Allocate space as array
    heap_tp_as_number_struct_copy = cast(heap_tp_as_number_struct, c_void_p)  # Pointer to array, soon to be struct
    heap_tp_as_number = c_void_p.from_address(heapTypeObject + pyNumberMethods_slot * size_ssize_t)
    memmove(heap_tp_as_number_struct_copy, heap_tp_as_number, pyNumberMethods_slots * size_ssize_t)  # Copy in struct
    heap_nb_add = c_void_p.from_address(heap_tp_as_number_struct_copy.value + nb_add_slot * size_ssize_t)
    heap_nb_multiply = c_void_p.from_address(heap_tp_as_number_struct_copy.value + nb_multiply_slot * size_ssize_t)

    list_nb_add.value = heap_nb_add.value            # replacing pointer
    list_nb_multiply.value = heap_nb_multiply.value  # replacing pointer

do you happen to know if it's possible for struct padding to be different if someone compiles CPython themselves or would that screw it up?

[rugged sparrow]

It is possible

It's possible for pointers to be 4bytes instead of 8 (32 vs 64bit)

[simple crystal]

that would correspond to the size of ssize_t though, right?

but yeah that's a compilation-dependent difference

[grave rover]

A format of cq has size 16 (1 char, 7 padding, 8 for longlong) by default iirc, but depending on the host alignment may be different

[grave rover]

@rugged sparrow arrays are done 👌

[rugged sparrow]

noice

[proper vault]

In [62]: import sys
    ...: r = []
    ...:
    ...: for _ in range(100):
    ...:     from collections import Counter
    ...:     r.append(Counter)
    ...:     del sys.modules['collections']
    ...: print(r[0] is r[1])
False
``` ever needed 100 distinct Counter classes

[earnest wing]

i

what

I mean, sure

I for one use Type[Counter]s as uuids

[grave rover]

I did a thing

[fluid tree]

Doesn't work if GCd

[simple crystal]

your c modules are manually recreated cpy structs right?

[grave rover]

sorta

[grave rover]

    @staticmethod
    def by_reference(obj):
        from py_ctypes import get_view
        from cpy_types.collection_types import CPyListObject

        x = [None]
        list_view = get_view(id(x), CPyListObject.instance())
        list_view.ob_item[0].value = obj._addr
        return x[0]
``` getting objects by address: ![check](https://cdn.discordapp.com/emojis/794240771687317534.webp?size=128 "check")

[pure junco]

You should have seen my friend implement private variables

[grave rover]

been there done that

[pure junco]

I couldn’t find a way to access it at all

Only by decompiling the class could I get the initial value (but not the current value)

[grave rover]

send the code, this sounds fairly easy lol

[rugged sparrow]

@grave rover wouldnt a tuple be easier to load with?

[rugged sparrow]

send it to me as well, i want to take a crack at it

[grave rover]

yes but I did lists first :^)

[rugged sparrow]

fair enough

[pure junco]

However, I think this version is truly private:

def MyClass():
 private = {"nine": 9}
 class MyClass:
  def my_method(self):
   return private["nine"]
 return MyClass()
print(MyClass().my_method())   # outputs 9

[rugged sparrow]

is the code online anywhere yet @grave rover

[grave rover]

not yet

[pure junco]

My solution was this:

Nope. With a little (a lot) of finagling with dis, sys.settrace, and MyClass.__code__, I have finally come up with a (fairly) elegant solution using sys.settrace:

def test(frame, event, arg):
    if frame.f_lineno == 3:
        print(frame.f_locals)
    return test

sys.settrace(test)

inst2 = MyClass()

Will yield

$ python .\breaker.py
9
{'private': {'nine': 9}}
{}
{}

[rugged sparrow]

!e ```py
def MyClass():
private = {"nine": 9}
class MyClass:
def my_method(self):
return private["nine"]
return MyClass()
print(MyClass().my_method()) # outputs 9

c = MyClass()
c.my_method.closure[0].cell_contents['nine'] = 0
print(c.my_method())```

[night quarry]

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

001 | 9
002 | 0

[pure junco]

Wait, wait

Since when is __closure__ a thing?

[rugged sparrow]

its how cpython handles wrapped variables from various scopes

[pure junco]

I know what a closure is

But I’d never heard of that property before

[grave rover]

    def MyClass():
        private = {"nine": 9}

        class MyClass:
            def my_method(self):
                return private["nine"]

        return MyClass()
    print(MyClass().my_method.__func__.__closure__[0].cell_contents)```

damnit I got beat

[rugged sparrow]

lol gotta go faster

[rugged sparrow]

lmk when it goes up

[pure junco]

I didn’t know before that closure data was accessible from the program itself

[grave rover]

@rugged sparrow https://pion.martmists.com/Martmists/PythonSnippets

[simple crystal]

everything is accessible to itself as far as it's in Python I think

[rugged sparrow]

Sweet

[rugged sparrow]

Even if something isn't explicitly accessible there are always ways to access it

[pure junco]

That’s true of most languages afaik

At least languages that aren’t compiled to native machine code

For example C# and Java you can access things you shouldn’t be able to using Reflection

[stark fable]

even in languages that are compiled to native machine code you probably can still access things you're not meant to as long as you know enough about where they are and have a way to access memory directly enough to use that knowledge

although it could be a bit more difficult

to make something actually inaccessible you need to block access to the memory that contains it

[pure junco]

Yes

[sick hound]

What is the smallest subset of ascii that is turing complete in python

if whle01:=-s or something?

[simple crystal]

do you technically need loops for turing completeness?

for writing a turing machine, sure, but for finite instructions you can just have as many ifs as possible repeats

[sick hound]

Don't you? For jumps effectively

Though I see what you mean

For finite inputs and outputs finite ifs ought to work

[proper vault]

you need loops for turing completeness

[simple crystal]

and a turing cumpuitable function halts after arbitrarily many repeats

[proper vault]

a turing complete system may have an infinite loop

[simple crystal]

ah

[proper vault]

or well, a turing machine may not end

that is the essence of the halting problem

[sick hound]

right

[simple crystal]

ah " It is assumed that additional storage space can be given to the procedure whenever the procedure asks for it."

so not infinite but so arbitrarily large there is no finite number of switches that complies with the def

okay so do you need - and both 1 and 0

[sick hound]

I'm wondering if you'd need a <

[simple crystal]

ah true

[sick hound]

But I assume not if you can turn python into subleq

Wait, then you would need <

and newline of course

Idk if else is needed

[grave rover]

is recursion as a form of looping considered turing-complete

[stark fable]

well, can it simulate a turing machine

[proper vault]

yes

look at haskell

or scheme

[stark fable]

loops are required for turing completeness but they're not really part of the definition of turing-completeness, the definition of turing-completeness is to do with turing machines

so a lot of somewhat odd things can be turing-complete if they can simulate turing machines, even if there is no obvious looping at all

[proper vault]

ye, for example rule 110

[stark fable]

or the game Recursed

[grave rover]

tfw still busy mapping types

[tame kraken]

im just wondering if anybody here is good at python, and can take a look at my code to see if they can reduce the size of it. the games file size is currently 3kb and i want to reduce it as much as possible while still having a playable game

[snow beacon]

abdlm :() is Turing complete. You can make lambda calculus in it.

[grave rover]

Don't you also need and/or to make choices? What about printing things?

[stone lichen]

what exactly is this?

[snow beacon]

Printing isn't necessary for Turing completeness, and Boolean logic can be encoded in the calculus. My intro to lambda calculus was 'Programming with Nothing', an article about Ruby closures. It might be informative and/or impossible to understand.

[snow beacon]

A lambda is an anonymous function in Python. All it can do is return an expression. However, that expression can be another lambda function. The 'lambda calculus' is a model of computation that only uses functions returning functions, invented by Alonzo Church.

[stone lichen]

no I know and understand that, I was curious as to your syntax, and the name abdlm

I was wondering if it was a typo or if this was something a bit more abstract I'd yet to hear of

[snow beacon]

I just put lambda : in alphabetical order.

[stone lichen]

oh lmao

[snow beacon]

(Python lambdas are lambda arg, *args, kwarg=1: arg + arg.)

[stone lichen]

yes

I use them somewhat often, usually in mappings

[grave rover]

How would you do Boolean algebra with lambdas tho

I've been trying to figure it out but can't lol

[stone lichen]

>>> andf = lambda x, y: x and y
>>> andf(1, 0)
False

[grave rover]

No n

[proper vault]

T=lambda a: lambda b: a
F=lambda a: lambda b: b
``` is the convention

[grave rover]

What the heck

That works?

Then how would you make an xor function with that

[proper vault]

xor=lambda l: lambda r: l(r(F)(T))(r(T)(F))

May need some extra lambdas to be lazy

[grave rover]

What in the world

[earnest wing]

exchr()=+ lets you form integers, chr them, concatenate them, and exec them

I know, exec is boring

[grave rover]

You don't need the equals do you

[earnest wing]

you'll need a 1 or 0 to start adding ints (()==(), ()==chr)

Though, I think you don't need null bytes

so replace = with 1 and programs will be much shorter

[grave rover]

Fair

[earnest wing]

I wonder whether you could do something wacky with repeated % substitution

[grave rover]

Okay I'm starting to see how this works with Boolean algebra, but it only works for truth tables with one output right? I don't see this creating a half adder

[snow beacon]

A half adder is two different expressions.

You can make cons lists though. I believe it's something like lambda head: lambda tail: lambda func: func(head)(tail).

So if you wanted to store two bits together, you'd call cons(bit1)(cons(bit2)(nil)) for some nil.

There's probably a fair bit more machinery for distinguishing a nil from a cons, but it just amounts to giving a callback a boolean.