esoteric-python

[floral meteor]

The blur method is average of pixels within a circular radius

restricting that radius for red, more than green, more than blue, as is the pattern consistent with a focal point in front of the retina instead of on it.

The code for it is ironically the code in above screenshot

me: I coded something hard to read!
the code:

If the definition of obfuscated code is code that is deliberately hard to read, then I think putting a short-sightedness filter over the code counts, right?

[severe canyon]

pain

[long hamlet]

What is that

Ok will share a few moar python generated art then an esoteric pure python 2D /3D Library I made

https://github.com/TechnoTanuki

That is pure python no libraries...

No classes all functional

Made my own bitmap fonts

Made my own graphics library from pixel addressing to 2D primitives like line , circle, etc
To 3D projection to 2D with hidden surface removal and shading

Made my own lin alg lib

[runic mirage]

damn

[long hamlet]

I think I promised to share this here months ago lol better late than never

[runic mirage]

also these circles up there look really cool

[long hamlet]

Can get funky

[runic mirage]

wohoo

how did you make these shapes?

[long hamlet]

Esoteric Math

[runic mirage]

😜

The one looks like an avocado 🥑

[long hamlet]

Mix of sines and cosines then I deform

[pure dew]

oh its not actual gpu drawing? got me excited

[long hamlet]

Yes pure CPU...sorry for late reply was sleeping...

I created an in memory byte array that serves as a virtual video buffer that has the byte layout that matches a mono, 16 color , 256 color or RGB windows bitmap with variable size...that I made addressible with bit or byte based operations via my own pixel routine...then built up from there ....

[finite blaze]

I wonder how fast it is?

[long hamlet]

I profiled it pretty fast most operations take less than a second

But it isnt real time and output is to memory then file totally bypassing the GPU lol

Ok a full set of tests done in 5 sec

Outputs the test bmp shown previously in which I tried to cram in most of the features lol...has 2D and 3D

BTW done on a i7 laptop with a naked board I need a new one

I short the pins to turn it on lol

[long hamlet]

Estoteric bare board laptop computing lol

I can also do filters, image resize, outline etc with the library ...it is there but I havent checked in the image manipulation unit tests

[sick hound]

Why is the pcb naked

[long hamlet]

Power switch no longer work and keyboard keys broke too

[sick hound]

What model?

[night quarry]

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

utf-8

[long hamlet]

Old ASUS i7 probably buy a new one ...like the bare board ....cpu fan ded too ...

Bare board dont overheat lol

[sick hound]

Lmao what

So what's dissipating the heat?

[long hamlet]

I play CIV 5 no overheat

Desktop fan

Eso hardware lol

[sick hound]

The more you tell me the more confused I am

What even happened there?

[long hamlet]

There is also a heat sink from an old laptop i slipped below

[sick hound]

Did someone nuke your laptop?

[long hamlet]

I did to make it useful...fan broke. Power Switch broke in middle of Covid lol

Macgyver

So i removed all broken parts and retained what is essential...Zen

[split salmon]

sorry took me a day https://paste.pythondiscord.com/qerowuxuxi, that code is incompatible with this program

[floral meteor]

what's the domain and range of the mandelbrot fractal?

well, the interesting parts of it, at least

This is an encrypted message.

[2240738, 5573346, 1449296, 4889873, 1828324, 4274806, 4841102, 4607466, 3890048, 5126385, 420300, 2687704, 1255302, 2727331, 813057, 2034996, 1788130, 1124028, 580034, 2031250, 5514529, 5534410, 187413, 5387283, 2443965, 4789045, 70840, 323349, 327840, 4990935, 4813300, 2292617, 756145, 2113802, 2673823, 5544930, 1088033, 1619516, 4902911, 4273008, 274757, 2571054, 3443850, 3954383, 4213766, 3136367, 1315753, 139190, 3926698, 1756197, 4093664, 822713, 3623924, 4306893, 4627029, 4387294, 2417394, 4490347, 1262744, 4135132, 1758104, 3652352, 2826844, 4154119, 4203715, 2112125, 2580893, 5127401, 2909679, 2545234, 4337444, 1103112, 5212413, 3574240, 3579850, 4291504, 3940594, 4091867, 329468, 271258, 1558640, 835649, 4886917, 5344415, 3710410, 3583840, 1083078, 4851517, 1604477, 2262315, 3745322, 1071431, 3014282, 2267304, 2688016, 1744634, 2680341, 5334260, 2916342, 3207715, 198719, 2069901, 2676107, 1722937, 2795632, 1015599, 65964, 3888581, 4251997, 5054434, 301474, 5410911, 1193709, 5452445, 1031311, 2683125, 892396, 824137, 4243221, 4871888, 1232370, 4558140, 3307947, 4561688, 588931, 913729, 1131667, 2931906, 3977081, 1817868, 3134927, 4364988, 2969286, 2671041, 1937563, 495556, 1385182]

This is the decryption method

def generate(seed:int,key:int):return''.join([*loadbar((*map(lambda x:int(x)<<1,str(seed)),),lambda x:str(x**key))]).strip('0')
def intfuck(c:list[int]):
  a=__import__("collections").defaultdict(int)
  i=p=t=0
  def add(i):
    a[i]=(a[i]+1)%256
    return i,0
  def sub(i):
    a[i]=(a[i]-1)%256
    return i,0
  def left(i):return i-1,0
  def rite(i):return i+1,0
  def dot(i):
    print(end=chr(a[i]))
    return i,0
  def com(i):
    a[i]=ord(__import__('sys').stdin.read(1))
    return i,0
  def loop(i):
    return i,not a[i]
  def end(i):
    return i,-bool(a[i])
  while 0<=p and p<len(c):
    if t:t+=(c[p]==7)-(c[p]==8)
    else:i,t=[add,sub,left,rite,dot,com,loop,end,lambda i:(i,0)][int(''.join(map(str,range(8))).find(str(c[p]-1)))](i)
    p+=1-2*(t<0)
  return int(bool(t))
def decrypt(array:list,seed:int,key:int):
  print("generating...")
  abstract_integer = generate(seed,key)
  print("decrypting...")
  del key
  numbers = [int(abstract_integer[int(this)])for this in array]
  return[intfuck(numbers),print("decryption complete.")][0]

[long hamlet]

https://github.com/TechnoTanuki/Python_BMP/blob/5b371a2e564a2c326901a4cc28f8284cfdd8c099/fractalsandart.py

[floral meteor]

Approximately how long is it likely to take to decrypt the message?

[long hamlet]

mandelparamdict(): return {'maxdefault':(1.75,-1.75,1.5,-1.5),'maxeqdim':(1.75,-1.75,1.75,-1.75),'middefault':(.75,-1.25,1.25,-1.25),'mindefault':(.75,-.75,.5,-.5),'mineqdim':(.5,-.5,.5,-.5),'custom1':(-.5,-.7,-.5,-.7)}

[floral meteor]

the os module is missing an important implementation

rmdir with the /s flag

although I guess that's to stop noobs from shooting themselves in the foot

but it's still annoying

[long hamlet]

Yes shooting one self in the foot is a common programmer joke

[floral meteor]

now I have to make an object s with a __rtruediv__ method that is a wrapper on an os method

the reason we're stuck with the walrus is just so this is invalid syntax

a = 1
while a = 1:
  if input('q to quit>')=='q':
    a = 0

The reason we can't

def destroy_dir(dir):
  import os
  os.rmdir(dir, recursive=True)

is rather obvious, but still annoys me

I'm gonna have to implement recursive

just for a decryption project 🤦

[proper vault]

!d shutil.rmtree

[night quarry]

shutil.rmtree

shutil.rmtree(path, ignore_errors=False, onerror=None)```
Delete an entire directory tree; *path* must point to a directory (but not a symbolic link to a directory). If *ignore\_errors* is true, errors resulting from failed removals will be ignored; if false or omitted, such errors are handled by calling a handler specified by *onerror* or, if that is omitted, they raise an exception.

[floral meteor]

I already made the s object

@type.__call__
class s:
      def __rtruediv__(self, other):
        def inner(dir, *, _rec=0>0):
          try:other(dir)
          except NotADirectoryError as e:
            if _rec:os.remove(dir)
            else:raise e
          except OSError:
            for entry in os.listdir(dir):
              try:os.remove(entry)
              except:inner(dir+os.sep+entry,_rec=1>0)
            other(dir)
        return inner
(os.rmdir /s)("dir_name")

[severe canyon]

i used x=-2 to 0.47 and y=-1.12 to 1.12 for my visualization

or, around there for shorter code :P

[floral meteor]

I prefer to round to
x in (-2, 0.5)
y in (-1, 1)

[severe canyon]

Round up rather than down then at least

[floral meteor]

I don't see anything happening on the imag axis further than 1 unit away

[severe canyon]

you do cut off

[sly root]

(as-> #@(bot.event (defn/a on_ready [] (print "Logged in as" bot.user))) on_ready
  (-> ace.events
    (.update {"on_ready" on_ready})))
(as-> #@(bot.event (defn/a on_command_error [ctx exception] (.error log exception))) on_command_error
  (-> ace.events
    (.update {"on_command_error" on_command_error})))```

i like this so much

there's no such channel for this topic so idk

[sly root]

compiles into ```py
@bot.event async def on_ready():
return print('Logged in as', bot.user)
on_ready = ace.events.update({'on_ready': on_ready})

@bot.event async def on_command_error(ctx, exception):
return log.error(exception) on_command_error = ace.events.update({'on_command_error': on_command_error})

[sly root]

with for loop

(for [(, name event) (zip ["on_ready" "on_command_error"]
                          [(fn/a [] (print "Logged in as" bot.user)) 
                           (fn/a [ctx exception] (.error log exception))])]
  (-> ace.events
    (.update {name event})))```

[sly root]

compiled

async def _hy_anon_var_1():
    return print('Logged in as', bot.user)
async def _hy_anon_var_2(ctx, exception):
    return log.error(exception)

for (name, event) in zip(['on_ready', 'on_command_error'], [_hy_anon_var_1, _hy_anon_var_2]):
    ace.events.update({name: event})```

[finite blaze]

whats this language?

[sick hound]

i’m working on a bot that reads json responses and most of the responses are in jeff:jeff2 format and i want to only jeff2 part and if it doesnt have a jeff2 part so only a jeff part i want to skip ig

[fleet bridge]

What is "hy"?

[restive void]

https://github.com/hylang/hy

[sly root]

hylang

basically just language created with ply (or sly)

[quartz wave]

i'm wondering where IFcoltransG went and why isn't he active any more

[long hamlet]

Maybe he is busy

[floral meteor]

Nooooooooooo

[severe canyon]

yessssssssssss

[long hamlet]

Yes

[floral meteor]

I want to make a float that doesn't float.

but can still execute float operations

including complex numbers

only need rational floats

although 1/3 might be a bit difficult

what if I stored rational floats as exact value integer fractions?

but then the mathematical algorithms to apply float operations...

!doc fractions.Fraction

[night quarry]

fractions.Fraction

class fractions.Fraction(numerator=0, denominator=1)``````py

class fractions.Fraction(other_fraction)``````py

class fractions.Fraction(float)```

[floral meteor]

that's what I need

[vestal thicket]

does anyone have any advice on grabbing data from output of cli tools like xrandr and xinput that is more approachable than using basic string manipulation and line reading techniques?

i'll give you an example of the kind of thing i'm fucking sick of having to write at this point

https://pastebin.com/ppuiWzYB

like, spending an entire afternoon just writing sections of code like that is torture

[floral meteor]

there comes a point where pep8 makes things harder to read, and I think you found that point

you use imports for a cli?

you use tools?

bro I have minesweeper where the only tools I use are msvcrt.kbhit and msvcrt.getch and random.choice and time.time. Maybe itertools.product.

and all of those I make fit into my design, not the other way around.

you do mean command line interface when you say cli, right?

[vestal thicket]

oh yeah i forgot uh

that isn't actually including the code where it GETS the data haha

[floral meteor]

if you're gonna talk about stuff like xrandr and xinput, preferably link to an explanation so we don't have to research and/or guess what you mean

[vestal thicket]

i'm just using subprocess and reading from stdout

[floral meteor]

what's wrong with input()?

[vestal thicket]

xrandr is a cli tool used to get information about the current X screen and connected X displays, xinput is a cli tool used to get information about connected input devices like keyboards and mice and their properties

primarily they're intended for interactive use

[floral meteor]

riiiiiiight.

my minesweeper game is very interactive, you use wasd keys to move a cursor around to select the squares

anyone who gonna use a cli aint the type of person who would use the mouse

[vestal thicket]

here's an example of the sort of output i have to grab data from
xrandr --listmonitors Monitors: 3 0: +*DVI-I-1 1920/518x1200/324+1280+0 DVI-I-1 1: +DP-5 1280/376x1024/301+0+0 DP-5 2: +HDMI-0 1280/339x720/190+3200+0 HDMI-0

[floral meteor]

I just play around with the terminal instance hosting my script, and I use ANSI escape sequences to do that

[vestal thicket]

it's a pain in the ass

[floral meteor]

don't need to know about monitors smh

a cli only needs to know about its terminal

[vestal thicket]

what i'm writing is a gui app that sets the input dimensions for drawing tablets, so you can set the tablet cursor to be confined to the size of a monitor

[floral meteor]

didn't you say cli before? now you're making a gui

[vestal thicket]

it uses cli tools to gather data about the system

[floral meteor]

why do you need to use these specific ones?

especially in this channel, the idea is to think outside of the box

[vestal thicket]

it's just kind of how you do a lot of things when scripting using linux

[vestal bramble]

mmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm

[floral meteor]

no spam pls

one m will do very nicely

don't need 51 of them

[vestal thicket]

the reason is that xinput and xrandr are both tools that will be installed on basically any desktop linux setup my app is intended for and i don't have to worry about native code at all

both heavily documented and relied upon

[floral meteor]

it won't be installed on windows desktops

probably not macos either

although macOS doesn't really count

[vestal thicket]

i don't really know of any other way get information about x displays and x input devices using python at the time being. also no, it is exclusively a tool for configuring drawing tablets on systems using X

there is no need for it on windows or mac os

[floral meteor]

why do you need to configure drawing tablets on systems using X?

[vestal thicket]

because there aren't any existing graphical apps to do the task when it comes to X

[floral meteor]

make a graphical app to do the task, then.

then a graphical app would exist, thus solving your problem

[vestal thicket]

just to setup the dimensions your drawing tablet is active in in relation to your monitors, you have to spend 30 minutes to an hour researching which configuration files you have to configure, calculate the transformation matrix, and either modify the x config file or use xinput to set it non-permanently constantly

that's what i'm doing

the app has to grab information about the currently active X displays and the attached input devices and the easiest way to do that it using the cli tools xrandr and xinput, at least as far as i'm aware. you run those applications as subprocesses and grab and interpret the output in the python app

just as you might have a gui app that has to make and interpret database calls; there are all sorts of different sources you can use for data

[floral meteor]

This is an extremely specialist topic, and unfortunately for you, it appears you are the highest authority on this topic currently present.

as for string reading, I'm assuming your python script takes as input the output of the xrandr command?

there does appear to be patterns in these lines

[vestal thicket]

yeah basically it reads the output of stdout, which is generally multiple lines of text

[vestal bramble]

ok

[vestal thicket]

i was just wondering if anyone knew of any tools to make picking out data from lines of text easier than

[floral meteor]

umm i need to know what these numbers and abbreviations mean

[vestal thicket]

having to do it manually using just the basic string manipulation functions

[floral meteor]

+HDMI-0 what does that mean?

what do the pluses do?

[vestal thicket]

it's just for display

[floral meteor]

what's the difference between the second attribute and the last attribute?

other than the prefix of +?

[vestal thicket]

that's the thing; this stuff is formatted for reading, not as a data structure

why it lists the name of each monitor twice i don't know

[floral meteor]

I know, but it can still be used to build a dataclass

[vestal thicket]

that's actually a good question

i just pick out the data i care about

[floral meteor]

what are the plus-separated values after the last forward-slash mean?

I can't discern their meaning straight from context

I'm gonna give them some dummy name for now

[vestal thicket]

what line?

[floral meteor]

all of them have it in the example

[vestal thicket]

oh that's offset data

x and y offset

[floral meteor]

what's the third value then?

I'll just store it all in offset_data as a tuple

[vestal thicket]

it's something like, (x resolution)/(x physical width)x(y resolution)/(y physical height)+(x offset)+(y offset)

that's my understanding anyway

[floral meteor]

well that doesn't make sense i assumed the middle axb would be a is screen width, b is screen height and the first value in the forward slash separated series is corner-to-corner size

[vestal thicket]

the offset being the displays offset in pixels from the virtual desktop's 0,0 starting coordinate

in the case of a/bxc/d+e+f , b is the physical width in millimeters

it's not a mathematical formula or anything

it's just formatted output

er, formatted data

[floral meteor]

This is the sort of dataclass I can build from that input (untested)

import attr
@attr.s
class Monitor:
  """default values are example values"""
  id:int = attr.ib(2)
  monitor_name:str = attr.ib("HDMI-0")
  monitor_identifier:str = attr.ib("+HDMI-0")
  resolution:tuple[int,int]=attr.ib((1280,720))
  size:tuple[int,int] attr.ib((339,190))
  offset:tuple[int,int]=attr.ib((3200,0))

n = int(input().partition(': ')[2])
monitors = []
for i in range(n):
  line = input().lstrip().split(' ')[1:]
  mon_id, yay, name = line
  width, this, that = yay.split('/')
  x, height = this.split('x')
  y, *o = that.split('+')
  monitors += [Monitor(i, name, mon_id, (x, y), (width, height), o)]

refactor at will

oh wait gotta store these instances somewhere

[vestal thicket]

interesting

[floral meteor]

assuming stdin is piped

[vestal thicket]

oh that's from attr ok. i was like "what is that"

[floral meteor]

boilerplate avoidance

[vestal thicket]

yeah

[floral meteor]

you likey?

oh wait

I just found a fatal flaw in my script

I forgot to do typing

[vestal thicket]

i don't know yet if you'll use that way of writing it but i'm certainly learning a lot just from reading through it and looking up the stuff i don't know/hadn't considered so thanks

[floral meteor]

import attr
@attr.s
class Monitor:
  """default values are example values"""
  id:int = attr.ib(2)
  monitor_name:str = attr.ib("HDMI-0")
  monitor_identifier:str = attr.ib("+HDMI-0")
  resolution:tuple[int,int]=attr.ib((1280,720))
  size:tuple[int,int] attr.ib((339,190))
  offset:tuple[int,int]=attr.ib((3200,0))

n = int(input().partition(': ')[2])
monitors = []
for i in range(n):
  mon_id, yay, name = input().lstrip().split(' ')[1:]
  width, this, that = yay.split('/')
  x, height = this.split('x')
  y, *o = map(int,that.split('+'))
  monitors += [Monitor(i, name, mon_id, (int(x), int(y)), (int(width), int(height)), o]

correctly typed this time(hopefully)

[vestal thicket]

it's certainly a lot more compact

[floral meteor]

I don't like scrolling unless the code has significant size

in this compact form, I can non-linearly analyse the code

[vestal thicket]

right

[long hamlet]

Ah yes accessing the display buffer

[floral meteor]

ah yes, interrogating the display for its personal information

[vestal thicket]

honestly, reading and writing to buffers is easier than text manipulation stuff imo lmao

to image/screen buffers i mean

[floral meteor]

to me, 'buffer' describes the Rock compared to Elon Musk.
The Rock is very buff. much buffer than Elon.

[long hamlet]

More efficient if you can do that

[floral meteor]

I might've gotten resolution and size the wrong way around, whoops

but refactor at will, as I said earlier

[vestal thicket]

thanks!

[long hamlet]

"The Frame Buffer Device API — The Linux Kernel documentation" https://www.kernel.org/doc/html/latest/fb/api.html

[vestal thicket]

yeah that's when i'd just use sdl, lol

or opengl

[long hamlet]

Because display buffer is moar predictable than text formats lol

[floral meteor]

for some reason it only works horizontally XD

nvm it doesn't work properly in any direction I algorithmed wrong

and I think using ctypes to mutate the cell into its parent class is why it's randomly crashing without an error message

[floral meteor]

it works now.

[wary shoal]

I have come to bleed my eyes

[wary shoal]

Beautiful

[pure dew]

why are you fetching monitor information for a CLI game?

[floral meteor]

apparently, cli tools are being used for a gui

so close...

wait

I had a valid move still

what's so confusing about not a or not b or a==b for x in self.matrix for a, b in (x[:2],x[1:3],x[2:]) that python cannot understand?

a.value==b.value might work better

and ```py
print ('\x1b[H\x1bc')

works really great to reinitialise the terminal.

[floral meteor]

140 lines of dense mf

and it looks like this and functions without bugs.

[finite blaze]

how are u rendering these colors?

[knotty delta]

ansi escape code ig

[knotty delta]

o damn cool stuff

esoteric source pls 🥺

[floral meteor]

https://paste.pythondiscord.com/mesikosexa

only works on windows

only tested on win11

[knotty delta]

[pure dew]

only works on windows?

this isnt what you were querying the monitor for, is it?

[wispy wedge]

any way I could make my own images with just an rgb code? I'd just like to generate a 32x32 image that's purely one singular color without 3rd party dependencies. I don't care about any metadata or filetype, it just needs to be renderable in a web browser

[severe canyon]

why not just to that in html and css?

[wispy wedge]

I don't actually have access to any html or css, the api just forces me to send image data

[long hamlet]

Lol my library might help

https://github.com/TechnoTanuki/Python_BMP

[wispy wedge]

I previously used PIL but that's an extra dependency for no reason

[wispy wedge]

damn mate, where the structure at

[severe canyon]

<svg width="32" height="32">
  <rect width="32" height="32" style="fill:rgb(0,0,255)"/>
</svg>

use svg?

[wispy wedge]

yeah I can't do svg, it doesn't count it as an image

idk why

first thing I tried

[severe canyon]

hm

[long hamlet]

https://github.com/TechnoTanuki/Python_BMP/blob/5b371a2e564a2c326901a4cc28f8284cfdd8c099/Features_Speedtest.py

Lots of things there have to trim it down lol

mx=1024 my=768 bmp=b.newBMP(mx,my,24)

How to make an in memory bmp

[wispy wedge]

yeah idk mate. At that point I could just keep PIL

also no offense but this isn't a lib I'd want to have in my dependencies tbh

[long hamlet]

No worries lol and yeah i understand

[long hamlet]

I can write custom code that will solve your issue thou one python file no dependencies

But it is lafe night here and I have to sleep...it will still be a bitmap

[wispy wedge]

I wonder if it even supports bitmaps tbh

like I was hoping to somehow figure out how tf to make a png or jpg

[long hamlet]

Is that for a favicon?

[wispy wedge]

an api requires raw image data to display an image

[long hamlet]

Jpg is more complex has lossy compresion

Png has alpha channel

[wispy wedge]

gif should probs work too but idk if it's harder

[long hamlet]

And compression

Use library if you want those formats

I can work at it but not for gratis lol

[rapid sparrow]

i figured a cool "trick" with context managers that abuses sys._getframe

!e

from sys import _getframe                                               
class AssignmentTracker:
  def __enter__(self):                                                  
    self._fr1 = _getframe(1)                                            
    self._fr1_orig = dict(self._fr1.f_locals)
  def __exit__(self, *err):                                             
    self.changed = {                                                    
      k:v for k,v in self._fr1.f_locals.items()                         
      if k not in self._fr1_orig
      or self._fr1_orig[k] is not v
    }
    print(f"New or changed variables: {self.changed}")

with AssignmentTracker():
  name = "John"
  import random
  age = random.randint(1, 50)
  print(f"{name}'s age is {age}")```

[night quarry]

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

001 | John's age is 47
002 | New or changed variables: {'name': 'John', 'random': <module 'random' from '/usr/local/lib/python3.10/random.py'>, 'age': 47}

[rapid sparrow]

there is a framework that uses this kind of "with block" to accomplish a similar outcome

but I doubt they are doing it with frames

it's like

with model:
  n = 3
  a = 7
  b = (2, 6)
assert model.n == 3
assert model.a == 7
assert model.b == (2,6)

how would you do that ?

[sly root]

maybe custom globals with context manager?

[rapid sparrow]

interesting - how would that work?

[sly root]

fatal helped me alot when i was writing brainfuck interpreter, custom globals were used to make it work

give me a sec

here https://paste.zeronet.space/p7B3nBJBvh.py

you basically need to add if or match check and then check if requested name == "model" or something else

[noble nexus]

hey guys, I was wondering
what happens when you do a = 2, where does the 2 go in terms of Python?
If I do __builtins__ = None then I can't use any pre-existing function or variable, but then if I do a = 2 it works and I can access it : a yields 2
I'd assume they go in locals / globals. They still do but it's just I don't have access to it anymore since my "locals" string can't resolve to the actual object through __builtins__?
How would I go about still somehow accessing locals or globals, or wherever variables are?
(don't hesitate to ping me)

[fleet bridge]

2 is int literal
you cant change behaviour of literals
2 always means instance if int type

when you do __builtins__ = None, you dont delete all builtin types and objects, you just remove reference to builtins.__dict__

!e

__builtins__ = type('',(),{'__getitem__':print})()
a
b
c
hello_world

[night quarry]

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

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

[fleet bridge]

🤔

it works in repl

[upbeat sonnet]

!e python print(" ".join([chr(i) for i in range(65, 65 + 26)]).strip(" "))

[night quarry]

@upbeat sonnet :white_check_mark: Your eval job has completed with return code 0.

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

[upbeat sonnet]

🏃‍♂️

[fleet bridge]

!e

print(' '.join(map(chr,range(65,91))))

[night quarry]

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

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

[upbeat sonnet]

pro

[noble nexus]

!e ```py
print(*map(chr,range(65,91)),sep=' ')

[night quarry]

@noble nexus :white_check_mark: Your eval job has completed with return code 0.

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

[noble nexus]

one char less 👀

[fleet bridge]

nice

[upbeat sonnet]

king fr fr

[noble nexus]

I still don't know how to monkey my way around a __builtins__ = None though 😔

(I can't write __ :kappa:)

[upbeat sonnet]

what is the * operator

[noble nexus]

unpacking
well I think it has some other weird official name

[fleet bridge]

>>> int
<class 'int'>
>>> __builtins__ = None
>>> int
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: 'NoneType' object is not subscriptable
>>> del __builtins__
>>> int
<class 'int'>

[noble nexus]

yeah but you used __

👀

I know, I have weird problems

[fleet bridge]

copy this: __ every time you need to write this

[noble nexus]

there are other various conditions, sadly
can't import stuff (like ctypes)

[upbeat sonnet]

nice

[noble nexus]

I'm also interested if you have some convoluted way to set a read-only attribute 👀
desperate

[fleet bridge]

change __setattr__

[noble nexus]

yeah but you used __ 👀

[fleet bridge]

[noble nexus]

hehehe

[fleet bridge]

!e use this to create function names

dunder = dir(0)[0][:2]
print(dunder)

[night quarry]

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

__

[noble nexus]

yeah but you can't call dir with builtins = None 👀
sorry that's a really tough setting don't bother

[polar plover]

so we got a challenge?
so rules are __builtins__ = None and no using __?

[noble nexus]

there are more but it's an actual challenge so I should just solve it by myself

I mean you can try with these 2 rules, no other globals(/locals) and no import

in fact you can import but that first rule (as long as it stands) de facto prevents it the way Python works

[polar plover]

!e interesting py __builtins__ = None class Foo: pass

[night quarry]

@polar plover :warning: Your eval job has completed with return code 0.

[No output]

[polar plover]

wait wat

[noble nexus]

yep

don't know why but it's how things are
various stuff breaks, and apparently class too

[polar plover]

differences between scripts and repel are annoying xD

>>> __builtins__ = None
>>> class Foo: pass
...
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: 'NoneType' object is not subscriptable```

!e py __builtins__ = None print("does this work still?")

[night quarry]

@polar plover :white_check_mark: Your eval job has completed with return code 0.

does this work still?

[polar plover]

okay, wtf

[noble nexus]

you only have the new stuff you define as a consequence, plus keywords

[polar plover]

why is it working differently depending on if you run it in a script, or if you are using the repl

[noble nexus]

I don't know how the repl works but seems fair it would make assumptions like "the environment isn't f-d up"

[upbeat sonnet]

bro i thought i had no life 💀

[polar plover]

the no __ is really mean ...

[noble nexus]

I have found numerous and very different approaches that all wind up in either __ or f-d up builtins unexpectedly f-ing some other things up
a bit disappointed
ultra-low-level things are behind imports
non-__ low-level things are read-only

[polar plover]

and all the fun attributes are behind __

[severe canyon]

you dont need the seperator? 🤨

[noble nexus]

I'm tired

[severe canyon]

understandable

[upbeat sonnet]

wanted the space

between letters

!e python print(*map(chr, range(65,91)))

[night quarry]

@upbeat sonnet :white_check_mark: Your eval job has completed with return code 0.

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

[upbeat sonnet]

oh ic

[severe canyon]

:)

[umbral prism]

Can any of you oneline this ?

def pack(data: dict):
    d = []
    for k, v in data.items():
        v = str(v).lower() if isinstance(v, bool) else v
        v = "undefined" if v == None else v
        d.append(f"{k}={v}")
    return "&".join(d)

I don't care if its hard ass complicated shit to read, all i care is that it works :^)

[dark wharf]

def pack(data:dict): return "&".join(f'{k}='+{None:"undefined",True:"true",False:"false"}.get(v,str(v)) for k,v in data.items())

seems to work pretty well

[umbral prism]

oh thanks

[maiden river]

!e

speak = print
speak("hi")

[night quarry]

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

hi

[maiden river]

i wish we could define new keywords that easily
inthething = in
candy = for

could there be some way 👀

some class function operator magic ?

[sick hound]

Hi Gentlemen i wanna ask a question now , i'm solving problems on codeforces but i always find very difficult problem for me and i saw the answer of others and i don't know how they are thinking in it in that way i want to learn that skill can any one help ?🤦‍♂️

[floral meteor]

probably the wrong channel to ask for help, see #❓｜how-to-get-help

[sick hound]

Oh sorry

[floral meteor]

that would be the parser. Unfortunately everything is parsed before anything runs, and nothing runs if a syntaxerror is determined.
It is possible to change some syntax rules before importing something containing the python code with the desired syntax.

[floral meteor]

It is, however, possible and fun to build esolangs that can parse within python syntax rules.

[earnest wing]

Because every repl line is a separately compiled unit

and builtins are "basically" propagated at compile time

er, wrong word

inhabited?

idk

[floral meteor]

for example I could make something like:

build_class. myObject[
  init_method (a, b)[
    setattr. a=a,
    setattr. b=b,
  ],
  method. add .magic="++" (other)[
    setattr. a=a+other.a,
    setattr. b=b+other.b,
  ],
  method. sub .magic='--' (other)[
    setattr. a=a-other.a,
    setattr. b=b-other.b,
  ],
  property. abs[
    ret. math.sqrt(a**2 + b**2)
  ],
] .endclass

function. main[
  set. x == myObject (3,4),
  set. y == myObject (1,2),
  x ++ y,
  delete. y,
  echo. x,
] .endfunction

main

to output

<MyObject x=4, y=6; at 0xc06198e3a012105>

[floral meteor]

all of the above should be technically possible to implement, but some of it would be very painful.

including building the methods out of the arrays of sentinal objects, and hacking globals.__getitem__ almost irrecognisably

[floral meteor]

conservatively written

from msvcrt import*
class Quit(RuntimeError):quit=lambda:(()for()in()).throw(Quit)
@type.__call__
class C:
  def __init__(s):s.c=[0,0,0]
  def __str__(s):return f'\x1b[H\x1b[38;2;{";".join(map(str,s.c))}m{s.c}   \x1b[0m\n\x1b[48;2;{";".join(map(str,s.c))}m{"": >{len(str(s.c))}s}\x1b[0m   \n\x1b[48;2;{";".join(map(str,s.c))}m{s.c}\x1b[0m   \n\x1b[48;2;{";".join(map(str,s.c))}m{"": <{len(str(s.c))}s}\x1b[0m   \x1b[A'
  def __add__(s, n):s.c[n]+=1;s.c[n]=min(255,s.c[n])
  def __sub__(s, n):s.c[n]-=1;s.c[n]=max(0x0,s.c[n])
try:
  print('[0, 0, 0]\n\n[0, 0, 0]')
  while 1:
    if kbhit():{'w':lambda:C+1,'a':lambda:C-0,'s':lambda:C-1,'d':lambda:C+0,'q':lambda:C-2,'e':lambda:C+2,'\x1b':Quit.quit}.get(getch().decode(),lambda:0)();print(C)
except Quit:print()

!e ```py
def concatenate(a:str, b:str):
return a + b

print((lambda f:setattr(f,'annotations',{'a':str,'b':str})or f)(lambda a,b:a+b).annotations == concatenate.annotations)

[night quarry]

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

True

[floral meteor]

demonstrating type hints for oneline code

you can also use a similar concept to "name" an anonymous function (lambda), by setting __name__

[floral meteor]

the algorithm behind 2048 can pretty much be summed up here

def _swipe(self,i,d):
    m=[[self[i,q]for q in range(4)],[self[i,q-1]for q in range(4,0,-1)],[self[q-1,i]for q in range(4,0,-1)],[self[q,i]for q in range(4)]][d];q=[*map(int,m)]
    for j in range(3):
      y=q[j:]
      while not y[0]and any(y[1:]):q+=[q.pop(j)];y+=[y.pop(0)]
    for j in range(3):
      y=q[j:]
      if not[*{*y[:2]}][1:]:q[j]+=q.pop(j+1);q+=[0]
    for j,n in enumerate(q):
      if n:self[m[j]]=Cell(*m[j],n)
      else:self.empty(*m[j])
    return q!=[*map(int,m)]

[floral meteor]

everything else is just interaction

[floral meteor]

I shrunk it to 80 lines, reguessed the colours and removed the frame inspection causing backwards dependency.
https://paste.pythondiscord.com/usilopojib.py

[severe canyon]

that code for drawing the board must be reducable

[sly root]

very painfully if 0deps

else with any table building lib it can be reduced to 5-9 lines of code

[sick hound]

I need help I’m new python

[vale tangle]

!e

for i in range(10):
    print(f"Hacking your mom... {(i+1)*10}% complete")
from random import randint as r
def n():
    return r(0,128)
print(f"Hacked successfully, ip: {n()}.{n()}.{n()}.{n()}")

[night quarry]

@vale tangle :white_check_mark: Your eval job has completed with return code 0.

001 | Hacking your mom... 10% complete
002 | Hacking your mom... 20% complete
003 | Hacking your mom... 30% complete
004 | Hacking your mom... 40% complete
005 | Hacking your mom... 50% complete
006 | Hacking your mom... 60% complete
007 | Hacking your mom... 70% complete
008 | Hacking your mom... 80% complete
009 | Hacking your mom... 90% complete
010 | Hacking your mom... 100% complete
011 | Hacked successfully, ip: 11.62.103.124

[knotty delta]

damn

y'all wizards

[floral meteor]

it most definitely is soft codable, I'm just feeling lazy.

[rapid sparrow]

cool!!!'=''';

[sick hound]

nice code

[worn lion]

why isnt the import at the top lol

[sick hound]

whats the *

[floral meteor]

unpacking an iterable into an indefinite amount of arguments to print

[floral meteor]

if you're new to python, any channel other than this one would be a great start

[floral meteor]

__enter__

• save copy of globals
• clear globals

__exit__

• save copy of new globals, and reset globals to original saved copy
• construct dataclass using last globals save.

[floral meteor]

because it doesn't need to be

[worn lion]

it looks wrong but then everything in this channel also looks wrong

[floral meteor]

!e ```py
@type.call
class model:
def enter(self):
import sys
self.reset = sys._getframe(1).f_globals.copy()
sys._getframe(1).f_globals.clear()
def exit(self, *_whatever):
import sys
k = sys._getframe(1)
r = type('model',(),k.f_globals)()
k.f_globals.clear()
k.f_globals.update(self.reset|{'model':r})
return 0

with model:
n, a = 3, 7
b = 2, 6
assert not[{(model.n,model.a),(3,7)}][1:]
assert not[{model.b,(2,6)}][1:]
print(model)

[night quarry]

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

<model object at 0x7f7f2061bee0>

[floral meteor]

that's one interpretation, anyway

!e ```py
def implicit_equals(ab):return not[{ab}][1:]
fork,in[(0**0,1,1>0),((),()),(0,0<0),((1,2),)*3,(1,2,3)]:print(implicit_equals(*k))

[night quarry]

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

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

[rapid sparrow]

is this basically what i did?

[floral meteor]

assuming model is allowed to preexist

what was your solution?

if it's supposed to be a dataclass constructor some globals getitem hackery using bytecode inspection might tell it to create an intermediate context manager to build the class from the parallel globals

[earnest wing]

I did that but with classes instead of context managers

and i guess literally plugging the contents into z3 lol

[rapid sparrow]

no mine was nothing like that, merely a context manager thing

[rapid sparrow]

sounds cool

you still have the code?

it's a cool idea but i think it's just beyond the way python is intended to be used

so might not be bulletproof

there's something called contextvar as well

!d contextvars

[night quarry]

contextvars

This module provides APIs to manage, store, and access context-local state. The ContextVar class is used to declare and work with Context Variables. The copy_context() function and the Context class should be used to manage the current context in asynchronous frameworks.

Context managers that have state should use Context Variables instead of threading.local() to prevent their state from bleeding to other code unexpectedly, when used in concurrent code.

See also PEP 567 for additional details.

New in version 3.7.

[rapid sparrow]

it starts off sounding interesting at least

the api is practically non-existent

!e
import contextvars
print(dir(contextvars))

[night quarry]

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

['Context', 'ContextVar', 'Token', '__all__', '__builtins__', '__cached__', '__doc__', '__file__', '__loader__', '__name__', '__package__', '__spec__', 'copy_context']

[rapid sparrow]

i wrote a similar function

to execute code in another namespace

like in the namespace of a module or class

it's not bulletproof either but useful

[twilit mountain]

Thanks for messing with my brain, half the brain cells are dead now

[earnest wing]

Uh it should be on github

https://github.com/RocketRace/easy_z3

yeeeah there it is

the benefits of using classes is that you don't need to make a mess with globals & locals awareness; classvars are relatively simple

the downsides is that it's uglier and editors hate it with a passion

class _(Model): vs with Model(...):

[vale tangle]

!e

print(True,True,True==(True,True,True))

[night quarry]

@vale tangle :white_check_mark: Your eval job has completed with return code 0.

True True False

[floral meteor]

looking forward to what I'm gonna do with the other half?

[floral meteor]

!e dummy integer ```py
@type.call
class a:value=0;index=int=hash=lambda s:s.value

a.value = 4
print([*range(3),*range(3)][a])

[night quarry]

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

1

[floral meteor]

!e which can make a really cursed ```py
@type.call
class a:
value=47;index=int=hash=lambda s:s.value;iter=lambda s:s
def next(self):
a.value+=1
if a.value==58:raise StopIteration
return print

*map(lambda :(chr(a)),a),

[night quarry]

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

001 | 0
002 | 1
003 | 2
004 | 3
005 | 4
006 | 5
007 | 6
008 | 7
009 | 8
010 | 9

[floral meteor]

if I didn't stop at 58 this would indefinitely iterate, like a really cursed event loop

wait did I just discover async mechanics?

!e or I could make accessing it as an integer increase its integer value ```py
@type.call
class a:
value=0;index=int=hash=lambda s:setattr(s,'value',s.value+1)or s.value;iter=lambda s:s

print(*map(list,map(range,[a]*10)),sep='\n')

[night quarry]

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

001 | [0]
002 | [0, 1]
003 | [0, 1, 2]
004 | [0, 1, 2, 3]
005 | [0, 1, 2, 3, 4]
006 | [0, 1, 2, 3, 4, 5]
007 | [0, 1, 2, 3, 4, 5, 6]
008 | [0, 1, 2, 3, 4, 5, 6, 7]
009 | [0, 1, 2, 3, 4, 5, 6, 7, 8]
010 | [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]

[stuck peak]

somebody posted a comment on a video I was watching yesterday saying that people who scraped stuff from the web might need 10's or even 100's of proxies to " avoid detection.. "
🤔
It just had me wondering what might this person be scraping that they would need so many proxies . .

[floral meteor]

I scraped a pet adoption website quite directly to determine which cat of theirs was the fattest. Shouldn't I have?

[rapid sparrow]

wow

()=_ # dump variables into globals```

that is truly a hidden gem

[rapid sparrow]

and people are doing this very thing

something where the site owner really doesn't want them to, in all likellihood

[rapid sparrow]

no no

if you were able to do it directly, that's a good sign you're not going against the owner's wishes

[pure dew]

i'm scared yet impressed

[floral meteor]

When I read that my mind says ```py
assert not len(_)

[quartz wave]

!e ```py
notﾠdefined

[night quarry]

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

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

[quartz wave]

yes

[twilit mountain]

Learnt a lesson: Never look into this channel again as long as there ain't any pings from here

[pure dew]

@twilit mountain https://gitlab.a-sketchy.site/AnonymousDapper/traitor/-/blob/master/examples/from_into.py

now you'll know not to look in this channel at all :^)

[twilit mountain]

No no no u just gave me anxiety of what's in the link but I ain't clicking it

[pure dew]

its just python

a couple small classes

[twilit mountain]

Wait, this wasn't that bad, but what's with the >>>

[pure dew]

Binds a trait to a type for the impl

it's the "target type", so to speak

this is basically the Rust trait system

minus a few things

[twilit mountain]

O

[south tundra]

!e py a = print a("test")

[night quarry]

@south tundra :white_check_mark: Your eval job has completed with return code 0.

test

[south tundra]

!e py a = for b = in c = range a _ b c(1,2): print(_)

[night quarry]

@south tundra :x: Your eval job has completed with return code 1.

001 |   File "<string>", line 1
002 |     a = for
003 |         ^^^
004 | SyntaxError: invalid syntax

[south tundra]

aw

[floral meteor]

this is python, not c

[floral meteor]

what you can do is write a turing complete string interpreter then the string can be whatever.

[sick hound]

!e

assert not len('l')

[night quarry]

@sick hound :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

[sick hound]

!e

assert not len('')

[night quarry]

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

[No output]

[sick hound]

can be fixed easy

#define a for
#define b in
#define c range
a _ b c(1,2):
    print(_)

[sick hound]

This

[unreal echo]

I come here occasionally when I'm feeling too smart and need to lower my confidence to humanly levels.

[fervent umbra]

I come here occasionally when I want to realise I have no idea how python works

[fiery crane]

i come here occasionally to realise I don't know how to program

[sick hound]

!e

#define a for
#define b in
#define c range
a _ b c(1,2):
    print(_) 

[night quarry]

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

001 |   File "<string>", line 4
002 |     a _ b c(1,2):
003 |       ^
004 | SyntaxError: invalid syntax

[sick hound]

What does _ look like here?

!e
test={"cool": "no way"}
() = test
print(globals())

[night quarry]

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

001 | Traceback (most recent call last):
002 |   File "<string>", line 2, in <module>
003 | ValueError: too many values to unpack (expected 0)

[astral rover]

class _(Model):

its in the reply chain

[sick hound]

And it will take the local vars of that model and put it into globals?

[sick hound]

Also what would be after this? Just a pass?

[astral rover]

no look at the code in the github repo

[sick hound]

It can be any target, if he wants it done fast then more proxies are needed

Sometimes scraping at 1k rps

[rapid sparrow]

yeah lol

i still dont get what it does .

[rugged sparrow]

() = obj calls obj.__iter__ so obj.__iter__ can dump values into globals

!e ```py
class O:
def iter(self):
print('i got called')
return
yield

obj = O()
() = obj```

[night quarry]

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

i got called

[fleet bridge]

if False: yield

[fleet bridge]

This code is used somewhere in std lib

iirc in the Iterable protocol

[rapid sparrow]

i got some esoteric code by accident

def some_beartyped_function():
  if __beartype_pith_0 is not __beartype_raise_exception and not isinstance(
    __beartype_pith_0, str
  ):
    __beartype_raise_exception(
      func=__beartype_func, pith_name="package_name", pith_value=__beartype_pith_0
    )
  __beartype_pith_0 = kwargs.get("module_name", __beartype_raise_exception)
  if __beartype_pith_0 is not __beartype_raise_exception and not isinstance(
    __beartype_pith_0, str
  ):
    __beartype_raise_exception(
      func=__beartype_func, pith_name="module_name", pith_value=__beartype_pith_0
    )
  __beartype_pith_0 = kwargs.get("func", __beartype_raise_exception)
  if __beartype_pith_0 is not __beartype_raise_exception and not isinstance(
    __beartype_pith_0, __beartype_object_9795712
  ):
    __beartype_raise_exception(
      func=__beartype_func, pith_name="func", pith_value=__beartype_pith_0
    )
  __beartype_pith_0 = kwargs.get("version", __beartype_raise_exception)
  if __beartype_pith_0 is not __beartype_raise_exception and not isinstance(
    __beartype_pith_0, __beartype_object_15338576
  ):
    __beartype_raise_exception(
      func=__beartype_func, pith_name="version", pith_value=__beartype_pith_0
    )
  __beartype_pith_0 = kwargs.get("hash_algo", __beartype_raise_exception)
  if __beartype_pith_0 is not __beartype_raise_exception and not isinstance(
    __beartype_pith_0, __beartype_object_14936032
  ):
    __beartype_raise_exception(
      func=__beartype_func, pith_name="hash_algo", pith_value=__beartype_pith_0
    )
  __beartype_pith_0 = kwargs.get("user_provided_hashes", __beartype_raise_exception)
  if __beartype_pith_0 is not __beartype_raise_exception and not isinstance(
    __beartype_pith_0, set
  ):
    __beartype_raise_exception(
      func=__beartype_func,
      pith_name="user_provided_hashes",
      pith_value=__beartype_pith_0,
    )
  __beartype_pith_0 = kwargs.get("registry", __beartype_raise_exception)
  if __beartype_pith_0 is not __beartype_raise_exception and not isinstance(
    __beartype_pith_0, __beartype_object_549598516568
  ):
    __beartype_raise_exception(
      func=__beartype_func, pith_name="registry", pith_value=__beartype_pith_0
    )
  __beartype_pith_0 = __beartype_func(*args, **kwargs)
  if not isinstance(__beartype_pith_0, __beartype_object_549601113792):
    __beartype_raise_exception(
      func=__beartype_func, pith_name="return", pith_value=__beartype_pith_0
    )
  return __beartype_pith_0```

[fleet bridge]

Wtf

[rapid sparrow]

the function I was using to look at tracebacks tries to decompile ones that don't have a source file, like this was File <string>

the function had a @beartype on it

was kind of shocked to see all that get printed out lol

[pure dew]

what is beartype??

[astral rover]

a runtime type checker

[flint nacelle]

Hey there, any good tips for how to upload a game on telegram written on Python? Thanks in advance for who will reply 🙂

[floral meteor]

@flint nacelle ❌ Your eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 1, in <module>
003 | WrongChannelError: please refer to the channel [#❓｜how-to-get-help](/guild/267624335836053506/channel/704250143020417084/) for more information

#❓｜how-to-get-help

[quartz wave]

>>> import discord
>>> guild = discord.get_guild('Python')
>>> channel = guild.get_channel('esoteric python')
>>> channel.get_help()
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Python310/Lib/site-packages/discord", line 344, in get_help
NoHelpException: cannot get help from channel; do you mean to be in channel "how to get help"?

i wonder why 🤔

[thorny tapir]

help I cannot be helped

[split salmon]

lol

[golden finch]

https://stackoverflow.com/questions/71315838/when-i-modify-builtins-to-a-custom-class-it-only-works-in-shell
This seems like the sort of thing you guys might be able to help with - reason I'm not posting it in a help channel is that it's very niche, and there isn't a strictly defined solution, just making __missing__ work... somehow

Full disclosure: This is my stackoverflow post

[knotty delta]

selfbot confirmed

[earnest wing]

Builtins are a bit special in that they persist / are cached throughout a single execution unit
In a repl each line is its own unit, while in a program you can only create new units with exec, etc.

You can obviously do some interior mutability tricks, though, such as .clear()ing, or using ctypes etc. to patch the actual object in memory

[fleet bridge]

How can i make Return() function?

Return = ...
def f():
  Return(42)
assert f() == 42

Maybe i should set trace function to parent frame and then execute return opcode?

Possible use case:

def some_decorator():
  @Return
  def wrapped(): ...

# instead of 
def some_decorator():
  def wrapped(): ...
  return wrapped

It can be used in ternary operator to return from function if some invariants are broken

[twilit mountain]

Huh?

[twilit mountain]

would have been sooo fun if this worked lol

[floral meteor]

make it a subclass of Exception and make a decorator to upgrade a function to support Return

[sick hound]

troll

[floral meteor]

!e ```py
class Return(BaseException):
def init(self,*data):
try:self.data,=data
except:self.data=data
super().init(repr(data))

def implementsReturn(f):
def inner(*a,**k):
try:return f(*a,**k)
except Return as e:return e.data
return inner

@implementsReturn
def main():
a, b = 60, 9
raise Return(a+b)

result = main()
print(result)

[night quarry]

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

69

[floral meteor]

with this pair of implementsReturn and Return, Return now returns from the next frame that implementsReturn, thus supporting returning from multiple frames at once.

!e another example ```py
class Return(BaseException):
def init(self,*data):
try:self.data,=data
except:self.data=data
super().init(repr(data))

def implementsReturn(f):
def inner(*a,**k):
try:return f(*a,**k)
except Return as e:return e.data
return inner

@implementsReturn
def main():
def inner(n):
raise Return(60+n)
inner(9)

print(main())

[night quarry]

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

69

[floral meteor]

also supports building a tuple just like builtin return

I'll leave making an example of that as an exercise for the reader.

[twilit mountain]

😔

[formal sandal]

https://paste.pythondiscord.com/lubifivuru

def foo():
    n = 0

    goto = Goto()
    while goto:
     with goto as label:
        if label("fizz"):  print("fizz", n)
        if label("buzz"):  print("buzz", n)
        if True:           n += 1
        if n == 5:         goto.hell()
        if n > 3:          goto("fizz")
        else:              goto("buzz")

foo()
``` ```
fizz 0
buzz 0
buzz 1
buzz 2
buzz 3
fizz 4
buzz 4

[knotty delta]

imagine using this in job interviews

[formal sandal]

if an interviewer asks me to invert a binary tree on a whiteboard, they 100% deserve this tbh 🙂

[rough mulch]

Woah

[rugged sparrow]

!e ```py
BYTES_HEADER = bytes.basicsize - 1
PTR_SIZE = tuple.itemsize
ENDIAN = ['big','little'][memoryview(b'\1\0').cast('h')[0]&0xff]

load_addr = type(m:=lambda n,s:lambda v:s(v)or n)(
(M:=m.code).replace(
co_code=b'\x88'+M.co_code[1:]
),{}
)(r:=iter(range(2**(PTR_SIZE*8-1)-1)),r.setstate)

memory_backing = bytes(PTR_SIZE)
+ id(bytearray).to_bytes(PTR_SIZE, ENDIAN)
+ bytes([255] * (PTR_SIZE - 1) + [127])
+ bytes(PTR_SIZE * 4)

memory = memoryview(load_addr(id(memory_backing) + BYTES_HEADER))

def getframe(level=0):
try:raise
except Exception as e:
frame = e.traceback.tb_frame.f_back
for () in [()] * level:
frame = frame.f_back
return frame

def Return(val):
frame = getframe(1)
offset = id(frame.f_code.co_code) + bytes.basicsize + frame.f_lasti - 1
memory[offset + 2: offset + 4] = bytes((83, 0))
return val

def test(x):
Return(x + 1)

print(test(1))```

[night quarry]

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

2

[quartz wave]

oh

!e ```py
BYTES_HEADER = bytes.basicsize - 1
PTR_SIZE = tuple.itemsize
ENDIAN = ['big','little'][memoryview(b'\1\0').cast('h')[0]&0xff]

load_addr = type(m:=lambda n,s:lambda v:s(v)or n)(
(M:=m.code).replace(
co_code=b'\x88'+M.co_code[1:]
),{}
)(r:=iter(range(2**(PTR_SIZE*8-1)-1)),r.setstate)

memory_backing = bytes(PTR_SIZE)
+ id(bytearray).to_bytes(PTR_SIZE, ENDIAN)
+ bytes([255] * (PTR_SIZE - 1) + [127])
+ bytes(PTR_SIZE * 4)

memory = memoryview(load_addr(id(memory_backing) + BYTES_HEADER))

def getframe(level=0):
try:raise
except Exception as e:
frame = e.traceback.tb_frame.f_back
for () in [()] * level:
frame = frame.f_back
return frame

def Return(val):
frame = getframe(1)
offset = id(frame.f_code.co_code) + bytes.basicsize + frame.f_lasti - 1
memory[offset + 2: offset + 4] = bytes((83, 0))
return val

def test(x):
Return(x + 1)

import dis
dis.dis(test)

[night quarry]

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

001 |  33           0 LOAD_GLOBAL              0 (Return)
002 |               2 LOAD_FAST                0 (x)
003 |               4 LOAD_CONST               1 (1)
004 |               6 BINARY_ADD
005 |               8 CALL_FUNCTION            1
006 |              10 POP_TOP
007 |              12 LOAD_CONST               0 (None)
008 |              14 RETURN_VALUE

[quartz wave]

so it just replaces the POP_TOP with a RETURN_VALUE except that it doesn't actually know that there's a POP_TOP in it

[rugged sparrow]

It doesnt need to know that its POP_TOP but python guarantees there will be at least one op after a function call.

[quartz wave]

i know that

[south tundra]

fr tho

[golden finch]

https://stackoverflow.com/questions/71332220/advanced-python-customise-handling-of-nameerror-on-referencing-undefined-variab
Y'all might be able to help with this

[golden finch]

!e

import sys
import dis
import ctypes

mutable_bytes=lambda i:memoryview((ctypes.c_uint8*len(i)).from_address(id(i)+bytes.__basicsize__-1)).cast('B')#from Python Discord

def tracefunc(frame,*useless):
    print(f'\nDisassembling frame {frame}... ')
    co_code=mutable_bytes(frame.f_code.co_code)
    #now, find the indices of all the LOAD_NAME operations in the frame
    indices=[i for i in range(0,len(co_code),2)if co_code[i] in (dis.opmap['LOAD_NAME'],dis.opmap['LOAD_FAST'])]
    names=[frame.f_code.co_varnames[co_code[i+1]]for i in indices]
    print(dis.dis(frame.f_code))
    print(names)

sys.settrace(tracefunc)

print('foo')

print(bar)

[night quarry]

@golden finch :x: Your eval job has completed with return code 1.

001 | foo
002 | Traceback (most recent call last):
003 |   File "<string>", line 20, in <module>
004 | NameError: name 'bar' is not defined

[golden finch]

that... is not the output on my machine

[rugged sparrow]

i have code that makes this work already one sec

!e ```py
from ctypes import py_object
import atexit, builtins, sys, dis

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

ob_base_p.value = missing_hook

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

init_missing_hook(sys._getframe().f_globals, lambda d, k:(k, 'var'))

print(foo)```

[night quarry]

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

('foo', 'var')

[rugged sparrow]

@golden finch ^ it changes the type of the globals dict in place and adds a __missing__ method

[golden finch]

damn, I tried exactly this thing but couldn't get it to work

how did you get it working?

by the way, that thing is intimidating

[rugged sparrow]

first off you need to make sure everything in the class your hotswapping in is being loaded with LOAD_FAST

[golden finch]

okay

[rugged sparrow]

you also need to make sure the new class has the same instance size __slots__ = () causes the class to lose all of the dynamic extra stuff (leaving it the same size as dict)

[golden finch]

remind me what slots does again?

[rugged sparrow]

it normally specifies the variable slots on an object

builtinexc was unnecessary to i removed it

[golden finch]

okay

[rugged sparrow]

youre welcome to just copy init_missing_hook, it takes in a dict instance and a func

[golden finch]

No, I'm curious how it works though - reading this:

• is presume ob_base_p is a pointer to the pyobject of the dict instance?

or is it a pointer to the pyobject of dict itself?

[rugged sparrow]

ob_base_p = py_object.from_address(id(dct) + 8) this line gets the pointer 8 bytes from the start of dct and interprets it as a pointer to a pyobject

that happens to be the tp_base pointer, which is the pointer to the base class of an object

so by swapping that, we effectively change the type(dct) from dict to missing_hook

[golden finch]

ah

and now your missing hook class

so why do you declare slots to be empty?

[rugged sparrow]

that causes python to skip adding the extra space on class instances that is normally used for a dictionary and an attribute cache

empty slots means we have no attributes, so no dictionary and no need for a cache

this is one of the things i used that missing_hook to do

[golden finch]

extra space?

[golden finch]

this is incredible, with all due respect

[rugged sparrow]

thanks, its the result of way to much time with python internals

[rugged sparrow]

yea like how normally a user class can have random attributes

[golden finch]

user class?

[rugged sparrow]

!e ```py
class a:
pass

A = a()
A.foo = 1

class b:
slots = ()

B = b()
B.foo = 1 # fails```

[night quarry]

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

001 | Traceback (most recent call last):
002 |   File "<string>", line 11, in <module>
003 | AttributeError: 'b' object has no attribute 'foo'

[golden finch]

classes can have whatever attributes - but my understanding is __slots__ limits them to a set number of attrs

ah

makes sense

[golden finch]

okay, but why do we need to make the number of attributes constant?

[rugged sparrow]

because the builtin class dict has no attributes

and we need our hotswapped class to have the same layout in memory as the dict class

[golden finch]

mhm

why?

isn't it just pointed to?

so the pointer should work regardless of the size (and yes, I know I'm wrong, but I'm curious why)

[rugged sparrow]

because the globals dict has a specific size. python knows this by checking obj.tp_base.__basicsize__ plus some other stuff. we change tp_base to a custom class without changing the actual size of the globals dict

without __slots__ = (), python checks obj.tp_base to see if there is a dict and an attribute cache, it sees that there should be, and it happily reads past the end of the original globals object into random memory

>>> class a:pass
... 
>>> a.__basicsize__
32
>>> #[8:refcnt, 8:base, 8:dict, 8:attr_cache]
>>> class a:__slots__ = ()
... 
>>> a.__basicsize__
16
>>> #[8:refcnt, 8:base]```

[golden finch]

okay, makes some degree of sense, terrifying though

[golden finch]

ah thanks

[rugged sparrow]

if you want to read more about stuff like that, look up type confusion vulnerabilities, its a similar concept

[golden finch]

okay so now you define a __missing__ method on missing_hook, but wait, it has no attributes, so how can you define a method on it?

[rugged sparrow]

attributes are stored on the instance, methods are stored on the class

[golden finch]

ah, okay

[rugged sparrow]

i actually found a second type confusion bug in python where you can coerce it into accessing an arbitrary object as a cell object

[golden finch]

TIL

once we finish discussing this code snippet, can we discuss what a cell is?

[rugged sparrow]

lol sure, ill warn you i have condensed this code a ton more

[golden finch]

intimidating

[rugged sparrow]

!e ```py
load_addr = type(m:=lambda n,s:lambda v:s(v)or n)(
(M:=m.code).replace(
co_code=b'\x88'+M.co_code[1:]
),{}
)(r:=iter(range(2**(tuple.itemsize*8-1)-1)),r.setstate)

print(load_addr(id(1)))

[night quarry]

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

1

[rugged sparrow]

all that does is let you load an arbitary address as a python object

[golden finch]

spicy

that is scary

[rugged sparrow]

but you can use it to make fake objects in memory (using bytes)

[golden finch]

okay so now how does __missing__ work?

ugh, I'm going to have to go for maybe 45 minutes now

but I'll be back

[rugged sparrow]

__missing__ is called on dictionary subclasses when the key is not found in the dictionary

I actually need to go now as well, its 1:30 am here

feel free to ping me with questions, ill answer when I wake up

[quartz wave]

sad that it only works for linux

(didn't mean to ping)

[golden finch]

oh, that's a useful bit of info

[sick hound]

Is there any hacky way to re-add python 2 style print statements back to Python 3? so ```py
print 989
print "foobar"
print 3*8-9

I gotta imagine no but you guys do amazing stuff <#esoteric-python message>

[knotty delta]

~~change t to + sneakily and add dunder add then no one will notice ~~

[golden finch]

@rugged sparrow what does that finally clause do? shouldn't it be ignored via return?

[quartz wave]

no but aph suggested a sneaky way to do it

[knotty delta]

wait.. i think so, i found a stack overflow answer, i might be able to tweak it somehow

[golden finch]

@rugged sparrow doesn't seem to work inside functions

[knotty delta]

hi, i came back to say that i failed

it's possible to read a file with the tweaked tokenize thing

but it'll error out whenever there's parsing error if i use import past

[quartz wave]

like what error

[knotty delta]

SyntaxError, idk how to avoid it, it'll always get one bcuz it really is an invalid syntax

however, my code w0rks a bit, it does weird stuff like 2to3 does

bUt it only works by running itself (gib filename as input, convert stuff, run code) but not at running it directly with python

bcuz python parses

:(

[prisma coral]

!e One issue is that print abc is a syntax error, so I think you would need to add an extra character to avoid this. The first thing that came to mind is this:

from __future__ import annotations
@type.__call__
class __annotations__(dict):
    def __setitem__(self, item, value):
        if item == "print":
            print(value.strip("'"))

print: 989
print: "foobar"
print: 3*8-9

[night quarry]

@prisma coral :white_check_mark: Your eval job has completed with return code 0.

001 | 989
002 | foobar
003 | 3 * 8 - 9

[prisma coral]

But obviously this doesn't fully meet the Py2 print statement syntax

[prisma coral]

!e Just realised that __future__.annotations was not actually what I was looking for.

@type.__call__
class __annotations__(dict):
    def __setitem__(self, item, value):
        print(value)
print: 989
print: "foobar"
print: 3*8-9

[night quarry]

@prisma coral :white_check_mark: Your eval job has completed with return code 0.

001 | 989
002 | foobar
003 | 15

[knotty delta]

i found this answer on google: https://stackoverflow.com/a/4572994/13884898

quite k00l ngl

[knotty delta]

btw readline only works on linux

[fleet bridge]

!e

class SomeMeta(type):
    __slots__ = () # note empty slots

class X(metaclass=SomeMeta):
    ...

# now X is instance of SomeMeta, X should have no dict, because SomeMeta.__slots__ is empty
X.a = 42 # but dynamic assignments still works
print(X.a)
print(X.__dict__) # and X has .__dict__ (wrapped in mappingproxy)

[night quarry]

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

001 | 42
002 | {'__module__': '__main__', '__dict__': <attribute '__dict__' of 'X' objects>, '__weakref__': <attribute '__weakref__' of 'X' objects>, '__doc__': None, 'a': 42}

[rugged sparrow]

Nah finally blocks run even after a return statement

[rugged sparrow]

What do you mean?

[sick hound]

Hey, that's pretty close I didn't know you could add dunders to annotations

[prisma coral]

Since __annotations__ is a dict, you can create a subclass of dict, override the dunders there, and then set __annotations__ to be an instance of that subclass (that's what the @type.call is for)

[sick hound]

!e how come the function annotations don't work ```py
@type.call
class annotations(dict):
def setitem(self, item, value):
print(value)
a: 111
b: 222
def f(a: 123, b: 456):
pass
f(1, 2)

[night quarry]

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

001 | 111
002 | 222

[astral rover]

cause they dont go through the global annotations

[prisma coral]

They get stored in f.__annotations__

Not the global annotations

[wise bone]

for something as primitive as a list of objects, how is it possible that

for elem in li:
  ...

could result in elem and li being the same value? i.e.

>>> elem
[{'foo': 'bar'},{'baz': 'quux'}]
>>> li
[{'foo': 'bar'},{'baz': 'quux'}]

so the iteration results in a TypeError? this is a strange one...

[earnest wing]

!e with recursive objects

x = ['']
x[0] = x
print(x) # [[...]]
for y in x:
    assert y is x
    print("wow")

[night quarry]

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

001 | [[...]]
002 | wow

[earnest wing]

@wise bone in general, a lot of simple ideas break down / hit edge cases with recursive objects

!e e.g. comparison

x=[]
y=[]
x.append(y)
y.append(x)
assert x == y

[night quarry]

@earnest wing :x: Your eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "<string>", line 5, in <module>
003 | RecursionError: maximum recursion depth exceeded in comparison

[wise bone]

makes sense, thanks!

[fleet bridge]

!e

class X:
  def __iter__(self):
    yield from (self,)

x = X()
print(x)
print()
for item in x:
    print(item)
    assert x is item

[night quarry]

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

001 | <__main__.X object at 0x7f2799fdfc10>
002 | 
003 | <__main__.X object at 0x7f2799fdfc10>

[floral meteor]

a self containing non-mutable iterable, brilliant!

[rugged sparrow]

If you abuse the garbage collector you can make a tuple that contains itself with a magic iterable

[floral meteor]

how about tuple unpacking with an iterable that grabs the tuple from frame inspection and modifies it?

[rugged sparrow]

!e ```py
import gc

class magic:
def length_hint(self):
return 1

def __iter__(self):
    for obj in gc.get_objects():
        if isinstance(obj, tuple):
            try:0 in obj
            except SystemError:
                yield obj
                break

weird = tuple(magic())
print(weird, weird in weird)```

[night quarry]

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

((...),) True

[floral meteor]

I woulda prefered

weird=*magic(),

[rugged sparrow]

That wouldn't work

The bug is specific to the tuple() constructor

[floral meteor]

ouch

it's just that *magic(), looks more magical

tuple(magic()) looks mundane

[rugged sparrow]

Yea sadly *magic(), makes a list internally, then makes a tuple from it

[floral meteor]

makes a list internally?

[rugged sparrow]

If I remember correctly yea

[floral meteor]

does that mean tuple(a) would be more optimised than *a,?

[rugged sparrow]

!e py import dis dis.dis('*a,')

[night quarry]

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

001 |   1           0 BUILD_LIST               0
002 |               2 LOAD_NAME                0 (a)
003 |               4 LIST_EXTEND              1
004 |               6 LIST_TO_TUPLE
005 |               8 POP_TOP
006 |              10 LOAD_CONST               0 (None)
007 |              12 RETURN_VALUE

[floral meteor]

This is an obfuscated snippet from a simple yet well known game I made, in the initialisation of the middle layer of code.
Who can guess what game this initialises?

class Engine:
  r=__import__('random').randint
  __getitem__=lambda s,i:(lambda i,j:s.m[j][i])(*i)
  def __init__(s):
    s.m=[[EmptyCell(i,j)for i in range(4)]for j in range(4)]
    r=lambda:(lambda l:(l(),l()))(lambda:s.r(0,3))
    a=lambda:Cell(*r(),2);a=a(),a()
    while-len({*a})+2:a[1].move(*r())
    *map(lambda i:[()for s[i]in[i]],a),

[rugged sparrow]

I'm pretty sure overhead from looking up a global name outways the internal list building

[rugged sparrow]

game of life?

[floral meteor]

game of life with 16 cells would be boring, so no

[rugged sparrow]

Minesweeper then?

[floral meteor]

again, 16 cells, boring, no

[rugged sparrow]

What's the game then

[floral meteor]

what can you think of with a 4x4 board?

and initialises with two non-empty cells?

[rugged sparrow]

No idea

[floral meteor]

so the obfuscation worked or you don't know the game

2048

the initialised board

the Engine is the middle layer, "where the algorithm is"

[rugged sparrow]

Ahh that makes sense

[floral meteor]

I layer my code like

Interface
Processor or Engine
Datatype or Cell

[rugged sparrow]

Nah I recognized the code, just forgot the game

[floral meteor]

so the code makes sense?

I can't remember if the game starts with only 2 2s or if the same random 2 or 4 initialisation applies to the starting cells

so I started with 2 2s so the user doesn't get a headstart by starting with 4s, or a setback by starting with a 2 and a 4

a few moves in

and here's the game when the goal is achieved

[floral meteor]

and here's the command line interface that's also builtin

although if I make this modification, the output can be interpreted by python as a list.

which makes this cursed option available:

[distant briar]

i am curious on why your terminal is red

[white verge]

its translucent

[floral meteor]

My desktop background is the colour of blood

[golden finch]

your desktop background is the colour of my tears after reading your code

[fleet bridge]

Wow, thats really weird trick

i hate it

It brokes invariant of tuples: tuple cant contain itself as item

[floral meteor]

I can now cheat, and use cli

[knotty delta]

python -i game !?

[knotty delta]

ooh

[floral meteor]

or python

>>> from _2048 import Interface
>>> interface = Interface()

play the game

>>> engine = interface.uninstall()
>>> engine.set(this, that, whatever)
>>> engine.up()
'swiped up   '
>>> interface.install(engine)
>>> interface.run()

continue to play modified game

[knotty delta]

ooh

[unreal echo]

would this belong here? py embed.add_field(name='Extra', value=f'Hash: {hash(member)}\nID: {member.id}\nColor: {member.color}\n{int(member.public_flags.early_supporter) * "Early supporter"}\n{int(member.public_flags.verified_bot_developer) * "Verified developer"}\n{int(member.public_flags.partner) * "Discord partner"} \n{int(member.public_flags.discord_certified_moderator) * "Certified moderator"}\n{int(not (member.public_flags.bug_hunter_level_2) and member.public_flags.bug_hunter) * "Level 1 bug hunter"}\n{int(member.public_flags.bug_hunter_level_2) * "Level 2 bug hunter"}\n{int(member.public_flags.staff and not (member.public_flags.partner)) * "Discord employee"}\n{int(member.public_flags.spammer) * "**Careful**, this user has been reported to discord for **spamming**"}')it's some code from a discord bot i made

or is my softwrap fucking with me?

[waxen stirrup]

Anyone here good with pytesseract and pyautogui dm me im tryna create a bot for something

[floral meteor]

It might help to use triple quotes so that you can use literal newlines

!e ```py
from ctypes import*
p=py_object.from_address(id(...)+8)
class SOS:
value=[*'Heo World!']
def neg(self):
self.value.insert(2,'l')
return self
def sub(self, also_self):
print(''.join(self.value))
()=self.value[2:4]=[]
p.value = SOS

...---...

[night quarry]

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

Hello World!

[unreal echo]

sounds like your trying to read text and paste it somewhere, is that right?

or maybe a spammer

[jovial monolith]

what is this cursed stuff

how does it work lol

[upbeat sonnet]

why is this more active than the help channels 😭

anyway watch and learn

!e py help(print)

[night quarry]

@upbeat sonnet :x: Your eval job has completed with return code 1.

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

[upbeat sonnet]

🏃‍♂️

[jovial monolith]

lmao what 😂

lmao

[fleet bridge]

!e

from pydoc import help
help(print)

[night quarry]

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

001 | Help on built-in function print in module builtins:
002 | 
003 | print(...)
004 |     print(value, ..., sep=' ', end='\n', file=sys.stdout, flush=False)
005 |     
006 |     Prints the values to a stream, or to sys.stdout by default.
007 |     Optional keyword arguments:
008 |     file:  a file-like object (stream); defaults to the current sys.stdout.
009 |     sep:   string inserted between values, default a space.
010 |     end:   string appended after the last value, default a newline.
011 |     flush: whether to forcibly flush the stream.

[jovial monolith]

lmao

[fleet bridge]

lmao

[jovial monolith]

>>> help.__class__.__module__
'_sitebuiltins'```

[upbeat sonnet]

mad

[floral meteor]

well the class comes from _sitebuiltins

but the instance is created in pydoc

[fleet bridge]

no

class comes from pydoc, but all code is in pydoc

[jovial monolith]

ah

wut

[fleet bridge]

there are only wrapper in sitebuiltins

[upbeat sonnet]

but why the bot cant use it normally tho

[jovial monolith]

also is it possible to have two variables pointing to the same memory address to achieve something like a pointer?

[floral meteor]

My desktop background is the colour of the tears of people who read my code:

[upbeat sonnet]

fr

[jovial monolith]

it's interesting

[fleet bridge]

🟥

[jovial monolith]

p=py_object.from_address(id(...)+8)```

what does this do?

[upbeat sonnet]

import asyncio as time
await time.sleep(10)```

[upbeat sonnet]

c stuff

[jovial monolith]

yeah but what

[floral meteor]

that accesses the memory position 8 addresses after the address of the builtin Ellipsis

oops

wrong quote

[upbeat sonnet]

it finds the memory pos of the object

[floral meteor]

this one

[fleet bridge]

gets pointer to pointer to type of ellipsis
and changes it to change type of ellipsis

[upbeat sonnet]

why no whitespaces tho

[fleet bridge]

this explanation is better

[floral meteor]

whitespace is for readability

I don't need readability, I just need funcionality

[jovial monolith]

ah

[floral meteor]

at least in this case, only the driver line needs to be read

[upbeat sonnet]

when the word functional is dysfunctional:

[jovial monolith]

so like a pointer

[floral meteor]

and the driver line is

...---...

and the function is

print("Hello World!")

[jovial monolith]

oh you're doing something with - operator

[floral meteor]

yes, the array of memory starting at ellipsis contains a pointer to the class, but I changed the pointer to the address of a user created class

which defines __sub__ and __neg__

[jovial monolith]

ah cool

lmfao

[quartz wave]

x---x
```is parsed like```py
x - --x

replace x with the ellipsis, you get```py
...---...

```py
from ctypes import*
p=py_object.from_address(id(...)+8)
```now this part, `p.value` is equal to `type(...)`
so we make a custom class ```py
class SOS:
  value=[*'Heo World!']
  def __neg__(self):
    self.value.insert(2,'l')
    return self
  def __sub__(self, also_self):  
    print(''.join(self.value))
    ()=self.value[2:4]=[]
```and set it to `p.value` ```py
p.value = SOS
```ok but how does this work? ```py
value = [*'Heo World!']
```well, it's basically the same thing as ```py
value = ['H', 'e', 'o', ' ', 'W', 'o', 'r', 'l', 'd', '!']
```remember how the thing is parsed?```py
x - --x
``` so before you subtract, you have to first evaluate the two operands, `x` and `--x`
`x` doesn't need to evaluate anything, so we don't focus on that
but `--x` we will focus on, and it basically just does this:`x.__neg__().__neg__()`
so remember that custom class from earlier? and its `__neg__` magic method?```py
def __neg__(self):
  self.value.insert(2, 'l')
  return self
``` ok so it's basically inserting `l` into the second position of `self.value`, just after the `e` right here ```py
value = ['H', 'e', 'o', ' ', 'W', 'o', 'r', 'l', 'd', '!']
                  ^
```so after the first `.__neg__()`, `self.value` is now equal to this:```py
value = ['H', 'e', 'l', 'o', ' ', 'W', 'o', 'r', 'l', 'd', '!']
```after the second `.__neg__()`, `self.value` is now equal to this:```py
value = ['H', 'e', 'l', 'l', 'o', ' ', 'W', 'o', 'r', 'l', 'd', '!']
```now, we'll perform the subtraction with the `__sub__` magic method:```py
def __sub__(self, also_self):  
  print(''.join(self.value))
  ()=self.value[2:4]=[]
```so remember `self.value`? we join it and print it now:```py
print(''.join(self.value))
```this results in the `Hello World!` being printed
then, we reset `self.value` by doing this:```py
()=self.value[2:4]=[]
```basically makes `self.value` the list it was before anything

[floral meteor]

hence, the SOS signal prints helo world

[quartz wave]

that took the whole discord text limit to type

[jovial monolith]

woah lmfao this is very comprehensive

yeah that makes sense

thanks 😂

[knotty delta]

scariest wpm

[floral meteor]

all that action in that cute little script.

[quartz wave]

i also had to delete some spaces for it to just fit

[floral meteor]

lmao

[upbeat sonnet]

imagine ''.join()

[floral meteor]

it was like reading my thought processes whilst I was writing it

[upbeat sonnet]

* when

[floral meteor]

* is the opposite operation

[jovial monolith]

value could be "hello world" and neg could just return self

it would work too right

[upbeat sonnet]

isnt it unpacking and printing?

[fleet bridge]

!e

print(*'hello')

[night quarry]

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

h e l l o

[floral meteor]

it packs and prints

[upbeat sonnet]

oh ic

[floral meteor]

it's already unpacked, it needs to be packed

[quartz wave]

yes

[floral meteor]

it would work, but I want to drive with ...---... instead of ...-...

hence I gotta make it dependant on neg

[quartz wave]

what surprises me is how this took just 4 minutes or less to type

[floral meteor]

keyboard warrior!

[jovial monolith]

!e py from ctypes import py_object p = py_object.from_address(id(...) + 8) class SOS: val = "hello world" def __neg__(self): return self def __sub__(self, other): print(self.val) p.value = SOS ...---...

fk

lmao

[floral meteor]

val != value

[night quarry]

@jovial monolith :white_check_mark: Your eval job has completed with return code 0.

hello world

[jovial monolith]

yeah that works too i guess

[floral meteor]

it doesn't look as good though

[jovial monolith]

as esoteric

[upbeat sonnet]

nothing here looks good

[jovial monolith]

!e
print(....value)

[upbeat sonnet]

!e print((True, True, True) is True, True, True

[night quarry]

@upbeat sonnet :x: Your eval job has completed with return code 1.

001 |   File "<string>", line 1
002 |     print((True, True, True) is True, True, True
003 |          ^
004 | SyntaxError: '(' was never closed

[upbeat sonnet]

!e print((True, True, True) is True, True, True)

[night quarry]

@upbeat sonnet :white_check_mark: Your eval job has completed with return code 0.

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

[jovial monolith]

!e
print(True, True, True == (True, True, True))

[night quarry]

@jovial monolith :white_check_mark: Your eval job has completed with return code 0.

True True False

[jovial monolith]

this was on youtube wasnt it

[upbeat sonnet]

i think so

i still dont grasp it yet

!e
print(True, True, True == *(True, True, True))

[night quarry]

@upbeat sonnet :x: Your eval job has completed with return code 1.

001 |   File "<string>", line 1
002 |     print(True, True, True == *(True, True, True))
003 |                               ^
004 | SyntaxError: invalid syntax

[upbeat sonnet]

ic

doesnt work like that does it

[quartz wave]

when you unpack, it needs somewhere to go to

[floral meteor]

!e ```py
from ctypes import*
p=py_object.from_address(id(...)+8)
class SOS:
value=iter(["World!","Hello "])
def neg(self):
return type('',(),{'neg':lambda s:next(self.value)+s.value,'init':lambda s,v:setattr(s,'value',v)})(next(self.value))
def sub(self, phrase):
print(phrase)
self.value = iter([phrase[6:],phrase[:6]])
p.value = SOS

...---...

[upbeat sonnet]

!e
print(True, True, True == list(*(True, True, True)))

!e
print(True, True, True == set(*(True, True, True)))

[night quarry]

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

Hello World!

[floral meteor]

@upbeat sonnet use the trashcan reaction to remove repeated error results pls

ty

[floral meteor]

with this modification, the driver is more dependant on the neg

[upbeat sonnet]

cant we do like a unicode iterator thing to find the letters in hello world then print them

[floral meteor]

why?

[upbeat sonnet]

cos yes

[floral meteor]

splitting in the middle works.

it's even more cursed because it joins it backwards, so the iterable has to be backwards

[upbeat sonnet]

😭

why do this

[floral meteor]

"I think this is my subconscious sending out a distress signal because it's stuck in my brain"

That would make a great docstring for this script

!e ```py
from ctypes import*
p=py_object.from_address(id(...)+8)
class SOS:
value=iter(["World!","Hello "])
def neg(self):
return type('',(),{'neg':lambda s:next(self.value)+s.value,'init':lambda s,v:setattr(s,'value',v)})(next(self.value))
def sub(self, phrase):
print(phrase)
self.value = iter([phrase[6:],phrase[:6]])
p.value = SOS

...---...
...---...

[night quarry]

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

001 | Hello World!
002 | Hello World!

[floral meteor]

it works multiple times too

[floral meteor]

I'm gonna make 4-dimensional 2048

just planning it is giving me a headache. The source code will be difficult to understand.

[jovial monolith]

how does 4d work in the first place anyway

[floral meteor]

There is 16 4x4 boards laid out in a 4x4 grid. You use wasd keys to swipe local vertical and horizontal, and you use arrow keys to swipe hypervertical and hyperhorizontal.

__format__=lambda s,f='1':s.colour(str(s)[int(f)*7:7+int(f)*7].__format__(f'{[" >"," ^","_<"][int(f)]}7s'))

I'm already going insane supporting larger numbers in a 7x3 square, due to the extreme terminal character cell height to width ratio

the conditional CSI5m wouldn't work, so I removed it

this is just to display a large integer (up to 21 digits) with wrapped center alignment.

any larger numbers just get the end chopped off, which isn't significant figures anyway.

I am plagiarising myself like a bad boii.

I am literally writing it in order from start to finish, also very recommended against.
Write stub code before the algorithm

[jovial monolith]

what the hell... you're writing it in esoteric code too lol

wouldn't it be better to write out the proper code first and then screw it over later?

[floral meteor]

I screw it as I go.
here's the engine init

class Engine:
  def __init__(self):
    self.matrix=[[[[BaseCell(i,j,k,l)for i in range(4)]for j in range(4)]for k in range(4)]for l in range(4)];r=lambda:(lambda l:(l(),l(),l(),l()))(lambda:randint(0,3));a=lambda:Cell(*r(),2);a=a(),a()
    while-len({*a})+2:a[1].move(*r())
    [()for x in a for self[x]in[x]]

4 dimensional matrix go brrrrr

2 dimensional script is less than 80 lines. I'm aiming for less than 44 lines in the 4 dimensional script

[jovial monolith]

what the actual hell is that lmfaoo

good luck tho

very interesting concept

[floral meteor]

the swipe algorithm upper level of code

def swipe(self,d):
    def _swipe(i):
      def __swipe(j):
        def ___swipe(k):
          ... # le algörithem
          return q!=[*map(int,m)]
        return any([___swipe(k)for k in range(4)])
      return any([__swipe(j)for j in range(4)])
    r=any([_swipe(i)for i in range(4)])
    return[f"swiped {['up','down','right','left','hyperup','hyperdown','hyperright','hyperleft'][d]: <10s}","Game Won","invalid move"][max(2*r,self.gamewon)].__format__(' <18s')

the lower level code is next in line to be written

I think this might end up more lines than the 2D script.

oops there's a glitch in the matrix.

return[f"swiped {['up','down','right','left','hyperup','hyperdown','hyperright','hyperleft'][d]: <10s}","Game Won","invalid move"][max(2*(not r),self.gamewon)].__format__(' <18s')

this time I will test the engine before I write the interface because the interface is gonna be chonkers.

the cli would be hard to write or view, having only one dimension to present data in, in order for the __repr__ output to also be able to be python interpreted to convert the matrix to a matrix of raw integers

[floral meteor]

it is now fully stubbed with a few labelled insert-heres

where the important stuff is yet to be done like algorithm, board display and event loop

it's coming along great!

\x1b[100m \x1b[49m{self.eng[0,0,0,0]:0}|{self.eng[1,0,0,0]:0}|{self.eng[2,0,0,0]:0}|{self.eng[3,0,0,0]:0}\x1b[100m \x1b[49m{self.eng[0,0,1,0]:0}|...
ok so maybe I should soft code this board else it's gonna take forever to write.

the board would be 129 characters by 65 lines, but still be a square.

16x16 cells

4x4x4x4, technically

[jovial monolith]

oh my god that is horrendous lmfao

send a vid of it running?

[lime bane]

I have a challenge for you people here:
Find the smallest set of characters that you need to make turing complete python scripts.
So for example with the characters a, = and 5
You can make a valid python script, but it isn't turning complete.
The opposite example would be the character @. It brings convenience features to the language, but without it you can still achieve everything in python.

[fleet bridge]

while + if + _ + 0-9 + -* + [] + <space> + =

!e

# memory initialisation
_=[0]*2**10 # 1k of ints

_[1]=3 # input data
_[2]=4 # input data

# i want to compute ceil((a**2+b**2)**0.5)
_[3]=_[1]**2--_[2]**2
while _[0]**2<_[3]:
    _[0]+=1

# result value is in _[0] cell
# print value
print(_[0])

[night quarry]

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

5

[fleet bridge]

while + if + 1 + -* + [] + <space> + = + < + : + \n

from 1's we can contruct any integer using only + and -

• we can implement as --

we dont need functions and function calls, so no ()

[sudden osprey]

You could use while as an if, right?

like

_[0] = <condition>
while _[0]:
    # do something
    _[0] = 0

[fleet bridge]

!e

# memory initialisation
i=1--1
i*=1--1
i*=1--1
i*=1--1
i*=1--1
i=[1-1]*i*i # 1k of ints

i[1]=1--1--1 # input data
i[1--1]=1--1--1--1 # input data

# we want to compute ceil((a**2+b**2)**0.5)
i[1--1--1]=i[1]*i[1]--i[1--1]*i[1--1]
while i[1-1]*i[1-1]<i[1--1--1]:
    i[1-1]-=-1

# result value is in i[0] cell
# calling side should get result from this cell
# print value
print(i[0])

[night quarry]

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

5

[fleet bridge]

yes

while + 1-*=<[]: + <space> + \n

i think we can implement < in other way

we can dynamically extend memory without limits, so our memory can be larger than any integer

[cloud fossil]

* is probably not needed. You can use repetitive addition

[fleet bridge]

yes

[cloud fossil]

You can get rid of < by simply subtracting two values until one of them reaches zero

This looks more and more like brainfuck

[fleet bridge]

!e

# i will use `+` and integer literals, but keep in mind they will be replaced with `--` and combination of 1's
# memory initialisation
i=[0]
i+=i
i+=i
i+=i
i+=i
i+=i

i[1]=3 # input data
i[2]=4 # input data
i[3] # = i[1]**2
i[4] # = i[2]**2
i[5] # = i[3]+i[4]
i[6] # = counter
i[7] # = while condition
i[8] # = i[0]**2

i[6]=i[1]
while i[6]:
  i[6]-=1
  i[3]+=i[1]

i[6]=i[2]
while i[6]:
  i[6]-=1
  i[4]+=i[2]

i[5]=i[3]+i[4]

i[7]=1
i[0]=0
while i[7]:
    i[0]+=1
    i[8]=0
    i[6]=i[0]
    while i[6]:
        i[6]-=1
        i[8]+=i[0]
    
    i[7]=i[8]<i[5] # sorry for `<`
    
# result value is in i[0] cell
# calling side should get result from this cell
# print memory and value
print(i)
print(i[0])

[night quarry]

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

001 | [5, 3, 4, 9, 16, 25, 0, False, 25, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
002 | 5

[sudden osprey]

i=[0]
i+=i
i+=i
i+=i
i+=i
i+=i

in this case you can't just replace with - though. not sure how you would create multi element arrays without *. +, or an additional character

[fleet bridge]

!e

# i will use `+` and integer literals, but keep in mind they will be replaced with `--` and combination of 1's
# memory initialization
i=[0]
i+=i
i+=i
i+=i
i+=i
i+=i

i[1]=3 # input data
i[2]=4 # input data
i[3] # = i[1]**2
i[4] # = i[2]**2
i[5] # = i[3]+i[4]
i[6] # = counter
i[7] # = while condition
i[8] # = i[0]**2
i[9] # for <
i[10] # for <
i[11] # flag
i[12] # res

i[6]=i[1]
while i[6]:
  i[6]-=1
  i[3]+=i[1]

i[6]=i[2]
while i[6]:
  i[6]-=1
  i[4]+=i[2]

i[5]=i[3]+i[4]

i[7]=1
i[0]=0
while i[7]:
    i[0]+=1
    i[8]=0
    i[6]=i[0]
    while i[6]:
        i[6]-=1
        i[8]+=i[0]
    
    # i[7]=i[8]<i[5] # sorry for `<`
    i[12] # -1 - >, 0 - =, 1 - <
    i[11]=1
    i[9]=i[8]
    i[10]=i[8]
    while i[8]==i[5]:
        i[8]=-1
        i[11]=0
        i[12]=0
        
    while i[11]:
        i[9]-=1
        i[10]+=1        
        while i[9]==i[5]:
            i[9]=-1
            i[11]=0
            i[12]=-1

        while i[10]==i[5]:
            i[10]=-1
            i[11]=0
            i[12]=1

    i[7]=i[12]==1

# result value is in i[0] cell
# calling side should get result from this cell
# print memory and value
print(i)
print(i[0])

[night quarry]

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

001 | [5, 3, 4, 9, 16, 25, 0, False, -1, 25, 25, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
002 | 5

[fleet bridge]

wow

it works

this is comparison

[fleet bridge]

🤔

while + 1-=[]: + <space> + \n + some character to construct large list

i can write i=[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,...] but it is finite memory and it cant be expanded a runtime

this is machine with infinite registers

[cloud fossil]

Instead of that, you would use -- to concatenate

[cloud fossil]

Why do you need infinite?

[fleet bridge]

because if memory is finite, our machine isnt turing complete

[cloud fossil]

!e

print([0] (--) [0])

[night quarry]

@cloud fossil :x: Your eval job has completed with return code 1.

001 |   File "<string>", line 1
002 |     print([0] (--) [0])
003 |                  ^
004 | SyntaxError: invalid syntax

[cloud fossil]

I realized this is ridiculous

You can't have infinite memory, it's theory

[fleet bridge]

i can make memory as large as i want

[sudden osprey]

If you added in the , character to do that you might as well add * or + instead.

[fleet bridge]

oh

i forgot ,

* is better than + or ,

this is multiplication

[cloud fossil]

This is brainfuck translated to Python

[fleet bridge]

pyfuck

[cloud fossil]

Kind of

[fleet bridge]

hmm

i can use infinite number of variables without array

i, ii, iii, iiii, ...

[cloud fossil]

That makes i a character

[fleet bridge]

and i can mix letters from while

[fleet bridge]

there is an i in while

[cloud fossil]

Right

Just add +

[fleet bridge]

+ - is one extra character

[fleet bridge]

this requires no extra characters

[cloud fossil]

What do you plan to do with all those variables

[fleet bridge]

use them

[cloud fossil]

Then yeah

[fleet bridge]

but if i use variables instead of list, i cant make and use pointers to memory(

[cloud fossil]

Yo

I just thought of something

Put new items using slicing

You already have colon as a character

[fleet bridge]

>>> i=[]
>>> i[0:1]=[0]
>>> i
[0]
>>> i[1:2]=i
>>> i
[0, 0]
>>> i[2:]=i
>>> i
[0, 0, 0, 0]
>>> i[4:]=i
>>> i
[0, 0, 0, 0, 0, 0, 0, 0]
>>> i[8:]=i
>>> i
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
``` yeah

[sudden osprey]

nice

[fleet bridge]

we should track memory size to do correct slices, but it is good idea

[cloud fossil]

Creativity 💯

[lime bane]

Now one question: do you really need 0-9? couldn't you just use 1 and construct the numbers like this:
0 : 1-1
1 : 1
2: 1--1
3: 1--1--1

[cloud fossil]

You mean

1--1--1

Yes

I think it was just to quickly script

[lime bane]

Yes. It's more readable with all 10 decimal digits, but we could just work with 1

So we are down to: ```
w
h
i
l
e
1

=
[
]
:

and space and `\n`

[cloud fossil]

That character's name is newline

[lime bane]

I know.