esoteric-python

[wary python]

writing a program in the least amount of characters you can

[austere mauve]

Write as short as possible code that solves a given task

[wary python]

and by least its like

least least

[austere mauve]

Most often counted in bytes

[vernal crest]

thank you

[untold socket]

how to make email automation generator.

[long fulcrum]

anyone got any ideas to golf this more?

import sys
z=sys.argv[1].replace(*" .")
exec('z=[[v," "][v=="."*(sum(y in"# "for y in z[x-52:x:51]+z[x+52:x:-51])>2)]for x,v in enumerate(z)];'*650)
print(*z,sep="")```

[subtle viper]

skipping the first replace and swapping the space and dot characters in the later part of the code?

would also free up the argv[1] to leave to an unpacking assignment

[long fulcrum]

i dont think i can do that without making it longer

if i dont replace at the start i have to use a placeholder so i can then change the to . and then placeholder to afterwards

i think

thats what i did before

[subtle viper]

are there other dots?

[long fulcrum]

at the end i need dots remaining

[subtle viper]

i see

[long fulcrum]

sample input output

#S#      E#       #S#  ....E#
# ### #####       #.###.#####
#   #     #   →   #...#.....#
### ##### #       ###.#####.#
#         #       #  .......#
###########       ###########```

youd need to change the slicing in my code to work on that sample though

after the first replace i check every square and change it back to a space if its a dead end

[subtle viper]

i think that equal-multiply-gt can be turned into a chain comparison

[long fulcrum]

can it?

i cant see any way to get it shorter cause i need the v=="." to check if its a . and then the rest is to count the surrounding tiles

the * used to be an and

6 chars saved

import sys
z=sys.argv[1].replace(*" .")
exec('z=[[v," "][2<sum(y<"$"<v<"/"for y in z[x-52:x:51]+z[x+52:x:-51])]for x,v in enumerate(z)];'*650)
print(*z,sep="")```

[subtle viper]

the v==s*(...>2) thingy can be made into v==s!=...>2 which saves a character

but this one's better

[long fulcrum]

I can't get any shorter with this I think I might need a different algorithm to get shorter

[unreal echo]

Use a lib

Also wrong channel

[terse pollen]

@wicked dagger a spoiler for the actual "102 bytes" part
⁨```py
i=lambda**k:lambda f:type(f)(f.code,type('',(dict,),{'missing':f.globals.getitem})(k))

[random juniper]

!e ⁨py print("".join([(0.5.__class__.__name__[0]), (9).__class__.__name__[0], "".join([dir(str)[-1][0] for x in range(2)])]) + "".join([True.__class__.__name__[0].upper() + ().__class__.__name__[1], "".join([dir(str)[-1][0] for x in range(2)])]))⁩

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

fizzBuzz

[random juniper]

lmao

[ocean fractal]

best code in my life

[terse pollen]

!e

a=lambda o,i:o.__class__.__name__[i];b=[dir('str')[-1][0]for _ in'ab'];print(a(1.,0),a(1,0),*b,a(True,0),a((),1),*b,sep='')

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

fizzbuzz

[onyx pendant]

542431903886606399

[subtle viper]

!e ```py
print('%x'%542431903886606399)

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

7871af15200003f

[subtle viper]

dunno how to interpret that

[earnest wing]

looks like a discord id

[subtle viper]

@fickle marsh

#542431903886606399

i see

<@&542431903886606399>

[versed eagle]

meow

[bitter sparrow]

hello i want to make openclaw work but every tim he sai permission denied

i try many thing nothing work plese help

[earnest wing]

lol

[sick hound]

hello

[muted cipher]

almost everything i write seems to be unreadable to anyone except me so here's a dictionary that gives items a timelimit

[night quarry]

Click here to see this code in our pastebin.

[hybrid trail]

Given that you have access to a mutable class object in python, is the only way to use to call a function without using parentheses setting __class_getitem__?

# given
class a: ...
# is this the only dunder we can trigger?
a.__class_getitem__ = exec
a['...']

context is restricted source challenges

[earnest wing]

You can call some functions parenless via decorators on a class statement

[hybrid trail]

restricted source challenge: using only printable characters (string.printable) but none of ('(', '.', 'c') in your source code, is it possible to print "Hello, World!" to stdout and nothing else?

[wary python]

!e print(import("string").printable)

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

001 | 0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ!"#$%&'()*+,-./:;<=>?@[\]^_`{|}~ 	
002 | 
003 | 
004 | 

[wary python]

huh does that not work anymore

ah

[austere mauve]

The

import __hello__

Works up to and including python 3.10

[hybrid trail]

Close, punctuation a bit off though (:

[versed eagle]

including the quotes?

[hybrid trail]

if you really want to, my solution works either way

[versed eagle]

curious

[mystic carbon]

They removed __hello__ in 3.11?

[austere mauve]

No, but it no longer prints the Hello, World! on import

You need to call __hello__.main()

[mystic carbon]

ohhhhhhh, that makes more sense

[austere mauve]

Noo, you are right 🙁

[subtle viper]

!e ||```py
getattr=path=print
all=["Hello, World!"]
from main import*

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

001 | Hello, World!
002 | Hello, World!

[hybrid trail]

oh wow, that is great

can it be fixed to print once?

[subtle viper]

no idea

[versed eagle]

!e ```py
getattr=path=exit
all=["Hello, World!"]
from main import*

[night quarry]

:x: Your 3.14 eval job has completed with return code 1.

Hello, World!

[versed eagle]

tada

actually idk if exit prints to stdout or stderr

[unreal echo]

!e ```py
getattr=path=print
all=["Hello, World!"]
from main import*

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

001 | Hello, World!
002 | Hello, World!

[cold lance]

import this
>>> this
<module 'this' from '/usr/lib/python3.12/this.py'>
>>> love = this
>>> love
<module 'this' from '/usr/lib/python3.12/this.py'>
>>> this is love
True
>>> love is True
False
>>> love is False
False
>>> love is not True or False
True
>>> love is not True or False; love is love
True
True

[terse pollen]

why is __path__ involved?

[restive void]

!e

__import__("fishhook").hook(bool)
def __str__(self): return "em truh t'nod ybab"[::-1]
love = 1 <3
what = all([1])

print(what is love)

[night quarry]

:white_check_mark: Your 3.13 eval job has completed with return code 0.

True

[restive void]

:(

works on my machine

[hybrid trail]

i think you're missing a @ for the decorator

[restive void]

!e Right.

@__import__("fishhook").hook(bool)
def __str__(self): return "em truh t'nod ybab"[::-1]
love = 1 <3
what = all([1])

print(what is love)

[night quarry]

:white_check_mark: Your 3.13 eval job has completed with return code 0.

baby don't hurt me

[subtle viper]

!e ```py
getattr=print
all=["Hello, World!"]
from main import*

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

001 | __path__
002 | Hello, World!
003 | Hello, World!

[subtle viper]

it's kind of needed for the star import, i guess?

[terse pollen]

oh, it wouldn't normally have a __path__ because it wasn't imported to begin with; the import logic looks for it; and the __getattr__ hook triggers

[keen fog]

yeah exactly, import machinery is wild when you start messing with hooks

[terse pollen]

isn't fishhook third-party?

[fallen heath]

yes

[restive void]

by @rugged sparrow

[restive void]

But there's a few libraries we install in the environment the bot is running code in.

[terse pollen]

ooh

[somber belfry]

!e e

[night quarry]

:x: Your 3.14 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "/home/main.py", line 1, in <module>
003 |     e
004 | NameError: name 'e' is not defined

[somber belfry]

!e ls

[night quarry]

:x: Your 3.14 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "/home/main.py", line 1, in <module>
003 |     ls
004 | NameError: name 'ls' is not defined

[restive void]

Please use #bot-commands for experimenting with the bot

[hybrid trail]

!e I guess this should work?

__path__ = None
__getattr__=print
__all__=["Hello, World!",0]
from __main__ import*

[night quarry]

:x: Your 3.14 eval job has completed with return code 1.

001 | Hello, World!
002 | Traceback (most recent call last):
003 |   File "/home/main.py", line 4, in <module>
004 |     from __main__ import*
005 |   File "<frozen importlib._bootstrap>", line 1421, in _handle_fromlist
006 |   File "<frozen importlib._bootstrap>", line 1417, in _handle_fromlist
007 | TypeError: Item in __main__.__all__ must be str, not int

[subtle viper]

i mean sure

if stderr doesn't matter

[somber obsidian]

This is genius

[somber elk]

!e print(chr(sum(range(ord(min(str(not())))))))

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

ඞ

[somber elk]

lets goo i still remember it

[twilit igloo]

HUHHUHHUHHUHHUH

!e print(chr(sum(range(ord(min(str(not())))))))

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

ඞ

[twilit igloo]

HOW DOES THIS WORK

ohhh

i kinda got it

this is like uhh

pythonfuck

[golden steppe]

Step by step its

True->'True'->'T'->84->3486->ඞ

[twilit igloo]

ahhhh got it

i also figured it out

the not() kinda freaked me out

but it was actually not ()

and since () is falsey

not () returns True, the bool

[long fulcrum]

is it possible to avoid assigning n?
"string"[n:='><+-.['.find(i)];a+=n//5

[fallen heath]

by duplicating the code

[swift turtle]

@long fulcrum Could you also do this instead?

n=((n-1)//2)*(n%2)or ((n+1)*3)*(n%2==0)
n=[(n+1)*3,(n-1)//2][n%2]

nvm codegolf made a better one

[gleaming linden]

:p

[long fulcrum]

this is my full thing

for n in range(1000):
    s=0
    while n:n=[-~n*3,~-n//2][n%2];s+=1
    print(s)```

oh I can do n>>1

[spiral forge]

C++ is better yell

[wary python]

this image feels like its missing something without the other panels

[karmic pumice]

the above panel is the message above the image and the bottom panel is the even more above messages

[wary python]

yay

[unreal echo]

.bm

[subtle viper]

i think they were trying to say C++ is better yellow

[long fulcrum]

is there a shorter version of

i=0
_,g=sys.argv```

or the whole thing actually

import sys
i=0
_,g=sys.argv
while i<1055:m=i%33==31;print(".#"[2<sum(g[i-67+x:][:3].count("#")for x in b"!Bc")<4+(g[i]<"$")],end="\n"*m);i+=1+m```

[wary python]

:3

[long fulcrum]

:3

[long fulcrum]

oh I can do i%33>30

I forgot I skip iterations where it's 32

[dry gulch]

does code golfed Tetris have a place here :o

it's pretty esoteric

[subtle viper]

code golfed anything belongs here

[dry gulch]

heck yeah

I made Tetris in 25 lines, I've seen people do one liners but I don't know how to do that

https://pastebin.com/i0cQNb2n

[sick bridge]

Is code golf even a concept in any other language other than python

[stable hornet]

yes, people have even made golfing languages

it is certainly not a Python-specific thing

[sick bridge]

Nice

[coarse void]

This python code converts roman numerals into decimals

roman=lambda _,__=lambda ___:[
[                           1,
5,                         10,
50,                       100,
500,                      1000
][                   "IVXLCDM"
.index(                      _
)                            ]
for                          (
_                            )
in                           (
___                          )
]:sum(                       [
[                          _i,
-_i                         ][
_i                           <
__i                          ]
for                          (
_i,                        __i
)                           in
(                         zip(
[                            0
]                            +
__(                          _
),                         __(
_                            )
+                            [
0                            ]
)                            )
]                            )
print(roman("MMMCCCLXXXVIII"))

[earnest wing]

I think you can expand it even further

for example "I" "V", sum (, etc

[restive void]

Should make it in the shape of a roman numeral (maybe "V"?)

[dry gulch]

awesome

[shadow cobalt]

g

[coarse void]

!e

p = lambda x,c=lambda n:[1,5,10,50,100,500,1000]["IVXLCDM".index(n)]:sum([c(i)*(-1 if c((x+"I")[y+1])>c(i)else 1)for y,i in enumerate(x)])
print(p("MCDLXXIV"))

Can anyone come up with any shorter code than this?

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

1474

[coarse void]

!e

p=lambda x,c=lambda n:[1,5,10,50,100,500,1000]["IVXLCDM".index(n)]:sum([c(i)*(1-2*(c((x+"I")[y+1])>c(i)))for(y,i)in enumerate(x)])
print(p("MCDLXXIV"))

130 characters

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

1474

[subtle viper]

!e

p=lambda x,c=lambda n:[1,5,10,50,100,500,1000]["IVXLCD".find(n)]:sum(c(i)*(c(f"{x}I"[y+1])<=c(i)or-1)for y,i in enumerate(x))
print(p("MCDLXXIV"))

forgot to remove end

ouch

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

1474

[long fulcrum]

!e

def f(i):
    t=0
    for r in i:n=10**(205558%ord(r)%7)%1999;t+=n-2*t%n
    return t
print(f("MCDLXXIV"))```

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

1474

[coarse void]

explain pls

[long fulcrum]

idk it's just some code I have saved, I imagine it's from stack overflow

[coarse void]

could we make it inline lambda?

[subtle viper]

not really

it's a for loop that updates an outside variable

[earnest wing]

the magic section looks to be found via pysearch?

(https://github.com/lynn/pysearch)

[long fulcrum]

yeah I couldn't find it on stack overflow so probably

is anyone able to figure how to get this code running like the first part

# import sys;e=enumerate
# for m in sys.argv[1:]:
#     m=m.split()
#     for y,j in e(m):
#         for x,i in e(j):
#             s=[]
#             for g in[(y,x)for y,j in e(m)for x,i in e(j)if i=="0"]:
#                 q=[(y,x)];d={(y,x):0}
#                 while q:
#                     v,*q=q;b,a=v
#                     if v==g:s+=d[v],
#                     if m[y][x]=="#":break
#                     for B,A in(b-1,a),(b,a-1),(b,a+1),(b+1,a):
#                         if -1<B<9>A>-1and m[B][A]!="#"and(B,A)not in d:d[(B,A)]=d[v]+1;q+=(B,A),
#             print(end=([chr(x)for x in range(150)if chr(x).isalnum()][min(s)]if s else m[y][x]))
#         print()
#     print()
import sys
for m in sys.argv[1:]:
    for x,i in enumerate(m):
        s=[]
        for g in[y for y,j in enumerate(m) if j=="0"]:
            q=[x];d={x:0}
            while q:
                v,*q=q
                if v==g:s+=d[v],
                if m[x]=="#"or m[x]=="\n":break
                n=[]
                if v>9:n+=v-10,
                if v<72:n+=v+10,
                if v%10>0:n+=v-1,
                if v%10<8:n+=v+1,
                for w in n:
                    if (m[w]=="-"or m[w]=="0")and w not in d:
                        d[w]=d[v]+1;q+=(w),
        print(end=([chr(x)for x in range(150)if chr(x).isalnum()][min(s)]if s else m[x]))
    print("\n")```

first code uses a 2d list but im trying to get it working with just the string that gets passed

input output

-#-------    -#6543212
-#-----0-    -#5432101
-#-------    -#6543212
-#-------    -#7654323
-#-------    -#8765434
-########    -########
-#-------    -#-------
-#-------    -#-------```

the new code fails on this input

---------
---------
---------
---###---
---#-#---
---###---
---------
---------
--------0```

[slate plover]

[long fulcrum]

that 72 should have been a 79

[austere mauve]

Why not? There's locals(), vars(), globals() and sys._getframe()...

I mean, it would be worse that way, but still...

[austere mauve]

!e ```py
f=lambda i:(t:=0)or[t:=t+n-2*t%n for n in[10**(205558%ord(r)%7)%1999 for r in i]]and t
print(f("MCDLXXIV"))

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

1474

[austere mauve]

It is possible

But that doesn't mean it's shorter

[austere mauve]

!e ```py
f=lambda i,t=0:[(n:=10**(205558%ord(r)%7)%1999)+(t:=t+n-2*t%n)for r in i]and t
print(f("MCDLXXIV"))

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

1474

[austere mauve]

Golfed it a bit

Two chars longer than def version.

[subtle viper]

that's what i was thinking

kind of communicated it wrongly though :p

!e ```py
f=lambda i,t=0:[t:=t+(n:=10**(205558%ord(r)%7)%1999)-2*t%n for r in i]and t
print(f("MCDLXXIV"))

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

1474

[subtle viper]

i don't know if that's shorter than the def version with one-space indents

[austere mauve]

I was counting with one space indents

[subtle viper]

oh okay

this must be shorter then

feels like magic

how does it even work

[austere mauve]

!e ```py
f=lambda i,t=0:[t:=t+(n:=10**(205558%ord(r)%7)%1999)-2*t%n for r in i][-1]
print(f("MCDLXXIV"))

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

1474

[austere mauve]

Another char down

It's finally shorter than original!

!e ```py
f=lambda i,t=0:max(t:=t+(n:=10**(205558%ord(r)%7)%1999)-2*t%n for r in i)
print(f("MCDLXXIV"))

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

1474

[austere mauve]

I was unsure whether I could do this, but it works out, so -1 byte

I think that shorter solutions would need to switch to a different algorithm altogether (if that's possible).

I really don't see where it could be improved

[woeful oasis]

this python program recognizes it's own sha256 hash and describes itself without any file reads. <redacted due to windows line endings making it finnicky>

[restive void]

Doesn't work for me, is there anything about the line endings I have to take care of?

[woeful oasis]

Are you getting 5b80b6f58d7062d48307e1dea85f35c31e94baca19642fa9c12d4727b3cf609e as the hash of the source code?

[restive void]

No, b059e490d17cb61e4794aefad78a652b2b973e0b9dd6fba1cc85094a3e82504f

Hexdump: https://paste.pythondiscord.com/YYGQ

[woeful oasis]

I believe it's the encoding

Hold on its still being weird....

[noble belfry]

i tried a few combinations, none seem to work:

% sha256sum *
c981fa78ac1134b4caa8af7d6a4e7c2defe2190c57b20df36fc6f156bfc48651  GO4Q.txt 
8513992ed9910ddc8cecd457274268958a4f635c71c6ec8f0bba8b617ca73c68  crlf_notrailing_nl_utf16.py
52695766fb61bed089b113ff13ea221938434d1248a665b2532e2ab662488895  crlf_notrailing_nl_utf8.py
2f626dfd858786a8269b86597631f3d42ffe618f523216b2b5cf0ecc0d9cb772  crlf_trailing_nl_utf16.py
d7873b97d6076571d3e8f2666d92023d5e0c9ae7ce852efbd91276a063430884  crlf_trailing_nl_utf8.py
4e830932e951220cab02f128f603b6963dd4b5bad099d5ff9427c07c0a632178  lf_notrailing_nl_utf16.py
92969b687732b9bf43655b1fae55d8d9107b67b46f933990c6183a42e2d9fe6e  lf_notrailing_nl_utf8.py
cf232524fb77f4e2ba8e15751826863c9a32d56041541bc269ec65a967333313  lf_trailing_nl_utf16.py
1819a93a06815a4b1be23c653b405f2a5d6d16f5a3e746e435b9e2c3da7277b0  lf_trailing_nl_utf8.py

[restive void]

Do you happen to be on Windows?

[woeful oasis]

Yes

[restive void]

Oh. And not on WSL I'm guessing.

[woeful oasis]

Oh, yeah thats the problem, when I go into wsl I'm getting the c981 hash you guys are getting

[restive void]

I converted it to DOS line endings, but now I'm getting f275268ba73ba946ddd0246bf796f3b05135cee467dd592a29b3734a5564045f...

Okay, and if I remove the trailing line ending but keep DOS line endings I'm getting c981...

Where did you get this Python file from?

[woeful oasis]

I made it

[restive void]

Well, it works for me (with a normal file on a non-weird OS) when I replace the string in the file with e5a4cda988c87e09681ce41b9ff30d82b7b8830138e1e23eeb1a969377f23e4f

[woeful oasis]

How did you even cook that up

[restive void]

Add a print(x(sys.argv[1])) somewhere and call it with its hash?

How did you?

[woeful oasis]

sed -i 's/\r//' spooky.py should fix the line endings* if you've got the c981 version!

[restive void]

That's just dos2unix

[woeful oasis]

It's giving me the correct 5b80b version

[noble belfry]

i thought it's supposed to be fe45fb1678176053679ede6188ed19e301214aeb22783f0acd6f480689910845

[woeful oasis]

No, 5b80b is what it should be

the fe45 is a magic constant

[noble belfry]

oh, i just realized my editor auto formated the file on save. without it, i get 5b80b

$ sha256sum lf_notrailing_nl_utf8.py
5b80b6f58d7062d48307e1dea85f35c31e94baca19642fa9c12d4727b3cf609e  lf_notrailing_nl_utf8.py

[woeful oasis]

the fe45 constant is important because if you set it to anything else, the "english checksum" breaks, plus the hash of the file changes

well really they are all important and triply constrain each other

[terse pollen]

... is sha256 really so weak that putting together something like this is feasible? spooky indeed

[woeful oasis]

thanks asm and l3viathan for helping me realize what was causing problems and why it was careless to finalize this with windows line endings. not very good for copypasting from discord! i went back and redid everything to (hopefully) fix that issue. LF only line endings. was not easy to make everything line up again lol

this python program detects it's own sha256 hash while simultaneously describing it's own character composition. everything depends on everything else. change one character, break everything!

import hashlib

def x(hex_string):
    r = 1
    for c in hex_string:
        r += int(hashlib.sha256(c.encode()).hexdigest(), 16)
    return hashlib.sha256(str(r).encode()).hexdigest()

if __name__ == "__main__":
    if len(sys.argv) != 2:
        print("usage: python3 spooky.py <sha256 hex digest>")
        print("the program detects if you give it its own sha256 hash.")
        sys.exit(1)
    
    input = sys.argv[1]
    if x(input) != "d694dd1b4a8c7b658240b201d313f933968f9b750dbb69d5bfbbe09018107b7c":
        print("That's not the hash of the source code. Verify you got everything: the full source file has exactly two j's, twenty six f's, fourteen b's, sixteen d's, fifty four i's, thirty two o's, fifteen c's, ninety one e's, thirteen p's, seventeen g's, twelve l's, two k's, six m's, thirty seven h's, twenty three a's, fifty one n's, one q, thirty four r's, eighty six s's, seventy t's, sixteen u's, sixteen v's, nine w's, seventeen x's, twenty four y's, one z, seven 0's, eleven 1's, eight 2's, eight 3's, four 4's, eight 5's, ten 6's, six 7's, five 8's and seven 9's. Not case sensitive.")
    else:
        print("Success!")```

The hash of the file should be 234d5dce0e9c5f785481a6711a65b0db962f794ef462cbfa82b00d8313ac73e9

[austere mauve]

Mind sharing how did you do it?

[woeful oasis]

it is meant to be confounding 😉 its like a ship in a bottle

[austere mauve]

I mean the x function is definitely lenient and any order of the letters in the hash would work as the input and this could be brute-forced, but the self-describing message is self dependent and I'm not sure how to calculate that.

[woeful oasis]

you're definitely on the right track, that is one of the tricks

[austere mauve]

You could improve brute force time by changing the order of the letters listed instead of changing the compared hash. There's a clear distribution of the number of letters and they don't seem to vary a lot between the two of the files.

I can somewhat imagine how it could be done, but nonetheless the self-describing part looks hard.

I'd try creating it by:

1. Generate hash of a file without the comparison hash and file description
2. Generate the comparison (of old version) and description (of old version).
   Repeat a few times (maybe once is enough) to get rough expected letter count.
   Lock in the comparison hash (after all it's just equivalent of the file description, but disregarding the order of the letters).

Then just permutate the order of how the letters are listed and check if the resulted hash of the file has the required count of each letter (based on comparison hash) until it's a match.

Considering that from n up, the letters + numbers are ordered and before theme there are 15 "randomly" arranged letters, I can estimate that it took ~10^12 tries. Definitely doable within one day time-frame using some fast implementation in paralel.

How close am I?

[woeful oasis]

The comparison hash is indeed locked in before the permutation search begins - even before the self-describing bit is built - and yes it takes about 10^12 tries! But the way I derived the "comparison hash" is different, and it was derived for a specific reason to make the bruteforce search as fast as possible. You're extremely close.

[austere mauve]

How long did it take to calculate?

[woeful oasis]

The self-describing bit (the autogram) takes <60 seconds, the comparison hash itself I can do instantly because it encodes a specific property of the program hash, and the final permutation search is the long one, generally a few hours

[gleaming linden]

six 7's

[long fulcrum]

any ideas to make this shorter?

import sys
e=enumerate
for m in sys.argv[1:]:
 o=[*m]
 for k,c in e(A:=[*filter(str.isalnum,map(chr,range(150)))]):[[-1<i+w<89and o[i+w]=="-"and o.__setitem__(i+w,A[k+1])for w in[-10,10,-1,1]]for i,x in e(o)if x==c]
 print(*o,"\n",sep="")```

the filter is to convert the distance to base62

[quartz wave]

!e ```py
from string import*;print(printable[:62])

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ

[quartz wave]

well

[quartz wave]

is o only 89 characters in length?

[long fulcrum]

yeah

[long fulcrum]

this

[quartz wave]

so first i think you can import string and do string.printable[:62].swapcase() instead

[long fulcrum]

I also changed the o[i+w]=="-" around to save a space

import sys,string
e=enumerate
for m in sys.argv[1:]:
 o=[*m]
 for k,c in e(A:=string.printable[:62].swapcase()):[[-1<i+w<89and"-"==o[i+w]and o.__setitem__(i+w,A[k+1])for w in[-10,10,-1,1]]for i,x in e(o)if x==c]
 print(*o,"\n",sep="")```

[quartz wave]

also instead of py for m in sys.argv[1:]: o=[*m] you can do ```py
for*o,in sys.argv[1:]:

[long fulcrum]

oh what I always thought you couldn't do anything to the variable in a for loop

[quartz wave]

for loop targets are like any assignment targets

[subtle viper]

you can collapse this into a single line now btw

[long fulcrum]

oh yeah

[subtle viper]

just put the for k,c in e(A) in the listcomp and assign to A before it

and also instead of print(*o,"\n",sep="") use "".join

there's another optimization that can be done with [-10,10,-1,1] but i might be away for a while

basically you can shorten it by doing enumerate(o,<some constant>) and making that constant list into a bytestring

[long fulcrum]

can this be shorter than what I have?

I tried this with bytestring but it ended up the same length

import sys,string
e=enumerate
A=string.printable[:62].swapcase()
for*o,in sys.argv[1:]:[[[-1<(j:=i+w-50)<89and"-"==o[j]and o.__setitem__(j,A[k+1])for w in b'(<13']for i,x in e(o)if x==c]for k,c in e(A)];print("".join(o)+"\n")```

[subtle viper]

oh wait you needed the extra newline

okay

print("".join(o)+"\n")
print(*o,"\n",sep="")

[long fulcrum]

oh yeah this is shorter I just forgot to change the print back

[subtle viper]

put the -50 as the second argument in e(o)

also -1<...==o[j] can be simplified

[subtle viper]

okay wait

[long fulcrum]

it ends up the same length

[subtle viper]

i see

[long fulcrum]

I tried 89>i+w>=0<("-"==o[i+w]) which is the same length

[subtle viper]

btw you can nest for clauses in a listcomp

[...for k,c in e(A)for i,x in e(o)...]

[long fulcrum]

221

import sys,string
e=enumerate
A=string.printable[:62].swapcase()
for*o,in sys.argv[1:]:[-1<i+w<89and"-"==o[i+w]and o.__setitem__(i+w,A[k+1])for k,c in e(A)for i,x in e(o,-50)if x==c for w in b'(<13'];print(*o,"\n",sep="")```

I can change the last and to ==

218

import sys,string
e=enumerate
A=string.printable[:62].swapcase()
for*o,in sys.argv[1:]:[x!=c or-1<i+w<89and"-"==o[i+w]==o.__setitem__(i+w,A[k+1])for k,c in e(A)for i,x in e(o,-50)for w in b'(<13'];print(*o,"\n",sep="")```

[subtle viper]

untested

import sys,string
e=enumerate
A=sorted(string.printable[:62])
for*o,in sys.argv[1:]:
 for j in[i+w-50for k,c in e(A)for i,x in e(o)for w in b'(<13'if x==c]:
  if-1<j<89!="-"==o[j]:o[j]=A[k+1]
 print(*o,"\n",sep="")

[long fulcrum]

k is undefined

[subtle viper]

oh right

212 untested ```py
import sys,string
e=enumerate
A=sorted(string.printable[:62])
for*o,in sys.argv[1:]:[o.setitem(j,A[k+1])for k,c in e(A)for i,x in e(o)for w in b'(<13'if-1<(j:=i+w-50)<89!=x==c>"-"==o[j]];print(*o,"\n",sep="")

[long fulcrum]

yep it works

[subtle viper]

191 untested ```py
import sys
A="123456789ABCDEFG"
for*o,in sys.argv[1:]:[o.setitem(j,d)for c,d in zip("0"+A,A)for i in range(89)for w in b'(<13'if-1<(j:=i+w-50)<89!=o[i]==c>"-"==o[j]];print(*o,"\n",sep="")

[cunning cedar]

I love coming into this channel time to time just to see the... whatever this is lmao

So funky

[subtle viper]

basically turns the input on the left into the output on the right

-#-------    -#6543212
-#-----0-    -#5432101
-#-------    -#6543212
-#-------    -#7654323
-#-------    -#8765434
-########    -########
-#-------    -#-------
-#-------    -#-------```

the basic idea is an ascending character value search that uses a 1D grid interpretation with hardcoded indexes to replace the orthogonal neighbors with the next character value if they're empty (-)

[long fulcrum]

doesnt work, the zeroes start spreading

A needs to run up to q

oh wait i did it wrong

it messes up the newlines and doesn't do any of the distances

I think I'm still doing it wrong actually

it works but it's one longer using string.printable for A and joining

this ends up the same length

import sys,string
A="".join(sorted(string.printable[:62]))
for*o,in sys.argv[1:]:[o.__setitem__(j,d)for c,d in zip(A,A[1:])for i in range(89)for w in b'(<13'if-1<(j:=i+w-50)<89!=o[i]==c>"-"==o[j]];print(*o,"\n",sep="")```

it's longer actually i was comparing to an older one

[subtle viper]

might not need the "".join for A=

[long fulcrum]

oh yeah

[long fulcrum]

was finally able to get this shorter using this

w=[1]
exec("print(*w);*w,=map(sum,zip(w+[0],[0]+w));"*20)```

[austere mauve]

Python: loop unrolling

[sick hound]

.l import sys
Z=sys.argv[1].replace(*" .")
Exec('z=[[v," "][2<sum(y<"$"<v<"/"for y in z[x-52❌51]+z[x+52❌-51])]for x,v in enumerate(z)];'*650)
Print(*z,sep="")

.l ```import sys
z=sys.argv[1].replace(*" .")
exec('z=[[v," "][2<sum(y<"$"<v<"/"for y in z[x-52:x:51]+z[x+52:x:-51])]for x,v in enumerate(z)];'*650)
print(*z,sep="")

!e import sys z=sys.argv[1].replace(*" .") exec('z=[[v," "][2<sum(y<"$"<v<"/"for y in z[x-52:x:51]+z[x+52:x:-51])]for x,v in enumerate(z)];'*650) print(*z,sep="")

[night quarry]

:x: Your 3.14 free threaded eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "/home/main.py", line 2, in <module>
003 |     z=sys.argv[1].replace(*" .")
004 |       ~~~~~~~~^^^
005 | IndexError: list index out of range

[long fulcrum]

https://pastebin.com/9QGiYZsa

[cursive sequoia]

Hi I'm new to esoteric python

!e

f=lambda x:(-1046017.47468*x**10)+(2623063.85928*x**9)+(-439870.30489*x**8)+(-3989739.38765*x**7)+(3407989.77792*x**6)+(721271.24084*x**5)+(-1991793.21887*x**4)+(722163.21951*x**3)+(58786.17727*x**2)+(-76753.93198*x)+(11021.04325)
d=lambda i,c:chr(round(f((i*2+ord(c)-86)/44))-i*2)
print("".join([d(i,c) for i, c in enumerate("Hello World")]))

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

I hate fish

[versed eagle]

finding polynomials that pass through given points is cute

[long fulcrum]

finally managed it shorter

import sys
z=sys.argv[1].replace(*" .")
exec('z=[[v," "][2<sum(y<v=="."for y in z[x-52:x:51]+z[x+52:x:-51])]for x,v in enumerate(z)];'*650)
print(*z,sep="")```

[cursive sequoia]

What does this do?

[somber glen]

fih

[austere mauve]

I'm late, but I just realised that you could partially pre-compute hash for the first part of the file as it doesn't change.

[long fulcrum]

solves a maze

[long fulcrum]

you can run it with an input like this

[long fulcrum]

the way it works is it fills the maze in with . which is used to mark the path to the finish. and then it just loops through the maze and if the . is a dead end it then it's changed to a . then I just do that 650 times until all the dead end paths are gone

[cursive sequoia]

wait did you put everything in a string and multiply it by 650 just to avoid a for loop??

[subtle viper]

well python doesn't have a loop 650: statement and it's shorter than both kinds of existing loops

such is the nature of codegolf

[cursive sequoia]

dang

where can I learn common codegolf techniques?

[versed eagle]

read golfed code

golf code yourself

i dont really care for stack exchange personally but codegolf.stackexchange.com is a decent repository of golfed code in a bunch of languages if you want to look at it

[terse pollen]

i dont really care for stack exchange personally
may I introduce you to https://codegolf.codidact.com/

[magic wraith]

for(i)in' '*650:f()
exec('f();'*650)

[subtle viper]

for()in[()]*650:f() is also unfortunately the same length as the former

[cursive sequoia]

ty bro

[heavy furnace]

hello

[austere mauve]

hi

[magic wraith]

TIL that () = [] is a valid statement

[dusk plover]

I was just thinking about maybe doing some python code golf but i just saw the posts here and i think im good...

[subtle viper]

we all started from a place of comfort, didn't stop some of us :D

[dusk plover]

LOL

Did u guys start this because u already wrote noodle-code or because u wrote clean code that even my Grandma could understand? or something inbetween?

[earnest wing]

esoteric python is something you do to reeeeally have fun with the python language once you understand it. kinda like how artists need to intimately know the rules to break them

[subtle viper]

i was interested in the python language's features and codegolf allowed me to discover most of it

plus other forms of esoteric python

it's basically knowing the language better than it knows itself

[cursive sequoia]

!e

() = []

[night quarry]

:warning: Your 3.14 eval job has completed with return code 0.

[No output]

[cursive sequoia]

me too

!e

def foo(_:(x:="bar")=0):
    print(x)

foo()

[night quarry]

:white_check_mark: Your 3.13 eval job has completed with return code 0.

bar

[cursive sequoia]

does anyone know if it is ever useful to put expressions inside of type hints?

[subtle viper]

well all typehints are valid expressions, for one

but for nonstandard use of typehints there's probably not much accepted use

besides putting strings in, though that was largely outphased by 3.14's lazily evaluated annotations

for unconventional uses it could be used to communicate another outside value while keeping the value hidden within the function object

similar to default values

[earnest wing]

use them as an alternative to globals 👍

[magic wraith]

day 77 of putting code in the weirdest places ever
behold, typehints

[tranquil yarrow]

it can be actually useful for Annotated types

Annotated[TypeForm, *Metadatas]

on the right side you just give something potentially useful at runtime. pydantic uses it to smuggle runtime objects representing validators and other things

by TypeForm i mean type form as defined in the i-love-that-it's-final pep https://peps.python.org/pep-0747/

oh man it's so great typeforms are there

i wish we had it earlier

[gray radish]

I'm trying to tightly integrate 3.14 GIL-less Python with Godot Engine. To eliminate the need for marshaling and make (Godot) Object zero copy passable via GE-Python boundary. The problem is that Object is virtual so it will have vtable at the start and PyObject_Head only after that (if Object will inherit from PyObject). Static casting down to PyObject* should give Python correct pointer (PyObjectType must mind that offset). The issue is with allocation - GIL-less requires to allocate Python Object using mimalloc. Correct allocation size seems trivial but that offset to PyObject_HEAD is problematic; C++ side frees Object by calling memfree on them and this is hijacked to instead participate in Refcounting.

[dusk plover]

codegolf: try defining a function(lambda doesn't count) that takes in a list of arbitrary integer values randomly shuffled and outputs the largest one in binary

output type must be int or string

[fallen heath]

def x(i):print(f"{max(i):b}")

[dusk plover]

!e def x(i):print(f"{max(i):b}")

._. Well i thought youd get it quit quickly but wow... i didnt event know:b was a thing

[earnest wing]

"lambda doesn't count" python code golf generally allows lambdas as full solutions

they are the exact same as functions after all, just different syntax and no name

[karmic pumice]

if the 0b prefix is acceptable then bin is 2 chars shorter

f"{x:b}"
bin(x)##

[magic cosmos]

I feel sorry for myself to open this chat

[reef mesa]

!e

[night quarry]

Click here to see this code in our pastebin.

[cursive sequoia]

tictactoe

b,t=[0,0],1
while not any((b[p]&m==m)and not print("XO"[p]+"wins!")for m in(7,56,448,73,146,292,84,273)for p in(0,1))and sum(b)<511:
    if (u:=int(input("> "))) not in range(9)or 1<<u&sum(b):break
    t^=1;b[t]|=1<<u
    for i in range(9): print("X "*(b[0]>>i&1)+"O "*(b[1]>>i&1)or ". ",end="\n"*(i%3==2))

expanded code if anyone wants to actually read it https://paste.pythondiscord.com/V7HA

[cursive sequoia]

I decided to put all of it inside of a type hint

_:(b:=[0,0],t:=1,[exit() if any((b[p]&m==m)and not print("XO"[p]+" wins!")for m in(7,56,448,73,146,292,84,273)for p in(0,1))or sum(b)==511 else (exit() if (u:=int(input("> "))) not in range(9)or 1<<u&sum(b) else ((b:=[b[0],b[1]|1<<u]) if t else (b:=[b[0]|1<<u,b[1]]),t:=t^1,[print("X "*(b[0]>>i&1)+"O "*(b[1]>>i&1)or ". ",end="\n"*(i%3==2)) for i in range(9)])) for _ in[0]*9])=0

[fallen heath]

SyntaxError: named expression cannot be used within an annotation

You can't use := insidie an annotation

Looks like this is only an error in 3.14 and 3.15

[noble jasper]

check out my obfuscator https://github.com/lyric0x10/Oblivion

give me some feed back : )

view some obfuscated code https://paste.pythondiscord.com/VRFA

[echo breach]

well it doesn't seem like a vm

which is a good thing

[noble jasper]

vm is obfuscated. read the github readme

[echo breach]

yea i meant the code u postedthat was obfuscated

prolly i'm just braindead

[karmic pumice]

pretty cool
would be interesting to see the deobfuscation of the example automated

[noble jasper]

yea, but I dont think anybody in this server can. I'd like to see someone attempt it.

[subtle viper]

<outdated>

[subtle viper]

interesting project if i manage to keep this motivation until i can code again

[crystal rain]

yall wanna calculate pi with 0 error in finite time

from __hello__ import TestFrozenUtf8_2 as __cod__;22/7.; print("\
" ""+
__cod__.__doc__
)

[restive void]

!e

from __hello__ import TestFrozenUtf8_2 as __cod__;22/7.; print("\
" ""+
__cod__.__doc__
)

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

π

[undone thicket]

.topic

[glass drum]

**What are the pros and cons of messing with __magic__()?**

Suggest more topics here!

[undone thicket]

what's magic

!d magic

[night quarry]

magic

%magic```
Print information about the magic function system.

Supported formats: -latex, -brief, -rest

Similar names: label.magic

[undone thicket]

doesn't seem to exist

[restive void]

It's badly formatted, it's supposed to be __magic__, which still doesn't literally exist.
What's meant is "magic methods" aka dunders aka special methods.

[undone thicket]

oh yes dunders

[quartz wave]

is there an option to keep names that need to be exported while obfuscating the code? or is that handled already

[quartz wave]

python_minifier also doesn't seem like an internal module

[noble jasper]

Your right it isnt an internal module. the github specifically says "No external pip packages required for core obfuscation". minifying your code isn't an obfuscation feature.

[noble jasper]

no, there isn't. i'll add that to the todos list

[quartz wave]

it's needed to run main.py for some reason though

[noble jasper]

Why do you not want to install it?

[quartz wave]

i mean it's an external pip package which probably should be mentioned but i can just do that i guess

[noble jasper]

yea i updated the readme to mention it. honestly i created the readme using ai and didnt read it before posting. it's my fault

[arctic skiff]

explain?

[restive void]

I just copied the message above me which hadn't been !e'd. But it looks like ||the 22/7 is just a distraction, it just reads a docstring in the __hello__ module||.

[arctic skiff]

I just realized why that docstring contains pi, its because its for testing utf8

I didn't think that

though I do know thay 22/7. is a distraction

[arctic skiff]

what about "\" "

which isn't also printed

[restive void]

It's not "\" ", but "\␤" "", where the \ is the line continuation backslash. So it's just "" "" (implicit concatenation) plus the pi.

[arctic skiff]

couldn't see it properly on mobile

[magic wraith]

@noble jasper didn't you have another discord account earlier?

do you remember this?

[noble jasper]

i did but it got hacked and deleted

that was my V1 obfuscator

[cursive sequoia]

its my attempt at golfing tictactoe

You have to enter values from 0 to 8* to play moves

with the board formatted like this

0 1 2
3 4 5
6 7 8

[vale tangle]

>>> def d(n):     
...     print(n)
...     return n
... 
>>> a = [15,10,2]    
>>> a[-1]-=d(a.pop())
2
>>> a
[15,0] # expected outcome: [15,8]

i know that's some edge case i'm hitting and in general i should not do this, but... what? (i just joined this place and became too confused due to amount of channels, so im gonna ask here)

the console i ran this on is 3.14.0a1, but this seems to do the same in 3.12.4.

[subtle viper]

a[-1] is evaluated first

that's how -= and all other augmented assignments get their first value

LHS eval, RHS eval, BINOP done, LHS store

[vale tangle]

yeah, but i was confused because a[-2] gives an error with a=[10,2]

[subtle viper]

that error is on the store part

[vale tangle]

ohh okay thanks

so it reevaluates the index on load and store?

[subtle viper]

it reevaluates the subscript

the index itself used for the subscript is copied

a[i] -= x

is equivalent to ```py
_idx = i
a[_idx] = a[_idx] - x
#^^^^^^ ^^^^^^

different operations

[vale tangle]

okay i get it now, big thanks

i kinda imagined there was an optimisation there because the metamethods are separate 😆

but wait, what does _isub_ do then?

[subtle viper]

__isub__ defaults to regular __sub__ when unimplemented :p

[vale tangle]

oh 😐

[subtle viper]

for methods that implement it, could be used to modify inplace

that's for the object that's getting subtracted from though

storing into an index is a different operation that's done to the list and not the element getting subtracted from

[magic cosmos]

Give me some esoteric python code as a beginner

As a beginner to esoteric python, not to programming in python

[subtle viper]

!e ```
print(False == False in [True])

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

False

[magic cosmos]

I have downloaded the whole precedence in my brain

Level 2

Or level 1, if we consider this as level 0

Wait

Is that not about preceding

what to hell

[subtle viper]

!e ```
a = {}
b = a
a[0] = a = {}
print(a, b)

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

{} {0: {}}

[subtle viper]

i mean..

what precedence do you think would make sense here :p

[subtle viper]

!e ```
print((False == False) in [True])
print(False == (False in [True]))

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

001 | True
002 | True

[magic cosmos]

That thing is scary

Wth is happening

[magic cosmos]

Surely this isn't about == and is difference?

[subtle viper]

well, there's only == and in here

[magic cosmos]

!e False in [True]

[night quarry]

:warning: Your 3.14 eval job has completed with return code 0.

[No output]

[magic cosmos]

!e print(False in [True])

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

False

[magic cosmos]

I am questioning my life choices

Is this a bug or something

[subtle viper]

it's a feature! ^^

a common one too

[magic cosmos]

😭

[magic cosmos]

Nnoo

I will leave programming in python

[subtle viper]

in fact this feature is widely known for being pythonic

[magic cosmos]

This can't be true

Am i really that bad in python

[subtle viper]

hint: it can be used to check if something is between two values

[magic cosmos]

This thing makes no sense 😭 i have used walrus but this thing is beyond science

[magic cosmos]

Is this going algebra or something

[subtle viper]

(this specific case is very esoteric, however)

[magic cosmos]

x in y

in will check x's membership in y

== will check left.eq(right) first

[subtle viper]

regular python feature. like beginners can use it too, most people just don't use it this way (or even know it can be used this way)

[magic cosmos]

saying from my memory

[magic cosmos]

I don't think i ever did it

😭 chat am i that bad in python

[subtle viper]

have you ever done something like m <= x < n?

[magic cosmos]

start inclusive range

math yes

Isn't that like our builtin range

[magic cosmos]

which excludes end

[subtle viper]

which in other languages needs to be done as m <= x and x < n

[magic cosmos]

but what it does have anything to do with this

False == False in [True]

[magic cosmos]

is it that hard for them to implement

I give up chat

😔

Tell me the secret

[subtle viper]

not sure, but most have had chain comparisons unimplemented since their birth

[magic cosmos]

False == False and False in [True]

True and False -> False

Is that what it is

[subtle viper]

in is the same precedence as other comparison operators

[magic cosmos]

I don't get it if so

[subtle viper]

and since comparison operators are chained in python..

[magic cosmos]

in basically is == with iteration right

[subtle viper]

mhm, but that's not the focus here

[magic cosmos]

i see

[magic cosmos]

Is that how it is translated internally?

[subtle viper]

yup

[magic cosmos]

But why

[magic cosmos]

I would expect either of this to happen

[subtle viper]

in is in the same category as comparison operators in the parser

[magic cosmos]

[subtle viper]

why? idk :p

but it makes for a fun feature

[magic cosmos]

but that goes against how i see it

it should be a separate check

Iterate over iterable for each item, target._eq_(item).

Isn't it?

[magic cosmos]

Against*

[subtle viper]

yes, most implementations follow that

[magic cosmos]

Python being unpythonic

Chat, we need to nuke this thing

[subtle viper]

it's a python-only feature so i'd argue it's very pythonic ^^

[magic cosmos]

this doesn't happen in practical things and works as we usually expect, right?

Or i will have anxiety next time i use un parenthesized expression which contains in

[subtle viper]

nobody usually uses the in as a chain

and you only need to parenthesize when you're using comparison operators with it and you don't intend to chain it

[magic cosmos]

Now this

I think i saw something like this a few days ago in python discussion channel

[magic cosmos]

I see. Very tricky bug that can go un noticed and horrible

[magic cosmos]

a[0] ??????

Ok wait i thought it was a set

Nvm

[magic cosmos]

Because of right to left (reverse) assignment, such that a = {}, a[0] = {}, b contains the original dict which a used to reference?

Was that it?

[magic cosmos]

Wait

Wth

Why would that be assigned to b

[subtle viper]

i mean that would mean a would be assigned then a[0] would be assigned..

[magic cosmos]

i will never use indexed assignment with multiple variables

[subtle viper]

it's a very simple idea though

[magic cosmos]

Why is that all happening to b than a

https://tenor.com/view/alejandro9ix-gif-13711685647150536594

[subtle viper]

because i assigned b to a before it?

[magic cosmos]

Then it should happen to a too print shows its empty

😭

Okay wait

Don't tell answer until i say i give up

Hints are welcome

1. a[0] = {}
   then reassigned a

Oh i didn't have to think right to left assignment

a[0] = {} # b also gets that
a = {} # becomes empty

Okay that was easy, shouldn't have thought it is tricky

[subtle viper]

yup!

[magic cosmos]

That was honestly what i initially would i have thought had i not been biased and thought that probably its tricky so it could be right to left than usual left to right

Next please

It was fun.

[subtle viper]

with the exception of the rightmost expression which is evaluated first, python assignments are evaluated left to right

okay

() = ()

[magic cosmos]

I see

Thanks for your explanation

I wonder why "explain" contains "i" but not "explanation"

English can't stay predictive

I hate Inconsistency

[subtle viper]

simplification, probably
it's like how "comfortable" doesn't pronounce "for," or how "pronunciation" doesn't have the "o" in "pronounce"

[subtle viper]

!e this btw ```py
[] = []
() = ()

[night quarry]

:warning: Your 3.14 eval job has completed with return code 0.

[No output]

[magic cosmos]

What is there

That i will read () at left side as a tuple assignment?

both sides are thought of as expressions than tuple, even though they're empty to be literal'ed as tuple

Not sure why it does that but makes sense if it contains expressions

So there's nothing to carry the empty tuple so it just does not get assigned

Hence it gets just... Forgetten

Is that all

[subtle viper]

there's no assignment being done, that's true

right side is always evaluated and left side is always a target in an assignment though

!e even though it looks like nothing... ```py
() = 1

[night quarry]

:x: Your 3.14 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "/home/main.py", line 1, in <module>
003 |     () = 1
004 |     ^^
005 | TypeError: cannot unpack non-iterable int object

[subtle viper]

they need to fix that error position :p

whatever

[magic cosmos]

Lol

Next

Gotta binge it

Why does it allow for [] = []

Tuple is understandable

Its not even a subscription

[subtle viper]

it's practically the same thing

[magic cosmos]

[] considered as an expression than a list literal?

https://tenor.com/view/alejandro9ix-gif-13711685647150536594

[subtle viper]

it looks like a list literal, but it's the same as a tuple literal in the LHS

[magic cosmos]

!e [a, b] = [1, 2]
print(a, b)

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

1 2

[magic cosmos]

Yo you're rigjt

Right

Why tho, shouldn't it only group by tuple for expressions than list literals

!e {} = {}

[night quarry]

:x: Your 3.14 eval job has completed with return code 1.

001 |   File "/home/main.py", line 1
002 |     {} = {}
003 |     ^^
004 | SyntaxError: cannot assign to dict literal here. Maybe you meant '==' instead of '='?

[magic cosmos]

now that's racist

discrimination

[], () allowed but not {}

[subtle viper]

a list is also sequence, so

[magic cosmos]

next please

gotta binge 'em all

[undone thicket]

ok wtf

[magic cosmos]

Valid crashout

[undone thicket]

this scares me

[magic cosmos]

@undone thicket see rhis

[undone thicket]

i will not sleep tonight

[magic cosmos]

😈😈

[undone thicket]

HUH?!?

why cant u do it for dicts

oh ig they're hashmaps so different but still jeez

[magic cosmos]

😭 but why even allow anything at all than ()

For whatever reasons Just don't

[fallen heath]

!e

a = 1, 2, 3
b = {}

for c, b[c] in zip(a,a):...
print(b)```

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

{1: 1, 2: 2, 3: 3}

[magic cosmos]

I think this because you can put any assignable name in for variables

[subtle viper]

sequences
[a, b, c] = 1, 2, 3 functions the same as (a, b, c) = 1, 2, 3

[magic cosmos]

https://tenor.com/view/alejandro9ix-gif-13711685647150536594

[magic cosmos]

But dict is also a sequence but with two pairs

[subtle viper]

it's a hashmap

[magic cosmos]

Damn

Mapping

This chat does not have slow mode?

[subtle viper]

nope

[magic cosmos]

Wow time to spam :)

[subtle viper]

i mean you still have the bot limits

but yea

[magic cosmos]

xD

At least i can send previewable CJ gifs

Send me next esoteric python if you got

[magic cosmos]

Mod 💀

[subtle viper]

i can't really think of other things.. hmm
-# not in the time i currently have, that is

[magic cosmos]

Ohk, thank you so much for today

[versed eagle]

!e
this is a little less esotericpython per se but its a fun side of python to know about

x = 0 # using x to prevent const-folding
print(True + x + True) # 2

from ctypes import *
cast(id(True),POINTER(c_int))[6]=100
print(True + x + True) # 200

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

001 | 2
002 | 200

[cursive sequoia]

wait what how does this work?

[versed eagle]

!e hint: ```py
from ctypes import *
cast(id(True),POINTER(c_int))[6]=100
print(int(True))

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

100

[sullen kayak]

hey there demons, it's me — ya boy:

import math; from rlbot.agents.base_agent import BaseAgent, SimpleControllerState;from rlbot.utils.structures.game_data_struct import GameTickPacket; from util.boost_pad_tracker import BoostPadTracker;from util.orientation import Orientation, relative_location;from util.drive import steer_toward_target;from util.vec import Vec3;Derived = type("Derived", (BaseAgent,), {"__init__": lambda self, name, team, index: (super(Derived, self).__init__(name, team, index), setattr(self, "boost_pad_tracker", BoostPadTracker()))and None, "initialize_agent": lambda self: self.boost_pad_tracker.initialize_boosts(self.get_field_info()), "get_output": lambda self, packet: (self.boost_pad_tracker.update_boost_status(packet), setattr(self, "target_location", min([[i.location, Vec3(packet.game_cars[self.index].physics.location).dist(i.location)] for i in self.boost_pad_tracker.boost_pads if i.is_active == True], key=lambda i: i[1])[0]), setattr(self, "controls", SimpleControllerState()),setattr(self.controls, "steer", steer_toward_target(packet.game_cars[self.index], self.target_location)), setattr(self.controls, "boost", True), setattr(self.controls, "throttle", 1.0),setattr(self.controls, "handbrake", abs(math.atan2(relative_location(Vec3(packet.game_cars[self.index].physics.location), Orientation(packet.game_cars[self.index].physics.rotation), self.target_location).y, relative_location(Vec3(packet.game_cars[self.index].physics.location), Orientation(packet.game_cars[self.index].physics.rotation), self.target_location).x)) > 1.5), )and self.controls})

anyone bored enough to take a stab at making this harder to read?

[magic cosmos]

Imagine debugging that

😭

[vast wave]

whats the memory layout of the bool type again

[magic wraith]

how about this

[night quarry]

Click here to see this code in our pastebin.

[dense wing]

Is there a way to ignore NameErrors/automatically initialize an uninit variable's value?
I'm trying to get the following to run:

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

hello(James) # prints 'Hello, James!'```

LLMs suggested import hooks but ideally I'd like to have everything in a single file :p

[magic wraith]

you want it to become a string?

[dense wing]

yup

without e.g. needing to initialize it to some arbitrary value and looking up globals

[magic wraith]

I think i've seen some code in this channel doing that by overriding __getattr__ or __getattribute__ on builtins somehow

[earnest wing]

you use ctypes to replace the globals object with your own dict subtype that implements __missing__

[dense wing]

i don't have much experience with ctypes, how does it allow for replacing globals?

[dense wing]

nevermind I got it

[magic wraith]

!e this is how one can do it in a function ig ```py
import builtins, types

class BuiltinsProxy(dict):
def missing(self, key):
return key

x = 10
def template():
return x, foo, len([1, 2, 3])

g = dict(template.globals)
g["builtins"] = BuiltinsProxy(vars(builtins))

weird = types.FunctionType(template.code, g, template.name, template.defaults, template.closure)
print(weird())```

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

(10, 'foo', 3)

[magic cosmos]

Guys, give me next esoteric python code as a beginner to esoteric python

Gotta grind 'em all

[earnest wing]

and inside a class scope you can also use a metaclass with a custom namespace type

[versed eagle]

refcnt - 8b
type ptr - 8b
num limbs - 8b
int value - num limbs * 4b

its just an int

with different tp ptr

[magic wraith]

isnt that python version dependent?

[subtle viper]

relatively the same for most of the recent versions

the only difference would be the platform or the kind of interpreter

[arctic skiff]

TIL that in faq section of python docs there is faq about one liners

[quartz wave]

officially mandated esoteric python

[terse pollen]

huh, I don't remember that entry...

[subtle viper]

part of the way there https://paste.pythondiscord.com/ZKMA

[karmic pumice]

who if not you would do such a thing

[subtle viper]

reduced like a hundred lines from redundant assignments and other stuff https://paste.pythondiscord.com/ZPXA

i know it's kind of obvious for anyone who read the obfuscator's source code fully but there might be a pattern for the header

i guess this part is the hard part

this is asking too much of a python hobbyist and not deobfuscation expert..

uu..

[subtle viper]

still some redundant parts left and i'm figuring out the other stuff. i'll work on this tomorrow

heh

wrong account soz

[arctic skiff]

I think python would be one the most popular languages to write esoteric code because of how it was designed to be intuitive and how hell if can be because of complex design

[subtle viper]

!e ```py
with open("_A__g.py", "w+") as f:
f.write("print('Hello, World!')")

class __A:
import __g

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

Hello, World!

[terse pollen]

Because of name mangling, yes?

[zealous aurora]

who here knows anything about array.array

looking into it as a possible data type to use over lists and dicts (im in a competition where third party libraries and c modules are banned)

[restive void]

Have you read its docs? What are you wondering about?

[zealous aurora]

any way of doing vector ops on them?

[gleaming linden]

not really afaik

[versed eagle]

not very much

[wintry oasis]

Hello, i am doing an side project. and i want to obfuscate the source code. is there good obfuscate services?
i tried to see some providers. and they almost all of them is being cracked by Low-experience reverse

[magic wraith]

then don't write it in python or keep it behind a web server

[magic wraith]

because obfuscated python is much easier to reveng than normally compiled go or rust, for example
also let me know what you think about this@wintry oasis

[wintry oasis]

🌚 alot of my friends told me python is not that good in obfuscate, and now i realize...

Thanks i think i'll go for rust or C++

[wintry oasis]

i'll change the project into rust or C++ Thanks for help

[wintry oasis]

Hello again, i found PyArmor. is it good? paid version ofc
my goal is keeping my source away from newbie reverse kids

[gleaming linden]

honestly if that's your goal then pyarmor is probably good enough

but it can also be reversed with some degree of knowledge/effort

[subtle viper]

all obfuscation is "reversible," like how most prisons are escapable, but the guarding and other manners of prevention are there to deter people from doing so

[subtle viper]

reversible as long as it contains the actual code logic, that is

[magic wraith]

as a rule of thumb, if the code is something you definitely don't want anyone to see ever, then offer that as a service over http for example

[vast wave]

as a rule of thumb, all obfuscation is reversible; all code that can be executed can be reverse engineered.

[noble jasper]

check out mine. its in early stages of development, but i promise its better than 99.9% of the other obfuscators out there. https://github.com/lyric0x10/Oblivion

honestly, i would still try to use rust or c++.

[tribal bronze]

I don't know if this is the right sub chat but I was doing some research on Lean and I was wondering if anybody in here has tried making any libraries or code in general to kind of work alongside that? I think it would be really interesting to try to integrate an agentic system to kind of formulate new theorems I'm sure it already exists but just to kind of see what already exists

[magic wraith]

@noble jasper please write some proper python code, apply the obfuscator and send it here as a challenge. I want to figure out how well LLMs work against it

again, i got this for you #esoteric-python message if you wanna play around with my obfuscated code from 2023

[craggy mason]

do u need dev?

[cursive loom]

AI can one shot deobf for it, but I assume it can for any open source obfuscator

[magic wraith]

it being open source doesn't mean anything because ai can't just search it based on the code

with that said, chatgpt managed to deobfuscate (into bytecode, not completely decompile) pjorion obfuscations when i gave it the obfuscator src code

but it still struggled against this one nevermind, it beat it this time lol

[cursive loom]

no like i got it to write the full decompiler. i got to the same point as u, just ask it to lift the bytecode to an AST then run ast.unparse on it and the results are clean

[noble jasper]

what do you mean proper? are you asking for a real world looking file? give me a little while, ill have chatgpt generate one, and ill give you the obfuscated file.

[magic wraith]

something nontrivial

[noble jasper]

@magic wraith

[night quarry]

Click here to see this code in our pastebin.

[magic wraith]

https://paste.pythondiscord.com/4VLPRTJC7PW5YCUTNDOTPHSY5A

btw I recommend you just call python-minifier in the end on the obfuscated script, and avoid hardcoded arithmetic that constant-folding just simplifies, since it doesn't add any obfuscation it just increases code size for no reason

[noble jasper]

yea good job

[chilly island]

@noble jasper What's the purpose of your Oblivion Obfuscator? Are you making it to learn more about Python or does it solve an actual problem?

[noble jasper]

I just made it for fun. I was never planning on posting it. It's still early on in development and can be deobfuscated, like any obfuscator can.

[chilly island]

That seems like a fun project. How much Python experience do you have to make such a high quality level obfuscator?

I'm not that good at Python and I have like no idea what any of your Features are.

[noble jasper]

I’ve been coding since 2021. It’s less about skill and more about enjoying it and sticking with it. You can be bad at first, but if you like coding, you WILL get better. I love it. It’s one of my favorite hobbies and I code a few hours a day.

Don't try making a calculator app or a snake game or any bullshit tutorial. Find an interest in coding and just start learning. Try to use LLM's as little as possible so you don't become a vibecoder.

[chilly island]

Dang. That seems like a pretty long time. I'm currently learning some python libraries right now and I probably should spend some more time on it.

[noble jasper]

Wanting to make a great project will never get you further than wanting to code. If you like to code and its a genuine interest, eventually you will make something great. Focus on small stuff like gaining an understanding of the language before you try to make a project.

[chilly island]

Yeah. I'm starting off with the basics and working my way up.

[wraith granite]

is there a way to skip indentations??

like

for t in range(10):

x = sin(t)
y = cos(t)

??????
i just need the x and y declaration part to be starting from newline

[hybrid trail]

good question

you can do it with a bunch of junk, im not sure if there's a cleaner way

#coding:u7
from math import cos, sin
for t in range(10):          #+AAoAIABc
x=cos(t)                     #+AAoAIABc
y=sin(t)                     #+AAoAIABc
print(f'{x**2 +- y**2:.2f}')

[magic wraith]

!e i hate every single bit of this py from math import cos, sin for t in range(10):\ (x:=cos(t)),\ (y:=cos(t)),\ print(f'{x**2 + y**2:.2f}')

[gleaming linden]

for t in range(10):
    exec("""
x = sin(t)
y = cos(t)
    """)

[magic wraith]

if it has to start with x = i think it can be done like this```py
from math import cos, sin

class A:
def init(self, *magic):
...
def eq(self, other):
magic: ...
return True
x: int = A('x')
y: int = A('y')

for t in range(10):
x == sin(t),
y == cos(t),
print(f'{x2 + y2:.2f}')``` so the entire bottom part is an expr, but the __eq__ dunder assigns a global variable or something

[fluid pumice]

!e

_=([...,[id]]!=[{...:...,...:...,...:id}]);___=_-_;__=_<<_;open(1,'w').write(bytes([_:=([...,[id]]!=[{...:...,...:...,...:id}]),____:=(___*(__**((__<<__)-_)))+(_*(__**((__<<__)-__)))+(_*(__*((__<<_)+_)))+___+(_*(__*(__+_)))+(_*(__**___))+(_*(__**__*__<<_<<_)-__**__-__-__),____-_][::-_][:-_]).decode())

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

hi

[magic wraith]

!e ```py
from math import cos, sin
xꘌy=1
yꘌx=2
for t in range(10):
xꘌy,
yꘌx,
print(f'{yꘌx2 + xꘌy2:.2f}', end=', ')

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

5.00, 5.00, 5.00, 5.00, 5.00, 5.00, 5.00, 5.00, 5.00, 5.00, 

[quartz wave]

!e ```py
from math import sin, cos
for t in range(10):

x = sin(t)

y = cos(t)

print(f'{x2 + y2:.2f}')

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

001 | 1.00
002 | 1.00
003 | 1.00
004 | 1.00
005 | 1.00
006 | 1.00
007 | 1.00
008 | 1.00
009 | 1.00
010 | 1.00

[magic wraith]

yeah they are all ones, it didn't quite work

also

[quartz wave]

!e ```py
from math import sin, cos
for t in range(10):
x = sin(t)
y = cos(t)
print(f'{x2 + y2:.2f}')

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

001 | 1.00
002 | 1.00
003 | 1.00
004 | 1.00
005 | 1.00
006 | 1.00
007 | 1.00
008 | 1.00
009 | 1.00
010 | 1.00

[magic wraith]

oh what

wait

[quartz wave]

faulty error detector

[magic wraith]

ok nevermind i'm totally dumb i had cos cos here instead of sin cos and i remembered various numbers

[quartz wave]

these for explanation
<links>

no actually

this explains better
https://github.com/python/cpython/blob/main/Parser/lexer/lexer.c#L532-L538

[night quarry]

Parser/lexer/lexer.c lines 532 to 538

else if (c == '\\') {
    // Indentation cannot be split over multiple physical lines
    // using backslashes. This means that if we found a backslash
    // preceded by whitespace, **the first one we find** determines
    // the level of indentation of whatever comes next.
    cont_line_col = cont_line_col ? cont_line_col : col;
    if ((c = tok_continuation_line(tok)) == -1) {```

[quartz wave]

!e so this is valid too ```py
from math import sin, cos
for t in range(10):

x = sin(t);
y = cos(t)
print(f'{x2 + y2:.2f}')

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

001 | 1.00
002 | 1.00
003 | 1.00
004 | 1.00
005 | 1.00
006 | 1.00
007 | 1.00
008 | 1.00
009 | 1.00
010 | 1.00

[wraith granite]

yoooo escaping nl works in python?? cool to learn

the best way i found that worked for me was to just use exec with multiline strings tho

thanks

[long fulcrum]

any way to shorten this?

import sys
for b in sys.argv[1:]:print("".join(f"\n{sum(b[p+i-89:][:p+i>88]=="X"for p in b"MNOXZcde")}F"[2+ord(b[i])&3]for i in range(109))+"\n")```

[long fulcrum]

-1

import sys
for b in sys.argv[1:]:print("".join(f"F \n{sum(b[p+i-89:][:p+i>88]=="X"for p in b"MNOXZcde")}"[ord(b[i])%4]for i in range(109))+"\n")```

[glass canopy]

i am havving a difficult time recereating a better sonification than john cramer did in 2013

[reef mesa]

.l

[long fulcrum]

.l?

[reef mesa]

idk

[echo whale]

I've been looking at this for awhile and I really cant come up with anything else that isn't the same length or slightly longer

[long fulcrum]

the ord(b[i])%4 part is for the inputs which can either be \n, X or _

[spice grail]

it looks like a grid problem and the grid size is fixed, it is possible that a regex solution would be shorter here but can't say without the problem

ah its minesweeper from c.g

without giving too much away: ||it is possible to get adjacent cells using some clever slicing after which you can use .count||

[keen lichen]

hi guys, i'm new to esoteric python. i programmed the game of life within pickle: https://linus.space/posts/2026-04-03-pickle.html

im convinced that you can do even crazier stuff within these constraints, and also build a simpler impl of the game of life

!e

p = """cbuiltins
print
(VHello, world!
tR.
"""
import pickle
pickle.loads(bytes(p, "utf-8"))

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

Hello, world!

[keen lichen]

!e

p="""ccollections
deque
((I0
I1
lI2
tRp0
0ccollections
deque
(cbuiltins
zip
(cbuiltins
map
(cbuiltins
print
cbuiltins
map
(coperator
itemgetter
(I0
tRcitertools
repeat
(g0
tRtRtRcbuiltins
map
(cbuiltins
getattr
(g0
Vappend
tRcbuiltins
map
(cbuiltins
sum
citertools
repeat
(g0
I10
tRtRtRtRI0
tR."""
import pickle; pickle.loads(bytes(p, "utf-8"))

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

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

Full output: https://paste.pythondiscord.com/SALTFB2EEU46C4BC7RB3Z74YF4

[restive void]

Cool! Is this the earliest pickle protocol?

[keen lichen]

yes, this is the protocol 0

[long fulcrum]

I can't figure this out but I've saved a few more

import sys
for*b,in sys.argv[1:]:b+="\n"*14;[print(end=f"F \n{sum(b[p+i-89]=="X"for p in b"MNOXZcde")}"[ord(b[i])%4])for i in range(111)]```

[terse pollen]

couldn't you just... write the program normally and then pickle the module?

but I applaud the effort to learn a format deeply enough to do that sort of thing. Reminds me of a younger me.

[keen lichen]

that would be writing python

[terse pollen]

yeah, that was my point; you can get the result you describe without working at the level you work at. But I guess working at that level was the point for you

[keen lichen]

yeah, not writing python is explicitly a constraint

[spice grail]

http://golf.shinh.org/p.rb?Minesweeper specifically leonid's python solution

[long fulcrum]

managed 131

import sys
for t in sys.argv[1:]:
 s=(' '*12+t)*2
 for x in t+"&&":print(end=f"F \n{(s[:32]*3)[::11].count("X")}"[ord(x)%4]);_,*s=s```

[magic cosmos]

...

[terse pollen]

yeah, that's not how it works

you wouldn't learn an unusual foreign language by "give me the next sentence" either

[spice grail]

!e was reading a bit of the backlog, combining loop stuff with assignment stuff

for () in [(), [], {}, set()]:
    print(1)

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

001 | 1
002 | 1
003 | 1
004 | 1

[spice grail]

!e

import sys
print(sys.path)

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

['/home', '/snekbin/python/3.14/lib/python314.zip', '/snekbin/python/3.14/lib/python3.14', '/snekbin/python/3.14/lib/python3.14/lib-dynload', '/snekbox/user_base/lib/python3.14/site-packages', '/snekbin/python/3.14/lib/python3.14/site-packages']

[spice grail]

!e not 100% sure this work on this environment, but an all time favorite of mine

from sys import*
s="guvf cebtenz cresbezf ebg13 ba f"
exec(open(path[2]+'/this.py').read()[4:])

yeah doesn't work unfortunately :(

[magic cosmos]

Lmao

[compact crest]

!e

from sys import* 
s="guvf cebtenz cresbezf ebg13 ba f"
exec(open(path[2]+'/this.py').read()[4:])

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

this program performs rot13 on s

[magic wraith]

!e py import sys;print(sys.path) i was curious

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

['/home', '/snekbin/python/3.14/lib/python314.zip', '/snekbin/python/3.14/lib/python3.14', '/snekbin/python/3.14/lib/python3.14/lib-dynload', '/snekbox/user_base/lib/python3.14/site-packages', '/snekbin/python/3.14/lib/python3.14/site-packages']

[long fulcrum]

-2

import sys
z=sys.argv[1].replace(*" .")
exec('z=[[v," "][2<sum(z[(x+i)%2651]<v=="."for i in(-52,-1,1,52))]for x,v in enumerate(z)];'*650)
print(*z,sep="")```

I spent so long figuring that out

[long fulcrum]

I now see that I can just do (z+z)[x+i] but unfortunately it's too slow

[stark granite]

has anyone got the snippet which makes HelloWorld("print") work?

i remember it was doing some cursery to globals or smth which makes variable resolution funky

[karmic pumice]

!e

from ctypes import py_object

class globols(dict):
    __slots__ = ()
    def __missing__(_, name, builtins=__builtins__):
        try:
            return builtins.__dict__[name]
        except KeyError:
            return lambda fn: builtins.__dict__[fn](name)

py_object.from_address(id(globals()) + tuple.__itemsize__).value = globols

HelloWorld("print")

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

HelloWorld

[karmic pumice]

(depends on object layout, so you need to figure out the offset for freethreaded builds separately)

[stark granite]

thanks!

[arctic skiff]

!e```py
import ast
from linecache import*
import sys as s
from traceback import*
s.excepthook=lambda t,v,tb:getattr(builtins,(c:=ast.parse(getline((e:=extract_tb(tb)[-1]).filename,e.lineno)).body[0].value).args[0].value)(c.func.id)if issubclass(t,NameError)else s.excepthook(t,v,tb)

Hello_world("print")

without globals

[night quarry]

:x: Your 3.14 eval job has completed with return code 1.

Hello_world

[arctic skiff]

it translates n('f', a, b=c) to f('n')
when n is a undefined identifier

[mystic vapor]

beautiful

[stark granite]

._.

[long fulcrum]

150

import sys
z=sys.argv[1].replace(*" .")
exec('z=[[v," "][2<sum(z[x-i]<v=="."for i in(1,52,2599,2650))]for x,v in enumerate(z)];'*650)
print(*z,sep="")```

[bright valve]

Is this passing the function name to the function name passed as parameter?

[subtle viper]

yes

an unassigned global basically defaults to an anonymous function that calls the builtin function referred to by the string argument (parameter fn) with that call's argument as the unassigned global's name

[bright valve]

Thats fucking smort

[arctic skiff]

what is the input supposed to be?

[arctic skiff]

148py import sys z=sys.argv[1].replace(*" .") exec('x,z=0,[[v," ",x:=x+1][2<sum(z[x-i]<v=="."for i in(1,52,2599,2650))]for v in z];'*650) print(*z,sep="")

[past jungle]

!e
py
import sys
z=sys.argv[1].replace(*" .")
exec('x,z=0,[[v," ",x:=x+1][2<sum(z[x-i]<v=="."for i in(1,52,2599,2650))]for v in z];'*650)
print(*z,sep="")

!e
import sys
z=sys.argv[1].replace(*" .")
exec('x,z=0,[[v," ",x:=x+1][2<sum(z[x-i]<v=="."for i in(1,52,2599,2650))]for v in z];'*650)
print(*z,sep="")

[long fulcrum]

x isn't defined here but this 149 works

import sys
z=sys.argv[1].replace(*" .")
exec('x=-1;z=[[v," ",x:=x+1][2<sum(z[x-i]<v=="."for i in(1,52,2599,2650))]for v in z];'*650)
print(*z,sep="")```

[long fulcrum]

there's a sample input here

[arctic skiff]

hm right

thought that would be 148?

oh wait you did -1

nvm

so yeah 149

[terse pollen]

is this still the maze solver?

[long fulcrum]

yeah

[long fulcrum]

146

import sys
z=sys.argv[1].replace(*" .")
for _ in z:x=-1;z="".join([v," ",x:=x+1][2<sum(z[x-i]<v=="."for i in(1,52,2599,2650))]for v in z)
print(z)```

[spice grail]

144

import sys
z=sys.argv[1].replace(*" .")
for _ in z:x=0;z="".join([v," ",x:=x+1][2<sum(z[x-ord(i)]<v=="."for i in"5ਨਜ਼")]for v in z)
print(z)

contains \x02 not sure if that copies right

[terse pollen]

are you counting bytes or characters now?

[spice grail]

bytes

[terse pollen]

ah but you want to use unicode just so that you have the wider range of values, ok

(I say "unicode" but I mean "code points 256 and up", I guess)

unpacking from an integer would perhaps avoid the ord call but is surely a net loss...

[long fulcrum]

I never thought of that

I tried using bytestrings but they only go to 255

[terse pollen]

Indeed.

[balmy condor]

Probably the wrong place to ask, but i cant find a haskell server, so ig this falls into functional programming in general:

how do implementations of functional languages optimise functions that cant be condensed into tail-recursive versions of themselves to avoid a stack overflow error?

[austere mauve]

I really don't know, I'm just guessing how I would attempt implementing something like this:

• In some cases it might be possible to change the code in a way that it uses tail call elimination (quite complex, only sub-cases)
• Optimise the stack as much as possible (reduce the cost of not eliminated recursions)

[austere mauve]

I suppose this is one of the ways to tackle the problem:
https://www.youtube.com/watch?v=JzMCwokUTl4

[karmic pumice]

cant find a haskell server
https://discord.gg/K6XHBSh

https://discord.gg/4Kjt3ZE would also be on topic

[karmic pumice]

the general answer is they dont, there is no magical silver bullet that would make functional languages avoid a call stack, e.g. f(x) = h(f(g1(x)), f(g2(x))) will almost always be horrible
but one step above tail calls is tail call modulo constructors
having a call-stack-on-the-heap helps raise the limit such that you wont overflow it in practice
trampolining and CPS are a thing too

[balmy condor]

Thanks! 🙂

[balmy condor]

I see, ill look into those - cheers

[coarse void]

event, setmdir, moveapple, move, draw, keymap, applepos, movedir, win, snake = lambda: [setmdir(movedir,keymap[event.key]) for event in __import__("pygame").event.get() if event.type == __import__("pygame").KEYDOWN and event.key in list(keymap)],lambda mdir,di: mdir.__setitem__(0,di), lambda: applepos.__setitem__(0, [__import__("random").randint(0,9)*80, __import__("random").randint(0,9)*80]),lambda: [snake.append([snake[-1][0]+movedir[0][0] if snake[-1][0]+movedir[0][0] != 800 and snake[-1][0]+movedir[0][0] != -80 else (0 if snake[-1][0] == 720 else 720),snake[-1][1]+movedir[0][1] if snake[-1][1]+movedir[0][1] != 800 and snake[-1][1]+movedir[0][1] != -80 else (0 if snake[-1][1]==720 else 720)]),snake.pop(0) if snake[-1][0] != applepos[0][0] or snake[-1][1] != applepos[0][1] else moveapple()],lambda: [win.fill((255,255,255)),__import__("pygame").draw.rect(win, (255,0,0), (*applepos[0], 80, 80)),*(__import__("pygame").draw.rect(win,(0,255,0), (*pos, 80, 80)) for pos in snake),*(__import__("pygame").draw.rect(win,(0,0,0), (x*80,y*80,80,80), 2) for x in range(10) for y in range(10)),__import__("pygame").display.flip()],{__import__("pygame").K_w:[0,-80],__import__("pygame").K_a:[-80,0],__import__("pygame").K_s:[0,80],__import__("pygame").K_d:[80,0]},[[320,560]],[[80, 0]],__import__("pygame").display.set_mode((800,800)), [[240, 240], [240, 320], [240, 400]]

[[event(),move(),draw(),__import__("time").sleep(0.5)] for _ in iter(lambda: 0,1)]

pygame snake in 2 lines

[subtle viper]

event, setmdir, moveapple, move, draw, keymap, applepos, movedir, win, snake = [[event(),move(),draw(),__import__("time").sleep(0.5)] for _ in iter(lambda: 0,1)][0] = lambda: [setmdir(movedir,keymap[event.key]) for event in __import__("pygame").event.get() if event.type == __import__("pygame").KEYDOWN and event.key in list(keymap)],lambda mdir,di: mdir.__setitem__(0,di), lambda: applepos.__setitem__(0, [__import__("random").randint(0,9)*80, __import__("random").randint(0,9)*80]),lambda: [snake.append([snake[-1][0]+movedir[0][0] if snake[-1][0]+movedir[0][0] != 800 and snake[-1][0]+movedir[0][0] != -80 else (0 if snake[-1][0] == 720 else 720),snake[-1][1]+movedir[0][1] if snake[-1][1]+movedir[0][1] != 800 and snake[-1][1]+movedir[0][1] != -80 else (0 if snake[-1][1]==720 else 720)]),snake.pop(0) if snake[-1][0] != applepos[0][0] or snake[-1][1] != applepos[0][1] else moveapple()],lambda: [win.fill((255,255,255)),__import__("pygame").draw.rect(win, (255,0,0), (*applepos[0], 80, 80)),*(__import__("pygame").draw.rect(win,(0,255,0), (*pos, 80, 80)) for pos in snake),*(__import__("pygame").draw.rect(win,(0,0,0), (x*80,y*80,80,80), 2) for x in range(10) for y in range(10)),__import__("pygame").display.flip()],{__import__("pygame").K_w:[0,-80],__import__("pygame").K_a:[-80,0],__import__("pygame").K_s:[0,80],__import__("pygame").K_d:[80,0]},[[320,560]],[[80, 0]],__import__("pygame").display.set_mode((800,800)), [[240, 240], [240, 320], [240, 400]]

-# pygame snake in 1 line

[tawny sierra]

Hello world (No strings, no prints, no imports, cross-platform)

[night quarry]

Click here to see this code in our pastebin.

[earnest wing]

"strictly increasing, but any number of them" => "increasing (not necessarily strictly)"

i would say the question is not well formed but it can be made so
namely, "as short as possible" doesn't really make sense in the context of arbitrary inputs: so long as every possible sequence of digits decodes to some input string, your solution is in some way "optimal" -- for every input string, the output string is the shortest one that's not already taken by a different input string
what would make the challenge work better is that the size of the output string should be similar in size to the input string, so short strings turn short.

[earnest wing]

it depends on your notion of reversible :) either the encoding is bijective or injective

[coarse void]

There is something about codes that are not one-liners, but look so thick and complicated.
This codes reads from a GPX file and creates an ascii art of it, including elevation with ASCII ramp.

[gleaming linden]

unnecessary space spotted

[subtle viper]

just looks like my average aoc rush code

[coarse void]

noo 😭

[serene stratus]

another one

[coarse void]

yeah there are two on the first line as well

with(open("Tubes_Ultra_Trail_5_laps.gpx","r"))as(readfile):data=readfile.read().split("\n")
data="".join([row.strip()for(row)in(data)])
data,points=data[data.index("<trkseg>")+8:data.index("</trkseg>")].split("</trkpt>"),[]
for(row)in((i)for(i)in(data)if(len(i))):
    num,long="",row.index("lon=")+5
    while row[long]!="\"":num+=row[long];long+=1
    long=float(num);num,lat ="",row.index("lat=")+5
    while row[lat]!="\"":num+=row[lat];lat+=1
    lat=float(num);elev=float(row[row.index("<ele>")+5:row.index("</ele>")]);points.append((lat,long,elev))
long_min,long_max=min(points,key=lambda x:x[1])[1],max(points,key=lambda x:x[1])[1]
lat_min,lat_max=min(points,key=lambda x:x[0])[0],max(points,key=lambda x:x[0])[0]
elev_min,elev_max=min(points,key=lambda x:x[2])[2],max(points,key=lambda x:x[2])[2]
gradient="$@B%8&WM#*oahkbdpqwmZO0QLCJUYXzcvunxrjft/\|()1{}[]?-_+~<>i!lI;:,\"^ "
elev_diff=elev_max-elev_min;elev_ratio=len(gradient)/elev_diff
width=int(int(input("How many characters wide should the representation be?\n\t> ")))
ratio,art_points,art_dict=width/(long_max-long_min),[],{}
for(long,lat,elev)in((int((i[1]-long_min)*ratio),int((i[0]-lat_min)*ratio),i[2])for(i)in(points)):
    if(long,lat)not in(art_points):art_points.append((long,lat));art_dict[(long,lat)]=elev
art_lat_max=max(art_points,key=lambda x:x[0])[0]
for(y)in(range(art_lat_max+1,0,-1)):
    for(x)in(range(width+1)):
        if(x,y)in art_points:print(gradient[::-1][int((art_dict[(x,y)]-elev_min)*elev_ratio)-1]*(1+x%2),end="")
        else:print("."if(x%2==0)and(y%2)else(" "*(1+x%2)),end=(""if(x!=width)else"\n"))

There you go play around with it, i removed all spaces except lambda x: I don't know, is there a way to get rid of that?

[karmic pumice]

(lat_min,lat_max),(long_min,long_max),(elev_min,elev_max)=[(min(points,key=(k:=__import__("operator").itemgetter(i))),max(points,key=k))for(i)in(range(3))]

something like that could work i think
oh also need to access the [i]th item so maybe just

(lat_min,lat_max),(long_min,long_max),(elev_min,elev_max)=[(min(xs:=[p[i]for(p)in(points)]),max(xs))for(i)in(range(3))]

thats also much less chars than

long_min,long_max=min(points,key=lambda x:x[1])[1],max(points,key=lambda x:x[1])[1]
lat_min,lat_max=min(points,key=lambda x:x[0])[0],max(points,key=lambda x:x[0])[0]
elev_min,elev_max=min(points,key=lambda x:x[2])[2],max(points,key=lambda x:x[2])[2]

respectively

art_lat_max=max(art_points,key=lambda x:x[0])[0]
#
art_lat_max=max(p[0]for(p)in(art_points))

also the double int on input is funny

python should get a minmax builtin

[fallen heath]

would that see enough use to be a build in?

[subtle viper]

maybe
the implementation is also already there; both min() and max() calls into a single function that gets either

[coarse void]

oh the double int was left there by accident 😛 Also can you ELI5 on this

(lat_min,lat_max),(long_min,long_max),(elev_min,elev_max)=[(min(xs:=[p[i]for(p)in(points)]),max(xs))for(i)in(range(3))]``` tomfoolery?

[terse pollen]

just unroll the (outer) list comprehension's loop:

(lat_min,lat_max) = (min(xs:=[p[0]for(p)in(points)]),max(xs))
(long_min,long_max) = (min(xs:=[p[1]for(p)in(points)]),max(xs))
(elev_min,elev_max) = (min(xs:=[p[2]for(p)in(points)]),max(xs))

the := is just being used to avoid repeating the inner comprehension for the min and max.

The next step of deobfuscation (assuming we don't care about the scope of xs, which we wouldn't as it's just a temporary for the calculation) would be something like

xs = [p[0]for(p)in(points)]
(lat_min,lat_max) = (min(xs),max(xs))
xs = [p[1]for(p)in(points)]
(long_min,long_max) = (min(xs),max(xs))
xs = [p[2]for(p)in(points)]
(elev_min,elev_max) = (min(xs),max(xs))

which is straightforward, just strangely punctuated

[karmic pumice]

well really it should be

(lat_min,lat_max),(long_min,long_max),(elev_min,elev_max)=((l:=([[inf:=float("inf"),-inf]for(_)in"___"]))and(s:=lambda*k:l[k[0]].__setitem__(k[1],k[2]))and(all((s(i,0,p[i])if(p[i]<l[i][0])else(s(i,1,p[i])if(p[i]>l[i][1])else()))or(1)for(p)in(points)for(i)in(0,1,2)))and(0))or(l)

to be single pass over points
(i really like single pass)

[karmic pumice]

oh also you can technically get around the space in lambda x:f(x) with lambda*x:f(x[0])
but it makes it a bit longer

[subtle viper]

or lambda*x:f(*x) for function calls specifically

things like lambda*x:*x+ would be such a cool feature though

at that point adding implicit lambdas would be nice too

[austere mauve]

I wonder how syntax should look like for inline functions to avoid ambiguities while keeping it pythonic.

[austere mauve]

Mainly as function arguments with other arguments afterwards:

print(" ".join(map(def(x):
    if x % 15: return "FooBar"
    if x % 5: return "Bar"
    if x % 3: return "Foo"
    return str(x)
  , range(20)
)))

Forcing dedent seems like the best way while avoiding brackets, which...:

print(" ".join(map((def(x):
  if x % 15: return "FooBar"
  if x % 5: return "Bar"
  if x % 3: return "Foo"
  return str(x)),
  range(20)
)))

(Should this be indented in this case, or is this valid:

print(" ".join(map((def(x):
if x % 15: return "FooBar"
if x % 5: return "Bar"
if x % 3: return "Foo"
return str(x)), range(20))))

Or this:

print(" ".join(map((def(x):if x % 15: return "FooBar"
if x % 5: return "Bar"
if x % 3: return "Foo"
return str(x)), range(20))))
``` IMHO this should not be valid)

First way seems the best way, maybe PEP-8 would be:
```py
print(" ".join(map(
  def(x):
    if x % 15: return "FooBar"
    if x % 5: return "Bar"
    if x % 3: return "Foo"
    return str(x)
  , range(20)
)))

But there's this hanging coma...

One line functions

f(def(x): x += 1; return x**2
, def(y): return y[-1])
f((def(x): x += 1; return x**2), def(y): return y[-1])

Annotations??

a = def(val: str, *args, *kw) -> bool: ...

Are there any cases that I failed to consider? When ambiguity could arise?

def callback(fun: def(val: str) -> bool = def(val): return val != ""):
  callbacks.append(fun)

Annotations

[austere mauve]

@def(func):
  return func
def test(): pass

a: def(): return "How to make assignment to this variable?"
= "I hope not like this"

b: (def(): return "This is bettter") = "for sure"

# This should be disallowed
def a(): b = 1; c = def():
  return False

# Same with
fun = def(): def(x):
  tot = 0
  for v in x:
    tot += v
  return tot

# But this fine:
def a():
  b = 1
  c = def(): return False

# And:
fun = def():
  def(x):
    tot = 0
    for v in x:
      tot += v
    return tot

# This is not:
b(def(x): if x == 0:
  return 1
return x)

# Neither is this (slightly more correct:
b(def(x): if x == 0:
    return 1
  return x)

# It should be done as (the rule that on-eline statements cannot contain multiline statements:
b(def(x):
  if x == 0:
    return 1
  return x)

If you want a task of any form: find ambiguities / cases where this break.

[molten herald]

!e

def mult(numbers):
    return math.prod(numbers)

list_of_numbers = [100,198209,181031]
mult(list_of_numbers)```

[night quarry]

:warning: Your 3.14 free threaded eval job has completed with return code 0.

[No output]

[molten herald]

!e

def mult(numbers):
    return math.prod(numbers)

list_of_numbers = [100,198209,181031]
print(mult(list_of_numbers))```

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

3588197347900

[shy willow]

"bettter"

[austere mauve]

explain

Ah, ok

Don't concern yourself with that

[terse pollen]

Obligatory cold water:
This has been tried and proposed on many separate occasions by many very smart people across many years. The fundamental objections haven't really been addressed. Simply defining something that can be unambiguously parsed, is not going to satisfy the core Python team.
Lambdas have the restrictions they do for a reason (although you can still write effectively everything that way if you have sufficiently little care for writing clear code). It's part of Python's recognition of a strong boundary between statements and expressions.

[austere mauve]

I'm well aware that it breaks a well-established boundary between statements and expressions and I'm not doing it to propose this as a change to Python. I see that it's hard to design it in a way that would result in clean, understandable code. You mentioned earlier attempts - that sounds interesting, can you give some links?

[rigid dune]

Yeah

[molten herald]

!e

def mult(numbers):
       return math.prod(numbers)

x = [10000, 18289]
y= mult(x)
print(f"{y:,}")```

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

182,890,000

[long fulcrum]

I spent so long looking for the golf for gray code encoder. I tried everything and then it turns out it was just two completely unnecessary brackets

I'm only 1 off the J gold I'm gonna try find that next

[north wraith]

What is that Code Golf problem?

[long fulcrum]

https://code.golf/gray-code-encoder

[magic wraith]

i got 67 ||py from sys import* [print(bin(a^a//2)[2:])for a in map(int,argv[1:])]||

[subtle viper]

oh wait

that's it

[long fulcrum]

just importing sys is shorter

[magic wraith]

ah right

it makes it 65

idk why i decided to use a comprehension, a normal for loop saves one more char

[subtle viper]

||format strings|| might be a good idea as well

[magic wraith]

64 ||py import sys for a in map(int,sys.argv[1:]):print(bin(a^a//2)[2:])|| i doubt ||strings|| help here, i feel like the 62 letter solutions ||somehow skip the sys.argv[0] which is a "-"||

[subtle viper]

||strings|| don't, but ||format strings|| do

[dry gulch]

i've been doing some programming art
on the TI 84 Plus CE Python

https://pastebin.com/AqYV7Q43

apologies for the terrible camera setup

[arctic skiff]

reminds me of a guy who did ray tracing on ti calculator

https://youtu.be/todarS6XTPc
https://youtu.be/rY413t5fArw

[dry gulch]

Holy crap!!!!

[arctic skiff]

he also ran super mario bros and minecraft

[dry gulch]

I did see the Minecraft one, cool stuff

[night quarry]

Sorry, an unexpected error occurred. Please let us know!

HTTPException: 429 Too Many Requests (error code: 40062): Service resource is being rate limited.

[vast wave]

xd

[smoky leaf]

i made this a while ago, how can i improve it

[night quarry]

Click here to see this code in our pastebin.

[smoky leaf]

#bot-commands message

[wary python]

there's no way this works

why does this work

[smoky leaf]

yes

[arctic skiff]

char: chr chr is not a type chr is a function and it returns string, so its str anyways

[smoky leaf]

oh

[smoky leaf]

new version at https://github.com/rock3tsprocket/hello-world.py/blob/main/hello.py

[mystic vapor]

you forgot to inherit from object

[smoky leaf]

inherit from what

i don't like OOP

[mystic vapor]

class Main(object):
  ...

[smoky leaf]

added that

[terse pollen]

this is not required and has no effect in 3.x.

[mystic vapor]

ik...

[wary python]

oh i didnt see ___PosixProcessStartup yeah that makes sense now

[smoky leaf]

exactly

[restive void]

!warn 1504136579709407374 Don't advertise

[night quarry]

:incoming_envelope: :ok_hand: applied warning to @urban fractal.

[red wren]

Hello

[shy willow]

Badabun

[mystic vapor]

hehehe

[warm crystal]

ʤ,ʣ=1.79175946923,1.94591014906;print(str(round(2.718281828 ** ʤ))+str(round(2.718281828 ** ʣ)))

!e

ʤ,ʣ=1.79175946923,1.94591014906;print(str(round(2.718281828 ** ʤ))+str(round(2.718281828 ** ʣ)))

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

67

[stable hornet]

!e ```py
import cmath;exit(abs(3-(2cmath.exp((1j11*cmath.pi)/6))**6))

[night quarry]

:x: Your 3.14 eval job has completed with return code 1.

67.0

[twilit meteor]

!e ```py
import cmath;exit(int(abs(3-(2cmath.exp((1j11*cmath.pi)/6))**6)))

[night quarry]

:warning: Your 3.14 eval job has completed with return code 67.

[No output]

[forest wasp]

!e

binary_search=lambda a,t,l=0,r=-2:-1 if l>(r:=(r,len(a)-1)[r<-1])else m if t==a[m:=l+r>>1]else binary_search(a,t,*[(l,m-1),(m+1,r)][t>a[m]])

print(binary_search([1,63,400,4001,5001,5002],4001))

[night quarry]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

3

[subtle viper]

!e ```py
from bisect import*;binary_search=bisect_left

print(binary_search([1,63,400,4001,5001,5002],4001))

[unique shoal]

p = lambda s: bool(len(s) < 2 or s[0] == s[-1] and p(s[1:-1]))

[buoyant agate]

⛳ 🏌

[heady meteor]

🐍

[brisk zenith]

yay!

[wooden mirage]

when you have no nitro and no golfing emojis to use

[dense spire]

🐐 lf

[brisk zenith]

im honestly surprised that this idea got past lemon

[sick hound]

we need some challenges in here

[dense spire]

I wasn't even opposed to it.

[brisk zenith]

oh damn, nice :D

[dense spire]

don't you remember my code jam qualifier code? I like obfuscation

I just hate golf

[lusty vector]

I have no idea what that code does but I think I know the answer is either 0 or 1 😛

[buoyant agate]

I'm pretty sure it recursively checks if a string is a palindrome

[unique shoal]

:+1:

[slim pecan]

Ding ding ding, we have a winner

[viral hedge]

The or is the exit condition, right?