esoteric-python

[karmic pumice]

f=orig
g=my_current_version

f([3],[<class '__main__.'>])='D'
g([3],[<class '__main__.'>])='D'

wtf
i gave it type('',(),{})

[grave grail]

It limit to certain data type otherwise it will fail I think

Or maybe first one doesn't matter 🤔

I am surprised that it work with anything

or nvm

yeah, it would work

since i==0

so it just completely bypass it if it just have 1 thing

[karmic pumice]

i mean..

f([4],[[], [1], [2], {}, {1}, <function <lambda> at 0x000001607022EA20>])='G'
g([4],[[], [1], [2], {}, {1}, <function <lambda> at 0x000001607022EA20>])='G'

im not sure how to get the ,,other" keypad characters from the 2nd argument

[grave grail]

!e
print(list(zip([1],[1,2,3])))

[night quarry]

@grave grail :white_check_mark: Your 3.12 eval job has completed with return code 0.

[(1, 1)]

[grave grail]

yeah, since x only have 1 element

so zip will only have length of 1

so also bypass check

it should crash if you do
f([1,1], [[], [1]])
ig

[karmic pumice]

yep

[grave grail]

And python output the most confusing error message I have ever seen 🤣

[karmic pumice]

def g(x, y):
    table = {
        0: [" "],
        2: ["A", "B", "C"],
        3: ["D", "E", "F"],
        4: ["G", "H", "I"],
        5: ["J", "K", "L"],
        6: ["M", "N", "O"],
        7: ["P", "Q", "R", "S"],
        8: ["T", "U", "V"],
        9: ["W", "X", "Y", "Z"],
    }
    d = [
        table[d[1]][
            (
                [
                    (i, b[0])
                    for i, b in enumerate(zip(x, y))
                    if i == 0 or b[0] != x[i - 1] or b[1] > y[i - 1] + 0.5
                ]
                + [(len(x),)]
            )[i + 1][0]
            - d[0]
            - 1
        ]
        for i, d in enumerate(
            [
                (i, b[0])
                for i, b in enumerate(zip(x, y))
                if i == 0 or b[0] != x[i - 1] or b[1] > y[i - 1] + 0.5
            ]
        )
    ]
    return "".join(d)

funny formatting doesnt help me understand it yet

[grave grail]

🤣

You shouldn't really use d outside when d is in the list comprehension

Even tho they most likely won't disturb each other

And I successfully make 2 people in another server hate one-linear (possibly)

[karmic pumice]

def g(x, y):
    table = {
        0: [" "],
        2: ["A", "B", "C"],
        3: ["D", "E", "F"],
        4: ["G", "H", "I"],
        5: ["J", "K", "L"],
        6: ["M", "N", "O"],
        7: ["P", "Q", "R", "S"],
        8: ["T", "U", "V"],
        9: ["W", "X", "Y", "Z"],
    }
    thing = [
        table[b_item[1]][
            (
                [
                    (c_idx, c_item[0])
                    for c_idx, c_item in enumerate(zip(x, y))
                    if c_idx == 0 or c_item[0] != x[c_idx - 1] or c_item[1] > y[c_idx - 1] + 0.5 # same thing?
                ]
                + [(len(x),)]
            )[b_idx + 1][0]
            - b_item[0]
            - 1
        ]
        for b_idx, b_item in enumerate(
            [
                (i_idx, i_item[0])
                for i_idx, i_item in enumerate(zip(x, y))
                if i_idx == 0 or i_item[0] != x[i_idx - 1] or i_item[1] > y[i_idx - 1] + 0.5 # same thing?
            ]
        )
    ]
    return "".join(thing)

im currently tryingg to understand that condition

[grave grail]

Ok lol

And for the comment, yes, they are same thing

But because I need to make it fully one-linear without semi-colon, so I cannot use a variable to store the result

So I put those things inside

[karmic pumice]

and no way for walrus yeah

def g(x, y):
    table = {
        0: [" "],
        2: ["A", "B", "C"],
        3: ["D", "E", "F"],
        4: ["G", "H", "I"],
        5: ["J", "K", "L"],
        6: ["M", "N", "O"],
        7: ["P", "Q", "R", "S"],
        8: ["T", "U", "V"],
        9: ["W", "X", "Y", "Z"],
    }
    something = [
        (idx, pair[0])
        for idx, pair in enumerate(zip(x, y))
        if idx == 0
        or pair[0] != x[idx - 1]
        or pair[1] > y[idx - 1] + 0.5
    ]
    thing = [
        table[s_item[1]][(something + [(len(x),)])[s_idx + 1][0] - s_item[0] - 1]
        for s_idx, s_item in enumerate(something)
    ]
    return "".join(thing)

okay thats great

[grave grail]

And now, you quite fully decompressed to when I code it

[karmic pumice]

ok right nice so i take idx and the first item in pair guaranteed for the first pair and then if the first item is.. not equal to the item in uhh, the first item in the previous pair or the 2nd item is higher than the (second item in previous pair + 0.5) [well, not really previous, negative indexes and stuff]

[grave grail]

lol

You are quite close I think

[karmic pumice]

f([2, 2, 2],[3, 2, 3])='BA'
g([2, 2, 2],[3, 2, 3])='BA'

thats interesting i finally got new letter

[grave grail]

lol

So now, you need to use this to create an output of "HAPPY BIRTHDAY"

[karmic pumice]

hmmh

phone keypad thinking mindset activated

i need to become the keypad

[grave grail]

🤣

[karmic pumice]

alright so x will definitely be [4, 2, 7, 7, 9, 0, 2, 4, 7, 8, 4, 7, 3, 2, 9]

now on to the second uhh magic thing

[grave grail]

Nope

It is more than just this

[karmic pumice]

hmmh

[grave grail]

Read this on how 3 element on x only result in 2 character

[karmic pumice]

yeah because the conditional

[grave grail]

So you have to consider that as well

[karmic pumice]

the only thing i am sure in right now is that x[0] needs to be 4 lmao

[grave grail]

Yes, that true 🤣

[quasi island]

i need help

menu={
"Baja Taco": 4.00,
"Burrito": 7.50,
"Bowl": 8.50,
"Nachos": 11.00,
"Quesadilla": 8.50,
"Super Burrito": 8.50,
"Super Quesadilla": 9.50,
"Taco": 3.00,
"Tortilla Salad": 8.00
}

while True:
try:
item = input("Item: ")
if item in menu:
order=0.00
order=float(order)+float(menu[item])
else:
break
except EOFError:

break

if item in menu:
print("Total: $" + str(order))
else:
print("")

[grave grail]

What issue

!code

[night quarry]

Formatting code on discord

Here's how to format Python code on Discord:

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

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

For long code samples, you can use our pastebin.

[quasi island]

im getting a

[karmic pumice]

why #esoteric-python ?

[quasi island]

menu={
"Baja Taco": 4.00,
"Burrito": 7.50,
"Bowl": 8.50,
"Nachos": 11.00,
"Quesadilla": 8.50,
"Super Burrito": 8.50,
"Super Quesadilla": 9.50,
"Taco": 3.00,
"Tortilla Salad": 8.00
}

while True:
try:
item = input("Item: ")
if item in menu:
order=0.00
order=float(order)+float(menu[item])
else:
break
except EOFError:

break

if item in menu:
print("Total: $" + str(order))
else:
print("")

idk what other server

[karmic pumice]

just #python-discussion or #1035199133436354600

[quasi island]

oh ok thx

[karmic pumice]

this place is for cursed code

[grave grail]

Yeah lol

[quasi island]

phhhh

ohhhh

thx

[grave grail]

Like this #esoteric-python message

[karmic pumice]

def g(x, y):
    table = {
        0: [" "],
        # 1 missing because idk
        2: ["A", "B", "C"],
        3: ["D", "E", "F"],
        4: ["G", "H", "I"],
        5: ["J", "K", "L"],
        6: ["M", "N", "O"],
        7: ["P", "Q", "R", "S"],
        8: ["T", "U", "V"],
        9: ["W", "X", "Y", "Z"],
    }
    something = [
        (idx, pair[0])
        for idx, pair in enumerate(zip(x, y))
        if idx == 0 
        or pair[0] != x[idx - 1] 
        or pair[1] > y[idx - 1] + 0.5
    ]
    thing = [
        table[s_item[1]][
            (
                keyidx := 
                (something + [(len(x),)])[s_idx + 1][0] 
                - (s_item[0] + 1) # very interesting expression
            )
        ]
        for s_idx, s_item in enumerate(something)
    ]
    return "".join(thing)

hmmh

[grave grail]

You can print something out from the code which might help 🤔

[karmic pumice]

yeah probally

[grave grail]

Since you have decompressed it

[karmic pumice]

the len(x) when calculating keyidx is confusing me very much

[grave grail]

Oh lol

It was to provide the length of x to calculate something

[quartz wave]

this is oddly familiar to a nokia or any old mobile phone typing system

[karmic pumice]

yeah it is a keypad

but the input output correlation is very weird

currently i have

p(
    [4, 4, 2, 7, 7, 9],
    [2, 2, 2, 7, 8, 7],
)

which gives me

something=[(0, 4), (2, 2), (3, 7), (4, 7), (5, 9)]
s_idx=0 s_item=(0, 4) part=(2, 2) -> keyidx=1 which is ((2, 2)[0] -> 2) - (((0, 4)[0] -> 0)+1)
s_idx=1 s_item=(2, 2) part=(3, 7) -> keyidx=0 which is ((3, 7)[0] -> 3) - (((2, 2)[0] -> 2)+1)
s_idx=2 s_item=(3, 7) part=(4, 7) -> keyidx=0 which is ((4, 7)[0] -> 4) - (((3, 7)[0] -> 3)+1)
s_idx=3 s_item=(4, 7) part=(5, 9) -> keyidx=0 which is ((5, 9)[0] -> 5) - (((4, 7)[0] -> 4)+1)
s_idx=4 s_item=(5, 9) part=(6,) -> keyidx=0 which is ((6,)[0] -> 6) - (((5, 9)[0] -> 5)+1)
thing=['H', 'A', 'P', 'P', 'W']
f([4, 4, 2, 7, 7, 9],[2, 2, 2, 7, 8, 7])='HAPPW'
g([4, 4, 2, 7, 7, 9],[2, 2, 2, 7, 8, 7])='HAPPW'

i dont understand how to uhh get a different letter

[grave grail]

Ok, do you know what is the thing on y

[karmic pumice]

kinda? i need to increase keyidx to 2

[grave grail]

Try to describe it

[karmic pumice]

like how

[grave grail]

Ok, I don't think you know what y is then 🤣

[karmic pumice]

you like select the 9th keypad button and get the 3rd character

[grave grail]

Ok, y accept a list of float data

[karmic pumice]

ah you meant the y argument

[grave grail]

Now try something

[karmic pumice]

not the letter y

[grave grail]

Yeah

[karmic pumice]

i actually got y right now

p(
    [4, 4, 2, 7, 7, 9, 9, 9],
    [2, 2, 2, 7, 8, 9, 9, 9],
)

gives me

thing=['H', 'A', 'P', 'P', 'Y']

i think i understand it

[grave grail]

Probably 🤔

[quartz wave]

>>> g([4,4,2,7,7,9,9,9],[1,1,1,1,2,1])
'HAPPY'

huh

[karmic pumice]

yeah i do

p(
    [4, 4, 2, 7, 7, 9, 9, 9, 0, 2, 2, 4, 4, 4, 7, 7, 7, 8, 4, 4, 7, 7, 7, 7, 3, 2, 9, 9, 9],
    [2, 2, 2, 7, 8, 9, 9, 9, 0, 2, 2, 4, 4, 4, 7, 7, 7, 8, 4, 4, 7, 7, 7, 7, 3, 2, 9, 9, 9],
)

thing=['H', 'A', 'P', 'P', 'Y', ' ', 'B', 'I', 'R', 'T', 'H', 'S', 'D', 'A', 'Y']

this is genius

[grave grail]

No S

[karmic pumice]

uh right my english is funny

p(
    [4, 4, 2, 7, 7, 9, 9, 9, 0, 2, 2, 4, 4, 4, 7, 7, 7, 8, 4, 4, 3, 2, 9, 9, 9],
    [2, 2, 2, 7, 8, 9, 9, 9, 0, 2, 2, 4, 4, 4, 7, 7, 7, 8, 4, 4, 3, 2, 9, 9, 9],
)

thing=['H', 'A', 'P', 'P', 'Y', ' ', 'B', 'I', 'R', 'T', 'H', 'D', 'A', 'Y']

[grave grail]

By follow this rule:

• The first element of y must be 0
• Each element in y must at least increment by 0.1 than the last value

What is the minimum value for the last element of y to get "HAPPY BIRTHDAY"

And also, all value in y must be 1 decimal point

Like you cannot have 0.91

Only 0.9 or 1

[karmic pumice]

i see

[quartz wave]

i just put a bunch of zeroes

[quartz wave]

0

unless negatives are allowed

[karmic pumice]

must at least increnment by 0.1 and start with 0

[quartz wave]

oh

[grave grail]

lol

[quartz wave]

well there's only one that needs a 0.6 increment so

[grave grail]

So, putting all 0 probably won't work either

[quartz wave]

2.7

[grave grail]

Type the whole answer so it would be easy to verify

[quartz wave]

>>> g([4,4,2,7,7,9,9,9,0,2,2,4,4,4,7,7,7,8,4,4,3,2,9,9,9],[0,0.1,0.2,0.3,0.9,1,1.1,1.2,1.3,1.4,1.5,1.6,1.7,1.8,1.9,2,2.1,2.2,2.3,2.4,2.5,2.6,2.7])
'HAPPY BIRTHDAY'

idk

[karmic pumice]

yeah im basically doing the same thing rn

[grave grail]

gg

You guys solve the question

I will just throw the question to people that barely learn python 🤣

And they would go crazy

[quartz wave]

i feel like the >0.5 increment was just to distinguish the consecutive characters

[grave grail]

Yeah

[quartz wave]

i think i understandn it now :D

[karmic pumice]

im trying to represent the keypresses funnier with the data

p(
    [*(4,)*2 , 2  , *(7,)*2    , *(9,)*3         , 0  , *(2,)*2    , *(4,)*3         , *(7,)*3       , 8  , *(4,)*2   , 3  , 2  , *(9,)*3],
    [*(0,0.1), 0.2, *(0.3, 0.9), *(1.0, 1.1, 1.2), 1.3, *(1.4, 1.5), *(1.6, 1.7, 1.8), *(1.9,2.0,2.1), 2.2, *(2.3,2.4), 2.5, 2.6, 2.7    ],
)

like that

[grave grail]

lol

[karmic pumice]

this was really interesting problem thanks

[grave grail]

lol

https://discord.com/channels/267624335836053506/1163499849434529843

You can refer to this

the origin of the problem

[quartz wave]

def keypad(inp, dist):
    keypad = {
        0: [" "],
        # there's no ASCII characters for 1 on a keypad
        2: ["A", "B", "C"],
        3: ["D", "E", "F"],
        4: ["G", "H", "I"],
        5: ["J", "K", "L"],
        6: ["M", "N", "O"],
        7: ["P", "Q", "R", "S"],
        8: ["T", "U", "V"],
        9: ["W", "X", "Y", "Z"],
    }
    idx_key_pair = [
        (idx, key)
        for idx, (key, dist_from_prev) in enumerate(zip(inp, dist))
                                                # add if:
        if idx == 0                             # - just started
        or key != inp[idx - 1]                  # - key is different
        or dist_from_prev > dist[idx - 1] + 0.5 # - distance from previous key is > 0.5
    ]
    chars = [
        keypad[key][
            # get the character based on the amount of keypresses
            # between the next one (or the length of the input)
            # like an old keypad
            idx_key_pair[s_idx + 1][0] - idx
            if s_idx + 1 < len(inp) else
            len(inp) - idx
        ]
        for s_idx, (idx, key) in enumerate(idx_key_pair)
    ]
    return "".join(chars)

de-esotericized, commented

[karmic pumice]

teamwork

[quartz wave]

now i'm gonna go golf it later

[karmic pumice]

lmao

degolf -> golf combo

[quartz wave]

i have school first

[karmic pumice]

i hab school in 7 hrs too
i might need to go sleep

[grave grail]

lol

I have school in 9 hours

[grave grail]

more golf

lambda x,y: ''.join([{0: [' '],2:['A','B','C'],3:['D','E','F'],4:['G','H','I'],5:['J','K','L'],6:['M','N','O'],7:['P','Q','R','S'],8:['T','U','V'],9:['W','X','Y','Z']}[d[1]][d[2][i+1]-d[0]-1]for i,d in enumerate([(i,b[0],[i for i,b in enumerate(zip(x,y))if i==0 or b[0]!=x[i-1] or b[1]>y[i-1]+0.5]+[len(x)])for i,b in enumerate(zip(x,y))if i==0 or b[0]!=x[i-1] or b[1]>y[i-1]+0.5])])

[grave grail]

and more

lambda I,L: ''.join([{0:[chr(0x20)],2:[chr(I) for I in [0o101,0x42,0b1000011]],3:[chr(i+3) for i in [0o101,0x42,0b1000011]],4:[chr(L+6) for L in [0o101,0x42,0b1000011]],5:[chr(I+9) for I in [0o101,0x42,0b1000011]],6:[chr(l+12) for l in [0o101,0x42,0b1000011]],7:[chr(l+15) for l in [0o101,0x42,0b1000011,0o104]],8:[chr(L+19) for L in [0o101,0x42,0b1000011]],9:[chr(i+22) for i in [0o101,0x42,0b1000011,0b1000100]]}[l[1]][l[2][i]-l[0]-1]for i,l in enumerate([(i,I[i],[i for i in range(1,min(len(I),len(L)))if (I[i]!=I[i-1])|(L[i]>L[i-1]+0.5)]+[len(I)])for i in range(len(list(zip(I,L))))if (i==0)|(I[i]!=I[i - 1])|(L[i]>L[i-1]+0.5)])])

Using confusing variable name, repeated variable name in list comprehension, reduce duplicate part in real code, remove or key word, obfuscate dictionary with list comprehensive with integer prefix

This probably didn't really increase much difficulty

But annoy people

[karmic pumice]

oh nice

lmao the octals [is this how base 8 is called?]

[grave grail]

And it only consist of 4 variable: i l I L

[karmic pumice]

haha

[grave grail]

Yeah, at least what I know

[karmic pumice]

you know what you could do

[grave grail]

Oh wait, I can also obfuscate the dictionary key

[karmic pumice]

!e

print(~-0)

this instead of x-1

[night quarry]

@karmic pumice :white_check_mark: Your 3.12 eval job has completed with return code 0.

-1

[grave grail]

Tf

How it work?

[karmic pumice]

~x (for ints) = -(x+1) #****

[grave grail]

Ok

[karmic pumice]

so ~-2 is -(-2+1) -> 2-1 -> 1

[grave grail]

!e
print(~-10)

[night quarry]

@grave grail :white_check_mark: Your 3.12 eval job has completed with return code 0.

9

[grave grail]

ok nices

[karmic pumice]

instead of l[1] you could do l[-~int()]

[grave grail]

!e
print(~-0b1010)

[night quarry]

@grave grail :white_check_mark: Your 3.12 eval job has completed with return code 0.

9

[grave grail]

ok, great

[karmic pumice]

and thing[0] -> thing[int()] maybe
or thing[type(0)()]
or thing[not __name__=="__main__"]

[grave grail]

ok, need to learn more about esoteric lol

[karmic pumice]

!e

print(~-len([('x',['d',('f',['u',('n',['n',('y',[([()])])])])])]))

[night quarry]

@karmic pumice :white_check_mark: Your 3.12 eval job has completed with return code 0.

0

[karmic pumice]

you can start expressing numbers as peano numbers i guess with successors

!e

s=lambda x=0:x+1
print(s(s(s(s()))))

[fleet bridge]

!e My favourite way to get number 3: ```py
import _testcapi
print(_testcapi.the_number_three)

oops

:(

[night quarry]

@karmic pumice :white_check_mark: Your 3.12 eval job has completed with return code 0.

4

[fleet bridge]

>>> import _testcapi
>>> print(_testcapi.the_number_three)
3

[karmic pumice]

lmao

does it have other the numbers

In [27]: print([t for t in dir(_testcapi) if t.startswith('the_number')])
['the_number_three']

😭

[fleet bridge]

the 3 is special enough

[grave grail]

lambda I,L: str().join([{~-1:[chr(0x20)],~-0b11:[chr(I) for I in [0o101,0x42,0b1000011]],(~-0+0xA)//0b11:[chr(i+3) for i in [0o101,0x42,0b1000011]],0b100:[chr(L+6) for L in [0o101,0x42,0b1000011]],0xA//~-0b11:[chr(I+9) for I in [0o101,0x42,0b1000011]],(~-4)*(~-3):[chr(l+12) for l in [0o101,0x42,0b1000011]],0b111:[chr(l+15) for l in [0o101,0x42,0b1000011,0o104]],0o10:[chr(L+~-20) for L in [0o101,0x42,0b1000011]],(~-0b100)**2:[chr(i+ord('\x16')) for i in [0o101,0x42,0b1000011,0b1000100]]}[l[~-0b10]][~-0+l[0b10][i]-l[int()]]for i,l in enumerate([(i,I[i],[i for i in range(1,min(len(I),len(L)))if (I[i]!=I[i-1])|(L[i]>L[i-1]+(~-0b10)/0b10)]+[len(I)])for i in range(len(list(zip(I,L))))if (i==int())|(I[i]!=I[~-int()+i])|(L[i]>L[~-(not __name__=="__main__")+i]+(~-0b10)/0b10)])])

Discord syntax highlight is hating me lol

[fleet bridge]

its pretty good on PC and awful on mobile

[grave grail]

Yeah

And also I am making the code slower and slower

Because it have to calculate all those stuff

[karmic pumice]

it doesnt

the stuff thats constant is constant, it doesnt get recalculated

[grave grail]

Mhh really? I thought it gonna recalculate the dictionary on each iteration

[karmic pumice]

well it does build the dict each time, but the stuff from which it builds it is constant

[grave grail]

Oh ok

[karmic pumice]

i think if you care enough you could write the table as an argument with a default value

lambda I,L,t={}:

but then write it normally probally 🤷 kinda incompatible ideas

[grave grail]

So if I make it access a variable from the iteration to do a pointless calculation, it have to recalculate every time?

Like +i-i

[versed eagle]

no

lists & list comps aren't constant

[karmic pumice]

that part - sure, but im talking about the added integer calculation stuff
~-1, ~-0b11

[versed eagle]

oh

yeah those will be constants

[grave grail]

Oh nah, I was talking about the entirety of the stupidly long dictionary

[versed eagle]

an easy way to check is to use dis

[karmic pumice]

all of that stuff is calculated at bytecode compile time btw, it doesnt get calculated even at first call

!e

f=lambda I,L: str().join([{~-1:[chr(0x20)],~-0b11:[chr(I) for I in [0o101,0x42,0b1000011]],(~-0+0xA)//0b11:[chr(i+3) for i in [0o101,0x42,0b1000011]],0b100:[chr(L+6) for L in [0o101,0x42,0b1000011]],0xA//~-0b11:[chr(I+9) for I in [0o101,0x42,0b1000011]],(~-4)*(~-3):[chr(l+12) for l in [0o101,0x42,0b1000011]],0b111:[chr(l+15) for l in [0o101,0x42,0b1000011,0o104]],0o10:[chr(L+~-20) for L in [0o101,0x42,0b1000011]],(~-0b100)**2:[chr(i+ord('\x16')) for i in [0o101,0x42,0b1000011,0b1000100]]}[l[~-0b10]][~-0+l[0b10][i]-l[int()]]for i,l in enumerate([(i,I[i],[i for i in range(1,min(len(I),len(L)))if (I[i]!=I[i-1])|(L[i]>L[i-1]+(~-0b10)/0b10)]+[len(I)])for i in range(len(list(zip(I,L))))if (i==int())|(I[i]!=I[~-int()+i])|(L[i]>L[~-(not __name__=="__main__")+i]+(~-0b10)/0b10)])])
import dis
dis.dis(f)

[night quarry]

@karmic pumice :white_check_mark: Your 3.12 eval job has completed with return code 0.

001 |   1           0 RESUME                   0
002 |               2 LOAD_GLOBAL              1 (NULL + str)
003 |              12 CALL                     0
004 |              20 LOAD_ATTR                3 (NULL|self + join)
005 |              40 LOAD_GLOBAL              5 (NULL + enumerate)
006 |              50 LOAD_GLOBAL              7 (NULL + range)
007 |              60 LOAD_GLOBAL              9 (NULL + len)
008 |              70 LOAD_GLOBAL             11 (NULL + list)
009 |              80 LOAD_GLOBAL             13 (NULL + zip)
010 |              90 LOAD_FAST                0 (I)
011 |              92 LOAD_FAST                1 (L)
... (truncated - too many lines)

Full output: https://paste.pythondiscord.com/SP2B2X5VS35S4EPZJB567NPBPU

[versed eagle]

!e ```py
f=lambda I,L: str().join([{~-1:[chr(0x20)],~-0b11:[chr(I) for I in [0o101,0x42,0b1000011]],(~-0+0xA)//0b11:[chr(i+3) for i in [0o101,0x42,0b1000011]],0b100:[chr(L+6) for L in [0o101,0x42,0b1000011]],0xA//~-0b11:[chr(I+9) for I in [0o101,0x42,0b1000011]],(~-4)*(~-3):[chr(l+12) for l in [0o101,0x42,0b1000011]],0b111:[chr(l+15) for l in [0o101,0x42,0b1000011,0o104]],0o10:[chr(L+~-20) for L in [0o101,0x42,0b1000011]],(~-0b100)**2:[chr(i+ord('\x16')) for i in [0o101,0x42,0b1000011,0b1000100]]}[l[~-0b10]][~-0+l[0b10][i]-l[int()]]for i,l in enumerate([(i,I[i],[i for i in range(1,min(len(I),len(L)))if (I[i]!=I[i-1])|(L[i]>L[i-1]+(~-0b10)/0b10)]+[len(I)])for i in range(len(list(zip(I,L))))if (i==int())|(I[i]!=I[~-int()+i])|(L[i]>L[~-(not name=="main")+i]+(~-0b10)/0b10)])])
import dis
dis.dis(f)

[night quarry]

@versed eagle :white_check_mark: Your 3.12 eval job has completed with return code 0.

001 |   1           0 RESUME                   0
002 |               2 LOAD_GLOBAL              1 (NULL + str)
003 |              12 CALL                     0
004 |              20 LOAD_ATTR                3 (NULL|self + join)
005 |              40 LOAD_GLOBAL              5 (NULL + enumerate)
006 |              50 LOAD_GLOBAL              7 (NULL + range)
007 |              60 LOAD_GLOBAL              9 (NULL + len)
008 |              70 LOAD_GLOBAL             11 (NULL + list)
009 |              80 LOAD_GLOBAL             13 (NULL + zip)
010 |              90 LOAD_FAST                0 (I)
011 |              92 LOAD_FAST                1 (L)
... (truncated - too many lines)

Full output: https://paste.pythondiscord.com/UBVHGIDPADRZYU53LXMGSSCORI

[karmic pumice]

lmao

[grave grail]

!e

import dis
a = lambda I,L: str().join([{~-1:[chr(0x20)],~-0b11:[chr(I) for I in [0o101,0x42,0b1000011]],(~-0+0xA)//0b11:[chr(i+3) for i in [0o101,0x42,0b1000011]],0b100:[chr(L+6) for L in [0o101,0x42,0b1000011]],0xA//~-0b11:[chr(I+9) for I in [0o101,0x42,0b1000011]],(~-4)*(~-3):[chr(l+12) for l in [0o101,0x42,0b1000011]],0b111:[chr(l+15) for l in [0o101,0x42,0b1000011,0o104]],0o10:[chr(L+~-20) for L in [0o101,0x42,0b1000011]],(~-0b100)**2:[chr(i+ord('\x16')) for i in [0o101,0x42,0b1000011,0b1000100]]}[l[~-0b10]][~-0+l[0b10][i]-l[int()]]for i,l in enumerate([(i,I[i],[i for i in range(1,min(len(I),len(L)))if (I[i]!=I[i-1])|(L[i]>L[i-1]+(~-0b10)/0b10)]+[len(I)])for i in range(len(list(zip(I,L))))if (i==int())|(I[i]!=I[~-int()+i])|(L[i]>L[~-(not __name__=="__main__")+i]+(~-0b10)/0b10)])])

dis.dis("a([4,4,2,7,7,9,9,9,0,2,2,4,4,4,7,7,7,8,4,4,3,2,9,9,9],[0,0.1,0.2,0.3,0.9,1,1.1,1.2,1.3,1.4,1.5,1.6,1.7,1.8,1.9,2,2.1,2.2,2.3,2.4,2.5,2.6,2.7])")

[night quarry]

@grave grail :white_check_mark: Your 3.12 eval job has completed with return code 0.

001 |   0           0 RESUME                   0
002 | 
003 |   1           2 PUSH_NULL
004 |               4 LOAD_NAME                0 (a)
005 |               6 BUILD_LIST               0
006 |               8 LOAD_CONST               0 ((4, 4, 2, 7, 7, 9, 9, 9, 0, 2, 2, 4, 4, 4, 7, 7, 7, 8, 4, 4, 3, 2, 9, 9, 9))
007 |              10 LIST_EXTEND              1
008 |              12 BUILD_LIST               0
009 |              14 LOAD_CONST               1 ((0, 0.1, 0.2, 0.3, 0.9, 1, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2, 2.1, 2.2, 2.3, 2.4, 2.5, 2.6, 2.7))
010 |              16 LIST_EXTEND              1
011 |              18 CALL                     2
... (truncated - too many lines)

Full output: https://paste.pythondiscord.com/H2V3GFKEKLIUG4OZNJIX2PXYLY

[karmic pumice]

ahaha

[grave grail]

Lmao, it doesn't show lambda

[versed eagle]

wdym

[grave grail]

It just show call

And nothing else

[versed eagle]

yeah

that's cause functions each have their own bytecode

[grave grail]

I thought it going to show what happen inside the function when it run

[versed eagle]

nope

[karmic pumice]

for that you'd disassemble the function itself, not a call to it

[grave grail]

!e
import dis
def a():
print("Hi!")
dis.dis("a()")

[night quarry]

@grave grail :white_check_mark: Your 3.12 eval job has completed with return code 0.

001 |   0           0 RESUME                   0
002 | 
003 |   1           2 PUSH_NULL
004 |               4 LOAD_NAME                0 (a)
005 |               6 CALL                     0
006 |              14 RETURN_VALUE

[karmic pumice]

nope

[versed eagle]

if passed a string, it compiles that string

[grave grail]

Oh

Ok

[versed eagle]

and then disassembles

[karmic pumice]

!e

def f():
  pass # load const None return whatever
  # yeah
import dis
dis.dis(f)

[night quarry]

@karmic pumice :white_check_mark: Your 3.12 eval job has completed with return code 0.

001 |   1           0 RESUME                   0
002 | 
003 |   2           2 RETURN_CONST             0 (None)

[versed eagle]

there's no way for it to actually do that

[karmic pumice]

do dis.dis(a)

[grave grail]

!e

import dis
a = lambda I,L: str().join([{~-1:[chr(0x20)],~-0b11:[chr(I) for I in [0o101,0x42,0b1000011]],(~-0+0xA)//0b11:[chr(i+3) for i in [0o101,0x42,0b1000011]],0b100:[chr(L+6) for L in [0o101,0x42,0b1000011]],0xA//~-0b11:[chr(I+9) for I in [0o101,0x42,0b1000011]],(~-4)*(~-3):[chr(l+12) for l in [0o101,0x42,0b1000011]],0b111:[chr(l+15) for l in [0o101,0x42,0b1000011,0o104]],0o10:[chr(L+~-20) for L in [0o101,0x42,0b1000011]],(~-0b100)**2:[chr(i+ord('\x16')) for i in [0o101,0x42,0b1000011,0b1000100]]}[l[~-0b10]][~-0+l[0b10][i]-l[int()]]for i,l in enumerate([(i,I[i],[i for i in range(1,min(len(I),len(L)))if (I[i]!=I[i-1])|(L[i]>L[i-1]+(~-0b10)/0b10)]+[len(I)])for i in range(len(list(zip(I,L))))if (i==int())|(I[i]!=I[~-int()+i])|(L[i]>L[~-(not __name__=="__main__")+i]+(~-0b10)/0b10)])])

dis.dis(a)

[night quarry]

@grave grail :white_check_mark: Your 3.12 eval job has completed with return code 0.

001 |   2           0 RESUME                   0
002 |               2 LOAD_GLOBAL              1 (NULL + str)
003 |              12 CALL                     0
004 |              20 LOAD_ATTR                3 (NULL|self + join)
005 |              40 LOAD_GLOBAL              5 (NULL + enumerate)
006 |              50 LOAD_GLOBAL              7 (NULL + range)
007 |              60 LOAD_GLOBAL              9 (NULL + len)
008 |              70 LOAD_GLOBAL             11 (NULL + list)
009 |              80 LOAD_GLOBAL             13 (NULL + zip)
010 |              90 LOAD_FAST                0 (I)
011 |              92 LOAD_FAST                1 (L)
... (truncated - too many lines)

Full output: https://paste.pythondiscord.com/ESNSEJG5QASGFNAG2FYRAGGYQA

[versed eagle]

since what the name refers to is unknown until runtime

[versed eagle]

probably you should use your own repl or #bot-commands if you want to do this a lot

[grave grail]

Nah, I was literally gonna go off and sleep

The only reason I run it here is I put my laptop on sleep mode

[lapis kettle]

Can i reassign var without =,@and ()?

[golden finch]

yeah I think so

lemme make an example rq

it should just be the decorator chaining stuff, right?

[lapis kettle]

Im sorry i forgot that is not possible too

[golden finch]

not sure then

[lapis kettle]

re.escape('\\(~}?>{&/%)<$|*=#!-+'0123456789;@, '` it would be like this for the regex

[versed eagle]

!e ```py
import main
for main.variable in ["value"]: ...

print(variable)

[night quarry]

@versed eagle :white_check_mark: Your 3.12 eval job has completed with return code 0.

value

[versed eagle]

:3

though it only does globals

[astral rover]

til you can do attribute access in for loops like that wtf

[versed eagle]

you can also do indexing

[lapis kettle]

How come you get this idea😭

[versed eagle]

you asked for it

[versed eagle]

.

[lapis kettle]

Yeah but i mean is i didnt expect using for loops for this. Youre genius!😭

[golden finch]

that's a smart way to avoid the assignment target being explicitly stated

bravo

Anyhow, what I'm here for is that I'm trying to implement SKI in python

Obviously, Python is eager, so the following won't work:

I = lambda x:x
K = lambda x:lambda y:x
S = lambda x:lambda y:lambda z: x(z)(y(z))

So I thought "hmmm. maybe thunks are a way to implement this"

Specifically, to implement lazy evaluation

At first I did this:

class Thunk:
  def __init__(self, f, arg):
    self.f = f
    self.arg = arg
  def __call__(self): # thunks take no arguments
    return self.f(self.arg)

But that's stupid, because that's just one of the following (which? I think it's the second)

lambda f,arg:lambda:f(arg)
lambda f,arg:lambda f=f,arg=arg:f(arg)

And well, that's where I got stuck

[versed eagle]

what's ski?

[golden finch]

https://en.wikipedia.org/wiki/SKI_combinator_calculus

[versed eagle]

why doesn't this work?

[golden finch]

because python is strictly evaluated

[versed eagle]

why does that mean it won't work

[golden finch]

because lambda calculus (which is what I am trying to implement here) uses a lazy evaluation model, for which Python's strict evaluation (of expressions) isn't suitable

so I need to implement lazy expression evaluation

which I think (?) can be done in a strict-evaluation situation using thunks?

[versed eagle]

do you have an example in which that makes a difference?

[golden finch]

not to hand, my lambda calculus isn't great

but off the top of my head - in strict evaluation, the fixpoint combinator is an infinite loop

in lazy evaluation it isn't

i think

[versed eagle]

is it not?

the recursion isnt bounded by anything

[golden finch]

for a lot of this I'm going off a conversation I had months ago in the esolangs discord where they told me this didn't work

by a very skilled functional programmer

maybe I should head over there and ask them again

[versed eagle]

i think i understand what you're wanting

you're wanting to be able to reduce functions step by step yes?

such that SII(SII) reduces back down to SII(SII) but stops there

[versed eagle]

did i understand right?

[golden finch]

maybe?

the combinators should take thunks (unevaluated expressions...?) and return the same, I presume

[lapis kettle]

Is there any python docs about this?

[golden finch]

for loops have structure

for assignment_target in iterable:
    ...

and they turn that iterable into an iterator

then run assignment_target = next(iterator)

so it's just doing __main__.variable = "value"

[versed eagle]

probably

[dreamy bridge]

can i golf https://open.kattis.com/problems/romanholidays further than this?

r=[(H+T+O).replace(' ','')for H in" .C.CC.CCC.CD.D.DC.DCC.DCCC.CM".split('.')for T in" .X.XX.XXX.XL.L.LX.LXX.LXXX.XC".split('.')for O in" .I.II.III.IV.V.VI.VII.VIII.IX".split('.')]
v=sorted(r)
for n in[int(input())for i in range(int(input()))]:a,b=divmod(n,1000);b=r[b];print([946*a+v.index(b),-54*a-1000+v.index(b)][bool(b and b[0]in'XV')])

i mean yes i have gotten shortest but just posting here for fun

# first 1000 roman numerals
r=[(H+T+O).replace(' ','')for H in" .C.CC.CCC.CD.D.DC.DCC.DCCC.CM".split('.')
    for T in" .X.XX.XXX.XL.L.LX.LXX.LXXX.XC".split('.')
    for O in" .I.II.III.IV.V.VI.VII.VIII.IX".split('.')]
v=sorted(r) # sorted
for n in[int(input())for i in range(int(input()))]: # loop thru input
    a,b=divmod(n,1000) # remove all Ms from the roman numeral for now
    b=r[b] # get the actual roman
    # simple multiplication as
    # M is at index 945, MC at 946 etc, then MM is at 1892 and MMC 1893
    # XXXVIII is at index -1, MXXXVIII -55, MMXXXVIII -109 etc
    if b and b[0]in'XV':print(-54*a-1000+v.index(b)) # return negative index
    else:print(946*a+v.index(b)) # return positive index

heres readable version yes

[elfin ether]

Any ideas on how to implement (hack) an automatically unpacked record type?
Basically I want to be able to do this: (lambda x, y: x + y)((lambda: xy)()) and auto-magically destructure xy similar to how multiple comma separated return values work in Go. I want to avoid decorating the outer function.

[dreamy bridge]

!e

xy = (1,2)
print((lambda x, y: x + y)(*(lambda: xy)()))

? not sure if thats what you want

[night quarry]

@dreamy bridge :white_check_mark: Your 3.12 eval job has completed with return code 0.

3

[elfin ether]

Not exactly. I'm essentially trying to store **kwargs in a variable; this expression throws a syntax error in most circumstances though. I suspect there's no way to do what I want without some absurd hacks.

[dreamy bridge]

yea i've never been able to store a variable with lambda before if thats what you want

[elfin ether]

Maybe this is more illustrative of what I want.

xy = *(2, 3) # this is the black magic i'm trying to replicate
assert (lambda x, y: x + y)(xy) == 5

[grave grail]

!e
print(type(True).real.class.init_subclass().sizeof()*4+1)

[night quarry]

@grave grail :white_check_mark: Your 3.12 eval job has completed with return code 0.

65

[grave grail]

nice

[unique heath]

!e py print(type(True).real.__class__.__init_subclass__().__sizeof__().__mul__(4).__add__(__builtins__.__dir__().__len__().__floordiv__(__name__.__dir__().__len__()))) kind of a lazy extra obf layer but yewah

[night quarry]

@unique heath :white_check_mark: Your 3.12 eval job has completed with return code 0.

65

[quartz wave]

202b ```py
f=lambda x,y,i=0,j=-1:bool(t:=[(j,a)for a,b in zip(x,y)if(j:=j+1)<1 or b>.5+y[j-1]or a-x[j-1]])''.join([" ",0,"ABC DEF GHI JKL MNO PQRS TUV WXYZ".split()][a][[*t,[len(x)]][i:=i+1][0]-j-1]for j,a in t)

-2b can be achieved by not caring about empty input

[last locust]

<1 or b --> <1or b to save an extra character

[quartz wave]

that'll be a SyntaxError sometime later in python's development

[last locust]

But not yet 😉

[quartz wave]

ok then -3b max recklessness

[low lynx]

[int(input())for i in range(int(input()))] could be [int(input())for i in[0]*int(input())] i think
in fact doing a,b=divmod(int(input()),1000) and putting the loop into exec('code;'*int(input())) might be shorter

print([946*a+v.index(b),-54*a-1000+v.index(b)][bool(b and b[0]in'XV')])
print(-54*a-1000+v.index(b)if b and b[0]in'XV'else 946*a+v.index(b))

also

r=[(H+T+O).replace(' ','')for H in" .C.CC.CCC.CD.D.DC.DCC.DCCC.CM".split('.')for T in" .X.XX.XXX.XL.L.LX.LXX.LXXX.XC".split('.')for O in" .I.II.III.IV.V.VI.VII.VIII.IX".split('.')]
r=[H+T+O for H in".C.CC.CCC.CD.D.DC.DCC.DCCC.CM".split('.')for T in".X.XX.XXX.XL.L.LX.LXX.LXXX.XC".split('.')for O in".I.II.III.IV.V.VI.VII.VIII.IX".split('.')]

another char save would be removing b=r[b] and doing if(b:=r[b])and b[0]in'XV' etc

saving a variable to v.index(b) should save a few chars

v.index(b)v.index(b)
w=v.index(b)ww

actually doing w=v.index(b)-54*a and then doing w-1000 and w+1000*a should be better

and turning r into a walrus should save a char as well

[quartz wave]

assign 1000 to some variable

[low lynx]

that doesn't save anything does it?

[quartz wave]

also don't use divmod()

[low lynx]

t=1000;tt
10001000

[quartz wave]

assuming your optimizations are used

[low lynx]

oh right yes

[quartz wave]

i don't feel right about all those repeated C/X/Is

[low lynx]

agreed

but avoiding duplication there is hard

[fleet bridge]

you can remove spaces from long strings and remove .replace(' ','')

[low lynx]

yeah that's my point there

[fleet bridge]

oh, indeed 😄

[dreamy bridge]

woah ok

ill try to do all of that

[versed eagle]

there's a way to do that yeah

[brave schooner]

hi

i'm from VietNam

May everyone help you:v

[grave grail]

?

Do you mean "May anyone help me?"

[grave grail]

apparently __builtins__.__dir__().__len__() is different between location where in interpreter, it is 161 but in code, it is 158

[cloud fossil]

159 on Python 3.12.0 REPL

Never mind, _ gets added to it

158

[grave grail]

more golf

lambda I,L: str().join([{~-1:[chr(__builtins__.__dir__().__len__().__floordiv__((__builtins__.__dir__().__len__().__floordiv__(__name__.__dir__().__len__())<<(__builtins__.__dir__().__len__().__floordiv__(__name__.__dir__().__len__())<<__builtins__.__dir__().__len__().__floordiv__(__name__.__dir__().__len__()))).__add__(__builtins__.__dir__().__len__().__floordiv__(__name__.__dir__().__len__()))).__add__(__builtins__.__dir__().__len__().__floordiv__(__name__.__dir__().__len__())))],~-0b11:[chr(I) for I in [type(True).real.__class__.__init_subclass__().__sizeof__().__mul__(4).__add__(__builtins__.__dir__().__len__().__floordiv__(__name__.__dir__().__len__())),0x42,0b1000011]],(~-0+0xA)//0b11:[chr(i+0b10+__builtins__.__dir__().__len__().__floordiv__(__name__.__dir__().__len__())) for i in [0o101,0x42,0b1000011]],0b100:[chr(L+6) for L in [0o101,0x42,0b1000011]],0xA//~-0b11:[chr(I+9) for I in [0o101,0x42,0b1000011]],(~-4)*(~-3):[chr(l+12) for l in [0o101,0x42,0b1000011]],0b111:[chr(l+15) for l in [0o101,0x42,0b1000011,0o104]],0o10:[chr(L+~-20) for L in [0o101,0x42,0b1000011]],(~-0b100)**2:[chr(i+ord('\x16')) for i in [0o101,0x42,0b1000011,0b1000100]]}[l[~-0b10]][~-0+l[0b10][i]-l[int()]]for i,l in enumerate([(i,I[i],[i for i in range(__builtins__.__dir__().__len__().__floordiv__(__name__.__dir__().__len__()),min(len(I),len(L)))if (I[i]!=I[i.__sub__(__builtins__.__dir__().__len__().__floordiv__(__name__.__dir__().__len__()))])|(L[i]>L[i-1]+(~-0b10)/0b10)]+[len(I)])for i in range(len(list(zip(I,L))))if (i==int())|(I[i]!=I[~-int()+i])|(L[i]>L[~-(not __name__=="__main__")+i]+(~-0b10)/0b10)])])

And now syntax highlight on desktop is not working 💀

!e

print((__builtins__.__dir__().__len__().__floordiv__(__name__.__dir__().__len__())<<(__builtins__.__dir__().__len__().__floordiv__(__name__.__dir__().__len__())<<__builtins__.__dir__().__len__().__floordiv__(__name__.__dir__().__len__()))).__add__(__builtins__.__dir__().__len__().__floordiv__(__name__.__dir__().__len__())))

And this is 0b101

[night quarry]

@grave grail :white_check_mark: Your 3.12 eval job has completed with return code 0.

5

[fleet bridge]

i guess thats because of quit() exit() copyright() license() help() shit

[grave grail]

Maybe

So my code probably being specific that you can only run it on a script

Not interpreter

[fleet bridge]

btw, __builtins__ is module in repl and builtin dict in script

that also may cause confusion

!e print(type(builtins))

[night quarry]

@fleet bridge :white_check_mark: Your 3.12 eval job has completed with return code 0.

<class 'module'>

[fleet bridge]

hmm, maybe it is other way around

also it is not clearly documented what exactly __buultins__ is

[grave grail]

Yeah, it is just used for creating 1 and 32

Good luck understanding this 🤣

[elfin ether]

Can you elaborate? Is what you have in mind CPython implementation specific?

[grave grail]

How to access any variable without specify it

[karmic pumice]

like?

[grave grail]

for example for the code yesterday

how to access I, L etc.

without specify it

without exec and eval

[karmic pumice]

define ,,without specify it"

[fleet bridge]

globals()['L']

[grave grail]

great, I forget this, this is helpful

[fleet bridge]

without attr name appearing in code object as name

[karmic pumice]

!e

x=1
print(next(filter(lambda v:isinstance(v,int),globals().values())))

[night quarry]

@karmic pumice :white_check_mark: Your 3.12 eval job has completed with return code 0.

1

[grave grail]

thx

!e
I = 1
print(builtins.getattr(builtins, 'globals')()[chr(((1<<3)*(0b11**2)).add(1))])

[night quarry]

@grave grail :white_check_mark: Your 3.12 eval job has completed with return code 0.

1

[grave grail]

now I only need to obfuscate the string

[versed eagle]

it's also different based on whether its from __main__ or not

in __main__, __builtins__ is builtins
in other modules, __builtins__ is builtins.__dict__

[grave grail]

Ok

[grave grail]

https://paste.pythondiscord.com/JXUXTS4WBB2BFTTV5BOEWNWLKA

anyone want to see this

[versed eagle]

not cpython specific
or at least doesn't have to be

though in practice it would be much easier to do in cpython than in other implementations

[versed eagle]

its like that, except its also a module in __main__

!e ```py
import os
print(os.builtins)

[night quarry]

@versed eagle :white_check_mark: Your 3.12 eval job has completed with return code 0.

001 | {'__name__': 'builtins', '__doc__': "Built-in functions, types, exceptions, and other objects.\n\nThis module provides direct access to all 'built-in'\nidentifiers of Python; for example, builtins.len is\nthe full name for the built-in function len().\n\nThis module is not normally accessed explicitly by most\napplications, but can be useful in modules that provide\nobjects with the same name as a built-in value, but in\nwhich the built-in of that name is also needed.", '__package__': '', '__loader__': <class '_frozen_importlib.BuiltinImporter'>, '__spec__': ModuleSpec(name='builtins', loader=<class '_frozen_importlib.BuiltinImporter'>, origin='built-in'), '__build_class__': <built-in function __build_class__>, '__import__': <built-in function __import__>, 'abs': <built-in function abs>, 'all': <built-in function all>, 'any': <built-in function any>, 'ascii': <built-in function ascii>, 'bin': <built-in function bin>, 'breakpoint': <built-in function breakpoint>, 'callable': <buil
... (truncated - too long, too many lines)

Full output: https://paste.pythondiscord.com/GWC4PMVEIHXIS72XPGFMZ4GG2Y

[versed eagle]

its very silly

[versed eagle]

in cpython you can do things like bytecode modification and tracing and hooking global name lookups and whatnot

[elfin ether]

Yeah, that's what I had in mind

[versed eagle]

but in any python implementation you can use codecs

[elfin ether]

codecs reference?

[versed eagle]

or at least, any conformant implementation

[elfin ether]

nvm, i see it

[versed eagle]

this gives you full control over what gets compiled

you could perform a translation like xy -> x, y

[elfin ether]

very cool. wasn't aware of this

[versed eagle]

what do you want this for?

xy -> x, y isn't really a common transformation

xy -> x * y or xy -> (x)(y) is more common

[elfin ether]

mostly just for fun. I want to be able to simulate multiple returns the way that it works in Go

[versed eagle]

that seems really annoying to work with

you wouldn't be able to return variables with multicharacter names

[elfin ether]

the main purpose is to make this possible:

xy = (2, 3)
(lambda x, y: x + y)((lambda: xy)()) # returns 5

[grave grail]

That seem correct

(lambda: xy)() # -> (2,3)
(lambda: x,y)((2,3)) # -> 5

!e
def a(x,y):return x+y
print(a((2,3)))

[night quarry]

@grave grail :x: Your 3.12 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "/home/main.py", line 2, in <module>
003 |     print(a((2,3)))
004 |           ^^^^^^^^
005 | TypeError: a() missing 1 required positional argument: 'y'

[grave grail]

Nvm

[versed eagle]

oh

for that you'd need both expansion and a *

[dapper cipher]

just kidding

[grave grail]

Anyone interested to build a obfuscating code generator

Basically it generate obfuscated code from integer and string

For a set complexity

[stark granite]

!e

something = type('thing', (), {'__init__': lambda self, s: self.value = s})("yes")
type(something).__iter__ = lambda self: iter(self.value)
for c in something:
  print(c)

[night quarry]

@stark granite :x: Your 3.12 eval job has completed with return code 1.

001 |   File "/home/main.py", line 1
002 |     something = type('thing', (), {'__init__': lambda self, s: self.value = s})("yes")
003 |                                                                           ^
004 | SyntaxError: invalid syntax

[stark granite]

raahhhh

[versed eagle]

self.__setattr__('value', s) or setattr(self, 'value', s)

lambdas have only expressions

no statements

[stark granite]

ah

!e

something = type('thing', (), {'__init__': lambda self, s: setattr(self, 'value', s)})("yes")
type(something).__iter__ = lambda self: iter(self.value)
for c in something:
  print(c)

[night quarry]

@stark granite :white_check_mark: Your 3.12 eval job has completed with return code 0.

001 | y
002 | e
003 | s

[stark granite]

i cant actually believe

i recall you could mess around with globals somehow so that this code would be like valid:

# messing around with globals

hello_world(print)  # prints «hello_world»

[versed eagle]

yes

you could do that

[stark granite]

but how

[rugged sparrow]

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

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

ob_base_p.value = missing_hook

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

init_missing_hook(globals(), lambda G, key: lambda func:func(key))
hello_world(print)```

[night quarry]

@rugged sparrow :white_check_mark: Your 3.12 eval job has completed with return code 0.

hello_world

[rugged sparrow]

basically makes the globals dict a subclass with an overridden __missing__ dunder

[elfin ether]

I have the following functions:

def add_(x, y, z):
    if x is None: return (z - y, y, z)
    if y is None: return (x, z - x, z)
    if z is None: return (x, y, x + y)
    
def sub_(x, y, z):
    if x is None: return (z + y, y, z)
    if y is None: return (x, x - z, z)
    if z is None: return (x, y, x - y)
    
def mul_(x, y, z):
    if x is None: return (z / y, y, z)
    if y is None: return (x, z / x, z)
    if z is None: return (x, y, x * y)

def div_(x, y, z):
    if x is None: return (z * y, y, z)
    if y is None: return (x, x / z, z)
    if z is None: return (x, y, x / y)

representing equations in three variables. If two of the three arguments are numeric and one is None, a tuple containing the values of all variables will be returned. I'm trying to write a function which composes these functions in the following way:

f(g(x1, y1, None), y2, z2), producing a function which accepts 4 arguments, and providing any 3 of them returns a tuple with the correct value of all 4.

The goal is for this composition to work not just for two three-variable functions, but also for functions returned by earlier compositions which may have arbitrarily many arguments.

Here's a non-working version of what I'm trying to code up.

# this doesn't work
def comp_(f, g):
    def inner(*args):
        args = iter(args)
        a, b, c = g(next(args), next(args), None)
        _, *args = f(c, *args)
        return tuple([a, b, *args])
    return inner

f = add_
ff = comp_(f, f)
fff = comp_(ff, f)
ffff = comp_(fff, f)
# these all pass
assert ff(1, 2, 3, None) == ff(1, 2, None, 6) == (1, 2, 3, 6)
assert fff(1,2,3,4, None) == (1, 2, 3, 4, 10)
assert ffff(1,2,3,4,5, None) == (1, 2, 3, 4, 5, 15)
# ff(None, 2, 3, 6) < -- this throws an error

Any help is appreciated.

[halcyon swallow]

Omg where is this

[static knoll]

isnt this the local function reference bug

nvm

ok i just stuck your code in a debugger and its immediately obvious where the problem is @elfin ether

you're calling g with "None, 2, and None"

[elfin ether]

Yes, that's the problem with a specific instance of calling the function, but the entire function logic is wrong. I only shared the function to give others somewhat of an idea of what I'm trying to achieve. The problem really is "what does the correct composition logic look like". The code I shared only works for a very limited set of functions; for example, g is assumed to only have three arguments. I think I need to either check the signature for the number of arguments or do something like:

from cytoolz.functoolz import curry
from itertools import chain

def comp_(f, g):
    def inner(*args):
        args = iter(args)
        curr = curry(g)
        while callable(curr):
            prev = curr
            arg = next(args)
            curr = curr(arg)
        args = chain([arg], args)
        curr = prev(None)
        _, *args = f(curr[-1], *args)
        return tuple([*curr[:-1], *args])
    return inner

[static knoll]

oh ok

[versed eagle]

def add_(*args):
    # it's assumed that you pass at least 3 things in, and only one of them is None
    *o, a = args
    if a is None:
        return (*o, sum(o))
    i = o.index(None)
    o.pop(i)
    o.insert(i, a - sum(o))
    return tuple(o)

that accepts more arguments

[elfin ether]

I think the variadic argument in inner is part of the problem. When composing (in the way I've described) variadic functions you can't say where the args of one end and the args of the other begin. It'd be nice to somehow wrap "inner" with a signature that is "isomorphic to" the signature of g (except the last arg) + the signature of f

[rugged sparrow]

if your goal is for composed functions to behave as f2(f1(arg1, arg2, arg3), arg4, arg5) then you either need to use keyword args or make them take 5 args. not 4

[elfin ether]

It's somewhat more like this (though not exactly):
A(B(C(v, w, None), x, None), D(y, z, None), q)
f(f(f(1, 2, None)[-1], 3, None)[-1], f(4, 5, None)[-1], None)

[rugged sparrow]

or do you want your arguments to act as a window? such that py compose(f1, f2) -> function(arg1, arg2, arg3, arg4) where arg[1-3] is passed into f1 and arg[2-4] is passed into f2

[elfin ether]

the last argument of the function is meant to represent the right hand side of some equation. for the sake of the composed function, it is an "intermediate result" that we don't care about

compose(add_, add_) should act as (w + x + y = z)

[quartz wave]

ff(None, 2, 3, 6) -> f(f(None, 2, None)[-1], 3, 6)

[elfin ether]

add_, mul_, div_, sub_, are all meant to act as x @ y = z, (replacing @ with the appropriate operation) where you provide two of the variables and you get the values for all of them

[rugged sparrow]

@elfin ether what tuple should this return?

cause is it supposed to be (1,2,3,6)?

(if it is, you would need to walk backwards from the final result tuple, and pass its args back into the first function to pull back the 1)

actually wait that would fail anyways because f(None, 2, None) would crash

[versed eagle]

you're right, their implementation is just wrong

[rugged sparrow]

it seems like the goal is some sort of equation solver, where you can make an equation with one variable and then find the variables value

[versed eagle]

mhm

[rugged sparrow]

but once you start working with multiplication and division you start having the possiblility for multiple solutions

[versed eagle]

for (None, 2, 3, 6) it should give (None, 3, 6) to the first add and then (None, 2, 3) to the second one
(*f(None, 2, f(None, 3, 6)), 6) is the desired result

[versed eagle]

only if the operation isnt known, or if there are multiple unknown values

but they specify what operation it is, and there's only one unknown

[versed eagle]

@elfin ether if this is your goal, you're going about it in a way that will be hard to do

[rugged sparrow]

yea, and the functions would need to move around for (1, 2, 3, None)

[elfin ether]

the simple three additions case looks like this:
(w + x) + y = z
and we're saying "(x=2, y=3, z=6), tell me w". By passing f(None, 3, 6) to the "outer" call to f, we can obtain (w+x) = 3, which we can then feed into the inner call to f.

[versed eagle]

1, 2, 3, None is the simplest use case

getting the result is trivial, you just evaluate all the functions with the values

[rugged sparrow]

my point there is that figuring out how to layout the functions is nontrivial once you have enough args

[versed eagle]

wdym

[rugged sparrow]

comp(add_, comp(mul_, div_)) -> x1 + (x2 * (x3 / x4)) = x5
comp(comp(mul_, div_), _add) -> (x1 * (x2 / x3)) + x4 = x5

and then adding in the fact that any of those args can be None

for example, if x2 for the second example is None then the first step would need to be None + x4 = x5

then with that result (r1) you would need to do x1 * None = r1

then with that one you would need to do None / x3 = r2

so simple composition would not work, you would need to build some sort of tree of operations and then intelligently chose branches based on arguments

[elfin ether]

idk, the tree is built for us when we compose the functions

[rugged sparrow]

but in order to know how to apply the arguments to the tree you need to know more information about each level

[versed eagle]

(f for function, o for operand, a for answer)

for c = comp(f₁, f₂, ... fₙ), we have c(o₁, o₂, ..., oₙ, o₍ₙ₊₁₎, a) -> a = o₁ o₂ f₁ o₃ f₂ ... o₍ₙ₊₁₎ fₙ
(i used reverse polish notation here because the order of operations is unambiguous)
(and also it's a chain, not a tree, which is useful here)

if a is unknown, just evaluate the expression
if a is known, then the unknown is in the expression

if it's in the expression, you can just walk up the expression, and evaluate until you find it
then when you reach the unknown you can start from the top and reduce down to the unknown
then you're just left with the the unknown, and it's not unknown anymore

[versed eagle]

you only need a tree if you use infix notation

[versed eagle]

yes but you'd still need to walk through it

and tree walking is complicated

[rugged sparrow]

fair

[elfin ether]

Does this assume each f takes three arguments?

[versed eagle]

yes, though it could be generalised

as long as the user passes the correct amount of arguments for the functions, and the functions have some way to expose how many arguments they want

[elfin ether]

ok, i think i can write this

[versed eagle]

and the functions have some way to expose how many arguments they want
to get this you can look at the code objects, or just have the functions expose an attribute

[elfin ether]

i'm just going to use the curry trick i posted above

[rugged sparrow]

you can get the arg count from function.__code__.co_argcount (for simple functions with non-variable argcounts)

[versed eagle]

mhm

presumably, none of the functions would be variadic

or if they are, they would be the last function

[sand heath]

slice = lambda to_slice, start=0, end=None, step=1: [(ret := [to_slice[i] for i in range(sorted((start if start >= 0 else start+len(to_slice), 0, len(to_slice)-1))[1],sorted((end if end >= 0 else end+len(to_slice), 0, len(to_slice)))[1], step)]), "".join(ret) if type(to_slice) is str else ret][1]

[halcyon garnet]

is the object returned by globals() the same in every module in the import chain, or different? and has this changed with python versions?

[fleet bridge]

it is the global namespace of current module

it is same for every call inside one module, different in different modules

i doubt it has been changed

[halcyon garnet]

yeah, i did some tests, and it is different in every module. is there a way to override the globals of a calling module?

i.e., if i import package z into package y, some code in z would update the globals of y

[fleet bridge]

you can mess with frame objects and achieve something like this

but why?

[halcyon garnet]

well, this is #esoteric-python

[fleet bridge]

fair enough

[halcyon garnet]

but basically my goal is to dynamically assign values to every undefined variable in a script, by simply importing the helper module

or not even assign, just intercept the name resolution process so i can do something with the names of the undefined variables

[restive void]

sys._getframe(1).f_globals should be the globals of the caller

[halcyon garnet]

it seems the fakeglobals is interfering with the import process now

__missing__: name=_imp
Exception ignored in: <function _get_module_lock.<locals>.cb at 0x000002330F137670>
Traceback (most recent call last):
  File "<frozen importlib._bootstrap>", line 177, in cb
AttributeError: 'str' object has no attribute 'acquire_lock'

first line is my trace print, last line happens because i return the name if it is not found in locals, globals, or builtins

any idea how to make it not interfere with the import?

oh, nvm, i fixed it by setting the getframe depth to 2 instead of 1

or maybe that didnt fix anything

yeah nevermind, the issue persists

[fleet bridge]

you can do nothing in these import-related frames

and you can detect them by looking at something, for example check if __name__.startswith('importlib')

not perfect, but will do the work most of the time

[halcyon garnet]

i messed around a bit, and im not sure how but i managed to silence most of the tracebacks

the weird thing is that i still see names like setattr and isinstance go through my __missing__ method

so i just grab their values from builtins so importlib stops complaining

now the real issue is that im still unable to modify the globals of the importing module

comparing the id's of the sys._getframe(1).f_globals in the imported module, and globals() in the importer, they are different

hm ok, through trial and error it seems that the actual importer module's globals are at frame depth 6...

[fleet bridge]

i guess that's because you cant change globals object through frame object

[halcyon garnet]

unbelievable. it works

[fleet bridge]

you can change type of global namespace with ctypes trickery

[halcyon garnet]

i kept the py_object.from_address(id(g) + 8).value = FakeGlobals code someone pointed to earlier

[fleet bridge]

yeah

[halcyon garnet]

where g = sys._getframe(6).f_globals

[halcyon garnet]

is there a way to find the real caller's frame, without hardcoding values?

[versed eagle]

yes

[versed eagle]

depends what you mean by real caller

and how much you know about it

[halcyon garnet]

like, the function or module or whatever stack that is one level above my code, excluding things like importlib

e.g. to get the importing module's frame, its at depth 6 during the import process, but how can i guarantee this 6 will work every time?

[versed eagle]

the module you want is one level above your code

[fleet bridge]

you can iterate over frames and skip whichever you dont like (with 'importlib' in its name for example)

maybe you can skip entire stdlib

[versed eagle]

if frame.f_globals["__name__"] in sys.stdlib_module_names or something like that

[fleet bridge]

you can also inspect module filename

[versed eagle]

mhm

[halcyon garnet]

yup, i will try these when im polishing up the code

[halcyon garnet]

one question though, why am i assigning the subclass of dict directly here, instead of an instance? if im not mistaken, globals() is an instance of dict

[fleet bridge]

because there you are not replacing instance, you are changing type of existing instance

it is like globals().__class__ = DictSubclass

but most types dont support changing types, so you are manually doing it by assigning new value to the pointer that stores type of globals()

all objects store their types somewhere in them
in recent versions they store it on offset 8 (so it is second pointer in their struct, first value is either refcount or __dict__, i dont remember)

[versed eagle]

yep, it's refcount stuff

[versed eagle]

__dict__ is stored variably iirc

oh wow i need to pay attention more to whats going on with python
apparently there's a new thing where the vm will manage the __dict__ for you

[quartz wave]

variably?

[versed eagle]

like not in the same place every time

[quartz wave]

i don't think that's the case

[versed eagle]

tp_dictoffset

[quartz wave]

it should be in the same place for as long as the class/instance is alive

unless i haven't studied enough

[versed eagle]

https://docs.python.org/3/c-api/typeobj.html#c.Py_TPFLAGS_MANAGED_DICT

[halcyon garnet]

so that is what the 8 is for, makes sense

[versed eagle]

yeah its always in the same place for the same type

by variably i meant [not in the same place for every object]

[quartz wave]

ok makes sense

[versed eagle]

whereas something like ob_type is always at a specific and constant offset (in this case, 8) for every object

[halcyon garnet]

everything works pretty much great, thanks everyone. i'll look into iterating and filtering the frames tomorrow as its almost 7am

one more question for the record, is it possible to make python accept x y z as valid syntax (with the spaces separating names)?

[quartz wave]

yes if you modify python's source code

[halcyon garnet]

well, im not going that far 😁

[quartz wave]

ok well here's a pseudo-space character ﾠ

!e ```py
xﾠyﾠz = 5
print(xﾠyﾠz)

[night quarry]

@quartz wave :white_check_mark: Your 3.12 eval job has completed with return code 0.

5

[versed eagle]

or if you modify the source of the program

[quartz wave]

codecs

[versed eagle]

with codecs or an import hook or something

[versed eagle]

yes

[halcyon garnet]

it should be typeable with a keyboard

[versed eagle]

i can type that with my keyboard

:3

[halcyon garnet]

and not with alt codes and what not

[versed eagle]

... i can type that with my keyboard :3

[quartz wave]

use codecs then

[halcyon garnet]

in other words, ascii space

[versed eagle]

use codecs or an import hook

[halcyon garnet]

can i do that externally?

[quartz wave]

?

[versed eagle]

wdym

[halcyon garnet]

cause the script with the spaces is being passed directly into the interpreter

[versed eagle]

yes

[halcyon garnet]

all the shenanigans happen during the import of my module

[versed eagle]

you can use .pth files

or site stuff

[halcyon garnet]

sorry its 7am, im not really making much sense

[versed eagle]

it's ok

[halcyon garnet]

what does a pth file do, or how could i use a codec/import hook?

[versed eagle]

if you're just running it with python3 path/to/file.py, then .pth files will work fine
so will messing with site stuff

[versed eagle]

its a file that python runs before anything else

[quartz wave]

https://github.com/thatbirdguythatuknownot/python-token-macros

reference

[versed eagle]

its meant to be used to mess with sys.path (hence .pth) but it can run arbitrary python

[versed eagle]

oo interesting

[halcyon garnet]

so i can use it to replace all the space characters in a script with commas before it is executed?

[quartz wave]

maybe

[fleet bridge]

or by implementing custom codec

codec is a thing that handles encodings, you can specify your own encoding in file # coding: my_stuff and register codec that can handle that
it should search file content for x y z patterns and replace them with whatever you want and return result which will be parsed by python parser

[grave grail]

!e
print(int().str().len())

[night quarry]

@grave grail :white_check_mark: Your 3.12 eval job has completed with return code 0.

1

[grave grail]

Simple way to get 1

with obfuscate

[rugged sparrow]

the first value is refcount, and the dict can sometimes be placed before the object in its header (which also holds things for the GC). tp_dictoffset afaik is used for the offset of the dictionary from the instance, and if it is set to -1 then python will manage the placement for you. (if it is 0 then no dict is created). When python is managing the dict for you it can sometimes be placed inside the header, or in a few other places.

[fleet bridge]

why binary ops priority is so weird

a + b | c is parsed as (a + b) | c

[dire galleon]

it makes sense

LHS | RHS where an addtion takes place in the LHS

as opposed to ... Term1 + Term 2 where X | Y takes places in Term 1

[fleet bridge]

i would like it to be parsed as a + (b | c) because bitwise ops are kinda multiplication, and should be evaluated before addition

[versed eagle]

its from C i think

[dire galleon]

It’s a logical operator still

Arithmetic operators take precedence

https://docs.python.org/3/reference/expressions.html#expression-lists

6.17

[fleet bridge]

and and or are also logical, but they are evaluated last

[dire galleon]

exactly

it makes sense for relops and logops to be done last

and by the looks of it the binary equivalents take precendence too

[fleet bridge]

hmm

yeah, now it makes a little bit of sense

🙂

[arctic elm]

bitwise ops have lower priority overall

but internally they follow sensible rules

& is like *

| is like +

^ is also like +

but apparently has higher precedence than |

bitshifts are also like *

[quartz wave]

more like power to me

since 1 << 5 is 1 * 2**5

[arctic elm]

I guess

matches with the precedence too

[grave grail]

& is like and, but instead of true and false (1 bit), it process multiple bit

| is or

^ is xor

bitshift is int(x*2**n)

[cloud fossil]

Wait

!e

print(f"{3:08b} {~3:08b}")

[night quarry]

@cloud fossil :white_check_mark: Your 3.12 eval job has completed with return code 0.

00000011 -0000100

[cloud fossil]

!e

print(f"{3:08b} {~(3 & 1 << 8):08b}")

[night quarry]

@cloud fossil :white_check_mark: Your 3.12 eval job has completed with return code 0.

00000011 -0000001

[cloud fossil]

!e

print(f"{3:08b} {~(3 | (1 << 8)):08b}")

[night quarry]

@cloud fossil :white_check_mark: Your 3.12 eval job has completed with return code 0.

00000011 -100000100

[cloud fossil]

I guess I have to do it with the good ol'
!e

print(f"{3:08b} {0x100 - 3:08b}")

!e

print(f"{3:08b} {0x100 - 3:08b}")

[night quarry]

@cloud fossil :white_check_mark: Your 3.12 eval job has completed with return code 0.

00000011 11111101

[grave grail]

???

[cloud fossil]

Tried to get bitwise operators to display two's complement

But ~ just prints the absolute value with a negative sign for negative integers

Um

[grave grail]

!e
print(~-1)

[night quarry]

@grave grail :white_check_mark: Your 3.12 eval job has completed with return code 0.

0

[grave grail]

Did you misunderstood ~ or I misunderstood you

As far as I remember, ~ is -(x+1)

[cloud fossil]

I forgot binary string formatting doesn't show two's complement

[grave grail]

Confused by my terrible English

What is complement

[cloud fossil]

Two's complement is a way of representing negative integers in binary. It also happens to play nicely with arithmetic

[grave grail]

ok

So what are you trying tbf

[cloud fossil]

Instead of using the actual formula for a negative integer's two's complement, I tried to get there with bitwise operators

And for some reason with ~ too

[grave grail]

!e
original = "10011"
length = len(original)
print(f'{abs(int(original, 2))^int("1"*length,2)+1:08b}')

🤔

!e
original = "10011"
length = len(original)
print(f'{(abs(int(original, 2))^int("1"*length,2))+1:08b}')

[night quarry]

@grave grail :white_check_mark: Your 3.12 eval job has completed with return code 0.

00001101

[cloud fossil]

!e

original = 0b10011
print(f"{0x100 - original:08b}")

[night quarry]

@cloud fossil :white_check_mark: Your 3.12 eval job has completed with return code 0.

11101101

[cloud fossil]

The top three bits are apparently wrong here

Oh but I guess you could use 1 as the padding character

!e

print(f"{abs(~2):1>8b}")

[night quarry]

@cloud fossil :white_check_mark: Your 3.12 eval job has completed with return code 0.

11111111

[cloud fossil]

That's wrong 💀

!e

print(f"{~-2:1>8b}")

[night quarry]

@cloud fossil :white_check_mark: Your 3.12 eval job has completed with return code 0.

11111111

[cloud fossil]

!e print(~-2)

[night quarry]

@cloud fossil :white_check_mark: Your 3.12 eval job has completed with return code 0.

1

[cloud fossil]

Makes sense

!e

for x in range(8):
    print(f"{~(1-x):1>8b}")

[night quarry]

@cloud fossil :white_check_mark: Your 3.12 eval job has completed with return code 0.

001 | 11111-10
002 | 111111-1
003 | 11111110
004 | 11111111
005 | 11111110
006 | 11111111
007 | 11111100
008 | 11111101

[cloud fossil]

Wait wrong side

!e

for x in range(8):
    print(f"{~(x-1):1>8b}")

[night quarry]

@cloud fossil :white_check_mark: Your 3.12 eval job has completed with return code 0.

001 | 11111110
002 | 111111-1
003 | 11111-10
004 | 11111-11
005 | 1111-100
006 | 1111-101
007 | 1111-110
008 | 1111-111

[grave grail]

???

I think you should go to#bot-commands

[cloud fossil]

True

Got it

Just one-pad the result of -~x/x+1

[fleet bridge]

or zero-pad

that depends on sign

[cloud fossil]

Ye

[fleet bridge]

!e ```py
pad=lambda n,k=8:''.join(f'{n>>i&1}'for i in range(k)[::-1])

for n in range(-5, +5+1):
print(f'{n:+} = {pad(n)}')

[night quarry]

@fleet bridge :white_check_mark: Your 3.12 eval job has completed with return code 0.

001 | -5 = 11111011
002 | -4 = 11111100
003 | -3 = 11111101
004 | -2 = 11111110
005 | -1 = 11111111
006 | +0 = 00000000
007 | +1 = 00000001
008 | +2 = 00000010
009 | +3 = 00000011
010 | +4 = 00000100
011 | +5 = 00000101
... (truncated - too many lines)

Full output: https://paste.pythondiscord.com/WBJAGVVC6P2Q7QNWGTRSFVLEXQ

[grave grail]

Tf went wrong with syntax highlight

[fleet bridge]

discord mobile moment

[versed eagle]

discord mobile syntax highlighting is Not Good™️

[grave grail]

*Terrible

Also don't ping pls

[winter siren]

is this the text channel to talk about technomancy in?

[versed eagle]

what?

[fleet bridge]

pretty much

[karmic pumice]

!e

from einspect import impl
@impl(str)
def thing(self):
  return 1
print(''.thing())

[night quarry]

@karmic pumice :x: Your 3.12 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "/home/main.py", line 5, in <module>
003 |     print(''.thing())
004 |           ^^^^^^^^
005 | AttributeError: 'str' object has no attribute 'thing'

[karmic pumice]

@dry mirage are you busy?

[dry mirage]

o/ yeah haven't gotten around to patching that yet, wonder what the issue is

!e

from einspect import view

view(str)["xyz"] = lambda: 5

s = "a"
print(s.xyz())

[night quarry]

@dry mirage :x: Your 3.12 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "/home/main.py", line 6, in <module>
003 |     print(s.xyz())
004 |           ^^^^^
005 | AttributeError: 'str' object has no attribute 'xyz'

[dry mirage]

hm...

[karmic pumice]

i tried to do that too hahah

i was getting some errors when some dictionaries stuff was happening
i dont have a clue on how it actually works so dont know either :(

[dry mirage]

!e

from einspect.structs import PyTypeObject

strType = PyTypeObject(str)

strType.tp_as_sequence[0].sq_concat = lambda self, other: "abc"

a = "dog"
print(a + "cat")

[night quarry]

@dry mirage :white_check_mark: Your 3.12 eval job has completed with return code 0.

abc

[dry mirage]

struct slots still work, but 3.12 fixed the old immutable flag hack it seems

[halcyon garnet]

i've been messing with codecs for the past day, and got regular encode/decode to work, but for some reason the incremental decoder isnt working, and i have no idea why

def _encode(input, errors="strict") -> tuple[bytes, int]:
    return utf_8.encode(_(input), errors)


def _decode(input, errors="strict") -> tuple[str, int]:
    decoded, size = utf_8.decode(input, errors)
    decoded = decoded.replace("_ ", "_,")
    decoded = "foo"
    return decoded, size


class UndersIncrementalDecoder(utf_8.IncrementalDecoder):
    def decode(self, input, final=False) -> str:
        self.buffer += input
        if final:
            buffer = self.buffer
            self.buffer = b""
            decoded, _ = _decode(buffer)
            return decoded
        return ""


def get_frame() -> inspect.FrameInfo | None:
    return next(
        (
            frame
            for frame in inspect.stack()
            if frame.code_context and "# coding: _" in inspect.getsource(frame.frame).splitlines()
        ),
        None,
    )


def register_executor(frame: inspect.FrameInfo) -> None:
    atexit.register(CustomGlobals.at_exit, frame.frame)
    py_object.from_address(id(frame.frame.f_globals) + 8).value = CustomGlobals


@register
def search_function(encoding_name) -> CodecInfo | None:
    if encoding_name == "_":
        if frame_info := get_frame():
            register_executor(frame_info)
        return CodecInfo(
            name="_",
            encode=_encode,
            decode=_decode,
            incrementaldecoder=UndersIncrementalDecoder,
        )
    return None

there is no error, python just seems to default to using utf-8 no matter what i do

my goal is to modify the source code python sees before execution

the decoded = "foo" line was supposed to test that theory, but it never shows up

[pearl pier]

what is ~ in python?

[fleet bridge]

bitwise inversion

[versed eagle]

do you need an incremental decoder/encoder?

[sly root]

!e

# test 

import math as M

F=lambda N: N*(N**2-1)
P=lambda K: K*F(2+K)+(2**K)*(F(K)**2)/F(2*K)
Q=lambda K: (1/(16**K))*((4/(8*K+1))-(2/(8*K+4))-(1/(8*K+4))-(1/(8*K+6)))

print(
    M.pi,
    4 * M.atan(1),
    355/113,
    22/7,
    Q(-0.001480581466888099),
    P(3),
    sep="\n"
)

[night quarry]

@sly root :white_check_mark: Your 3.12 eval job has completed with return code 0.

001 | 3.141592653589793
002 | 3.141592653589793
003 | 3.1415929203539825
004 | 3.142857142857143
005 | 3.141592653589792
006 | 24.942857142857143

[cloud fossil]

You can also get it from 2 * math.acos(0) or 2 * math.asin(1)

[sly root]

noice

[cloud fossil]

And don't forget math.radians(180)

[sly root]

i find it funny that you can get E with (1+2e-17)**(2**52)

2**52 casually sitting here

and oops wrong channel

[versed eagle]

from math import atan
atan(1) + atan(2) + atan(3) # pi

[cloud fossil]

Doesn't work on my machine, 1 + 2e-17 evaluates to 1.0 (probably because 2e-17 is such a small number so it doesn't even try to calculate this)

[vast wave]

floating point math moment

[hard spoke]

something isn't right here. it should be (1+1/n)**n for large n, e.g. (1+1e-10)**1e10

[fleet bridge]

!e ```py
print(
(1j**1e-9).imag*2e9,
)

[night quarry]

@fleet bridge :white_check_mark: Your 3.12 eval job has completed with return code 0.

3.141592653589793

@sly root :x: Your 3.12 eval job has completed with return code 1.

001 | Traceback (most recent call last):
002 |   File "/home/main.py", line 5, in <module>
003 |     E=C**P
004 |       ~^^~
005 | OverflowError: (34, 'Numerical result out of range')

[sly root]

!e ```py
import math as M

print(1+2e-16, 252)
print(
(1+2e-16)(2**52),
M.e,
sep="\n"
)```

[night quarry]

@sly root :white_check_mark: Your 3.12 eval job has completed with return code 0.

001 | 1.0000000000000002 4503599627370496
002 | 2.718281828459045
003 | 2.718281828459045

[hard spoke]

ah, interesting indeed. 1/2**52 is actually 2.22e-16, not 2e-16, but the error gets cut off by the float precision limits.

[weak imp]

how to change the return type of print()?

[versed eagle]

hook it

[weak imp]

yeh but which method?

[fleet bridge]

!e ```py
import sys
print = lambda x, y: x + y

sys.stdout.write(str(print(3, 5)))
sys.stdout.write(str(print('a', 'b')))

[night quarry]

@fleet bridge :white_check_mark: Your 3.12 eval job has completed with return code 0.

8ab

[weak imp]

import sys
from fishhook import hook

@hook.cls(print.__class__)
class __NewPrint:
    def __call__(*args, sep=' ', end='\n', file=sys.stdout):
        output = sep.join(map(str, args)) + end
        file.write(output)
        return output

a = print('a')
print(a)

a
None
<built-in method __enter__ of _thread.lock object at 0x000002AEC3F22000>
<built-in method __exit__ of _thread.lock object at 0x000002AEC3F22000> None None None
<built-in method __enter__ of _thread.lock object at 0x000002AEC3F22000>
<built-in method __exit__ of _thread.lock object at 0x000002AEC3F22000> None None None

why?

[rugged sparrow]

that hooks __call__ for all <class 'builtin_function_or_method'>

[weak imp]

A

makes sense

<built-in function print> a
<built-in function print> <built-in function print> a
```lol

[versed eagle]

there are lots of ways to hook things

[halcyon garnet]

im following https://shailist.github.io/posts/python-source-preprocessor-custom-encoding/

using python 3.8. if needed i can switch to a newer version

[halcyon garnet]

nvm, got it to work after some painful debugging

[halcyon garnet]

is there a way to parse code into an ast without checking if it is valid?

[quartz wave]

nope

[halcyon garnet]

i ended up solving my issue with a regex, but it has the side effect of not understanding the code context and interpreting symbols in strings as any other symbols

[quartz wave]

there's the tokenize module to tokenize python code

but i guess regex can be used too

[halcyon garnet]

hmm, maybe that would work as well

nevertheless, my abomination is almost complete

the only issue i have now is putting a .pth file in site-packages consistently

it works in a venv, but the path is wrong when installing in user-specific site

[hybrid field]

guys help me how do i fix this? I already have pygame downloaded

[torn tulip]

can you run py -0 in the terminal?

[hybrid field]

lemme chack

[cloud fossil]

You meant to ask this question in #editors-ides most likely

[hybrid field]

i can't

[hybrid field]

i'm new on the server

[cloud fossil]

👍 then this channel will seldom be helpful to you

[torn tulip]

why not?

[cloud fossil]

Did you install Python from python.org?

[hybrid field]

yes

[cloud fossil]

Did you tick the box which adds the py launcher in the installation process?

[hybrid field]

i'm using vs code

[hybrid field]

no

i started python a few days ago

[torn tulip]

type in cmd in that terminal, then run the command again

[cloud fossil]

It's the second checkbox from the bottom

Don't mind the Python version

[hybrid field]

i didn't

[cloud fossil]

Rerun the installer and check that box and go through the process again

[hybrid field]

so do i have to uninstall python?

[cloud fossil]

I think you don't have to

[hybrid field]

ok but i use python extesion in vs code

[cloud fossil]

The py launcher is independent of the editor

[hybrid field]

not the actual python

[hybrid field]

oh

ok

[cloud fossil]

If you're using the extension and it sets up the configuration file correctly, there's no need for the py launcher

You just have to run the project with the extension

[hybrid field]

also this warning didn't appear yesterday

i saves it yesterday and opened it today

[cloud fossil]

There's like a ▶️ button at the top or something if the extension is enabled

[hybrid field]

hold on

it's not green anymore

it's green again

this is the warning

nvm i'll finfish the code and see if it works

[vast wave]

every few weeks

[cloud fossil]

That checkbox is for a different configuration

Also what I posted was a screenshot from python.org, I didn't install Python like that don't worry

[unique heath]

lmao

[versed eagle]

the best installer is ./configure && make && make test && make install

:3

[grave rover]

Did the size of opcodes in the bytecode change somewhere in 3.11? I'm seeing some weird stuff in both generating and parsing bytecode that seems to suggest some opcodes are now 4 bytes maybe?

[quartz wave]

well they're the normal 2 bytes

but with CACHEs

following them

and it's necessary to know exactly how many bytes to skip after an opcode because the CACHE is literally only for storing info

[grave rover]

The co_code for a 26-instruction method is 118 bytes

[quartz wave]

that means that it doesn't always stay a CACHE

[grave rover]

Hmm, but then why doesn't dis show the CACHE ops if it does support showing them?

[quartz wave]

needs show_caches=True

[grave rover]

What do I pass to _unpack_opargs to include them?

And equally importantly, do jumps count CACHEs as size 0 or size 2

Because I'm trying to do bytecode replacement but I'm seeing weird behavior from my bytestrings for some unclear reason

[versed eagle]

2 probably

since it'd be easier to make it work that way

[grave rover]

Yeah so apparently if you remove all CACHE instructions from co_code, it completely alters the behaviour and output from dis

Because my generated co_code is identical to the original co_code pretty much

Alright I think I've got it

I just had to measure the difference in index to detect CACHEs

In next versions I'll remove them and add them again during serialization but for now I guess the user has to add them manually

[vast wave]

CACHE is important

the spec says they’re required, if they’re not present, the interpreter wont know if it should skip the next n bytes or read them

[grave rover]

gotcha

[versed eagle]

the interpreter wont know if it should ...
im pretty sure it just skips them

[vast wave]

i know

it expects them to be there

[versed eagle]

yes

[vast wave]

i was just making a point that it cant know if cache exists or not

[versed eagle]

wdym

[meager zinc]

has anyone found any interesting esoteric usages of the new 3.12 generics

def fun[T: restriction](a: T, b: T) -> T:
  return a + b

[south totem]

I haven't read the changes in 3.12 but WTH does that do (ik type hinting)

[halcyon garnet]

from what i understand, it lets you define typevars inline without the use of TypeVar

[fleet bridge]

@meager zinc

[craggy hamlet]

Golfing: https://www.codewars.com/kumite/6534637c5396a5f1b9d193c6?sel=6537510a49d3ef7767de1717
this is the best I managed to get:

Greeting=lambda n,rank='',formal=0:f"He{['y','llo'][a:=formal]},{f' {b}'if(b:=rank)*a else''} {n}{'!.'[a]}"

[next flame]

do you need the long parameter names

[next flame]

oh they do ugh

[quartz wave]

cool

i managed to find a shorter one

oh damn

i found another

[grave rover]

I got to use pyasm at work today because upstream package has a bug and the PR has been open for over a week :)

[quartz wave]

nice

[versed eagle]

Greeting=lambda n,rank='',formal=0:f"He{['y','llo'][a:=formal]},{f' {rank}'*a} {n}{'!.'[a]}"

[quartz wave]

that doesn't work when rank is empty

[vast wave]

oh wow mom i made it

i was the one who did the first “golfed” iteration of that

[thin idol]

print "hello world"

[subtle viper]

0x3c50?

[vast wave]

yes

[subtle viper]

cool

i have the latest

Greeting=lambda n,rank='',formal=0:f"He{['y','llo'][f:=formal]},{rank*f and' '+rank} "+n+'!.'[f]

[versed eagle]

oh

[hybrid field]

i don't get it which button do i have to press?

[hybrid field]

nvm i fixed it

[meager zinc]

!e I just made this seems kinda neat to me (it stores the state inside the TypeVar)

def _[__:lambda _=[]:_.append(_)or _](____:__):print(___:=len(__.__bound__()))or ___<____ and _(____)
_(5)

[night quarry]

@meager zinc :white_check_mark: Your 3.12 eval job has completed with return code 0.

001 | 1
002 | 2
003 | 3
004 | 4
005 | 5

[meager zinc]

!e this is a similar concept but extended for a different result

def _[__:lambda _=[]:_.append(_)or _]():print(end=chr(___:=len(__.__bound__()))*(-~___%0x69<=1))or ___<0x80 and _()
_()

[night quarry]

@meager zinc :white_check_mark: Your 3.12 eval job has completed with return code 0.

hi

[turbid dragon]

What kind of sorcery syntax is this?

Python 3.12 stuff?

How are you doing def _[…]

[quartz wave]

python 3.12 stuff

[versed eagle]

they changed the type annotation syntax

[fleet bridge]

they didn't change old one, they added additional syntax

[meager zinc]

yeah its new type annotation syntax:

def function[T: restriction](a: T, b: T) -> T:
  return a + b

def function[T](a: T) -> T:
  return a

https://peps.python.org/pep-0695/

[turbid dragon]

I still dont understand typevars and their purpose

[meager zinc]

it's supposed to help connect the types of different arguments and the type of the return

[turbid dragon]

So its sort of like an any but links argument and return type?
Whats the point of [T:str] instead of just str then?

[meager zinc]

I think there’s a difference in how subclasses/superclasses are handled

E.g. if T is a subclass of str then all cases of T must be that subclass

whereas just using str would allow them to be different

[meager zinc]

^

[proper vault]

consider an example like

def f[T: Animal](a: T) -> list[T]:
    return [a]
def g(a: Animal) -> list[Animal]:
    return [a]
f(Cat()).append(Dog()) # bad, can't add a dog to list[Cat]
g(Cat()).append(Dog()) # good, can add a dog to list[Animal]

[versed eagle]

addition is a type of change

[meager zinc]

(the second call to f is supposed to be g I think) [fixed]

[proper vault]

true

[turbid dragon]

So the main purpose of type vars is dealing with inheritence?

[turbid dragon]

Nice example

[earnest wing]

https://mypy.readthedocs.io/en/stable/generics.html#generics

[north wraith]

do you like... donuts?

while ('o_' in dir()) or (A := (0)
               ) or (B := 0) or print((                           
            "\x1b[2J")) or not ((sin := ((                        
         __import__('math'))).sin)) or not (                     
      cos := __import__('math').cos) or (o_ := (                   
     True)): [pr() for b in [[(func(), b) for ((z                 
    )) in [[0 for _ in range(1760)]] for b in [[ (                
   "\n") if ii % 80 == 79 else " " for ii in range(               
  1760)]] for o, D, N in [(o, D, N) for j in range((              
 0), 628, 7) for i in range(0, 628, 2) for (c, d, e,(             
f), g, l, m, n) in [(sin(     i / 100), cos(j / 100),(             
sin(A)), sin(j / 100),          cos(A), cos(i / 100) ,            
cos(B), sin(B))] for              (h) in [d + 2] for (            
D, t) in [ (1 / ( c               * h * e + f * g + (5
)), c * h * g - f *                e)] for (x, y) in [            
(int( 40 + 30 * D * (              l * h * m - t * n)),            
int( 12 + 15 * D * ( l           * h * n + t * (m))))]            
 for (o, N) in [(x + 80 *     y,int(8 * (( f * e -  c              
  * d * g) * m - c * d * e - f * g - l * d * n)))] if             
  0 < x < 80 and 22 > y > 0] if D > z[o] for func in              
   [lambda: (z.pop((o))), lambda: z.insert((o), (D)                
    ),(lambda: (b.pop( o))), lambda: b.insert((o),                 
     ".,-~:;=!*#$@"[ N if N > 0 else 0])]][ 0][1]                 
       ] for pr in [lambda: print("\x1b[H"),                    
         lambda: print("".join(b))] if (A :=
            A + 0.02) and ( B := B + 0.02)]                        
               #..--~~EvanZhouDev:~~--.#                           
                    #...,2023,...#

Run this bad boy

[meager zinc]

you stole that from that video

[north wraith]

There is literally the programmer's name on the donut, so he gets the credit.

not me lol

[meager zinc]

whoops

I didn't mean steal

I meant like its from

https://www.youtube.com/watch?v=tzpfyTFvU6M

this video

[unreal echo]

Impressive, very nice

Lets see paul allens script

[robust plank]

on a 1 to 10 cursed scale where would you put a holonomic_turtle subclass ?

[elfin ether]

Ok, I think this works:

def add_(x, y, z):
    if x is None: return (z - y, y, z)
    if y is None: return (x, z - x, z)
    if z is None: return (x, y, x + y)
    
def comp_(*fs):
    def inner(*args):
        curr = None
        ret = [*args]
        if args[-1] == None:
            for f in fs:
                n = getattr(f, '__argcount__', f.__code__.co_argcount) - 2 + int(not bool(curr))
                curr = f(curr[-1], *args[:n], None) if curr else f(*args[:n], None)
                args = args[n:]
            ret[-1] = curr[-1]
            return tuple(ret)
        
        gs = iter(fs)
        index = 0
        g = None
        curr = None
        seen_stop = False
        
        for f in gs:
            n = getattr(f, '__argcount__', f.__code__.co_argcount) - 2 + int(not bool(curr))
            if None in args[:n]:
                g_args = list(args[:n+1 - int(not bool(curr))])
                j = args.index(None)
                args = args[n:]
                index += j
                lhs = curr
                g = f
                break
            index += n
            curr = f(curr[-1], *args[:n], None) if curr else f(*args[:n], None)
            args = args[n:]
            
        gs = iter([*gs][::-1])
        curr = None
            
        for f in gs:
            n = getattr(f, '__argcount__', f.__code__.co_argcount) - 2 + int(not bool(curr))
            curr = f(None, *args[-n:], curr[0]) if curr else f(None, *args[-n:]) 
            args = args[:-n]
            
        if lhs:
            if curr:
                g_args[j] = lhs[-1]
                ret[index] = g(g_args[0], None, curr[0])[j+1]
            else:
                ret[index] = g(lhs[-1], *g_args)[j+1]
        else:
            ret[index] = g(*g_args, curr[0])[j]
        return tuple(ret)  
    n = 2
    for f in fs:
        n += (f.__code__.co_argcount - 2)
    inner.__argcount__ = n
    return inner

these all pass at least..

f = comp_(add_, add_)
g = comp_(f, f)
answer = (1, 10, 100, 1000, 10000, 11111)
assert g(1, 10, 100, 1000, 10000, None) == answer
assert g(1, 10, 100, 1000, None, 11111) == answer
assert g(1, 10, 100, None, 10000, 11111) == answer
assert g(1, 10, None, 1000, 10000, 11111) == answer
assert g(1, None, 100, 1000, 10000, 11111) == answer
assert g(None, 10, 100, 1000, 10000, 11111) == answer

[cloud fossil]

You're breaking a rule which you've agreed to follow

[halcyon garnet]

in a python package's setup.py, is there a way to find out which exact site-packages path it is being installed to?

decided to ask here since theres no packaging channel, and this ties into the esoteric stuff i've been working on

my end goal here is to put a .pth file into a site-packages where it will be executed by the interpreter regardless of how the package was installed

[fleet bridge]

from pathlib import Path
import pip 
site_pkgs = Path(pip.__file__).parent.parent

not the best idea in the world, but we are in #esoteric-python

are site packages always installed in .../lib/site-packages/ ?

[earnest wing]

see docs for the site module (responsible for reading pth files)

[dreamy bridge]

i made funny working interpreter for programming language in python

make number integer ask (comments are in brackets)
make count 0
make count add count 1 (increment)
zap 10 if modulo count 3 (zap just teleports to a line)
zap 8 if modulo count 5 (conditions can only be used in zap lol)
say "FizzBuzz"
zap 14
say "Fizz"
zap 14
zap 13 if modulo count 5 
say "Buzz"
zap 14
say count
zap 3 if count less number (loop till end)

its called portal and completely functionless classless for and while loopless
only has ints and strings for now but im happy

[restive void]

Only works if the pip you're installing with is contained in the same site-packages you're installing to, right?

[fleet bridge]

yes

if you know at least one package that is installed in site-packages, you can do that (replace pip with that module in my example)

[earnest wing]

you can fetch all the appropriate directories

just follow site.py

[unique heath]

so basically asm with more features

[halcyon garnet]

i can get the user site packages and global site packages with site.getusersitepackages() and site.getsitepackages(), but how do i know which one to copy the pth file to during installation?

[dreamy bridge]

i dont know whats asm, just got this idea and decided to do it

[unique heath]

assembly

with more features

but no fucntion

s

[dreamy bridge]

i guess

never saw assembly before

i guess its more high level assembly

yea

[hearty saddle]

loool

https://tenor.com/view/american-psycho-impressive-very-nice-coping-patrick-bateman-gif-26518058

btw guys how could i do this shorter?
while 1:w=[int(input())for _ in' '*8];print(w.index(max(w)))

[quartz wave]

while w:=eval('int(input()),'*8):print(w.index(max(w)))

[hearty saddle]

thanks dude

[mint hinge]

def fib(N):
  f = [0, 1]
  for i in range(1, N+1):
    f.append(f[i-2] + f[i-1])
  return f[N]

def fib(N):
  f=[0,1]
  for i in range(1,N+1):
    f.append(f[i-2]+f[i-1])
  return f[N]

def fib(N):
  f=[0,1]
  for i in range(N):
    f.append(f[i]+f[i+1])
  return f[N]

def fib(N):
  f=[0,1]
  for i in range(N):f.append(f[i]+f[i+1])
  return f[N]

def fib(N):
  f=[0,1]
  for i in range(N):f+=[f[i]+f[i+1])
  return f[N]

Could this be shortened any further?

[low lynx]

fib=lambda n:n<2and n or f(n-1)+f(n-2)

[gleaming linden]

Yeah that

[restive void]

!e

import fishhook
fishhook.hook(str, name="__index__")(lambda x:["zero", "one", "two", "three", "four"].index(x))

print(range(0, 100, 3)["three"])

[night quarry]

@restive void :white_check_mark: Your 3.12 eval job has completed with return code 0.

9

[fleet bridge]

!e

import fishhook
fishhook.hook(int, name="__repr__")(lambda x:dict(zip(range(5), ["zero", "one", "two", "three", "four"])).get(x, fishhook.orig(x)))

print(*range(10))

[night quarry]

@fleet bridge :white_check_mark: Your 3.12 eval job has completed with return code 0.

zero one two three four 5 6 7 8 9

[fleet bridge]

hehe

[versed eagle]

depends if you want the package to be user-specific or not

[distant salmon]

Just a nitpick, you are recursively calling a function that doesn't exist

[halcyon garnet]

i think the user or pip decides that, for example based on whether the user ran the installation command as admin or not

[meager zinc]

!e ```py
from fishhook import*
hook(int,name="repr")(lambda x:o if(o:=orig(x))and x>9999 else''.join(["zero","one","two","three","four","five","six","seven","eight","nine"][int(v)]+["","ty"," hundred"," thousand"][~-len(orig(x))-i]+" "for i,v in enumerate(o)))

print(~-1)
print(9261)
print(4 + 890)

[night quarry]

@meager zinc :white_check_mark: Your 3.12 eval job has completed with return code 0.

001 | zero 
002 | nine thousand two hundred sixty one 
003 | eight hundred ninety four 

[meager zinc]

if you have it recurse you can get it to 999999 but I was lazy

!e (also it doesn't handle teens very well and doesn't handle tens very well)

from fishhook import*
hook(int,name="__repr__")(lambda x:o if(o:=orig(x))and x>9999 else''.join(["zero","one","two","three","four","five","six","seven","eight","nine"][int(v)]+["","ty"," hundred"," thousand"][~-len(orig(x))-i]+" "for i,v in enumerate(o)))

print(11, 12, 13, 14, 15, 16, 17, 18, 19)
for i in "123456789":
  print(int(i * 4))

[night quarry]

@meager zinc :white_check_mark: Your 3.12 eval job has completed with return code 0.

001 | onety one  onety two  onety three  onety four  onety five  onety six  onety seven  onety eight  onety nine 
002 | one thousand one hundred onety one 
003 | two thousand two hundred twoty two 
004 | three thousand three hundred threety three 
005 | four thousand four hundred fourty four 
006 | five thousand five hundred fivety five 
007 | six thousand six hundred sixty six 
008 | seven thousand seven hundred seventy seven 
009 | eight thousand eight hundred eightty eight 
010 | nine thousand nine hundred ninety nine 

[fleet bridge]

onety one onety two onety three onety four onety five onety six onety seven onety eight onety nine
i love that ❤️

wait

they all are like that

😄

[low lynx]

not a nitpick, I missed that, good point

f=fib=lambda n:n<2and n or f(n-1)+f(n-2)

[unique heath]

onety

twoty

threety

fourtu

fivety

[fleet bridge]

tbh, this naming system is better than actual one

[unique heath]

yeah

welcome to eengleesh

[cloud fossil]

Someone in a recent reform wanted to spell machine as masheen

[fleet bridge]

eengleesh moomeent

[earnest wing]

1840 is pretty recent

[unique heath]

eengleesh

[unique heath]

compared to several hundred thousand years yeah

[halcyon garnet]

is there a way to detect if my package is being installed in the user site packages or system site packages? i have a .pth file that is included as a data file, with the path Lib/site-packages. it works fine when installed in a venv or system site-packages, where the path matches, but in the user site-packages the path should be PythonXX/site-packages instead. if the pth file is placed in the user site-packages under Lib/site-packages it does not work

or maybe the better question is, is there a way to determine the exact path to the site-packages of installation, so that i don't have to handle all the cases separately going off just sysconfig.get_path("purelib")

i can get either the system site-packages or the user site-packages, but i cant determine which to copy my file to during installation, as the user could install into either

[compact raven]

I hate the walrus.

From datetime.py in the standard library, original:

    s = _format_time(self._hour, self._minute, self._second, self._microsecond, timespec)
    tz = self._tzstr()
    if tz:
        s += tz
    return s

supposed improvement:

    s = _format_time(self._hour, self._minute, self._second, self._microsecond, timespec)
    if tz := self._tzstr():
        s += tz
    return s

My suggestion:

    return \
      (
            _format_time(self._hour, self._minute, self._second, self._microsecond, timespec)
        +
            (("", tz)[tz] for tz in (self._tzstr(),))[0]
      )

[halcyon garnet]

anyway, here is the result of my hard work and wasted time lol
https://github.com/lemonyte/_

[gleaming linden]

Most of it comes from experience, the more you golf the better you get

and along the way you'll learn the useful tricks

[serene stratus]

is there a shorter way of saying None in python?

like in less than 4 chars

or a solution that would allow an if to be put against it

[fleet bridge]

can you show code?

[static knoll]

0?

what is the usecase

the empty string is technically 0 characters

[serene stratus]

The exercise is just to return None, I have a solution that doesn't return anything so implicitly returns None but I was wondering if there's a way to return None if ... But then change None into something that allows if to be stacked against it

[fleet bridge]

# 1:
return a if c else None

# 2:
if c:return a