wow I made a load of mistakes lmao
#esoteric-python
gilded orchid
(probably cos' I coded this all in one go and then tested it afterwards, probably not the smartest idea)
sick hound
del(inp[0])``` you can't do that with a string, do `inp = inp[1:]` insteadnow it actually works 😃 (after I put the whole thing in a try except to avoid the index error from codePos at the end)
now I need to find a way to implement cells before the first one, so doing < at the begining works
west hornet
Which channel here is for talking about Python design patterns ?
vocal oyster
for general discussing, #python-discussion
if it's related to a problem you have and you need specific help, then any of the help channels will do
gilded orchid
this channel is for people who take python way too far
west hornet
I am taking python way to far :), but the issues is not too far 😄
onyx jacinth
This channel is also for people who take python too far to make it small
brisk zenith
well, we're more than just golfers :D
marsh void
we're
special = [chr(c) for c in (*range(33, 48), *range(58, 65), *range(91, 97), *range(123, 127))]
[Credits][Eviee#0666]
sick hound
we're py special=[chr(c)for a,b in((33,15),(58,7),(91,6),(123,4))for c in range(a,a+b)]
tepid glacier
we're
special=[*__import__('string').punctuation]
gilded orchid
I'm currently working on a stack-based esolang, and this is the syntax I have so far
| = Pops the number off of the top of the stack
. = Pushes 0 to the stack
; = Pushes X on to the stack. If X is not passed, pushes 5 to the stack
+ = Adds X to the top number of the stack
- = Subtracts X from the top number of the stack
* = Multiplies X by the top number of the stack
# = Square roots the number at the top of the stack
/ = Divides the top number of the stack by X, then rounds it to the nearest whole number
( = Begin a loop. If after it there is a number, repeats by that number. If after it there is a =, followed by a number,
the loop will end once the end is reached and the top number of the stack is equal to X.
) = End a loop.
> = Swaps the top value of the stack, with the X from bottom of the stack
< = Swaps the top value of the stack, with the X from top of the stack
[ = Begins a function that gets added to the function slot. There is only one space for functions, and it'll get replaced if a new one is began
] = Ends a function
^ = Clears the function
% = Runs function X. If function X is not passed, runs the function in the function slot
? = Runs the function in the function slot if the number of the top of the stack is equal to X
= = Runs the single instruction after it, if the top value of the stack is equal to X
~ = Roll the stack by X
' = Outputs the number X. If X is not passed, outputs the number at the top of the stack
" = Outputs the unicode character of X. If X is not passed, outputs the unicode character of the number at the top of the stack
(Note that 'X' refers to a number directly after the instruction, and it can be any positive integer)
any suggestions?
marsh void
@gilded orchid fucks brains well
gilded orchid
I'm thinking of having a seperate stack for functions, but I don't know how I'd implement that
(I'm thinking up the syntax before actually starting making it for context)
calm rampart
suggestion: operands are for cowards, just make digits operators that pop the stack, multiply by 10, add the number, and push
😛
gilded orchid
🤔
calm rampart
and every other operator's X is just popped from the stack
(this is arguably less esoteric, but more stack-based)
anyway the other thing i'd do is have a way to create strings and a way to evaluate them - dc builds loops and functions on this
gilded orchid
Could I maybe have a string slot that can be appended to, cleared and outputed
calm rampart
how about hollerith style strings: H: Reads the next [top of stack] bytes of input and places them on the stack as a string object
to build loops and functions dc-style, you need string objects that can be placed on the stack and string variables, not just a single string slot
gilded orchid
maybe I can have 3 different stacks, the number stack (the main one), the string stack and the function stack
and then have 3 different instructions that switch what stack the commands are running on
calm rampart
for conditionals you could have an operator that's "If the number on the top of the stack is non-zero, pop two items. Otherwise, pop the number and remove the third item from the stack, leaving the second"
honestly you can probably build loops from that and an eval function without even having an explicit loop operator
oh, a way to copy any given item from the stack can probably usually be used in place of variables.
gilded orchid
to replace X?
calm rampart
no i meant for the features i've been suggesting that require variables
gilded orchid
oh right
calm rampart
what if the language were stringly typed
then you could just have one stack and all elements on the stack are strings
math operators just treat the string as a number
gilded orchid
quick question, is the second number in the stack second from the top, or second from the bottom?
snow beacon
Top. Always top.
marsh void
Heh, more stack fun incoming? @gilded orchid
gilded orchid
I'm making the language way more limiting, to make it a lot harder to code in
. = Pushes 0 to the stack
# = Pops the number off of the top of the stack
; = Duplicates the value at the top of the stack
< = Swap the two values at the top of the stack
> = Roll the stack by the top number of the stack
^ = Makes a string out of the top number of the stack characters below the top character of the stack, and then pops all of those, and pushes that string to the stack
~ = Runs the string at the top of the stack. If it isn't a string, do nothing.
? = Run the string at the top of the stack, if the second number of the stack is non-zero
- = Subtract one from the number at the top of the stack
+ = Add one to the number at the top of the stack
" = Outputs the string at the top of the stack, and pops it off of the stack. If it is a number, do the same with that number corresponding unicode character.
how does this look, I got rid of a load of stuff and added @calm rampart's string suggestion, to make it way more annoying to code in
calm rampart
always the top
in most stack languages each operator only cares about however many operands on the top of the stack, and ignore an unlimited amount below
gilded orchid
I'll do that after I quickly write an interpreter (also I'm gonna add a * command to make it so you don't have to do 40 or more +'s to start a loop lmao)
marsh void
lol
gilded orchid
what's an easy way to swap the two things at the end of a list?
wait nvm, figured it out
marsh void
a[-1], a[-2] = a[-2], a[-1]
Anyway >.>
with my subclass of list called nlist, I can do:
from ntools import nlist # not on pypi yet
a = nlist(1,2,3,4)
a.swap(-1, -2)
So it’s cool :)
gilded orchid
you should find a way to modify lists so they function as nlists when using []
gilded orchid
ok, just finished the interpreter (I haven't actually tested it yet so it probably won't work lmao)
stack = []
def run(code):
codePos = 1
while codePos<=len(code):
exec({'.':'stack.append(0)',
'#':'stack.pop()',
';':'stack.append(stack[-1]',
'<':'stack[-1],stack[-2]=stack[-2],stack[-1]',
'>':'[stack.append(stack.pop(0))for i in range(stack[-1])',
'^':'[stack.append("".join([chr(i)for i in stack[-1:-stack[-1]:-1]));del(stack[-2:-(len(stack[-1]))-2])'
'~':'(run(stack[-1])if type(stack[-1])==str else pass)',
'?':'((run(stack[-1])if type(stack[-1])==str else pass)if stack[-2]!=0)',
'-':'stack[-1]-=1'
'+':'stack[-1]+=1'
'*':'[stack.append(stack[-1]*stack[-2]),del(stack[-2:-3:-1])]'
'"':'(print(stack[-1]if type(stack[-1])==str else print(chr(stack[-1])))'}[code[codePos-1]])
codePos += 1
(I used a really long dict cos' I was lazy)
gilded orchid
ok, I've done that
here's a (I think) working version
codeDict = {'.':'stack.append(0)',
'#':'stack.pop()',
';':'stack.append(stack[-1]',
'<':'stack[-1],stack[-2]=stack[-2],stack[-1]',
'>':'[stack.append(stack.pop(0))for i in range(stack[-1])',
'^':'[stack.append("".join([chr(i)for i in stack[-1:-stack[-1]:-1]]))];del(stack[-2:-(len(stack[-1]))-2])',
'~':'(run(stack[-1])if type(stack[-1])==str else pass)',
'?':'((run(stack[-1])if type(stack[-1])==str else pass)if stack[-2]!=0)',
'-':'stack[-1]-=1',
'+':'stack[-1]+=1',
'*':'stack.append(stack[-1]*stack[-2]);del(stack[-2:-4:-1])',
'"':'print((stack[-1]if type(stack[-1])==str else chr(stack[-1])),end="")'}
stack = []
def run(code):
codePos = 1
while codePos<=len(code):
exec(codeDict[code[codePos-1]])
codePos += 1
run('.+++++++++++.+++*"')
brazen geyser
you can use this syntax for dicts too:
some_dict = {
'key': 'value',
'key': 'value',
}
if you want it a little cleaner
gilded orchid
that's basically what I'm already doing, except it looks stupid because of short discord line lengths
(here's how it should look)
gilded orchid
i'm working on it
marsh void
Oki
gilded orchid
god damn it, I just made 'Hello corld!'
ok got it, hello world is .++++++.++++++*;.++*"#;.+++*;-------"#""+++"#;----"#;.+++*---------------------"#;.+++*+++"+++"------"--------"#;---" @marsh void
Imma try and create the fibonacci sequence using it
marsh void
noms a cookie
gilded orchid
I'm gonna add a new instruction that outputs the number, just to make this not take an eturnity to make
jade dust
What can you do with the ~ or ? operator?
Are you limited to executing single instructions, or is there some way to set up a chain?
gilded orchid
it runs a string, and the string can be as long as you want
it can be used, for example, for loops by using a string that keeps on running itself
gilded orchid
@marsh void I just realised to do that, I’m gonna need nested loops, so i need to make a strong that makes another string
oh god
jade dust
Time to make a quine!
gilded orchid
Quines actually aren’t possible lmao, because it uses eval so it breaks if you try to output "
lmao
jade dust
Wait, how?
gilded orchid
yeah maybe
I was sort of just bodging it together, so readability didn't really matter when I was making it
I didn't even do any testing until after I made it, and then I fixed the 9 different syntax errors lmao
gilded orchid
Currently my roll instruction is moving everything down, and moving the bottom item of the stack to the top
is that what roll is meant to do?
calm rampart
what do you mean by down
sorry, my brain's always twisted up by this part because the first stack-based language I learned was HP48 calculators, which displays the top of the stack at the bottom of the screen
anyway, the HP48 description of ROLL is
ROLL
Moves a specified level to level 1
( e.g. 4 ROLL will move level 4 to level 1 )
i.e. it moves the X'th item of the stack (after popping X) to the top.
so I think yours is right then
swap is just 2 roll; rot is 3 roll; and dup is just 1 pick, so it might be worthwhile to make your available primitives roll/rolld/pick instead of dup/swap/roll
@gilded orchid
gilded orchid
I think mine's wrong, because doing roll on [1, 1, 2, 1, 2] gives [2, 1, 2, 1, 1]
and I think it should give [1,2,1,1,2]
frank hazel
@gilded orchid what does "roll" do?
gilded orchid
'Rotate: Also referred to as “roll,” specifies the number of elements in a stack which are rotated in their order. For example, rotating the top four elements of a stack would move the topmost element into the fourth position while the next three elements move up one position.'
frank hazel
so that should rotate the top 2 elements on the stack, then?
wouldn't that be [1, 1, 1, 2]
gilded orchid
my roll thing rolls the entire stack by the top value on the stack
gilded orchid
now that I think about it, I might change it to roll the top X of the stack (where X is the number at the top of the stack)
frank hazel
and I think it should give [1,2,1,1,2] this makes no sense
why isn't it popping the top number on the stack?
gilded orchid
good point, I should probably change that
gilded orchid
I'm trying to implement the 'proper' definition of rol, but I'm really struggling to actually get it to work
I currently have [stack.insert(-i,stack.pop(-1))for i in [stack[-1]]*stack.pop(-1)], but that rolls the entire stack, instead of only X from the top of the stack
nvm, I got it to work
brazen geyser
top of the stack is index 0?
brisk zenith
well if we consider how we can use lists as stacks, doing list.append(x) is like pushing to the stack and list.pop() is obviously popping from the stack
brazen geyser
im just a lil confused as to how rolling [1, 1, 2, 1, 2] should give [1,2,1,1,2]
thought maybe the top of the stack was index 0 in this case
brisk zenith
i'd expect it to be -1 🤔
gilded orchid
I used [stack.insert(-i+1,stack.pop(-1))for i in [stack.pop(-1)]] to make it work like that definition I sent earlier
(pop the top object off of the stack and roll X below it by one)
marsh void
hm
brisk zenith
not really an #esoteric-python question, but probably unpacking
sick hound
I wonder what features of Python can be turing complete
Lambdas and function calls are because lambda calculus (although technically you do run into a recursion limit)
but what else?...
grave rover
metaclass abuse for sure
wind maple
almost all operators are turing complete because you can put anything you want inside the operator overload 😃
gilded orchid
Can one-liners be Turing complete?
grave rover
Anything can be a oneliner
distant wave
Anything can be anything
behold:
>>> def a():
... print("1")
...
>>> a()
1
>>> a.__qualname__ = "b"
>>> a
<function b at 0x10a79d0d0>
>>> a.__code__ = compile("print('c')", "<string>", 'exec')
>>> a()
c
That function you passed me is now mine :P
gilded orchid
Is there a list of all the different double underscore things in python?
brisk zenith
look at the python data model in the documentation
there's an entire page on it, but it shows and explains all dunders as well as other things
specifically section 3.3
gilded orchid
are you able to override all dunders?
eg: could you override new to run something whenever a new instance of a class is made
gilded orchid
Would you just create a class called new?
marsh void
I don't think so
gilded orchid
After doing some research, I’m pretty sure you just create a class that derives from object and then define new in that
wait nvm I’m actually an idiot
marsh void
@gilded orchid yep, you define __new__ in class, and actually run something in it lol
gilded orchid
how would you do that?
marsh void
hm
class ntuple(tuple):
def __new__(cls, *args):
do_some_stuff()
return super().__new__(cls, args)
little example I guess
gilded orchid
what is super()?
marsh void
I am bad at explaining, honestly
gilded orchid
I just googled it and ‘
super() gives you access to methods in a superclass from the subclass that inherits from it.’
I don’t understand any of that lmao
brisk zenith
there's hardly anything esoteric about this :D
gilded orchid
(I’m just bad at python)
marsh void
>>> class nlist(list):
... def __init__(self, *args):
... super().__init__(args)
>>> n = nlist(1,2)
>>> dir(n)
['__add__', '__class__', '__contains__', '__delattr__', '__delitem__', '__dict__', '__dir__', '__doc__', '__eq__', '__format__', '__ge__', '__getattribute__', '__getitem__', '__gt__', '__hash__', '__iadd__', '__imul__', '__init__', '__init_subclass__', '__iter__', '__le__', '__len__', '__lt__', '__module__', '__mul__', '__ne__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__reversed__', '__rmul__', '__setattr__', '__setitem__', '__sizeof__', '__str__', '__subclasshook__', '__weakref__', 'append', 'clear', 'copy', 'count', 'extend', 'index', 'insert', 'pop', 'remove', 'reverse', 'sort']
so super() would represent list in that case? (For the class at the top)
marsh void
don't really understand what you mean >.>
gilded orchid
Nvm
rugged sparrow
@gilded orchid this make any sense?
gilded orchid
so super just adds the mainClass's code onto it?
rugged sparrow
Super() gives you the instance of the class that the methods are inherited from
Sort of
But it's specifically that instance
gilded orchid
oh ok
rugged sparrow
Yeah
gilded orchid
so super().test() is an instance of mainClass().test()?
bronze idol
Sorta, it's actually more like mainClass.test(self), giving the method of main class the current instance of subClass
super() is just shorthand for getting the parent class and passing the current instance.
sick hound
class MainClass:
def print_me(self):
print(self)
class SubClass(MainClass):
def print_me(self):
print('hello!')
super().print_me()
main = MainClass()
main.print_me() # <MainClass object at 0x......>
sub = SubClass()
sub.print_me() # hello! <SubClass object at 0x......>```completely random, but what would you guys say is the worst esolang hello world?
the Malbolge Unshackled one is absurd https://paste.pythondiscord.com/pobivugini.py
(it's longer then 2k chars so it can't fit in a discord message lmao)
wind maple
definitely fugue
gilded orchid
wow that is ummm... interesting code
also verbose is quite something
AND OUTPUT IT FOR THE CURRENT PERSON USING THIS PROGRAM TO SEE lmao
wind maple
naming your language verbose and deciding to use roman numerals instead of unary
:pathetic_skinner:
gilded orchid
I find it hilarious that you can't see the text "Hello World!" in 99% of these
calm rampart
i'd say the numbers should be spelled out in english
marsh void
>>> def nrange(start, end, step):
... ret = []
... while start < end:
... ret.append(start)
... start += step
... return ret
...
>>> nrange(1,2,0.1)
[1, 1.1, 1.2000000000000002, 1.3000000000000003, 1.4000000000000004, 1.5000000000000004, 1.6000000000000005, 1.7000000000000006, 1.8000000000000007, 1.9000000000000008]
Can someone explain a bit?
marsh void
It feels quite weird
bronze idol
That's floats for ya. If you really need decimal math to work with exact precision check out the Decimal module that comes with Python https://docs.python.org/3/library/decimal.html
brisk zenith
it's not really esoteric, but there's a website for it :D https://0.30000000000000004.com/
bronze idol
With a domain like that it's esoteric 😝
marsh void
That’s actually quite neat 
gilded orchid
wow, Minifuck is quite something
< : Same as Brainfuck
. : Move to next bit and invert it and output letter stored in first 8 bits, if none then input
[ : Move to next bit and invert it, skip next instruction and invert next bit if zero
I feel like it'd be fun to try and golf that
gilded orchid
Is there a short way to convert a list of binary into a denary number? (eg: [1,0,1] would become 5)
polar plover
int("".join(map(str, my_list)), 2) might work
!e ```py
my_list = [1, 0, 1]
my_number = int("".join(map(str, my_list)), 2)
print(my_number)```
night quarryBOT
@polar plover Your eval job has completed.
5
Ves Zappa#3787 (not gonna ping) just gave me an identical solution in #help-coconut lmao, thanks though
(also the reason I asked here is because someone said I should because I was trying to golf it)
polar plover
np 😃 xD
gilded orchid
oh wow, 2/3 char brainfuck exists
< = Move pointer one cell to the left. (If pointer is currently over the first bit, do nothing.)
( = Move to next bit and flip it, skip next instruction if zero.
any other character = No-op instruction.
I'm working on a compiler to Subleq, which only has one instruction. It's a little different though because it has operands.
gilded orchid
even after looking at esolangs.org/wiki/Subleq, I'm still really confused how it works lmao
frank hazel
@gilded orchid hello do you have a minute to talk about our Lord and savior
@snow beacon what does it compile to?
snow beacon
No, It compiles to Subleq. No way am I going to write any programs with one instruction.
snow beacon
Just some syntax I threw together, with a lot of prewritten compiler macros in Python.
frank hazel
cool
snow beacon
I've got a bnf, but no language name.
I think I'll upload it somewhere once I'm happier with it. I was planning on using it for the pride esoteric python golf challenge. I'd have some golfed Subleq code, a golfed Subleq interpreter, and probably the compiler as a reference.
gilded orchid
@frank hazel ?
frank hazel
@gilded orchid you said you didn't know how subleq works
gilded orchid
Oh right
so subleq (the instruction not the language) takes 3 arguments
but what do they do?
gilded orchid
Does the third run the 3 numbers that are there as a subleq instruction?
frank hazel
all 3 are memory addresses, shockingly
it takes address 1
wait no
subleq x, y, z is like y = x - y; if y <= 0: jmp z
it sets y to x subtracted by y then if y is leq (less than or equal to) 0 it jumps to z
pretty simple
gilded orchid
What’s does jump do?
frank hazel
jumps the instruction pointer to that memory address
are you familiar with how control flow in assembly works?
gilded orchid
not really
frank hazel
it's all about jumping
time for pseudoassembly
say I want to check if two addresses contain the same number
I might do something like
00 cmp $F0E2 $F0E3
10 jeq 30
20 jeq 40
30 <if code>
40 <rest of code>
not real assembly, but
the idea is to structure it around jumping around since that's the only option you have
gilded orchid
So jump changes what instructions are running?
frank hazel
yes
a while True loop with jumps would look like
0xE000: uhh do something
0xE001: do another thing
0xE002: jmp 0xE000
jumping based on addresses is more of a machine code thing, though
in assembly you have something called labels to make it make much more sense
:loop
wow
stuff here
wow code
jmp loop
this is more similar to something a real assembly language would have
Python has gotos, by the way, guess what
pip install goto
then you can import it
frank hazel
sort of, yeah
but in machine code there are no "lines of code"
only instructions
it would jump to the instruction stored at X memory address
gilded orchid
Would the first instruction be instruction 0?
frank hazel
yes
well that's not exactly defined by subleq itself, it might begin executing at some other memory address
I think most subleq emulators start at 0
gilded orchid
Also, does it jump to value X, or jump to the value stored at address X (where X is the first value)
frank hazel
it jumps to address X
the way execution works is there's an instruction pointer, which is an address
each time it reads the 3 values the pointer points to
executes that as subleq
then increases the pointer and goes again
jumping just sets the instruction pointer, basically
gilded orchid
So basically subleq has control flow, if statements and subtraction
in one instruction
frank hazel
yes, that's basically the idea
frank hazel
wdym
memory addresses, like list indexes, always start at 0
but some of the lower numbers might be reserved, like with CHIP-8
where the program actually starts at 0x200
gilded orchid
I meant does their value start at 0?
frank hazel
wdym value
gilded orchid
I’m so confused lmao
gilded orchid
I was thinking kind of like brainfuck
gilded orchid
are memory addresses nothing until you try and use it for something
and then it gets assigned to 0 and runs that instruction?
frank hazel
@gilded orchid it's always something
snow beacon
Your code can and should modify itself. Your code is in memory starting from zero generally. After that it's probably zeroes, or it could throw errors. My interpreter has accessing unset values terminate the program.
That way I can put everything in a list
The first thing I put in my compiler was labels, so I didn't have to bother with absolute addresses in code I'm going to change later.
marsh void
I know, kinda not esoteric, but how can I make the following code faster?
x, y = 1280, 720
a = []
for i in range(x):
for j in range(y):
a.append(
(i, j)
)
I found one though
from itertools import product as f
x, y = 1280, 720
a = [(i, j) for i, j in f(range(x), range(y))]
Is there a faster one?
nocturne saddle
I think you saved one character with that as f import
edgy kelp
simple list comp is considerably faster than the original
distant wave
Smaller, bit slower
x, y = 1280, 720
a = [(i // y, i % y) for i in range(x*y)]
That's... slower?
from itertools import product
a=[]
a.extend(product(range(x),range(y)))
How does python x, y = 1280, 720 a = [(i, j) for i in range(x) for j in range(y)] compare?
snow beacon
Or perhaps using python list(itertools.product(range(x), range(y)))?
edgy kelp
the extend of product or list-ing it are the fastest, and probably the same (with list being more readable)
gilded orchid
I challenge you guys to golf the UBFIM interpreter(https://esolangs.org/wiki/Ultimate_bf_instruction_minimalization!)
It's suprisingly complicated because of the how input output works in it
marsh void
@gilded orchid ultimate bf sounds like ultimate boyfriend lol
gilded orchid
lmao
gilded orchid
is there a short way to do 1-liner assignment to variables?
marsh void
wym?
gilded orchid
i'm trying to do it in a list
marsh void
oh
give me an example please, I am bad at abstract stuff
you can write it normally, and I can try to one-line
gilded orchid
I was just wondering for it generally
I'm not actually doing anything atm
(apart from my failed attempt to make that esolang interpreter I mentioned earlier)
polar plover
!e py foo = [1, 2, (locals().__setitem__("bar", 5), 3)[1], bar * 2, bar * 3] print(foo)
night quarryBOT
@polar plover Your eval job has completed.
[1, 2, 3, 10, 15]
yeah that works
polar plover
but, why?
night quarryBOT
marsh void
yup
print(a:=10)
def _():
try:
idx0 = width_height.index(t0)
except ValueError:
idx0 = 0
try:
idx1 = width_height.index(t1)
except ValueError:
idx1 = 0
how to make that shorter?
gilded orchid
for i in [(idx0,t0), (idx1,t1)]:
try:
eval(f'{i[0]}=width_height.index({i[1]})')
except:
eval(f'{i[0]}=0')
not really pythonic but this works I think
@marsh void
edgy kelp
could be a bit shorter sinceo nly the numbers change on idx and t
marsh void
lol, it gives NameError obv
gilded orchid
oh yeah
edgy kelp
but when you get to eval the question if you really want it to be shorter comes up
edgy kelp
well there are plenty variables that are terribly long in there 😄
gilded orchid
for i in [0,1]: #shorter then range(2) lmao
try:
eval(f'idx{i}=width_height.index(t{1})')
except:
eval(f'idx{i}=0')
(that's shorter btw)
@marsh void
marsh void
my ass was lazy so for readability I just made a function for that lol
gilded orchid
not really related but I just realised you can do 0,1 instead of [0,1] for golfing
neat
snow beacon
It says t{1} rather than t{I}
gilded orchid
oh yeah
oops
https://esolangs.org/wiki/English I love that this has a page lmao
'Many people are familiar with it even if they don't know any other programming language.'
unique kettle
huh. i never knew python even had a unary positive operator: https://docs.python.org/3/library/operator.html#operator.pos, https://docs.python.org/3.7/reference/datamodel.html#object.__pos__
i'm having trouble thinking of what i might even want to overload it to do
worthy ridge
well, maybe it would be useful if you had a lib to do particle physics too
bronze idol
Invert could be used for doing fuzzy checks ~my_fancy_thing == my_var, __invert__ would need to return an object that overloads __eq__... Would be way more clear with a named method lol
unique kettle
the one time i've seen ~ overloaded unconventionally is in pyparsing: https://pyparsing-docs.readthedocs.io/en/latest/pyparsing.html#pyparsing.NotAny
bronze idol
Wow lol
brazen geyser
@marsh void you can use two loops in one list comp
[(i, j) for i in range(w) for j in range(h)]
not sure if faster though
one cool thing you can do with the invert operator is inverting sequence indexes
>>> def palindromic(sequence):
for i in range(len(sequence)//2):
if sequence[i] != sequence[~i]:
return False
return True
>>> palindromic('abba')
True
>>> palindromic('abbo')
False
>>> palindromic('aboba')
True
>>>
i like how much semantic sense it makes
plus you probably save some cycles over -i-1
marsh void
@brazen geyser this one is shorter?
def palindromic(seq):
return reversed(seq) == seq
mortal ingot
would that really work? doesn't reversed returns an iterator?
candid haven
It works, since I tested it yesterday
distant wave
seq[::-1] == seq
mortal ingot
In [1]: def palindromic(seq):
...: return reversed(seq) == seq
...:
In [2]: palindromic('abba')
Out[2]: False
>>> a = "abba"
>>> a[::-1] == a
True
that does beg the question though, what's the shortest way to check palindromeness for any type of iterable?
candid haven
mmm maybe I tested it yesterday and I didn't use it because it returned an iterator 😐
sorry
mortal ingot
no worries
distant wave
[*a][::-1] == [*a]
marsh void
@mortal ingot wait weird lol
mortal ingot
any iterable, including iterators
brazen geyser
is the [:] necessary
distant wave
no
mortal ingot
(though it's probably just a=[*i];return a[::-1]==a or something)
sick hound
@marsh void nope
>>> def test():
... yield 1
... yield 2
... yield 1
...
>>> g = test()
>>> [*g]
[1, 2, 1]
>>> [*g]
[]```ugh I forget about that
sick hound
the best I can do is py def p(s):l=[*s];return l[::-1]==l
wait no
def p(s):*l,=s;return l[::-1]==l```the shortest one is where you add a special palindromic check operator to cpython and use that 😛
sick hound
that's cheating though
brazen geyser
😠
sick hound
now you have a closed relationship
brazen geyser
homewrecker
mortal ingot
relationship.seek(0)
sick hound
ValueError: I/O operation on closed file.```relationship jokes? lol
@sick hound yeah, I got what is the problem; generator objects, for example, are not subscriptable
brazen geyser
relationships: dict
sick hound
they're not subscriptable, and if you iterate them twice you might get different things
worthy ridge
checking an iterator is a palindrom certainly requires casting it into a list or tuple, as checking it forward will consume it, and you don't know the length in advance, and you can't look forward anyway
sick hound
from ctypes import*;pythonapi._Py_Dealloc(py_object(meta))```meta has been deallocated now
they don't exist
sick hound
no
worthy ridge
damn, that's abusing the unpackaging syntax 😄
sick hound
>>> *a, = range(20)
>>> a
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19]```@sick hound importing * is a sin apparently, even worse than ; in normal code
brazen geyser
def p(s): return (a:=[*s])[::-1]==a
sick hound
well it's less characters in this case :P
marsh void
yep
worthy ridge
@marsh void it's not an import of *, it's using the unpacking syntax
sick hound
from ctypes import*``` this is an import of `*`@worthy ridge lol I know that; was kinda slow; I replied to ^
worthy ridge
ah ok
marsh void
sorry for inconvenience 
marsh void
Hey guys, really need your help
import itertools
a = [*itertools.product(range(1280), range(720))]
b = [*itertools.product(range(10, 100), range(20, 200))]
is there a fast way to remove all elements in b from a?
numpy?
edgy kelp
something with sets ought to be faster than list manipulation
list(set(a)-set(b))
but you lose how it was arranged
marsh void
yeah, and actually
c = list(set(a)-set(b))
c.sort()
@edgy kelp why not sort it after operation?
edgy kelp
not sure how long sets with sort will take compared to some kind of list comprehension
marsh void
yeah, that might take long
edgy kelp
almost a second on my pc
marsh void
it actually did sort quite fast lol
>>> a = [*itertools.product(range(1280), range(720))]
>>> b = [*itertools.product(range(10, 100), range(20, 200))]
>>> c = list(set(a)-set(b))
>>> c.sort()
took around a second, even less probably
edgy kelp
well second's a lot but trying list comprehension now is taking horribly long compared to that
you can replace the initial lists with sets too if you're not doing anything with them
marsh void
set comprehension isn't a thing?
edgy kelp
just set(itertools.product(range(1280), range(720)))
but there is set comprehension with {} iirc
marsh void
yeah, didn't know about setcomp tho
edgy kelp
still going for least chars?
brazen geyser
you can use a SortedSet from https://pypi.org/project/sortedcontainers
marsh void
no, not really @edgy kelp
snow beacon
I've given up on implementing a stack based language I designed with two or three modes for almost all of the punctuation. It's a lot of instructions to implement.
But once you get into non-ascii?
I had enough trouble reading APL.
marsh void
@gilded orchid brainfuck is cool tho
gilded orchid
yeah
but I started hating it after trying to do anything difficult with it
for my 6000 character 12 days of Christmas in Brainfuck, I resorted to using python to generate it for me
(after a failed attempt doing it in brainfuck myself)
marsh void
the good thing is that you can put other chars in code and the get ignored
(at least in online compiler tho)
gilded orchid
what compiler do you use for brainfuck?
personally, I mainly use Boof (https://brandly.github.io/boof/) as it updates in real time, however it bugs out if your program is too long. if that happens, I switch over to using Jelly (https://btzy.github.io/jelly)
marsh void
@gilded orchid why not write your own tho
Dunno, just googled and found one online compiler, don’t remember it’s name
gilded orchid
I tried writing my own but got stuck at nested loops
i'll have another shot at it sometime though
marsh void
@gilded orchid can you send a link of bf instructions here? I forgot one of them >.>
gilded orchid
> Move the pointer to the right
< Move the pointer to the left
+ Increment the memory cell under the pointer
- Decrement the memory cell under the pointer
. Output the character signified by the cell at the pointer
, Input a character and store it in the cell at the pointer
[ Jump past the matching ] if the cell under the pointer is 0
] Jump back to the matching [ if the cell under the pointer is nonzero
@gilded orchid kinda don’t get what , does
worthy ridge
let user type a character? like input() in python, but limited to one char
i don't get what's interesting about writting brainfuck (or ook or whichever equivalent) code, sure, it's neat to know it's turing complete ande hence you can write any program in it, but why would you? if you want a very constrained env, you can write assembly code, or try to do parallelized code in gpu, or serverless code using distributed erlang code, things where the constraint brings you something
at least piet code is nice to look at
gilded orchid
Basically brainfuck programmes can have an input buffer, and , takes the ascii value of that input and puts it onto the current cell. Then that input gets removed from the input buffer
@worthy ridge because it’s a fun challenge, and it’s a difficult logic puzzle (at least that’s why I like brainfuck)
marsh void
@gilded orchid ok in my mind I have implementation of around 40-50 lines of code
marsh void
@worthy ridge I would say, esolangs are made for fun, and saying "why would you?" is not a very good question here imo.
worthy ridge
probably
i mean, there is no shortage of interesting problems, i rarely see a point in making them harder, but that might just be me
marsh void
Hey guys, how can I modify variable with exec?
def run():
out = ""
exec("out += '0'")
print(out) -> ""
# I'd like it to be "0" that way
>>> exec('a = 5', globals(), locals())
>>> a
5
>>>
>>> def run():
... out = "0"
... exec("out+='1'", globals(), locals())
... print(out)
>>> run()
0
@brazen geyser
candid haven
is this cheating? py def run(): out = "0" _locals= locals() exec("out+='1'", globals(), _locals) out = _locals["out"] print(out)
marsh void
@brazen geyser lol I was pissed off and have just written a class that holds that variable
marsh void
@gilded orchid
NeKitDS $ bf
++++++++++++[-<+++++>]<.
p+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
while arr[p]:
arr[p]-=1
p-=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
p+=1
p-=1
store.output+=chr(arr[p])
Output: '<'
So yeah, it works but it’s kinda bad at indentation
frank hazel
@marsh void bf -> python compiler?
frank hazel
I made a thing
from collections import*
from operator import*
from functools import*
f=defaultdict(int)
def q(n):
f[n]+=1
for x in filter(partial(gt,n),[0,1]):q(n-x-1)
return n
for i in range(q(int(input())),0,-1):print(f[i])
I made an esoteric fibonacci
and then golfed it, I guess
marsh void
@frank hazel yeah, bf -> python compiler
marsh void
NeKit $ bf
>++++++++[-<++++++>]<+..-.+.
.bf -> .py
p+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
while arr[p]:
arr[p]-=1
p-=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
arr[p]+=1
p+=1
p-=1
arr[p]+=1
store.output+=chr(arr[p])
store.output+=chr(arr[p])
arr[p]-=1
store.output+=chr(arr[p])
arr[p]+=1
store.output+=chr(arr[p])
-----Output-----
1101
you're adding to store.output (which isn't even defined in the code?) instead of printing
p and arr aren't defined either
marsh void
It is not actual code
Lemme send the real one (it’s bad tbh)
@frank hazel there -> https://paste.pydis.com/ogeguxeguj.py
marsh void
Ok so I needa write so people can copy+paste code that was compiled and it can run? @frank hazel
frank hazel
I would do that
marsh void
@frank hazel sorry for mention, there you go -> https://paste.pydis.com/ekicutikat.py
marsh void
Okay
NeKit $ bf test.bf
.bf -> .py
from collections import defaultdict
p = 0
arr = defaultdict(int)
output = ""
... # compiled BF code goes here
@frank hazel okay updated ^
frank hazel
nice
marsh void
https://paste.pydis.com/fetibesida.py haha first serious brainfuck program
@frank hazel ok some actual program up here ^
also, added new command:
$ - add value of ceil to output
e.g. >+$<$ -> 10
gilded orchid
@marsh void can it handle this program I wrote a while ago? (https://pastebin.com/WJVxQnvp)
marsh void
Let me see
My terminal lagged a bit, I need to make debug: bool thing for it not to print stuff
gilded orchid
It's meant to output the lyrics to the song 12 Days of Christmas, but I accidentially used the wrong new line char (I used 13 instead of using 10 lmao)
marsh void
On the First day of Christmas
My true love sent to me
A Partridge in a Pear Tree.
On the Second day of Christmas
My true love sent to me
Two Turtle Doves, and
A Partridge in a Pear Tree.
On the Third day of Christmas
My true love sent to me
Three French Hens,
Two Turtle Doves, and
A Partridge in a Pear Tree.
On the Fourth day of Christmas
My true love sent to me
Four Calling Birds,
Three French Hens,
Two Turtle Doves, and
A Partridge in a Pear Tree.
... # discord char limit
On the Twelfth day of Christmas
My true love sent to me
Twelve Drummers Drumming,
Eleven Pipers Piping,
Ten Lords-a-Leaping,
Nine Ladies Dancing,
Eight Maids-a-Milking,
Seven Swans-a-Swimming,
Six Geese-a-Laying,
Five Gold Rings,
Four Calling Birds,
Three French Hens,
Two Turtle Doves, and
A Partridge in a Pear Tree.
wow it actually works lmao
marsh void
It did print :)
Yeah lol
I can give you the interpreter code if you’d like (it’s quite messy rn)
gilded orchid
ok
If you're wondering how it works, it basically multiplies the number 32 over around 200 different cells, then it goes through all of them and adds/subtracts to make them into the correct number, so the cells are formatted (numbers for string, 0, numbers for other string, 0, etc.)
then I spam [.>] and [<] to print them all in order
i'm gonna try and golf a brainfuck to python compiler
gilded orchid
ok thanks
marsh void
np, have fun :)
marsh void
@gilded orchid I’m looking for a way to detect if a brainfuck loop is while True:... any suggestions?
gilded orchid
I don’t really know how you’d do that
frank hazel
it actually is impossible
it's the halting problem, exactly
@marsh void basically, it comes down to if it were possible to do that it would be possible for any program, including the program that checks if the loop goes forever
so you can make a program that checks itself to see if it goes forever
if the answer is yes, end the program
if the answer is no, loop forever
no matter what the answer is wrong
so the program that checks if the program loops forever does not exist
snow beacon
You can check if the program exactly repeats itself
Or simple things like a monotonically increasing cell
(That last one would actually be pretty difficult. You'd need to be sure that it doesn't affect other cells.)
But for the former you can just store a bit of a history of program states, and if the program repeats itself exactly, break.
gilded orchid
I’m gonna have a shot at
https://code-golf.io/brainfuck#brainfuck later
although I probably won’t be able to do it
marsh void
How the fuck should brainfuck in brainfuck work? @gilded orchid
gilded orchid
a couple of them already exist lmao
I have an idea of how I'm gonna do everything
so basically I'm gonna start of taking input, which will consist of, using , (obviously) and then repeatedly subtracting from it to work out what brainfuck command (if any) it is
then I'm gonna split the into into lots of 8 bit chunks, where each bit coreesponds to a command
and between each 8 bit chunk there's gonna be an empty gap, which I can put 255 into
255 will represent where it is in the program (and you can always jump to it by using +[->+]-)
and then I'll have the cells, which will be 2 cells each
one of the cells will be that cell's value when running the brainfuck programme, and the other one will be either 0 of 254 (254 represents the position in the cells)
marsh void
So you need to have an other Lang interpreter already to execute interpreter of brainfuck
gilded orchid
yeah basically lmao
gilded orchid
wdym?
gilded orchid
depends on what implentation of brainfuck you're using
marsh void
hm
gilded orchid
wait no i'm dumb
marsh void
😂
gilded orchid
yeah you would have to enter a null character for that to end
marsh void
@gilded orchid I mean, do you take inputs with ,, and then execute them?
gilded orchid
no, I'm gonna take all the inputs and then execute after that
because executing them straight away wouldn't work for [ or ]
marsh void
Okay, then how?
gilded orchid
I haven't actually started doing it yet but I've planned it out
I think + will be +[->+]->+ because +[->+]- will jump to the current cell and then it'll just add to it
marsh void
@gilded orchid I found this (not mine):
>>>,[>++++[-<-------->]+<-[----------[>]>[<+<+>>>>]<<<-[>]>[<+<+>>>>]<<<-[>]>[<+<+>>>>]<<<-[>]>[<-<+++>>>>]<<<--------------[>]>[<++<+>>>>]<<<--[>]>[<-<+++++++>>+>>]<<++++[-<------>]+<+[>]>[<++<+>>>>]<<<--[>]>[<<+>>>>]<<-<[+]<[>]>,>]<]>>[->]>+<<<<-[<]>[-[<<]>[<+[>]>>[<]<<[>>+[<<[<]<<-[>>]<[>>>>[>]>+<<[<]<]<-[>>]<[>>>>[>]>-<<[<]<]<++[->>+<<]>>[>]>]<]<[<]>-<]>-[<<]>[<++[>]>>[<<]>[<<+[<<[<]>[-<<+>>]>--[<<]>[[>]>+<<[<]<]>+[<<]>[[>]>-<<[<]<]>+[>]>]]<<[<]>--<]>-[<<]>[[>]>>.<<<[<]<]>-[<<]>[[>]>>-<<<[<]<]>-[<<]>[[>]>>,<<<[<]<]>-[<<]>[[>]>>+<<<[<]<]>-[<<]>[[>]>>>[>>]>[<<<[<<]<+>>>[>>]>-]>[-]<<+[<[->>+<<]<]<[->>+<<]<[<]<]>-[<<]>[[>]>++[-->[-<<+>>]>]+<<-[++<<]<[->>>[>>]>+<<<[<<]<]<[<]<]<++++++++>>[+<<->>]>]
In my implementation raises IndexError if I give it \n
fallow cairn
Ooh make a programming language that compiles to brainfuck
It would have to be basically brainfuck but readable with comments and ascii as ints, ontop of a "smart array" system that groups stuff together over the virtual 30 something array you have in brainfuck
marsh void
@fallow cairn any ideas how to make this work even if user hits enter?:
input(": ")[0]
wind maple
sys.stdin.read()
marsh void
Oh yeah
fallow cairn
getchar()
marsh void
this ain’t python lol
wind maple
getchar() isn't a standard function
fallow cairn
It is C though
marsh void
^
marsh void
sys.stdout.read() gets one character?
wind maple
I have no idea what the goal here is
, in BF is basically the same as doing os.read(sys.stdin.fileno(), 1)
gilded orchid
I find it absurd that Slashes (https://esolangs.org/wiki////) has been proved to be turing-complete despite the fact that it only has string replacement
marsh void
NeKitDSS $ bf --debug p.bf
-----Input-----
, # get input
. # print input
[-] # clear cell
.bf -> filter
,.[-]
.bf -> .py
from collections import defaultdict
p = 0
arr = defaultdict(int)
output = ""
def f(k):
n = input(f"In [#{k}]: ")
return chr(0) if not n else n[0]
arr[p]=ord(f(p))
output+=chr(arr[p])
while arr[p]:
arr[p]-=1
print(output)
In [#0]: 2
-----Output-----
2
@gilded orchid the thing is, you can replace text with more string replacement
fallow cairn
👏 @marsh void
I want to make that programing language a reality
Need some syntax ideas though
It will basically be a improved brainfuck that compiles to the original
marsh void
@fallow cairn wym compiles to the original?
fallow cairn
Transpiles back to brainfuck
But I want to add a complicated mechanism so it makes a shit load of brainfuck
marsh void
Hm I have ideas how to improve BF, but it won’t be translatable back
fallow cairn
Like a function has to be duplicated in brainfuck text over and over whenever compiled
Like to make the ir (brainfuck) much larger in source then original
Basically make a monstrosity that is nice to work with and makes very, very ugly brainfuck
Like py x=0 for i in range(3): add(x) def add(x): x += 1 Would have to have the add function duplicated every time it ran
:D
edgy kelp
NameError: name 'add' is not defined
marsh void
>+++[-<+>]< @fallow cairn monstrosity?
gilded orchid
^ why lmao
fallow cairn
Functions that go into other functions and recursion would be fucking annoying
And would make the source code massive for something other then like the hello world of bf
Well time to learn basic bf!
marsh void
@fallow cairn bf I simple, you just need to keep track of what cell are you pointing at
fallow cairn
Yep
gilded orchid
(and also nested loops can get kind of insane so you need to keep track of that lmao)
marsh void
^
gilded orchid
even printing out a 3 digit number from a cell is difficult
fallow cairn
Or strings
gilded orchid
strings are relatively easy compared to printinng out a number from a cell
because you need to do division and modulo only using +1 and -1
fallow cairn
Right
I need to learn language dev
Then i'll do it
I can't think straight right now, way too tired 😅
gilded orchid
just to prove a point
[->+>-[>+>>]>[+[-<+>]>+>>]<<<<<<]
>>>[>>+<<-]>>>>++++++++++<<
[->+>-[>+>>]>[+[-<+>]>+>>]<<<<<<]>>>>>
++++++[>++++++++<-]>[<<+<+<<<<+>>>>>>>-]
<<<<<<<.>>>>>.<.
this prints a 3 digit number from a cell (assuming you have [number, 0, 100] and your pointer is on the number)
(also I stole that divmod algorithm from https://esolangs.org/wiki/brainfuck_algorithms#Divmod_algorithm)
marsh void
@gilded orchid I don’t think that you stoled it tbh, just took it lol
wind maple
@gilded orchid that was actually kind of a problem in a coding competition I took part in
gilded orchid
oh wow
wind maple
we had to create a program that would determine whether a string could be constructed by repeating a smaller string n times
using only x->y string substitution
e.g. k=2 and the string is 1010 would output true
but 0110 would output false
marsh void
+[,]
# Transpile update: Hitting Enter on input sets the cell to 0
also storing/doing arithmetic number above 255 is a nightmare (assuming you're using an implementation that wraps around)
I think you'd either have to store it in binary or store it in base 255
snow beacon
All this talk of compiling to brainfuck makes me thing of my original goal for my compiler to Subleq: make really incomprehensible code from much nicer code.
It's technically finished, but I'm going to keep adding macros.
silk yew
Im afraid of the darkness in the soul of all of you
frank hazel
@fallow cairn sounds good, consider checking out other projects that try to do the same thing
brain comes to mind
also me and two others tried to do this with a language based on C
fallow cairn
Dude I still have to learn it :P
frank hazel
maybe you can get further than everyone else
fallow cairn
¯_(ツ)_/¯
Gave myself 2 weeks to learn how to make a very basic language
Like very basic
frank hazel
how basic
frank hazel
https://esolangs.org/wiki/Ly made this a while ago
fallow cairn
Damn
500 lines
Ok how the fuck did you make this when you where 12
@frank hazel
frank hazel
shrug
fallow cairn
Chirst
gilded orchid
Not really related but how were you able to get your language added to tio.run? @frank hazel
frank hazel
@gilded orchid asked the creator, he lets anyone add their lang
use the code golf stack exchange chat
gilded orchid
You need 20 reputation for that
rip me
How do you actually get reputation on stack Exchange?
snow beacon
Ah, that's what I've been doing wrong.
marsh void
@gilded orchid woah that brainfuck in brainfuck actually works with my new transpiler:
In [#3]: +
In [#4]: [
In [#5]: ,
In [#6]: .
In [#7]: ]
In [#8]:
In [#10]: a
In [#10]: b
In [#10]: c
In [#10]:
-----Output-----
abc
On #8 I hit Enter
Also takes exclamation mark as an end of input
frank hazel
@marsh void why is the input that way?
marsh void
@frank hazel what? You can’t input more than one char per cell
It is brainfuck in brainfuck
sick hound
What if you write a python compiler in brainfuck and a brainfuck compiler in python
gilded orchid
marsh void
@sick hound python compiler in brainfuck? Think about this shit first
gilded orchid
(I think it was a joke)
snow beacon
You can simulate decimal math. It's Turing complete
marsh void
@snow beacon while being Turing complete, it takes hella shit amount of time to calculate something
wind maple
why did this channel turn from esoteric python into brainfuck
tropic gulch
marsh void
@wind maple don’t know honestly
frank hazel
somewhere down the line it turned into esolangs
iz ok I got it
from collections import*
from operator import*
from functools import*
f=defaultdict(int)
def q(n):
f[n]=-~f[n]
for x in filter(partial(gt,n),[0,1]):q(~-n-x)
return n
for i in range(q(int(input())),0,-1):print(f[i])
boom new and improved obfuscated-golfed fibonacci thing
brisk zenith
i'm back from being away from home, and ava makes a good point. your esolang projects are cool, but i think it's best to get this channel back to being true #esoteric-python. :D
marsh void
while not print("Hello", end=" "):
print("World!")
Would esolang projects be good for the off topic channels? @brisk zenith
marsh void
@gilded orchid I’m guessing so :)
brisk zenith
sure, but also #303934982764625920 could be used when you want to show them off.
gilded orchid
If it’s something we created in an esolang, and not an esolang written in python, would it be ok for #303934982764625920?
brisk zenith
not really, no.
grizzled cloak
@marsh void that will never exit the loop right?
grizzled cloak
yeah
grizzled cloak
But it will never exit
marsh void
yes
gilded orchid
(unless it increases memory usage each time, meaning it’d eventually run out of memory)
wind maple
well
in a theoretical world that argument doesn't work
computers wouldn't be turing complete if you account for the fact that they have limited memory
but we say they are because, well, for all intents and purposes they are
so for all intents and purposes a while True loop does indeed go on forever
in practice no if you were to do something silly like allocate 10gb of memory each iteration
but if you just no-op each time then you will indeed run "forever"
until your machine fails
marsh void
yep
formal sandal
@frank hazel
n,a,b=[],[],[[]]
for i in b*int(input()):n=[n]
while n:print(len(b));a,b,n=b,[*a,*b],n[0]
@grizzled cloak
[0, [...]]
frank hazel
@formal sandal thanks, it's even more awful
wind maple
have you seen my list comp one
formal sandal
@frank hazel python eval('exec("n,a,b=[],[],[[]]\\nfor i in b*int(input()):n=[n]\\nwhile n:print(len(b));a,b,n=b,[*a,*b],n[0]")')
It seems i'm good at making terrible and unreadable code.
My favorite on is the Brainfuck interpreter
import sys
for p in sys.argv:
s,a='',[0]*32;j=t=0
for i in p:s+=' '*t+['j-=1','a[j]+=1','j+=1','a[j]-=1','print(chr(a[j]),end="")','while a[j]:',0,'',0,0][ord(i)%18-6]+'\n';t+=(i>'Z')-2*(i>'[')
exec(s)
good, that sucks
formal sandal
It's 204 characters long, and someone figured out how to make it 118 characters long.
I feel helpless.
fallow cairn
The slowest bf interpreter :D
marsh void
++[>+++++[>+++++++>+++>++++++++++<<<-]<-]>>+.>>[>+>+<<-]>>[<<+>>-]<<+.++++++++++.--.>+.<+++++++.--.+++++++.<++.[-]<---.[-]>>>----.<------.[-]>+++++++.[-]
Here’s the name of one of my favorite games
formal sandal
[+]
marsh void
@formal sandal oh shit here we go again
also, it doesn’t do anything
+[+] is the real deal
formal sandal
Right.
marsh void
+[-+] this is more sad
brisk zenith
i go away for another weekend and we're back off the rails for #esoteric-python!
worthy ridge
the size of code is really the worst metric one could think of to judge a software on
snow beacon
I don't know about that. If your storage capacity is limited, you don't want your code taking up much space.
distant wave
Larger code takes longer to read, from a certain point onwards
gilded orchid
If it’s really short code then it could become very unreadable
snow beacon
Code golf is to programming what poetry is to literature.
Poems are useful ways of conveying information when you have limited pages. Golfed code is the same for pages of memory.
formal sandal
@distant wave Not really. Well, technically, a shorter piece of code might be faster to literally read, but it might take a lot of time to process it.
A poem might be short, but the shorter the poem is for the same amount of meaning, the harder it is to understand.
And there are also awful poems.
marsh void
And there are awful short codes
snow beacon
Seems more and more like a good analogy to me.
formal sandal
Many poets die at the age of 27.
Likewise, ninja coders die at 27 because their colleague programmers are so pissed off.
marsh void
@snow beacon same
@formal sandal ninja code yes
I’m kinda having a strange feeling when typing help('modules') in IDLE
gilded orchid
I challenge you guys to obfuscate the number 0 the best you can
I came up with [not[[]]][not[[]]]
or, even better, [[[[not[[]]]]]][not[[]]][not[[]]][not[[]]][not[[]]]
gilded orchid
>>> not[]
True
>>> not[[]]
False
🤔
hot crypt
def _(__):
return (lambda _ : _.__code__.co_argcount)(__)
def __():
return 1
_(__)```(globals().__setitem__("sys",-1),len(__import__("sys").argv)+sys)[sys]```(globals().__setitem__("sys",-1),len(__import__("sys").argv)**0+sys)[sys]```
@polar plover (so it’s going to be always 0, no matter how many args there are)hmmmmm,,,,,,, no xD
marsh void
Explain
polar plover
ohh wait, you changed it
thought you meant mine
yes yours will always be 0
from what I can see
marsh void
yup
And your won’t always be 0
stupid python, thinks that 0**0=1
But len(sys.argv) won’t be equal to 0 though
Or can it be 0?
Kinda interested 🤔
polar plover
I knew that it would not always be 0, but never thought of the **0
marsh void
Same
I was kinda sitting trying to think of how to operate on x to make it equal to 1
polar plover
I got another idea: py (<super long code to get any number>) * 0
hot crypt
int(str(ord(str(type(1))[0]))[1]) 🤔
marsh void
import sys
from collections import defaultdict
class Container:
def __init__(self):
self.output = ""
def clear(self):
self.output = ""
store = Container()
def run(code: str):
p, indent = 0, 0
arr = defaultdict(int)
pycode = (
"""def f(k):
n = input(f"In [#{k}]: ")
return chr(0) if not n else n[0]
"""
)
rdict = {
"+": "arr[p]+=1",
"-": "arr[p]-=1",
">": "p+=1",
"<": "p-=1",
".": "store.output+=chr(arr[p])",
",": "arr[p]=ord(f(p))",
"?": "print('Array:', list(arr.values()))",
"[": "while arr[p]:",
"]": ""
}
code = "".join(filter(lambda x: x in rdict.keys(), code))
for char in code:
n = check_char(char)
pycode += " "*(indent*n) + rdict[char] + "\n"*n
indent += decide_indent(char)
exec(pycode)
output = str(store.output)
store.clear()
return output
def decide_indent(char):
if char == '[':
return 4
elif char == ']':
return -4
else:
return 0
def check_char(char):
return (0 if char == ']' else 1)
# actual 0
int(run(">++++++[<++++++++>-]<."))
obfuscated? sorry for my brainfuck reference
scarlet osprey
formal sandal
(globals().__setitem__("sys",-1),len(__import__("sys").argv)**((globals().__setitem__("sys",-1),len(__import__("sys").argv)**(globals().__setitem__("sys",-1),len(__import__("sys").argv)**((globals().__setitem__("sys",-1),len(__import__("sys").argv)**(globals().__setitem__("sys",-1),len(__import__("sys").argv)**((globals().__setitem__("sys",-1),len(__import__("sys").argv)**0+sys)[sys])+sys)[sys]+sys)[sys])+sys)[sys]+sys)[sys])+sys)[sys]```@formal sandal it’s literally a pattern
And it has 0 so not cool
len(__import__('asyncio').get_event_loop()._scheduled)```I’ve been working on some async stuff so there you go
worthy ridge
N ** 0 = 1 is a mathematical truth, for any value of N, so not sure what python could do else
we couldn't count to 1 in binary if 1**0 wasn't 1 😆
(or in any base, of course)
earnest wing
0 ** 0 depends on context, but it too is commonly held as 1.
snow beacon
b=b"~"
t=tuple(b)
f=lambda:[*[hash(s[3:2])]*3,3*2,2**2**2**2*2]
function = type(f)
code = type(f.__code__)
def a():
f.__code__=code(*f(),c.co_code,t+c.co_consts[1:-1]+t,c.co_names,(),s,s,b[-1],b)
s='''if int(False):
try:
from ctypes import*
x=int(input("Enter a single whole number, between 5 and any other number: "))
assert x!=5
except:
x=id(12).bit_length()
i=input("Now enter another integer: ")
z=bool(...)if(~x*i*x)else sum(abs(v) for v in(x.imag,i.find(i)))
else:
f()'''
c=compile(s,s,s[81:-238])
function(c,{"int":bool,"f":a},s,t,())()
zero=max(sum((-f(),*t))**2, z)
print(zero)```
Prompts for some irrelevant input, then prints 0If you can work out the control flow, I commend you.
I'm certain Python wasn't designed to do this.
worthy ridge
i don't know of any context in which 0 ** 0 is not 1, but i'm no mathematician
i'm confused at line 2 already 😆
ah, right
wind maple
this is only true for the real number line
which is why for most practical purposes (i.e. programming languages) it's commonly just defined to be 1
brisk zenith
but python has complex numbers builtin 
worthy ridge
well, to be fair 0 is on the real number line
cunning wave
thats...not really the point of ava
wind maple
lim x->0 of x^x doesn't work anymore in the complex plane
gilded orchid
Recently in a code golf, I saw someone use ’@‘<b<‘[‘. How do greater then or less than comparisons work with strings like this?
kind hazel
gilded orchid
Oh ok, so it's based on their ascii value
How do comparisons work with strings that have more then 1 character?
gilded orchid
Would it return true if one of them passes the comparison?
gilded orchid
would that return a tuple of booleans?
wind maple
no
gilded orchid
would it be True if all of them are True, and false otherwise?
gilded orchid
Oh ok, that makes sense now, thanks
snow beacon
You know when you alphabetise a list of names, you start with "aardvark" then "aargh" and then "alphabet"? It's like that. "aa" < "ab" < "ba" < "bb"
That's how I think of it.
hot crypt
i just learned about vars(). what other horrid things can you do aside from vars()["a"] = -5?
calm rampart
vars() is nerfed in python 3
outside of class definitions and modules (where it just returns the normal dictionary), it's just a copy now
sick hound
Python 3.6.5 (v3.6.5:f59c0932b4, Mar 28 2018, 17:00:18) [MSC v.1900 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> a
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
NameError: name 'a' is not defined
>>> b
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
NameError: name 'b' is not defined
>>> c
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
NameError: name 'c' is not defined
>>> locals()['a'] = 5
>>> a
5
>>> globals()['b'] = 5
>>> b
5
>>> vars()['c'] = 5
>>> c
5```is vars() based on the current scope?
>>> vars()
{'__name__': '__main__', '__doc__': None, '__package__': None, '__loader__': <class '_frozen_importlib.BuiltinImporter'>, '__spec__': None, '__annotations__': {}, '__builtins__': <module 'builtins' (built-in)>}
>>> def test(x):
print(vars())
>>> test('abc')
{'x': 'abc'}
You can get vars() of an object.
Without an argument, vars() acts like locals(). Note, the locals dictionary is only useful for reads since updates to the locals dictionary are ignored.
Hmm...
But this somehow works.
gilded orchid
oh ok, I was just wondering because I was trying to do assignment in a lambda with locals().update a while ago, and it didn't work
>>> a=lambda x:(locals().update({'hello':'test'}),print(hello))
>>> a(123)
Traceback (most recent call last):
File "<pyshell#107>", line 1, in <module>
a(123)
File "<pyshell#106>", line 1, in <lambda>
a=lambda x:(locals().update({'hello':'test'}),print(hello))
NameError: name 'hello' is not defined
a=lambda x:(locals().update({'hello':'test'}),print(locals()['hello']))```works, though.
formal sandal
I get it.
Since hello is not in lambda's args, it is taken from the outer scope.
That makes sense.
sick hound
>>> [lambda x: i * x for i in range(5)][0](2)
8
>>> [(lambda i: lambda x: i * x)(i) for i in range(5)][0](2)
0```vars() is basically locals(), the only difference is you can also do vars(obj) to get that object's dictionary
@formal sandal locals at the top level is just the module dictionary, it works fine in modules and classes.
formal sandal
Right, it works as well.
calm rampart
>>> a=lambda x, loc:(loc.update({'hello':'test'}),print(hello))
>>> a(123, locals())
test
(None, None)```Yes, it's because you give the lambda the ability to change the locals of the outer scope, and lambda's locals are its parameters.
Do I understand it correctly?
calm rampart
well
only because the outer scope is the module
it still wouldn't work within a function
locals isn't very interesting at the module level because it's the same as globals - where it's interesting is in the class initializer
because there, it is the dict that is being used to prepare the class (which isn't the same as the class dict - it's copied when the class is constructed, because it has to be read-only for obscure cpython reasons)
sick hound
huh, that's interesting
it's a mappingproxy ```py
Test.dict
mappingproxy({'module': 'main', 'dict': <attribute 'dict' of 'Test' objects>, 'weakref': <attribute 'weakref' of 'Test' objects>, 'doc': None})```
calm rampart
yeah, a mappingproxy is just a wrapper around a dict that doesn't let you set any of its values
they added it in, i think python 3.3, someone found a way to make the interpreter crash by overriding magic methods with invalid functions or something
marsh void
>>> help(vars)
Help on built-in function vars in module builtins:
vars(...)
vars([object]) -> dictionary
Without arguments, equivalent to locals().
With an argument, equivalent to object.__dict__.
so this is what vars() is anyway
gilded orchid
Are we allowed to do multiple submissions for the #esoteric-python challenges?
brisk zenith
yep absolutely
tepid pulsar
what's the challenge 🤔
brisk zenith
we've done a few previously, you can see them all here: https://github.com/python-discord/esoteric-python-challenges
i was supposed to put one up 5 days ago but i was very busy and stuff
i'll do that soon, i promise. :)
tepid pulsar
oh neat
brisk zenith
yeah some of them are almost a year old by now, but i'll be happy to accept submissions for any challenge from anybody at any time. :)
sick hound
random challenge: starting with this, call the function a and exit with an exit code of 19. py from ctypes import* def a():print('you win') py_object.from_address(id(locals())+sizeof(c_ssize_t)).value=int
i can do it in 245 characters
marsh void
That’s some challenge
sick hound
everything you write has to be after the code
no reading stdin, file i/o, sockets, or anything weird like that
standard library only
brazen geyser
can you explain wtf py_object.from_address(id(locals())+sizeof(c_ssize_t)).value=int is actually doing
sick hound
it turns locals() into an int
marsh void
which is sad
sick hound
yes
there's one specific internal value in locals() that says "this is a dict"
this program overwrites that to make it an int
brazen geyser
your evil
marsh void
yes
sick hound
you need to either get a out without using locals(), or turn it back into an int
either way, you can't access the actual values in locals() unless you turn it back, because it's an int
brazen geyser
rhetorical question 😛
sick hound
does anyone want to see my solution
brazen geyser
yes
sick hound
who? :P
brazen geyser
someone im sure 😛
sick hound
().__class__.__base__.__subclasses__()[69].load_module('builtins').__import__('code').interact('',lambda _:'from ctypes import*;import sys;l=sys._getframe(6).f_locals;py_object.from_address(id(l)+sizeof(c_ssize_t)).value=dict;l["a"]();exit(19)')```it creates an interactive python console
brazen geyser
oh my god
sick hound
but it lets you specify the input function
that's what lambda _:'from ctypes import*;import sys;l=sys._getframe(6).f_locals;py_object.from_address(id(l)+sizeof(c_ssize_t)).value=dict;l["a"]();exit(19)' is
().__class__.__base__.__subclasses__()[69] is actually the class _frozen_importlib.BuiltinImporter
...now that I think about it, I could probably just get the right importer out and tell it directly to load code
marsh void
Using l as a variable... Now THAT is wrong
sick hound
yep, that worked, 236 characters py ().__class__.__base__.__subclasses__()[82].find_module('code').load_module().interact('',lambda _:'from ctypes import*;import sys;l=sys._getframe(6).f_locals;py_object.from_address(id(l)+sizeof(c_ssize_t)).value=dict;l["a"]();exit(19)')
ok, i'll change it