esoteric-python

[near gust]

!e

print(hex(hash(max(enumerate(reversed(list(enumerate(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(list(enumerate(list(enumerate(list(enumerate(list(enumerate(list(enumerate(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(list(enumerate(list(enumerate(list(enumerate(list(enumerate(list(enumerate(reversed(list(enumerate(list(enumerate(reversed(list(enumerate(list(enumerate(reversed(list(enumerate(list(enumerate(reversed(list(enumerate(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(list(enumerate(list(enumerate(list(enumerate(list(enumerate(list(enumerate(list(enumerate(list(enumerate(reversed(list(enumerate(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(reversed(list(enumerate(reversed(range(len(str(list()))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))

[night quarry]

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

0x1337133713371337

[near gust]

nice

[hybrid trail]

that is really cool

how long did it take to bruteforce the hash?

hmmm, it is exactly 64 bits...

[earnest wing]

this is awesome

[earnest wing]

The hash should be easy to compute if you know your data shape

It's not cryptographic

[digital dirge]

hash of an int is the int, so that's extra easy

!e

print(hex(hash(0x133713371337)))

[night quarry]

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

0x133713371337

[digital dirge]

Still cool 👍

[distant wave]

!e ```py
x=9
s=[1]x
v=[s[p:x:p]:=~-x//p[0]for p in range(3,x,2)if sorted({*str(p)})==[*str(p)]]
print(2,*v,sep="\n")

I guess I can't star a :=?

[night quarry]

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

001 |   File "/home/main.py", line 3
002 |     v=[s[p:x:p]:=~-x//p*[0]for p in range(3,x,2)if sorted({*str(p)})==[*str(p)]]
003 |        ^^^^^^^^
004 | SyntaxError: cannot use assignment expressions with subscript

[distant wave]

bleh

[long fulcrum]

most annoying thing ever 😭

[bronze sorrel]

the hell is this channel, the word estoric alone tells me not to try understand

[grave grail]

the painful part lul

[rugged sparrow]

!e ```py
import gc

class magic:
def iter(self):
yield magic
for obj in gc.get_objects():
if isinstance(obj, frozenset) and magic in obj:
yield obj
return

weird = frozenset(magic())

print(f"{weird = }")
print(f"{weird in weird = }")
print(f"{hash(weird) = }")

[night quarry]

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

001 | weird = frozenset({<class '__main__.magic'>, frozenset(...)})
002 | weird in weird = True
003 | hash(weird) = -7903945776645685223

[versed eagle]

huh

[karmic pumice]

damn, the hash of a paradox

[rugged sparrow]

Recursive frozenset

[versed eagle]

so cool

[rugged sparrow]

Similar trick works for tuples

!e ```py
import gc

class magic:
def length_hint(self):
return 1

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

weird = tuple(magic())
print(f"{weird = }")
print(f"{weird in weird = }")

[night quarry]

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

001 | weird = ((...),)
002 | weird in weird = True

[rugged sparrow]

Trying to think of a way to make a recursive frozen set without gc

[fleet bridge]

I wonder how that even works
how does that know hash of weird if it is not even fully created?

[rugged sparrow]

frozenset must account for recursion in its hash?

[fleet bridge]

why would it?

[rugged sparrow]

Tuples hash function doesn't

[rugged sparrow]

No clue, I'll look tho

[fleet bridge]

hash(x) == hash({x})

[versed eagle]

oh thats delicious

[rugged sparrow]

Ah, the hash of a frozenset is cached

So when it's added, it's hash is computed and stored

[fleet bridge]

ah, now I get it

[karmic pumice]

!e

import gc

class magic:
  def __iter__(self):
    yield magic
    for obj in gc.get_objects():
      if isinstance(obj, frozenset) and magic in obj:
        yield obj
        return
        
for _ in range(2):
  print(hash(frozenset(magic())))

:( the paradox is not consistently hashed

[night quarry]

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

001 | -2598414521264488544
002 | 7128057917496661008

[fleet bridge]

so this hash is not really consistent
if you recompute hash again, you will get different value

[fleet bridge]

that makes sense
you include first set into second, and it changes the hasg

run gc.collect() in the loop, it should produce same hashes

!e

import gc

class magic:
  def __iter__(self):
    yield magic
    for obj in gc.get_objects():
      if isinstance(obj, frozenset) and magic in obj:
        yield obj
        return
        
for _ in range(2):
  gc.collect()
  print(hash(frozenset(magic())))

[night quarry]

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

001 | -4958887202234241987
002 | -4958887202234241987

[fleet bridge]

yay, I eyeballed it right

[karmic pumice]

hmh
what is not being collected though?

[fleet bridge]

set references itself, so it is not gc'd right away

and magic is in it, so you yield it into second set

[fleet bridge]

!e

import gc

class magic:
  def __iter__(self):
    yield magic
    for obj in gc.get_objects():
      if isinstance(obj, frozenset) and magic in obj:
        yield obj
        return
        
for _ in range(2):
  print(frozenset(magic()))

[night quarry]

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

001 | frozenset({<class '__main__.magic'>, frozenset(...)})
002 | frozenset({<class '__main__.magic'>, frozenset({<class '__main__.magic'>, frozenset(...)})})

[fleet bridge]

ah, i missed a detail
when you found the set, you return immediately

so the first set contains itself
and the second one finds first one first, and then ends. so the seconds one also contains the first

idk how that changes hash though

[humble beacon]

If your aim is to create a recursive tuple, you can do it more easily

[rugged sparrow]

I mean there are plenty of ways to make a recursive tuple

[rugged sparrow]

You can also do tuple->list-tuple

But its more fun to do tuple->tuple

[serene stratus]

You can use marshal load but I'd consider that cheating

[rugged sparrow]

Also only works in the versions where they have that bug fixed lol

!e could obviously just do this: ```py
from fishhook import hook

@hook(set)
def hash(self):
return hash(frozenset(self))

weird = set()
weird.add(weird)
print(f"{weird = }")```

[night quarry]

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

weird = {set(...)}

[rugged sparrow]

But that's just fishhook doing the magic behind the scenes

[long fulcrum]

is there a shorter way i can be doing this to generate primes

x=7**9
s=[1]*x
for p in range(2,x):
    if s[p]:s[p::p]=~-x//p*[0]```

it cant be much slower unfortunately

[hybrid trail]

what does your code do

oh sieve i guess

you just don't do anything with the primes?

[long fulcrum]

i check if each digit is larger than the last and if so i print

[hybrid trail]

hmmm, i struggle with that hole a lot

[quartz wave]

wait what hole is this

[hybrid trail]

seems to be ascending primes

[long fulcrum]

yeah

[hybrid trail]

68 bytes is so short I bet that they're not doing actual primality checking

probably some other conditions that just happen to work for the 100 outputs

[sick hound]

Does anybody know if a regex match strings can be compared directly without checking if they are equal for all possible strings? maybe there is a way to reduce them to finite automata and compare the automata directly?

I want to know if regex expression r'x' and 'x' are equal for all possible strings x when passed to re.findall(x, text)

[night quarry]

xy-problem

The XY problem can be summarised as asking about your attempted solution, rather than your actual problem.

Often programmers will get distracted with a potential solution they've come up with, and will try asking for help getting it to work. However, it's possible this solution either wouldn't work as they expect, or there's a much better solution instead.

For more information and examples, see http://xyproblem.info/.

[versed eagle]

i dont think this is an instance of the xy problem

[fleet bridge]

!pypi interegular

[night quarry]

interegular v0.3.3

a regex intersection checker

Released on <t:1704582082:D>.

[restive void]

Regexes can't be reduced to finite automata (at least not all of them can). I'm confused about the r'x' and 'x' example — raw strings are only a parsing thing. But you just mean checking whether two regexes will match the exact same set of strings, right?

Or do you literally mean "Can a regex ever match something else when I add/remove the r prefix?", in that case the answer is yes (minimal example I can think of: re.findall(r"\b", "x") == ["", ""], but re.findall("\b", "x") == [])

[fleet bridge]

if a regex match strings can be compared directly
you can compare contents of matches

[sick hound]

yes, but I need to compare the equivalence between regexes themselves, not compare the results of matches, which do not prove the general case for arbitrary strings

[fleet bridge]

why do you want that? are absolutely sure that you need to know that?

[sick hound]

because I asked in #python-discussion and nobody knows for sure if r'y' is equivalent to 'y' for all arbitrary regexes y where y is a placeholder

[fleet bridge]

r-prefix has nothing to do with regexes

[sick hound]

it's used in regexes

[fleet bridge]

nope
it has nothing to do with regexes

[sick hound]

why is it used in regexes, in this case?

[fleet bridge]

it is used to disable escaping in strings, which is convenient when there are a lot of backslashes in strings

and it just that happens that regexes have a lot of backslashes

[fleet bridge]

you already got your answer in #python-discussion
'\b' and '\\b' are different regexes, they will match different strings

[sick hound]

import re
text = '  \n foo  \n  '

match_0 = re.findall('[^\n\r\t\f\v ]', text, re.DEBUG)
match_1 = re.findall(r'[^\n\r\t\f\v ]', text, re.DEBUG)

print(match_0==match_1) # True
print(match_0)   # ['f', 'o', 'o']
print(match_1) # ['f', 'o', 'o']
print(id(match_0)==id(match_1)) # False

Does this mean r'y' and 'y' regexes are equivalent for all arbitrary strings and for all arbitrary regexes?

[fleet bridge]

print(id(match_0)==id(match_1)) # False
how is this related to the question?

[sick hound]

ignore this bit

[fleet bridge]

nope
once again, '\b' and r'\b' are different regexes

[sick hound]

interesting counterexample. Would you say regex r'[y]' is equivalent to '[y]' for arbitrary strings and regexes?

[lofty comet]

It's the same string

[fleet bridge]

interesting counterexample
you already seen it in #python-discussion

Would you say regex r'[y]' is equivalent to '[y]' for arbitrary strings and regexes?
no
consider y = x]\b[y

for arbitrary strings and regexes?
i have no idea what you mean exactly by that

[sick hound]

i see, the instructions are different. so the r does change the regex after all in many cases?

import re
text = '  \n foo  \n  '

match_0 = re.search('[x]\b[y]', text, re.DEBUG)
print("--------------")
match_1 = re.search(r'[x]\b[y]', text, re.DEBUG)

print(match_0==match_1) # True
print(match_0)
print("--------------")
print(match_1)

"""

LITERAL 120
LITERAL 8
LITERAL 121

 0. INFO 12 0b11 3 3 (to 13)
      prefix_skip 3
      prefix [0x78, 0x8, 0x79] ('x\x08y')
      overlap [0, 0, 0]
13: LITERAL 0x78 ('x')
15. LITERAL 0x8 ('\x08')
17. LITERAL 0x79 ('y')
19. SUCCESS
--------------
LITERAL 120
AT AT_BOUNDARY
LITERAL 121

 0. INFO 8 0b1 2 2 (to 9)
      prefix_skip 1
      prefix [0x78] ('x')
      overlap [0]
 9: LITERAL 0x78 ('x')
11. AT UNI_BOUNDARY
13. LITERAL 0x79 ('y')
15. SUCCESS
True
None
--------------
None

"""

I wonder why the results show as equal even though regex used different instructions to arrive at the result in both cases

[fleet bridge]

did you even looked at the results?

they are empty, no wonder they are equal

two completely unrelated regexes can both dont match, and results would be equal

[sick hound]

just because both the results are None doesn't mean regexes should be equal with __eq__ if they do different things

[fleet bridge]

how is __eq__ involved in this?
None == None
there are no regex object (whatever that means) in your code

!e print(None == None)

[night quarry]

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

True

[sick hound]

you are too focused on this particular case of results. I am asking if there is a way to check if a regex is equivalent to a different regex. e.g. we know that r'y' and 'y' are different via counterexample. Now, how would you check if \S is equivalent to [^\n\r\t\f\v ] or r'[^\n\r\t\f\v ]'. We are back to square one, because we have no general checker (unless you know a tool that can directly compare regexes?)

[fleet bridge]

i linked the tool above

[sick hound]

cant find a link anywhere in your replies

[fleet bridge]

cant help you with that

the tool works great
it can build FSM's from regexes, and then you can compare FSM's

[sick hound]

sorry I missed a section of replies above, found it

[sick hound]

import interegular

parse_pattern_0 = interegular.parse_pattern('[^\n\r\t\f\v ]')
parse_pattern_1 = interegular.parse_pattern(r'[^\n\r\t\f\v ]')
fsm_0 = parse_pattern_0.to_fsm()
fsm_1 = parse_pattern_1.to_fsm()
print(fsm_0)
"""
  name final? \t \n \x0b \x0c \r   anything_else 
-------------------------------------------------
* 0    False                       1
  1    True
"""
print(fsm_1)
"""
  name final? \t \n \x0b \x0c \r   anything_else 
-------------------------------------------------
* 0    False                       1
  1    True
"""
direct_comparison = interegular.compare_patterns(parse_pattern_0, parse_pattern_1)
print(list(direct_comparison)) # [(Pattern(options=(_Concatenation(parts=(_CharGroup(chars=frozenset({'\t', ' ', '\n', '\x0c', '\r', '\x0b'}), negated=True),)),), added_flags=<REFlags: 0>, removed_flags=<REFlags: 0>), Pattern(options=(_Concatenation(parts=(_CharGroup(chars=frozenset({'\t', ' ', '\n', '\x0c', '\r', '\x0b'}), negated=True),)),), added_flags=<REFlags: 0>, removed_flags=<REFlags: 0>))]

Python thinks they are the same, even though we established they can have different results on identical text searches.
Did I compare the regexes correctly, or should I have used:

parse_pattern_0 = interegular.parse_pattern("'[^\n\r\t\f\v ]'")
parse_pattern_1 = interegular.parse_pattern("r'[^\n\r\t\f\v ]'")

?

[fleet bridge]

Python thinks they are the same
python has no opinion on this

we established they can have different results on identical text searches
did we? do you have an example where these two regexes produce different results?

parse_pattern_0 = interegular.parse_pattern("'[^\n\r\t\f\v ]'")
parse_pattern_1 = interegular.parse_pattern("r'[^\n\r\t\f\v ]'")
``` these two are obviously different, no tool is needed for that

[sick hound]

'[x]\b[y]' and r'[x]\b[y]' give different results

[fleet bridge]

🤦‍♂️

[fleet bridge]

your code shows equivalence of [^\n\r\t\f\v ] and [^\\n\\r\\t\\f\\v ]
we didnt establish that they are different, because they are equivalent

[sick hound]

you think there exists no "nice" string that will fail to yield a different search result between these 2 regex patterns?

[fleet bridge]

if you use the same regex, result will always be the same

otherwise it would be weird if the same regex produced different results

[sick hound]

parse_pattern_0 = interegular.parse_pattern("'[^\n\r\t\f\v ]'")
parse_pattern_1 = interegular.parse_pattern("r'[^\n\r\t\f\v ]'")

If I specify my regex patterns to be like this, the patterns are no longer deemed idential by interegular

[fleet bridge]

of course

because they are different

first one matches only strings that start with ', second one matches only strings that start with r

[sick hound]

r is not supposed to be a string to find. It's a symbol that makes the regex pattern nice

[fleet bridge]

!e print("r")

[night quarry]

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

r

[fleet bridge]

it is just a letter in this context

[sick hound]

not supposed to be a letter. r makes the string a repr, right?

[fleet bridge]

"r''" is not the same as r''

[fleet bridge]

strings here have no prefix before them

we should stop, this is clearly off-topic to this channel

open a help thread if you need further help with understanding of string literals

#❓｜how-to-get-help

[versed eagle]

well no

r'\\y' and '\\y' are two different things

for example

thats trivial

but for an arbitrary regex determining whether they're equivalent is not trivial

[versed eagle]

no, it makes it a raw string

[versed eagle]

obviously, they are different strings

but generally its not trivial to determine whether two regexes match the same set of strings i think

if it were there wouldnt be a library for it lol

[sick hound]

yeah I think the general case of this problem is undecidable for arbitrary regex and strings. But doesnt library interegular offer limited insight into some simple regex comparisons like '[^\n\r\t\f\v ]' and r'[^\n\r\t\f\v ]'?

[versed eagle]

not undecidable

well it depends maybe

its certainly not undecidable for actual regular expressions, which match against regular languages

but "actual regular expressions" python's regexes are not

nor most modern regex implementations

since they include things like backreferencing capture groups

[sick hound]

!e print("sigma sigma")

[night quarry]

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

sigma sigma

[sick hound]

!r

!r print ( "sigma sigma" )

!e print("у")

[night quarry]

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

у

[sick hound]

the reason why I think arbitrary regex equivalence comparison is undecidable is that with complex enough regex patterns the problem will be able to be reduced to a halting problem. Wouldn't more and more recursive and self-referential behaviour appear for more and more complex regexes, making the problem just as hard as halting problem?

[fleet bridge]

vanilla regex always halts

[sick hound]

isn't regex turing complete?

[fleet bridge]

no

[sick hound]

ok, that's interesting

[sick hound]

but don't regex backreferences require a turing machine to be fully supported?

[fleet bridge]

your computer is a Turing machine
but that does not mean that anything your computer does is Turing complete

[versed eagle]

recursive and self-referential behaviour are not part of actual regular expressions

formal regular expressions, anyway

with formal regular expressions, you can enumerate the set of strings that each describes (using some sentinel for * instead of expanding it) and then compare those sets

[proper vault]

regular expressions as implemented in Python do not necessarily describe a regular language, so that's not all that applicable

[versed eagle]

well yes, you're repeating what i've said earlier

[hybrid trail]

Another installment in the programming with only a-z and parens: given a string with space separated integers, compute their sum (eval and exec disallowed as well)

[fleet bridge]

if there is no restriction, is there a shorter way to do this?:

s='1 2 3'
print(sum(map(int,s.split())))

[hybrid trail]

print(eval(s.replace(*' +')))```

[hybrid trail]

my best is currently 335 (using input)

[quartz wave]

we can create an among us in less than 50 characters with those restrictions but can't even do that on this challenge??

world works in mysterious ways i guess

[hybrid trail]

maybe you can ;)

[quartz wave]

no idea where to start-

[digital dirge]

Positive ints of arbitrary size?

[hybrid trail]

Yes

[proper vault]

@hybrid trail is A-Z allowed?

[hybrid trail]

nope

[proper vault]

Noted

[proper vault]

!e

from operator import methodcaller
from operator import itemgetter
print(sum(int(x)for x in methodcaller(itemgetter(len(str(help)))(dir(str)))('1 2 3 4')))

Pretend the string is input()

[night quarry]

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

10

[proper vault]

Not sure if the imports are cheating

[hybrid trail]

very interesting method, but it does have forbidden chars space and newline

[proper vault]

Oh space is bad

That makes this approach much harder

[gleaming linden]

No commas either?

[quartz wave]

yup

[grave grail]

Challenge: write a prime number generator with only 5 [a-zA-Z0-9 \n] character, gl

[karmic pumice]

as in, len(set(code)) <= 5? or 5 characters in general, len(code) <= 5?

[quartz wave]

P

done

no language specified so i assumed math

[grave grail]

*Challenge: write a prime number generator in python with only 5 [a-zA-Z0-9 \n] character, with no import, gl

[earnest wing]

prime

[hard spoke]

2

[unborn sphinx]

not sure if that can go here, but why not

[karmic pumice]

!e

class _arg:
  def __getitem__(self, n):
    new = _arg()
    new.n = n
    return new
  def __mul__(self, val):
    return lambda *args: args[self.n] * val
arg = _arg()
class fn:
  def __init__(self, _):
    ...
  def __rshift__(self, thing):
    return matcall(thing)
class matcall:
  def __init__(self, f):
    self.f = f
  def __matmul__(self, args):
    if not isinstance(args, tuple):
      args = (args, )
    return self.f(*args)

double = fn(int) >> arg[0] * 2

println = matcall(print)

println @ f"{double @ 4}"

[night quarry]

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

8

[unborn sphinx]

bro has done this in 5 seconds 😭

[grave grail]

It doesn't generate

[karmic pumice]

its pretty wrong if one tries to use it for something else, e.g. a function of 2 arguments would not work (something like arg[0] * arg[1])
i basically just replicated it for this specific example

[unborn sphinx]

ik, ik

i'll do several args later today

but also i just realised that my initial idea doesnt work, like add @ 2, 3 evaluates to... a tuple of add @ 2 and 3, which i forgor about

[karmic pumice]

i think it would be cool to make it curried, add @ 2 @ 3
you can use the amount of thing in fn to know how much args does it have

[unborn sphinx]

yep, but then... oh wait, it does add @ 2 first, right? or 2 @ 3 ?

[karmic pumice]

i think its (add @ 2) @ 3

[unborn sphinx]

should be that way cuz... same operator same priority, ig

[restive void]

Pretty sure it's left-associative

[hard spoke]

why not? it evaluates to a prime number

[karmic pumice]

enddo functors?? monads?

!e

class q:
  def __init__(self, x):
    self.x = x
  def __matmul__(self, other):
    return q((self.x, other))
print((q(1) @ 2 @ 3).x)

[night quarry]

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

((1, 2), 3)

[grave grail]

*Challenge: write a prime number generator that infinitvely generate prime number and print it until Ctrl+C in python with only 5 [a-zA-Z0-9 \n] character with no import, gl

[versed eagle]

with only 5 characters?

[grave grail]

Well, you can use unlimited amount of character that isn't those

[restive void]

Instructions unclear, please provide regex.

Like I have to pick five from this set but can use them as often as I like, or I get five uses of those chars total?

[karmic pumice]

unicode abuse time (nfkc normalization)

[earnest wing]

truly ascii with a top hat

[grave grail]

Five use of those character in total

[grave grail]

Can I not have unicode abuse?

Also btw, eval/exec is allowed

:)

[restive void]

Ah, too bad. I have a solution without exex/eval, relatively short, but using those chosen chars several times.

[grave grail]

^(?:[^a-zA-Z0-9 \n]*.[^a-zA-Z0-9 \n]*){,5}$

[grave grail]

Well, the planned way of doing the challenge is how to assemble a string with 1 of those character, and use either eval/exec to run it

But other way are welcome :)

[restive void]

45 chars, with the other interpretation:

iter((int()**int()+int()**int()).__int__,int)

(itern)

[grave grail]

That's a nice one if it work (probably test it later

[restive void]

!e probably not what you meant, but it does infinitely produce primes

x = iter((int()**int()+int()**int()).__int__,int)
print(next(x))
print(next(x))
print(next(x))
print("...")

[night quarry]

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

001 | 2
002 | 2
003 | 2
004 | ...

[grave grail]

...

*a prime number generator that infinitely produce unique prime that wasn't generated before

[craggy hamlet]

most primes have been generated already

(see https://en.wikipedia.org/wiki/List_of_prime_numbers)

[grave grail]

*a prime number generator that infinitely produce incremental prime number from 2(inclusive)

[craggy hamlet]

!e sure!

x = iter(x for x in (n for n in range(2, 10**6) if all(n % d for d in range(2, int(n**0.5) + 1))))
print(next(x))
print(next(x))
print(next(x))
print(next(x))
print(next(x))

[night quarry]

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

001 | 2
002 | 3
003 | 5
004 | 7
005 | 11

[grave grail]

You missed the outer part of the question

[craggy hamlet]

i dont see any restraints

https://tenor.com/view/cat-smile-funny-evil-smile-cat-cat-meme-gif-5500986885420487358

[grave grail]

Basically now:

Challenge: write a prime number generator that infinitely produce incremental prime number from 2(inclusive) and print it until Ctrl+C in python with only 5 [a-zA-Z0-9 \n] character with no import, gl

[hard spoke]

i don't think there's much of anything you can write with only 5 characters

[versed eagle]

exec with %c is boring though

[grave grail]

Well, that's what I want lul

[earnest wing]

why bother with a challenge if you only allow one solution

just show it off

[grave grail]

Well, you might able to find other way

Such as lambda suggest, unicode abuse, I didn't block that (just don't like it

Because it is still reasonably fun

[vast wave]

is that brother calm

[unborn sphinx]

@karmic pumice there i go (also i couldnt come up with a name so coffeesnake it is 💀)

[grave grail]

Nice

[fleet bridge]

can you golf this?

import collections.abc
import typing as t

def collapse_numbers_into_ranges(
    nums: collections.abc.Iterable[int],
) -> collections.abc.Iterator[list[int]]:
    '''
    takes an iterable of ints (not necessrily in sorted order)
    returns an iterable of [x,y] or [x] (for single-element ranges)
    '''
    # reference implementation:
    import itertools

    rng = []
    for x in itertools.chain(sorted(nums), [t.cast(int, ...)]):
        if not rng:
            rng = [x, x]
            continue
        if x == rng[-1] + 1:
            rng[-1] = x
        else:
            yield rng if rng[0] != rng[1] else [rng[0]]
            rng = [x, x]

# tests:
def fmt(nums: list[int]) -> str:
    return ','.join('-'.join(map(str, rng)) for rng in collapse_numbers_into_ranges(nums))
assert fmt([]) == ''
assert fmt([1]) == '1'
assert fmt([1, 2]) == '1-2'
assert fmt([1, 2, 3]) == '1-3'
assert fmt([1, 2, 3, 5]) == '1-3,5'
assert fmt([1, 3, 5]) == '1,3,5'
assert fmt([1, 3, 5, 6, 7]) == '1,3,5-7'
assert fmt([1, 7]) == '1,7'
assert fmt([5, 4, 7, 6, 3]) == '3-7'
assert fmt([5, 4, 7, 3]) == '3-5,7'

[subtle viper]

def collapse_numbers_into_ranges(s):
 t=z=''
 for x in*sorted(s),2j:
  if''!=t!=x-1:yield(z,t)[z==t:];z=x
  t=x
``` untested

[fleet bridge]

>>> collapse_numbers_into_ranges([1, 3,4,5,6]) // list
[('', 1), (3, 6)]
``` almost works i guess

[subtle viper]

def collapse_numbers_into_ranges(s):
 t=z=''
 for x in*sorted(s),2j:
  if''!=t!=x-1:yield(z,t)[z in(t,''):];z=x
  t=x

actually no wait-

[fleet bridge]

>>> collapse_numbers_into_ranges([1,2]) // list
[(2,)]
>>> collapse_numbers_into_ranges([1, 3,4,5,6,7]) // list
[(1,), (3, 7)]

[subtle viper]

def collapse_numbers_into_ranges(s):
 t=z=''
 for x in*sorted(s),2j:
  if''!=t!=x-1!=(yield(z,t)[z==t:])or''==z:z=x
  t=x

[fleet bridge]

works 🎉

good job, here is a frog for you: 🐸

[subtle viper]

ty ^^

~~1b down:

def collapse_numbers_into_ranges(s):
 t=z=''
 for x in*sorted(s),2j:
  if''!=t<x-1!=(yield(z,t)[z==t:])or''==z:z=x
  t=x
```~~ nvm.

assert fmt(func([1, 2, 3, 3, 4, 5])) == '1-3,3-5' fails- :<

[fleet bridge]

idk if repeated numbers should be allowed

originally I needed that function to nicely display set of numbers, so there were no duplicates possible

[subtle viper]

well the -1b version should work if so-

[clear venture]

!e

a = [1,2,3,4,100]
b = a.copy()
[next(...for a[i]in b[len(a)-i-1::-1])for i in range(len(a))]

print(a)

[night quarry]

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

[100, 4, 3, 2, 1]

[quartz wave]

121b -> 114b py def collapse_numbers_into_ranges(s,z=.5): for x in*sorted(s),2j: if z%1or t!=x-1!=(yield(z,t)[z==t:]):z=x t=x can be 113b with t<x-1

[unborn sphinx]

@karmic pumice here's more

[fleet bridge]

>>> 8 // 2 .__mul__ // print
16

[real leaf]

Hey all, is there a "division" of esoteric python for minimum lines?

I'm not sure it's considered esoteric, but it feels like a challenge that fits in.

[fleet bridge]

that is called "code golfing" - minimizing code size (not the number of lines, but rather the number if characters)

[real leaf]

I juset found out about the compile() function python has, and was wondering if you could make any single program one line by just compressing everything into one string and compiling, then executing it?

[karmic pumice]

you can do that by exec'ing a string, yes
compile is unnecessary here

[real leaf]

this seems like a more worthy use of time, yeah.

[fleet bridge]

you can indeed compress anything into one line
exec('multiline code goes here') is even simpler than your suggestion
but you even can do that without using exec/evval/compile/etc...

[real leaf]

oh, cool.

[real leaf]

hmm

alrighty

thank you both :)

[fleet bridge]

[long fulcrum]

there's a 3:1 and 2:1 packer here
https://code.golf/wiki/langs/python

[versed eagle]

minimal lines (1) is easily achieved via exec et al.

[grizzled rose]

Hi

!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.

[grizzled rose]

from pygame import*;from random import*;init();j=display;z=j.set_mode((400,300));c=time.Clock();q=255;t=10;r=draw.rect;S=R,L,D,U=range(K_RIGHT,K_UP+1);s,d,f=[(100,50),(90,50),(80,50)],R,(n:=lambda:(randint(0,39)*t,randint(0,29)*t))()
while 1:
 for e in event.get():
  if e.type==QUIT:quit()
  if e.type==KEYDOWN:d=k if(k:=e.key)in S else d
 x,y=s[0];x+=(d==R)*t-(d==L)*t;y+=(d==D)*t-(d==U)*t;s=[(x,y)]+s[:-1];f=n()if s[0]==f else f
 if not(0<=x<400 and 0<=y<300)or s[0]in s[1:]:break
 z.fill([q]*3);[r(z,(0,q,0),(*p,t,t))for p in s];r(z,(q,0,0),(*f,t,t));j.update();c.tick(t)
quit()

Can anybody help me modify it to be lessweight to be only 400 bytes?

[hybrid trail]

do you need if e.type==QUIT:quit()?

[hybrid trail]

also your snake does not seem to get longer, not sure if that is intentional

[gleaming linden]

I think you can do this for event handling py for e in event.get():e.type==QUIT==quit();e.type==KEYDOWN==(d:=[d,e.key][e.key in S])

[quartz wave]

do you really need to do quit() at the end of the program?

never did pygame before so i don't really know- :p

[quartz wave]

shaved off 29 bytes but i'm not sure about making it to sub-400 =-=

[quartz wave]

*40b now ```py
from pygame import*
from random import*
init()
j=display
z=j.set_mode((400,300))
c=time.Clock()
q=255
t=10;T=t,t
r=draw.rect
S=R,L,D,U=79,80,81,82
s=(100,50),(90,50),(80,50)
d=R
g=randrange
f=g(40)*t,g(30)*t
while 1:
for e in event.get():e.type==QUIT!=quit();e.type==KEYDOWN!=(d:=[d,k:=e.key-2**30][k in S])
x,y=s[0];x+=(d==R)*t-(d==L)*t;y+=(d==D)*t-(d==U)*t;k=x,y;*s,l=k,*s
if k==f:s+=l,;f=g(40)*t,g(30)*t
if not 300>y>=0<=x<400 or k in s[1:]:break
z.fill([q]*3);[r(z,(0,q,0),p+T)for p in s];r(z,(q,0,0),f+T);j.update();c.tick(t)
quit()

is discord highlighting broken-

.

gosh

i just found a new bug with highlighting :D

*g ```py
from this import *

this is out of the codeblock!!

whyy--

[karmic pumice]

that doesnt work for me :( maybe its pygame-ce vs pygame diff

[quartz wave]

wait i found a problem-

- d=r
+ d=R

[karmic pumice]

yep

[quartz wave]

it's also possible to have an apple spawn with collision to the snake but i'm not adding that check in- :p

the snake growing was just added in too..

[karmic pumice]

its so minimalistic and cool with rounded borders and stuff

[versed eagle]

it renders correctly on mobile

[hybrid trail]

482 ```python
from pygame import*
from random import*
j=display
z=j.set_mode((400,300))
t=10;T=t,t
r=draw.rect
S=R,L,D,U=b'OPQR'
f,*s=(100,50),(90,50),(80,50)
d=R
g=randrange
while 1:
for e in event.get():e.type==QUIT!=quit();e.type==KEYDOWN<(d:=[d,k:=e.key&255][k in S])
x,y=s[0];x+=(d<L)*t-(d==L)*t;y+=(d==D)*t-(d>D)*t;k=x,y;*s,l=k,*s
if k==f:s+=l,;f=g(40)*t,g(30)*t
if not 300>y>-1<x<400or k in s[1:]:break
z.fill(-1);[r(z,5**9,p+T)for p in s];r(z,'red',f+T);j.flip();time.Clock().tick(t)

[floral meteor]

It's an old screenshot. Sometimes I am confused by the thoughts of my past mind

[rugged sparrow]

is that interned strings modification to cause 'f' to equal ''?

[floral meteor]

yes

another gem from old screenshots related to this channel

[hybrid trail]

seems like init() is not even needed for the program 🤔

[gilded briar]

How to get Discord spam tool

[versed eagle]

??????

lmao

[fleet bridge]

set_mode() calls init()

[signal dock]

Please read our #rules and #code-of-conduct : we don't help with 'spam tools' (see Rule 5), and keep conversations relevant to the channel topic.

[quartz wave]

463 ```py
import pygame as p,random as r
j=p.display
z=j.set_mode((400,300))
t=10;T=t,t
g=r.randrange
r=p.draw.rect
S=d,L,D,U=b'OPQR'
f,*s=(100,50),(90,50),(80,50)
while 1:
for e in p.event.get():a=e.type;a<257!=quit();a<769<(d:=[d,k:=e.key&255][k in S])
x,y=s[0];x+=(d<L)*t-(d==L)*t;y+=(d==D)*t-(d>D)*t;k=x,y;*s,l=k,*s
if k==f:s+=l,;f=g(40)*t,g(30)*t
(300>y>-1<x<400)<1>r;k in s[1:]>r;z.fill(-1);[r(z,5**9,p+T)for p in s];r(z,'red',f+T);j.flip();p.time.Clock().tick(t)

[hybrid trail]

if we're assuming we only need to handle the two events you could remove the 769 condition

also could you could replace quit() with something like Z to trigger a namerror?

[quartz wave]

second i can do

but first one doesn't make sense to me-

[hybrid trail]

hmm, you're right

i assumed the only events were quit and keydown 🤦

[gleaming linden]

Use randint instead of randrange?

Nvm randint takes 2 args

[gleaming linden]

are the >rs meant to nameerror?

[hybrid trail]

not namerror, r is defined (it is aliased to random)

but it will error, yes

[quartz wave]

i just thought NameError was too boring :p

[gleaming linden]

oh right comparisons can error

import pygame as p,random as r
j=p.display
z=j.set_mode((400,300))
t=10
g=r.randrange
r=p.draw.rect
S=d,L,D,U=b'OPQR'
f=t,5;*s,=(9,5),(8,5)
while 1:
 for e in p.event.get():a=e.type;a<257!=quit();a<769<(d:=[d,k:=e.key&255][k in S])
 x,y=s[0];x+=d<L;x-=d==L;y+=d==D;y-=d>D;k=x,y;*s,l=k,*s
 if k==f:s+=l,;f=g(40),g(30)
 (30>y>-1<x<40)<1>r;k in s[1:]>r;z.fill(-1);[r(z,5**9,(x*t,y*t,t,t))for x,y in s];r(z,5**15,(f[0]*t,f[1]*t,t,t));j.flip();p.time.Clock().tick(t)
``` untested 461

[quartz wave]

seems to work

[hybrid trail]

change 'red' to 5**15 for another byte

[quartz wave]

it's 0b..

[hybrid trail]

oops

[gleaming linden]

added it anyway because it looks cooler

[hybrid trail]

doesn't work for me

    f=t,5;*s=(9,5),(8,5)
          ^^
SyntaxError: starred assignment target must be in a list or tuple

[gleaming linden]

oh oops

hmm

[quartz wave]

455 459?? ```py
import pygame as p,random as r
j=p.display
z=j.set_mode((400,300))
t=10
g=r.randrange
r=p.draw.rect
S=d,L,D,U=b'OPQR'
f=t,5;s=(9,5),(8,5)
while 1:
for e in p.event.get():a=e.type;a<257!=quit();a<769<(d:=[d,k:=e.key&255][k in S])
x,y=s[0];x+=d<L;x-=d==L;y+=d==D;y-=d>D;k=x,y;s,l=k,s
if k==f:s+=l,;f=g(40),g(30)
(30>y>-1<x<40)<1>r;k in s[1:]>r;z.fill(-1);[r(z,5**9,(xt,yt,t,t))for x,y in s];r(z,'red',(f[0]*t,f[1]*t,t,t));j.flip();p.time.Clock().tick(t)

[hybrid trail]

change the definition of s to just (9,5),(8,5) for -2

[quartz wave]

i somehow got OverflowError: Python int too large to convert to C long

oh

it's the 5**15

[gleaming linden]

so what's the best now?

[quartz wave]

i think

[gleaming linden]

ah

[quartz wave]

oh my god it's the 5**15.

[hybrid trail]

453

import pygame as p,random as r
j=p.display
z=j.set_mode((400,300))
t=10
g=r.randrange
r=p.draw.rect
S=d,L,D,U=b'OPQR'
f=t,5;s=(9,5),(8,5)
while 1:
 for e in p.event.get():a=e.type;a<257>b;a<769<(d:=[d,k:=e.key&255][k in S])
 x,y=s[0];x+=d<L;x-=d==L;y+=d==D;y-=d>D;k=x,y;*s,l=k,*s
 if k==f:s+=l,;f=g(40),g(30)
 (30>y>-1<x<40)<1>r;k in s[1:]>r;z.fill(-1);[r(z,5**9,(x*t,y*t,t,t))for x,y in s];r(z,'red',(f[0]*t,f[1]*t,t,t));j.flip();p.time.Clock().tick(t)

[gleaming linden]

alr

[hybrid trail]

changed 5**15 back to 'red'

[quartz wave]

oh

i accidentally reverted the !=quit() -> >b change

[gleaming linden]

import pygame as p,random as r
j=p.display
z=j.set_mode((400,300))
t=10
g=r.randrange
r=p.draw.rect
S=d,L,D,U=b'OPQR'
h=f=t,5;s=(9,5),(8,5)
while 1:
 for e in p.event.get():a=e.type;a<257>b;a<769<(d:=[d,k:=e.key&255][k in S])
 x,y=h;*s,l=h,*s;x+=d<L;y-=d==L;y+=d==D;y-=d>D;h=x,y
 if h==f:s+=l,;f=g(40),g(30)
 (30>y>-1<x<40)<1>r;h in s[1:]>r;z.fill(-1);[r(z,5**9,(x*t,y*t,t,t))for x,y in s];r(z,'red',(f[0]*t,f[1]*t,t,t));j.flip();p.time.Clock().tick(t)
``` 452?

[quartz wave]

crashes for some reason :p

[gleaming linden]

i guess not then

[quartz wave]

452 works ```py
import pygame as p,random as r
j=p.display
z=j.set_mode((400,300))
t=10
g=r.randrange
r=p.draw.rect
S=d,L,D,U=b'OPQR'
f=t,5;h=9,5;s=h,(8,5)
while 1:
for e in p.event.get():a=e.type;a<257>b;a<769<(d:=[d,k:=e.key&255][k in S])
x,y=h;x+=d<L;x-=d==L;y+=d==D;y-=d>D;h=x,y;s,l=h,s
if h==f:s+=l,;f=g(40),g(30)
(30>y>-1<x<40)<1>r;h in s[1:]>r;z.fill(-1);[r(z,5**9,(xt,yt,t,t))for x,y in s];r(z,'red',(f[0]*t,f[1]*t,t,t));j.flip();p.time.Clock().tick(t)

[gleaming linden]

~~```py
import pygame as p,random as r
j=p.display
z=j.set_mode((400,300))
t=10
g=r.randrange
r=p.draw.rect
S=d,L,D,U=b'OPQR'
f=t,5;h=9,5;s=h,(8,5)
while 1:
for e in p.event.get():a=e.type;a<257>b;a<769<(d:=[d,k:=e.key&255][k in S])
h[0]+=(d<L)-(d==L);h[1]+=(d==D)-(d>D);s,l=h,s
if h==f:s+=l,;f=g(40),g(30)
(30>y>-1<x<40)<1>r;h in s[1:]>r;z.fill(-1);[r(z,5**9,(xt,yt,t,t))for x,y in s];r(z,'red',(f[0]*t,f[1]*t,t,t));j.flip();p.time.Clock().tick(t)

[quartz wave]

you forgot the x and y in the bound checks

[gleaming linden]

oh right

[quartz wave]

also h is a tuple

[hybrid trail]

does 'red' -> U**8 work

actually it is bigger than 5**15 so probably not lol

[gleaming linden]

import pygame as p,random as r
j=p.display
z=j.set_mode((400,300))
t=10
g=r.randrange
r=p.draw.rect
S=d,L,D,U=b'OPQR'
f=t,5;h=9,5;s=h,(8,5)
while 1:
 for e in p.event.get():a=e.type;a<257>b;a<769<(d:=[d,k:=e.key&255][k in S])
 x,y=h;x+=d<L;x-=d==L;y+=d==D;y-=d>D;h=x,y;h in s>r;*s,l=h,*s
 if h==f:s+=l,;f=g(40),g(30)
 (30>y>-1<x<40)<1>r;z.fill(-1);[r(z,5**9,(x*t,y*t,t,t))for x,y in s];r(z,'red',(f[0]*t,f[1]*t,t,t));j.flip();p.time.Clock().tick(t)
``` 448

[quartz wave]

on the edge case that h moves to the tail's coordinates in the same tick that the tail moves away

that would fail

that is unless the snake also consumes an apple in the same period

[hybrid trail]

version control via discord lol

[quartz wave]

so i don't think it's valid

okay after further testing that's a galactic edge case

448 valid

[gleaming linden]

[quartz wave]

yellow apple variant for t**8

[hybrid trail]

i keep trying and failing to use hash instead of random

[gleaming linden]

id maybe?

[hybrid trail]

I don't think id changes between runs

so the apple would spawn in the same place every time

[gleaming linden]

i thought it was memory address in cpython

which would be the same if you don't do anything in between runs ig

[hybrid trail]

you can get 429 but it depends on the position of the snake as a hash so often it is the same

[quartz wave]

also the snake can reverse directions apparently

[restive void]

The id() of an object will typically change between runs, but yeah, it's not guaranteed.

[quartz wave]

as in it's going right but you can press the left arrow to make it kill itself

same with going up and you can press down

[versed eagle]

some impls of snake allow this

[hybrid trail]

421, pretty sure it should be legal

from pygame import*
j=display
z=j.set_mode((400,300))
t=10
r=draw.rect
S=d,L,D,U=b'OPQR'
f=t,5;h=9,5;s=h,(8,5)
while 1:
 for e in event.get():a=e.type;a<257>b;a<769<(d:=[d,k:=e.key&255][k in S])
 x,y=h;x+=d<L;x-=d==L;y+=d==D;y-=d>D;h=x,y;h in s>r;*s,l=h,*s
 if h==f:s+=l,;f=id(e)%40,id(e)%30
 (30>y>-1<x<40)<1>r;z.fill(-1);[r(z,5**9,(x*t,y*t,t,t))for x,y in s];r(z,'red',(f[0]*t,f[1]*t,t,t));j.flip();time.Clock().tick(t)

id should make it completely different between runs, and python tuple[int] hash should have good rng

[quartz wave]

oh

i just used id(e)

[hybrid trail]

that works

[quartz wave]

there's an obnoxious increase in the amount of apples spawning on edges but other than that it seems pretty random

[hybrid trail]

is it fine that the apple can never spawn in some squares

[quartz wave]

i came up with a pretty weird thing

428 ```py
from pygame import*
j=display
z=j.set_mode((400,300))
t=10
r=draw.rect
S=d,L,D,g=b'OPQR'
f=t,5;h=9,5;s=h,(8,5)
while 1:
for e in event.get():a=e.type;a<257>b;a<769<(d:=[d,k:=e.key&255][k in S]);g^=id(e)//9
x,y=h;x+=d<L;x-=d==L;y+=d==D;y-=d>D;h=x,y;h in s>r;s,l=h,s
if h==f:s+=l,;f=g%40,g%30
(30>y>-1<x<40)<1>r;z.fill(-1);[r(z,5**9,(xt,yt,t,t))for x,y in s];r(z,'red',(f[0]*t,f[1]*t,t,t));j.flip();g^=time.Clock().tick(t)

user dependent randomness...?

[hybrid trail]

whats the point?

[quartz wave]

what point

generate numbers that look random

[subtle viper]

now we just need to shave 29 more off that :p

[gleaming linden]

make a pygame fork designed for golfing

pygm or something

[grave grail]

lol

[hybrid trail]

410

from pygame import*
j=display
t=10
r=j.set_mode((400,300)).fill
S=d,L,D,g=b'OPQR'
f=t,5;x,y=h=9,5;s=h,(8,5)
while 1:
 for e in event.get():a=e.type;a<257>b;a<769<(d:=[d,k:=e.key&255][k in S]);g^=id(e)//9
 x+=d<L;x-=d==L;y+=d==D;y-=d>D;h=x,y;h in s>r;*s,l=h,*s
 if h==f:s+=l,;f=g%40,g%30
 (30>y>-1<x<40)<1>r;r(-1);[r(5**9,(x*t,y*t,t,t))for x,y in s];r('red',(f[0]*t,f[1]*t,t,t));j.flip();g^=time.Clock().tick(t)

[earnest snow]

g^=time.Clock().tick(t) is wild

[rough moat]

madness

the most cursed part of that entire thing is the from pygame import*

[earnest snow]

S=d,L,D,g=b'OPQR'?
x+=d<L;x-=d==L;y+=d==D;y-=d>D;h=x,y;h in s>r;*s,l=h,*s?

[hybrid trail]

the second line is what moves the snake

the first line instantiates the arrow key directions

[earnest snow]

yeah ik, I'm just amazed on why there's an in between those

just saw it

g^=id(e)//9

g^=id(e)//9

how can something not work without a g^=id(e)//9

[hybrid trail]

to check if the snake is self intersecting

[rough moat]

could have sworn // would still return a float in some cases

oh it's if you divide floats

[earnest snow]

Oh it's for random

just noticed, you use g for randomness

[hybrid trail]

you might be on to something here

actually .tick() is probably not random enough

[earnest snow]

id(s) should be

from pygame import*
j=display
t=10;W=Clock()
r=j.set_mode((400,300)).fill
S=d,L,D,g=b'OPQR'
f=t,5;x,y=h=9,5;s=h,(8,5)
while 1:
 for e in event.get():a=e.type;a<257>b;a<769<(d:=[d,k:=e.key&255][k in S]);
 x+=d<L;x-=d==L;y+=d==D;y-=d>D;h=x,y;h in s>r;*s,l=h,*s;
 if h==f:s+=l,;f=id(f)%40,id(e)%30
 (30>y>-1<x<40)<1<r;r(-1);[r(5**9,(x*t,y*t,t,t))for x,y in s];r('red',(f[0]*t,f[1]*t,t,t));j.flip();W.tick(t)

414

[hybrid trail]

that rng seems random enough to me

[karmic pumice]

what are the a<257>b doing

[hybrid trail]

if event.type < 257, exit

checks for the quit event

[karmic pumice]

if a<257, it will check for it being >b, and that will error?

[hybrid trail]

398 which I assume you meant to type

from pygame import*
j=display
t=10
r=j.set_mode((400,300)).fill
S=d,L,D,g=b'OPQR'
f=t,5;x,y=h=9,5;s=h,(8,5)
while 1:
 for e in event.get():a=e.type;a<257>b;a<769<(d:=[d,k:=e.key&255][k in S])
 x+=d<L;x-=d==L;y+=d==D;y-=d>D;h=x,y;h in s>r;*s,l=h,*s
 if h==f:s+=l,;f=id(f)%40,id(e)%30
 (30>y>-1<x<40)<1<r;r(-1);[r(5**9,(x*t,y*t,t,t))for x,y in s];r('red',(f[0]*t,f[1]*t,t,t));j.flip();Clock().tick(t)

[hybrid trail]

correct

[earnest snow]

yeah

[earnest snow]

how putting time.Clock is shortert than W=Clock()?

even then, for my pygame just Clock() works though

[hybrid trail]

hmm, it doesn't work for me

NameError: name 'Clock' is not defined

I wonder why

[earnest snow]

pygame-ce 2.5.0 (SDL 2.30.3, Python 3.10.5)

yours?

[hybrid trail]

pygame 2.6.0 (SDL 2.28.4, Python 3.10.12)

I just installed it yesterday for this

[earnest snow]

better install pygame-ce though as pygame is not longer in support and -ce is the comunity edition

maybe it's a comunity edition feature or smth

[rough moat]

i wonder if you could oneliner this with walrus hackery and the cursed inline import

[earnest snow]

I can one-line it without walrus, but after a while though

[rough moat]

i know its possible to do that sort of stuff without walrus but frankly i can never be bothered

walrus logic is already fucked

[rough moat]

and what i mean by the cursed inline import is doing __import__()
i did it in my mandelbrot with color in one line function to do a system call so i could use ansi codes

[earnest snow]

every walrus is just taking a lambda out of it,
w:=s(a) is equivalent in a cursed way of (lambda w:...)(s(a))

[karmic pumice]

!e

print((d:={},exec("import math as _",d))[0]["_"])

__import__ without __import__

[night quarry]

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

<module 'math' from '/lang/python/default/lib/python3.12/lib-dynload/math.cpython-312-x86_64-linux-gnu.so'>

[rough moat]

the main use i have for walruses is just changing external variables in a list comprehension

which is basically required for something like mandelbrot, afaik

>>> n = 0
>>> [n:=n+1 for i in range(10)]
[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
>>> n
10

[karmic pumice]

!e

n=0
l=locals()
[0 for _,l["n"]in zip(range(10),iter(lambda:n+1,None))]
print(n)

[night quarry]

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

10

[rough moat]

that is fucked up

ngl i didnt even know you could use iter() like that, ive never really used it

why does that work with the sentinel set to None but it doesn't without it??

[karmic pumice]

single argument iter is for, well, iterating iterables, usually __iter__
i suppose they could make something like

def iter(thing):
  if callable(thing):
    while True:
      yield thing()

but then iter(callable & iterable) would be weird

[rough moat]

oh

i see

[karmic pumice]

nowadays walrus could be used here instead

while block := f.read(64):
  ...

[rough moat]

i find it funny that i found the locals() manipulation the least fucked part

[rough moat]

speaking of walrus i looked up when it was added (3.8) and apparently this is also a thing ??

i dont think ive literally ever seen this used

[jolly rock]

Yeah I don't really see the point

[rough moat]

ive had a few cases where i imagine it’d be handy

sometimes py will treat a positional one like a keyword one and then when i declare that keyword argument somewhere else it errors

not sure if this would prevent that though

[karmic pumice]

something like def __init__(self, **kwargs): or __call__ in functools.partial could make use of it, to avoid collision of the self that is being passed as the partial arg with the partial instance itself (the "implicitly" passed one)

[mortal thorn]

I have used both * and /.

If you don't want to expose the internal names of the variables or want to be able to change the parametrar names without having to maybe update other code that calls that function.

[quartz wave]

391 ```py
from pygame import*
j=display
t=10
r=j.set_mode((400,300)).fill
S=d,L,D,U=b'OPQR'
f=t,5;x,y=h=9,5;s=h,(8,5)
while 1:
for e in event.get():a=e.type;a<257>b;a<769<(d:=[d,k:=e.key&255][k in S])
x+=d<L;x-=d==L;y+=d==D;y-=d>D;h=x,y;h in s>r;*s,l=h,s
if h==f:s+=l,;f=id(f)%40,id(e)%30
(30>y>-1<x<40)<1<r;r(-1);[r(5**9(p!=f)or'red',(p[0]*t,p[1]*t,t,t))for p in s+[f]];j.flip();Clock().tick(t)

oh wait a minute

390 ```py
from pygame import*
j=display
t=10
r=j.set_mode((400,300)).fill
S=d,L,D,U=b'OPQR'
f=t,5;x,y=h=9,5;s=h,(8,5)
while 1:
for e in event.get():a=e.type;a<257>b;a<769<(d:=[d,k:=e.key&255][k in S])
x+=d<L;x-=d==L;y+=d==D;y-=d>D;h=x,y;h in s>r;s,l=h,s
if h==f:s+=l,;f=id(f)%40,id(e)%30
(30>y>-1<x<40)<1<r;r(-1)
for p ins,f:X,Y=p;r(5**9(p!=f)or'red',(Xt,Yt,t,t))
j.flip();Clock().tick(t)

388 ```py
from pygame import*
j=display
t=10
r=j.set_mode((400,300)).fill
S=d,L,D,U=b'OPQR'
f=t,5;x,y=h=9,5;s=h,(8,5)
while 1:
for e in event.get():a=e.type;a<257>b;a<769<(d:=[d,k:=e.key&255][k in S])
x+=d<L;x-=d==L;y+=d==D;y-=d>D;h=x,y;h in s>r;s,l=h,s
if h==f:s+=l,;f=id(f)%40,id(e)%30
30>y>-1<x<40or b;r(-1)
for p ins,f:X,Y=p;r(5**9(p!=f)or'red',(Xt,Yt,t,t))
j.flip();Clock().tick(t)

okay i'm doing my assignments bye

[hybrid trail]

376 ```python
from pygame import*
j=display
t=10
r=j.set_mode((400,300)).fill
d,L,D=b'OPQ'
f=t,5;x,y=h=9,5;s=h,(8,5)
while 1:
for e in event.get():a=e.type;a<257>b;a<769>78<(k:=e.key&255)<83>(d:=k)
x+=d<L;x-=d==L;y+=d==D;y-=d>D;h=x,y;h in s>r;*s,l=h,s
if h==f:s+=l,;f=id(f)%40,id(e)%30
30>y>-1<x<40or b;r(-1);c='red'
for X,Y in f,s:r(c,(Xt,Yt,t,t));c=5**9
j.flip();Clock().tick(t)

[quartz wave]

simultaneously fixed the feature(?) where if the apple spawns in the snake the apple is on top of it

[somber raptor]

guiys i wanna make this code work my redfininf the meaning of the keyword false

False = 1
print(False)

this should print 1

someone help

[versed eagle]

use codecs

[somber raptor]

python crashed

[grave grail]

Idk if using ctypes would be possible cuz they a different in bytecode

[somber raptor]

it worked

🙏

python crashed

[versed eagle]

that will not work

[somber raptor]

it did

[versed eagle]

well no

obviously not

[somber raptor]

well urm aktchully

[versed eagle]

you're screwing with refcount stuff (which is why python is crashing)

what you are NOT doing is changing the value

[somber raptor]

should i not screw with it

[versed eagle]

correct

at least not how you're doing

[somber raptor]

oh no

my python file got corrupted

ANYWAYS

[karmic pumice]

!e

import ctypes as c
(c.c_char * 12).from_address(id(True) + 16)[:] = c.string_at(id(False) + 16, 12)
print(True == False)

[night quarry]

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

True

[versed eagle]

!e ```py
x = 0 # using x to prevent const-folding
print(True + x + True)
from ctypes import *
cast(id(True), POINTER(c_int))[6] = 2
print(True + x + True)

[night quarry]

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

001 | 2
002 | 4

[versed eagle]

@somber raptor im guessing that's something like what you want

[somber raptor]

oh wow

hope my pc doesnt crash

[karmic pumice]

so the 24th byte is the one with the long value?

[versed eagle]

python int structure looks like this

struct {
    uint64_t refcount_stuff;
    void *ob_type;
    uint64_t ndigits;
    uint32_t digits[];
}

its different in the actual code, but that's an equivalent struct wrt. object offsets

[karmic pumice]

typedef struct _PyLongValue {
    uintptr_t lv_tag; /* Number of digits, sign and flags */
    digit ob_digit[1];
} _PyLongValue;

struct _longobject {
    PyObject_HEAD
    _PyLongValue long_value;
};

is what i found

[versed eagle]

yeah

PyObject_HEAD has the refcount stuff and the ob_type

theres actually stuff before the objects also

gc tags and whatnot

[karmic pumice]

so (int*)[6], or, well, 4*6 would be where digits starts, and you write a single int to that? nice

struct {
    uint64_t refcount_stuff; // 0 -> 8
    void *ob_type; // 8 -> 16
    uint64_t ndigits; // 16 -> 24
    uint32_t digits[]; // 24 -> ?
}

[karmic pumice]

hmh, i guess for code that relies on lv_tag sign you'd need to change it from zero to positive if you change false to 1, and vice-versa for true to 0

it should be possible to pretty easily replace the repr and other dunders with ctypes

[versed eagle]

yeah

[karmic pumice]

!e

import ctypes as c

ftype = c.CFUNCTYPE
obj = c.py_object

tp_repr_offset = 88
reprfunc = ftype(obj, obj)

@reprfunc
def new_bool_repr(self):
  if self is False:
    return "True"
  return "False"

(reprfunc * 1).from_address(id(bool) + tp_repr_offset)[0] = new_bool_repr

print(True)

[night quarry]

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

False

[versed eagle]

this hooks it for all bools

you can replace it for just one

[karmic pumice]

how do you replace a method for a single value? the method is only part of the type, no? or you mean that i could define a new type and set True's type to that?

[versed eagle]

!e ```py
from ctypes import *

class b (int):
slots = ()
def repr (self):
return "False"

py_object.from_address(id(True) + 8).value = b

print(True)

[night quarry]

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

False

[versed eagle]

exactly that

[karmic pumice]

yeah i see that makes sense

oh, also, if you keep an eye on peps,

!pep 667

[night quarry]

**PEP 667 - Consistent views of namespaces**

Status

Accepted

Python-Version

3.13

Created

30-Jul-2021

Type

Standards Track

[karmic pumice]

when thats implemented, that should make it pretty easy to modify the locals of the calling function, right?
like just modifying inspect.stack()[1].frame.f_locals should work
is there an easy way to do that currently?
like

import inspect
def var(name, value):
  inspect.stack()[1].frame.f_locals[name] = value

def f():
  var("x", 42)
  print(x)

or that wont work even with it? because at parse time x is not seen defined in f so it will not count as local?

!e

def f():
  exec("x = 1")
  return x
print(f())

[night quarry]

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

001 | Traceback (most recent call last):
002 |   File "/home/main.py", line 4, in <module>
003 |     print(f())
004 |           ^^^
005 |   File "/home/main.py", line 3, in f
006 |     return x
007 |            ^
008 | NameError: name 'x' is not defined

[karmic pumice]

like here
(i think thats whats happening here)

[versed eagle]

sys._getframe(1).f_locals

[somber raptor]

m=lambda A,B:[[sum(a*b for a,b in zip(r,c))for c in zip(*B)]for r in A]
p=lambda M,n:(R:=[[1]*len(M)for _ in range(len(M))],(R:=m(R,M)if n%2 else R,M:=m(M,M),n:=n//2)[0])[-1]if n else 0
f=lambda n:p([[0,1],[1,1]],n-1)[0][1]if n else 0
print(f(100))

outputs 2 why

based on the binet formula with matrix multiplication

[karmic pumice]

maybe you wanted to have some form of a loop in p? because you just do the thing once and end up with [[1, 2], [1, 2]]

[clear venture]

What's the most cursed way to reverse a list?

I posted this earlier but it got buried

!e

a = [1,2,3,4,100]
b = a.copy()
[next(...for a[i]in b[len(a)-i-1::-1])for i in range(len(a))]

print(a)

[night quarry]

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

[100, 4, 3, 2, 1]

[versed eagle]

in place?

[clear venture]

The example is in place but any way as long as it's cursed

[somber raptor]

def matrix_fib(n):
    def m(A, B):return[[sum(A[i][k]*B[k][j]for k in range(len(B)))for j in range(len(B[0]))]for i in range(len(A))];
    def pow(M,p):
        if p==0:return[[1,0],[0,1]];
        half=pow(M,p//2);
        if p % 2 == 0:return m(half,half);
        else:return m(m(half,half),M);
    if n == 0:return 0;
    elif n == 1:return 1;
    result=pow([[0,1],[1,1]],n);return result[0][1];
print(matrix_fib(100));

cpp moment

[fleet bridge]

you don't really need matrix multiplication here

[somber raptor]

proof

[somber raptor]

matrix_fib = lambda n: (
    (lambda m: (lambda pow: (
        lambda M, p: (
            [[1, 0], [0, 1]] if p == 0 else
            (m(pow(M, p // 2), pow(M, p // 2)) if p % 2 == 0 else m(m(pow(M, p // 2), pow(M, p // 2)), M))
        )
    ))(
        lambda A, B: [[sum(A[i][k] * B[k][j] for k in range(len(B))) for j in range(len(B[0]))] for i in range(len(A))]
    )
    )(
        [[0, 1], [1, 1]], n
    )[0][1] if n > 1 else (0 if n == 0 else 1)
)
print(matrix_fib(100))

help

[karmic pumice]

what did you expect? (lambda m: (lambda pow: (...)))(whatever) gives you the lambda pow, which only takes a single argument, and you are passing it two with the ([[0, 1], [1, 1]], n)

[somber raptor]

what the almighty hell have i got myself into

matrix_fib = lambda n: (
    (lambda m: (
        lambda pow: (
            pow([[0, 1], [1, 1]], n) if n > 1 else 1 if n == 1 else 0
        )(lambda M, p: (
            (lambda x: m(x, x))(M) if p % 2 == 0 else m(m(M, M), M) if p > 1 else M
        )(M, p // 2))
    ))(lambda A, B: [[sum(A[i][k] * B[k][j] for k in range(len(B))) for j in range(len(B[0]))] for i in range(len(A))])
)
print(matrix_fib(100))

why

fibonacci = (
    lambda n: (
        (lambda matrix_mult: (
            lambda matrix_pow: (
                matrix_pow(
                    [[0, 1], [1, 1]], 
                    n
                )[0][1] if n > 1 else (0 if n == 0 else 1)
            )
        )(
            lambda M, p: (
                [[1, 0], [0, 1]] if p == 0 else
                M if p == 1 else
                (lambda half_pow: (
                    matrix_mult(half_pow, half_pow) if p % 2 == 0 else
                    matrix_mult(M, half_pow)
                )(matrix_pow(M, p // 2)))
            )
        )(
            lambda A, B: [[sum(A[i][k] * B[k][j] for k in range(len(B))) for j in range(len(B[0]))] for i in range(len(A))]
        )
    )
)
print(fibonacci(100))

[karmic pumice]

missing parens, but also, you're using matrix_pow inside lambda M, p: ...: it isnt bound at that point
i dont know how are you even writing this, are you just doing random stuff?

[somber raptor]

im using code to dranslate it

🤷

wtf is this:
<function <lambda>.<locals>.<lambda>.<locals>.<lambda>.<locals>.<lambda> at 0x7fb8a5621040>

[karmic pumice]

!e

print((lambda:(lambda:(lambda:(lambda:()))))()()())

[night quarry]

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

<function <lambda>.<locals>.<lambda>.<locals>.<lambda>.<locals>.<lambda> at 0x7fc113225440>

[fleet bridge]

from __future__ import annotations
import typing as t
import collections.abc as t_abc


def c3[X](x: X, get_bases: t_abc.Callable[[X], list[X]], /) -> list[X]:
    def merge(mros: list[list[X | None]]) -> list[X | None]:
        if not any(mros):
            return []
        for candidate, *_ in mros:
            if all(id(candidate) not in map(id, tail) for _, *tail in mros):
                return [candidate] + merge(
                    [tail if head is candidate else [head, *tail] for head, *tail in mros]
                )
        else:
            raise TypeError("no legal mro")

    def _c3(x: X) -> list[X | None]:
        if bases := get_bases(x):
            return [x] + merge([_c3(base) for base in bases])
        else:
            return [x, None]

    m = _c3(x)
    assert m[-1] is None
    assert all(x is not None for x in m[:-1])
    return m[:-1]  # type: ignore


def test():
    # fmt: off
    assert c3(1, {1:[],2:[],3:[1,2]}.get) == [1]
    assert c3(2, {1:[],2:[],3:[1,2]}.get) == [2]
    assert c3(3, {1:[],2:[],3:[1,2]}.get) == [3, 1, 2]

    assert c3(1, {1:[],2:[],3:[],4:[1,2],5:[2,3],6:[3,1]}.get) == [1]
    assert c3(4, {1:[],2:[],3:[],4:[1,2],5:[2,3],6:[3,1]}.get) == [4, 1, 2]
    assert c3(5, {1:[],2:[],3:[],4:[1,2],5:[2,3],6:[3,1]}.get) == [5, 2, 3]
    assert c3(6, {1:[],2:[],3:[],4:[1,2],5:[2,3],6:[3,1]}.get) == [6, 3, 1]
    try:
        c3(7, {1:[],2:[],3:[],4:[1,2],5:[2,3],6:[3,1],7:[4,5,6]}.get)
    except TypeError:
        pass # ok
    else:
        raise Exception(f"expected error")

    assert c3(7, {1:[],2:[],3:[],4:[1,2],5:[2,3],6:[1,3],7:[4,5,6]}.get) == [7, 4, 5, 6, 1, 2, 3]
    # fmt: on

    return True
assert test()
``` C3 linearization
how far can you golf it?

[versed eagle]

(i assume we dont have to golf the tests and whatnot?)

[subtle viper]

tests and whatever imports they use

[versed eagle]

my version doesnt use any imports :3

all the import stuff is for typing lol

[fleet bridge]

of course not
the only required part is a c3 name definition, it should be a callable that takes two args and behaves as shown in tests

[fleet bridge]

what's the shortest way to raise TypeError?

raise TypeError
1+()
()()
id()
{[]} # not sure what exception this will raise, iirc unhashability errors are TypeError

is it possible in less than 4 chars? what are other fun ways to do it in pretty low number of chars?

+()

f() # only if you have variable f that is not callable or does not allow passing 0 args 
+x # only if you have non-numeric variable x
~x # non-int x

[gleaming linden]

0@0

[fleet bridge]

0[0]
0()

[mortal thorn]

-""
-()
-[]
-{} 

[versed eagle]

alright then

here

def M(*m):
 if not any(m):return[]
 for c,*_ in m:
  if 1^any(id(c)in map(id,t)for _,*t in m):return[c]+M(*[([h,*t],t)[h is c]for h,*t in m])
 else:0@0
c3=lambda X,G:(_:=lambda x:[x]+(M(*map(_,b))if(b:=G(x))else[0]))(X)[:-1]

i dont know what c3 linearization is so i dont know how to golf the underlying algorithm at all lol

but that passes the tests

[long fulcrum]

is it possible to simplify these to be shorter?

*2//10
*2%10

[fleet bridge]

n = 123
x=n*2//10;y=n*2%10
x,y=n*2//10,n*2%10
x,y=divmod(n*2,10)

[long fulcrum]

and then to add them just sum() isnt it

or would that be longer

[quartz wave]

yup

[versed eagle]

n*2//10+n*2%10
sum(divmod(n*2,10))

¯_(ツ)_/¯

[long fulcrum]

yeah 😭

[karmic pumice]

the thing that flattens and orders class hierarchies to form an MRO
usually it just ends up flattening things

# there is an implied "object" at the end of each one
class A:
  ... # A.mro = [A]
class B(A):
  ... # B.mro = [B, *flatten(A.mro)] = [B, A]
class C:
  ... # C.mro = [C]
class D(B, C):
  ... # D.mro = [D, *flatten(B.mro), *flatten(C.mro)] = [D, B, A, C]

but when there are multiple parents with the same thing in the mro it gets funny, how i understand it is

class A: ...

class B: ...

class C: ...

class D: ...

class E: ...

class K1(C, A, B): 
  ... # K1.mro = [K1, C, A, B]

class K3(A, D): 
  ... # K3.mro = [K3, A, D]

class K2(B, D, E): 
  ... # K2.mro = [K2, B, D, E]

class Z(K1, K3, K2):
  ... # Z.mro = ?
  # well, obviously start with Z itself,
  # then K1,
  # K1 has     [K1, C, A, B      ],
  # K3 has     [K3,    A,    D   ],
  # K2 has     [K2,       B, D, E],
  # so,
  # we take K1,
  # from it we take C because its unique in it, but we dont take A, B because they are also met in the mro's of the later things in the inheritance tree
  # then, from K3 we take K3 and A (its the last one with it), and dont take D,
  # then, from K2 we take K2 and B, D, E
  # Z.mro = [Z, K1, C, K3, A, K2, B, D, E]

https://en.wikipedia.org/wiki/C3_linearization

[versed eagle]

and with n*2//10+n*2%10 you can actually do n//5+n*2%10 to get a numerically equivalent result

so if you just want the result of adding them, that's probably the best for you

[long fulcrum]

oh shit that works?

i thought it would fuck with the floor

[versed eagle]

it works because 2 is a factor of 10

[worldly trail]

I saw something similar in this channel, few weeks ago:
Inspired me to do this:

def main() -> exec('import random;print(random.random())'): return "Hello"
print(main())

[versed eagle]

?

[night quarry]

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

001 | 0.2733108063088745
002 | None

[worldly trail]

!e

def main() -> exec('import random;print(random.random())'): return "Hello"
print(main())

[night quarry]

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

001 | 0.09043965249620067
002 | Hello

[versed eagle]

whats your point

[fleet bridge]

!e

def main() -> print(__import__('random').random()): return "Hello"
print(main())

[night quarry]

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

001 | 0.17067252666070132
002 | Hello

[worldly trail]

no point, just wanted to share this 😦

i thought it would fit the channel

[versed eagle]

sharing it is fine

im just curious which part of it is interesting to you

because i might be able to offer resources into further interesting things

[worldly trail]

the type annotation part and the exec part.

[versed eagle]

have you looked at hooking __annotations__ to do silly things?

one of the simplest things you can do is just give it globals

!e ```py
annotations = globals()
test: "hello, world"
print(test)

[night quarry]

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

hello, world

[versed eagle]

but you can also give it any object with custom __setitem__ behaviour

[worldly trail]

honestly, im kinda drunk. But truly, i will take a look over what you wrote. You also gave me some ideas in playing with globals() 🙂
Thanks

[versed eagle]

being drunk and playing with esoteric code is always fun :)

have fun! and make sure you drink plenty of water, alcohol is dehydrating

[worldly trail]

thx
GL & HF

back again.
This is nuts:

!e

asd = '''def main() -> print(__import__('random').random()): return "Hello"
\nprint(main())'''
zxc = [ord(x) for x in asd]
qwe = [chr(x) for x in zxc]
rr = "".join(qwe)
print(exec(str(rr)))

[unique heath]

oh that does make sense

in a way i hate

[versed eagle]

its quite lovely

i think

[unique heath]

esopy in a nutshell

[fleet lintel]

@versed eagle Reminds me of when I made visualizations for a manual MRO resolution challenge #esoteric-python message As long as you keep the graph ordered correctly left to right it's fairly easy to see what's going on in MRO.

Also I just said "Method Order Resolution resolution" smhing my head

[unique heath]

tree mro when????

[distant salmon]

I have a challenge to you guys, try to list the 12 constants of Python listed under https://docs.python.org/3/library/constants.html.
Hint: the first two are False and True.
No cheating

[subtle viper]

there's 12-?

||True, False, None, ..., "", (), 0, 1, math.e, math.pi, 1j, 0.0?||

oh-

[distant salmon]

Only the first 4 are correct

Still 8 missing

[gleaming linden]

||__builtins__||?

[subtle viper]

there's actually only, like, 6

the rest are ||only in specific environments (i.e. the REPL)||

[gleaming linden]

does ||Ellipsis|| count separately?

[unique heath]

||-12 (?)-255||

||__builtins__||

[subtle viper]

wrong both sides

[subtle viper]

not there

[unique heath]

atcualy maybe nto

[unique heath]

||-256/257?||

[subtle viper]

still wrong :3

[unique heath]

uh this is hard

[unique heath]

||-253/254||

[subtle viper]

also ||no ints||

[unique heath]

oh

waitwaht

3:

oH

||None||

how did i not think of that

idk i give up

imma do math now

[gleaming linden]

does ||__file__|| count?

[unique heath]

(in math class lmfaoo)

[subtle viper]

||nnope||

||actually gonna go check agai nope||

[distant salmon]

||Ellipsis is one, but not builtin nor file||

[gleaming linden]

ok at least I got one

[subtle viper]

fun unrelated fact there are cached strings

just lile cached ints

[gleaming linden]

are the repl ones ||quit, exit, copyright, credits, license, help, _||?

[subtle viper]

||last one is not part of it||

[gleaming linden]

actually the last one shouldn't be a constant

yeah

||open|| then?

[distant salmon]

Nope

[gleaming linden]

hm ok

[subtle viper]

god i hate typing on phone-

[gleaming linden]

||interestingly they're all listed at the end of __builtins__.__dir__() in the repl||

[subtle viper]

why are they even there ||on the constants||

[gleaming linden]

idk it just does that

I gtg soon I'm gonna check the answers

OH ||NotImplemented|| is a constant?

til

||__debug__|| is also really sneaky

[unique heath]

imma check

oh i didnt know ||notimplemented|| was a const

i refuse to believe ||exit/quit|| are consts

[subtle viper]

ikrrr??

[grave grail]

Oh, my challenge 💀

[gleaming linden]

||typing them in the REPL prompts something along the lines of "use exit() or ctrl+d to exit"||

[subtle viper]

but they're literally just like built-ins

[gleaming linden]

||It triggers when you access the name at all||

[subtle viper]

..you mean it repr()s properly

[gleaming linden]

||hm, it's the repr?||

[subtle viper]

yup

i'll check the source later

i guess.

[gleaming linden]

I thought it would be hardcoded somewhere, or it doesn't need to be a constant

[quartz wave]

https://github.com/python/cpython/blob/main/Lib/_sitebuiltins.py#L17-L18

[night quarry]

Lib/_sitebuiltins.py lines 17 to 18

def __repr__(self):
    return 'Use %s() or %s to exit' % (self.name, self.eof)```

[quartz wave]

https://github.com/python/cpython/blob/main/Lib/site.py#L437-L438

[night quarry]

Lib/site.py lines 437 to 438

builtins.quit = _sitebuiltins.Quitter('quit', eof)
builtins.exit = _sitebuiltins.Quitter('exit', eof)```

[unique heath]

cursed

[versed eagle]

||True, False, None, NotImplemented, Ellipsis (...)||
and ||the repl has quit, exit, license, copyright, help||
that's 10/12, but i have no clue what im missing lol

[distant salmon]

You are missing one from repl

[versed eagle]

yeah

iirc there's 6 in each

so missing one from repl and one from the main ns

[distant salmon]

yes

[versed eagle]

but i dont know what they are lol

ill probably figure it out in a few minutes and then realise how silly i am for not remembering them

[fleet bridge]

||__debug__|| is from main ns

|| _ || from repl

[distant salmon]

That one is not it

[versed eagle]

|| _ || isnt a constant i dont think

[fleet bridge]

but it is a builtin name that appears whhen you do stuff in repl

[versed eagle]

doesnt count as one

[fleet bridge]

ah, right, not a constant

[versed eagle]

because it swooshes around to the result of the previous thingy

yeah

[fleet bridge]

|| credits ||

[earnest wing]

my thoughts
||

True, False, None, ..., Ellipsis, NotImplemented, __debug__
# maybe the repl specific builtins / dunders?
# there was also a dunder constant that was specific to like 3.10 but it's probably not listed in the documentation

||

it was ||__peg_parser__||

:3

[flint hollow]

it's not in 3.12 😔

[fleet bridge]

it was in the last version that had peg parser

all versions after it use ll1 (is it called like that?) parser

LL1
LL(1)

[flint hollow]

i thought python switched from ll1 to peg

[fleet bridge]

hmmm

[flint hollow]

!pep 617

[night quarry]

**PEP 617 - New PEG parser for CPython**

Status

Final

Python-Version

3.9

Created

24-Mar-2020

Type

Standards Track

[fleet bridge]

I don't know, tbh
you might be correct

[flint hollow]

the dunder got nuked along with the older parser in 3.10: https://bugs.python.org/issue40939

[versed eagle]

backwards

[earnest wing]

!e ```py
for i in 1,...,3:
print('looping :)')

[night quarry]

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

001 | looping :)
002 | looping :)
003 | looping :)

[karmic pumice]

now uh
hook into tuple constructor and make that a range

[earnest wing]

it goes straight into co_consts unfortunately

[earnest snow]

Why does everything have to have so much sintactic sugar? range() was perfectly fine

[versed eagle]

there's no sugar here

[grave grail]

It just happen to work (as far as I aware

!e ```py
for i in -1,...,30000:
print('looping :)')

[night quarry]

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

001 | looping :)
002 | looping :)
003 | looping :)

[grave grail]

Ye, I'm not wrong

[earnest snow]

Oh, got scared for a moment

mb

[somber raptor]

def tobindec(x):
    w,b=int(x),'';f=x-w
    while f and len(b)<10:f*=2;b+=str(int(f));f-=int(f)
    return f"{bin(w)[2:]}.{b}"

print(tobindec(1.4))

theres no way to golf this code

[fleet bridge]

but tuple is still created
so you have to hook before the compilation happens

[distant salmon]

At f-=int(f), you should be able to do f-=f//1 instead

[somber raptor]

damn

DAMn

howww

[karmic pumice]

- f"{bin(w)[2:]}.{b}"
+ f"{w:b}.{b}"

[earnest snow]

also, to many int(x)s, make it a variable

def tobindec(x):
    i=int;w,b=i(x),'';f=x-w
    while f and len(b)<10:f*=2;b+=str(i(f));f-=i(f) # f-=i(f) is 4 chars, as f-=f//1
    return f"{w:b}.{b}"

print(tobindec(1.4))

[somber raptor]

brooo

[earnest snow]

since b grows linearly, you can substract len(b) by 10 and will have the same effect as len(b)<10

but with len(b)-10 since it's zero on len(b) = 10, you can multiply f instead of use an and

def tobindec1(x):
    i=int;w,b=i(x),'';f=x-w
    while f*(len(b)-10):f*=2;b+=str(i(f));f-=i(f)
    return f"{w:b}.{b}"

sitting at 122, test it just in case I'm wrong

[somber raptor]

124

!e
print(len("""def tobindec1(x):
i=int;w,b=i(x),'';f=x-w
while f*(len(b)-10):f*=2;b+=str(i(f));f-=i(f)
return f"{w:b}.{b}"
"""))

[night quarry]

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

124

[earnest snow]

you're counting the last new line and identating an empty line

[somber raptor]

oh

[earnest snow]

!e

print(len("""def tobindec1(x):
    i=int;w,b=i(x),'';f=x-w
    while f*(len(b)-10):f*=2;b+=str(i(f));f-=i(f)
    return f\"{w:b}.{b}\""""))

[night quarry]

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

119

[earnest snow]

tf 119

even better that what sublime text said

don't know how, I copy/pasted it

[somber raptor]

def tobindec1(x):
 i=int;w,b=i(x),'';f=x-w
 while f*(len(b)-10):f*=2;b+=str(i(f));f-=i(f)
 return f"{w:b}.{b}"

golfed it even more

just removed some spaces

would doing lambda remove char count

[earnest snow]

you just removed some identation

if you're gonna do that, put tabs, they're big enough their comfortable to look at and only 1 char

[hybrid trail]

does something like

bindec=lambda x:f'{int(x):b}.{int(x%1*2**10):010b}'.rstrip('0')
``` work

[somber raptor]

holy s***

wow

[earnest snow]

good one

[somber raptor]

why does the og have more persicion

[distant salmon]

looking at this, I'm kind of confused why there is both x and f, when you could just have reused x

[somber raptor]

explain it

63

bro removed 44 from the og function

[distant salmon]

Would save one more character from using 4**5 instead of 2**10, or just using 1024.

[unique heath]

ripoff rust ranges

[subtle viper]

bindec=lambda x:f'{int(x):b}.{int(x%1*1024):010b}'.strip('0')

[hybrid trail]

hint: there is one more byte that i can see be saved 😉

[distant salmon]

I wish this would work (it doesnt)

bindec=lambda x:f'{int(x):b}.{int(x%1<<10):010b}'.strip('0')

[hybrid trail]

actually i am wrong, my bad

[rapid temple]

I pretty much stopped using inheritance thanks to __getattr__ and ducktyping

Literally my favorite thing ever

[rugged sparrow]

!e ```py
from fishhook import hook, orig

@hook(tuple)
def iter(self):
for sz in range(3,6,2):
if len(self) == sz and all(isinstance(v,int) for v in self[::2]) and all(v==Ellipsis for v in self[1::2]):
yield from range(*self[::2])
return orig(self)

for i in 1,...,10,...,2:
print(i)```

[night quarry]

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

001 | 1
002 | 3
003 | 5
004 | 7
005 | 9

[rugged sparrow]

@fleet bridge ^ can always hook tp_iter

!e ```py
from fishhook import hook, orig

@hook(tuple)
def iter(self):
for sz in range(3,6,2):
if len(self) == sz and all(isinstance(v,int) for v in self[::2]) and all(v==Ellipsis for v in self[1::2]):
yield from range(*self[::2])
return orig(self)

for i in 0,...,5:
print(i)```

[night quarry]

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

001 | 0
002 | 1
003 | 2
004 | 3
005 | 4

[fleet bridge]

oh, right, didnt think of that 👍

[hybrid trail]

challenge: how low of a number can you get the output of the following program

def score(s):
    if 'eval' in s or 'exec' in s:
        raise ValueError('No silliness')
    if eval(s) != eval(s) == eval(s):
        return len(s)
    else:
        raise ValueError('Invalid input')
    

print(score(input()))

sorry for editing im thinking of how to best patch some cheese

[subtle viper]

||(s,s:='')[0]||?

oh wait-

||(s,s:=0)||

[hybrid trail]

hmm, actually both don't work for me

seems like reassigning variables in an eval is not that simple

[subtle viper]

nvm ||(s,s:='')||

nvmnvm-

[grave grail]

||(s:=0 if "s" not in dir() else 1,)||

Oh wait

[subtle viper]

ikrr

[grave grail]

||Change s to t or something like that||

||Did notice it was used to replace the variable, that variable should use in standalone||

[subtle viper]

||[t:='t'in dir()]||

-?

[hybrid trail]

yeah, that works

A sequel:

def safe_eval(s):
    return eval(s, {}, {})

def score(s):
    if 'eval' in s or 'exec' in s:
        raise ValueError('No silliness')
    
    if safe_eval(s) != safe_eval(s) == safe_eval(s):
        return len(s)
    else:
        raise ValueError('Invalid input')
    

print(score(input()))

[restive void]

I feel like the syntax should be 1, 3, ..., 9

[restive void]

(this is not safe)

[unique heath]

lol

[karmic pumice]

first, second, ..., last # step = second - first

a-la haskell?

[earnest wing]

1, 0.5, 0.25, ...

[restive void]

1, 1, 2, 3, ...

[wooden sail]

Hello friends, please help me with the task

And the last practical task, a more complex format:
Parsing the Marvel website (https://www.marvel.com/characters) - collect the following information about Marvel heroes:
Hero name,
Link to the page with him,
Create a dataset and save it to a CSV file using Python methods.
Use DevTools to find out the names of tags.

[earnest wing]

1, 6, 21, 107, 47176870, ...

[craggy hamlet]

is this homework lol

Hello friends, please help me with the task

And the last practical task, a more complex format:

[graceful prism]

https://paste.pythondiscord.com/T6NMPNOIW2N6ZOZQ23BLSD7E7U
it prints
0
26
-5
1
-10
-24

[zenith forum]

it's not hard to turn this into something easy to understand tho
start with the longest _ variable and just do some find-replaces

[versed eagle]

!e ```py
def score(s):
if 'eval' in s or 'exec' in s:
raise ValueError('No silliness')
if eval(s) != eval(s) == eval(s):
return len(s)
else:
raise ValueError('Invalid input')

inp="(globals().update({'ev' 'al':lambda*:True,'len':lambda*:float('-inf')}),)"

print(score(inp))

[night quarry]

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

-inf

[fleet bridge]

!e ```py
def score(s):
if 'eval' in s or 'exec' in s:
raise ValueError('No silliness')
if eval(s) != eval(s) == eval(s):
return len(s)
else:
raise ValueError('Invalid input')

inp="type('',(),dict(init=lambda*:import('gc').get_referents(str.dict)[0].setitem('len',lambda*:float('-inf')),eq=lambda*:True,ne=lambda*:True))()"

print(score(inp))
print('a'.len())

[night quarry]

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

001 | 173
002 | -inf

[fleet bridge]

I'm confused

need help
why doesn't it work?

!d sys

[night quarry]

sys

This module provides access to some variables used or maintained by the interpreter and to functions that interact strongly with the interpreter. It is always available.

[fleet bridge]

!e ```py
def score(s):
if 'eval' in s or 'exec' in s:
raise ValueError('No silliness')
if eval(s) != eval(s) == eval(s):
return len(s)
else:
raise ValueError('Invalid input')

inp="type('',(),dict(init=lambda*:[import('gc').get_referents(str.dict)[0].setitem('len',lambda*:float('-inf')),import('sys').clear_type_cache()][0],eq=lambda*:True,ne=lambda*_:True))()"

print(score(inp))

[night quarry]

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

216

[fleet bridge]

fuck it

I will go enjoy cat pictures

[versed eagle]

denballl im sorry to tell you this but,

that is not a cat

[earnest wing]

😭

[worthy blade]

Yoo that's one cute cat

[worthy blade]

Oh

Your're

righ

It's a lemon... How could I see it a cat

[worthy blade]

!e

import time
reminder=input("What do you want to be reminded of? ")
seconds=int(input("Seconds (0-60)"))
minutes=int(input("Minutes (0-60)"))
hours=int(input("Hours"))
if seconds>60:
    print("You have given more than 60 seconds. Restart the program")
    exit()

if minutes>60:
    print("You have given more than 60 minutes. Restart the program")
    exit()

if seconds==60:
    seconds=0
    minutes=minutes+1

if minutes==60:
    minutes=0
    hours=hours+1
    
minutes=minutes*60
hours=hours*60
hours=hours*60
total=hours+minutes+seconds

start=input("press enter to start your timer ")
print(reminder, "in", total, "seconds")

while total!=0:
    total=total-1
    print(total)
    time.sleep(1)

if total==0:
    print(reminder)

[night quarry]

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

:warning: Note: input is not supported by the bot :warning:

001 | What do you want to be reminded of? Traceback (most recent call last):
002 |   File "/home/main.py", line 2, in <module>
003 |     reminder=input("What do you want to be reminded of? ")
004 |              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
005 | EOFError: EOF when reading a line

[worthy blade]

ey then

nvm

[mild wren]

You could just hardcode it in your code you run here

[worthy blade]

How?

[mild wren]

reminder="my reminder"#input("What do you want to be reminded of? ")
seconds=123 #int(input("Seconds (0-60)"))
minutes=0#int(input("Minutes (0-60)"))
hours=0#int(input("Hours"))

the bot will most likely time out though at some point

Also, this isn't really esoteric 😄

[rugged sparrow]

len is handled by a slot method, so just modifying the dict isn't enough. Hook it with fishhook or similar code and that would work

[fleet bridge]

yeah I get it

but why clearing type cache didn't help?

what is this cache for? what does it contain

I have memories that clearing it helped with modifying builtins

[rugged sparrow]

I think it just clears the dict cache

Not the slot functions

[rugged sparrow]

At least for the native types

[somber raptor]

p=lambda n:n>1 and all(n%i for i in range(2,int(n**.5)+1))
print([x for x in range(2,int(input())+1)if p(x)])

try to make the code smalelr

[fleet bridge]

!e print({*{}})

[night quarry]

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

set()

[fleet bridge]

lmao

another empty set literal

[sage lintel]

what makes it a "literal"?

[quartz wave]

# literal
[] ; [a, b, ...]  # list
() ; (a, b, ...)  # tuple
{} ; {k0: v0, k1: v1, ...: ...}  # dict
{*()} ; {a, b, ...}  # set
"" ; "..."  # string
r"" ; r"..."  # r-string
b"" ; b"..."  # bytes
f"" ; f"a{b}c..."  # f-string
# and all the other string literal combinations...
...  # like literally just `...`
0 ; 1 ; ...  # int
0. ; 1.5123 ; ...  # float
1j ; ...  # complex
# constant
None
True
False
NotImplemented
__debug__
... ; Ellipsis
0 ; 1 ; ...
0. ; 1. ; 0.12 ; ...
0j ; 1.5j ; ...
() ; (1, 2, ...)
"" ; "..."
r"" ; r"..."
b"" ; b"..."

[sage lintel]

ok hmm

just to see if i understand

[quartz wave]

literals have dedicated syntax

[sage lintel]

is something like {*{*{*{*{*{*{}}}}}}} a literal?

[quartz wave]

you could technically say that, yup

[sage lintel]

i think i'm following

func(3) is not

3 + 3 is (wait maybe not because of the +?)

[quartz wave]

well it's 2 integer literals with an operation

[sage lintel]

ok makes sense 👍

in func(*args), *args is a literal i guess

just testing myself lol

[quartz wave]

*args is unpacking a variable args

[sage lintel]

hm

[quartz wave]

the empty set "literal" is just using the literal of the set with an unpacking of another empty literal to make an empty set

[sage lintel]

what about [*args], and that vs [*[]] and [1, 2, c]

[sage lintel]

would you call the whole combined thing a "single literal"?

[quartz wave]

well it's contained within a dedicated syntax, so yes

[quartz wave]

all of those are list literals

[sage lintel]

and it's ok to have variables in a literal? like [1, 2, b]

[quartz wave]

yup

[sage lintel]

ok

then *args can be one too, i think :p

i'm not sure

maybe not here

i get your point

[quartz wave]

it can be used inside of a literal but it's not a literal itself

[sage lintel]

ok that's true

[*args] is but *args not really

[quartz wave]

mhm ^^

[sage lintel]

interestingg

thanks

[quartz wave]

the definitions may differ for each language or person you ask but that's how i define it

[sage lintel]

ok that's good to know too, thank you

your definitions seem consistent though

[quartz wave]

i've looked it up on SO and someone defines a literal as "something that represents itself" which i think is a pretty good way to term it

[sage lintel]

i see how that leaves it open to interpretation :p

but yeah it does get the idea across

[fleet bridge]

alternative definition: literal is an expression that evaluates to the same thing no matter what (assuming no weird stuff happened)

# literals:
0
[]
(1,2)
{*()}
1+2
[1]+[2,3]
[0][0]

# not literals:
len('abc') # relies on len var not being shadowed
x
[x, y]
x + 1

I'm sure you can come up with weird examples that will prove this definition to be not very good

[sage lintel]

would that mean [print, 1][1] is a literal?

your definition is also interesting regardless

[fleet bridge]

strictly speaking [1,2] is not a literal, it doesn't represent a single value
it produces new value each time it is evaluated
it is called "list display" in the documentation

[fleet bridge]

I guess so 😄

[sage lintel]

ah that's a good point too

!e print([1, 2] is [1, 2])

[night quarry]

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

False

[fleet bridge]

same with dict/set displays
not sure about tuples:
(1,2) is the same object each time, but (x,y) produces new tuple each time

[sage lintel]

!e print((1, 2) is (1, 2))

[night quarry]

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

001 | /home/main.py:1: SyntaxWarning: "is" with 'tuple' literal. Did you mean "=="?
002 |   print((1, 2) is (1, 2))
003 | True

[sage lintel]

whoa

i didn't know that

[fleet bridge]

compiler is smart and noticed a duplicate constant
so it reused the same constant in two places
same happens with string literals, for example

[sage lintel]

that's very cool

that does make sense too, strings are similar

!e print((1,) + (2,) is (1, 2))

[night quarry]

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

001 | /home/main.py:1: SyntaxWarning: "is" with 'tuple' literal. Did you mean "=="?
002 |   print((1,) + (2,) is (1, 2))
003 | True

[fleet bridge]

compiler also can perform some operations on constants at compile time, as you can see from your snippet

[sage lintel]

!e print({*{*{*{*{*{*{}}}}}}} is set())

[night quarry]

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

False

[fleet bridge]

it is not smart enough to do that

[sage lintel]

hm makes sense

[fleet bridge]

try dis('x in {1,2}')
you will see that instead of creating new set each time, only one frozenset is created and reused

[sage lintel]

!e print({*()} is set())

[night quarry]

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

False

[fleet bridge]

set() must produce new set each time

[sage lintel]

!e print(dis('x in {1,2}'))

[night quarry]

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

001 | Traceback (most recent call last):
002 |   File "/home/main.py", line 1, in <module>
003 |     print(dis('x in {1,2}'))
004 |           ^^^
005 | NameError: name 'dis' is not defined. Did you mean: 'dir'? Or did you forget to import 'dis'?

[sage lintel]

sorry i haven't actually used dis before 🪦

[sage lintel]

ah

[gleaming linden]

you need from dis import dis

[sage lintel]

!e print(set() is set())

[night quarry]

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

False

[quartz wave]

so does any other way of creating a set

[sage lintel]

sorry if i'm doing too many bot calls

[fleet bridge]

so does any other way of creating a mutable object

[quartz wave]

yup

[sage lintel]

!e

from dis import dis
print(dis('x in {1,2}'))

[night quarry]

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

001 |   0           0 RESUME                   0
002 | 
003 |   1           2 LOAD_NAME                0 (x)
004 |               4 LOAD_CONST               0 (frozenset({1, 2}))
005 |               6 CONTAINS_OP              0
006 |               8 RETURN_VALUE
007 | None

[quartz wave]

now if it creates the exact same object in the entirety of the program it's a constant

[sage lintel]

👍

[quartz wave]

that's my definition

[fleet bridge]

by that logic ().__class__.__base__ is a constant
since it always evaluates to object

[quartz wave]

yup

but python doesn't optimize that out

yeah hold on i'm gonna fix my terms a bit

literal has dedicated syntax that, under normal circumstances, evaluates to pretty much the same thing in every runtime

constants are those that are the exact same object in the entirety of the program

and constant literals are those like (), (1, 2), 5, 7j, etc.

now drawing the line on "dedicated syntax" are operations applied on a literal

but not in the literal

[fleet bridge]

is [x,y] a literal?

[quartz wave]

to me, yup

[fleet bridge]

then it is not very clear what "to pretty much the same thing" means exactly
it can produce different values depending on x and y

[quartz wave]

non-literals are considered constant in "pretty much the same thing"

that sounds like a circular definition for some reason

[x, y] is a literal for me because it uses the [] syntax for lists

x[1::5] is a variable x being subscripted with a slice literal

[fleet bridge]

hmm, okay
"structurally the same" maybe?

[quartz wave]

that's a better way to term it, yes

[quartz wave]

1+5j is an example of a constant but technically not a literal since it's an operation applied on 2 literals

[fleet bridge]

is print('foo') a constant? it evaluates to None pretty much always

[quartz wave]

well not if print is redefined

[fleet bridge]

then replace print with some weird reliable way of getting it, and ask the question again

[quartz wave]

getting "it"?

[fleet bridge]

__builtins__['print']('foo') or something like that
I assume that nobody is touching built-in namespace

is __import__('os')._exit(0) a constant? 🤔

[quartz wave]

any variable to me is by default a non-constant

and anything obtained from a constant is technically a constant

so that's why ().__class__.__base__ is a constant

but print('foo') i have a hard time defining

[fleet bridge]

so constant expression also must use only constant expressions as parts of it
by that logic [].__len__() is not a constant, because it uses non-constant as a part of it

[quartz wave]

under normal circumstances that's a constant

print('foo') is also a "constant" under normal circumstances but it does something else other than return a value

it's like an undefined error happening in my brain when function calls are involved T^T

[fleet bridge]

(0 for x in []).gi_frame.f_globals['__builtins__']['print']('foo')

[quartz wave]

constant UNC

it's like the halting problem now

"determine whether or not this line will do the same thing for the entirety of the program"

[fleet bridge]

is (a thing that returns 1 if Riemann hypothesis is true otherwise returns random number) a constant