#voice-chat-text-0
1 messages · Page 349 of 1
why are we muted
!voice
Can’t talk in voice chat? Check out #voice-verification to get access. The criteria for verifying are specified there.
where r u from alisa
I learned how to solder when I was in primary school
it was ok
Didn't do a good job
same
but I have unlearned it since
but didn't hurt anyone either
I think it's quite easy to guess
nope
just that once lol haha
never touched it again
Sure, what I meant was, it is easy and safe enough to start on your own.
(I might've done it before school, but wouldn't really consider those times learning)
but like anything, it will be difficult to master
A look at the Record Runner novelty record player. A new take on the 1970s Sound Wagon - AKA the Vinyl Killer, the little VW van that drives around a record, both playing and destroying it at the same time.
CORRECTION (updated 15/3/18)
Replace the word ‘Tally’ with the word ‘Tare’
---------------SUBSCRIBE------------------
http://www.youtube...
ohno I forgot to do the obligatory mention of XML literals when Scala was brought up in the conversation
I love this!
lol i saw the esoteric python channel and things were so cursed
!e
print(any(sum((a)for(a)in(range(I))if(I)in(sum((a)for(z)in(bytes(b)))for(b)in(range(I)))or(a)is(int(not())))in((I)for(z)in(dir()))for(I)in(int(input())for(i)in(chr(not())))))
h
@wind raptor can I steam?
!stream 1028671867785068574
✅ @whole bear can now stream until <t:1725030442:f>.
!e
print((lambda: lambda f: lambda: (lambda: lambda x: lambda: f()(x()(x))())()(lambda: lambda x: lambda: f()(x()(x))())())()(lambda: lambda g: lambda: lambda x: lambda: 1 if x() == 0 else x() * g()(lambda: x() - 1)())()(lambda: 100)())
:white_check_mark: Your 3.12 eval job has completed with return code 0.
93326215443944152681699238856266700490715968264381621468592963895217599993229915608941463976156518286253697920827223758251185210916864000000000000000000000000
note how it actually works
nested lambdas are for lazy evaluation
that one is quite systematically written
either lazy lambda: not_lazy or not lazy lambda arg: lazy
uh
destructors?????
they, like
exist
the only reason to manually close is to catch errors on sync
fsync
manual close is only for sync error handling
!e
code
i.e. you need it for writes
not reads
sooooo
when you end the main function by calling return does it call the destructer of the ifstream?
and does it know what the buffer is
if its heap allocated
does it not need to be specefied
yes, close is called when ifstream is dropped
or does the bit insertion from getline work
and the buff gets deleted if its heap allocated
right?
and how does it recognize what the buff is
buffer there is an implementation detail
or does it manually auto de-alloc at main func end
since you're not doing direct IO
ohno
another thing added to C++ without me knowing
what is native hande?
so each open file gets a file handle right?
no
wait
you get a ifstream that you can add into
then you can close it
what site you using for cpp docs?
mostly
ifstream is a stream not a handle
yes
soo what is the handle
it's a wrapper/abstraction around a native handle
on Linux, file descriptor is just a number that the kernel recognises as pointing to an open file
or a TCP socket
or a UDP socket
ahh
or an eventfd
or a pipe
or an epoll fd
or whatever you make it to be
on Windows, handles are a bit more limited
i didnt know that youre into this lol
average gigachad code golfer:
and what does this compute?
@whole bear using relative path for file would be better then hard coding fullpath.
what the shit
all three are well-structured in their original form, they just look weird
agreed but for now Im using full path since im sometimes running this in a higher folder
relative paths would break
how would I access the handle
wait until C++26 releases
!stream 732291998274420828
✅ @sour imp can now stream until <t:1725031975:f>.
imo discord.py is completely fine if you know what you're doing and prefer higher level of control by default
!stream 1028671867785068574
✅ @whole bear can now stream until <t:1725032039:f>.
e.g. when no automatic sync is the intended behaviour
and for linux/unix/posix, whichever manpage that is more complete/up to date
bro is literally talking like this: https://youtu.be/_ls_BpNFm-c
greetings
I feel Osyra's SSD is just a graveyard of forgotten LLMs.
The ones he downloaded and no longer uses and fogort to delete.
Dev, how you doing my good sir?
!paste
If your code is too long to fit in a codeblock in Discord, you can paste your code here:
https://paste.pythondiscord.com/
After pasting your code, save it by clicking the Paste! button in the bottom left, or by pressing CTRL + S. After doing that, you will be navigated to the new paste's page. Copy the URL and post it here so others can see it.
Is Kastien lagging, or is it because my CPU is being tortured again?
issue's on your end, if you're asking about audio
Yeah, he sounds robotic.
50 shades of grey between me and my CPU.
"Oh you're a bad computer, calculate TEBD faster. Yeah, you naughty CPU, running slow even with njit."
I just noticed that the code from quimb that's taking most of runtime is already Numba jit enabled.
quimb/tensor/decomp.py lines 324 to 332
@njit # pragma: no cover
def svd_truncated_numba(
x, cutoff=-1.0, cutoff_mode=4, max_bond=-1, absorb=0, renorm=0
):
"""Accelerated version of ``svd_truncated`` for numpy arrays."""
U, s, VH = np.linalg.svd(x, full_matrices=False)
return _trim_and_renorm_svd_result_numba(
U, s, VH, cutoff, cutoff_mode, max_bond, absorb, renorm
)```
I love this feature for the bot. Just pastes the block of code. Chef's kiss.
Is there anything I can do about this?
Dr. Gray's code is already running with numba apparently?
@verbal zenith I have added an altered version of that codewars exercise to the list
https://parrrate.github.io/exercises/exercises/cwf.html
I really like your website
I'm curious, is it supposed to show some implementation or just the example
ah show hidden lines
oh is it supposed to be an exercise?
some have initial (broken) code
some only have tests
some don't even have tests
I should make Linked exercise harder by obscuring what lines need to be fixed
||```rs
fn with_slice<T>(f: impl FnOnce(&str) -> T) -> T {
let s = "te".to_string() + "st";
f(&s)
}
fn main() {
let string = with_slice(
|s| String::from(s)
);
assert_eq!(string, "test".to_string());
}
Is this thee intended solution?
one of many, yes
the point is to understand why original doesn't compile
@sour imp in the first mini-game add some "bombs" that shouldn't be clicked or they will make you lose points
compiling
@upper basin yes but not much
@njit # pragma: no cover
def svd_truncated_numba(
x, cutoff=-1.0, cutoff_mode=4, max_bond=-1, absorb=0, renorm=0
):
"""Accelerated version of ``svd_truncated`` for numpy arrays."""
U, s, VH = np.linalg.svd(x, full_matrices=False)
return _trim_and_renorm_svd_result_numba(
U, s, VH, cutoff, cutoff_mode, max_bond, absorb, renorm
)
most of the time you don't need @njit
because numpy is already responsible for high performance
is the performance mostly same with and without the profiler enabled?
I don't know the specifics of that function, so can't say how to optimise it
import cProfile
sequential_encoder = Sequential(TKETCircuit)
def run_sth():
mps = MPS(statevector=state, bond_dimension=512)
circuit = sequential_encoder.prepare_mps(mps, num_layers=1024)
print(circuit.return_cx_count())
cProfile.run("run_sth()", "sequential_encoding_optimized.prof", sort="tottime")
so many hacks to make incorrect code be displayed as correct
https://parrrate.github.io/exercises/exercises/linked.html
bet will do!
Testing this now on a smaller example.
@verbal zenith this one is quite useful in practice if you solve it the right way
https://parrrate.github.io/exercises/exercises/get_functions.html
custom version of this?
https://doc.rust-lang.org/beta/std/iter/struct.Peekable.html
what does it currently return when the stream is over?
EOF token?
let mut tokens = Vec::new()
while let Some(char) = self.current {
...
}
tokens.push(Token { kind: TokenKind::EOF, ..})
return tokens
Token::new(TokenKind::EOF, self.span_from(self.current_index), "")
if self.current.kind == TokenKind::EOF {
None
} else {
self.current = self.lexer.lex_next();
Some(self.current)
}```
and implement FusedIterator
that just returns that flag
oh wait
I confused with FusedStream
FusedStream has a method that says if it's done
This trait should be implemented by all iterators that behave this way
Iterator::fuse has a specialisation for FusedStreams
wild idea:
implement Clone
this way you can peek however far ahead you need to, then go back
if you're reading from String not file
pub struct LexerIterator<'contents> {
exhausted: bool,
lexer: Lexer<'contents>,
}
impl<'contents> Iterator for LexerIterator<'contents> {
type Item = Token<'contents>;
fn next(&mut self) -> Option<Self::Item> {
if self.exhausted {
return None;
}
let token = self.lexer.lex_token();
if token.kind == TokenKind::EOF {
self.exhausted = true;
}
Some(token)
}
}
Split for strs implements Clone
pub struct Lexer<'contents> {
filename: &'static str,
source: &'contents str,
char_indices: std::iter::Peekable<std::str::CharIndices<'contents>>,
current_char: Option<char>,
current_index: usize,
pub tokens: Vec<Token<'contents>>,
reporter: ReportSender,
}
seems fine
macros
though that too is invoked by parser
pub struct ReportSender {
sender: Sender<Box<Report>>,
}
thats looking great!
but multiple consumer functionality is hidden
I have a question.
Thanks.
current implementation of std::sync::mpsc is based on crossbeam::channel
@vocal basin
I was thinking of making my error report function implement google translate so that you can set a preferred error language. if one is not found it uses English.
sudo code:
errorReport(string str)
{
if lang == null || lang == english:
// do normaly
}
else {
if internet() == true:
try:
translate msg
// do rest
except:
// do normaly
// mention that there is no internet so other langs are not supported
}
Syntax:
/*Normal multi line comment*/
// one line comment
/@ErrorLang: English@/ (special comment congaing debug info for the interpreter)
// rest of the code
this should only be opt-in (disabled by default)
well it is
when you enable a Error lang it checks for internet
then makes a http request to translate a msg
by default it uses internet and does not check anything
Brb
this one currently has no error reporting?
Option<Result<Token>>
i cant deal with my own shit anymore
Traceback (most recent call last):
File "/Users/andreas/Desktop/Code/RemoteFolder/TensorOps/examples/tensorops/mlp.py", line 35, in <module>
y = model(X)
^^^^^^^^
File "/Users/andreas/Desktop/Code/RemoteFolder/TensorOps/tensorops/model.py", line 96, in __call__
return self.forward(input_node)
^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/andreas/Desktop/Code/RemoteFolder/TensorOps/examples/tensorops/mlp.py", line 24, in forward
input_node = layer(input_node)
^^^^^^^^^^^^^^^^^
File "/Users/andreas/Desktop/Code/RemoteFolder/TensorOps/tensorops/model.py", line 201, in __call__
return self.forward(forward_inputs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/andreas/Desktop/Code/RemoteFolder/TensorOps/tensorops/model.py", line 196, in forward
activation(self.input_nodes)
File "/Users/andreas/Desktop/Code/RemoteFolder/TensorOps/tensorops/model.py", line 291, in __call__
return self.forward(forward_inputs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/andreas/Desktop/Code/RemoteFolder/TensorOps/tensorops/model.py", line 287, in forward
forward(self.context.nodes)
File "/Users/andreas/Desktop/Code/RemoteFolder/TensorOps/tensorops/node.py", line 358, in forward
node.compute()
File "/Users/andreas/Desktop/Code/RemoteFolder/TensorOps/tensorops/node.py", line 191, in compute
self.value = self.node1.value * self.node2.value
~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~
TypeError: unsupported operand type(s) for *: 'float' and 'method'
least shit stack trace:
If my code takes 285 seconds, and with cprofiler takes 403, is that an issue?
missing () somewhere
nice ty
330232382 function calls (318645293 primitive calls) in 403.795 seconds
Ordered by: internal time
ncalls tottime percall cumtime percall filename:lineno(function)
33825 99.376 0.003 99.592 0.003 decomp.py:324(svd_truncated_numba)
56353 52.557 0.001 55.944 0.001 numeric.py:932(tensordot)
359852 27.571 0.000 89.730 0.000 autoray.py:30(do)
6961264/26625 7.983 0.000 22.532 0.001 copy.py:128(deepcopy)
35858754/35809602 5.195 0.000 9.338 0.000 {built-in method builtins.isinstance}
33814 4.932 0.000 5.069 0.000 decomp.py:682(qr_stabilized_numba)
67650 4.390 0.000 129.483 0.002 tensor_core.py:415(tensor_split)
that's expected
Otherwise seems consistent.
Don't have the patience to do a massive experiement right now.
Ngl, I would feel betrayed if the cprofiler messed with the profiler run.
"I trusted you to show me the truth, and you corrupted it."
Thanks, I'll try it now.
built-in search in any definition no longer sounds outstanding for AI
fml
X = [Node(random.uniform, requires_grad=False, weight=False) for _ in range(10)]
random.uniform(-1,1) fml
you were spot on though
I've just looked at 7 separate cargo projects that were recently updated and had 10~50K downloads
all have 450~550 downloads per version ever released
time to start a conspiracy theory about how blindly assume that all those downloads are from automatic tools that just regularly download everything
@verbal zenith that thing I've shown on stream came from this
https://docs.rs/tracing/latest/tracing/level_filters/struct.LevelFilter.html
where funny thing is how, despite the implementation being private, you can exhaustively match on all values
.expect("lmao")
AF what do I do with this?
http://localhost:11235/
as in
it might require some instrumentation on python's side
marking what to instrument
@lavish rover iterators and streams have options of results
because they expect the option to be outer
i see
@verbal zenith
(just so it's written down)
enum Token {
Lt,
Gt,
}
type TokenOrEof = Option<Token>;
const EOF: TokenOrEof = None;
const LT: TokenOrEof = Some(Token::Lt);
const GT: TokenOrEof = Some(Token::Gt);
you need separate Eof to know where it is, right?
let token = self.current;
match token.kind {
kind if predicate(kind) => {
if kind != TokenKind::EOF {
self.advance();
}
Ok(token)
}
TokenKind::EOF => Err(UnexpectedEOF.make(token.span).with_message(message).into()),
kind => Err(UnexpectedToken(kind)
.make(token.span)
.with_message(message)
.into()),
}
writing to string?
afaik String implements Write
curl CLI or libcurl?
only used wrappers around it
you need to ensure both header and library are accessible
@vocal basin What is MakeFile?
instructions for building the project
https://github.com/lucianoforks/tetris-os/blob/master/Makefile
btw is that for Cmake or ninja or smth
@verbal zenith it implements a different Write
you can write! to it
@verbal zenith what crate is that?
String: std::fmt::Write
ah
I just realised
you need Vec<u8>
String can't std::io::Write because the latter operates on bytes
Is this correct?
Python by default : Sequential. Runs line by line.
Threading : Concurrent. Switches back and forth between threads really fast.
Multiprocessing : Parallel. Abides the GIL rule by spawning multiple interpreters to run threads separately in parallel.
nogil Multithreading : True parallel. Doesn't follow GIL rule, and uses a single interpreter to run multiple threads at once.
@verbal zenith have you imported the trait?
@verbal zenith you're using wrong reference
seems like
you need &mut not &
Sorry to bother you @amber raptor. Could you kindly let me know if this is accurate if you get a chance please?
where asyncio
I am trying to sort of explain different modes of executing a python code for my understanding.
Of course assuming the explicit modification using the appropriate package.
I'm sorry for asking this again and again.
Looks right. No idea about no Gil since I haven’t looked into it.
Thank you very much.
My reference is this.
From pycon 2024.
And the pycore podcast. Now that I'm saying it, perhaps I should have first read the PEP regarding this.
I sort of get confused between threading and multithreading prior to nogil.
I feel like multithreading is more appropriate for nogil threading, aka the image on the right.
idk how to order them properly
sequential
asyncio
pre-3.13 threading
3.13 threading
3.12 many interpreters
multiprocessing
not entirely
it's still a single process
you can share some data within that process
multiprocessing existed before 3.12
Is the left image multiprocessing or 3.12 many interpreters?
latter
So, how would you draw multiprocessing compared to 3.12 many interpreters?
"Let it snow"
since you're in IntelliJ IDE, stdout/stderr are very fast
What is asyncio?
!d asyncio
Hello World!
import asyncio
async def main():
print('Hello ...')
await asyncio.sleep(1)
print('... World!')
asyncio.run(main())
```...
Yeah, but in relation to modes of execution.
not much different
Like compared to threading?
most of asyncio runs on a single OS thread
So, not concurrent?
But you're saying it's on a single OS thread.
Doesn't concurrent basically need two threads and would switch back and forth real fast?
concurrency:
A starts
B starts
A finishes
B finishes
asyncio achieves that without OS threads
tasks not threads
I thought a thread was a sequence of tasks.
asyncio allows tasks be executed non-sequentially within a thread
I have to review my terminology then.
!e
import asyncio
async def run():
for i in range(3):
print(i)
await asyncio.sleep(0.1)
async def main():
await asyncio.gather(run(), run())
asyncio.run(main())
:white_check_mark: Your 3.12 eval job has completed with return code 0.
001 | 0
002 | 0
003 | 1
004 | 1
005 | 2
006 | 2
So, concurrent but within a single thread?
Instead of switching left and right (where left and right are two threads), it switch back and forth (up and down) until it finishes execution?
yes
Threading:
T1 : -A-B->
T2 : -C-D->
Asyncio:
T1 : -A-B->
?
!e
import asyncio
async def run():
for i in range(3):
print(i)
await asyncio.sleep(0.1)
async def main():
await asyncio.gather(*(run() for _ in range(1000)))
asyncio.run(main())
:white_check_mark: Your 3.12 eval job has completed with return code 0.
001 | 0
002 | 0
003 | 0
004 | 0
005 | 0
006 | 0
007 | 0
008 | 0
009 | 0
010 | 0
... (truncated - too many lines)
Full output: https://paste.pythondiscord.com/QHSJ6BIG77V4EH7I55IGNTO3KY
!e
import asyncio
async def run():
for i in range(3):
print(i)
await asyncio.sleep(0.1)
async def main():
await asyncio.gather(*(run() for _ in range(10000)))
asyncio.run(main())
I wonder how far can it go
:white_check_mark: Your 3.12 eval job has completed with return code 0.
001 | 0
002 | 0
003 | 0
004 | 0
005 | 0
006 | 0
007 | 0
008 | 0
009 | 0
010 | 0
... (truncated - too many lines)
Full output: unable to upload
asyncio implements its own task switching instead of relying on the OS
true threading (assuming no GIL)
---------->
---------->
asyncio
-- - --->
--- - >
I see. So concurrency also depends on how many threads you have. It can switch back and forth between threads, or do so within a single thread. Difference being that one uses a single thread and uses its own logic for switching, and the other uses multiple threads and uses the OS logic?
Uhh, you have two threads though.
No?
test your code as you go. It’s way easier to catch bugs early than to fix a mess later. Writing tests alongside your features saves you from headaches down the road and gives you confidence that everything’s working as expected.
I thought threading with nogil basically runs two threads in parallel.
second one has two tasks and single thread
Threading with gil would switch back and forth instead.
Above one has two tasks and two threads?
Sorry, got a lil bit confused.
threading (blocked on GIL in red)
--[31m---[0m-[31m-[0m--->
[31m--[0m---[31m-[0m-[31m---[0m>
asyncio
-- - --->
--- - >
(I remembered Discord can do ANSI)
task is a sequence of instructions in this case
asyncio can interleave them
tasks can contain other tasks, and thread as a whole can be viewed as a task
Also, ran a smaller example, here's the result
359677464 function calls (347195532 primitive calls) in 1175.105 seconds
Ordered by: internal time
ncalls tottime percall cumtime percall filename:lineno(function)
36900 363.062 0.010 363.318 0.010 decomp.py:324(svd_truncated_numba)
36888 278.150 0.008 278.404 0.008 decomp.py:682(qr_stabilized_numba)
61476 177.170 0.003 183.327 0.003 numeric.py:932(tensordot)
393682 65.371 0.000 258.464 0.001 autoray.py:30(do)
Without profiler it took 1045 seconds.
two threads, each running asyncio
- --[31m-[0m-- >
-[31m--[0m-- [31m---[0m-->
-[31m----[0m- >
[31m--[0m- ---[31m--[0m>
but A can take longer right? like A can finish after B too?
(that was just an example of two cases possible)
yeah lol
the proper is:
A starts
B starts
A finishes
(this sequence is enough to demonstrate concurrency)
and you'll have to keep polling until b finishes?
since technically an async func would return a future on its own?
wait, no, actually B being able to finish after A finishes there is somewhat important
so that one is not strictly contained within the other
because
AAAAAAAAAAA
BBBBBB
there B might be just a subcall of A
sure
@vocal basin is math.ceil or __ceil__ better? Or are they the same?
math.ceil is more idiomatic
Any reason not to use math.ceil then?
there might be a performance difference, not sure
Cool, thank you!
Support Primer on Patreon: https://www.patreon.com/primerlearning
Check out 3Blue1Brown: https://www.youtube.com/c/3blue1brown
Books that inspired this video and have more information:
- The Selfish Gene, Richard Dawkins, https://amzn.to/2LpffQl
- An Introduction to Behavioral Ecology, https://amzn.to/2YspPQz
(If you decide to buy one of these ...
Hi
Markdown Test
!paste
If your code is too long to fit in a codeblock in Discord, you can paste your code here:
https://paste.pythondiscord.com/
After pasting your code, save it by clicking the Paste! button in the bottom left, or by pressing CTRL + S. After doing that, you will be navigated to the new paste's page. Copy the URL and post it here so others can see it.
rounding
?
9.5 more hours
you joined 3 days ago, you cannot have completed 72 hours
of days
you have been hereo n 3 different days, that's true
yesterday, today, the day before
?
look at my oldest message
not 3 whole days yet, since last joining
Doesn't matter, you joined barely 3 days ago
it has been 3 days
not quite
and ive been here 3 days
nope
like there are a span of 3 days ive been here
you joined on wednesday
ive also been here the last 3 days
3 days there means 72 hours
no like
You joined 2 days ago, on the 28th
today is the 30th
why is this so complicated?
you joiend 2 days ago
ive been here for several hundred hours
that's great
why does that matter
yea i dont have any clue what that has to do with it
it checks when you joined, has it been 72 hours? no? not 3 days
really simple stuff
thats not what it says
i joined 6 years ago
it actually does
and this week
i did
2 days ago you joined
i also did yes
as far as the Discord is concerned
It is the last time you joined.
i joined both more than 3 days ago, and 3 days ago
three days of duration
yea
i've vc-ed here before
3 is 3
yes i did
Your most recent join is what is checked
None of us can do anything about it, not sure why I'm wasting my time
you elft
you lost your roles
i didnt have a roll
why did you leave?
WTH is happening
If you have VC you had a role
if you did not have a role, you are lying about being verified previously
verification brings with it a role
!e
from datetime import timedelta
seconds = timedelta(days=3).total_seconds()
minutes = seconds / 60
hours = minutes / 60
print(hours)
:white_check_mark: Your 3.12 eval job has completed with return code 0.
72.0
this is what enables me to speak
they wnt to talk, they do not want to wait
i have talked before
that's great
zulu?
you've talked a lot today haven't you
yea why is that punished
Its not a punishment
you lost a privilege because you left
it 100% is
It's really not
it limits my ability to communicate
I can't get back in the same spot
you get back to the end of the line
welcome back~!
yea thats stupid
That's life
no
welcome to reality
its actually not
they're basic rules in place that prevent spam
that's pretty much how it works everywhere
cool
go join your country's discord where it is relaxed if we're too not relaxed here
there's rules for a reason
take more than 3 days time and relax as much as you want
keeps VC happy
no it keeps it an in-group
lol
because joining is such a... tedious exclusive tasak
that is comical, an in group? Yes, the people who have spent a few days in teh server and sent messages
so uh... people who want to spend time here get to speak
I spent several months and 2000+ msgs
not really an 'in-group' by most definitions since it's so easy to get in
cool ,then you elft
why is leaving to be punished
you've got the message count
i also have the time
oh wait I remember who this is
what a 🦜 reference
if I burn my passport, why can't I have a new one immediately?
not consecutive though
the ship it as soon as its lost
I remember them too
No, you need to file paperwork, wait for things
i have done it
nothing is immediate
that's great, but you elft
lost your permission, soon you can get it back
yes usually 2-3 days
i know this because my passport was destroyed and i got a new one the next monday
that's how long you have to wait after joining this server before yo uapss the time requirement to get VC
@whole bear tl;dr: this person likes to argue dumb stuff without any backing reasoning, as many people have had to deal with previously
Question, are you arguing with us because you believe we will issue you permissions to talk, or because the principle annoys you?
since you asked for explanation
you lied
thats it
i am saying the system is broken here
and why
why allow for a broken thing
cool, I'm saying it works as designed
because that's how it was designed, and it works, per the spec
I agree.
seems like we're at an impasse here
maybe ytou can come back tomorrow and we can tlak it out in VC
based on interactions from ~year ago we may just assume they're a troll lmao
you are just kinda dumb here
this is wrong
this is the not it
you do not interact kindly
you take my words for lies
i do not know why
i show you proof of what i mean
you do no such thing
insisting something about 72
and i dont know why
what proof are you talking about
What the actual fuck? 72 consecutive hours from joining must be had
why is that so hard to comprehend?
imagine being a troll and begging for attention at the same time
You must be in the discord for more than 3 days from your most recent join date. That's the only fact we need, what else do you want?
why is it this way
who's mean?
just a reminder: if you try to use existence of your message history for arguments, be ready that people will read it
100% it was a ride going back through those arguments
why are you a mod
They are not
you seem kinda mean
why do you not read roles?
oh i see
If oyu've been here so much, you should know yellow = helper
because they are listed as having moderation capability
yellow what
yes i use discord
color blind?
- I'm not a mod
- I'm only mean to people who deserve it a lot
how do you get it to do that
How do you get it to not do that?
you have compact mode on
My best guess is you changed a setting
I did not
Discord has always shown members in role colors for me across all my devices
it does it on my phone
so I can't tell you what you changed, because I did not change it for myself
o dear
well, seems like it's a you thing
does not disable colours
oh
yea im kinda lost here
ew you too
find the nearest map
I enabled it for tests
see why i think people are mean
took me a few tries to even find it
they dont take my words to be good
that's a fun one
no like
in accessability it turned it to sync with computer
think
yeah that's an opinion you can have
lol
i was trying to make it less pointed, which you used as a weapon
Less pointy weapon? That's gotta hurt more
blunt trauma is a real headache
I'd take a sharp blade impaling me over a lead pipe
okay but i was trying to make it a non-weapon stmt
granted either option seems pretty shit
i want to de-escalate
that's tough when you already escalated it
you need to take my words not disingenuinly
like a whole, "nevermind, please play nice now" after starting a scuffle at the playground
yea
i dont want to hurt you
i dont want to make fun of you
i want to be able to understand why the rules exist as they do
now that i understand that they do
@rapid crown
Nice, glad we're clear on the verification 🙂
yea
This channel is locked for peoples mental health
-# This message was approved by jo mama [Discord]
i dont want to make enemy
do not abuse markdown
k
Rules exist as a way to establish order, fairness, and predictability within societies, organizations, and various systems. They are shaped by a combination of factors, including human behavior, cultural norms, historical precedents, ethical considerations, and practical needs. Here are some reasons why rules exist the way they do:
1. **Maintaining Order:** Rules help prevent chaos by setting clear expectations for behavior, which is essential for any group or society to function smoothly. Without rules, there would be no agreed-upon standards, leading to conflict and disorder.
2. **Promoting Fairness and Justice:** Rules aim to ensure that everyone is treated equally and fairly. They are designed to prevent discrimination, favoritism, and unjust practices, creating a level playing field for all.
3. **Protecting Rights and Safety:** Many rules are established to protect the rights, safety, and well-being of individuals and communities. For example, traffic laws exist to prevent accidents and injuries, while laws against theft protect property rights.
4. **Reflecting Cultural and Ethical Values:** Rules often reflect the moral and ethical beliefs of a society or organization. They encode what is considered right and wrong, acceptable and unacceptable, based on shared values.
5. **Ensuring Functionality and Efficiency:** In organizations and systems, rules are necessary to ensure that processes run efficiently and effectively. For instance, rules in a workplace might govern how tasks are completed or how decisions are made to maximize productivity.
6. **Encouraging Accountability:** Rules establish clear responsibilities and consequences, holding individuals and entities accountable for their actions. This helps in preventing abuse of power and ensuring that people adhere to agreed-upon standards.
7. **Adapting to Change:** Rules can evolve over time to address new challenges, technologies, and social changes. This adaptability helps societies and organizations remain relevant and effective in a changing world.
8. **Providing Predictability:** By setting expectations, rules create a predictable environment where people know what to expect and how to behave, reducing uncertainty and confusion.
Rules are not static; they evolve in response to changing circumstances, challenges, and collective learning, ensuring that they remain relevant and effective.
Rules provide structure and order in society. They help establish clear expectations for behavior and promote fairness and safety for everyone.
MMHMHMHh
context from VC, as to not !rule 10: this was generated
technically doesn't violate rules, but might be confusing (rule 10)
but why does this that
Well it's not disabled. only your mic is, and based off this conversation I'd say the rule was a well worth it investment
i dont know you
you are just being mean to me
please, stop
bot not recognising prior messages as counting towards the total timer because:
- single source of data
- easier to implement
if you want to fix that behaviour, make a PR to the bot's source
okay
!source voiceverify
Unable to convert 'voiceverify' to valid command, tag, or Cog.
idk where it's now
but couldnt save itself
!source voice
Unable to convert 'voice' to valid command, tag, or Cog.
Run Python code and get the results.
hmm
Indeed this is a bad argument
so it works for !e but not for !voice
if everyone in conversation understands it's AI-generated, it's fine
42
don't we only allow media?
there is !paste for text files
I did encountered some very cursed syntax issues
can't remember an example right now, but it is possible to run into something non-trivial
happens more often near async, generators and comprehensions
object.__new__
or just call to super()
!d pathlib
Added in version 3.4.
Source code: Lib/pathlib.py
This module offers classes representing filesystem paths with semantics appropriate for different operating systems. Path classes are divided between pure paths, which provide purely computational operations without I/O, and concrete paths, which inherit from pure paths but also provide I/O operations.
If you’ve never used this module before or just aren’t sure which class is right for your task, Path is most likely what you need. It instantiates a concrete path for the platform the code is running on.
Pure paths are useful in some special cases; for example:
Lib/pathlib.py lines 343 to 351
def __new__(cls, *args, **kwargs):
"""Construct a PurePath from one or several strings and or existing
PurePath objects. The strings and path objects are combined so as
to yield a canonicalized path, which is incorporated into the
new PurePath object.
"""
if cls is PurePath:
cls = PureWindowsPath if os.name == 'nt' else PurePosixPath
return object.__new__(cls)```
note: __init__ is called even if you return an object which has been __init__ed before
my_first_calculator.py lines 19264 to 19265
print("35*37 = 1295")
if num1 == 35 and sign == '*' and num2 == 38:```
constructor calls __new__, then, if it's an instance of the type, calls __init__
hmm
I need to check something
!e
class A:
def q(self): print("a")
class B(A):
def q(self): print("b")
A.q(B())
:white_check_mark: Your 3.12 eval job has completed with return code 0.
a
no magic
I only now noticed the cropping on the first image lmao
no, not anymore
I was just trying to remember what I was going to say
and went a bit too far back in message history
infraction count is conveniently not included
it won't fit
what was that infinite binary search called
exponential search
like
you start at 1
then try 2
then 4
until you hit a reasonable upper limit
then just binary search
when values are more likely to be towards the lower limit
e.g. when you need to check between 0 and 2**65536
Can anybody help me please
it is mostly for cases when there is no list
big integer arithmetic
also post screen shots not photots 😭
Alright
please please please read the logs yall they are pretty helpful!
the logs: tracing::warning!("lmao")
I don't think I've ever done that in messages of less-than-ERROR errorness
next week?
λ
just refresh
@verbal zenith have you seen what node-ipc did? do that but for syntax error instead of geolocation
hello selenium my old friend.. time to break out the automaton
why is geckodriver on arch repo out of date
meanwhile python where everything is both by reference and by value
nothing is preventing you from implementing methods other than next
oh
I remembered what I was searching for
(after, what, an hour?)
brb
@sour imp 200 billion NDAs
going back to discussion of line count in codebases,
some time I ago I made a tool which visualises how long unique lines of code live for within a codebase;
just as another way of displaying project history, especially rewrite rates
(afaik KastienDev hasn't seen those graphs yet)
the only good thing about it is that it turned out more or less exactly how I imagined it to be
longest-compiling dependency in biggest Rust project I'm working on is written in C++ not Rust
tbf, yeah, Go compiles faster than Python starts up lmao
if you go for compile times, then, yes, Go
@verbal zenith why do you have stored_space not equal to indent on initialisation
I have not
here width per commit is adjusted by lines changed
so slope corresponds to added-deleted ratio
I should rewrite it in Rust at some point, because now it's, like, 10 separate python scripts
pickling data in and out
in a >20 stage Dockerfile
@remote imp
!d ast.unparse
pip install pyaes
ast.unparse(ast_obj)```
Unparse an [`ast.AST`](https://docs.python.org/3/library/ast.html#ast.AST) object and generate a string with code that would produce an equivalent [`ast.AST`](https://docs.python.org/3/library/ast.html#ast.AST) object if parsed back with [`ast.parse()`](https://docs.python.org/3/library/ast.html#ast.parse).
Warning
The produced code string will not necessarily be equal to the original code that generated the [`ast.AST`](https://docs.python.org/3/library/ast.html#ast.AST) object (without any compiler optimizations, such as constant tuples/frozensets).
Warning
Trying to unparse a highly complex expression would result with [`RecursionError`](https://docs.python.org/3/library/exceptions.html#RecursionError).
Added in version 3\.9\.
wow that took a long time
both Python bot response and introducing the function to the module
pip install pyaes
ye but its still red
another reason might be multiple python interpreters
!rule 5
5. Do not provide or request help on projects that may violate terms of service, or that may be deemed inappropriate, malicious, or illegal.
nope
I did not read the screenshot the first time
that not good
k
just went for rule 5
lol
after reading
well its just a py lib
Wait is that a Bank graber or a blank grapper as in a whitespace?
Uhhh, yeah we cannot help with any token grabbers.
oh wait
it's zip-downloaded from github lol
not even cloned
I always forget that it's a feature
my brain does not insta-process when I see such stuff on screenshots
can you close the help thread?
helpers can close threads
uh
no work
Oh wow
Make the help post this command was called in dormant.
Helpers could close our old help system channels
bot/constants.py line 306
cmd_whitelist: tuple[int, ...] = Guild.moderation_roles```
@verbal zenith *(index+array)
but also must work for *(key+map)
@verbal zenith [](){}()
[]<>(){}() if C++ allowed empty generics
[]<T>(){}(), I think, is valid
force parentheses when mixing bitwise and arithmetic
make precedence ordering non-total
@verbal zenith -x ** 2
from maths
.wa -3^2
@verbal zenith ;capture;parameters;body;arguments;
so this becomes ;;;;;
capture nothing, take no parameters, do nothing, pass no arguments
let x = ;:;:;:1;;;()()()
print(x) // 1
how would factorial work
#voice-chat-text-0 message
huh just do
print(x)```
@verbal zenith 226
probably the most I have on a repo
though there's a nuance that this one's a screenshot of a tag that's an amalgamation of three projects' histories
!e
print(23+428+1313)
:white_check_mark: Your 3.12 eval job has completed with return code 0.
1764
and 9 commits which come from nowhere
sleep schedule? what's that?
(it's 4 a.m.)
website on my IP is safe
it just redirects to a rick roll
6 servers is for hidden services
@raven gale attacks have happened before in the dumbest way possible:
just own so much nodes that it's statistically probable the user picks only ones you control
https is free enough
send your website i wanna see if i can dox it or smth
!stream 1028671867785068574
✅ @whole bear can now stream until <t:1725068320:f>.
Podolsk, Russia
Yerevan, Armenia
which one do you like more
russia
!stream 373121697957543946
✅ @wind raptor can now stream until <t:1725068591:f>.
!stream 732291998274420828
✅ @sour imp can now stream until <t:1725068613:f>.
Thanks for this! I have saved it
@sour imp iirc Docker Desktop requires payment somewhere around $1M too
$10M annual revenue
Sacha Baron Cohen made a laughing stock out of some U.S. lawmakers with his new series on Showtime called 'Who is America?' Our Steve Langford has the story.
I found a PR I have open since 2019
as for PRs that actually matter I so far only waited 4 and 2 months
Angular
*dramatic pause*
JS
so it's a Jinja2 question
templating
I only know how to do it on front-end side, not in Flask's templates
(is Flask still using Jinja2 or is it Jinja3 already?)
rn trying to find how to iterate in Jinja2
I'm reading through that and having Angular flashbacks
with its pipes for data transformation
what are we hashing again
if you're making your own hashing:
- stop, use existing schemes instead
- if you don't stop, learn how it works: use slow hashing of salt+password and store salt+hash[:however], then compare for equality
!e
from os import urandom
from hashlib import sha256
salt = urandom(16)
password = salt + "clementina♡".encode()
for _ in range(10_000):
password = salt + sha256(password).digest()
print(password.hex())
:white_check_mark: Your 3.12 eval job has completed with return code 0.
3c17bdfb4067f43073eb78c6950a67acf3daf293e015c36a104131cd0658c9d67ca74ef4d2d0238fb1083df24ed456d8
!e
from hashlib import sha256
hashed = bytes.fromhex("3c17bdfb4067f43073eb78c6950a67acf3daf293e015c36a104131cd0658c9d67ca74ef4d2d0238fb1083df24ed456d8")
salt = hashed[:16]
password = salt + "clementina♡".encode()
for _ in range(10_000):
password = salt + sha256(password).digest()
assert password == hashed
:warning: Your 3.12 eval job has completed with return code 0.
[No output]
initial hashing
validation
why you shouldn't use that exact method I used there:
low memory usage, and a relatively cheap algorithm as a whole
🤖 Download the Source Code Here:
https://brandonhancock.io/claude-crash-course-templates
Don't forget to Like and Subscribe if you're a fan of free source code 😉
📆 Need help with CrewAI, join our FREE Skool Community:
https://skool.com/ai-developer-accelerator/about
📣 New Premium Fullstack AI Course:
By the way, I’m currently pre-launching my...
AI can solve it in quite compact manner
https://parrrate.github.io/exercises/exercises/rcchars.html
sure it will have questionable semantics
but it will work and be safe from the outside
I'm quite sure I learned almost the entirety of stable Rust
(not mastered but at least aware of)
and unlike C++, there are no hundreds of inadequate weird edge cases like std::vector specialisation
kernel in what sense
Go works good enough
many design mistakes but overall not critical
OS kernel in Rust
😦
wsg
cant speak TwT
I don't lol
gotta be a member for 3 days min
yeahh
honestly just joined to see if I could get some help on a project im working on 😭
oof lol
to scared
I actually don't remember what they use for main servers they build
TwT
either Linux or illumos
👋
it is indeed not Linux
I see they too are fans of big workspaces
https://github.com/oxidecomputer/crucible/blob/main/Cargo.toml#L1-L27
rather quiet voice chat tonight
It's probably for the best. I should be studying, not chatting.
@young matrix hello sir
do i need to voice verify to talk in voice chat?
Yes
!voice
Can’t talk in voice chat? Check out #voice-verification to get access. The criteria for verifying are specified there.
nvm i need to fulfill the requirements
can you use py to make code for a signal jammer?
interesting
wdym by signal jammer?
exactly what it sounds like, I was looking them up previously for fun to see how they work and there IS A company that makes them as small as every day people use to $45k military, then I looked into making one and a guy was showing how it works and he showed his code and I have no idea what the script was.
but it wasn't py
like what this code actually do?
jams signals and stuff, like radio waves and cell phones.
owh
what did you think I meant?
i thought something like a ddos attack script
That is completely different
well still you can do something like that
I wonder if it would be easier
yes and no
i can make a script that can interact with radiowaves
It's a grey area deepening on your use
budget?
anything you can get at Microcenter
If I ever make one I'll be getting all my parts from there
i live in bangladesh and i have no idea what is a microcenter
Microcenter is a giant toy store for nerds of all kinds
I take it you dont watch Linus Tech Tips then
I wonder if I could reverse signal jam tho