Thread 106971061

>keeps out-of-order designated initializers out of sepples to make constructors seem useful (they have been present in c since c99)
What eas his endgame

>>106971061 (OP)
He's coding C++ compilers in C# + AOT
>look what they did to my boy

>meme arrows
>"sepples"
>"endgame"
either be serious or go full meme bait.
never do half measures.
enjoy your dead thread.

>>106971061 (OP)
C++ is the least shitty shit if you have to build complex systems with performance in mind.
Yeah I hate C++ but it's still the best option.
But of course for random small apps C# is ten times more comfy.

>>106971061 (OP)
>language designed by european
>it's a mess
lol

>>106971737
>C++ is the least shitty shit if you have to build complex systems with performance in mind.
Rust exists today.

>>106971737
>C# is ten times more comfy
honestly? how? it just looks like OO java garbage. all the examples I see just look much worse than Rust even, which is surprising since C# is supposed to be a GC'd VM language.

>>106971881
and to be concrete, even premeir shit like ASP.NET Core just looks assy.

Handling multipart in Axum:
https://github.com/tokio-rs/axum/blob/main/examples/stream-to-file/src/main.rs#L88

Handling multipart in ASP.NET Core based on the M$ examples:
https://github.com/dotnet/AspNetCore.Docs/blob/main/aspnetcore/mvc/models/file-uploads/samples/9.x/FileManagerSample/Program.cs#L29

why do you even need to get the MIME boundary? like what? and all the stupid _private name vars and everything? really?

>>106971881
>it just looks like OO java garbage.
Most of the code I write in C# is functional code.
And what's wrong with OOP? It's just an optional tool for safe state management. It has it's own place.

I bet you think that you have to include every language construct you've ever read about in your code.

>>106971861
>Rust exists today.
They recreated all the problems C++ has.

>>106971953
>Most of the code I write in C# is functional code.
ok, so show me some snippets or something. I just don't believe you. when people say "muh functional" and they show me F#, Ocaml or babby's first anonymous function use I just immediately scowl and walk away.

>>106971960
Async is shit - true.
But I bet async in C++ is unmanageable hell, knowing even the very basics are fucked up in this disaster of a lang "designed" by a (committee of) retard(s).

>>106971964
Watch the videos of this Chad:
https://www.youtube.com/@zoran-horvat/videos

>>106972014
>Async is shit - true.
it's literally one of the best, cross-platform abstractions possible though. there really was no other way to do it. not sure why people seethe about it.

C++ coroutines are a meme.

>>106972021
can you just link to one? ffs. I ain't clicking random videos.

>>106972025
Extremely poor orthogonality.
It doesn't fit well with the other language features.
It fits so poorly they literally had to develop a workaround (pinning) to the key feature of the language - its move semantics - to make it work at all.

>>106971061 (OP)
>What eas his endgame
cashing "committee" checks until he dies. and it's not an endgame, it's the whole game.
>>106971737
seepeepee was and will always be the most shit option.
the choices were:
C (-> Pascal) -> C++
now it's:
Rust -> C -> C++
in three years, it will be:
Rust -> Zig/C -> Carbon -> C++
and after that, something better than Rust may emerge, and seepeepee will drop off the list completely. it is already dead as a "new project" language in many fields already, thankfully.

>>106972071
>1% of types cannot be moved in memory, that means it's over
>even though self: to all the appropriate trait impls can just be Pin<&mut Self> and the hardest you have to think about it is projecting pin on members correctly.
like, I don't get it. are people still really seething about Pin? you're going to have hard problems no matter what approach you take and move by default was objectively correct for 99% of cases, and even without it, you'd still have problems of specifying things that shall not be molested.

>>106972089
Try to write something more complicated than your school projects!

>>106972089
>C (-> Pascal) -> C++
man, I still deal with Pascal stuff. It's not that good either.

>>106972143
this is hilarious because "school projects" and "blog posts" represents 99.9% of /g/eets's experience with programming (if any).
>>106972129
you're conversing with a /g/eet (possible llm) who web-scraps for "rust problems" and spam them endlessly. "it" doesn't actually comprehend anything.

>>106972149
system languages don't have to be very "good" vs. C++, because no matter what. they will have infinitely less layers of shit than C++.
no one can catch a language that isn't designed, but rather had a committee with the sole incentive to keep adding shit to it for decades.

>>106972089
>rust
>performance
errr, no.
just no.
its decently fast but its way slower than c
unless you write c. in rust. which completely defeats the purpose of the language.

to be fair it may find a use in osshit and servers, if you have to deal with a constant personnel churn.
i wouldnt use it as it is now, but theoretically rust could bring new stuff to the table.
still, its not a solution desirable to any degree when performance is your chief concern

seriously
why the fuck would i bother with safety mechanisms only to promptly ignore them?
why would i care about a standard library if im gonna reimplement everything myself anyways?
makes negative sense. also bc if you min max into memory retention by learning everything there is under the sun, well, you minmaxed into memory retention.
at the cost of your problem solving.

>>106972089
nothing useful and actually real is done in rust, only toy showcase projects and virtual signaling endeavors inside big companies that will get dropped when they realize what a waste of time it was to build an inferior version of something that already existed.
meanwhile c++ powers the world's most important and relevant projects and that alone will keep the demand for it alive.

>>106972555
well at least the "branchless" dot finding retard finally moved on to simd intrinsics.

>>106972752
>t. seething at superior skills
kek. best compliment you could have given me

>>106972752
btw
this IS branchless
while in the past two years or so i graduated from poor mans sse to full blown simd
you still didnt even learn the basics

>>106972089
Thing is every time I check on these memelangs, they don't have the same kind of libraries C++ does.

>>106973507
you're still on that. what kind of one-trick-pony tardation is that.
i think you disappeared before when this was linked:
https://docs.rs/memchr/latest/memchr/

>>106972660
what do you use for OTP

C++ is so bad at organizing code even python is better

>>106973550
>they don't have the same kind of libraries C++ does.
At least you can use them in Rust, C++ has atrocious package and dependency management.

>>106973909
i had to reimplement an atoi/f/l and i dont have the fast boi code on my machine anymore
so instead of retyping half of an inferior solution based on a screenshot i just implemented it properly with my current level of skills
>memchr
you use sized strings.
i overallocate
we just arent the same

the practical difference is that i deal with everything using simd intrinsics only.
i dont have a tail end i have to iterate through, everything is treated as a 16/32 byte long vector
its a wholly different beast mr branchless
diffreent constraints, different solutions

>>106973909
>>106974125
the fast boi solution is already 3x faster than the libc btw
so if you benchmarked your memchrs and atois and such against the libc
my poor mans SSE is already faster than your standard shit

>>106974125
btw btw
>the practical difference is that i deal with everything using simd intrinsics only.
>i dont have a tail end i have to iterate through, everything is treated as a 16/32 byte long vector
which is gonna make an enormous difference with my data
most of my shit is ~5-8 chars for the integer part, and 6 for the decimals
so with a sized string, simd wouldnt even properly kick in most of the time

>>106971061 (OP)
He made the language overly complex on purpose.

>>106974646
>ideology instead of engineering concerns
this never works out
i agree that the current problem is people who are undertrained
but turning a language into a steaming pile of bloat is not a solution to that

>>106974646
>>106974845 cont.
he should have stuck with c principles
keep the lang as simple as possible
and fuck the midwits, theyre gonna filter themselves out
with none of that safety shit sepples seems to be interested in nowadays
or 30 different containers for that

fundamentally, sepples should have never turned its back on raw pointers
its should have been C with more powerful interfaces
as opposed to C but with standard, separate interfaces for everything
including scratching your balls or passing a gas

its just my unsollicited 2 eurocents, but that do be what im thinking

>>106975298
C ain't simple retard.

>>106971711
Apologize

>>106975325
its syntax is
the complexity emerges from how complex the machine youre building, is
and the subtelty you have to put in the syntax.
and if you use intrinsics its because of autism
autism can be a major factor too in completely unnecessary difficulty

but at its root
c is dead fukken simple

>>106975461
(pigrel is a WIP btw. in case the printouts werent explicit enough. its just to illustrate i do program in c. if i speak its not bc im a fanboi or bc its a culture war talking point. also i find switch/case fallthrough kinda neat, as a pattern)

>>106974646
he isn’t wrong about pythonniggers infecting everything, but his solution reminds me of the Terry Davis quote
https://youtu.be/k0qmkQGqpM8

Every pythonnigger gets filtered by C, even though C has like 5% of Python’s features.

>>106975566
>Every pythonnigger gets filtered by C, even though C has like 5% of Python’s features.
c is the true measure of a man
c is exactly as complex as the machine one's building
not more, not less

>>106975566
>>106975581
also the deal with features is replacing brains with procedures
instead of needing to understand stuff
brainlets can now learn by heart some prosneedures that do things on their behalf

notice how its mainly autists who are into man pages shit
kinda fukken dismantles the idea they have a superpower, doesnt it?
kinda explains why theyre DISTURBINGLY SUSCEPTIBLE to propaganda, doesnt it?

>>106975566
>>106975581
>>106975614
t. pythonigger btw
its a webshitter who taught me programming

i had to relearn everything with c
but i started with paiton/php/js/html/css
your most vanilla webshitter stack possibly imaginable ~ 2010

today i do c and opencl and anything else makes me miserable.
i could accept glsl and cuda, or vulkanshit whatever
even metal would do
but i dont want to deal with high level shit at all if i can help it
FUKK learning others spaghetti. mine's a handful enough

>>106974242
whatever you're doing is not faster than this:
https://docs.rs/memchr/latest/memchr/memmem/struct.Finder.html
provide numbers to prove otherwise.

>>106973929
authy

>>106975461
>its syntax is
No, it's not. You are mistaking simplicity by lack of features and even in that case it's not simple because there are implicit rules and UBs for almost everything. I would not define C as simple but as rudimentary (Go lang fits this category too).

>>106975581
Nothing created in C can be elegant. https://rentry.co/g7aofwhc

>>106975503
What is that doing? Converting a string to float? Do you know that converting a string to float is more complicated than that, right?

>>106975784
>authy
lmao

>>106975986
retard

I started a new C project the other day and decided to target c23 and who the fuck thought it was a good idea to bring constexpr, nullptr and auto (but severely limited) to C? it looks like shit.
you'll pry NULL and #define from my cold dead hands.

>>106975765
ye, benchmark it bc im not dling 20 gigs of framework to compile crabcode
fukken 20 gigs but also bc fuck crabcode
https://files.catbox.moe/0t62xe.c
i didnt implement decimals yet
which is a goto. and assignments to pointers bc despite trying to turn my vecs into literals
its still a pointer
anyhoo
test it against integer floats bc the code as it is just doesnt take decimals
its still gonna be 2 orders of magnitude faster htan your fukken generic fukken slop

>>106975825
u sound like youre seeting
yes, you have to pay for proper spelling for "seethnig"
i have no idea how to do that anonymously though...
>>106975895
dl the catbox
krsh krsh krsh and such

yeah im expecting you to kneel
fkn hoers
youre below me
its only fitting that you will serve
this is my fizzbuzz btw
this is the code i make public
bc this is the minimum level of competency i expect from anyone who says "i kinda know c"
only fukken kinda btw

>page 2
unacceptable
this discussion
the fucking retards in it
should be fucking stickied
so were clear
that fucking crabs
are b class citizens
and should ask for permission when they talk

right?
its only just that this happens, right?
why would we want fucking retards on our board?
if it isnt t make fun of them
and to demean them

to treat them like the fucking garbage that they are, right?
fucking crabs
whit their fucking crustacxean brains, amirite?

keep trying crab fucking retard
theres something like 10% perf to be gained on top of what i do
when your niggershit is fucking ORDERS OF MAGNITUDE slwer than my code
fucking go ahead
take your fucking time
use all the flags you want

use all of them
>page 2
how?
im still here
ill keep fucking bumping
post your fucking code, crab NIGGER
im still here
fuckng post it

>page 2 gain
ok
crabniggers should shut the fuck up forevermore bc they cant even compete for a fucking atoi
im sorry but htats the real truth
rust = internlang
if not a complete meme
>inb4
yea i dont care

fukken
wo a challenge ima think ima c virtuoso
lads, keep the discourse at a high level or the seltzer is gonna get to my head

and crabs
i fucking expect more of you
i come here to be challenged
this is the big fucking secret
idfc about you reflecting my own competency
i know exactly what im capable of

what im here for is to be challenged

>>106976001
>have opinions about programming languages
>using non open-source "app" on a phone for otp
>calling others retards
quintessential /g/eet stuff

>i never compiled "crabcode" and was just making shit up in >>106972555
good concession

>>106976697
>just making shit up
i posted code though
feel free to do the same
or bear the indignity
yuo fukken crab

code or gtfo

>>106976675
this is the most pathetic attempt at an argument I have ever seen.

>>106976697
>>106976713
>or gtfo
actually dont
who am i gonna call "crab" in a hateful manner otherwise...

btw
the code is beautiful, isnt it
the only actually ugly part is the cases but thats whats the fastest. i think
and the fukken cpu autists cant create decent interfaces...
please, understand
i literally cant do better withouth turning the code into macro hell

i think

aaaand the compatibilization bw cpu endianness and c endianness
which includes the manual masks
which are also ugly, admittedly

lets say my code is perfectly imperfect
the beauty of the sum imperfections of all the craftsmen that worked on it
unless you find a more elegant, at least equally fast way, of doing things

>>106976713
>stuck for months on one function for some reason
>"rust is slow"
>won't post numbers
>"i'm not going to compile crab code"
>continues to spam confidently
my autistic homies are not like this.
https://crates.io/api/v1/crates/memchr/2.7.6/download
that's a .tar.gz source package with zero dependencies ("framework" or otherwise).
post the numbers. or keep owning your autistic+retarded self with thread bumps.

also actually instead of inverting, then applying right shift, i think i can apply left shift
thats ~5 ccles of around 250 that can be shaved off
also i use reverse asssigment elsewhere but i dont think it has a cost compared to orderly assignment
but thats something that still can be streamlined

and then actually thinking about the scopes to avoir register pressure even if i think i fit within the 16 i have access to
also i can still clean up the load ops
actually give it a fink
but it all amouts to, idk, 10% of the runtime?
in something that already exceeds design specs bc the idea is to latency-hide reads from a drive?
ill profile that, but i think im alrdy 10x faster than my iops
its likely that i should rationally close the subject and move on

>>106976723
tard maxed: bloated closed-source OTP phone and/or desktop app, possibly with "cloud" features

full tard: same as tard maxed, but the desktop bloated solution is possibly open source, and no "cloud" features

half tard: FreeOTP or some other f-droid app

non tard: cotp

smart: writing your own solution based on a reliable and safe library (probably in rust)

----
you know your score

I think this thread proves that C fans are mentally ill people. While C may be an okay language, it still sucks.

>>106976848
>>stuck for months on one function for some reason
>poorfag sse vs proper power
i mean
are you dum?
you must be dum
that would check out bc in the past 1.5 year i went from doing poorfag sse to actually be comfy with avx 2
but you still dont know what branchless means.
in 1.5 years.

>>106972555
>>rust
>>performance
>errr, no.
>just no.
Interpretive dance perhaps. Or maybe mime.

>>106976948
>having to deal with the details is a bad thing
unless thats what you need to get 95% performance for 5% of the effort
which seems to be something you never needed...

>>106976941
who the fuck cares about it being open source. free software is a meme and nothing good ever came out of it. the only good projects are ones that started as free software but got adopted by (a) large corporation(s) like blender and maybe linux (the kernel only).
you are retarded because you can't refute what I said about your meme language, rust, and decided to randomly pick an irrelevant topic you think you can gotcha me but that has nothing to do with what I said

>>106976962
when its the difference bw waiting several hours and several months for you computation to finish
it has nothing performative to it
dance, maybe, but the performer is the only audience
nobody gives a shit, otherwise
maybe except the crabs
bc prosneedure is like dancing with your left hand tied to your right leg

>>106976991
>i'm a neet and there is nothing of value to lose from others having access to my accounts for me to be concerned

>>106976982
To be honest I prefer to code SIMD in assembler, I only use intrinsics for prototyping.

>>106977154
completely valid
if your brain bends in that dirction, more power to you
i dont have as much granular control as you, but using c i get to use a wider brush

>brush
i think its an apt analogy
i just use a wider brush than you
>would you paint mona lisa with a wide brush?
ofc not
but would you paint a park bench with a fine brush?
that would be irrational

>>106977154
that >>106977187
was just a fancy way of formulating amdahls law i realized. kek.

>>106977154
>>106977207
also: moment of honesty-
the difference bw the how the processor sees things and how things are presented in c is a kinda fuck in the ass
i had to debug this shit before in understood whats up.
twice.
that was completely unnecessary.
i absolutely understand why you stick to only one notation
that exact thing was kinda fucking bothersome

>>106977033
I work in the industry, unlike you who thinks an app being open source is of any relevancy and wastes his time learning a dead toy language instead of doing actual work

>>106975339
yeah it didn't go down quite as i expected...

>>106977488
>tech illiterate and retard
>"whether an app on my phone connected to the internet and having all my credentials is open-source or not is irrelevant"
>still opining about <literally anything>
lmao