>>109227374
I don't know, you tell me. I want to work on something cool.

How do I force myself to stop caring about QoC, Features, and Performance, and just complete a project?

First off there is code quality
>tell Big Pickle to write js
>it works, but it's written poorly
>arrow functions everywhere
>everything in one js file
>bad variable names
>not DRY enough

Then there is the scope creep
>image search using hashes
>oh how about a color filter
>oh how about a nsfw filter
>width, height, file size

Then there are performance optimizations
>0.5s queries? I want millisecond queries
>render time is 0.5s? I want milliseconds for that too.

All this makes a hobby project extend from a day into 1 or 2 months.

Should I blaze through a shitty fully functional, slow prototype, and then "fix it up" if I want?

>>109228021
I have a similar issue lol. Want to make resume quality projects but get caught up on if the code quality is good enough and lose motivation and stop working on them.

>>109228021
if it's a learning project write it and refactor it but don't spend too much time perfecting it. nobody is going to read the code so only do shit which you think are a good exercise.

>>109228275
Almost same here.
But when I do not care about quality and have fun, then I hack something together and as soon as it works I drop it.
I just don't have the motivation to do a minor refactor/cleanup and write up afterwards.
I could use AI for that, but i really don't want to

>>109229037
true if I get something close to working I realize I don't want to maintain it and just completely drop it. Have not once written a readme for anything I've made but I AM changing that this next project I'm doing.

>>109227374
im retarded and im still looking at 3d graphics math to try and use raylib to make a water simulation, i just really want a local program that i made that is as good as webgl water, a small pond that i can drop pebbles into and watch the waves move around

... and that's why /dpt/ will NEVER manage to write a registry dumper that finished in less than 10 seconds. Because these autists are completely and utterly unwilling to put the balls to the walls.

>>109229693
I made a Brainfuck interpreter which wins me good boy points with Jesus.

Now if only Jeshua cares, but unfortunately he died a little bit over 19 centuries ago.

>>109229730
No, He resurrected. He clearly cares about my Brainfuck interpreter.

>He resurrected
Autists say the darndest things.

File: 1783544404308422.png (269 KB, 2560x1414)

269 KB PNG

I ported my timer to ncurses because there was a slight start up delay in SDL3. Not the fault of SDL it's just too heavy for such a simple tool.

>SDL
Come on now.

>>109229893
yes it's effectively an entire game engine now.

>>109229913
>Simple
>Direct
>Layer
What did they mean by this

apparently writing a compiler is the best way to learn a language.
im learning rust. well i kinda know the basics, i went trough the book and can write some basic shit including; sending requests, streaming & reading bytes, parsing json, reading & writing files. i can do some basic gui with the iced framework. i somewhat get how to do async/parallelism with tokio. i know the fundamental theory kinda and yea but also the things Im interested in doing next seem brutally above my level. Making my own gui framework using wgpu/vulkan, making a pdf -> sheet music parser transcriber, steam dlc unlocker, maybe a mod manager ? But then I feel like there's no point because everything that I might need has already been made, and polished over time so I'd just be reinventing the wheel.

>>109229961
Reinvent the wheel is all AI does. It just looks for what you want on the Internet and checks if it's already there to present you with.

>>109229961
Why not write something that would be useful to you?

>>109229770
>He resurrected.
He'd found a +1UP somewhere in the desert, but was keeping quiet about it.

>>109229961
The whole point of Rust is it's memory safety meme built into the compiler.
If you're writing your own compiler you don't need Rust.

File: lifewater_flasks.png (5 KB, 1280x960)

5 KB PNG

>>109230029
You'd have to be pretty fucking retarded to not find them in the first level.

>>109230042
wtf are you even talking about. There is many reasons to use Rust for a compiler, from security and performance to developer experience.

>>109227374
>What are you working on, /g/?
I'm working in nothing, just doing some algorithm exercises from the internet. If i finished it in this week, i will study C or something like it.
And i want some recommendations of programming languages, in my list have: C, Lisp and Ruby.

>>109230068
About Rust, is it really good? I can't understand the point of it. To me is more good just use the classic C or C++ to do it.

File: 2026-07-08-223112_644x409(...).png (39 KB, 644x409)

39 KB PNG

I made my own PRNG in Tcl. How do you feel about that?

>>109230091
Go fuck yourself.

>>109230091
Why don't you just spent few evenings and see it for yourself? If you are already good at C++ and up to date with the new editions, you should have no problems with Rust.

There is a reason why it is popular among both hobbyists and big IT corpos. Yes, within its usecases it is very good, in many ways.

>>109230042
swift and rust are currently the top two languages to write a compiler in btw

>>109230125
>There is a reason why it is popular among both hobbyists and big IT corpos. Yes, within its usecases it is very good, in many ways.
Hm, i will get a time to do it. Thanks by the idea, a new underground programming language to learn for laugh.

>>109230068
It's literally a compiler with guard rails that you need to remove to do anything useful the language itself is a meme.

Security... bwhahahahahah
How's those crates.io exploits going? How many layers of nested shit do you need to audit to make sure your Hello, World! does not steal your crypto keys.

>>109230092
base. tcl is based, but auitism

>>109230153
I hate the rust trannies, too, but c'mon guys. Don't be so silly.
Rust does solve a serious Problem - or at least tries to. And that's memory leaks. It does not solve logic bugs.
There are probably people that pretend something else, but to most it is kown that that's the case.
So just don't be stupid? Hate the grumpy rust trannies. Hate the rust compiler. But hating the theoretical concept behind rust is stupid

>>109230153
>It's literally a compiler with guard rails that you need to remove to do anything useful the language itself is a meme.
It doesn't if you know what you are doing. Git gud.

>How's those crates.io exploits going?
No idea what you are referring to. A quick search of "exploit" in rust-lang/crates.io show nothing noteworthy. Are you sure you mean crates.io website and not some actual rust crate?

>How many layers of nested shit do you need to audit to make sure your Hello, World! does not steal your crypto keys.
Wut?

File: 1778298417816074.jpg (81 KB, 862x1024)

81 KB JPG

>>109230186
>Rust does solve a serious Problem - or at least tries to. And that's memory leaks
No it doesn't? It prevents buffer overflows, heap corruptions, dangling references, not memory leaks.
What are you even talking about.

Wtf is going on with all these nocoders talking random shit they have 0 idea about in this thread lmao

>>109230186
free() solved memory leaks 1975. Just stop hiring retarded third world jeets.

>>109230199
TrapDoor.

>>109230211
>No it doesn't? It prevents buffer overflows, heap corruptions, dangling references, not memory leaks.
>What are you even talking about.
autistic framing.
Yes, you might be right, but you also got my point obviously.
And that's all what rust is about mainly. It's really not that crazy

File: 1409901609892.jpg (43 KB, 372x372)

43 KB JPG

>>109230228
>TrapDoor
?

>>109230233
>autistic framing
?

>>109230242
https://socket.dev/blog/trapdoor-crypto-stealer-npm-pypi-crates
Jesus fuck do you not know how to use the internet outside of reddit.

Every crate depends on another crate so one exploit in the system can cascade down to thousands of crates. The whole system is putrid.

>>109229886
https://github.com/mackron/miniaudio
Should I use miniaudio instead of system()? The way I am playing sound is kinda shit cause I need control over stop start and detecting if a sound is playing.

>>109230263
I don't see where this supposed "crates.io exploit" is described there. It seems that the attacker just uploaded malicious package to crates.io and then used it to steal stuff, inject into AIs and such.
Calling this "crates.io exploit" is like saying you exploit HTTP when you send someone a link to a malware.

>Every crate depends on another crate
That's not true obviously. Otherwise you would need a circular dependencies.
And if you want to put an exploit into someone's else crate then you need to convince the developer to include your malicious code there. This is no different from any other language.

File: 1783548509029860.png (431 KB, 2437x1038)

431 KB PNG

Why does AI lie like this.
Like I knew it was bullshitting me so I asked to clarify it said the same thing to me about linenoise and tinyexpr.

File: 1783548927871453.png (23 KB, 934x229)

23 KB PNG

JESUS FUCKING INDIAN CHRIST OHHHHHHHHHHH NOOOO NOT A 4MB FILE HOW WILL OUR SERVERS EVERY HANDLE THIS.

>>109230363
thta's just casual llm hallucination.
The LLM cannot know how this shitty lib is compiled or added to the project.
And since there are tons of ways which are popular in C/C++ it just picks one of them.
These are exactly the points that come up EXTREMEL:Y often, are frustrating as hell, and somehow AI shills straight up ignore them and do not count them as errors.
This is literally the reason why i dont want to use LLMs anymore. It's just frustrating

>>109230529
no the AI is right, he's just retarded, if you opened miniaudio.c it's just a 2 line file. with the IMPLEMENTATION macro
The reason that .c file exists is that if you copy the miniaudio.c to your source tree, it acts as a single TU (c/cpp) that holds the implementation. In other header only libraries, like stb, you need to choose which file it will go into.
in extras/miniaudio_split you can get a pre-split version. But I think if your goal is fixing compile times, I think it wouldn't actually help as much as you would think, and if you were including miniaudio.h into many files, you are doing something wrong, use an opaque pointer or even PIMPL.

File: checkpoint_4C.webm (2.88 MB, 1920x1080)

2.88 MB WEBM

My 3D mesh modeling project is coming along nicely.

I kinda find it funny that my laziness to learn Blender but a desire to make 3D asset has driven me to create my own editing program.

I don't know what that says about Blender's learning curve for making 3D models.

>>109230186
memory leaks are a low hanging fruit non-problem

File: 1768270714830343.png (208 KB, 588x680)

208 KB PNG

>using variable shadowing to statically assert nesting restrictions within block macros

Rust is based because it solves memory errors. Who gives a frick what sjws think?

>>109231165
unsafe

Java is memory safe.

>>109231165
Rust is a Microsoft invention to turn all GPL code into MIT.

>>109231187
DELETE this goy.

>>109231128
Memory leak detection is an undecidable problem.

What's the benefit of having a garbage collector? Wouldn't it be easier to just use smart pointers?

File: 1783560492661626.png (345 KB, 2560x1414)

345 KB PNG

Audio now properly implemented, this was completely painless.

>myfunc is inaccessible due to 'internal' protection levelSourceKit
>it's public
>the class is public
>restart lsp
>still errors
>restart ide
>still errors
>delete everything in the source file
>save
>crl-z
>save again
>it starts working
swift has to be one of the most dogshit and buggy languages there is

>>109231300
a language has a garbage collector because the language becomes a lot easier to use when you stop caring about performance (and compiled GC languages still run 300x faster than Python on very specific benchmarks), and you also don't have to worry about leaks from cyclic references (you can recover from errors and clean up parts or the whole VM without needing to start a new process).
From a datastructure POV, there are lots of use cases for using a garbage collecting as a method of cleaning up data (you would NOT call this a garbage collector, it's just cleaning up holes in your array).
For example, if you had an array of objects (around 16 bytes~) that don't have references/index pointing to it, and you loop + delete the objects frequently, you can just set the object to a free state (ex: id = -1), and free it during the loop.
In C++ this is can be done with std::remove_if(), and since the pred callback is made for every single object, you can just do your usual logic inside of the pred (remove_if is efficient because if you delete multiple at the same time, the cost of the next element is closer to O(1) but real world performance depends heavily).
For big objects, it would likely be faster to keep the holes and use a freelist at some point (for me, a freelist is just a single linked list, optionally you can sneak the node into the deleted object without extra space), but if you need the array to be sorted, to avoid moving the data, you need to traverse a linked list (you can use a single linked list if you are OK with only looping backwards, or a careful xor list which is a bit annoy to delete from, a double linked list can allow you to delete without traversing/collecting the garbage).
You might be
And if you DID have references, if performance was not a concern

>muh memory leaks
>muh memory safety
99% of all problems could've been solved with better allocator interfaces that _force_ developers to be smart about them. There was never any reason for yet another shitlang, it's ALWAYS been about autism.

Muh shitlangs

Reddit spacing this autism

>muh shitlangs
Yes.

>>109231489
>You might be
>And if you DID have references, if performance was not a concern
OH I was probably going to say something about sending an event to clear the reference, this is pattern I have seen in some games I read. The performance concern is probably because every single reference holder will need to be alerted about every single deleted object (you need to delay deleting the object for 1 extra pass).
shared_ptr is probably faster what I mentioned above, and if you only have weak pointer (AKA 1 strong pointer), you could try tagged pointers (store and compare a unique ID inside a object before reading the data, but it requires you to keep the memory alive and reuse the slot, in C++ you can use placement new to make your own pool allocator + address sanitizer to poison deleted memory so you don't accidentally access the memory)

I love using shitlangs, especially Rust.

File: autism.jpg (115 KB, 828x649)

115 KB JPG

>I love using shitlangs, especially Rust.

>>109231300
>just use unique
>std::move std::move std::move std::move std::move std::move std::move std::move
>just use shared
>oops, cycle

>rust
go is THE shitlang

>there can be only ONE (shitlang)
>>109231590

>>109231633
It's Haskell.

File: EE348A64-4E25-4638-8D11-8(...).png (2.88 MB, 1254x1254)

2.88 MB PNG

Rust is the German idealism of shitlangs.

File: 1685190660684825.jpg (364 KB, 1470x1913)

364 KB JPG

Herr du mein Lieber mein Vater

I am tearing my hair out trying to force the kernel to run a fault handler on stack overflow. I can touch the guard page and it works, but the overflow never does. It just does a default crash and core dump. sigaltstack() doesn't do shit to help

function void FaultHandler(int Signal, siginfo_t* SignalInfo, void*) {
  if (Signal == SIGSEGV) {
    byte* Address = SignalInfo->si_addr;
    fprintf(stderr, "Fault at %p\n", Address);
    fprintf(stderr, "In guard page? %s\n", (Address >= StackBoundary && Address < StackBoundary+StackBoundarySize) ? "true" : "false");
  }

  exit(1);
}

function void OverflowTest(void) {
  uint64_t _;
  OverflowTest();
  return;
}


function void* Entry(void* Argument) {
  fprintf(stderr, "Faulting...\n");
# if 0
  { // custom fault triggers in guard page
    int _;
    byte* Address = (byte*) &_;
    Address -= StackSize;
    Address[0] = 0;
  }
# else
  { // custom fault does not trigger
    OverflowTest();
  }
# endif
  return (null);
}

The kernel must be doing something I'm not understanding.

>>109231693
What are you expecting to happen? Overflow basically means that an operation on the stack pointer register yields more than the previous value.

>>109231737
An overflow means rsp points outside of the established stack memory range. In this case, at the guard page setup to have PROT_NONE permissions. I am expecting the fault to trigger on that page and from within the same thread, execution is dropped into the handler with rsp set to the alternate stack setup by sigalitstack().

>>109231755
>An overflow means rsp points outside of the established stack memory range.
No, an overflow LITERALLY means that an operation has wrapped around.

>>109231759
Why do you try to talk so authoritatively about things you clearly don't understand? Overflow means overflow. In the context of an integer it means wrapping, but not a stack or a buffer or anything involving memory. Do you think the stack pointer is wrapping around the entire 48-bit virtual memory range?

No longer care, autist, don't bother replying, I actively hope you NEVER (!) figure it out.

>>109231788
whatever dude, I genuinely hope you get off mount stupid someday and get more humble and less ignorant

File: fucking_awesome_the_trash(...).png (45 KB, 1060x250)

45 KB PNG

Didn't read, but can you imagine how AWESOME it will be when you just fucking bash your autistic, retarded head in, never figuring it out, and in the end just kill yourself because that's objectively what all autistic people deserve? Makes me wake up with a fucking bonver every single morning.

File: IMG_3540.jpg (36 KB, 600x600)

36 KB JPG

Autist anon is feeling raspberries in his butthole.

As long as autists get to smell it, sure.
Their sensory overload suffering is out of this world.

>>109231809
Okay, you're just a lolcow.

>>109231693
just copy that into godbolt and read the assembly.
I assume you forgot to hook to sigtrap because it calls that instruction ud or int I forgot which.
I now wonder if there are IDE's that have godbolt style assembly viewers.

As long as autists get to suffer and die I don't care.

>>109231827
There might be a vscode extension.

Shut the fuck up Ryan.

>Ryan
Lol
Lmao
Autists have ALWAYS been the scum of the earth and ALWAYS deserved to be tortured torturously.

>>109228275
>Want to make resume quality projects but get caught up on if the code quality is good enough
here is a tip: nobody actually gives a shit about quality, they only care if the project is useful or not.

Ryan will keep lolcowing on.

>>109230186
>memory leaks
does it solve the halting problem too?

>>109232077
Rust uses the double shitlang maneuver to solve the halting problem.

>>109232087
I shouldn't have doubted

File: cereal-experiments.jpg (411 KB, 1754x1240)

411 KB JPG

https://github.com/ran-sama/nftables-rsyslog-iptables

It's not a big deal and maybe could be done better, but I made the actual effort to do ingress chains. Either way, I prefer my logs silent.

>>109227374
I have never written erlang

>>109231128
>>109232077
cope.
What's so hard about appreciating that rust at least try to improve some important part of critical software?
Yes, it has all pros and cons, but it's just pointless arguing. ]g] is really stupid

>>109232474
Why not, anon?

>>109232487
you are saying incorrect things. you claim Rust solves something it doesn't, which makes your post completely pointless and reveals your unfamiliarity with the subject

>>109231971
>nobody actually gives a shit about quality
Try telling that to my former employer. I *still* remember him complaining to me about their contract with another company.
>and guess why I quit

ncurses is the dumbest thing in existence

if you want a gui make an application not a terminal based on one. i don't want graphics in my fucking terminal

>>109233165
works for me

>>109232632
Ok no problem. I wanted to get back into C++ anyways

the threadly argument about rust

FASCINATING

File: comfortable.png (24 KB, 1280x576)

24 KB PNG

>>109233426

boards:g;/adhd|autis|food|factorio/i

Stop giving it attention.

Why is this forum not moderated?
I just want to talk about code. I don't like autists

>it
And that's exactly why outists WILL be turned into foood.

File: awj24e.jpg (370 KB, 1080x1749)

370 KB JPG

>>109234083
go find a code forum for whatever faggot ass language (python prob) that u write in?

>>109227374
I am currently making a little website that runs locally that just monitors your docker containers you are running locally. Only doing it to look good on a resume and show I know how to use docker. In the end you'd just have to clone the github repo, and run docker compose to run it on your local machine. Trying to grind resume projects for tech I don't use at work to show competence and hit keywords for job descriptions

File: PXL_20260709_142739170.TS-vp9.webm (3.88 MB, 1080x1920)

3.88 MB WEBM

>>109234099
At the moment I am tinkering a bit with the pi pico. So C and holyC mostly

>>109231693
Holy Skill issue.
Windows lets you handle this via SEH.

>>109231776
He obviously doesn’t know the difference between a hardware stack and a stack in DSA.

>clear skill issue

>>109234354
>boomer language 1
>boomer language (useless) 2
when are you larpers going to pick up a real language like rust

>>109231540
>he needs to use a handicap for oop to even match what rust gets out of the box
you are right about one thing, your banal complaints about an objectively better language are ALWAYS autism...

>>109231827
oh I was probably thinking about SIGILL, not sigtrap.

why is glibc's allocator so bad still?
I forgot how bizarre it behaves and had to switch to jemalloc.

>objectively better language
Autism.
Autism.
Autism.

File: 1775226888759062.png (496 KB, 660x720)

496 KB PNG

https://leetcode.com/problems/path-existence-queries-in-a-graph-i/description/?envType=daily-question&envId=2026-07-09

implementing 2fa with authenticator (yeah webshit) and this stuff is insanely complex for what it does. i'm using cognito and there's like 6 http calls between client and server just to login. i can see why login pages are so shit and slow.

>>109235266
>6 http calls
idk shit about that but as a mobile monkey its simple
>set up cognito
>send username + password to cognito
>get a response asking for 2fa
>send 2fa code
>get auth token
ez

>>109235363
if you want email as fallback (no preferred method in settings) the whole process changes and requires an additional 4 calls.

>>109234952
There’s no reason for a C systems programmer not to pick up Rust instead of C++.

>>109235266
ya, it's really insane just how full retard the whole thing is and we keep ignoring that stuff like ssh auth (including ssh cert auth) exists and is so painless in comparison, but nope. need to hyperbloatmaxx webauthn spec that forces you onto fallable, fiddly bullshit hardware. Hell, even Kerberos would be less retarded if it could evolve away from its heavy dependence on DNS and central KDC.

>checking project from 3 years ago when I was a complete retard
>to my surprise I was doing concurrency correctly and preventing data races
what the fuck has happened to my brain I've turned to MUSH

>>109235363
>>set up cognito
already too bloated. do not pass go.

>>109235629
how do you know it is correct anon? C++ memory model is notoriously annoying. No, spamming SeqCst doesn't fix it.

>>109235090
>I'M A SURGEON
>I'M A SURGEON
>I'M A SURGEON
Sure bud, great arguments.

Aotists don't deserve arguments, only pain.

>Aotists
They're evolving.

>only pain
>pOOPers projecting this much
you get what you deserve

>you get what you deserve
Exactly, tist.

>>109235629
too much reliance on ai agents everyone's brain is fried

>>109232474
Same, but I've written Elixir. Didn't care much for it.

>>109227374
I think I might be retarded lol. Every time I want to make something if I think it needs some sort of like auth I get scared to try and even use like oauth2 or something because i dont want to publish something that isnt 100% secure and get whatever users I get (most likely just myself) screwed by not having proper security. Always end up trying to keep everything local

How do you answer the question "Is there a function in this library that does X?" without google/AI/stack overflow?

>>109236273
Same, but I've never written Elixir. Didn't care much for it.

>>109231590
I will choice the technical things and political things, muhahahhaha. And you will never stop me.

>>109231640
Holy shit, is more good just learn Lisp.

>>109236353
> open doxygen or other reference sheet
> ctrl-f
> type some keywords
> open eyes
> ????
> Profit!

https://www.youtube.com/watch?v=UNo2-viKfW8

>>109236365
Based

File: 1766195976588458.png (34 KB, 623x385)

34 KB PNG

Stop replying to the mexijeet spammer

Fuck. Lost the joy of coding again :(
Somehow I want to get that sweet little 12h/day hyperfocus, but i don't have a project idea that sparks joy.
What do? Just do something and hope for joy?

Fuck. Gained the joy of coding again. This time it's with C.

>>109236971
My condolences.

>>109237027
>>109236971
His condolences

wtf is virtual memory

>>109237417
There's no other good way to deal with fragmented memory.

File: Virtual_memory.png (79 KB, 500x791)

79 KB PNG

>>109237417

>>109237470
>here's your contiguous array bro

File: 260709.png (23 KB, 666x460)

23 KB PNG

>>109235126
Not the fastest, but easy.

File: 1783635250323848.png (170 KB, 1400x954)

170 KB PNG

>>109231346
miniaudio blew out my compile time and annoyed me but I realised I can just use SDL_Audio on it's own in ncruses and initialization remains virtually instantaneous.

Cool little tid bit this works perfectly fine in a tty.

File: think.jpg (28 KB, 510x392)

28 KB JPG

Say if i wanted to buy books about various tech related stuff, what are some must-haves.

I'm interested in low level languages, kernel dev, graphics programming and computer architecture .
Also where do you guys buy physical books for cheap in europe?

>>109237680
>buy physical books
You mean rent digital books, right anon? You're not some kind of terrorist are you?

>>109237470
We have enough memory. Why add such overhead?

File: 1783636256680821.png (321 KB, 1483x659)

321 KB PNG

>>109237680
I'd like to buy physical books but the pricing is absolutely jewish.

after getting over the initial learning hurdle of all the major core library features, syntax and stuff it feels more and more to me that programming is just reusing smarter people's algorithms and learning other people's libraries if I actually want to make something functional/useful.
Is this really it ?

File: 51D6TQNYBML.jpg (47 KB, 377x475)

47 KB JPG

>>109237680

>>109237680
>Also where do you guys buy physical books for cheap in europe?
Second hand bookstores. They still exist.
More of a problem is that they don't usually carry non-fiction books worth learning from.

>>109237696
you never know lel
>>109237746
This. i hate pirating books and reading them on my computer it hurts my eyes and i get distracted easily and can never get a decent deep session, unironically the reason my brain is deteriorating can be linked to this
>>109237752
Thanks anon, noted
>>109237796
exactly, you hardly find any decent tech books its just garbage and outdated c# and java shit

>>109237742
One: directly contradicts all the complaining about muh RAM prices.
Two: even IF we did: caches.
Three: ou-tism.

>>109237823
How are scientists and disinformation experts supposed to update your books if they're physical?

>>109237823
>Thanks anon, noted
iirc you can't really find a good pdf online and it's hard to get as physical book.
Also very old, but it explains computer architecture really fucking good

>>109237748
That is my conclusion as well

>>109238050
>>109237748
You could use Java swing for GUIs or just C/C++ for terminal interfaces.

File: 178364109242639.png (67 KB, 1294x1143)

67 KB PNG

Just created an account on codeberg, what am I in for?
https://codeberg.org/headlesschook

I now spend 90% of my time debugging shitty AI code "written" by the juniors. Fuck my life.

>>109227374
They named a coding language after a football player woaw

>>109238386
i mean thats sounds like a cushy job no ?
you don't have to come up with anything original. just read trough stuff and fix mistakes.

>>109238406
Cushy? I guess, but it's hard to get out of bed for. I used to have genuine love for programming and that part of me is dying.

>>109238393
Do coffee lovers and niggerballers just not know what to do with themselves? Who can like this shit?

>>109238386
I spend 100% of my time debugging code from juniors. And seniors. And clankers (from both of them).

>>109237752
Tanenbaum or OSTEP is better.

>>109238379
Hopefully you used a temp email and VPN anon?

AI is just a portable penis chopper.

>>109238580
email yes, VPN no.
How come?

File: 1755833964716545.gif (1.54 MB, 230x230)

1.54 MB GIF

https://leetcode.com/problems/path-existence-queries-in-a-graph-ii/description/?envType=daily-question&envId=2026-07-10
>Use binary jumping with a sparse table data structure.

>>109239282
I'm gonna broot.

>>109238393
No worse than naming it after a chemical reaction of iron and oxygen, or just naming after som stupid snake

>>109238421
just don't treat your job as anything important lol
nothing stops you from having fun on your own time

>Vulkan setup is already atrocious
>pipeline caching is a headache waiting to happen
>shader languages are full of inconsistencies and legacy nonsense (location vs binding)
>L1 is getting to the point where SSBOs are as fast as uniforms
I fucking hate graphics rendering so much.

>>109238379
>what am I in for?
for an extremly slow website.
But apart from that it just works. I guess github fucked up too much and now they got too much traffic

>>109240120
They're being slopped by jeets.

>>109238379
erm... your license??? you can't just host all rights reserved code on codeberg

>>109240330
You can do what ever you want with my code idgaf.

File: DeaDBeeF.png (41 KB, 830x423)

41 KB PNG

has anyone ever wrote a plugin for this bad boy?
I started to do some more old school music listening and now i have tons of playlists, but picking them on deadbeef is crap.
A vertical playlist tab would be nicer and I am too lazy to get into those cool cli music players or generally get used to a new one rn.
Can I do it? Can AI do it?

File: toot.png (159 KB, 2540x781)

159 KB PNG

how will non-coders ever recover?
https://mastodon.social/@mitchellh@hachyderm.io/116892134982408422

>>109240402
eh, nothing personell kiddo ;-)

typedef struct DB_gui_s {
    DB_plugin_t plugin;

    // returns response code (ddb_button_*)
    // buttons is a bitset, e.g. (1<<ddb_button_ok)|(1<<ddb_button_cancel)
    int (*run_dialog) (ddb_dialog_t *dlg, uint32_t buttons, int (*callback)(int button, void *ctx), void *ctx);
} DB_gui_t;

bro just implement that. It's easy!!
(it probably is easy, just looks weirdo)

>>109240502
uh what? This is valid C kode? I have never seen this before, I think.
It kinda makes sense for readability tho..

static DB_gui_t plugin;


static DB_gui_t plugin = {
    .plugin.api_vmajor = DB_API_VERSION_MAJOR,
    .plugin.api_vminor = DB_API_VERSION_MINOR
};

>>109240502
>>109240517
Ok looks like it's a bit harder. I thought i might jsut be able to rearrange things in the gui but
>There can only be one GUI plugin running at the same time, selected using the "gui_plugin" config file option.
And the GTK3 frontend is a plugin. so I'd have to modify that entire stuff and then compile it as new plugin. Hmm will see

>>109240517
>he has never seen a function pointer that takes a function pointer as argument
ngmi

>>109240720
I probably have, but i have never seen a variable declared twice and it not throw an error because one of them has different initilization

Question for all professionals. Is CS still worth it? I'm 28 and thinking about going getting a degree. I do feel too old to do it but what the hell.
One thing that makes me take a step back is AI and it's hype train. I have coworkers that cant code without AI fullstop. Is this truly the direction of CS?
>First time coming to /g/ in years and the first thing I see is a vibe coding general

cool but kinda ugly new meme lang which makes ECS a first-class feature.
https://flint-lang.github.io/

>>109241340
I guess we are back in 2016 again.
You should probably only do it if you really want to. It's not money printing and it's not infinite jobs. But it can pay off, if you love it.
With all the faang hype we just got too many people and I guess it will slowly adjust itself, but this can take quite a bit. The real experts will always find something

>>109241340
if you're going to pay for a degree then no period, otherwise gaining knowledge is always a net positivity

>>109241377
Thanks. I don't expect to make bank. Those times are long gone by now. I except not wanting to blow my brains out. My CEO, coworkers see it like the second coming of christ.
>>109241400
Just in it for the knowledge and piece of paper really. In my country (Portugal) degrees are a requirement.

postgres rewritten in rust, passing all tests

https://github.com/malisper/pgrust

Cniles.. it's over

>>109241575
very nice.
now lets see those benchmark results and any regressions

>>109241575
>Claude
Into the trash it goes

Is it true that people are migrating en masse from GitHub to Codeberg? wtf did they do to fuck up this bad? I'm out of the loop and I'm struggling to keep up with recent events lately.

>>109241587
>>109241609
To concisely give an overview of the project, I've been experimenting with using LLMs to build a better version of Postgres. Postgres is 30 years old and we've learned a lot about databases since hten. A lot of the techniques that work for doing a rewrite are also useful for doing a rearchitecture.

I'm now working on a new, not yet published version of pgrust that incorporates a lot of techniques. Currently the new version:

- Passes 100% of Postgres regression suite
- Implements a thread per connection model instead of the process per connection model Postgres does
- Is 50% faster than Postgres on transaction workloads
- Is ~300x faster than Postgres on analytical workloads. Right now it's 2x slower than Clickhouse on clickbench and I think it's possible to get faster than Clickhouse

>>109241624
>en masse
lol no. maybe 0.0000001% of users

>>109241628
>been experimenting with using LLMs to build a better version
oxymoron

File: 1775913942074957.png (274 KB, 736x981)

274 KB PNG

>>109241637
>Is 50% faster than Postgres on transaction workloads
>Is ~300x faster than Postgres on analytical workloads

>>109241575
It's a good thing if it's faster and safer.
What's your point? AI replacing humans? No, the intelligence is in the test suite, not the code itself. AI can't write that.

>>109241624
AI training and typical microsoft shenanigans

>>109241624
Github feels noticeably worse than it used to be + rate limit nags you if you aren't logged in.
Sometimes it will randomly hang on loading basic things.

How do I extract the data from an open source website using python but without the title and all the other stuff that's not the data? I just want to extract the car names, makes and their details.

Why wouldn’t you create a GitHub/Codeberg portfolio using a temp mail and VPN?

Once your projects are up, they can be traced to you. Future employers will be able to discover that.

>>109243053
I don't wanna work for Jews.

File: file.png (169 KB, 1058x439)

169 KB PNG

noob here.
i know allocating memory here isnt the most optimal approach but also I just want to fuck around.
why doesnt this work. the output is some memory garbage.

>>109243154
ptr.add probably sends you to the next u64 not the next u8, also on some systems the less significant bits are at higher memory addresss

Also on some systems ur mom is gay

What are some non-autistic things I could build?
I am just spinning around in circles and project hoping too much. I need a big boi project

>>109243661
>I am just spinning around in circles and project hoping too much.
>"WEEEEEEEEE I hope for a project."

>>109243661
compiler

>>109243681
You're absolutely right!
Maybe I should go through my git and polish old projects to finally have some sort of portfolio.
>>109243706
This was literally the last thing I was thinking about, but then my non-autistic brain goes crazy on what exactly do I want and shit like that.
I did this once in C++ https://interpreterbook.com/
Maybe I should work through the second book, make the compiler and then see how to modify the language from there

Long story short: Lets watch some video

>>109243681
>>109243763
PS: What drives me nuts: Rn I am wasting too much brain power and thinking about if I should once write something "useful" instead of re-inventing the wheel all the time.
Tho re-inventing the wheel is kinda fun

>>109243661
Non-compliant standard library in your own personal style with all the cool stuff you ever wanted.

>>109230092
I'm sad there's no autists screeching about how you should never make your own rng

>>109237417
modern operating systems cooperate with hardware to construct page tables from made up address space to physical ram. when you look at a pointer, the bottom 12 bits translate to an offset into the page while the rest of the pointer are mappings into some weird tree structure that tells the kernel how to get back to the physical ram address

File: 142454919_p0_3.jpg (2.02 MB, 3600x1532)

2.02 MB JPG

>>109227374
question about arithmetic in c++: if i have an expression like double x = (a + b + c + d) * 0.1 where all of abcd are uint32_t, does this overflow since i'm adding four 32 bit types together resulting in a max 128 bit value, or does scaling it by 0.1 actually avoid that situation? how can i guarantee this doesn't overflow?

opinions on Coding Jesus ?

File: difference.png (505 KB, 1920x1080)

505 KB PNG

>>109244161
>you should never make your own rng
Why do people not understand the difference between 'making your own' and 'using your own in production'?
It's not that hard, if you ask me. There is a very clear difference

>>109244651
Yes, it overflows, and you prevent it by making one of a, b, c, or d a uint64_t.

>>109244651
It may overflow. Cast all the values to double before adding. Strictly, you don't need to explicitly cast them all due to automatic promotions, but it's a good plan anyway because it's much clearer.
(You could also cast to uint64_t, but since it's going to double anyway and that's plenty large enough to do as a perfect a job as possible with the use of 0.1, there's no point in postponing it. All the inaccuracy in that calculation comes from multiplying by 0.1, nowhere else.)

>>109245285
i could avoid this problem by not chaining the additions right? just adding them to the double one-by-one?

File: 1783723571996980.png (14 KB, 283x252)

14 KB PNG

I can't decide.
The top means I can fit more code on the screen but I find the bottom more readable.

Trying out Goal because I can't get any K's to work very well on Windows. Most K code looks very aesthetic to me so I can't wait. Its a nice blend of lisp and array language paradigm. I mostly pick and choose languages by how the code looks to me. Very dumb for a number of reasons, but it is what it is. I would also recommend BQN or APL. J seems fine to me, but many people don't like it, same with Uiua.

>>109245695
definitely go with the one that's less readable to you, so you can have more code to struggle to read at once

>>109245695
I do top because most of the time functions are not long enough to warrant the extra line. However, when it is long / there are a lot of arguments, or it's something like a constructor I do this

int foo(
    argument1,
    argument2,
    argument3
) {
    /*code*/
}

Obj(
    argument1,
    argument2,
    argument3
) :
    member1,
    member2,
    member3,
{
    /*code*/
}

because it's easier to change, comment out a line, view in a diff, etc

File: Screenshot_20260711_011421.png (325 KB, 2546x1214)

325 KB PNG

>>109243661
Ok I think I got it. Back to fpga stuff again. That was a fun journey beginning of this year.
Lets polish that and add some cool features like direct camera image manipulation (sobel filter, gaussian and so on). Already noticed that my readme is missing required packages

File: Screenshot_20260710-204156~2.jpg (61 KB, 720x693)

61 KB JPG

>>109227374
This simple command gave me almost two pages of errors.

>>109243661
you are spinning in circles because you're making dump projects you don't actually care about

make something you genuinely need or want to exist

>>109245840
damn, I almost forgot how weird embedded linux dev is.
I somehow have to get a bit more structured, but I am still confused how I should have a decent protoyping env where i can tweak shit on linux kernel, buildroot, microchip device overlay and finally my userspace software. Oh and also from time to time the fpga configuration.
Somehow gotta streamline that. But i think that's a fun project to get back in. I should make youtube videos for that. There are not even jeet videos for this

File: webstream-vp9.webm (2.27 MB, 1410x769)

2.27 MB WEBM

>>109246076
Yes this is a big part of it.
And for the fun stuff I want too much sometimes.
In january I did this CSI camera interface for the microchip polarfire soc, because there was literally no example on the internet for the discovery kit, just for their more expensive video kit.
That's both fun (and also fucking sucks, tooling wise, h3h) and challenging. I tweak this more and do cool shit

>>109246138
PS: also should upgrade to raspi cam v2, which should have much much better picture

I'm a Haskell wizard!

rev (h:s) = rev s ++ [h]
rev [] = []

main = print $ rev "hola mi amigo"

>>109246286
Rewrite this in Haske- oh wait nevermind

Is Rust a good languange to learn if only for the pedagogic value of doing it or should i just learn C#? i don't have a strong motivator, maybe porting over very ugly bash scripts to Python or something that works in Termux too, i just feel that if i don't go deeper than stiching together argparse with the world, i'll become stagnant. I'm using a very shitty laptop for my learning (because if i use my main rig i'll procrastinate)

>>109246366
Congrats, Rust will run faster on your shitty laptop.

Stop using Rust, faggots.

>>109246533
>TRIGGERED

>>109246555
Im fine sitting on C++ while Rust faggots learn a language thats irrelevant and will never be used in anything serious.

>>109246563
>t. can't learn more than 1 programming language

>>109246577
I know a lot of languages. Except for fake ones like Rust.

>>109246592
What about real ones like Rust?

>>109246615
Like trannies, it never will be.

>>109246623
Rust is compiled not transpiled.

>>109246563
As a superior C programmer, I chose Rust instead of C++.

>>109246076
>make something you genuinely need or want to exist
This advice is useless if you're a boring person and everything you need is already out there and is better than you could ever hope to make

>>109246677
based
C++ trannies btfo

>>109246763
If something being already out there is your metric for not doing anything, then you'll never be able to do anything in your life.

>>109246763
Make it again but better.

File: 1757822375214853.png (1.24 MB, 1273x718)

1.24 MB PNG

It's over for gamedevs. Look at these amazing graphics. Look at the crisp UI and pixel perfect sprites. Imagine the infinitely autogenerated fetchquests and tower of hanoi puzzles and copypaste slop dungeons. Truly we are entering the golden age.

>>109247659
Zero monetizable content.

>>109247659
Funny how these "games" look exactly like the templates I've seen in every game engine.

File: 1783752108700331.png (423 KB, 2560x1416)

423 KB PNG

Is there a word for this style? It's like modular or containerized programing? It's like OOP but it's not true OOP as you can't instance a timer and would need to create an array of timers and change your logic.

Please don't be mean and call it retard programming.

File: 1769707904859116.png (550 KB, 1213x571)

550 KB PNG

>>109247765
Fuck the haters. Look at this incredible scene. And it only cost me $50 and 48 hours of processing!

>>109247810
Funny how it looks exactly like Tiny Wheels from Roblox.

File: 1756271234644452.png (12 KB, 285x96)

12 KB PNG

>>109247777
I guess you could call it functional programming. Also, what is going on here?

>>109247659
~S~T~Y~L~E~

>>109247858
Capping the timer to 99:59:59 hours as the output only has 6 digits.

>>109247810
This will all change with 5.6!!! Devs are so cooked now. All other models were trash, but 5.6 is conscious

>>109247777
Checked.
Also: why isn't Timer an opaque handle?

File: garbage.png (561 KB, 1920x1080)

561 KB PNG

>>109247659
And it runs at 30 FPS, if that.

File: 8mY9wx_iMSU-HD.jpg (160 KB, 1280x720)

160 KB JPG

>>109247858
check'em

>>109248185
Why are people STILL paying to whatever Asians say after FF13?

>look at my funny racism xD

>funny
I wish.

>>109248203
it's not paging

>>109248213
I neither know nor care what that means.

File: Screenshot_20260711_104521.png (170 KB, 1283x1015)

170 KB PNG

Damn, I never noticed that. He really was a genius. Unironicaly based

File: 1739435724033306.png (118 KB, 640x466)

118 KB PNG

>a schizo is based for ignoring all the lessons DOS taught us

File: cia.png (26 KB, 200x200)

26 KB PNG

>non-schizos stil don't get it

>>109241340
It is not at all clear why you would pay a lot of money and time to learn a skill and get certified in it when you can ask a clanker to teach you whatever you want to get good at when it comes to coding
You need better taste in programs to be useful to people who want solutions that programs can provide
Not necessarily be able to write a trie yourself

>>109241624
GitHub has had a 14x increase in traffic because of vibe coding and they’re struggling to keep up

>>muh CIA
>at this point no longer sure if tist or schizo

File: asfd.jpg (74 KB, 560x560)

74 KB JPG

>>109248493
This should be enough proof that AI makes you, at minimum, a 10x engineer.
Snailcats still coping

File: 1767569233096269.png (202 KB, 640x640)

202 KB PNG

>Snailcats

File: 1783764407666482.png (409 KB, 2560x1415)

409 KB PNG

>>109247858
I can see the appeal, this is comfy.

>>109241624
>this trail is filled with ATV drivers nowadays, making it difficult to walk on from all the tire tracks
>trail maintenance put up a bunch of gates and poles you have to walk around
>ATVs just drive on the sides slower leaving even worse tracks from driving on uncompacted soil
wouldn't you change trails at this point?

>>109237796
you can buy used books online in Europe, that's how i got the C book and SICP

>>109227374
I want to build something using rust. give me ideaz

>>109247659
You just know

>>109248221
Go ask AI then

>>109248221
https://www.youtube.com/watch?v=u38QZYjcLGs

File: Screenshot_20260711_135815.png (436 KB, 1937x600)

436 KB PNG

What did they mean by this? I wonder if there is a reason behind choosing exactly this picture

>>109249143
only indians and artificial indians are allowed jobs, chud

File: 1766384798101761.gif (1.45 MB, 320x285)

1.45 MB GIF

Had to check the hints to get this one
https://leetcode.com/problems/count-the-number-of-complete-components/description/?envType=daily-question&envId=2026-07-11

    func countCompleteComponents(_ n: Int, _ edges: [[Int]]) -> Int {
        var dsu = DisjointSet(n) 

        for edge in edges { 
            dsu.union(edge[0], edge[1])
        }

        var ans = 0 
        //A union is valid if there are nC * (nC - 1) / 2 edges, where nC is the node count
        for parent in 0...n - 1 where dsu.find(parent) == parent { 
            let nodeCount = dsu.ranks[parent]
            let expectedEdgeCount = nodeCount * (nodeCount - 1) / 2
            if expectedEdgeCount == dsu.edgeCount[parent] { ans += 1 }
        }

        return ans
    }

    //DSU implementation, also counts how many edges in each union
    struct DisjointSet { 
        var parents: [Int]
        var ranks: [Int]
        var edgeCount: [Int]

        init(_ n: Int) { 
            self.parents = Array(0...n-1)
            self.ranks = Array(repeating: 1, count: n)
            self.edgeCount = Array(repeating: 0, count: n)
        }

        public mutating func find(_ x: Int) -> Int { 
            if parents[x] != x { parents[x] = find(parents[x]) }
            return parents[x]
        }

        public mutating func union(_ x: Int,_ y: Int) { 
            let xParent = find(x), yParent = find(y) 
            if xParent == yParent { 
                edgeCount[xParent] += 1 
            }

            else if ranks[xParent] > ranks[yParent] { 
                ranks[xParent] += ranks[yParent]
                edgeCount[xParent] += edgeCount[yParent] + 1
                parents[yParent] = xParent
                edgeCount[yParent] = 0
            } else { 
                ranks[yParent] += ranks[xParent]
                edgeCount[yParent] += edgeCount[xParent] + 1
                parents[xParent] = yParent
                edgeCount[xParent] = 0
            }
        }
    }

File: images(1).jpg (6 KB, 254x198)

6 KB JPG

>>109248742
>erlang thread
>immediately goes offtopic with some nonsense

>>109249726
why check hints instead of using ai? boomer brain

>>109249143
https://www.canadianlawyermag.com/resources/practice-management/firm-launches-east-asian-and-south-asian-affinity-groups-to-promote-diversity-and-inclusion/360387

>>109250060
>shitlang thread is filled with shitposters
WaOw!

>ur lang
>isn't the ur language

Web search seems fucked and I don't trust clankers.

What is a good and easy algorithm to find a restricted random integer set partition?
e.g:
my_f(5,3) = [3,1,1]

>>109250836
>integer set partition
what do you mean? from a particular set? or do you mean N numbers between 0 and M?