Laughing at certified autists who still claim that compilers aren't garbage.
>they use C++

Erlang is King

>>109145747
Poor thing.

>>109146021
Deflection of responsibility.

>>109145475
Working on something is one thing, but are you or someone else actually using your programs?

>>109146094
Of course. That's why every cycle is so important.

>cycle autism
Rarely there's a clearer marker of a "program" that does absolutely nothing of value and is written only for the sake arguing about implementation technicalities. Such case is the polar opposite of a usable program.

>autist believe it's about implementation technicalities
>immediately project his own autism onto others
ngmi

>projecting autist ignores facts because they make them uncomfortable
>like last time the thread went dead
>which is why the cows will be coming for him too

>ignores posts because they make him feel uncomfortable
Autism confirmed.

Let's get some (You)'s in here you autistic fucks so I can see who is talking to whomst.

you can tell how not linking to his posts absolutely bruised the schizo's ego since he's now doing it himself all the time

File: you dopamine hit.gif (443 KB, 480x238)

443 KB GIF

Here's your (you), faggot

got a rockchip sbc
thinking of looking into what it would take to get driver support on the current kernel as practice
i know arm shit tends to use something called device trees as basically a substitute for acpi so i need to figure that out too

I just spent days figuring out git fixme and rebase. It's frustrating how there's like 0 help on using git outside of the GitHub workflow.

Is there another chan that's more programming focused btw?

>>109146748
/prog/ is dead
there's lambda on whats it called chan

You're going to lose every time against a C++ user because it's a game of being more complicated and they went over an event horizon of complexity in like 2020.

Arguments like "is this a good idea" or "is it that compelling" mean nothing.
I checked out of this years ago. Maybe if they made a release when they needed to instead of a new Harry Potter film every 3 years.

>>109147951
It's called C++26 because it will take 26 years to implement

Working on a data parsing algorithm that will parse thousands of Starcraft II and Quake replays, for the multithreading I use Go and for the data extraction into graphs I use imGUI using C++26.

>>109147951
I have "you just have to use modern C++" fatigue.

>>109146476
>i know arm shit tends to use something called device trees as basically a substitute for acpi so i need to figure that out too
For the sort of things you use a sbc for, that makes a lot of sense. Why spend ages autodetecting attached hardware on every boot when you can just tell the kernel exactly what's there and how it's connected?
Yes, that means they're not intended to be consumer devices as-is. But for hacking your own hardware? IT'S GREAT!

>>109146094
>are you or someone else actually using your programs?
Depends which programs.
Some of my programs are probably installed on your computer, and you might be using them without knowing it, every day.
What I'm working on at the moment? Nobody's using that yet. It's not yet released (and it's way too prototype-y for that to change for the next few months; need the companion ASIC back from manufacturing first).

File: file.png (17 KB, 535x99)

17 KB PNG

im a c++ programmer and just starter learning csharp. wtf is this syntax lol

>>109150513
lol that looks really bad but it's probably some syntax sugar and standard && (or "and") should work right?

>>109150742
C# has the classic shortcircuiting "x > 10 && x < 20.
And then the "is" pattern matching operator.

static bool IsFirstFridayOfOctober(DateTime date) =>
    date is { Month: 10, Day: <=7, DayOfWeek: DayOfWeek.Friday };

int i = 34;
object iBoxed = i;
int? jNullable = 42;
if (iBoxed is int a && jNullable is int b)
{
    Console.WriteLine(a + b);  // output 76
}

int[] empty = [];
int[] one = [1];
int[] odd = [1, 3, 5];
int[] even = [2, 4, 6];
int[] fib = [1, 1, 2, 3, 5];

Console.WriteLine(odd is [1, _, 2, ..]);   // false
Console.WriteLine(fib is [1, _, 2, ..]);   // true
Console.WriteLine(fib is [_, 1, 2, 3, ..]);     // true
Console.WriteLine(fib is [.., 1, 2, 3, _ ]);     // true
Console.WriteLine(even is [2, _, 6]);     // true
Console.WriteLine(even is [2, .., 6]);    // true
Console.WriteLine(odd is [.., 3, 5]); // true
Console.WriteLine(even is [.., 3, 5]); // false
Console.WriteLine(fib is [.., 3, 5]); // true

File: 1759491203029047.gif (3.63 MB, 211x374)

3.63 MB GIF

>leetcode contest is at 10:30 pm my time
fucking jeets

File: goatse_hands_by_mopeloves(...).png (229 KB, 523x350)

229 KB PNG

>>109150899
Well ain't that fancy?

>>109149913
oh good
i guess i should stop putting it off, the little bit i'd worked with acpi hacking had me dreading that working with device trees would be a similar nightmare

>>109151396
device trees are comfy

>>109150899
doesn't really explain why " > 0 and < 10" would be a good syntax to add to the language

>5 new posts
It's dead Jim.

Being able to program computers is a lost arcane art.
You've earned the right to be able to say you can through all the time you've ever invested.

I'm going to learn how to freestyle ski.

>>109151637
Pattern matching consistency.

if (person is { Age: >= 18 and < 65 })
{
    ...
}

return x switch
{
    < 0 => "negative",
    >= 0 and < 10 => "small",
    >= 10 => "large"
};

>>109151921
> < 0 => "negative",
> >= 0 and < 10 => "small",
> >>= 10 => "large"
Nope, I can only feel mild disgust when seeing this, I can't quite put my finger why

>>109152684
there are cheap supermarket bike computers that have no connectivity and are thus botnet-free.
but you can just get some ESP or other embedded chip, a hall sensor, a magnet and a screen to display stuff, and start hacking your own

>>109154324
>the goal would be to have one with navigation or at least a map
in that case the no-effort solution would be to have a cheap bike computer and a phone holder, and read GPS/map data off the phone

>Yes I could start with an esp32, but to get gps and maps on there might be challenging.
might be easier to write a phone app that would read GPS and map data and send it to your device via wi-fi or bluetooth

File: Zminok16.png (230 KB, 1339x872)

230 KB PNG

I should stop overengineering

>>109154918
>const field
terminally retarded

>>109154918
Why boost::shared_ptr instead of std?

Trying to get a Decky plugin together. One of the bigger roadblocks, besides me not being all that familiar with TypeScript, is that their documentation is very sparse and broken. Would it have killed them to give sensible defaults or examples for dropdowns?

If they want people to not use AI for shit (and AI tends to be very iffy on what to actually do, probably due to the shitty documentation), they really need to fix that.

>>109154959
The boost::shared_ptr needs to be returned because its passed into boost::program_options::options_description.add()

>>109146052
I don't see why. You can point plenty of major known climate fluctuations in the past and easily say that people aren't responsible now.

reading disassembled C++ is killing my soul, how do I cope

So what's the deal with switches? Are they not just glorified if statements?

What scenario would you use switches over ifs?

Started on a side project this weekend. Need to reread how curl multi & how glib does event handling.

GLib has got to have some of the least informative documentation there is. Each function is documented, but I feel like I have too little information about the object lifetimes and what whether it grabs/adds references.

>>109156649
A switch statement is just a glorified compiler hint that you want to turn something into a jump table. Although it also does happen to look less like shit when you want to just compare a value against a set of values.

practicing my rust by making a mod manager lite with primary focus on unreal 5 games, eventually would like to support unreal 4, zip and rar drop in, not ass looking gui

File: Screenshot from 2026-06-2(...).png (100 KB, 792x590)

100 KB PNG

>>109156649
Depends on the language. Python lets you put regex in them for string matching iirc.

For just reading code, it lets you avoid if else if else if else blocks, which all depend on the same variable. That makes it immediately clear that this variable decides what is done next from many possible states. We often use enums for that.

If you have a method with many response codes (probably an enum), you can make any decisions fairly clear.

This is old code I was using to read from a barcode scanner. Notice the lack of break; statements. I have intentionally allowed fall through for BUFFER_FULL & BUFFER_COMPLETE.
FULL would mean we can't read any more data and need to switch the buffer anyways.
COMPLETE means we received a control code to stop reading & start a new buffer.

So FULL & COMPLETE are essentially handled the same, but I could add validation or other checks when a full buffer is reached & then swap the buffers.


I enjoy switch statements for the fallthrough behavior, but also when you are making a basic statemachine.

while(true) {
/// perform some operations or read some data & set a flag.
switch(flag) {
/// cases can be used to decide on how execution will continue.
}

// use if statements for checking conditions with more than one variable or that would be applied in multiple states.
// e.g., if (thread_wants_to_kill_itself) quit()

}

>>109156818
a basic google search told me ifs are best for ranges of possibilities so x > 10

while switches are best for specific results x = 1 etc

File: 1755419551918139.jpg (58 KB, 976x850)

58 KB JPG

>participated in yesterday's weekly leetcode contest
>90 minutes for 4 questions
>3 mediums, 1 hard
>finished 2 mediums and 1 hard
>placed in the low 2000s / 37k contestants
>check the rankings
>everyone 300 and above cheated
>#1 guy is a jeet that finished the contest in 3:34 by pasting in all the chatgpt answers

>>109156861
Case statements (in C) only support constant-expressions. So they can't handle ranges or complex statements on their own. Switch statements are best for state machines. Each language implements them a bit differently, so your mileage will varies.

>>109156945
Please refrain from posting images of jira on the weekend

>>109145695
how can compilers be garbage when they're the only tool left in existence that can properly parse sepples?

my program stop behaving as intended, couldnt figure out what recent changes could have done that, asked the clanker to review my code, it just broke it more
its fucking joever

>>109157460
You're not gonna make it

>file_stem() method returns the file name before the LAST .

im gonna kill myself

>>109157608
>Rust user
>is going to kill >her(male)self because it will never be a real language
Checks out

embarrassing post

>>109157460
what are the intended and actual behaviors?

>>109158032
managed to fix it but basically it was caused by me changing a regular string strip suffix to a file_stem() on a file path with name foo.bar.ext and that returned foo.bar instead of the desired foo
tldr; clankers bad

I've never really touched AI coding tools, and I wanted to try them. I've tried code auto completion in vscode, but I didn't find it very helpful, if anything it makes me slower because I have to read its code to make sure it actually does what I wanted.
So, how do you guys use them, if you use AI tools at all? Or is AI assisted coding a meme unless you're vibe coding everything and letting the computer do its own thing?

>>109158189
I don't use them and neither do most of my coworkers.

>>109158189
I don't use them to write code. They're best as a search engine for documentation references.

I have had claude write code a couple times and it wasn't great. One was a visual basic script for outlook, no more than 50 lines. Unnecessary function wrappers + fucked up file tree traversal as well.

File: blue.gif (25 KB, 704x720)

25 KB GIF

>>109151637
Is not good syntax, it only shows they are irredeemable jeeted.

>>109158333
checked

File: terminal_tumor.png (27 KB, 571x618)

27 KB PNG

>>109154924

>>109156635
You can get that syntax candy with C++ by using a custom type and overloading operators, but is not worth unless you want to brain fuck people.

>>109154918
That shit looks slow AF.

File: elon-musk-nod-cap-white-h(...).gif (739 KB, 165x188)

739 KB GIF

>>109158189
First thing I did when autocomplete started being added to IDEs was to turn it the fuck off, its distracting.

You really want to use some sort of agent/harness (codex, CC, opencode, whatever) because these automate a lot of what makes in browser ai coding a pain, e.g. copy pasting context, iterating on console output, exploring the filesystem, etc.

The biggest issue with AI is that it creates complexity faster than you can understand it and internalize it. There are largely two schools of thought on this: not checking the output in any meaningful sense ("vibe coding") and treating each output like you're reviewing a PR. Which to use when is something you learn over time, and is largely a matter of circumstance and preference.

Personally, I use the former to validate concepts with prototypes that would have taken me a long time to write. I only do the latter with codebases with which I am intimately familiar with, because otherwise its more exhausting than just writing it myself. There is of course some standard stuff that AI is almost always reliable for, for example: deserializing APIs, writing docstrings, code microstructure ("what's the best way to implement this data structure in this language"), you'll discover these over time.

>>109159436
>its distracting
it really depends on the IDE, and in some cases, a specific plugin or configuration. a good one isn't intrusive at all (or at least can be configured as such) and just massively speeds up typing code

Hello here's my daily C# dummy question contiuning from yesterday about switches:
>>109156649

So case 1: do something
break;

Can you use switches with strings and chars? So instead of an int being 1 or 2, can you do a string/char being Y or N?

>Do you want to continue? press Y/N

case Y: continue
break;

case N: stop
break;


At least with If I can literally test if user input is Y or N
If input == Y
If input == N
and so on

So I guess what I'm asking is can switch case be anything but integers?

>>109160976
Of course, why wouldn't you be able to?

>>109161149
Oh shit it does work I'm just retarded, perfect.

Can you evaluate the same things in switches as you can in ifs?

>>109161219
So theres this thing called documentation...... (google C# selection statements or just google C# switch statements)

Reverse engineering MMBN2. Game is basically a giant switch between scenes. Almost done with the save system.

the job market for junior rust devs isnt looking great where i live...

>>109156649
You use them when:
>When you have a bunch of conditions
>When some of the mutually inclusive conditions can be represented in a linear hierarchy
Tagged unions are an obvious choice.
A switch statement is much nicer to work with than a bunch of ifs.

>>109161893
pro-tip: learn how to use AI and become an AI engineer.

Does anyone know of good forums where programming is discussed? When I was in college I remember I couldn't find any forums that were even semi-active & not geared towards noobs... besides one that was used for game cheat development lol

>>109162288
there are none.

best zed theme and why is it Rosé Pine

File: HCJb2DqbQAArsyH.jpg (61 KB, 680x508)

61 KB JPG

>>109162366
That's what I thought....

>>109162288
what is there to discuss? you are either: a noob, working on something you arent allowed to discuss, working on something so niche nobody else will care or understand if you try to discuss, or making game mods

>>109162827
im a noob what does this do

>>109161149
You can't in C/C++

>>109162288
Here, sadly.
>not enough programming discussion
Be the change you want to see.

>>109163882
Characters are integers in both C and C++.
In C++ you can calculate a hash and use that in the switch statement, if there is a collision the compiler will tell you:

constexpr auto h(const char* s, int z) {
    unsigned r = 0;
    for (int i = 0; i < z; ++i)
        r ^= s[i] << (i & 3) * 8;
    return r;
}

template <unsigned N>
consteval auto h(const char (&a)[N]) {
    return h(a, N);
}

#include<iostream>
int main() {
    std::string s;
    for (std::string s; std::cin >> s;) {
        switch (h(s.c_str(), s.size())) {
            case h("yes"):
                puts("YES"); break;
            case h("no"):
                puts("NO"); break;
            case h("maybe"):
                puts("MAYBE"); break;                
            default:
                puts("??");
        }
    }
}

>>109163158
> you are either: a noob, working on something you arent allowed to discuss, working on something so niche...
That is absurd.
There are plenty of people working on and discussing code publicly. They just happen to do so on mailing lists, irc, in this thread, &c.

You could open youtube right now and listen to someone talk about their code, their analysis from reading someone else's code, what they think about specific architectures, &c. On twitter I have seen deeper technical discussion than I had ever seen in a college classroom -- where the majority are there for the degree and not for the education.

When crawling project forums you will find people looking for discussion. Unfortunately, those forums are mostly for bug reports & are essentially a place for non-contributors to discuss their issues and expectations. I would rather see discussion occur on a broader forum (where it would be encouraged) than seeing them crop up after a bug report.

There are far more topics than "what am I working on today" or "help me with this problem".
> working on something so niche nobody else will care...
It is a forum, not social media. Not everyone will be interested, that is fine. "nobody" is a stretch, plenty of people would be interested in seeing a project and knowing about its' design.
> or understand if you try to discuss
That is the author's problem. If you can understand it, you should be able to explain it. We write for a purpose, audience, and context. It is up to the author to correctly identifies who they are talking to, why, and whether they are informed or not. If you know your field is niche, you know you will have an uninformed audience and must either simplify or extend your explanation.

>>109164127
Well.. that is why I crawl back here periodically. But 4chan doesn't feel like the place for this.
- Threads can only live for so long.
- Anonymity is nice, but you can't really have a long term debate, discussion, &c without somehow losing track of who is who between threads.
- imageboards really aren't made for this, I'm not coming back here that often and I will inevitably miss your replies. It's akin to eating soup with a fork.

I'm not on here every day, week, or month even. I would rather check my email and get a summary like 10 new posts under <topic> or 10 new comments on <post> since you last visited... Then I could follow up, make my posts and go dormant for a week.

The closest I get to that is remembering to keep this tab open before I close out for the day..

>>109145475
Reading through HW datasheets, planning how to write device drivers for it.

Started learning typst. The scripting language is... interesting. No global mutable state allowed, but it's got a state object that can be fast-forwarded/rewinded to different contexts within the document generation.

Kinda fun. And a fair bit more intuitive than Latex, which I've been working with for way too long.

>>109164692
The thread watcher helps with finding replies later. Even if you don't come back for a few days, you can still see replies in the archives and extensions like 4chan-xt will take you there automatically.
>Anonymity is nice, but you can't really have a long term debate, discussion, &c without somehow losing track of who is who between threads.
True enough. If you find a decent forum, drop it here, though some discord is probably your best shot.

What's the goto pastebin site for anonymously sharing code these days?

raylib 6 or sdl3?
Help me decide?

>>109165592
just make the triangle bro, it's not that hard

File: 1755514977145.gif (78 KB, 200x200)

78 KB GIF

Big ups to the anons from last thread giving tips on JSON parsing. I'm reading the UTF8 network stream buffer directly with Utf8JsonReader and saving just the parts we need. Straight from the stream, into the final data structure, with no intermediate storage. Peak RAM usage for each process went from 350MB to 90MB. So we can run almost 4x the threads now.

It was so high because the legacy code was storing a UTF16, JObject, and a slightly transformed UTF16 version of a giant JSON payload all at the same time in a function. So 3 intermediate storages simultaneously.

There's probably more RAM to save elsewhere, but I don't think there's anything else I can do for the JSON parsing.

>>109164692
>Threads can only live for so long.
>Anonymity is nice, but you can't really have a long term debate, discussion, &c without somehow losing track of who is who between threads.
This is a huge blessing for me. Getting locked into an argument with a smug retard for days or weeks on end was horrific for my productivity. On here it's just you just call each other retards and the thread dies with no long term memory or animosity.

Redditors, in particular, use the most disingenuous rhetorical tactics humanly possible. Amazing if you want practice picking apart rhetoric and logical fallacies. But horrible if you want to have good faith discussions or get anything done IRL.

>>109145475
Making a text editor with Raylib. Having fun before it all goes to shit.

>write C interpreter
>don't need scripts anymore
>can compile things when you're done working on them
>replace all macros with compile-time evaluated C
>???
It's 2000+26 why can't I just do that? I bet we can blame (((lispfags))) somehow.

>>109167383
You are fully capable of doing this. Though few people would actually use C as a scripting language. It tends to have a decent amount of boilerplate just for string operations.

>>109167383
tcc lets you put #!/usr/local/bin/tcc -run at the top of your file and run it like a script

also sean barrett has a video for being proficient at using C this way
https://www.youtube.com/watch?v=eAhWIO1Ra6M

>>109167715
not anon, but super interesting. I've got some older computers this might be fun to work with, probably going to try this soon.

>>109166550
Yeah, it really can go both ways. I'd just mute & move on if it were on a forum. I don't really engage that deeply once it is a pointless discussion.

blog:
I've had a long term project I've been working on for almost 10 years and it's pretty close to completion but the fact it took so long makes me feel stupid enough that I don't want to complete or publish it, but at this point I'm in a sunk cost fallacy so I guess I have to.

I've done practically nothing else in that time beside contribute patches, any original projects I made inbetween are small and just for myself (unpublished).
But, I also think I've learned how to program well in that time. My next program shouldn't take nearly as long since I know the practice now. But I still don't feel good about how much time has passed for very little result. Maybe I shouldn't have picked something big, complicated, and unique to work on.

>>109168394
Sometimes the journey is more important than destination. Ya know?

>>109168394
Sounds like you spent 10 years building something that you wanted to build, so I don't see the problem. I'm sure you learned a lot. If you post your project I will check it out, just use the code phrase "WALTER CRONKITE" in your post so I know it is you.

>use a git repo to keep my macbook and mac mini in sync
>any minor inconvenience or conflict and xcode shits itself
>have to pull out the command line
very cool

>>109168983
use syncthing

>>109168394
It's not unusual for a real project (one that's getting used) to take a long time to finish.
My personal one is now over a year in development, mostly because I'm more ok with it being unfinished than to risk breaking it or its data, so I mostly just use it as-is.

>set up .gitignore
>xcode ignores it
>stashes my secrets config
>undo the stash
>deletes the config

>>109169501
>using xcode
>using GUI to stash instead of git
your fault && skill issue

File: 1777602439140319.png (80 KB, 348x348)

80 KB PNG

>haven't used python in years
>coming back to refresh my memory

>trying to figure out how to mutate shit I pass into a function
>apparently that only works for dictionaries, arrays, sets, and classes
>but not for integers or strings or other shit, those get copied
>there's no structs either

how is python a real language? in swift it's
>everything passed into a function is immutable except classes
>if you want to mutate them, you pass an explicit reference

i'm working on a generic session process for unix-like operating systems in posix shell, fun enough it also had to be a generic implementation of user services since a session process requires user services.

anyway i've put out some releases and i am daily driving the program as well as dogfooding it so all breakages happen first in my system

as for now the program just works on linux but i'm also interested in doing the work so it can just work on freebsd tho it seems the crusty old sticklers at the freebsd forum got no interest

here's the main mirror check it out: https://github.com/eylles/shed

>>109170640
>This program is something that should have been written 10, 15 maybe even 20 years ago by someone much smarter but it was only when i could no longer bear the situation of there being no generic program with the capabilites of shed that i forced myself into writing it.
There's nothing I appreciate more than seeing software being produced out of necessity and frustration.

>>109167715
How long does a script have to be before it starts beating regular scripting languages?

>>109170552
In C, it's
>ALL parameters are copied into functions
>but some of the copied things are addresses to get to the caller's copy
Python is utter shit. I seriously have no idea how people can learn the reference semantic bullshit without already knowing what pointers are and what Python concepts map to them.

>>109168394
Why wouldn't you open source it earlier?
Then you actually might get some real users and feedback. Or not, but at least it's available out there.

>>109171862
It is open source and I have gotten user feedback, that's what I'm actively working on / why it's taking so long. So far I have a lot of bugs and requests but no contributors besides me. I agree with all of the feedback but implementing it has been difficult because the task is pretty complex and there's also not many people experienced in what I'm working on to even consult for help let alone get code contributions. The ones I know work for proprietary companies and never have time to consult with me about it.

Right now, I'm effectively rewriting the entirety of the last published version. Same API externally but this it should work correctly and the code should be more comprehensible despite the inherent complexity of the requirements.

I think these things combined (complex esoteric system + complex code) is why I have no contributors so I'm putting a great deal of effort into having good abstractions in the source code.
It doesn't just have to execute well it has to read well. And doing that is hard normally let alone for a big complex system of systems.

>>109171923
Then just work on the thing that you need for yourself, why would you maintain this for others for free?
Unless you want to show it off in some portfolio

>>109171970
>I agree with all of the feedback
I'm writing it because I want the software too. I use it.

File: 1765455857178437.png (59 KB, 671x256)

59 KB PNG

oh my god bruh

File: 1778194578073112.png (74 KB, 671x497)

74 KB PNG

>>109172307
OH MY GOD BRUH

>>109172307
>>109172314
No way to declare constants either.
Meme language.

>>109172307
Kind of based desu

>>109172307
Classes are actually bad and you can encapsulate private data using closures like a pro
>>109172314
Actually the cleanest approach is the multi ling string literal

File: howIthink.jpg (262 KB, 1418x850)

262 KB JPG

Hi, how do I quickly (fte 3days) learn about creating fastcsi backends that are prod grade? Book? Tutorial series?

Can be even python as long as I get to deliver quickly.

>>109172732
Start by spelling it right, after that, it depends how "from scratch" you want. Either start reading specs or pick a library.

I hate coming up with a cool solution to something only to later come up with a better but less interesting solution. Deleting the clever code and never using it.

File: deutschland-tired.png (1.46 MB, 1913x1717)

1.46 MB PNG

>>109170750
it is a special kind of hell to want a piece of software to exist then find it has not been written even tho the concept exists as a "standard" and that to write it you need to reference the architecture from a program form 40 years ago and use 20 year old standards from one of the most limited "programming" languages ever (posix shell does not have arrays), and in a matter of 4 years as a solo dev going in and out of employment you manage to go from barely a proof of concept to a daily driver program, it really begs "why none of these stupid programming geniuses made this 20 years ago?" why is it that me, a stupid autist that doesn't even like to code and cannot write anything but a memory corrupter in C has to be the one writing this fucking program... it is so utterly frustrating

>>109173079
I empathize completely. I deal with file systems for multiple operating systems, and that includes POSIX. And of course I end up using the API from the perspective of a user too, not just an implementer.
Since I know that toil and frustration particularly, I respect your effort. Good luck with your endeavors.

It's obvious that these things don't get built without fanatical people who demand better tools to exist for the benefit of us all, no matter the cost.

>>109173098
the most frustrating to me about my stupid program is that it is just shell script, not even 2000 lines of actual code...

>>109173109
In my experience, I've come to learn that making small simple solutions can be difficult, especially when you don't control every aspect of the API.
It has taken me a long time and much practice to get to a point where I can write pragmatic code like that while also making it legible to other developers besides just myself (but also including myself in the future too).

That said, I share your sentiment. There's not enough other people putting in the effort or care to get to that point and then produce them.
I get it though, its obviously tough and unpleasant at times, and of course thankless and less rewarding in open source projects.

>>109173124
>pragmatic code like that while also making it legible to other developers besides just myself (but also including myself in the future too).
man... writing code i will understand in the future is the bane of my existance, no matter how much time i spend in some piece of code, if i leave it alone without touching it for just 3 weeks i will come back to asking "wtf this does?", the best way i've managed to cope with that is trying to write good docstrings for every function and var i can

will learning C give me anything valuable besides a better understanding of "whats going on under the hood" ? From what I've seen, the core/standard library barely gives you anything to work with and you have to implement your own data structures/types before you can do something meaningful, which just seems stupid.

>>109175086
There are different levels of "under the hood".
Actual assembly instructions, calling conventions, etc.: no.
How code actually operates without a huge runtime underneath you without lots of hidden code: yes.

>>109175086
C is a /g/ meme, it's only useful if you have to use it for embedded stuff or interfacing with other languages.

>>109162288
there is unironically a lot of substantial programming discussion on bluesky kek

what happened to the MAIDS daily programming threads?
i loved them a lot they were very kindly worded and attached cute anime maid pictures to their posts

>>109175697
They just stopped posting as soon as normal programmers made a couple of the threads.

>company says we can subscribe to some LLM
>get claude
>try claude code
>hate the terminal
>make it implement claude-code features in emacs
>claude API has separated billing
>switch to chinese provider
well that was silly

Is it worth it to focus on C++ in the big 2026 after been learning Rust for the last few months?

I think Rust is awesome and love everything about it, except the community. But, I also really want (and more importantly, NEED) a job.

>>109177178
Yes, C++ will give you better job security. Also understanding Rust and why it does what it does can make you a better C++ programmer.

>>109177178
C++ finally has static reflection so it finally has everything I personally want from it
>But, I also really want (and more importantly, NEED) a job.
then reskill into a different industry
but being a proper polyglot can't hurt
maybe you could say you specialize in migration from C++ to rust and cite that cybersecurity bulletin the rustfags in biden's govenment had issued that basically tried to ban C and C++

>>109177178
For sure, I mean if you're proficient at Rust you should be able to find a job too...
But learning C++ will also make you a better programmer, so there's no "don't do it" answer to the question of "should I learn more C++?"

>>109175697
I remember them being here last year.. Then I dissapeared for a bit, came back and they made a couple threads late last year. I thought they got banned by the jannie

>>109175226
Is there? I find good discussion on twitter every now and then. Mostly when I am looking at ML/DS topics.
I found bluesky to be borderline impossible to use. Important settings hidden on different pages, I swear one of them was on their fucking bluesky moderator account to tell them I don't want anything filtered.
Maybe I'll log back in and take a look

>>109175086
learning assembly will actually teach you that if that’s your goal. C is abstracted away from the hardware enough to actually be portable. It’s good if you want to get fundamentals down for straightforward procedural programming without foot gunning yourself with a million retarded abstractions. It’s also very small and simple in feature set compared to modern languages. And yes, you basically do have to reinvent the wheel because libc is trash.

>>109175086
C doesn't teach you what's going on under the hood. It teaches you just the basics of systems programming and the facto standard for FFI.

How do I stop procrastinating and doing other things when I'm supposed to be programming? I think I might have undiagnosed ADHD since I constantly start thinking about different things when programming. How do I make programming more interesting? I'll usually wait like 2 hours in the morning procrastinating.

>>109179181
change your environment

File: GMhHdgAbgAAe7Q7.png (115 KB, 1178x1381)

115 KB PNG

>>109145475
>>109145475
I often hear people say the advantage of exceptions is the lack of runtime checks everywhere. But in that case, don't you still have to check for the erroneous condition to know when to throw an exception? What then is the difference exactly?