Seriously. Why need something else when there's modern picrelated?

Because languages that try to do a little bit of everything are always complete abominations.
Or they just reinvent Lisp.

>>105987059
Bruce won.
Sharty won.
/g/ lost.

>>105987044 (OP)
c++ is an absolute abomination and painful to read
It really deserves the diseased footless rat as its mascot

>design modules so we can finally get rid of the whole header/source rigmarole
>dogshit support by the common compilers and tooling
just.

>>105987044 (OP)
because the committee keeps adding random shit to the language but refuses to fix fundamental flaws. like proper unicode support. there's not even a fucking function in the std to split a string by delimeter characters. like jesus fucking christ. and don't even get me started on things like differentiating between a pointer that points to a single thing int* vs a pointer that points to a bunch of things and knows it's size. something like this int[]* would be tremendously helpful and would make your intent more clear to other programmers while also catching a large amount of bugs. manual memory management could be so nice and easy if these old fucks actually tried to improve it

>>105987044 (OP)
C++ would be better if it was simpler and more functional. C++ is headed in the functional direction, but the fact that the language is still essentially OOP sucks. Ocaml with manual memory management would be the ideal C++ replacement

>>105987059
90% of these complaints are just skill issue. does this idiot not know C make has a gobal_recursive command to get all the hpp and cpp files so you don't have to edit the cmake over and over? why would raii guards offend a cnile? why does he use bad testing practice and decide halfway through to implement it and be surprised he is a fucking retard?

>>105987044 (OP)
Because no matter how hard you want to larp as le serious programmer, managing memory manually is a pain in the ass and is unnecessary in 90% of usecases.

>>105989734
if you care about performance and lifetimes it's a must. not every project has to be a web page or shitty calculator

>>105987044 (OP)
std::cout << "hiv";

>>105987044 (OP)
overhead

>>105989734
C style memory management is optional. unique_ptr, shared_ptr, and references mean u dont have to new/delete everywhere. you can also take advantage of RAII with constructors/destructor if u want to have easier memory management
>>105990001
the only significant overhead is polymorphic functions because of the vtable pointer and exception handling, but these are both optional.

>>105987205
the standard refuses to address unicode becaue C++ is a system programming language. All the unix world uses utf-8, while all the nt world uses utf16le. It would be really difficult to unify both so they prefer to leave it to the programmer to know what they're doing.

Actual proper unicode support is not something that exists out of the box in a language. It's usually in a library and you have to manually specify exactly how you want to handle it. The tables required to canonicalize user input to a sane, normalized format take up multiple megabytes once statically linked. You don't want that anywhere in your language. Rust only lets you iterate over valid unicode strings (Strings in rust must be valid utf-8), that's just the very basic of handling unicode. For instance rust has no support for iterating over grapheme clusters, only codepoints.

>>105989734
C++11 literally fixed every memory handling problems by introducing move semantics and smart pointers. Post C++11, it's malpractice to ever have to call new/delete by yourself (unless dealing with older code/libraries). unique_pointer has 0 overhead and trying to copy a unique_pointer will generate a compile-time error instead of silently succeeding if you use raw pointers. Now I know that shared_ptr is not perfect, especially since it's 'always atomic' excepts when libstdc++ decides otherwise.

>>105991579
runtime polymorphism is not an overhead because there's no other way to achieve dynamic dispatch without using either function pointers or vtables. There are some times where you'll absolutely need dynamic dispatch (even in C), so you've gotta have some way to call the right function. When you work with a proprietary library whose source code you don't have access to, dynamic dispatch lets you decide what the function called by that library will do, it's not like you could modify the library's code and force your handlers to it.

>>105987205
unicode support support has no business being part of the language (or any language). Unicode support is a complex functionality that should be implemented as a library.

C++ already has too much shit in its standard library. Fixing c++ is not adding more to the standard library rather the standard library should be gotten rid of entirely. They need to move the standard library into a separate standard, and then focus on actual language features.

>>105987044 (OP)
That's like asking why would anyone speak anything other than Arabic.

>>105987044 (OP)
Because browns can't use c++ and they Will replace us

>>105987059
>manually rewriting functions/methods
Use an IDE that does it automatically, tard
>Manually add files to CMake
Same
>Manually manage your resources
More control is a benefit not a drawback. ie: You get the opportunity to finely control how your application functions from a memory perspective
>pitfalls in...
Such as?
>Write 5 different ctors for each class
You're doing it wrong if you're doing this. Now we know it's a skill issue.
>Every major update makes previous idioms old and obsolete
Only happens for the really major updates that are released decades apart. C++98 vs C++11 vs C++20. The old idioms are still fine - the new stuff is syntactic sugar on top of them.
>Manually write scripts to... dependencies...
Build tools are not a part of the language.
>No testing framework
I've used at least 3 different ones in my decade long career so far
>Resource leakage
lol, only if you're a shit coder. And why would you have to "rewrite" to fix bugs like that? Because your design was fundamentally broken in the first place? And wtf is that about "destructors not being invoked because of an exception". That's like the whole point of destructors.
>smart pointers
"hyper modern" LMAO
>docgen
Uh, these exist as well...?
>stdlib doesn't have X, so you need an additional library
Ok...?
>template tricks
Such as?
>12 hour compile time
You did something wrong
>Pages of STL error messages
Ok this is annoying. One valid complaint. Although in practice it's fine because you only need to read the first error anyways, and your IDE can redirect you right to that. So in practice, not actually an issue.

if this language is so shit why hasn't there been a competitor to displace it yet in the past few decades?

>>105992377
>>Pages of STL error messages
just use -Wfatal-errors to terminate compilation upon the first error

>>105992391
Rust just werks.

>>105991823
the std contains all kinds of shit that you can't use when writing operating systems or embedded systems. those people just use some kind of -no-std flag anyway. unicode is here to stay. why c++ shouldn't support it?

>>105992377
>Just use a bunch of bloated software to work around the flaws of a shitty language
No thanks.

>>105992976
You fail to realize that you are asking for the "bloated software" to be integrated into the language.
ie: "There's no build system" just means "I want to be forced to use the bespoke build system"

>>105987059
mentally ill

>>105993146
I would rather use a language that solves these issues on an architecture level instead of having to rely on some third party IDE that tries to fix the mess later.

>>105987059
Based bruce dabbing on seplets even to this day

>>105992391
if javascript is so shit why hasn't there been a competitor to displace it yet in the past few decades?

>>105992377
>>pitfalls in C++/templates
>Such as?
Very dishonest. First look at vector<bool> tells you about the rest of the language

>>105992900
We have established unicode libraries. Trying to implement some half baked garbage into the language spec would just be giving us another crippled white elephant that nobody will even use. The language already has enough garbage in the standard library that nobody uses.

>>105987044 (OP)
The people who manipulate the markets control the need, and it so happens that jeetcoding on something else less efficient is cheaper and gets the job done for NPC's who can't discern the quality of what they're consooming anyways.

Say what you will about C++, but some of the best software ever written was written on this clunkyass incel frankenlanguage, and this is coming from a straight C guy.

>>105993379
>but some of the best software ever written was written on this clunkyass incel frankenlanguage
That's because at that time there was no real alternative.

>>105993386
And I don't see a better alternative just yet. It would be fun to start over again and get everyone to jump on the Dlang wagon, though.

>>105993403
D has a GC baked into the stdlib, it would not be a suitable C++ replacement. Today we have Rust and soon Zig.

>>105993403
Rust is a good alternative if you ignore the trannies.

>>105993414
>D has a GC baked into the stdlib, it would not be a suitable C++ replacement.

That's one of the reasons I said there isn't a better alternative just yet, and also the reason I'd like to see D taking over in an imaginary scenario where you don't have to mind anything C++ whatsoever.

>Today we have Rust.

Ack.

>and soon Zig.

Dead on arrival. Too complicated to maintain for the little potential benefit it provides, assuming it will actually grow into something more c++ like.

>no suitable C++ replacement

odin

>>105987044 (OP)
This >>105992976
I can't stand having to fight the language to get shit done.

C++ can be a pain to get dependencies linked and working, but once it finally compiles the flood of serotonin is something no other language can deliver

>>105994230
but odin is a modern c not c++

>>105987044 (OP)
Until modules are usable I won't be touching C++.

>>105994230
Buy an ad you subhuman nigger.
No one likes your meme language.

>>105992377
>Build tools are not a part of the language.
>>stdlib doesn't have X, so you need an additional library
>Ok...?
Holy shit kill yourself you fucking subhuman.
>manually copying header files into my build directory and writing 2000 lines of cmake is totally normal

>>105987059
FBBP

>>105993496
>I said there isn't a better alternative just yet,
There is and it's called Rust.

>>105995611
>Better
>Rust
I would rather use Javascript for low level programming than get the borrow checkers cuck.

Modern D beats modern C++ in every possible way, is safer, and gets in your way less.

>>105992391
There is, it's called D. It can natively import and compile C if you must. It can interface with C++. It's the best language you aren't using.

>>105995762
It's a good IQ filter

>>105995958
>>105995951
and yet nobody uses it

>>105996114
They should.

>>105995958
D would be better if they stopped trying to be Java and went back to marketing to C++

>>105996109
Since a lot of the people that couldn't handle C++ went to Rust? Would Ruby-developer-to-Rust-evangelist Steve Klabnik fit that category?

Considering the bugs I've cleaned up in Rust developers' Rust code, I can't say that I'm impressed by the quality of all Rust developers' code.

Unless you are referring to the people that like something new, novel and interesting. Or researchers like Ralf Jung, that appear to benefit from the broken type system of Rust, that one can spend many years doing research on trying to fix. Or the difficulty of wrangling unsafe Rust, which is not yet specified, and which even the Rust stdlib developers can't handle.
https://materialize.com/blog/rust-concurrency-bug-unbounded-channels/

>>105997411
Steve made it past the BC, you couldn't.

>>105989734
>>105991656
>C++11 literally fixed every memory handling problems
this helps against memory fragmentation... how?
it is absolutely a problem when you have software intended to run uninterrupted 24 hours per day for the entire year

if anything we need private heaps in C++

>>105997411
>Since a lot of the people that couldn't handle C++ went to Rust?
Rust seems to be popular among developers who did get into C++ and realized how much it sucks. A lot of Rust design decisions are a response to stupid shit you deal with in C++.

>>105997428
Yet I fixed bugs in Rust developers' Rust code that they failed to figure out and fix themselves.
And Steve Klabnik is famous for being an evangelist. Does he have any technical achievements to his name?
Are you Steve Klabnik?

>>105997446
True, C++ is indeed burdened by backwards compatibility, though that backwards compatibility does have benefits. And Rust introduced a lot of baggage and issues and fundamental problems as well. I was hopeful about Rust in the past, but Rust turned out significantly worse than I hoped for, and the Rust community is overall not honest about even glaring issues and aspects of Rust.
Especially Steve Klabnik loves to lie. And that is anathema to a technical community.

https://materialize.com/blog/rust-concurrency-bug-unbounded-channels/
https://media.defense.gov/2022/Nov/10/2003112742/-1/-1/0/CSI_SOFTWARE_MEMORY_SAFETY.PDF
https://bpb-us-e1.wpmucdn.com/sites.gatech.edu/dist/a/2878/files/2022/10/OSSI-Final-Report-3.pdf
https://archive.ph/uLiWX
https://archive.ph/rESxe
https://lkml.org/lkml/2025/2/6/1292

l
>>105991656
>You don't want that anywhere in your language. Rust only lets you iterate over valid unicode strings (Strings in rust must be valid utf-8), that's just the very basic of handling unicode. For instance rust has no support for iterating over grapheme clusters, only codepoints.
And that's good enough. Much better than having no support whatsoever like in C++. Every other popular language already do this.

>C++11 literally fixed every memory handling problems by introducing move semantics and smart pointers.
And it couldn't even do this right because the non-destructive move semantics add pointless overhead.
Also overusing smart pointers adds even more pointless runtime cost over lifetime based compile time checks. You might as well just use Go at this point.

>unique_pointer has 0 overhead
False.
Moving out unique pointer has to leave previous pointer in a valid state that sometimes has to be checked for null even if it always gets moved out. That's an overhead that doesn't happen with raw pointers or Rust's move semantics.

>>105997490
>fundamental problems
What are the fundamental problems of Rust exactly?

>>105997452
Yes, I am Steve Klabnik. I am far more famous, rich and successful than you will ever be. lmao.

>>105997516
Have you asked that before, and gotten good answers, and now pretending to not know?

>>105997534
I have asked it multiple times and never heard about any problem of Rust that are fundamental.


You are yet to answer this question.

>>105997519
>far more famous, rich and successful

>>105997431
Than make your own allocator. The entire stdlib supports using custom allocators for all kind of purposes. Even the new keyword lets you specify an already allocated page that you mmaped/virtualallocated to use for your newly created object.

>>105997490
I agree, rust has many holes in the language/stdlib. For instance you can't (yet) constrain a closure to a specific lifetime. It's trait system prevents SFINAE bullshit, but it has problems of it's own, for instance HRTB is a real pain to deal with. I feel like every time I want to use some nice feature in the language, I find out it's not currently stable or fully implemented.

>>105997509
I still don't know why they didn't add destructive moves yet. That would improve moving by a lot. Also, it takes like ~20 lines to make a utf-8 forward iterator, add ~10 lines if you want to iterate backward. I think it's better to ship no support rather than ship broken or partial support. At least it gets people to research on the topic instead of blindly trusting the standard.

>>105997542
I see, you're that incompetent, schizophrenic Rust evangelist that hangs around here. You got triggered early, so it's a fair bet that you're actually Steve Klabnik or a friend of his.

>>105997550
>I agree, rust has many holes in the language/stdlib. For instance you can't (yet) constrain a closure to a specific lifetime. It's trait system prevents SFINAE bullshit, but it has problems of it's own, for instance HRTB is a real pain to deal with. I feel like every time I want to use some nice feature in the language, I find out it's not currently stable or fully implemented.
Nice try, Rust evangelist.
https://materialize.com/blog/rust-concurrency-bug-unbounded-channels/

>>105991579
>Exception handling is optional
I hate this fucking meme so much it's unreal.

libc is also optional in C too.

>>105997550
>I think it's better to ship no support
I think it's not.
A lot of people think it's not.

>>105997555
>can't answer simple question
>gets mad
>throws ad hominem
>nooo, you are the one triggered!!!111
Kek

>>105997560
>basic response without substance
>reposting some unrelated tech blog over and over
Bot or advertising or both?

>>105997623
>>105997651
Blatant lying, as well? Yup, you're that dishonest, incompetent, schizophrenic Rust evangelist that hangs around here.
Let's call you Triggcrab.

>>105991656
>this thread
C++ apologists are literally unhinged..
I literally do not understand why people carry water for something so bad. same goes for C too, btw.

>>105991656
>>105997694
Monologue.

>>105997685
>can't answer simple question
>accuses of dishonesty and incompetence
>continues with ad hominem
>no, you are the one being triggered!!111

>>105997542
https://github.com/lcnr/solver-woes/issues/1
>Even worse, there may be changes to asymptotic complexity of some part of the trait system. This can cause crates which start to compile fine due to the stabilization of the new solver to hang after regressing the complexity again. This is already an issue of the current type system. For example rust-lang/rust#75443 caused hangs (rust-lang/rust#75992), was reverted in rust-lang/rust#78410, then landed again after fixing these regressions in rust-lang/rust#100980 which caused yet another hang (rust-lang/rust#103423), causing it to be reverted yet again in rust-lang/rust#103509.

will polonius get faster than 5000x slower?

>>105997714
The only one lying here is (You), Triggcrab. And you even lie repeatedly.

>>105997735
if this is considered a fundamental problem, then C++ is basically "unusable" by such admission.

>>105997746
>ask a question
>nooo, stop this you are lying
>also you are triggered!

>>105997735
How exactly is this a fundamental problem if it's not even added to the language yet?

>>105987044 (OP)
Deprecated since 2020 anon. It's time to get rusty!

>>105997713
what are you implying? same fag? why would I say retarded C++ babble and then call myself unhinged?

the reality is, I can
cargo init .and start writing CODE

in C++, I have to first copy one of many different CMake templates or choose to lock myself into Visual Studio before I can even start programming. On top of that, I also have to fight with over 10 different CMake-based package managers to replicate what cargo does in one line of configuration.

Rust and rust-analzyer is an editor and IDE agnostic powerhouse that just fucking works.
I still have to fight with gcc/clang compatibility issues for clangd, deal with idiotic defaults like "automatic header insertion" and I still need a build system that either makes compile_commands.json or use some insane cope tool like bear.

again, the reality is C++ is shit. anyone who's actually interested in making software should plainly see this, unless they're so far up their ass in babby duck they can't cope with the reality that Rust has completely won.

>>105997713
what are you implying? same fag? why would I say retarded C++ babble and then call myself unhinged?

the reality is, I can
cargo init .


and start writing CODE

in C++, I have to first copy one of many different CMake templates or choose to lock myself into Visual Studio before I can even start programming. On top of that, I also have to fight with over 10 different CMake-based package managers to replicate what cargo does in one line of configuration.

Rust and rust-analzyer is an editor and IDE agnostic powerhouse that just fucking works.
I still have to fight with gcc/clang compatibility issues for clangd, deal with idiotic defaults like "automatic header insertion" and I still need a build system that either makes compile_commands.json or use some insane cope tool like bear.

again, the reality is C++ is shit. anyone who's actually interested in making software should plainly see this, unless they're so far up their ass in babby duck they can't cope with the reality that Rust has completely won.

>>105997749
>>105997754
kek, try again

>>105997786
skill issue. cmake is really easy to use. rust's cargo is a pain when the deps aren't available on official repos or when you're forced to use non-repo deps

>>105997787
see >>105997786

seems like Rust isn't fundamentally broken to me. I am able to shit out code in less than a week and wagies kneel to me because my shit is "just an .exe" they can double-click without autism.

where is the "fundamental" issue? I know C++ has a shitload, mainly that the grammars of the language aren't even deterministic + lack of cohesive build system makes it hard to actually use.

>>105997801
>cmake is really easy to use
not according to posters ITT like >>105995383

sounds like you're a baby duck stockholm syndrome retard.

>>105997750

>>105997806
kek
see >>105997735

>>105997861
Ok, what about it? Can you use words or are you retarded.

>>105997786
>>105997801
>>105997815
Too much incompetence in each of these posts, must be a monologue.
Also: >>105997779
Fucked up the formatting, more incompetence by (You).

>>105997801
>deps aren't available on official repos or when you're forced to use non-repo deps
dependency = { path = "path/to/crate" }

>>105997877
you forgot a question mark, retard

I prefer C#

>>105996956
Jeremy Bicha.
>SEX BAT BY JUVEN/VCTM UNDER 12; F.S. 794.011(2) (PRINCIPAL - 2 COUNTS)
Rust developer.
PPA for Rust.
https://launchpad.net/~jbicha/+archive/ubuntu/rust

>>105997787
>rust has fundamental issues
>what are the fundamental issues?
>a new, work in progress solver has performance issues
>how is this a fundamental problem?
>kek, try again
What?

>>105997887
answer the question, nigger.

>>105997912
try reading comprehension, retard >>105997735

>>105997937
How is this a fundamental problem if the new solver is not stabilized yet?

>>105997919
try asking a non-retarded question, retard

>>105997886
that's assuming the dep is in crate-compatible format

>>105997945
>I am incapable of critical thought and can only regurgitate unrelated details that has nothing to do with the discussion at hand nor explain how they're relevant
that's literally (You) and you think (You) have the right to call the question retarded?

>>105997940
try reading comprehension before asking retarded questions, retard >>105997735

>>105997947
Why would you want to use cargo with non-cargo rust projects?

What is the exact scenario are you talking about. I never ever had to use an external Rust project as a dependency that isn't a crate.

>>105997976
How is a solver performance problem fundamental?

>>105997976
>still can't answer.
being schizophrenic should be a global rule b& desu.
I'm guessing this supreme retard is reading something that none of us non-schizos can see.

>>105997956
TriggCrab, why so triggered?

>>105997937
>>105997912
affine types have exponential inferring complexity. if you've ever used a large rust project you might have noticed the absolute hell that it is to compile and link. for instance, linking against bevy, which has to happen after all compilation takes around 5 minutes in debug mode. it's even longer in release mode with lto enabled.

>>105997947
>copy code into project
ok. done.
>b-but what if it's not structured correctly as a module?!?!
include!() exists.

>>105998004
if you do that, you loose the ability to modularize the code in such a way that only what's changed gets recompiled.

>>105998002
this has nothing to do with the linked issue and C++ has the same problem.

>>105998013
>I'm retarded and do retarded things.
build.rs can cover every single retarded use case you can invent retard, including last-modified tracking of your insane non-crate code.

>>105997956
>>105997990
>>105997992
>no u
failed reading comprehension or willful ignorance? (You) are retarded either way

>>105998002
>>105998014
Monologue.

>>105998040
>failed reading comprehension
being unable see things that aren't there isn't failed reading comprehension, however, it is a sign of schizophrenia. take your meds schizo, or ideally, kys since meds rarely work without talk therapy, supposedly.

Steve Klabnik being called incompetent sure triggered the hell out of him and his friends. He could have just shut it down by pointing to his technical achievements. Steve Klabnik does have technical achievements to point to, right? Right?

>>105998068
>Steve Klabnik
literally who?
is this one of those random people /g/ draws up to cover their own incompetence? is this the next ebussy or something?

>>105998002
>for instance, linking against bevy, which has to happen after all compilation takes around 5 minutes in debug mode
What?
bat src/main.rs
$ cat src/main.rs
use bevy::{color::palettes::basic::PURPLE, prelude::*};

fn main() {
App::new()
.add_plugins(DefaultPlugins)
.add_systems(Startup, setup)
.run();
}

fn setup(
mut commands: Commands,
mut meshes: ResMut<Assets<Mesh>>,
mut materials: ResMut<Assets<ColorMaterial>>,
) {
commands.spawn(Camera2d);

commands.spawn((
Mesh2d(meshes.add(Rectangle::default())),
MeshMaterial2d(materials.add(Color::from(PURPLE))),
Transform::default().with_scale(Vec3::splat(128.)),
));
}
$ cargo run
Compiling bevytest v0.1.0
Finished `dev` profile [unoptimized + debuginfo] target(s) in 1.89s
Running `target/debug/bevytest`

Compilation + download + linking from scratch took exactly 1 minute (Finished `dev` profile [unoptimized + debuginfo] target(s) in 1m 00s)

>>105998104
most people who complain about Rust compile times literally never wrote anything nontrivial in C++ (or even C). It's extremely common. At least in Rust, the build system will usually do the right thing(tm) by default instead of given the rope of CMake to force you to autistically tweak your build targets til you can take your 2 hour long release build down to about 30 minutes.

t. have done the above and broke code because of static init ordering bugs in the code base.

>>105998053
>take your meds schizo, or ideally, kys since meds rarely work without talk therapy, supposedly.
>supposedly
lol
lmao
do you have personal experience with being treated for being schizo or something?
the only schizo here is (You). try again, schizo retard. or follow your own suggestion to schizos and kys

>>105998129
no, I am friends with an immensely successful schizo who currently lives a long and fulfilling life on medication and is also a Rust programmer nowadays, for M$.

>>105993336
>>>105992377
>>>pitfalls in C++/templates
>>Such as?
>Very dishonest. First look at vector<bool> tells you about the rest of the language
...
>"Very dishonest!"
>goalpost - "metaprogramming is <broken/bad/made me shit my pants/etc>"
>nvm
>goalpost - "I HAVE A PROBLEM WITH STD VECTOR THAT DOESNT BOTHER A SINGLE SOUL"
uh-huh.

This entire thread is nothing more than first year compsci students sperging into some combination of thin air and GPT. Sad.

>>105998098
Post your technical achievements, Steve, if you have any, instead of trying to defer your incompetence to everyone else.

>>105987189
is ai at the point today when it can portray this mascot rat in a provocative pose showing feet on camera as if its an anime woman or onlyfans person

>>105998141
NTA, but I struggle to understand your post. He clearly claimed that there are pitfalls, not that the system is broken. vector<bool> is a good example of C++ pitfalls. This quirk seems to be quite widely criticized too. I personally also think this was a really retarded move from C++ designers.

If we want to talk how C++ metaprogramming is broken, all we really need is to point at SFINAE.

>>105997904
>>105998138
Fucking hell, schizos, trannies and convicted pedos, is that all that Rust developers are?

>>105998141
vector<bool> is actually a huge problem thougheverbeit and the fact that you don't understand it tells me you don't actually understand C++.
such a specialization breaks a lot of assumptions, especially when it comes to shared mutable access. i and i+1 aren't necessarily independent. A lot of code bases have to cope with <bool> when using vectors or just remove them.

>>105998184
>t.

>>105998182
>all we really need is to point at SFINAE
kek. I still am on Visual Studio 2019 because of a SFINAE bug. I think they finally fixed it, but why bother upgrading? I should probably rewrite it in rust to make the schizo C++ tranny ITT mad.

>>105998194
https://www.offenderradar.com/offender-details/jeremy-c-bicha-of-florida-176493
https://launchpad.net/~jbicha/+archive/ubuntu/rust

>>105998208
>more schizo-posting
take your meds retard. no one can read your demented mind. maybe when you're medicated you'll transition to Rust programming and get that M$ job as well.

>>105998208
Based

>>105998219
>By the time she was 9, Jennifer’s brother Jeremy, then 12, began a pattern of regular sexual abuse that grew aggressive, then violent, and escalated into rape.
Are you denying >>105998208 ? Stop letting your actual schizophrenia spill into this thread. And be more gentle when referring to yourself.

>>105998128
I mean, compile times are not the best, but if your program takes 5 minutes to link you are doing something really wrong or are just making shit up.
90% of my build process is flashing the MCU anyway so who cares if compilation takes 5s or 30s. I do not run compile every minute, I have language server to do checks for me in the background.

>>105998182
>>105998193
>If we want to talk how C++ metaprogramming is broken, all we really need is to point at SFINAE.
Sure, go ahead. Meanwhile I have no problem supporting MSVC2010 and early GCC 4.x with just a handful of SFINAE macros. Everything I could possibly want a test for just werks.

Let's see if I'm following the argument
>reee there are million pitfalls in "templates"
>>reee youre being dishonest, muh std::vector<bool>
(because apparently, something something a standard generic type, this equates to "millions of flaws in templates")
>>>i dont understand
>>>>actually SFINAE is the problem
You troons are utterly demented. I'm sorry you got filtered by std::vector<uint8_t> and equate that to SFINAE and templates themselves being broken.

C++ has the most poorly designed standard library to ever exist.
Not only this, it is barren and pathetic in comparison to Java and C# standard libraries.

>>105998193
>>105998250
Monologue.

>>105998250
>SFINAE macros
What even is a SFINAE macros

>>105998068
I guess Steve Klabnik has no technical achievements to show. Guess that's why he became an evangelist. I prefer evangelists that also have technical achievements under their belts.

>>105998296
>"what are macros"
As I said, "This entire thread is nothing more than first year compsci students sperging into some combination of thin air and GPT. Sad."

>filtered by std::vector<uint8_t>
>filtered by retarded <= c99 preprocessing
>filtered by post c++11 static_cast SFINAE
>believes ref counted pointers are "hyper modern"
>"rustchads are totally ex-c++ experts and not an echochamber of incompetents circlejerking their 50 lines of functional slop, 200 lines of broken tests"
Actual dementoid behavior

>>105998343
I know what are macros, but I never heard of "SFINAE macros".

I am just curious what kind of eldritch abomination could come from combination of preprocessor style macros and SFINAE hack. Whatever that is, it must be a sight to behold.

>>105998250
>MSVC2010
>and early GCC 4.x
how to spot the cuck who has to maintain a woefully broken code base that doesn't work. many such cases in C++ land. thank god this isn't a problem for us Rust chads. Just updated to rustc 1.88 using Rust Edition 2024. Keep seething though.

>>105998369
It's really not that complicated.

Every compiler after 2012 will accept the same: "does child type", "does static variable", "does static variable of static cast", [...] similar tests. Write a macro for each use case where the end user as a struct derived from integral_constant. Then simply Ifdef for early msvc, replace with __if_exists for type/member detection; and ifdef for early C++03 compilers, replace the some of the member tests with mixin checker instead.

Claiming "muh SFINAE is a problem" is peak midwit. If you look at boost and any major project that takes older compiler support serious, 99% of the grief is around <type_traits>-like replacement utilities. Specifically finding compiler specific intrins and mostly sane fallbacks that mostly work on era correct tooling. Once you've got these core type_traits utilities down, perhaps in combination with some "does child type" or "does constant exist" tests, you're done. SFINAE isn't problematic at all.

>>105998343
>200 lines of broken tests
I ain't the one stuck on outdated and shitty GLIBCXX ABI because your code base violates the C++ standard and GNU's performance improvements breaks your extremely invalid code.

nothing is more immensely humorous than seeing these KWAB giga boomers or demented zoomers defend their shit ass language. They literally do not know C++.

>>105998423
Nobody is seething. I do find the psychology behind the zoomer interesting.

>decades of compiler technology bad
>rust latest version - 2 good
>must updooot
>ewww old thing bad
>ewww old thing totally never worked

>>105998465
NTA
>decades of compiler technology bad
Rust uses the same compiler backend as clang. LLVM is already 17 years old.

>>105998482
>Rust uses the same compiler backend as clang. LLVM is already 17 years old.
Show me your rust code that compiles on a 17 year old compiler

>>105998449
>Every compiler after 2012 will accept the same: "does child type", "does static variable", "does static variable of static cast", [...] similar tests. Write a macro for each use case where the end user as a struct derived from integral_constant. Then simply Ifdef for early msvc, replace with __if_exists for type/member detection; and ifdef for early C++03 compilers, replace the some of the member tests with mixin checker instead.
All of this instead of just using typeclasses.

>>105998491
Just because a compiler backend existed 17 years ago doesn't mean the frontend existed as well.
In the previous post you were talking about "decades of compiler technology" but now you are asking for 17 year old rust compiler.

What is your point?

>>105998449
Why not use C++ concepts instead? Everyone, including Bjarne, dislike preprocessor macros.

>>105998491
>>105998518
Sweet monologue, Triggcrab.

>>105998518
>>105998520

>>105998465
>I do find the psychology behind the zoomer interesting.
>>decades of compiler technology bad
>>rust latest version - 2 good
>>must updooot
>>ewww old thing bad
>>ewww old thing totally never worked

>>105998550
What?

>>105998040
>>no u
Never said so.
I only asked
>How is a solver performance problem fundamental?

>>105997940
>>105998609
>Never said so.
samefag memefag
>I only asked
>>How is a solver performance problem fundamental?
failed reading comprehension or willful ignorance? (You) are retarded either way)

How can Steve Klabnik be so incompetent and triggered? Is he really that useless at software development?

>>105998520
Concepts were originally proposed like twenty years ago, and then got thrown out by the retard committee for no reason. Then Bjarne back-pedalled and decided we need concepts again but gave us a shitty dollar store version of the concepts proposal from 20 years ago.

>>105999401
Source? Didn't Bjarne want something like concepts since decades ago?

>>105995354
But I like his meme language.

>>105987059
I use langs that can best accommodate my coworkers. If you are competent and like working with big tech firm corp, then c++ is the standard. Problem is that joe average likes to browse reddit a little too much so he gets fed propaganda from the 70s and 80s.

>>105987059
the header files was meant to be fixed by modules, but modules turned out to require interfaces anyway, fucking retarded coomite

>>105999778
http://www.drdobbs.com/cpp/the-c0x-remove-concepts-decision/218600111

>>105999873
Source on this? Are module interface declarations necessary or can you get away with not doing that?

>>106000303
Also the link here doesn't work, use Internet Archive

>>105999401
>>106000303
Bjarne was saddened by the removal, he had worked on it for years by 2009, if I'm reading the first page out of 3 correctly.
https://web.archive.org/web/20140110122738/https://www.drdobbs.com/cpp/the-c0x-remove-concepts-decision/218600111

>>105987044 (OP)
This is 28 programming languages in a trenchoat and it unifies the worst of all of them, plus some things unique to it so bad that nobody else could have ever come up with them.

>>106000533
Yet principles like zero-cost abstractions have inspired and guided successor languages.

I like that using concepts and requires, I can write classes with less concern for inheritance. It's really nice.
IMO inheritance should be an implementation detail more than an API concern

>>106000518
What was the reason for the removal?

>>106000577
Managed to dig them out.
Page 2:
https://web.archive.org/web/20150907061721/http://www.drdobbs.com/cpp/the-c0x-remove-concepts-decision/218600111?pgno=2
Page 3:
https://web.archive.org/web/20141217175208/http://www.drdobbs.com/cpp/the-c0x-remove-concepts-decision/218600111?pgno=3

Related:
http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2009/n2906.pdf

Bjarne mentions nominal typing and structural typing. I have not read the sources in depth, though.

>>106000846
why does bjarne write his reports in microsoft word unlike everyone else who write proposals in latex

>>105998141
astral-level projections

>>105994948
>it's good I have to jump through thousand hoops to achieve what should be the default, because it gives me a sense of achievement
Sounds like Stockholm syndrome.

>>106001063
Some use online webpage/HTML proposals. Who cares, as long as the end result is useful?

>>105987044 (OP)
deprecated by rust, chud

>>105987059
/Thread

>>105993496
>Dead on arrival. Too complicated to maintain for the little potential benefit it provides, assuming it will actually grow into something more c++ like.
Interfaces, closures, anonymous functions, operator overloading, traits, RAII, private fields, constructors, inheritance, and much more have all been explicitly rejected for Zig. They avoid adding new features like the plague, and are actually removing features when they feel like it. Mixins and async were removed from the language, and async will just become part of the standard library instead of a language feature.
They're focused on things like replacing LLVM, improving compile times, fully incremental compilation, safe recursion, improved debug info, a compile server for tooling, etc.

>>106002693
Pinnacle of worse is better

>>106002500
>>105995605
>>105993302
>generated documentation being *static* html is bad
>it should be dynamic with javascript instead
that guy is an idiot lol

>>106003923
>JS is le hecking bad because... BECAUSE IT JUST IS MKAY??
Enjoy 0 search functionality cnile

C++ is pointless
If you want to carefully manage memory use C.
If you want to do normal shit use go , lisp or lua or whatever

>>105987044 (OP)
Nothing else is needed

>>106004014
>go
opinion discarded
poo in the loo jeet

>>106002729
Not "less is more"?

>>106004014
Skill issue.

>>105987059
If you can't code in C++ do something else, we don't need more retards faking it.

>>105993336
That is a library, if you don't like it code your own.

>>106001953
>deprecated
Take your meds, schizo.

>>106004374
It's not just "a library" it's the standard library. Another dishonest gymnastics by /g/

>>105995762
Based AF

>>106004399
So what? Stop crying and get some other tool if the standard tool hurts your feelings.

>>106004450
>what's wrong with C++
>stdlib has pitfalls
>stop complaining
Looks like I struck a nerve. Why are you so emotionally attached to C++?

>>105987203
Only the try hard n00bs who have no abstraction skills asked for modules because their shit is leaky AF.

>>106004450
Interesting, so what alternative library do you propose we should use other than the stdlib? Boost?

>>106004472
>>106004515
Good morning saars!

>>105987059
bruce won

>>106004515
>>106007010
Is that not a Reddit comment?

>>106004515
abseil to go with your GoogleTest, but it really still won't save you. C++ is just that bad.

>>106003965
but it is

>>105987059
A lot of humor, but not always that much substance or logic.
Concepts, modules? How old is this post?
UB? Like this? https://materialize.com/blog/rust-concurrency-bug-unbounded-channels/

>>105987059
>>106007010
Is Bruce the "gnome gives you bitches" guy? If so, I respect his opinion, otherwise not.

>>106002693
>Interfaces, closures, anonymous functions, operator overloading, traits, RAII, private fields, constructors, inheritance, and much more have all been explicitly rejected for Zig
I will now not use your language