>Still the best language >And it keeps getting better Why

it keeps getting shittier, especially the syntax. by 2040 it will be an unreadable pile of special characters

File: C++.png (72KB, 1016x98px) Image search: [Google]

72KB, 1016x98px

I don't really like the idea of a language so utterly complex that it lets you do things 10 different ways, none of them are compatible with each other except when they ARE, and you're actively discouraged from using more than one of them at a time.

Meanwhile C just keeps getting better~

>>60563312
Does it support emojis yet?

>>60563312
it's the Ω language. God power!

/g/ is for discussion of video game hardware only, what are you guys are talking about

less maintainable than PHP

>>60563416
>template
>auto
god i absolutely resent these
why did they had to introduced them there a shit

File: download.jpg (7KB, 250x250px) Image search: [Google]

7KB, 250x250px

>>60563312
What a oldfag.

>>60563416
Wtf does that code do? It looks evil

>>60563466

>hating on templates
Also, auto makes smart pointers much nicer.

>>60563494
>smart pointers
why did they have to add these too?

C++ is one long extended sequence of mistakes

>>60563416
>give me less choice
>I'm too retarded to use the right tool for the right job
C++ is like an expert's toolbox. 10 different tools for hundreds of different problems.

Meanwhile shit like C is one hammer for every problem

>>60563511

>why did they have to add these too?
Because of your famous double free and memory leaks.

>>60563494
>smart pointer
whats wrong with regular pointers? are people to retarded to malloc and free? or new and delete?

>>60563312
It will have a usable standard library by C++ 75

>>60563555
Most people these days are afraid of "raw" pointers and allocating arbitrary amounts of memory and then casting to the type you want, because that's literally what malloc does.
They're usually webdevs.

C++ is what you use when you try to use C but get really insecure about newer languages' built-in features while suffering from inferiority complex.

>>60563555
Apparently. Insecure code is everywhere.

Don't worry anon, you can still malloc/calloc/free to your heart's content. You don't HAVE to use them.

>>60563555

>are people to retarded to malloc and free?
Clearly they are or you wouldn't have weekly reports of double-free exploits.

>>60563583
>They're usually webdevs.
cant they just keep to themselves? why do they have to ruin OUR language

>>60563622
>You don't HAVE to use them.
i do though, one of my superiours is wanting me to start using those because im the only one still used malloc and free in our work

>>60563638
>not being aware of your own code

>>60563655
are you this guy?
https://warosu.org/g/thread/S60303803

>>60563655

>I never call any code I didn't write

>>60563524
One hammer that works in a predictable way, that everyone knows how to use if you need help with it, and that swings the same way on every platform.

File: SimplicityVsSpeed.png (21KB, 480x339px) Image search: [Google]

21KB, 480x339px

The only reason C/C++ still exists because of majority of the code in today's operating systems are written with them and nobody is brave enough to rewrite the entire thing with the next high-level language. So if you want to interact with the OS in a natural way, C/C++ is way to go. At least until we create code-monkey AIs that can rewrite the entire thing in a modern language.

I like how it is fast and optimized, but it is also ugly as fuck and requires so much effort.

Why use C++ for everyday applications when I can just use Cython, pypy or whatever and get close to C-speed just in a couple of days instead of couple of weeks? (Excluding mission critical operations like satellite systems etc.)

>>60563638
RAII also works for C, once you hide your memory management behind

something_something_new
something_something_free

you won't have too much problems.

And with attribute cleanup you can autofree stuff once it goes out of scope.

>>60563672
kek no though im an other autist who does this
>technical debt

>>60563678
right forgot about other people i work low level implementations for embedded systems firmware so nothing is not in house code or how its called

>>60563701
> RAII also works for C
You don't know what RAII is, do you?

>>60563312
> C++
> How to do less with far more

If it gets any more bloated it will explode.

>>60563699
Many people are faster at writing C++ code than whatever scripting shit is hot at the moment.

>>60563699
Python speed is a meme, its good for some minimal scripting and thats it. The majority of software is already slow as fuck while throwing more resources down the drain.

>>60563749

I somehow doubt that.

It's funny that every somehwat serious c++ projects tends to stay way from every ""nice"" feature like overloading or metaprogramming and ends writing it's own lightweight stl version

>>60563749
lol bullshit

>>60563740
Useless comment. Feel free to point out where I'm wrong or stop wasting my time

>>60563749
The same fucking deal. If you know C++ it is fine, but if I am a newcomer why the fuck should I learn it when there is an easier cost-efficient option? (Except javascript, fuck javascript and all of the webdevs and their bitch ass frameworks and shit)

>>60563791
It is not a meme, you can literally compile Python with Cython compiler. People are just fucking lazy, they are using some simple if-else, while, for constructs w/ default interpreter and all of a sudden they see themselves as Python Overlords. There is bash scripting for that kind of shit.

Python =/= CPython

>>60563312
I am still writing in C++98...

>>60563999
LOL Look at this faggot!
I bet you are as retarded as this retard
https://gist.github.com/bkaradzic/2e39896bc7d8c34e042b
Seriously if you cannot write sepples or change to better language you should just kill yourself.

>>60563699
>when I can just use Cython, pypy or whatever and get close to C-speed
a. you can't
>inb4 muh trivial, numeric benchmark
b. why Python if there are languages that are faster to write and to run

>>60563312
Well...
>still no modules, a feature that should have been in there from day one, using headers only for legacy C support
>JUST NOW they have std::variant included when they just should have supported std::string in unions from the beginning
>there are at least 3 ways to pass function references, but they are either slow or ugly
>compilers are still slow, a big part because of headers
>still no functional package manager
>all kinds of template shit instead of simple compile time execution + AST manipulation API
>one of you faggots told me that the random number generators are garbage

>>60564124
You can. There's cython and c-api in python.

QT is kinda making me like C++. So far it seems superior to GTK.

>>60564382
Your post indicate that you're a C++ developer. Stick with that language please.

>>60564382
QT is pretty based

>>60564342
You can't.
By the time you get remotely fast in cython you are no longer dealing with semantically correct Python. Similar applies to PyPy, except the situation is more hopeless in general.
And bindings won't help you, as function calls easily become a bottleneck and you have to touch C again.

>>60564382
It is. By a long way.

>>60564423
C is great and all but I'm starting to realize that it should stick with very low level processes, ones that the user doesn't have to interact with.

>>60563312
>C++/CX C++/CLI C++11 C++17
They themselves don't know what they are doing.

C++ is so tough that I gave up.

>>60564462
And C++ for everything else?

>>60564525
If we're talking about outside of the terminal, yeah.

>>60564564
GUI?

>>60564590
Definitely

>want to make small application for personal use
>figure "I don't have a ton of C++ experience, so why not get some practice?"
>decide to use C++
>Google "C++ get files in a directory"
>Stackoverflow thread top result
>click it
>"C++ doesn't support this."
>say fuck it and write it in C# in 20 minutes

I just checked and the SO thread was updated to inform people that C++ finally supports directories in C++ 17.

It only took 40 years...

>>60564604
There are better language that C++ for that.

>>60564441
>You can't.
For 80% of the everyday programming scenarios, it is more than enough. And if your needs fall in that 20%, and you need to squeeze the performance out of your CPU then it can be defined as a mission critical operation. Then yeah, C might be a better option.

Say I am programming a bruteforcer program for some pentesting scenario, I don't need to write everything in C. I can just use a C library for key generation, and Python for everything else. And if I am losing 10% of the performance due to some bottlenecks for the sake of simplicity, so fucking be it.

And if you are really fucking crazy when it comes to performance, there is always the Assembly option. In theory, it might be even faster than C. But yeah...

So I guess it all comes down to your needs and priorities etc.

>>60564686
Or you just use a modern compiled languages that is not only less error prone to write than Python - and let's not even talk about C because it's a garbage language - but also will never run into this 20% problem that requries rewriting your application since it usually is such a core part of the application like a spatial/rendering viewport with a lot of data behind it that the user has to interact with in realtime. (Screw your nonsense pentesting input work output scenario.)
Instead of reverting back to ignoring the problem and hoping your next C library/browser update will speed it up 10% so the user might be less inclined to complain, but actually this improvement will never come because you can't fix a broken, bottlenecked layer architecture.

>>60564850
Tell me a compiled language that has the same simplicity as Python and is not a meme.

>inb4 Rust

>>60563474
>feel for the Rust meme
Nice one pajeet

>>60564917
D or golang(if you are degenerate), for example. C# or Java work as well. All of those are better options than Cython. OCaml. Swift.
>inb4 Rust
Rust.
>meme
Python is a meme, along with your pentesting example, data ""science"", deep learning and simple IO cycles that were a thing 30 years ago.

>>60564931
>feel

File: .gif (1MB, 600x600px) Image search: [Google]

1MB, 600x600px

>>60565009

>>60563312
The language is a botnet, meme and CIA nigger.

>>60565079
>t. rustfag

I unironically love c++ but

https://en.m.wikipedia.org/wiki/Greenspun%27s_tenth_rule

>>60565142

>en.m.wikipedia.org
Leave, phoneposter.

>>60563524
Call me crazy, but I think that's a point in C's favor.

>>60565252
ok, jerry

>>60565142
>https://en.m
stopped reading right there

>>60563416
Can't they just create an easy readable language?

>>60564626
>man readdir

>>60565535
it already exists, it's called C

>>60565629

Write equivalent code to >>60563416 in C, let's see how readable it is.

>>60565672
First explain to me what that monstrocity does then I'll write you something.

>>60563436
I wish this weren't true

>>60563312
You cant have an intelligent conversation about programming languages with a C++ developer because they are all fanatics who just want to shoot down every other language.

Bjarne Stroustrup is the ONLY language designer I know who spends all his time in language wars trying to prove his language is better. Every other designer of every other major language are humble guys who can accept other languages. Bjarne wants everyone to stop using other languages and just use his language because he is on an ego trip and simply wants to maintain C++ as the ONLY available option.

I dont have time for language wars, I use whatever language is the best, and C++ is being over-used simply because major companies are afraid to try replacement languages

>>60565605
>Linux only

>>60563699
>be me
>learn programming C->C++->C#/Java
>use only OOP languages
>try Python
>it's fucking hard
How can I adapt to non-POO programming, everything feels weird

>>60565956

>C
>POO

What's hard about Python, anyway?

>>60565983
that wasn't what I meant, C was just the first programming language I've been taught, as a way of introducing programming and have c++ and POO as a follow up.

>What's hard about Python, anyway?
After using such extensive usage of POO concepts in all of my programming experience, having such a rudimentary POO system makes programming in Python harder for me, I haven't been able to learn it properly

>>60563416
>pic
that's not C++17.

>>60565687
it's an unnecessarily complicated way to allocate a unique_ptr with a custom deleter.

>>60563655
Web developers don't use c++. Also, our language? You mean basement dwellers?

And a lot of times you have to maintain code you didn't write. Have you ever programmed anything other than autistic console applications?

>>60565956
>try Python
>it's fucking hard
That's because Python is insanely idiosyncratic, its not a good fucking language to use unless you're only ever going to use Python.
Also it is a POO language, everything can be treated as an object in Python.

Basically, don't fucking bother with it.

>>60564626
What? I distinctly remember writing a program in C++ that does a tree display of all files on a partition as a homework in college in 2010.

>>60564626
It has been in boost in a long time.
Why do you need a language specific way to do something that will always be OS dependent?

>>60564465
>CX, CLI
That's Microshit's thing

Daily reminder if you use boost/Qt/awesomium or any other framework/lib that is the size of the entire andromeda galaxy you are worse than pajeets writing Java/Python because they accept they're retarded.


C is godmode, C++ is bloated piece of shit where people absorb planets and then that overhead gets halved just because its native and compiled with C++ so you end up being 10% faster than Java.

>>60566522
>not using QT
It's pretty based.

>>60566580
QT is probably the worst designes lib.

>>60566598
Name a better one.

>>60563699

> mission critical operations like satellite systems

Spook detected.

Why you'd want to use a language that isn't memory safe on a satellite is beyond me, anon.

>>60563312
For embedded and safety critical, there is C. For everything else, there's rust

>>60566522
c suck for bigger systems

>>60566884
Because apart from Ada/SPARK, it's the only viable choice for writing formally verified software, cuck.

>>60566888

> rust

Get the fuck out of here with that SocJus-pushed crap.

>>60563416
most C++ programmers wont even touch code like your picture so fuck off.

>>60566936
It's superior to C++ in every way and language≠community.

>>60566888
>>60566973
Go back to >>>/leddit/ plz.
Ryst babbies are not welcome here.

>>60566973
>It's superior to C++ in every way
List them.

>>60566932

> cuck

I said spook as someone that works for an Intelligence Agency, not the retarded political science wankery definition /pol/ uses it for.

> only viable choice for writing formally verified software

Maybe for very simple programs? Formal verification for complicated programs isn't a thing. Satellites strike me as something complicated.

>>60566973

You promote the language? You promote the community. They'll dictate what norms are permitted in all governance and community and convention spaces.

I'm not touching SocJus Rust.

>>60566580
no, qt is easy, qt is in all modern cars, qt gives you 10 FPS lag on two simple gauges that indicate speed, thats because Qt is EASY to write, because it has shit ton of overhead.

I will never understand how people can spend 100k+ on a car and then turn it on just to see 10 FPS display speed gauge, all because of codemonkeys who are unable to use OpenGL and simple trigonometry. Instead they use Qt and absorb entire planet into something that has complexity of a task made for 1st grade. All because of hurr durr my time to market and ease of use.

>>60567072
Because good code isn't valued in the software industry, they want a working product NOW and qt just happens to be quick to use and shit out something semi-functional.

>>60567009
The core systems are always formally verified. In extreme cases, you choose subset of C such that there are some properties satisfied already (see Saab avionics division), it greatly helps with writing proofs. Formal verification for complex programs is a thing. At least if you consider flight control systems or even microkernels complex. There is a very good reason C is so widespread in avionics and it's not only because 'embedded'. Ada is being replaced by C++ in some shops (by C in others) partly because of how hard it is to hire competent developers.

>>60563416
Well you're trying to generalize a resource management problem using templates so it's actually pretty compact, in C it would take at least 2000 lines of code to reach this level of generality and you would lose track of the core idea.
I think it's great that C++ offers so many high level tools while remaining close to the metal, you could do your brainlet C stuff if you wanted to as well.

>>60567182

I don't have a high opinion of avionics systems - especially those run by the military.

Something about an NMAP scan crashing their software and causing a kneejerk panic against NMAP scans in general.

>>60567204
What's wrong with malloc/free?

>>60563701
> RAII also works for C
What are u doing on this board

>>60567236

The hundreds of thousands of bugs that emerged over the years because C programmers think they're smart enough to free by hand, but aren't.

>>60567275
>i'm too stupid to free my own memory
>everyone else is as stupid as me

classic autism

File: 1493921509736.jpg (403KB, 1426x950px) Image search: [Google]

403KB, 1426x950px

>>60567119
The term "software industry" encourages people to imagine that software is always created by a sort of factory and then delivered to "consumers." The free software community shows this is not the case. Software businesses exist, and various businesses develop free and/or nonfree software, but those that develop free software are not run like factories.

The term "industry" is being applied as propaganda by advocates of software patents. They call software development "industry" and then try to argue that this means it should be subject to patent monopolies. The European Parliament, rejecting software patents in 2003, voted to define "industry" as "automated production of material goods."

>>60567072
>>60566598
>>60566522
Wtf I hate QT now. What should I use?

>>60567297

I'm not responsible for the daily security bugs caused by cocky Ctards who freed one time more than they were supposed to.
Why are you trying to make this about me? Are you maybe projecting?

>>60567217
Quality of military software depends on where you are from. I'm yuropoor so that leaves only very small amount of options. We've never had a bug in C code of production-ready systems. Most people wouldn't like the subset of C we use, but most people wouldn't be able to prove their quicksort implementation is correct anyway. The US military tech i've had the opportunuty to play with was pretty alright (all Ada), but afaik wasn't as thoroughly verified.

>>60567236
There is nothing wrong with malloc/free if it is the appropriate solution for your problem, even C++ resource management in the end has to somehow allocate and deallocate storage by using the concepts behind malloc. But by creating abstractions we can hide a lot of boilerplate code, that's why we can summarize 2000 lines of C in a couple lines of C++. These abstractions also give us an intuitive way to reason about programs: constructors and destructors define a state where data has to be valid, templates make it possible to generalize behaviour and structure, smart pointers delimit the lifetime of data to its scope, etc.
And using these abstractions is a choice, you could always go back to C-style procedural programs.

>>60567344
Nothing? This is the "Modern C++" experience, enterprise galactus absorbing planets aka Java 2.0 - this time native!

Roll your own UI lib, write your own networking lib. Enjoy being in complete control like you should, instead of getting an error message filled with template garbage and length of the great wall


Always remember Anon, if its easy to write, it's probably slow as hell.

>>60567344
HTML5, obviously.

>>60567344
Vulkan + glfw, it's not that hard

>>60567297
C++ also improves on the semantics of memory ownership.

void Foo(struct Bar* b); // does Foo own b now?, do I need to check for null ?

vs.

void Foo(Bar& b); // non-owning, non-nullable reference to b
void Foo(std::unique_ptr<Bar> b); // pass ownership of b to Foo
void Foo(std::shared_ptr<Bar> b) // share ownership with Foo
void Foo(Bar* b); // non-owning, nullable reference to b

Still not perfect, but the semantics express intent much better.

>>60567755

you could just use hungarian I guess

>>60567799
Fuck hungarian, why use types without using types

>>60567755
You are legit retarded. Stick to your meme C++ language and never write C because just the thought that you and I share the same language makes my neurons commit suicide.

Thanks.

>>60567861

He's right though.
As soon as you call any code you haven't personally written in C, you better cross your fingers and hope it's fully documented or you'll be up to your neck in memory management issues in no time.

>>60567887
Can you really trust the pajeet that he wont cast the unique_ptr into a void * ?

:^)

>>60565721
>afraid to try replacement languages
why would you replace it with something slower?

>>60567844

Because you're using a shitty language with an inexpressive type system that makes it difficult to have any semantic value, and you need a band aid to cover it.

>using int *p = malloc instead of int p[1]

>>60567936

int *foo()
{
    int bar[1];
    return bar;
}

>>60567861
I program embedded C for work, dear anon. Even though I like it, I wish it had better semantics and stronger type checking.

>>60563871
RAII guarantees that the memory will be destructed when its scope is destructed. doing this by hand is not RAII because there is no guarantee that you will call free correctly, i.e. exactly once on each exact path from the scope. you can say "well i always call free correctly," but that's not RAII, because a key feature of RAII is to do that correctly on your behalf

>>60565956
with python, you basically just s/^/^self\./g
Variables can also randomly change format, so you need to make a few extra checks when using them.
It is not especially hard.

>>60563416
1. that's not how C++ code looks in practice.
2. C is a pile of shit that can't even have true generic containers

>>60568031
>2. C is a pile of shit that can't even have true generic containers
M4 MACROS MAN

STRING INTERPOLATION IS JUST LIKE TEMPLATES

>>60568031

>true generic containers
inb4 "put a void* in your struct it's totally the same"

>>60567969
Well, you can forget to put a free, but doing that is just being obtuse.

>>60567925
Why would it be better without it? At least errors are caught at compile time this way.
Also why is it inexpressive and why doesnt it have semantic value? A type by itself is empty of meaning, it's just a grouping of data, only the programmer decides what it means

>>60568124

You misunderstand.
Using Hungarian is an indication that you wish you had strong and flexible typing, but are using a language that provides neither.

>>60568124
types aren't just data, they have concepts. you can add two integers, or two floats, or two strings; but even though all three operators are different, integers and float addition implements an algebraic concept and strings do not. this matters if you write a generic mathematical algorithm

it is nice when the language allows you to express these things and test them e.g. at compile time. C++ isn't there yet but it's getting there, in many languages you cannot do this at all

>>60568181
So like haskell with typeclasses?
Also for simple operators like '+', you can just overload them in C++ for specific behaviour

>>60563383
It will become Java then.

>>60568181
I think he means that, ultimately, the language allows you only to create the syntactic (types and possible operations), but, ultimately, it is the programmer who brings the semantic. Their relation is not unlike the one shown in Frege's triangle.

As you say, I agree that it is indeed good when language allows to bring those two parts as close as possible, but they can never fully coincide.

[spoilers_on_g]
But who actually cares about semiotics of programming languages[/spoilers_on_g]

>>60568874
java's syntax is very simple

>>60569208

to a fault imo

>>60563416
>Current state of C++

C++ IS A GARBAGE

File: dlang logo.png (22KB, 1024x1024px) Image search: [Google]

22KB, 1024x1024px

>>60563312
>Package manager and build system
>Seamless import system
>Fast compiler
>Vastly improved syntax, including...
>Built-in basic data structures (arrays, maps)
>Simple array/string slicing
>Better control structures (foreach)
>Simpler, extended, vastly more powerful templating system
>Compile-time evaluation
>Actually readable (just look at a library's source to see what it does, documentation not necessary)
>Extensive and convenient standard library for Python-like versatility out of the box

C++ can fuck off. D is like a scripting language that compiles for native performance. It's the holy grail endpoint of the C family of languages.

>>60569511

how's library support?

>>60569522
You have to be more specific. The standard library has most of what people need, vibe-d is a pretty full-featured web server. If you're talking about scientific computing, sure, it's not quite there yet, though there's work being done on the Mir library. And there's a bunch of experimentation going on with game engines and real-time stuff.

>>60569522
Oh, and there's good GUI support. DlangUI is fully-featured, and the GTK bindings are 100% excellent according to the author of Tilix/Terminix, which was built in D.

>>60569610

I'll have to check it out sometime.

>>60569511
This, D is amazing

>>60564465

Only ISO C++ and GNU C++ should be used. Of these, the latest standard (C++17) should be preferred.

Ruby master race.

File: 1485212514097.jpg (25KB, 460x651px) Image search: [Google]

25KB, 460x651px

C++ is ugly, unreadable, and inherently worse than C

>>60563701
>lets hide the call to free using another function, that will surely make it safe!!
lmao

File: brit2.jpg (13KB, 631x394px) Image search: [Google]

13KB, 631x394px

>>60566888
>safety critical
>c

>c
>safe

>>60570299
That guy doesn't understand what RAII is. It's much closer to Go's "defer".

>>60563699
Enjoy spending 1000$+ on a PC just to run a browser on your shity high-level OS

>>60564382
Qt apps look like shit on everything but Windows and Plasma

>>60567072
I think you mixed up two coincidental things.

Your main complaint is that Qt tries to do everything at once AND it coincidentally has well designed interfaces which allows you to easily write and fully use it, but with bad perfomance.

Your "easy to write implies large overhead" is just false. You should always strive to make external APIs as simple and easy to use as possible. Overhead usually implies that your module is trying to do too much at once.

>>60570240
Team I'm in uses C++14 in its full extent, yet we write normal, easily readable code even if it concerns really low level stuff like custom filesystems on external devices. I guess you had some bad experience with C++ programs, but you can write awful unreadable code in any language, really.

D master race

>>60570744
D is to programming languages what Void is to linux distros, a cute meme that nobody uses.

>>60570778
Void is trash, D isn't

File: hqdefault (1).jpg (19KB, 480x360px) Image search: [Google]

19KB, 480x360px

>>60570744
so you like D?

>>60570830
https://github.com/search?l=d&q=stars%3A%3E1&s=forks&type=Repositories

Literally the only active D project is the D standard library.
Stop spamming this useless dead language.

File: Unofficial_JavaScript_logo_2.svg.png (53KB, 2000x2000px) Image search: [Google]

53KB, 2000x2000px

>Still the best language
>And it keeps getting better
Why do brainlets hate on JS? Do they think we're still all writing Ecmascript3?

File: 1367818709357.jpg (25KB, 500x375px) Image search: [Google]

25KB, 500x375px

>>60563312
I've spent the past few years programming just in lisp and recently started learning C++. It's much better than I expected. I thought I was going to have to deal with pointer arithmetic and forced OOP programming. Since there's containers, for_each and range loops, etc. I haven't had to deal with any of that. Templates aren't that bad, but that could be because they're a duck-typed and mostly-functional metaprogramming system and once you get that it's not hard to grasp coming from lisp. It's still kind of inconvenient to do functional programming in a language without return-by-default, but it's been good so far. I haven't to deal with any of things the majority of detractors have had to say about C++.

Overall I'd give it a 8/10 to be honest. It's a statically typed language that emphasizes programmer flexibility and runtime speed, and I like that.

>>60570855
>I don't know what I'm talking about but I'll pretend I do anyway

>>60570943
Oh yeah, might as well mention: the biggest con is package discoverability. There's no default package manager so I end up having to google and hope for the best. Pretty annoying, but it looks like there's a couple of competing solutions coming out now (Buckaroo/Conan) so I guess we'll see.

>>60570855
D can call C++ natively, you idiot

>>60570778
Void is for autistic hobos,
D is for patrician programming geniuses.

>>60563312
>int x=1, y=2;
>double failboat[2]={x,y};
>compile error

>>60570854
>White man
No wonder she's not impressed

>>60571083
>>60571072
>>60570830
>>60570778
>>60570744
D has a cute grill

https://www.youtube.com/watch?v=endKC3fDxqs

>>60571135
is this supposed to be a problem?

>>60571318
yes, because it was legal prior to C++0x

>>60564994
case in point
https://dlang.org/blog/2017/05/24/faster-command-line-tools-in-d/

>>60563312
C++ is like racial/cultural diversity.

You throw a whole bunch of topics and features in a single pot, mix it up, give it to a community of politically driven idiots, and what you get is a piece of shit that does a lot of things okay but needs to be heavily augmented if you really care about performance and many other issues.

>>60571161
kek

File: do_you_have_a_single_fact_to_back_that_up.png (21KB, 207x239px) Image search: [Google]

21KB, 207x239px

>>60571475
>implying you need to "heavily augment" (whatever that means) c++ to get decent performance

>>60569511
Yeah you wish..
> Garbage collector
Goodbye to a big part of the C/C++ audience, if you cannot finetune resource management it's not suitable for deterministic performance. RAII is impossible because disabling the GC removes a great part of core functionality.
> No stack allocation of class instances
This also severely limits performance as you cannot make a quick, small object on the stack.
In general D gives you less control of your program and there is not much that you cannot already do in C++, especially since C++11.
In my opinion D doesn't really have a purpose to exist, it's too verbose to compete with scripting languages like python and it's too unpredictable and unflexible to replace C++. It's just another flavour of the month language really

>>60569511
>Language is a letter
When will they stop doing this? Nobody will take you seriously

>>60563312
>Still writing C++98
>Not C89
kys

>>60563312
Rust has had for ages what C++17 is just getting, and has what C++ will probably not get until C++22.

>>60571805
>No stack allocation of class instances

scope c = new Class()

>>60571805
>Garbage collector
Finetuning allocation is a disproportionate labor sink for most project cases. For the specific cases, disabling the GC and using an allocator like the std library one is fine.
>Stack allocation
There's std.typecons.scoped!(). >>60571960 will be deprecated.

>>60572035
Almost all libraries are assuming the GC is enabled and there is the risk of memory leaks

>>60571897
shame rust stable is unusable, rust nightly could be a good featureful language someday

bumping for more C++ opinions and discussion

>>60574249
You will mostly find unemployed NEETs defending their language-fu.

>>60563312
>2017
>C++ 17
Ouch they use the silly Microsoft like naming convention!
It's already a huge problem!

File: YvgTM4Uw.jpg (84KB, 496x496px) Image search: [Google]

84KB, 496x496px

>>60566973
>language≠community
That's what Rusticians believe. In reality, the community shapes the language. Fuck off already.

>>60575563
>FORTRAN 66
>FORTRAN 77

>>60563532
That's a problem with bad programmers. As long as what you're doing is well structured, this should never happen.

>>60563511
Unique pointers are actually one of the best additions in C++11. You just make_unique whatever you want and not care about the memory, it WILL get released when it's time.

Now, shared_ptr are useless to me and in their case I'd rather control the memory manually.

>>60576023
Smart pointers are an elegant solution, it doesn't add much overhead and it ensures there is no memory problems.
Obviously, you can still intentionally fuck yourself, but it is a lot harder.

>>60563383
>I don't want to learn the new stuff so i just shit on it

>>60572211
It's a new language, so no surprises there.

>>60576023
>true Scotsman

>>60564263
>one of you faggots told me that the random number generators are garbage
I watched a talk by the C++11 (or 14?) random guy and he said the old random was shit for anything where random was important, and most people implemented it badly on top of that.

The new stuff seems academic-tier if implemented properly which isn't hard but does require like 4 lines of code to setup an "engine".

>>60568054
>comparing the streaming pile of shit of macros to templates

>>60563416
C++ template metaprogramming is a black art even withing C++ programmer circles.

All the new features in C++ actually both makes already achievable stuff in TMP easier (constexpr functions are awesome) and adds more helper stuff for more complex stuff.

DESU template metaprogramming is almost detached from C++, at that point you are writing in a different language that runs compile time.

>>60567072
Somehow I doubt that if a speed gauge is updating at 10Hz that it's because they use Qt. There's just no way Qt is burning 300M cpu cycles on drawing a gauge.

>>60564263
>still no functional package manager
>still developing on Windows
apt is a good package manager for c++

File: 1494277841075.jpg (103KB, 600x800px) Image search: [Google]

103KB, 600x800px

>>60576328
>apt is a good package manager for c++

>>60563312
Yeah man, I'm with you.

At work, I'm using C++17 features already in some of the subsystems I'm working on. if constexpr, automatic return type deduction, destructuring, sfinae for pattern match, I'm pretty much programming as if I already have the syntax for full concepts semantics... feels like I'm almost in Lisp or Haskell. Almost.

Man, just wait, just fucking wait until the new compile-time reflection TS gets implemented, combined with tuples and destructuring, no more fucking serialization boilerplate it'll be even better than how the meme languages have been doing it all these years.

Meanwhile, the junior devs and some of the old dogs at work are still programming like it's 1998 and can't even wrap their fucking heads around move semantics. Had to rewrite some pleb's shitty code the other week, and did it in one-third the lines, just because he didn't know the true power of the language. Seems like they'd rather be programming in C# or Javascript.

Oh well, at least that means more money for me.

>>60576078
std::shared_ptr is very useful once you get into multi-threading and passing objects to asynchronous tasks. Releasing the references on the shared object is thread-safe, ensuring only one thread will ever destruct the object.

>>60571897
Rust this. Rust that. One thing Rust won't achieve is International Standardization. If they every get their shit sorted out and become ready for standardization, it'll be 2026 and C++ will have everything that Rust does and more. Including ARC (there's already third-party ARC implementations for C++).

>>60576496
Sure, just like C++ rapidly stepped forward in adding modules.

>>60564636
Like?

>>60576423

>Meanwhile, the junior devs and some of the old dogs at work are still programming like it's 1998 and can't even wrap their fucking heads around move semantics.
Every time I try to see it explained, it's unclear.

struct DumbString
{
    int size;
    char* data;
    ...
    DumbString(const DumbString& d)
    {
        size = d.size;
        data = new char[size];
        memcpy(data, d.data, char*size);
    }
    DumbString(DumbString&& d)
    {
        size = d.size;
        data = d.data;
    }
};

Is this a proper usage of move semantics? Just stealing heap pointers? Please ignore the garbage style.

>>60576596
You also have to make sure "d" does not own the data anymore. So you also have to null d's size and data.

>>60576293
No, C++ templates are seriously not that hard and they are absolutely not considered black magic by intermediate C++ programmers.
It could be challenging to wrap your head around concepts like template templates, SFINAE, recursion or policies but the average programmer will only use templates to generalize some functions

>>60563312
Isn't HolyC even better?

>>60576308
QString, QInt, QPolygon :^)

>>60576860
>QString
MyString is better?

>qint8...64
Just typedefs.

>>60576868
deff not MyString, just by your naming convention i can figure out you're a retard.

>not just using ada

>>60577002
"MyString" was poking fun at homemade implementations of strings that were made simply to not use the standard one, if you couldn't tell.

QString has its own use -- localisation and unicode. You wouldn't want to use it in a console application, but any serious GUI work requires those features.

>>60577027
There are no strings anon.

If you need to absorb entire framework +class to display a number there is something wrong with you, and anyone who buys a car made in past 5 years is a retard, especially Tesla.

Go to your local car dealer and ask to see Audi R8, enjoy the smooth door opening animations ;)

File: reiuji utsuho (touhou) drawn by e.o. - Danbooru - 0ca94ccdff8ec365784d4ed3a3db534f.jpg (255KB, 1000x1398px) Image search: [Google]

255KB, 1000x1398px

>>60563416
you are entirely correct. c++ is a pile of garbage, the garbage being all the new things that add quality of life.

ideally quality of like should be THE standard iteslf, alas qol conflicts with rare instances where we do need performance.

you just shouldn't use c++. if it's necessary to write high-performance code use someone who actually enjoys c++ spaghetti and can keep up with standards. that's why on every c++ interview they ask about stl, because you are assumed to know how to program and expected to know the spaghetti. unlike say python programmer about whom we assume nothing but expect him to know how to program.

and no, c code isn't more comprehensible than c++ code in any real project.

>>60567755
Actually your last example may be owning, you don't really know unless there is documentation or source.

>>60577432
Well you shouldn't let a bare pointer own data, always use a smart pointer so if the last case is owning then it's bad software design

File: 1462308840929.jpg (8KB, 265x265px) Image search: [Google]

8KB, 265x265px

>>60564626
>get files in directory
>20 minutes
HAHAHAHAHAHA
OOOOOOOOOOH
WOOOOOOOOOW

>>60563701
>RAII in C
typical C faggot

Why is regex support so shit in c++?

>>60577011
from garbagebin with trash
begin suffering
cry
end suffering

>>60576023
90% of people who can program are bad at it.

of the remaining 10%, 1% have enough time and resources to write the beautiful code they would like. The rest are time constrained and budget constrained.

Low overhead high impact features like that are exactly what you want in a language. If you don't like it you are a) free to not use it and b) free to use a more pure language (except you can't because none of your pure languages are widespread anywhere that earns you money)

>>60577432

Isn't C++17 introducing observer_ptr to eliminate this ambiguity?

>>60563312
Isn't C++ slow compared to Java?

>>60578385

No.

>>60578385
Yes, at times Java is actually faster than C++

File: 15489289444.jpg (16KB, 600x600px) Image search: [Google]

16KB, 600x600px

>>60578385
>>60578573

>>60563699
I'm content with prototyping in Python and finalizing the program in C++. Fuck all you people who think one language solves every problem. Its a tool /b/ro, you gotta be flexible! Once you have written a large program in raw assembly, anything seems like magic.

>>60578624
I'm content with prototyping on paper and writing the algo in clojure.

>>60564263
What? C++ has modules. I just used them the other day in a project.

The problem with C++ having so many ways of doing things is everyone has their preferred "subset" of C++ that they use and it's completely different from everyone elses, even within the same teams.

Basically, whatever features they were taught, they use those no matter how shitty they are.

>>60580642
That's also a strength though, different problems having different approaches that may work better than others. Most projects only use some subset of C++, but there's always niche cases where everything's always available.

It's weird that a language has its own sub-specializations but it's not that difficult to move between them compared to jumping to some other language or design system.

Guy's you're going to love module syntax when it comes out:

At the November 2016 C++ standards meeting, the C++ standards committee changed the syntax for exporting a module (see Module Issue #1) from

export module Bank;

to

export import Bank;

>>60581255
If this is true, it is disgusting.

But it is fine, as we are used to the cpp commitee producing steaming piles of shit since at least 2011 minus 10 years.

I wish I was able to used all those nice "new" features such as smart pointers, constexpr, auto func(int asd) -> return, and tons of others.
I mean, look at the stl. Look at the way you would make a unique_ptr of a std::map containing an std::string and another pointer. Now tell me, is this readable?

It is not. Its not even black magic, nor Mayan alphabet.

Its __unreadable__. Period.

auto foo = make_unique<map<string, bar*> >();

What's the matter with this?

>>60581543
What the fuck does this even do?

Can someone explain nested template types?

>>60581577

foo is a unique pointer to a map of strings to bar*s. Isn't that clear?

>>60581591
strings to bars?
So every element of the string is a pointer to bar *?

>>60581614

It looks like you don't understand what a hashmap is.

(*foo)["some string"] // returns some bar*

>>60581639
No, I really don't understand the concept of using a string as an array index.

What allows this to work?
How is the string interpreted?
I've written hashmaps in C, the key was always an integer though.

>>60581577
Fucking brainlet, but alright.
First off he is using the std namespace which is a big red flag.
He is making a map from string to a bar pointer and he allocates this map on the heap with the default constructor (make_unique()).
The lifetime of the map instance is taken care of by the smart pointer so it will be automatically destroyed and deallocated when the scope ends.
Nested templates are not that hard, it's like nesting functions but with types (or constants).

>>60581673

there is literally nothing wrong with

using namespace std;

in a .cpp
prove me wrong

>>60581349
You can make templated code readable by providing scoped types:

using mapper = std::unordered_map<std::string, Bar*>;
auto m = std::make_unique<mapper>();

>>60581692
>.cpp
>not .cc
Plebs, all of you

>>60581255
It's stupid but I understand their decision, they are reusing the import keyword to reduce the amount of restricted keywords. I guess a lot of people still use 'module' as variable name.

File: Yakumo.Ran.full.1876155.jpg (385KB, 725x1024px) Image search: [Google]

385KB, 725x1024px

>>60563312
It doesn't get any better than C++11 desu

>>60581749
Are you kidding? C++17 will bring nested namespaces, shit's gonna be fire fa~m

>>60581692
Well what if you're including a function with the same name?

>>60563655
You should still use the std libraries though. Smart pointers aren't that bad.

>>60581795

You shouldn't be importing such functions to the global namespace, even if you're using nothing from std.

>>60581814
>implying anyone on /g/ is smart enough to use raw pointers without fucking up

>>60581661
Can someone answer this?

How the fuck do hashmaps let you use a string instead of an hashed integer key as an array index?

>>60581897
Are you stupid?

>>60581897

You can hash and compare strings. What more do you need?

File: ran_alert.png (64KB, 161x285px) Image search: [Google]

64KB, 161x285px

>>60581770
>fa~m
How do I know I can trust your opinions? Post your programming socks so that I know you're a real programmer.

>>60581908
>>60581916
How are you using a string as an array index without an intermediate function hashing the string in C++?

File: 1493071583753.jpg (2MB, 3264x2448px) Image search: [Google]

2MB, 3264x2448px

>>60581927
o-okay...

File: sakuya_nosebleed.jpg (389KB, 740x928px) Image search: [Google]

389KB, 740x928px

>>60581944
Intriguing. Tell me more about C++17.

>>60581928
>how does a hashmap work
>hashmap
>hash
you're a real dummy

>>60581928

C++ allows you to overload operator[] to take and return any type you like.

struct map_char_to_int
{
    int operator[](char c)
    {
        // some hashing bollocks
        return some_int
    }
};

If you prefer, you could use .at() or .find().

>>60581977
like r-right here? in front of all the other a-anons?

File: pair_programming.png (318KB, 593x535px) Image search: [Google]

318KB, 593x535px

>>60582012

>>60582002
This is thoroughly disgusting.
Does C++ actively encourage this level of language ricing?
what's wrong with using a function to return all N entries found within the particular cell pointed to by the key?
How does this even deal with hash collisions?

why /g/ don't like popular things.

>>60582064
The language only specifies that you can overload operators. The details on hashmaps (which use this language feature) can be found in the STL reference: http://en.cppreference.com/w/cpp/container/unordered_map

>>60582064

>what's wrong with using a function to return all N entries found within the particular cell pointed to by the key?
You mean you do this every time you want to pull something from your hashmap?

cell_id = hash(key);
while (key != map_storage[cell_id].key)
    cell_id++;
val = map_storage[cell_id].val;

>How does this even deal with hash collisions?
You can put any logic in your operator[], it's just sugar for a function call whenever you overload it.
The shit I just typed out can be in your operator[] definition.

>>60581715
At this point I feel like C++ has become a black hole with answers to every problem. And I mainly work with cpp

>>60563482
Basically, it turns

make_resource(creator, destructor, 1, 2, 3, ...);

into

std::unique_ptr(creator(1, 2, 3, ...), destructor);

>>60582478
To add to that:

Basically it's a wrapper for C stuff, like:

make_resource(fopen, fclose, "my_file", "r");

Would handle calling fopen paired with fclose for you by returning a std::unique_ptr<FILE*>.

>>60582508

I like it.

>>60581834
>implying anyone human can use raw pointers without fucking up enough to allow for exploits
If the Kernel guys can't manage it, it's likely nobody can.

>>60581724
>not .definition

>>60581897
Literally everything ever on a pc is an integer "deep down"

>>60583521

What about floats?

>>60576589
Any scripting language for things that don't need performance.

Call C (super easy) when you need it

>>60567437
Nuklear

Wrap it around any shit you want

>>60583560
They don't use some magic quantum memory. Interpretation of integers can be different, but that doesn't matter.

>>60583978

In that case, they're not really integers either. They're all just bitfields.

>>60584056
Or brainlets.

File: 4thEnglish.jpg (50KB, 591x722px) Image search: [Google]

50KB, 591x722px

Anyone know any good books for learning C++?

>>60584504
http://www.icce.rug.nl/documents/cplusplus/

if you know C

pdf on github

>>60584504
Good first books, in order of programmer experience:
>Programming, Principles and Practice
>A Tour of C++
>C++ Primer (Not C++ Primer Plus)
>Accelerated C++ (the above, but faster)

how do i into C++ from a C# basis?

>>60563416

Your attempt at trolling fails horribly. You can leave out the trailing return type in C++14. You can leave out the template parameters in C++17 (they will be deduced automatically). Your piece of code becomes:

template <typename Creator, typename Destructor, typename ...Arguments>
auto make_resource(Creator c, Destructor d, Arguments&&... args)
{
    return std::unique_ptr(c(std::forward<Arguments>(args)...), d);
}

>>60584973

learn C first.

>>60584973
>>60585014

> Should I learn C before I learn C++?

Don’t bother.

The common subset of C and C++ is easier to learn than C. There will be less type errors to catch manually (the C++ type system is stricter and more expressive), fewer tricks to learn (C++ allows you to express more things without circumlocution), and better libraries available. The best initial subset of C++ to learn is not “all of C”.

The go-to book to learn C++ is Stroustrup’s A Tour of C++. Run, don’t walk, to read and recommend this core overview.

If you want additional options: See Stroustrup’s Learning Standard C++ as a New Language for a discussion of the choice of C++ constructs, techniques, and libraries for early learning. For an example of books that takes that approach systematically, see Stroustrup’s Programming: Principles and Practice using C++ and Koenig & Moo’ Accelerated C++ from Addison Wesley’s C++ In Depth series.

If your ultimate goal is to learn C++ and you don’t already know C, reading books or taking courses in C will not only waste your time, but it will teach you a bunch of things that you’ll explicitly have to un-learn when you finally get back on track and learn C++ (e.g., malloc(), printf(), unnecessary use of switch statements, error-code exception handling, unnecessary use of #define macros, etc.).

If you want to learn C++, learn C++. Taking time out to learn C will waste your time and confuse you.

https://isocpp.org/wiki/faq/how-to-learn-cpp#learning-c-not-a-prereq

>>60584973
Read any of the books about it, maybe one of the shorter intro books and then skip directly into a Best Practices/intermediate book. So you can start writing good C++ instead of C#-as-C++.

>>60585042
>>60584901
Are all of those books available in moderner C++11/14/17 though? Pre-0x C++ is completely different in mentality.

>>60585072

They should all be aware of C++11, which is where the major break in style comes.

>>60585072
C++ Annotations is continually updated. It's even updated to C++17 right now.

>>60584973
>>60585072

The way I quickly got a foothold into C++ as a C and Java guy was by going on this site : learncpp.com

It's updated for C++11/14. has short explanations then a code example for each topic.

Then I watched a bunch of cppcon videos on youtube.

if c++ is so bad why does every game engine use it?

>>60585179
It's not bad but it doesn't prevent bad programmers from shooting themselves in the foot. Newer standards are better about that.

>>60585179
namespaces is a big help

>>60585179

Most people who hate it are too dumb to understand it. Which isn't surprising, the language is an absolute clusterfuck.
That said, plenty of people are informed and hate it. They're just not the majority of haters.

>>60585039
>Koenig & Moo
>Moo

>>60585039
>What should I tell people who complain that the word “FAQ” is misleading, that it emphasizes the questions rather than the answers, and that we should all start using a different acronym?

>Tell them to grow up.

kek

>>60570346
Found the nigger who didn't into formal methods in uni. Guess what runs your airplanes? Hint: it's c and ada

does c++ have auto garbage collecting??

>>60585759

no, use RAII

>>60585759
>>60585804
Sort of with the "new" smart pointers, but RAII is best.

>>60565252
>try to unscrew something
>smash it with a hammer
Ok m8

>>60585759
No, but you don't really need to manage memory manually either unless you choose or the environmental constraints demand it.

>>60585980

This.
If you're writing new or malloc, you're doing something wrong.

>>60585759
it does if you want it too.
this for example, http://www.hboehm.info/gc/
there are probably a bunch of other ways to do it too.

>>60587590
>source code stored in a naked source directory
This is the worst part of C++. Half of the things you find are just scattered around the web and have zero source control. You just download a tar.gz and hope for the best.

>>60587885
People just use linux package managers to handle C/C++ dependencies because it made sense at the time of it's inception.

I don't understand why every fucking language these days has it's own build system, package manager, dependency resolver, online package repository, etc.

>>60587934
Because package managers and repositories solve real problems.

>>60587934
well it makes sense if they're good, static linking is not harmful though.