ITT: explain what you have against Rust WITHOUT (and this is

>>56043132
are these not enough reasons?

>>56043146
no

corroded language

Too young.

>>56043132
I love the ideas behind it and the feature set but it's too ML-ish for my taste

It's not Haskell.

I don't have anything particularly against it, but I don't really enjoy writing in it for some of the same reasons I don't really like Java.

The language enforces tons of rules and restrictions, which is understandable because it's doing so to guarantee safety, but it's annoying. There's nothing unique about the language other than that set of restrictions meant to stop you from shooting yourself in the foot, and it annoyingly gets in the way when I know what I intend to do with something better than the compiler does.

You can get the same effect by slapping a whole bunch of restrictions onto a C compiler.

>>56043247
unsafe {

}

No community compared to other languages.

>>56043132
It doesn't yet have a "secure" flag for generating code that zeroises a variable on destruction but also applies that to all intermediates or temporary copies. Partly because LLVM doesn't have that capability yet.

So I have to use assembly to handle things which won't play sensitive data onto the stack (nothing I can do about CPU-internal register renaming).

It's no worse than any other high level language in that regard however.

Syntax is a bit meh.

I'll use it.

>>56043277
Then why bother with rust at all?

You're like those people who respond to complaints about Go's lack of generics with a recommendation to use interface{} everywhere and basically bypass the type system entirely. Why are you using a language if you have to work around its major design choices?

Psh, everyone on /g/ knows that the only programming languages acceptable to learn are C, Lisp, Algol, and Haskell. Basically, if your programming language doesn't require a PhD in math, soldering cpu buses, and was made after 1975 it's a "meme language" and not worth your time.

Don't tell me you've ACTUALLY been taking advice from /g/ have you anon? They're memeing you, that's why they hate rust. For the maymays

File: 2.jpg (15KB, 261x213px) Image search: [Google]

15KB, 261x213px

>>56043132
nothing really. i just dont understand why everyone has to make their own language now.

memory management is kek

just use GCs, it's not 1970 anymore

>>56043479
NOOO WHAT ARE YOU SAYING!

Gcs are worst evil in this world you might as well use java! And performance, man, even highest end computers cant handle average programs with gc, dont get me started on vms and huge ram overhead.

>>56043537
Use your gc language on an embedded device or real time system like a missile guidance system

>>56043664
Those areas use C or C++ nobody uses rust for that. If you need embedded you use C otherwise might as well make it easy for yourself and use language with GC unless you do AAA games then there is C++

Why not just use C instead?

>>56043697
>Those areas use C or C++ nobody uses rust for that.
Yet.

Rust WILL be used when it's ready, as it's way safer than C/C++ and just as fast.

>>56043132
All the libraries are written for C/C++ and I don't want my code to pay FFI overhead.

>>56043468

Langauges are just Tools..

20 years ago everybody builded Frameworks, but nowadays people are often tired of the shit you find in the "industry languages".

Also it's a kind of Evolution, it's not like new languages do things completely different. There's a lot of repition especially at the syntax.

The differneces are more about the language features (what you can do and - even more important - what you CAN'T do).

If you are a Develloper with some years of Java and C experience you will pick up Go in a few days to weeks..

That's also the reason why modern langauges often complete so much (i.e. Rust and Swift): they are designd for very specific purposes, not "one size fits all"-langauges like Java or C++.

>>56043664
The missile knows where it is at all times. It knows this because it knows where it isn't. By subtracting where it is from where it isn't, or where it isn't from where it is (whichever is greater), it obtains a difference, or deviation. The guidance subsystem uses deviations to generate corrective commands to drive the missile from a position where it is to a position where it isn't, and arriving at a position where it wasn't, it now is. Consequently, the position where it is, is now the position that it wasn't, and it follows that the position that it was, is now the position that it isn't.
In the event that the position that it is in is not the position that it wasn't, the system has acquired a variation, the variation being the difference between where the missile is, and where it wasn't. If variation is considered to be a significant factor, it too may be corrected by the GEA. However, the missile must also know where it was.

The missile guidance computer scenario works as follows. Because a variation has modified some of the information the missile has obtained, it is not sure just where it is. However, it is sure where it isn't, within reason, and it knows where it was. It now subtracts where it should be from where it wasn't, or vice-versa, and by differentiating this from the algebraic sum of where it shouldn't be, and where it was, it is able to obtain the deviation and its variation, which is called error.

>>56043719
There's no overhead from Rust/C FFI.

>>56043718
>yet
okay maybe in future but till then there is no reason is there?

just RAII shit out of everything in C++ and its pretty safe or use c++14

>>56043731

I have seen this pasta before..

>>56043741
>just RAII shit out of everything in C++ and its pretty safe
C++fags, everyone.

>>56043762
Well it adresses unsafety issue somewhat doesnt it?
Besides for low level embedded stuff u need to use unsafe in rust anyhow so no point in using it, might as well use C if you hate C++

>>56043132
No easily usable cross-platform GUI library
QtQuick does not count

>>56043733
You're not fooling anyone, Rust semantics are different from C, this isn't like C++ where exactly the same code will compile.

>>56043733
even if there is no function call overhead there is data type transformation overhead so you in fact do have overhead.

>>56043468
Because making JavaScript framework was cool two years ago

Having an SJW community behind it is enough reason to not use it.

>make an awesome framework/library in rust
>"OH WOW ANON'S SUCCESS IS DUE TO THE AMAZING LANGUAGE MADE BY SJWs! SEE GUISE? SJWs ARE THE COOL KIDS IN TOWN :^)"
nope, fuck yourselves

>>56043132
>and go
Go is just fine. I don't see why people compare them.

>>56044084
Yep i agree Go is great at what it does and actually has a purpose and great standard library. Sane syntax, fast compile and its speedy enough. While rust barely has any libraries and if you want something proper u have to compile C as well...

>Boycott X because bigots use it!
>Boycott Y because SJWs use it!
You people are no different from SJWs.

http://aturon.github.io/blog/2016/08/11/futures/

What other languages have zero-cost futures?

>>56044168
Seriously. I have never seen a group of people get more triggered than those that are afraid of "SJWs"

>>56044176
c++ ??

>>56044105

But Go is boring. It's like the overlap of the most common langauges. Sure, it has an easy syntax, but it has no powerfull features. It's basically "C for Pajeets". Mark my words, Google will attract even more low cost devellopers with Go.

Rust on the other hand has some real interesting features and concepts. People really put some effort in to think about a way to make C a better langauge.

>>56044176
Futures are a meme.

>>56044223
Really? Find me a C++ concurrency library that does no allocations.

>>56044261
great argument bro

>>56043334
>get a runtime error which is clearly linked to memory unsafety
>debug 10,000 lines of code in unsafe{ } instead of the 100,000 lines of code youur whole project contains
and 10% unsafe ist something I would only very rarely expect any project to have, unless it's something like computer graphics where you can't easily avoid dangling pointers without performance loss

>>56044256
Okay maybe boring and easy but there is nothing wrong with that. Let me give u an example i use dwm and i wanted to have random wallpaper every hour so as good guy i wrote it in C and added random image downloading as well with curl. It took me quite a bit to fix all quirks and parsing returned html was a pain so i basically used some dirty hack. I decided to check out GO couple of days ago and rewrote that program within 20 minutes of first touching the language and its shorter with no hacks and easily extendible. Compiled binary is bit too large but there is no cpu usage at all which is all that matters.
I think of GO as compiled python with sane syntax, i hate python so much.... Its good general purpose language.

While rust its like not as bloated little C++ brother with not as weird but still weird syntax and significant learning curve but not as high. It provides memory safety but lacks libraries. I really cant see it being used anywhere other than some nieche.

>>56044381
or no reason why not just use c++14 for safety if C is unsafe

>>56044381

I'm not saying Go is bad.

I just want to point out that you can't really compare those two. Go is like "hey, let's make a simple langauge without bells and whistles that anyone can use." Rust is like "hey, lets spend some years to think about the best possible way to create a langauge that could replace C".

It's just a differnt design purpose.

>>56044453

aah okay, why rust over C++14?

>>56044465
1. Safety
2. If you prefer functional style to OOP

>>56044453
>hey, lets spend some years to think about the best possible way to create a langauge that could replace C
Well, they failed. Maybe they should have tried improving C instead.

>>56044484
1) You do get safety with RAII
2) You can whack up some functional kind of c++
okay i agree with 2)

>>56044331
Good point

It has a weird syntax. The if or switch clause described in the book felt alien, I could never really get into it. I've tried.

Also can I now declare a python-esque dictionary on compile time? I wanted some sort of O(1) access to known values in something that wasn't a bunch of variables.

Productivity. Not that it matters in the fields where Rust should be used (crypto, high performance, ...) but that's what I personally have against it. I'd rather write my code quickly and refactor a bit later than having to care about traits, ownership, etc.

>>56044520
Thatt's pattern matching, from functional languages. Look it up.

>>56044465

Well' Im not a C++ guy, so just two liks:

>http://katecpp.github.io/rustlang/

>http://cantrip.org/rust-vs-c++.html

At the end of the day, it just a matter of what you like. If you are super familiar with C++ and like the way C++ does things - hey, good for you!

But if you would like a small and more readable language, maybe you should look into Rust.

>>56043132
Lots of overhead, without any real revolutionary changes. Literally the same as C with snippets and a dash of static analysis. It's not great, it's not bad, who gives a shit. Consistency is more important, so let's stop with all the new dumb languages. Okay?

>>56044533
So can I use that with strings to find regex?

>>56044596
Different concept. Look up functional pattern matching.

>>56043132
Nothing. Haven't tried it yet, so don't have an opinion.

How does Swift compare to Rust? Swift has its own safety features too.

>>56044647
swift is garbagio bro

Immature.

I'm okay with Rust's size and overhead for the most part but I don't understand linking in jemalloc by default. System allocators are good for 99% of the people programming and people not fine with it will use their own custom allocators.

There's also the fact that the syntax is way too verbose at times and trying to use it like C with unsafe is worse than using C itself with how much undefined behavior is involved.

That being said, it's still an impressive language.

So conclusion is that while rust provides some benefits over C it lacks libraries and has to bind to unsafe code for pretty much anything.It doesnt provide anyhing over C++14 which is one of most used languages with bunch load of libraries except personal taste of syntax.

>>56044654
Why?

>>56044711
Its useless for anything outside of OSX and apple sphere

Supposedly, you can get the exact same benefits in C++ with static analysis.

>>56043718
>just as fast
(citation needed)

>>56043132
I have nothing against Rust, I think it is a fine language.

>>56044747
No. Look up borrow checking.

>>56044695

Yes. It will take time but I only see Rust displacing C++ if the proposed Core Guidelines are very deficient in scope which is not the case.

Even if it's 'good enough', people won't switch.

>>56044788
how borrow checking more memory safe than std::unique_ptr<> ?

>>56044711
just is bro

>>56044730
[citation needed]

>>56044711
https://spin.atomicobject.com/2016/04/26/swift-long-compile-time/

People also forget that Rust makes concurrency easier and safer than c++ (no data races).

>>56044256
>but it has no powerfull features.

go fuck_yourself()

>inb4 haskell/literallywho/malbolge also have an equivalent to goroutines and channels
But none of these are as convenient to use.

>>56044890
ur butt isn't convenient to use but it sure gets rammed by beefy bara men all day

>>56044820
That's out of date. It's fixed.

File: 1441400747184.png (321KB, 540x588px) Image search: [Google]

321KB, 540x588px

>>56044902
Yes, but this doesn't invalidate my argument.

>>56044914
seriously swift?
Lets just use ruby for embedded systems!
in what way is it better than java or c# or whatever?

>>56044505
>You do get safety with RAII
You don't.
Unless you don't get raped by the compiler with checks, it isn't safe.
RAII in C++ is a comfort feature, not a security one.

>>56045078
explain?

uniqe_ptr = something

when do i get memory leak?

>>56044674
>There's also the fact that the syntax is way too verbose at times
This is the big thing for me. I don't want my code to look like

.unwrap().unwrap().unwrap().unwrap().unwrap()

>>56043733
Rust doesn't have anything like C unions.

>>56043399
You forgot about assembly.

Rust is a language for nigger cattle.

>>56044890

>But none of these are as convenient to use.

Go? Serisouly?

>Not extensible.
>A crash (unhandled panic) of goroutine crashes the entire system. A whole system may halt due to an individual bug.
>There is shared memory in Go. Hence, if one goroutine crashes, even if you catch the panic, it might leave some permanent in-memory junk i.e. corrupt data that will compromise other parts of the system.
>It's impossible to unconditionally terminate a goroutine. You can do something manually, but this amounts to sending a message "please kill yourself", and hoping this will happen. If due to some bug, you have a goroutine that ends up in a tight CPU bound loop, you can't possibly stop it, and it will waste your resources forever.

Rather use Elixir, it is superior in every single way:
-> It's purely functional
-> Super nice syntax
-> amazingly scalable
-> the most fault-tolerant langaue you'll find (since it's running an Erlang VM)
-> up to 10x faster than Go:
>https://www.quora.com/Concurrency-computer-science-Should-I-use-Go-Erlang-or-Elixir-for-a-mobile-app-backend


Elixir is love, Elixir is life..

>>56045413

>up to 10x faster than Go

Sorry, didn't read that correctly..

It was up to 10 times faster than Ruby, but still twice as fast as Go.

>>56045413
1)Does it have as many learning resources and libraries?
2)It requires learning new paradigm which is much slower hence its more expensive ($man hours > cpu hours)
3)Is it as productive? refer to 2)

>>56044168
Good! Reverse SJWs will save the White race.

>>56045476
but seriously it looks like neat language i want it to have good learningresources
what IDE you use for it?

>>56045394
Get off 4chan Terry

>>56043132
Isn't the community one of the most important parts of a programming language?

>>56045100
If you forget to do that or have some other asswhipe smuggle a new or malloc into the companies codebase. Or if you somehow fuck up the custom deleter you need all the time, because you need C libraries half the time, because C++ isn't great for libraries.

>>56045568
isnt that also true for rust?

>>56045593
Well Rust can't be any better than C++, so I'm sure the same mistakes are possible in Rust as in C++.

>>56045476
1) https://doc.rust-lang.org/stable/book/
http://rustbyexample.com/

>>56045622
I was talking about elixir haha

>>56045476

>1) Does it have as many learning resources and libraries?

Yes a lot of.
Elixir extends the Erlang langauge, therefore you can directly use Erlang libraries:
>http://elixir-lang.org/getting-started/erlang-libraries.html

Furthermore ther's the Phoenix Web Framework, which is pretty easy.


>2) It requires learning new paradigm which is much slower hence its more expensive ($man hours > cpu hours)

If you want to be decent in Elixir, you gotta learn Erlang first, but I wouldn't consider this a downside. Erlang is a very clean and small functional language..
When Haskell is C++, Erlang is C.

For exmaple, this is quicksort in Erlang:

sort([Pivot|T]) ->
    sort([ X || X <- T, X < Pivot]) ++ [Pivot] ++ sort([ X || X <- T, X >= Pivot]);
sort([]) -> [].

> 3) Is it as productive?

If you go to the Phoenix web site, the first word you read is: "productive".
>http://www.phoenixframework.org/

Elixir and Phoenix are incluenced by Ruby and Rails, but IMHO without the bad parts and with a lot of good parts.

For example Ecto (DB system of Phoenix) is way better than ActiveRecord (Rails):
>https://www.amberbit.com/blog/2016/2/24/how-elixirs-ecto-differs-from-rubys-activerecord/

Phoenix outperforms Rails often about 10 times.

You can even do AI in Elixir/Erlang:

>http://www.automatingthefuture.com/blog/2016/3/21/ai-generate-and-test-elixir

>http://www.springer.com/us/book/9781461444626


It is seriously the best language ever created.

>>56045593
No, because it is compile time checked. Without the additional pointer look-up at runtime.

If you mean the library thing, I overheard that they had plans to stabilize/standardize their ABI, but who knows what will actually happen. I'll just fast-forward some years before using rust.

>>56045645
That means, single thread, unparalleled performance can't be better than golang ones, because BEAM isn't great in that regard.

>>56045645
Okay im interested sounds like better option than rust.

What IDE or setup do you use?

>>56045645
So i should learn erlang first then i can start with elixir?

>>56045645

Just one more thing to add:

I wouldn't consider Rust and Elixir as rivals, since they are for different purposes.

Rust is cool for low level speed tasks where you want to optimize your code by fucking arround with the memory. Basically what C or C++ are used for today.

Elixir/Erlang are a tad slower if they run in as a single process. Only if when it comes to concurrency they can really unleash their power.

But the fault tollerance of Erlang alone is reason enough to use it whenever it comes to concurrency.

>>56045522
but SJWs are the whitest people I know

>>56045456
>twice as fast as Go.
This means literally nothing. Fast at what? Under what conditions? Is the code compared fairly in all cases or is one of the two optimized/retarded?

File: ElixirSublime.gif (302KB, 633x437px) Image search: [Google]

302KB, 633x437px

>>56045670
>>56045671

See:
>>56045729

I wouldn't compare Rust and Elixir since they have different pros and cons, differnent areas where they shine.

But I see no sense in using Go for concurrency when you can choose to use Elixir.

>What IDE or setup do you use?

vim-elixir or ElixirSublime, see picture..


>So i should learn erlang first then i can start with elixir?

You can start directly with Elixir.
But I highly recommend learning Erlang first. It's a small langauge, but does some things in a very special way. It's easier if you get some concepts first..

I really like "Programming Erlang" by Joe Armstrong or Armstrongs thesis about Erlang:
http://erlang.org/download/armstrong_thesis_2003.pdf

Or If you don't want to read lengthy books, the "Learn you some Erlang for great good" online tutorial is pretty good..

>>56045830
one more question is erlang/elixir general purpose language or is it more domain specific? For e.g can i make (while remaining sane) 2D game engine? Gui application? Or is it mostly used for web backend?

>>56045645
>I like go because it's convenient
>dude just use elixir
>you must learn Erlang first, though
No thanks.

This quicksort looks as readable as perl 5 code.

>For example Ecto (DB system of Phoenix) is way better than ActiveRecord (Rails):
>recommending using an ORM

>Phoenix outperforms Rails often about 10 times.
Rails is slow as shit so this isn't really an achievement.

>You can even do AI in Elixir/Erlang:
You can do AI in any language you want.

>>56044919
oh you

>>56045887

>is erlang/elixir general purpose language or is it more domain specific

Well, you can do almost anything with it, but it's strengths are centered about the web..

For examle you have wxErlang, a binding from Erlang to the wxWidget library (written in C++) so you can could plattform indipendent GUIs with it.
But most of the times you want Elixir/Erlang as backend, taking requests in XMPP, REST or SOAP.

If you want to be a cool Kid, you use Elixir as backend and Elm as frontend - therefore you have even mix Erlang and Haskell, haha.. :D


>This quicksort looks as readable as perl 5 code.

Because my formatting was broken, sorry.
Step by step:

sort([Pivot|T]) ->

Function is called "sort", then you do a "functional-languae specific" stuff, by taking the first element from the list (called Pivot). A list is writte like that: [ ]

sort([ X || X <- T, X < Pivot]) ++ [Pivot] ++ sort([ X || X <- T, X >= Pivot]);

[ X || X <- T, X < Pivot] is a list comprehension (similar to Python): "take all X from T which are smaller than pivot"

Then you just add the lists with the "++ operator".

sort([]) -> [].

If the list is empty, give back an empty list (end recursion).

>>56043132
Shit productivity.
No fun, write Rust sucks the life out of me.

I don't need 100% static safety.
If I need speed I just write C/C++ and just don't act like a drooling retard.
If I have a bit more budget I write in a language that takes care of memory and such for me.

>>56045898

>you must learn Erlang first, though

You don't have to.

To get some nuances you should, but you can also just play arround with Elixir and if you want to know the details, take a dive into erlang.


>You can do AI in any language you want.

Can you run a language without operation system? Because in Elixir you can:
>http://elixir-lang.org/blog/2013/05/02/elixir-on-xen/

Oh well, I'm not herre to preach. Go is a small nice language, but comapred to the power of Erlang it's a toy langauge.

Erlang runs "WhatApp" for example. It's a completely differnt beast..

>>56046115
For proper web server do i have to learn elixir or can i just use erlang?

>>56046141

Most people feel that Elixir has a nicer syntax than Erlang. Even the creator of armstrong found that elixir adds a lot of nice features.

But yes, if you don't need a shiny backend framework but want to be closer to the metal, Erlang works perfectly fine on it's own, since 20 years.

>>56046183

> the creator of armstrong

Hahaha..

I meant: "armstrong, the creator of Erlang"

>>56046141

Here, have a small webserver in Erlang:

-module(hello).
-export([start/1]).

start(Port) ->
    spawn(fun () -> {ok, Sock} = gen_tcp:listen(Port, [{active, false}]), 
                    loop(Sock) end).

loop(Sock) ->
    {ok, Conn} = gen_tcp:accept(Sock),
    Handler = spawn(fun () -> handle(Conn) end),
    gen_tcp:controlling_process(Conn, Handler),
    loop(Sock).

handle(Conn) ->
    gen_tcp:send(Conn, response("Hello World")),
    gen_tcp:close(Conn).

response(Str) ->
    B = iolist_to_binary(Str),
    iolist_to_binary(
      io_lib:fwrite(
         "HTTP/1.0 200 OK\nContent-Type: text/html\nContent-Length: ~p\n\n~s",
         [size(B), B])).

i fell for the CS meme and i have some basic experience
i'm working on learning haskell right now and in school i'm going to try to focus on C

am i making a mistake
what languages should I be learning so I don't get out of college and find myself making $50k/year at a pajeet tier job

>>56046056
Yeah, this is easier to understand now that you explained it to me but still, I come from C so I feel like my brain is now permanently wired to write and read imperatively.

>>56046115
>You don't have to.
But in this case I'd forever be a mediocre elixir programmer. On the other hand, since I have been writing C for a very long time it didn't take me very long to be extremely confident in my go skills.

>Can you run a language without operation system? Because in Elixir you can:
Do you honestly think that you'll ever need to do that? That's cool technically but if I had a list of reasons to use a language this wouldn't even be on it.
I know that elixir is a great language, /g/ and HN have been sucking elixir's dick for years now so they must have a good reason. The fact still is that, for me, go is the most convenient language to use in which I get efficient concurrency for free and it lets me be productive with my current set of skills. Which is 98% of what I care about.

>>56046229

C and Haskell are fine.

If you know Haskell, other functional langauges (i.e. like Erlang) will be very simple for you to pick up.

If you want a Job, learn Java, there's many many Java jobs out there.. But Java is an ugly language. :)

But with C + Haskell + Java you would already be a dangerous person.

If you want to go web, then you have to learn the basics (HTML, CSS, JavaScript) and a backend language of your choice, i.e. Ruby, Python, Node.JS, PHP (--> please don't) or Erlang/Elixir.

With Python or Ruby you also have nice langauges for everyday scripting, doing simple tasks you want to do automatically, so I'd say learn one of them.


My personal way was:

VBA
Java
C
Ruby
Erlang
Elixir

I only regret doing VBA, but it was for work. The rest was pretty nice.

>>56046354
what's your salary look like if you don't mind me asking?
i'm concerned that i'm fucking up by not going engineering

>>56046354
>But with C + Haskell + Java you would already be a dangerous person.
tfw the only three languages I know well are C, Haskell and C#

>>56046229
>what languages should I be learning so I don't get out of college and find myself making $50k/year at a pajeet tier job
Not Java
Not PHP
Not JS
Not C#
These are all languages that will land you a job for sure but that job will mostly consist of being a codemonkey implementing shitty UML classes diagrams.

>>56046384

Sorry, I was AFK..

Well, I get along. But if you really want to make a creer with good salary, your priorities should be:

1. Java
2. JavaEE

(long pause)

3. C
4. Python
5. Scala


But if you have passion what you do, you will be more successfull in the long run. No matter what your degree is, if you are really damn good, you will find work.

>>56046395

Also pretyt good combination IMO, with C# you can make up for Java jobwise. Or let me put it like this: C# is pretty close to Java, only without the fancy stuff. But you should have no problems switching to Java pretty fast.


>>56046399

This.

I migt not have the best well payed job ever, but before I code in Java again I'd rather learn Malbolge.

>>56047878

> C# is pretty close to Java, only without the fancy stuff

Eh, I meant: "C# has a lot of fancy stuff that Java doesn't have".


I also wanted to say that there are not many Elixir jobs right now, nowwhere as for exmaple Node.JS. But I believe we must code langauges we like, the industry will create job oportunities sooner or later.

If we only do what everyone does the way will be much easier, but maybe not as rewarding in the long run.

>it's an pointless waste of effort and resources
>it's solution to an non-existing problem
>when mozilla goes belly up, so will rust

>>56044053
>using the smiley with a carat nose

>start discussing Rust
>other functional language fags hijack the thread

every time

>ITT: Explain what you have against rust WITHOUT mentioning rust's actual problems

Can I play this game

ITT: explain what you have against C++ WITHOUT (and this is going to be extremely difficult for some of you):

>retarded users
>standard too big
>OOP

>>56049023
the OP excluded 3 things that often come up but have nothing to do with the language itself.

>>56045236
Usually you can replace this by stuff like

do_stuff().unwrap_or(12).and_then(|x| x.foo())

But I agree, this is the part about Rust that gets me mad sometimes. There should be some syntax sugar for this, like in Kotlin:

foo?.bar()?.lol() ?: 123

Although I heard there's talk about sugaring the try! macro

>>56045277
Yet, the RFC for unstagged unions has been merged, it just has to be implemented.

>>56043132
Syntax is a bit retarded in some places. and the organization of it's equivilant to classes\namespaces are pretty unintuitive

Wants to be an efficient systems language, but for all it's checks and bounds it defeats the purpose. it has some good benefits but... It's a great system for security tho. But at the same time, why not use Java or C#?

Literally they don't know how to make their shit modular. the base lib is fucking enormous.

>>56043132
the compiler is a mess [spoiler]because of the SJWs[/spoiler]

I don't quite like mozilla
I don't like it's development model (that cargo thing hat makes the project layout on your behalf)
It's not very engaging. I tried it, was nothing but drillwork with a rather plain language (yes yes ownership, SAFETY), boring nonetheless

Any Haskellfags here? What do you think of Rust in comparison?

It's just trying too hard to be different. It wants to be a typesafe, more concise C/C++, but it ends up being a syntactic clusterfuck.

Sometimes looks like C, sometimes a functional language, and sometimes like a scripting language. For anyone who knows another language, it's a huge amount of translating what you already know to an unnecessarily "special" syntax.

In terms of "it makes things safe", my response is just "okay, cool." What's your point? That comes with benefits and drawbacks. That's like preferring C# over C++ all the time because C# is safer. Both have their place, but something in between is more likely to have bad aspects of both than to have the best of both.

The whole "cargo" thing is just dumb from the ground up. Just let the programmer figure out how to set the shit up and compile.

>>56043132
Mozilla has already proven their utter incompetence with Firefox

>>56051108
I'm lispfag and don't care about either.

>>56043132
used by beta cucks

>>56043132
It's not C.

File: Marmont.jpg (180KB, 471x622px) Image search: [Google]

180KB, 471x622px

>>56043146
mfw anti-rust plebians still stuck in their Blub ways

>>56043132
Racket is better in every conceivable way.

>>56050722
>Wants to be an efficient systems language, but for all it's checks and bounds it defeats the purpose.

Checks such as what? Its checks and bounds are all done at compile time, that's the entire design ethos and rationale of the language.

>>56049023
C++ is great except all of the retarded garbage it inherited from C: text macros, header files, forward declaration, piles of undefined behavior, and the resulting need to design around all the legacy crap the language was built off of.

>>56043247
The restrictions are literally a codified, enforce version of C++ good practices. It won't restrict you if you don't code like a dumbass.

>>56043277
>>56043334
This is used to get around certain things like writing to locations in memory or if you're trying to call external functions.

>>56043282
Rust has one of the largest communities for a new language and has an unprecedented accelerating growth rate for such a "niche" language (systems programming is not common)

>>56043479
>GC
>good
kill yourself

>>56043708
Because Rust has cool features and borrows from many modern concepts.

>>56043719
The only valid reason with Rust is that you actually have to write C FFIs and shims.

>>56043762
Voluntary C++ users are legitimately mentally disturbed and should be gassed immediately.

>>56044465
Rust, much like C, only has a very small set of primitive concepts you need to know (traits are so powerful that it covers multiple C++ concepts) and it's easier without having to memorize semantics over 20,000 different features of the language.

>>56044711
It's a copy of Rust, and it doesn't have production-ready perf.

>>56045236
This is like complaining about callback hell in Javascript. Only a problem if you're just doing quick n' dirty shit, and not really a problem if you do everything right.

>>56052352
Uh, no. You shouldn't be "translating" anything in any language, let alone Rust. Do you write Python by first thinking in C and translating C to Python? Are you retarded?
And about that safety thing, it's like you've never done systems programming before. Almost all errors are due to memory mismanagement, and Rust solves them at compile time.
Cargo is amazing, and possibly the best package manager a language has to offer. You can set it up manually as well, too.

>>56043731
kek

>>56043132
Not as straightforward as C, not as useful as contemporary C++, not as expressive or concise as [Lisp/Python/Smalltalk/other-probable-meme-language-here]. Sorry, just can't see any reason to swap over.

It's not bad, it's just not good enough.

>>56053939
>It won't restrict you if you don't code like a dumbass.

The developers have repeatedly said Rust's borrow checker necessarily rejects many valid programs. They see this as undesirable and they've been regularly updating it since the release of 1.0 to make it more intelligent and expansive. If I were you I'd be more cautious about advocating for a language designed by people I thought "coded like dumbasses".

Dealing with dangling pointers and use after free is less infuriating than dealing with the rust compiler.

>>56053997
Having real closures and a mature static type system puts it well ahead of Python in expressiveness, in my opinion.

terrible inconsistent syntax, way too verbose and tedious to write (literally worse than java!), was invented for one project (servo) that is likely not going to actually end up being finished. rust will probably die as a result.

>>56054121
I only put that in there because I'd incite butthurt for leaving it out. I prefer Lisp languages a lot more.

>>56054118
Except when your use-after-free costs money.

>>56051108
no HKT, has mutable, not lazy, not expression based

the syntax is beyond retarded

>>56043132

Nearly every complaint I used to have has been fixed to a satisfactory enough degree. The community is actually really helpful if you steer clear of politics.

>>56043132
But anon, it IS a meme language.

>>56053939
how much does Mozilla pay to push its meme language on /g/?

>>56052379

B A S E D
A
S
E
D

>>56043132
I'd rather use Ada.

>>56056921

/g/ does this for free..

Just come up with the most autistic langauge you know and /g/ will love it.

>>56043132
>garbage collection
Not even oncef

>>56055712
>no HKT
nobody needs that outside of Haskell
>has mutable
hat is ST and IO
>not lazy
laziness is a bad default

>>56057987
It's not GC'd.

I wish there was a reason to not use C for everything.

lol no i dont

There's no need to bash Go at the same time as liking Rust, they're both decent languages and it's good they exist

>>56043132

Rust is new, and I'm a lot better of a C++ programmer. I don't hate it, but I'm not going to invest too much time into it. I will, however, keep an eye on it to see where it goes.

>>56043884

The only native data type that needs transformation is strings, since Rust does not do zero terminated strings. An i32 to a C int is zero overhead cost. Just cram it into a register like anything else and you're golden. Pointers, similarly, can be cast for free.

>>56059170

>they're both decent languages
Rust is an okay language.
Go is a bad language.

>>56057966

Is this a statement about how much you dislike Rust, or how much you like Ada? Because I've had a professor or two that actually liked Ada.

>>56055712

>has mutable
It's a system's language. It is expected that mutable state should be as easy as possible.

No HKTs

>no HKTs
GET OUT NORMIE LANGUAGE

https://github.com/Immington-Industries/way-cooler

For you bored fags that have nothing to do this weekend.

>>56059298
>Rust does not do zero terminated strings
Rust has std::ffi::CString for this. You do your text-splicing operations on a regular String, and then convert the result to a CString.

Does it have untagged unions yet?

>>56059298
>Is this a statement about how much you dislike Rust, or how much you like Ada?
Ada's kind of godawful. Rust is worse. At least Ada can get you a job.

>>56043731
What the fuck are you talking about? Holy shit you can't be this stupid.

>>56059170
Go is literally worse than Algol

>>56061221

Soon:
https://github.com/rust-lang/rfcs/blob/master/text/1444-union.md
https://github.com/rust-lang/rust/issues/32836

Sjw's are reason enough to avoid or attempt to destroy ANYTHING. They are vile and hateful.

>>56061374
>he doesn't know
https://youtu.be/F4Dvc1NrZJI

>>56060228
>todo
>electron for menu bar
fucking dropped

>>56043132
It's good in principle, but it was ruined by the shitty C++ syntax. I want to write stuff in rust, but the syntax is too awful.

File: 1454972996696.gif (2MB, 480x270px) Image search: [Google]

2MB, 480x270px

>>56043132
>numales
What the fuck is that????
Either you are a macho or a fucking faggot

>>56059298
>Is this a statement about how much you dislike Rust, or how much you like Ada? Because I've had a professor or two that actually liked Ada.
It's a statement of how much I like Ada.
If you have any questions about Ada, I'd love to answer them.

I don't know enough about Rust to make a qualified judgement.
Fell free to disregard my posts as typical /g/ shitposting.

I'm glad I've finally found a language that won't hurt my feelies :^)

>>56063926
Why is Ada considered more niche than Rust by distro package maintainers?

>>56064243
There are a couple of reasons for that.

1)
The biggest one I think is, that there simple are not a lot of well maintained open source Ada packages out there.
A lot of this comes from the specific niche Ada lives in.
Ada projects tend to be embedded, low level, safety-critical, very specific (both application and hardware wise) and proprietary.

2)
Builds are very non-uniform. There is no "default" build system in Ada.
Projects use gprbuild, make, automake, custom shell/python scripts or whatever.
So packaging them is more work than your typical C automake project.
And there isn't something like cargo/pip/npm/cabal for Ada.

3)
Full on Ada programs tend to not have dependencies on Ada packages at run-time.
They statically link that Ada dependencies, and dynamically link the C ones.
So Ada packages only end up as a build time dependency.
Like ghdl for example (a really good open source VHDL simulator).
So in the end there is no need to package the Ada dependencies.

4)
Importing (and exporting!) to/from C is so easy/good that you don't need these typical binding packages.

5)
It's seen as old, crusty and uncool by a lot of people.
Very much like Fortran, Cobol and Pascal.

>>56064616
Okay, so what distros would you recommend for getting started with Ada?

>>56065250
Any decent distro will do. gcc supports Ada.
Either get FSF's compiler (packaged as gcc-ada in Arch) or AdaCore's (available at libre.adacore.com). Both will be fine if you just want to learn the language.

>>56066477
The AdaCore one contains tool like gnatstub (create impl file stubs from headers) and and gprbuild (pretty good Ada/C build system).
But for just learning Ada, gcc-ada is enough.

The syntax is too esoteric
It should take a page from c#

>>56066519
Yeah, it also comes with the gps IDE, which is quite good.

>>56066552
C# is boring.

>>56066585
I never gave that one a chance, because the emacs Ada mode is pretty darn good.
It can even do "goto definition".

>>56066644
Ada support on vim is so bad. I don't mind it right now because I'm just learning, but I'll eventually switch to emacs+evil+ada-mode.

>>56066756
>emacs+evil+ada-mode
That's exactly what I do, and it's glorious.

>>56044809
what?
borrow checking confirms what will happen to your variables when you compile the program.
smart pointers do all the stuff at runtime

>>56050166
>Although I heard there's talk about sugaring the try! macro
Yep! In fact it has already been added. (I'm not sure if you can use it in nightly yet, though).

File::create("foo.txt")?.write_all(b"Hello world!")

>>56066829
unique_ptr adds no runtime costs.

>>56043132
ugly name and logo

File: doge-tiny.png (13KB, 100x100px) Image search: [Google]

13KB, 100x100px

>>56043132
this is their flagship project

>>56067341
much safe, such fast, wow

I think the only good guys on the Servo team are pcwalton and gw, those guys have done some cool shit with WebRender.

>>56044223
link or stfu

>>56044261
So is concurrency, I suppose

>>56044465
Because C++14 is a supserset of all the C++'s before it. If you could program everything in the world from the bottom-up in C++14 then you'd have a point.

>>56043132
>Explain why you don't like this thing
>Except for these valid reasons!
Stop that.

>>56043132
unwrap().something().unwrap().otherthing().unwrap()

>>56043132
it has potential but it's too early to tell. the three things you mentioned are also not great

>>56045830
aye dat preap theme

>>56053939
>Do you write Python by first thinking in C and translating C to Python?
since the average python code i have to deal with has garbage like

# Apply f() to all members in src_list
dst_list = [0]*len(src_list)
for i in range(len(src_list)):
    dst_list[i] = f(src_list[i])

he probably does. there are a lot of programmers only capable of thinking in C or Java, even at "top" companies

>>56043132
Got raided several times so I spent the rest of the time inside a tiny armored 1x2 and sometimes snuck out to build more layers of protection around it.

>>56043399
>dubs
>still can't code in XSLT

>>56044053
How stupid are you to actually believe that?

I guarantee you've never written a line of code that has been used anywhere but your basement.

>>56043157
why not?
it sure made me NOT want to learn this piece of fuck.

>>56043537
java is actually quite fast, especially now that we have hotspot.

rust was created by MIT niggers

File: sussman shig.jpg (71KB, 500x375px) Image search: [Google]

71KB, 500x375px

>>56053798
>rust
>not blub

>>56043277
basically this

rust is like "you can have only one mutable reference to object, but since it's impossible to write anything with this restriction you are doomed to use unsafe directly or indirectly on every second line"

I tried it but I couldn't figure out how to make clothes and people kept shooting me.

I've got no problems with Rust, but does it really have an SJW community?

Otherwise, I'm, not learning rust because it won't put food on the table yet.

>>56077698
The scary truth (for /g/, not me, I don't give a shit about someone's political standing as long as they don't shove it in my face) is that every programming community is full of SJWs/liberals, being liberal has always been correlated with a higher intelligence (in actual scientific research) and however much we want to meme about this, programming does require a minimum of an above-average intelligence.

If a SJW community bothers you that much, you can always fuck off back to /pol/ and shitpost all day like they do.

>>56077698
There are some oddballs in the Rust community, though not as much as you would expect. True SJW do not participate in the Rust community. They tend to stick with languages like Ruby and JavaScript. Remember: a SJW does not program, it 'codes'.

>>56077866
Yeah because things like the term SLOC were all invented by SJWs
Fuck off with your semantic pedantry. We've always used the two interchangeably, only bikeshedding fuckwits like you get hung up on what the right term to use is. Literal fucking bikeshedding.

>>56069984
how cancerous, fucking codemonkeys should be gassed

>>56054048
so what if it's a little overrestrictive. it's worth it in the long run. things like having shared memory in different threads when you *know* they are memory safe isn't memory safe, because you could be wrong.

also can't believe this thread is still up

>>56077493
Actually there are a bajillion cases where it's actually an impossible to handle restriction.

Those cases mostly happen with datastructures, but that's why they're in libstd, the invariants in the small unsafe code are verified through other means the compiler isn't smart enough to respect.

Besides that if you need internal mutability you have RefCell and Cell, and if you need to mutate something in multiple places at once you're likely doing it multi-threaded, so you use Mutex or RwLock and an Arc.

Most of the code you write typically isn't datastructures, it's using them, and the unsafe surface is small enough to verify either at compile time by leveraging the borrowck or at runtime because there's no other way to do it.

I've written several daemons and tools where I didn't need to use unsafe directly.

>>56043146
They're not reasons based on the language itself, which makes them kinda worthless.

>>56069984
what would be the proper way?
what does [0]*len(src_list) do, how is [0] valid in python?

>>56044674
why not use your own allocator if it's consistently better than the system allocator?

File: 1852930503.png (3MB, 10000x8640px) Image search: [Google]

3MB, 10000x8640px

>>56052352
>The whole "cargo" thing is just dumb from the ground up. Just let the programmer figure out how to set the shit up and compile.

The Rust community purports that a language should be defined by its community.

I don't like the community, ergo I don't like Rust.

>>56078119
My point is that putting such a strict restrictions and then bypassing them everywhere with dirty hacks is pointless.
In general rust barely offers any improvements in this field compared to c++ despite of all that "safe languge" hype.

>>56078534
>he uses a language because of the community
Good, you wouldn't have programmed anything anyway, in ANY language. Now fuck off and kys my man.

File: Screenshot 2015-06-27 00.58.44.png (1MB, 1694x932px) Image search: [Google]

1MB, 1694x932px

>>56077729
>tfw conservative and work with other conservative programmers

The best part is the liberal programmers on our team are too timid to call us out on our shit. I imagine they go home and cry on tumblr instead.

>>56066616
>I wish the floor was harder

>>56078585
Or maybe they just don't give a shit and just work to make some money and not to get into your pointless unrelated-to-programming political shit.

>>56078573
You post directly contradicts the statements of the Rust language project, being that community is the most important and strongest part.

Which is it then?

>>56045413
>>56045456
>Elixir
>faster than Go
https://benchmarksgame.alioth.debian.org/u64q/compare.php?lang=go&lang2=erlang
and for webdevs:
https://github.com/mroth/phoenix-showdown/blob/master/RESULTS_v3.md

>>56078614
>talk about safe spaces, gun control at work
>a co worker dares to disagree with you
>instead of defending position, look at floor and dismiss yourself

Keep on keepin on there buddy

>>56077729
>>56078585
(((conservatives))) are almost as bad as (((liberals)))

Community? What is that?
I just write c++ shit whole day long and barely ever communicate with other devs and anything.

>>56078634
and this excludes the performance gains merged in Go 1.7

>>56078555
Except they're not used everywhere as I just explained.

And that statement about C++ is just wrong, even using modern idioms with C++14/17 is prone to the same kind of retarded memory errors, it just happens.

With Rust it just does not, that's a fact. And even if 10% is unsafe code, you still have the certainity that the 90% doesn't suffer from it.

I repeat, you're inflating the actual need of using unsafe code in most codebases, just check yourself.

>>56078616
>being that community is the most important and strongest part
That's just lip service, no one actually believes that. The main reasons people have been drawn to Rust are the C-like performance, simplicity and memory safety and protection from programmer-related fuckups. And the fact that a giant like Mozilla is backing it.

>>56078634
Everyone leave, Rust coming through https://github.com/alexcrichton/futures-rs/tree/master/futures-minihttp

>>56078640
I don't give a shit, anon, I really could not care less either way, I don't really fall into conservative or liberal brackets. If people come to work for the political bants then that's up to them. I work for the money and to stay relevant and learn some cool stuff I might not know just reading /g/ or /dpt/.

>>56043132
The only valid argument against Rust, as many already commented in this thread, is its the horrendous syntax: I personally love C, including its syntax, but I think it only works because C is relatively small. Rust would be 10 times better if it used a ML-like syntax.

>>56078640
Liberals are traitors to the Constitution. I would fire an employee for being too leftist, it's legal.

>>56078738
Please don't pretend that if you are using unsafe indirectly than your codebase is "unsafe free".
And please don't "fact" random bullshit. Rust is not even mature enough to carry a single production application to start talking whether it's much safer or otherwise superior to c++ in any way. Write something and support it for at least 10 years. And then we'll listen to you.

>>56078877
When that unsafe code is small and verified, then yes, it's practically "unsafe free" because every external usage of it is verified and enforced by the compiler.

It is a fact, since the compiler enforces it, it can only happen within unsafe code, but if you're using unsafe code when you don't need to then you deserve it.

And Rust is being used in production by Dropbox and other companies, it's even inside Firefox.

>>56078751
rust hipsters finally managed to make a webserver, and it's supposedly 50% faster than the fastest Go server.
minihttp seems quite young, though. let's wait until they add more features to their server...

>>56044176
>I/O performance
for high-performance IO(load balancers, proxys, caches) you'll want C because no system call overhead on Linux(any other OS is simply not relevant when it comes to IO performance*)
consider HAProxy and LVS, both of which are written in C, and highly reliant on IO performance.

File: codeblocks_2016-08-14_20-29-10.png (9KB, 801x117px) Image search: [Google]

9KB, 801x117px

>>56044520
>I wanted some sort of O(1) access to known values in something that wasn't a bunch of variables.
constant array?

>>56044520
You want phf: https://github.com/sfackler/rust-phf

>>56079008
The fuck are you talking about.

>>56079156
I/O, which fast futexes are a part of

>>56079259
And how are those not available to Rust?

>>56079298
the point is that using anything other than C when IO performance REALLY matters is pointless

>>56078634

>Elixir
>faster than Go

Yes, defiantely!
But not if we talk about only a few threads..

In your benchmark they used 4 threads and 100 connections, that's not the place where you would need Erlang. But even here Erlang outperformed one of the Go frameworks (Martini Framework). So I would guess that the other Go Framework was pretty small. And that's kinda unfair. I can also put up a Rack site and say "Hey, look, Ruby is super fast!", but then my site is only suited for small tasks.


It's like comparing a frigate (Go) with an nuclear-powered aircraft carrier (Erlang):
Sure if just look at them maneuvering in a harbour you wonder why they built the carrier so big. But if you see them unleashing their full potential, play their game..


Again, I'm not saying Go is a bad langauge. I think it's pretty nice for small to medium tasks where I wouldn't even consider Erlang..
But Erlang in incredibly robust and has very lightweight processes. All those processes use a very simple yet powerfull communication system. So if you need to spawn a few hundred, thousand or even million processes - Erlang can handle this. And if you have a clever implementation that makes good use of concurrency (i.e. mapreduce) than you can expect a really good speedup.

So all I'm saying is for concurrency I would never use Go when I can choose Erlang instead.


>>56078751

Again, this is only a proof of concept for a mini-framework. I like Rust as a language, but please stop comparing apples and oranges.

>>56079379

Dat delusion of grandeur!


In 2008 there was a study about the
impact of distributed functional programming languages in the telecom sector. They were especially comparing C++ and Erlang..

>Nyström, J.H., P.W. Trinder, and D.J. King. Concurrency and Computation: Practice & Experience, 20(8),
2008.

What they found was:


1. Erlang was more robust and availible than the C++ code.

2. Erlang programms were less than one-third of the size of their C++ counterparts.

3. Due to ERLANG's lightweight processes, the implementation was three times faster than the C++ implementation.

>http://onlinelibrary.wiley.com/doi/10.1002/cpe.1223/abstract


Of course this is an area where C++ was not perfectly suited for (high availibility big concurrency), but that "C is always the best/fastest/coolest solution" is a bad meme. It ALWAYS depends on the task, otherwise we would only have one progarmmin language on earth..

>>56079944
you sound like an "internet marketer". how much are you being paid to shill for Erlang?

>>56069984
>I don't know what .map() is

>>56080149

Sorry, you are probably right that I sound like a fanboy..
I'm still in the process of learning Erlang and so I might got carried away a littel bit here. But I just really really like this language, it's so simple and yet powerfull..

Furthermore I just replied to anon because I think every langauge has a certain pros and cons.


For example Rust, it's a systems language that exists because (according to their website):

>There is too little attention paid to safety.
>They have poor concurrency support.
>There is a lack of practical affordances.
>They offer limited control over resources.


Now you can say C is perfect for "control over resources", but "little attention paid to safety" applies definately.
Or for Erlang, it's defiantely the best langauge for "concurrency support" but if you want "control over resources" to get super optimized code for your system software, Erlang is not among the top 10 for that.


Langauges are TOOLs, with very specific pros and cons. I hate overgeneralizations like "C is the fastest langue" - languagescan never be fast, only implementations. C is "fast" because you work pretty close to the machine and do a lot micoroptimization. But this comes at a cost (dev time, LOC, safty issues) so somtimes C is the perfect tool for your job and sometimes you definately don't want to use it, as every language..