This language is compiled and as fast as C while being as si

>>60628489
Program a game with it. Get it to work with apis

>not Nim

>>60628572
These things have already been done in Ruby, so it can be done just as easily in Crystal with the added benefit of being ridiculously faster.

>>60628489
Post programs that use it.

>>60628642
Have a look through https://github.com/veelenga/awesome-crystal

>>60628692
Oh great, now a question, how do i compile the code, with LLVM?

>>60628905
https://crystal-lang.org/docs/using_the_compiler/

post benchmarks fucker

>300 c++-like languages later, we still have not solved the problem
>maybe if we make it again
not to imply that the same thing doesnt sometimes happen with functional programming, but i don't see erlang having problems like people flat out refusing to use it because of the kind of shit that c++ people do. (and erlang's syntax is about as esoteric as you can get without actually writing prolog/forth/whatever)
maybe we don't need operator overloading, generics, and strong typing.
hell, maybe we don't need TCO.
maybe there is something else you can do to make a language good.

also llvm is shit lmao

>>60631159
Operator overload is nice if you do need to calculate stuff in your program.
If you can't comprehend that, you shouldn't be programming at all.

>>60628489
> as fast as C
prove it !

also, it's not portable, x86 only, etc .. doesn't sound very usefull

>>60628489
>another fucking language that says it's nicer than a standard
well shit, sure. the difference between this and ruby/c is that the latter two have established communities and infrastructure. it doesn't matter if this shit can shove a horse cock up its ass at record speed, it's not going to be very practical if you have to write every single shitty thing yourself when you could do it faster with c or ruby.
wait a few years until people start giving a shit

>>60628489
Produce a formally verified binary in it. Then we'll talk

>>60631225
>operator overloading is nice
>if you can't understand that, ur shit
reading comprehension is at a serious deficit these days
>maybe we don't need operator overloading
i don't know how this implied to you that i don't understand operator overloading or that i don't think it is a feature of enough value to call "nice", but whatever, i'll elaborate
"calculating stuff" (whatever that means) is all well and good, and i'm sure you 100% fucking REQUIRE it to add your vectors/angles/quats/multiprecision numbers/whatever (because who would want to use a dirty fucking pleb-tier inline'd function), but if you honestly can't even think of a world without it, you are at the minimum lacking in creativity.
and, just a reminder, operator overloading isn't remotely required to use a single operator to add multiple different types of data, especially when they should already be in the standard library.
all operator overloading is is letting users fuck with the operators from within the language, and while you are correct in saying that that is 'nice', we don't live in a perfect world, and everything has a cost.
matrix_add( m1, m2 ) is more readable in many cases, and with no operator overloading, you can be assured that anything you're reading can only do things the language says they can do, which i believe is the main reason we now have 50 competing standards to replace java and c++.

>>60628489
>uses GC
Dropped.

>>60628489
You forgot the most important part: You can spin the logo.


Seriously, it's not a system programming language. It's a better golang, though.

>>60628489
http://colinm.org/language_checklist.html

>>60631900
That list lacks quite a few options.

Another perfect langauge that fixes every problem C++ has? Okay, I'm sure this will be the one.

Literally who

File: 4L_cGJMfz0x.jpg (105KB, 454x603px) Image search: [Google]

105KB, 454x603px

Is this the rust thread?

Also
>ruby
>elegant
ISHYGDDT

>>60632568
>ruby
>not elegant
It's as good as dynamic anything languages get. Basically the Python for grown ups.

>>60628489
>gc
>as fast as C

>>60628489
>>60628607
>>60628692
>>60628954
No one cares about your language

>>60628489
Crystal doesn't use an actor model for its concurrency.

Ergo, it is stale. No reason why you should be using it over Go, apart from slightly better syntax.

>>60628489
Pro: Rich standard library
Con: No windows support, messy metaprogramming design, no file-scoped import, GC...

It could be better.

>>60631380
the benefit of using crystal is its almost identical to ruby, and calling c functions is pretty trivial.

https://crystal-lang.org/docs/syntax_and_semantics/c_bindings/

>>60628489
next time you r planning on shilling your shit on g atleast make a better logo

>>60628489
What benefit does this language give you?

Rust's whole selling point is a systems lang that has extra safety.

What does Crystal do that C or C++ doesn't?

>>60628489
> elegant as Ruby

That is not a good thing and judging by the example code on the website it is far from savable.

>>60638530
>What benefit does this language give you?
DESU, at this point, none. Not until they deliver that windows support.

But I can imagine when it can, then you can use it to write some trivial standalone cli/gui application.

C is too primitive, Rust is too crippled for generic application, go syntax is painful.

I can already imagine myself writing my own version of markdown or sass in it (for my own purpose of course)

>>60631511
>and, just a reminder, operator overloading isn't remotely required to use a single operator to add multiple different types of data, especially when they should already be in the standard library.

That only works for primitive data types, dumbass nigger.

>2017
>Shilling for a language with GC
Why tho.

>>60628584
>Nim
I'm curious how it compares to Nim

>>60639727
Nim is WHITESPACE SENSITIVE.

>>60628489
>Ruby syntax
>elegant

Pick one

>>60639761
ruby was elegant before a lot of new languages copied it

still, it's elegant enough compare to rust or go.

>>60639760
I don't care about that. Post actual important differences please.

>>60639822
too long to post here.
what do you particular care about?

File: caring_about_crystal.jpg (424KB, 1000x667px) Image search: [Google]

424KB, 1000x667px

>>60639791
>Caring about the appearance
>Ignoring the substance
Rust is conceptually superior since it introduces an actual innovation by bringing linear/affine typing into the mainstream, while Crystal/Nim/D bring nothing new to the table and are basically the same old C# with a different syntax.
You have to be a hipster to ignore that and to care more about syntax being `elegant`.

>>60631900
>Unfortunately, your language (has/lacks):
I'm not sure I get this question.
Is having and not having that feature BOTH mean wrong?
And then after that
>[ ] The most significant program written in your language is its own compiler
>[ ] The most significant program written in your language isn't even its own compiler
Is this bait, I can't tell?

>>60638530
Rust isnt a good language for desktop applications, neither is C, neither is C++. Leave the actual end user programs to Crystal and have fun writing libraries in your meme language - that's what it's for.

>"github" is a configuration key under package management
That's dumb, and worse it's advocading dumb shit like "I use github as version control".

>>60639836
>linear/affine typing
I care shit about it

>bring nothing new to the table
while it not really new, but crystal indeed support some modern features: type inference (with union types support), generic, native C bindings, unicode, etc.

I think it's doing enough.

>>60639692
GC is far preferably in any scenario where the value of productivity and saved man-hours trumps saved cpu-cycles and memory.

File: Consider_the_following_Red.png (242KB, 979x617px) Image search: [Google]

242KB, 979x617px

>>60639872
The first is kinda weird, but the second isn't ambiguous. Either the biggest project written in the language is it's own compiler, or even it's own compiler isn't written on that language.

>>60639888
more like: "we are too poor to have a proper package host, so we abuse github instead".

But it's understandable since they have like 2000 "shards" in total.

>>60639901
> some modern features: type inference (with union types support), generic, native C bindings, unicode, etc.
All these "modern" features are at least 25 years all, ML-family languages had them in the 90s, and they've been in the mainstream for another 10 years since C# 2.0.
I mean, you can say Haskell is a different beast being pure and shit, but OCaml had all these features in a user-friendly package for 20 years now.
>>60639916
Or you can minimize the difference in man-hours by using actually innovative approaches.

>>60639956
>Or you can minimize the difference in man-hours by using actually innovative approaches.
Name 1 (one) language which is competitive in productiveness and maintainability while exposing the intricacies of memory management to the programmer?

>inb4 some funny guy responds with C++

>>60639979
Rust.
>exposing the intricacies of memory management to the programmer
I said nothing about that, only that GC isn't the only alternative out there.

File: 1490484678175.png (435KB, 720x720px) Image search: [Google]

435KB, 720x720px

>>60628489
>as fast as C
Why the hell does everything claim that, but it's never actually the case?

>Fast as C
[citation needed]

>>60639956
ruby/python is 20 years old, and they are still considered "new".
What's your point?

>>60640001
>>60640002
they revoked that claim last year or so. pathetic.

File: 2017-05-29 03_38_22-The Crystal Programming Language.png (19KB, 671x660px) Image search: [Google]

19KB, 671x660px

>>60640038
Says it on the homepage

>>60640000
Rust is neither productive nor maintainable for application programming compared to e.g. C#, a 17 years old language.

>>60640064
That's the question of infrastructure, not the language itself.

>>60640000
>Rust
>I want all the bad parts of C and all the bad parts of C++ combined with all the mental illness of Ruby users

>>60640047
I didn't check the new website, they did revoke it in the previous design though.

>>60640075
It's most certainly the language itself.

return new String("Hello World");

How does that work out in Rust?

>>60640080
>everything I know about Rust I've learned from /g/
I mean, there's come legit criticism of Rust, but this makes no sense.

>>60640104

"Hello World".to_owned()

>>60640106
No, I actually tried using it, and it's basically all the feature problems of C, all the (ironically) feature creep of C++, all the DISGUSTING syntax of C++, and all the SJW insanity of the ruby community.

It's a shit language with no niche to fill. It serves no purpose other than to be a hipster language.

>>60640120
There you go: you have to manually tell the compiler that you want to return a copy of "Hello World" to avoid running afoul of the borrow checker. This of course has literally zero value for the end user, beyond the potential to save cpu cycles or memory through increased control over memory handling.

In this case you also felt compelled to returning a copied value to avoid compilation errors, where in a GC'd language it would be quick and painless to return a reference.

>>60640227
"Hello World" is a statically allocated string. You don't "own" a statically allocated string, and you cannot modify it.
In order to to have the same semantics as your code, I had to do that. I can easily return a (static) reference if I wanted to. What I'm doing is the exact same shit as your code.
>This of course has literally zero value for the end user, beyond the potential to save cpu cycles or memory through increased control over memory handling
You literally just said the value it is to the programmer, you fucking idiot.

>>60640295
Assembler has the same value to the programmer, and like Rust, C and C++ it is also loses out in productivity and maintainability to actual high level languages.

My god, you fucking script kiddies who progressed from PHP to Ruby to Golang and now to Rust. All you know and care about is that it's the new kewl, then you will gladly proceed to prune your fruit tree sapling with an axe and write web applications in fucking Rust. Kys, right now.

>>60631159
Crystal isn't C++-like. The only reason you may think it is C++-like is because you have not been exposed enough to natively compiled, object oriented languages outside of C++.

>>60638911
Wrong. You don't need operator overloading because you can instead use multimethods or typeclasses.

>>60632072
see
>>60640364

>>60640227
And it's funny how mostly Rustfags complain about the "repetitive error handling" in Go when writing Rust code feels like

x.unwrap().unwrap().to_owned().to_option().unwrap().to_error().to_str().unwrap().unwrap().map_error(|_| Error::new("lose all information here because error types suck"))

>>60640389
Go only has repetitive error handling when you pretend it's not an easy to write version of C.
Any time you need to check for errors in Go is the exact same time you'd be making a NULL check in C. But because you always use the same var name it looks stupid.

>>60640344
I honestly don't give a shit about Rust, but your complains are fucking retarded. C, Rust, and similar languages are designed for systems programming, not some ultra-high level VM garbage. This is the level of control that people expect from such languages.
Also, if your "garbage collector" and "high-levelness", why isn't your snippet of code just

return "Hello World"

? That's a trick question; it actually has absolutely nothing to do with garbage collectors at all.

>>60640389
Each time some dev Lang tried copy funcional language without a lot constructions, end up like unusable shit, unwrap values and chain of function are easy and pretty on FP.

>>60640389
It's funny how your made up example is still more readable and safer than `if err != nil {}` after each of the method calls you'd do in Go.
>>60640446
It's only easy when you're in the Either monad, and Rust has it since the ? operator was added.

>>60640433
>systems programming
well it's because people keep mention Rust here or fucking everywhere, like it's a language design for generic purpose or something, but clearly it's not.

Meanwhile crystal is perfectly for that purpose, only has the disadvantage of lacking windows support.

GC is fine for generic application, which is what crystal is aiming for.

>>60640591
>lacking windows support
That's a pretty goddamn massive disadvantage

>>60640599
indeed.

>>60639836
I think you totally miss the issue of programmer ergonomics. If you want proof that ergonomics matter, look at Elixir. It is 15 years younger than Erlang and is built on the same tech, the Erlang VM, but it already has more users. Why? Because it fixes Erlang's major warts: the weird syntax, the poor text support, the meager standard library, and limited documentation.

Linear types are cool if you work on game engines or device drivers, but not everybody needs linear types. For a lot of programming tasks a GC is fine and will let you develop faster. For those tasks a "C# with a different syntax" that was more pleasant to use than actual C#, compiled to native (static) binaries and had libraries would be a big deal. It could get programmers to abandon inefficient but nice-to-use scripting languages like Ruby and the RAM-hungry enterprise bullshit of Java and C# (fuck Entity Framework). This is why people care about Crystal, Nim, and D, why IBM is betting on Swift, etc.

>>60640424
Does Go have setjmp/longjmp? These helps with error handling in C.

>>60640840
Please don't use setjmp/longjmp.

File: eJwNyMERgyAQAMBeKAAOGMSzkHwdRAJOhDNwvjLpPdnnfsTdT7GIwnyNRan9GJH6LgdTDznJTJTPFK5jyEhVBeYQS02NhzKIzrnZGZiN0RN6oyw4tN5PaAEAEf7Fd93Ovlbqr-do_GD9Lnu2jfTqAOTVsvj-AJjrKxc.zoBjz257X6BzXvx4YvW4eemqLvI.png (122KB, 455x374px) Image search: [Google]

122KB, 455x374px

>>60638911
as i very clearly implied, you can make more of your data types primitive, retard
i know of 2 dsls that i can write vector( 1, 5 , 3 ) + vector( 3, 1, 5 ) in, and they don't even have anything related to oop. it has nothing to do with operator overloading if the operators are "overloaded" by the language itself, because operator overloading is the act of doing it inside of the language, unless you want to say that intel-style x86 assembly has operator overloading because it has addb, addw, addl, and addq aliased to a single 'add'

>>60640364
>a natively compiled oo language outside of c++
..which happens to be attempting to compete with c++ and java..
i think i've been exposed to plenty of compiled oop languages, not that being compiled has anything to do with the language itself.

regardless, i don't know if i 100% count a language as "natively compiled" unless it compiles itself, and considering this is just another llvm-dependent language, it really makes you wonder how anyone will ever replace c++ when they write their compilers using it.
not to say this language's particular goal is that, but it is interesting.

>>60640840
>Oh look, here is a ultra-cancerous abysmal idea.
>Hope that other shit tier language supports it as well.
Better jump into a cancer tornado.

>>60628489
Tested it. It's really comfy, but it simply isn't ready yet.

That said "fast as C" is bait, although it may be the case in some cases where the compiler is able to lower abtracts/escape analyse/exploit stack structs.

>>60631159
>also llvm is shit lmao
LLVM is great by the fact alone that it basically ended the dynamic language bullshit trend.

>>60638530
Being comfy like golang, but at the same time it's not for brainlets that can't into metaprogramming.

>>60640001
It's almost always the case in some scenario. Being faster than C in numeric benchmarks isn't hard.

>>60641115
How else would you implement fibers in C? ucontext is slow and notn-standard.

>>60641280
>llvm is great for ending the 'dynamic language' trend
yes i would have to agree that that is noble enough, just bothers me that academics can't write their own compilers these days and have to use baby's first compiler toolkits.

>>60641415
maybe because except for swift, programmers who chose LLVM aren't payed to implement their language?

>> Google for "Crystal language GUI".
>> Get no meaningful hits on the first page. No QT or GTK bindings, nothing.

Dropped.

>>60641280
>>LLVM is great by the fact alone that it basically ended the dynamic language bullshit trend.
You should check out Julia.

>>60641623
>> Get no meaningful hits on the first page. No QT or GTK bindings, nothing.
how about using "crystal-lang gui" instead?

I use visual basic,nobody things your "special" for writing c syntax. Well they do, but not in the way you want.

>>60641623
>You should check out Julia.
julia startup time is slow, therefore worthless.

Stop making every programming related thread in to a fucking Rust thread.

>>60641623
>You should check out Julia.
I did, another project that figured out afterwards that LLVM isn't a great JIT compiler, just like any other dynamic language implementers so far.

>>60641415
>these days
They couldn't before.
It's not like there are many organizations that can put that much money for optimisations in a modular, versatile compiler; in these days it's just more obvious because there is even less money in reimplementing the functionality found in LLVM.

>>60641415
>just bothers me that academics can't write their own compilers these days and have to use baby's first compiler toolkits
How many years into your CS major are you?

>>60641705
Isn't VB.NET just C# with XML literals?

>>60628489
>New language
>Fast as x elegant as y
Fuck off

>>60642624
>tfw C
>almost as fast as Fortran, modern C++
>elegant as COBOL
>safe as Perl

>>60642651
You seem to not know what "safe" means in the context of programming languages. Perl is memory-safe.

>>60642696
>reference counting
>memory-safe

>>60642721
https://en.wikipedia.org/wiki/Memory_safety

>>60642761
>it even says you either need a tracing garbage collection or statical analysis like Rust or some external checkers employs
Nice work proving my point, feggit.

>>60642807
>it even says you either
>either
It doesn't say that. You should read that Wikipedia article rather than skim it.

>>60643018
It does say that. This paragraph of the approaches part is literally the only thing of that article that does matter, which is actually quite a good quota when it comes to WP articles regarding software engineering. Reference counters that can't retrive cycles at runtime - and the Perl one can't, like most - aren't memory safe. All other RC approaches like proposed by D. Bacon are just tracing GC with extra steps.
And even that is purely academical, because if Perl was using a tracing garbage collector that would make it only memory safe, but not all the other kinds of safe.
Pro-Tip: If a language has an extra option like use strict it's determined to be unsafe garbage.

>>60628489
https://crystal-lang.org/api/0.22.0/Char.html
>A Char represents a Unicode code point. It occupies 32 bits.
>It is created by enclosing an UTF-8 character in single quotes.
>Unicode
>32 bits
>UTF-8
Dropped

>>60643368
Your point?

>>60643368
I hope you didn't want an u8 char.

>>60643368
what are you talking about?

>>60643498
UTF-8 is a particular implementation of Unicode and has variable length characters. There are other implementations that use 32 bits for every character like UTF-32. What crystal-lang actually uses matters (UTF-8 vs. UTF-32 vs. their own standard) if it's being positioned as a replacement for C/C++. Kyselves brainlets

>>60628489
Why not use actual C or Ruby then?

Also it seems kinda impossible. Or most likely it's C that kinda looks like ruby

>>60641747
>julia startup time is slow, therefore worthless.

So is Java, but no one cares. It is still the most popular language in the industry.

>>60643592
That's because Java is not for scripting.
>>60643558
>if it's being positioned as a replacement for C/C++
It isn't though.

>>60643558
>it's being positioned as a replacement for C/C++
no it isn't. crystal is for people who enjoy writing ruby but want better speed, that's why it's almost identical to ruby.

>>60643258
Honestly, it is kind of silly to extend the definition of memory safety beyond access errors.

>>60643643
>It isn't though.
The topic of this thread disagrees with you:

>>60628489
>There is no excuse to use dated inferior languages such as C(++) etcetera.

File: images.jpg (5KB, 284x177px) Image search: [Google]

5KB, 284x177px

>>60643684
uh-huh
>>60643688
>The topic of this thread disagrees with you:
Like it's my fault OP is always a(n illiterate) fag.

>>60643577
ruby is slow, c standard library + syntax isn't good enough.

>>60643717
>Like it's my fault OP is always a(n illiterate) fag.
Yes, it's your fault for going off topic.

>>60643577
Because [dynamic scripting shit] + [shitty C bindings] isn't a solution for software that actually matters.

>>60643749
O-Ok, I'll just pretend to be a retard like OP from now on.

>>60643681
So what you're really saying is that OP is retarded?

>>60641747
How fast is Matlab startup time?
Or are you not aware of the literal purpose of Julia?

>>60643765
>pretend

>>60643873
>semi-shitposting is now not retarded

>>60643892
>taking things this literally
kek

>>60643368
And your problem is? Go and Rust do it like this as well...

>>60643906
>/g/ is now my favorite shitposting board
I'm well aware, but still, stop triggering my autism.

>>60643934
I apologize anon, have a kiss as compensation

>>60643968
Thanks anon, that gives me the energy to continue 1000% hypocritically shitposting in other thread.

>>60643917
Same argument applies to both of these meme languages, but what does that have to do with crystal-lang specifically being positioned as replacement for C/C++ in the topic of this thread?

File: IMG_4073.jpg (74KB, 1024x962px) Image search: [Google]

74KB, 1024x962px

>>60643968
I want a kiss

>>60644031
smooch :3
>>60644006
good luck

>>60644058
thanks, kek

File: IMG_4280.jpg (119KB, 485x946px) Image search: [Google]

119KB, 485x946px

>>60644058
I'll take anything, b-but are you a femanon?

>>60644084
I can be ;)

File: IMG_4272.png (44KB, 710x577px) Image search: [Google]

44KB, 710x577px

>>60644103

>>60644148
:3

>>60628489
Or, instead of learning a new language that I've never heard of before, I could just stick with tried and proven C.

>>60641280
>LLVM is great by the fact alone that it basically ended the dynamic language bullshit trend.

Bait thread?

Bait thread.

>>60643844
Julia purpose is irrelevant to my purpose of using a language.
V8 is JIT and it's blazing fast. LuaJIT is even faster!
Julia sucks.

>>60628489
>ruby but with an AoT compiler

why. why didn't you do something useful like make a better ruby vm?

meme bullshit LLVM garbage of the week.

>>60644322
you already learned it if you know ruby, which you should have known already just to prove that you are not a ignorant retard who only knows java or python because he learned it in school.

File: 1495923224721.jpg (12KB, 238x183px) Image search: [Google]

12KB, 238x183px

>>60631290
> mfw when x86 only...

>>60646595
Only for macOS and Linux as well. Apparently no one on the core team uses Windows.

>>60644483
>why. why didn't you do something useful like make a better ruby vm?
Not him, but those exist. But there is only so much you can archive for a dynamic language.
In practise, for real software static MOST THINGS languages that are AOT compiled with a sufficient optimization phase, possibly PGO involved will always be faster and predictable.

>>60646595
its x64 only though

>>60644483
rubinius, truffle already do the job

all meme languages compared to golang

>>60628489
>C while being as simple and elegant as Ruby
Change it to python syntax then we'll talk

>>60648258
Well they don't want a toy for non-programmers and data """""scientists""""" they want a programming language for grownups and professional programmers.

>>60648439
Good thing python's syntax instead of ruby's wouldn't change that

>>60648520
Hate to tell you, but you're wrong.
CPython isn't even implemented by proficient people.

>>60648554
What the fuck are you talking about?
python SYNTAX. I'm not talking about how python is.

>>60648725
These people can't implement a proper interpreter. You shouldn't trust them with the specifics of the interface to it.

If you meant a generally indentation based syntax, there is little reason to prefer it to a mostly keyword based language.
The only reason one'd strongly prefer it is being a keyboard cripple.
You wouldn't happen to be a keyboard cripple now, weren't you anon?

>>60649212
>there is little reason to prefer it
hahahahahaahahahahaahahahaahahahaahahahaa

>>60649323
>typelet

>>60631159
Golang is love.

>>60650388
it is indeed
http://golang.web.fc2.com