Crystal

ctrl + f "crystal" -> 0 results

Where are all of the crystal shills at?

Other urls found in this thread:

tiobe.com/tiobe-index/
twitter.com/SFWRedditImages

why would anyone think cloning Ruby syntax is a good idea?

Because it's the best. Except when it isn't; funny how they ruined it by putting Perl parts into it.

I'm still waiting for it to be ready for action.

stdlib contributor here, back for round two I guess.

Feel free to ask any non-meme questions, I'll answer eventually.

It's quite a ways away from being production ready, ignoring all the other obstacles (some of which were discussed in the last thread), the stdlib is not stable enough for major projects to depend on.

I'm currently writing an irc server with it, just to get better. Without knowing any ruby, it was pretty hard to get started because their documentation is so shitty

oh well, in another 5 years, I guess
golang 2.0 won't be a thing any time soon and you are basically competing with nim only, so no haste

Well since I'm a complete newbie to ruby/crystal, tell me how are the compile times? I'm mostly looking to a more light-weight rust alternative since I'd prefer to code rather than wait 10 minutes for the program to compile

>tell me how are the compile times?
The llvm amount, but way better than rust.
Basically, it's fast once it starts up.

>go
Why won't the meme die already?

We are also competing directly with Ruby now. Our primary reasons for creating Crystal were execution speed and easy, familiar syntax. Ruby has been experimenting with a stack based JIT VM that may mainline in 3.x, which will threaten the usefulness of Crystal.

Mediocre. File size for compiled binaries are also very large because we statically compile the entire stdlib and LLVM.

Oh hey static compilation, that is actually useful.

Would be nice if it only included the code it actually needed for a given program.

Great. Another programming language? What's this one about now?

Fuck. I seriously need to change professions.

>stack based JIT VM
I shmell hax and vulnerabilities.

Making a fast, productive programming language that is well prepared for multi-thread use and isn't for little babies, like go is.

>Ruby has been experimenting with a stack based JIT VM that may mainline in 3.x, which will threaten the usefulness of Crystal.
So what, they are compiling to C, so their warm-up times will be beyond catastrophic. And since you can't fix broken semantics with a JIT compiler and at best mitigate, their throughput will be lousy. They even aim for 3x improvements, only. I mean, the best effort is still TruffleRuby but even they are minimally better than PyPy.

So nothing to worry here, unless Crystal focusses on server software, which would be a stupid idea anyway because one can just use some JVM/.Net language and beat up anything there in terms of rough throughput anyway.

The current compiler doesn't perform proper deduplication of code across source files either last I checked.

Eh, maybe it's for the best. Wouldn't want to make the compile time rust-tier

It's true nobody who cares that much about throughput will use Ruby or Crystal, but the more the gap closes, the more it loses its place as the middle contender.

I've been told the YARV-MJIT implementation is surprisingly good, but I've not sifted through it myself. I also stated earlier it was a stack implementation, which is incorrect as it is actually register based.

What are the use cases for multi-thread use? Realtime engines and such?

I'm a backend/fullstack dev and I've barely ever had a use for it, except maybe optimising generation of reports. Multi-processing works just fine for a lot of general use cases.

>company behind it is barely making it
HAHAHAHA
it will be ded in 5 years

Ruby is ranked #11 in the tiobe index
tiobe.com/tiobe-index/
if the Crystal devs ever made a version of Crystal that was fully compatible with Ruby it would take off in popularity over night and probably replace MRI as the default Ruby implementation

If we made Crystal source compatible with Ruby syntax, we wouldn't bother making Crystal in the first place. Just skip the middleman and compile the Ruby code directly.

Even if it does die, the source is available so the knowledge isn't lost.

Crystal is actually in the 51-100 range on that list, if I'm reading it correctly.

Well whatever happens, I'm going to learn Crystal. Looks good enough for my scripting language of choice.

LLVM Masterrace

It will be always niche language like D.
Go is the future

not when *anything else* exists. Go is bad, and you should feel bad.

...

>muh no generics
I laughing all the way to the bank