Subscribe now

Installation and Introduction to Crystal [08.18.2017]

Greetings! In this episode, we will introduce Crystal, a statically-typed and compiled language, learn how to install it, compile, and run a simple program.

So, first of all, what is Crystal? Crystal is a language with a syntax heavily inspired by Ruby, with a focus on programmer happiness. Even if Crystal can compile some simple Ruby programs, keep in mind that it is an entirely different language.

Unlike Ruby or JavaScript, which are interpreted languages, Crystal is a compiled language, making it much faster and with a lower memory footprint. Under the hood, it uses LLVM for compiling to native code.

Crystal is also statically typed, which means that the compiler will help you catch type errors in compile-time. But more on this later: let's now dive into installing Crystal.

We will install Crystal in Mac OS using homebrew, but as you can see, there are also instructions for Debian, CentOS, and other Linux distributions.

brew update
brew install crystal-lang

Installing Crystal on Mac is as easy as installing the crystal-lang homebrew package. After installation, we can test that it installed successfully.

crystal -v

We’ll write a simple hello world program.

puts "Hello world"

Now run it by passing the file to the command crystal:

$ crystal
Hello world

Under the hood, this is compiling our program to a binary and running it. We can do each of those steps separately. First let us compile the program:

$ crystal build
$ ls
$ file hello

As we can see, there is a hello executable.

And then we can run it:


This binary already has all of its required dependencies embedded, so we can easily copy it to any other machine with the same architecture and run it there, making deployment a very easy task.

Speaking of deployment, we can also compile our program in release mode, which takes longer to compile, but enables several optimisations that greatly increase its speed in runtime for production:

crystal build --release

Now, since we are working with a very small program, we won't notice any differences. Let’s set up the scaffolding for a larger project:

rm *
crystal init app hello
create  hello/.gitignore
create  hello/LICENSE
create  hello/
create  hello/.travis.yml
create  hello/shard.yml
create  hello/src/
create  hello/src/hello/
create  hello/spec/
create  hello/spec/

This will create the basic structure for a full-blown project, with an already initialised git repository, license and readme files. It also comes with stubs for tests, and the shard.yml for describing the project and managing dependencies, also known as shards.

As a quick example, let’s add the faker shard, which generates random names.

cd hello
vim shard.yml
    github: askn/faker

Note that Crystal does not have a centralized repository, like rubygems or npm. The way to specify a dependency is through its git repository, which in this case sits in Github.

We install the dependencies listed running crystal deps.

crystal deps

Now we can require the library and use it from our code.

vim src/
require "./hello/*"
require "faker"

And if we run our project...

crystal src/ prints out a random name.

So, now that you know how to install Crystal, and create and run your first project, in the next episode, we will get started with the language basics. See you then.