Carlo Pecchia

A dive into Reia language (part 1)

Ok, I’m a curious guy …so why not spend some hours in learning another language?

I’m particularly interested in Reia after looking for some (safe) environment for distributed computation.

Reia has a syntax very similar to ruby, but it runs on Erlang virtual machine (that’s probably one of the most advanced and safe environment for distributed applications). A like a lot the idea: powerful of Erlang with (friendly) syntax fo ruby. Why not giving Reia a chance?

Disclaimer: I’m not a Reia expert, I’m just approaching it at the time of writing this post.

Installation

Prerequisites are small: you need ruby and erlang installed on your computer.

1. download Reia from Github
2. unpack it
3. enter its directory
4. hit a rake command
5. hit a rake install command

Done.

First program

In order to see that everything worked well, let’s try to hit some commands through Reia interactive shell:

$ ire
Reia Interactive Shell (prerelease)
Running on Erlang (BEAM) emulator version 5.6.5 [source] [async-threads:0] [kernel-poll:true]

Loading standard library from /opt/local/lib/erlang/lib/reia/lib... done.
>> puts("Hi, guys!")
Hi, guys!
=> [nil]
>> 2+2
=> 4
>> 2*5
=> 10
>> 
BREAK: (a)bort (c)ontinue (p)roc info (i)nfo (l)oaded
   (v)ersion (k)ill (D)b-tables (d)istribution
a

It works like the Erlang shell (so, use “CTRL+C” to leave it).

Now let’s try the must-do-it thing: Fibonacci serie.

$ ire
Reia Interactive Shell (prerelease)
Running on Erlang (BEAM) emulator version 5.6.5 [source] [async-threads:0] [kernel-poll:true]

Loading standard library from /opt/local/lib/erlang/lib/reia/lib... done.
>> module Fib
..   def calc(0)
..     0
..   end
.. 
..   def calc(1)
..     1
..   end
.. 
..   def calc(n)
..     calc(n - 1) + calc(n - 2)
..   end
.. end
=> Fib
>> 
.. 
.. puts(Fib.calc(10))
55
=> [nil]

Do you see the first nice thing? We can define functions by pattern matching on arguments list, like we usually do in Erlang.

Incidentally I’ve found a similar thing applied directly in Ruby with the use of Functor (used, for example, in Waves ruby framework).

Basically, with a functor we can (in Ruby) define somthing like that:

#!/usr/bin/env ruby

require 'rubygems'
require 'functor'

fib = Functor.new do |f|
  f.given( Integer ) do |n|
    f.call( n-1 ) + f.call( n-2 )
  end   
  f.given( 0 ) { 1 }
  f.given( 1 ) { 1 }
end

puts fib.call(20)

Language reference

Ok, now let’s summarize the language reference. Maybe it should be a little boring, but quickly you’ll have a big look at the picture.

Modules contain functions, so we can call it using the form module.function as seen above.

Data types: 10 (integers), 3.1415 (floats), true and false (booleans), nil (no value).

Atoms: :oranges, :"Hi people"

Strings: "", '', 'foo', "2 + 2 = #{2+2}"

Lists: [] (empty list), [1,2,3,4,5]

Tuples: () (empty tuple), (x,) (single element tuple), (x, :foo, fib(10)) Tuples are immutable value sets.

Hashes: {} (empty hash), {:name => 'Your name', 'age' => 32}

Binaries: <<>> (empty binary), <<0>>, <<0,1,2>> Binaries are list of integer in 0-255 range

As you can see most of the things are identical tu ruby.

Operators can be:

• boolean: and, or, not and the equivalents &&, ||, !
• comparison: ==, !=, <, >, <=, >=
• numeric: +, -, *, /, **, %

For branching and loops we practically have all ruby constructs:

if condition
  ...
elseif condition
  ...
end

unless condition; ...   end

x = 2 if condition
x = 1 unless condition

case condition
when :first
  ...
when :second
  ...
when _
  ...
end
# notice the use of the catch-all '_'

while condition; ... end

until condition; ... end

do ... while condition

for x in a_list
  ...
end

for x in 1..100
  ...
end

For array types we have the powerful each method too.

Lambdas:

>> times_2 = fun(n) { n * 2 }
=> #<Fun erl_eval:-expr/5-fun-3->
>> times_2(8)
=> 16

Don’t forget that state mutation inside lambda does not affect the outer scope.

module MyModule
  def a_method(a, b, &block)
    a = a+1
    b = b+1
    block(a, b)
  end
end

ret = MyModule.a_method(2, 3) {|a,b| a * b}
puts(ret)

Surely list comprehensions is the most appreciated thing coming from the Erlang “ancestor” side:

>> [x * 2 | x in 1..10]
=> [2,4,6,8,10,12,14,16,18,20]

We can also add multiple generators/filters:

>> [x * 2 | x in 1..10, x % 2 == 0]
=> [4,8,12,16,20]

Ok, maybe it’s enough for your curiosity. We’ll continue in a future post (maybe).