Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Intro to Crystal Language

Sponsored · Your Podcast. Everywhere. Effortlessly. Share. Educate. Inspire. Entertain. You do you. We'll handle the rest. →

Intro to Crystal Language

What is Crystal Language?

Talk presented online at RubyConfOnline December 2015.

Luis Lavena

December 02, 2015

Transcript

36 slides
  1. Overview of features • LLVM-backend • Ruby-like syntax • Compiles

    to native code • Global type inference • Compile time error checking • Go-inspired concurrency design*
  2. LLVM-backend • Used by many new languages (ie. Rust, Swift)

    • Simplifies architecture porting and cross compilation • Advanced branching and native code optimizations • Stands on the shoulders of giants
  3. Ruby-like syntax • Reduces learning curve • Inspired but not

    directly copied • many ways to do things vs one way (no aliases) • Compiled, no VM or bytecode to interpret • No eval or crazy meta-programming • Self-hosted (Crystal is written in Crystal)
  4. Ruby-like syntax, on steroids • Initializer shortcuts • Method overloading

    • Macros • Generics • 21st century Standard Library
  5. Initializer shortcuts # Ruby class Greeter def initialize(entity) @entity =

    entity end end # Crystal class Greeter def initialize(@entity) end end
  6. Method overload # Ruby def guess(value) case value when Integer

    puts "Integer!" when String puts "String!" else puts "No idea!" end end # Ruby guess 10 # => "Integer!" guess "hello!" # => "String!" guess true # => "No idea!"
  7. Method overload # Crystal def guess(value : Int32) puts "Integer!"

    end def guess(value : String) puts "String!" end def guess(value) puts "No idea!" end # Crystal guess 10 # => "Integer!" guess "hello!" # => "String!" guess true # => "No idea!"
  8. Macros • Executed before type inference • Transform AST at

    compilation time • Used for minimal meta-programming and callbacks ◦ inherited (Class) ◦ included (Module) ◦ method_missing ◦ delegate ◦ setter/getter/property ◦ respond_to?
  9. Type inference • Always know what is being used and

    where • Compile time checks beats runtime errors ;-)
  10. Types • String is not the same as Char ◦

    “a” vs ‘a’ • Int32 and Int64 • Float32 and Float64 • Nil • Boolean • Array of types (Array(T))
  11. Types a = [1, 2, 3].map { |x| x +

    3 } pp a # => a = [4, 5, 6] pp typeof(a) # => Array(Int32) b = [1, "a", 'b'] pp typeof(b) # => Array(String | Int32 | Char) # b << 12.5 # => Compile error (Float32)
  12. Modern Standard Library • No legacy libraries ;-) • HTTP

    Client and Servers built-in • OAuth2 and Websockets • YAML, JSON, XML, Markdown parsers • Growing ecosystem (shards)
  13. HTTP Server require "http/server" server = HTTP::Server.new(9292) do |request| HTTP::Response.ok

    "text/plain", "Hello world!" end puts "Listening on http://0.0.0.0:9292" server.listen
  14. HTTP Server (benchmark: wrk) $ ./wrk -c 10 -t 2

    http://localhost:9292/ Running 10s test @ http://localhost:9292/ 2 threads and 10 connections Thread Stats Avg Stdev Max +/- Stdev Latency 97.90us 212.99us 13.47ms 99.67% Req/Sec 55.93k 4.88k 112.00k 90.55% 1118367 requests in 10.10s, 107.72MB read Requests/sec: 110733.34 Transfer/sec: 10.67MB
  15. HTTP Server (prefork) require "http/server" server = HTTP::Server.new(9292) do |request|

    HTTP::Response.ok "text/plain", "Hello world!" end puts "Listening on http://0.0.0.0:9292" server.listen_fork(3)
  16. HTTP Server (benchmark: wrk) $ ./wrk -c 10 -t 2

    http://localhost:9292/ Running 10s test @ http://localhost:9292/ 2 threads and 10 connections Thread Stats Avg Stdev Max +/- Stdev Latency 128.90us 525.80us 12.09ms 97.18% Req/Sec 87.02k 17.70k 117.77k 63.50% 1731802 requests in 10.01s, 166.81MB read Requests/sec: 173042.01 Transfer/sec: 16.67MB
  17. HTTP Server (memory) # fresh $ pmap -x $(http_server PID)

    Address Kbytes RSS Dirty total kB 69772 4408 776 total kB 111224 5116 1316 total kB 70264 4956 1156 Start wrk GC
  18. JSON Parsing class Person # … mapping … def initialize(@name,

    @age) end end user = Person.new("Aaron", 34) user.to_json # => {"name":"Aaron","age":34}
  19. No

  20. No

  21. Web frameworks • Amethyst • Kemal • Moonshine • Carbon

    • Chocolate • Beryl* https://github.com/veelenga/awesome-crystal
  22. Simplified deployment • Single executable • Small footprint and size

    • Fast compile times • Easy Heroku deploy* https://github.com/scaint/heroku-buildpack-crystal
  23. Community • Small, friendly • Code of Conduct • New

    language, fewer libraries • 3rd party libraries starting to show • Less drama ;-) https://github.com/veelenga/awesome-crystal
  24. Useful resources • Docs: http://crystal-lang.org/docs/ • API: http://crystal-lang.org/api/ • Group:

    https://groups.google.com/group/crystal-lang • Twitter: https://twitter.com/CrystalLanguage • Awesome: https://github.com/veelenga/awesome-crystal

More Decks by Luis Lavena

Other Decks in Programming