TITLE

Inline::Ruby

SYNOPSIS

    use Inline::Ruby;

    EVAL 'puts "Hello!"', :lang<Ruby>;
    say EVAL('Time', :lang<Ruby>).now.to_s;

    # Method calling, some simple params, cast to Str
    say ~'Time':rb.now;
    say ~'[2, 6, 8, 4]':rb.sort.slice(1,2);

    # More advanced, this is Ruby's map and each_with_index
    # This shows the :rb postfix-operator sugar instead of EVAL
    "[1,2,3,4]":rb
      .map(-> $n { +$n + 1 })
      .each_with_index: -> $n, $i { say "$n @ $i" }

    # Import and wrap ruby classes
    use csv:from<Ruby>;

    CSV.foreach: 'examples/hiya.csv', -> $row {
      say "   Raw row: {$row.gist}";
      say "Name field: $row[1]";
    }

DESCRIPTION

Module for executing Ruby code and accessing Ruby libraries from Perl 6.

BUILD / DEV

In theory you can get it from zef. Let me know if that's true :)

I'm running Ubuntu with ruby2.3-dev installed from apt. Then:

./configure.pl6  # Creates Makefile and runs make
make test        # or prove -e 'perl6 -Ilib' -v

STATUS

Master (released) branch

Develop branch

• This only barely works!

  • Lots of missing features

  • Sometimes segfaults!

  • Only tested on my machine!

• You can currently EVAL code

• Some return types converted to Perl 6 values:

  • TRUE

  • FALSE

  • NIL

  • FIXNUM

  • STRING

  • FLOAT

  • Objects are wrapped in an RbObject

• You can call basic methods!

• Some param types converted to Ruby values:

  • Int

  • Str

More examples:

# Use the :rb postfix to eval a string
# In a string context, .to_s is called in ruby
# In perl6, .gist is called during printing, which wraps native-ruby values
# in «...»:rb. So when you see that, you know you are looking at a wrapped
# native Ruby object

say '5':rb;      #=> «5»:rb

# If you do some basic math (+,-,*,/), they will auto-convert

say '5':rb + 2   #=> «7»:rb

# Do it the other way around and you'll get Perl6 values instead

say 2 + '5':rb   #=> 7

# Experimental native 'use'. Tries to import things

use csv:from<Ruby>;
my $data = CSV.read('examples/hiya.csv')
#=> «[["id", "name"], ["1", "andy"], ["2", "bella"], ["3", "chad"], ["4", "dua"]]»:rb

# That gets importing wrong sometimes, so you can do it more directly
# Here we'll slurp the file in Perl6, feeding the resulting string to Ruby JSON

BEGIN { ruby_require 'json', :import<JSON> };
my $data = JSON.parse("examples/slide-up.json".IO.slurp);
#=> «[{"type"=>"ClutterGroup", "id"=>"actor", ... }]»:rb

# Now $data contains a ruby Array with nested hashes, wrapped in a P6 proxy
# object. You can call methods and some operators, such as []. Note that ruby
# uses [] and not {} for hash access! But we alias {} so you can still use it.

say $data.length       #=> «2»:rb
say $data[0]["type"]   #=> «ClutterGroup»:rb
say $data[0]<type>     #=> «ClutterGroup»:rb

# The value there is still a RbObject (proxy object). You can force Str or
# Num context

say $data[0]<children>[1]<depth>    #=> «20.0»:rb
say ~$data[0]<children>[1]<depth>   #=> 20.0

# Can call methods with blocks!

"[1,2,3]":rb.map: -> $n { 1 + $n }         #=> «2, 3, 4»:rb

use csv:from<Ruby>;
CSV.foreach: 'customers.csv', -> $row {
  say $row[2];
}

NOTES/TODO - Brainstorming and such.

• Nice reference https://silverhammermba.github.io/emberb/c/

• A big trick is deciding when and how much to auto-convert between langs

  • It's nice to leave things in Ruby if they start there, so we don't have to copy it all over

  • Final values are nice to have as native Perl 6

  • Maybe we should only explicit-convert

• Mmm... maybe there should be two layers of .to_p6

  • One would do simple types -- strings, numbers

  • Second would do complex types -- Array, Hash

  • The simple would be called implicitly, the second explicit

• Write up Lang Integration Guide

• Separate out reusable roles

LICENSE

The Artistic License 2.0 -- See LICENSE file.

AUTHOR

Brock Wilcox [email protected] Some code from Stefan Seifert [email protected]