Question

I’m using Ruby 1.9.3 on Windows 7 with Tk interface. In the following simple example, if i click a button, GUI will return me a “??????” string instead of “привет” displayed. Is it possible to get back actual unicode string entered?

#!/usr/bin/env ruby
# coding:utf-8 vi:et:ts=2
require 'tk'
TkRoot.new.tap { |o|
  $edit = TkEntry.new( o ).tap { |o|
    o.pack( :side => 'left' )
    o.insert( 0, "привет" )
  }
  TkButton.new( o, :text => "click me" ).tap { |o|
    o.pack( :side => 'left' )
    o.bind( '1' ) {
      ##  In this place i want unicode, but got garbage :(
      puts( $edit.get().encoding.name )
      puts( $edit.get().inspect )
    }
  }
  o.mainloop()
}

Answer 1

So I checked in Windows and got it working. I’d recommend you put this at the top of your file:

#!/usr/bin/env ruby -Ku
require 'tk'
# etc.

The # coding bit is unnecessary; the -Ku flag tells Ruby to use the Unicode codepage. While my testing on Mac doesn’t seem to have this issue (either appending -Ku or using # coding: utf-8 will work), it is indeed occurring in Windows. I’m running the same versions as you, just in Parallels.

Alternatively, you could delete the shebang part and run the file with ruby -Ku test.rb

Original answer:
Yes, although the solution I’m accustomed to would be using UTF-8. You just have to put # coding: utf-8 on the top line of your file and Ruby will mystically switch over to processing strings in UTF-8:

# coding: utf-8
require 'tk'
TkRoot.new.tap { |o|
# etc.

For further reading, I’d suggest this link which goes over how Ruby thinks about encoding.