shell_split_spec.rb

class String
  # Splits text into tokens the way a shell would, handling quoted
  # text as a single token. Use '\"' and "\'" to escape quotes and
  # '\\' to escape a backslash.
  # 
  # @return [Array] an array representing the tokens
  def shell_split
    out = [""]
    state = :none
    escape_next = false
    quote = ""
    strip.split(//).each do |char|
      case state
      when :none, :space
        case char
        when /\s/
          out << "" unless state == :space
          state = :space
          escape_next = false
        when "\\"
          if escape_next
            out.last << char
            escape_next = false
          else
            escape_next = true
          end
        when '"', "'"
          if escape_next
            out.last << char
            escape_next = false
          else
            state = char
            quote = ""
          end
        else
          state = :none
          out.last << char
          escape_next = false
        end
      when '"', "'"
        case char
        when '"', "'"
          if escape_next
            quote << char
            escape_next = false
          elsif char == state
            out.last << quote
            state = :none 
          else
            quote << char
          end
        when '\\'
          if escape_next
            quote << char
            escape_next = false
          else
            escape_next = true
          end
        else
          quote << char
          escape_next = false
        end
      end
    end
    out
  end
end

describe String do
  describe '#shell_split' do
    it "should split simple non-quoted text" do
      "a b c".shell_split.should == %w(a b c)
    end
    
    it "should split double quoted text into single token" do
      'a "b c d" e'.shell_split.should == ["a", "b c d", "e"]
    end
    
    it "should split single quoted text into single token" do
      "a 'b c d' e".shell_split.should == ["a", "b c d", "e"]
    end
    
    it "should handle escaped quotations in quotes" do
      "'a \\' b'".shell_split.should == ["a ' b"]
    end
    
    it "should handle escaped quotations outside quotes" do
      "\\'a 'b'".shell_split.should == %w('a b)
    end
    
    it "should handle escaped backslash" do
      "\\\\'a b c'".shell_split.should == ['\a b c']
    end
    
    it "should handle complex input" do
      text = "hello \\\"world \"1 2\\\" 3\" a 'b \"\\\\\\'' c"
      text.shell_split.should == ["hello", "\"world", "1 2\" 3", "a", "b \"\\'", "c"]
    end
  end
end