require 'gherkin/core_ext/array'

module Gherkin
  module RbLexer
    class <%= @i18n.underscored_iso_code.capitalize %> #:nodoc:
      %%{
        machine lexer;
 
        action begin_content {
          @content_start = p
          @current_line = @line_number
        }

        action start_pystring {
          @current_line = @line_number
          @start_col = p - @last_newline
        }
        
        action begin_pystring_content {
          @content_start = p
        }

        action store_pystring_content {
          con = unindent(@start_col, data[@content_start...@next_keyword_start-1].utf8_pack("c*").sub(/(\r?\n)?( )*\Z/, '').gsub(/\\"\\"\\"/, '"""'))
          @listener.py_string(con, @current_line) 
        }
 
        action store_feature_content {
          store_keyword_content(:feature, data, p, eof) { |con| multiline_strip(con) }
          p = @next_keyword_start - 1 if @next_keyword_start          
          @next_keyword_start = nil
        }
 
        action store_background_content {
          store_keyword_content(:background, data, p, eof) { |con| multiline_strip(con) }
          p = @next_keyword_start - 1 if @next_keyword_start                    
          @next_keyword_start = nil
        }

        action store_scenario_content {
          store_keyword_content(:scenario, data, p, eof) { |con| multiline_strip(con) }
          p = @next_keyword_start - 1 if @next_keyword_start                    
          @next_keyword_start = nil
        }

        action store_scenario_outline_content {
          store_keyword_content(:scenario_outline, data, p, eof) { |con| multiline_strip(con) }
          p = @next_keyword_start - 1 if @next_keyword_start                    
          @next_keyword_start = nil
        }

        action store_examples_content {
          store_keyword_content(:examples, data, p, eof) { |con| multiline_strip(con) }
          p = @next_keyword_start - 1 if @next_keyword_start                    
          @next_keyword_start = nil
        }

        action store_step_content {
          con = data[@content_start...p].utf8_pack("c*").strip
          @listener.step(@keyword, con, @current_line)
        }
 
        action store_comment_content {
          con = data[@content_start...p].utf8_pack("c*").strip
          @listener.comment(con, @line_number)
          @keyword_start = nil
        }
 
        action store_tag_content {
          con = data[@content_start...p].utf8_pack("c*").strip
          @listener.tag(con, @current_line)
          @keyword_start = nil
        }
 
        action inc_line_number {
          @line_number += 1
        }
 
        action last_newline {
          @last_newline = p + 1
        }

        action start_keyword {
          @keyword_start ||= p
        }
 
        action end_keyword {
          @keyword = data[@keyword_start...p].utf8_pack("c*").sub(/:$/,'')
          @keyword_start = nil
        }
 
        action next_keyword_start {
          @next_keyword_start = p
        }

        action start_row {
          p = p - 1
          current_row = []
          @current_line = @line_number
        }

        action begin_cell_content {
          @content_start = p
        }

        action store_cell_content {
          con = data[@content_start...p].utf8_pack("c*").strip
          current_row << con
        }

        action store_row {
          @listener.row(current_row, @current_line)
        }

        action end_feature {
          if cs < lexer_first_final
            content = current_line_content(data, p)
            raise LexingError.new("Lexing error on line %d: '%s'." % [@line_number, content])
          else
            @listener.eof
          end
        }

        include lexer_common "lexer_common.<%= @i18n.underscored_iso_code %>.rl"; 
      }%%
 
      def initialize(listener)
        @listener = listener
        %% write data;
      end
 
      def scan(data)
        data = (data + "\n%_FEATURE_END_%").unpack("c*") # Explicit EOF simplifies things considerably
        eof = pe = data.length
 
        @line_number = 1
        @last_newline = 0
 
        %% write init;
        %% write exec;
      end

      CRLF_RE = /\r\n/
      LF_RE = /[^\r]\n/
      CRLF = "\r\n"
      LF = "\n"

      def multiline_strip(text)
        crlf_count = text.scan(CRLF_RE).size
        lf_count = text.scan(LF_RE).size
        eol = crlf_count > lf_count ? CRLF : LF
        text.split(/\r?\n/).map{|s| s.strip}.join(eol).strip
      end

      def unindent(startcol, text)
        text.gsub(/^ {0,#{startcol}}/, "")
      end
 
      def store_keyword_content(event, data, p, eof)
        end_point = (!@next_keyword_start or (p == eof)) ? p : @next_keyword_start
        con = yield data[@content_start...end_point].utf8_pack("c*")
        @listener.send(event, @keyword, con, @current_line)
      end
      
      def current_line_content(data, p)
        rest = data[@last_newline..-1]
        rest[0..rest.index(10)||-1].utf8_pack("c*").strip
      end
    end
  end
end