require 'nokogiri'
require 'cgi'

module Deface
  class Parser
    # converts erb to markup
    #
    def self.erb_markup!(source)

      #all opening html tags that contain <% %> blocks
      source.scan(/<\w+[^<>]+(?:<%.*?%>[^<>]*)+/m).each do |line|

        #regexs to catch <% %> inside attributes id="<% something %>" - with double, single or no quotes
        erb_attrs_regexs = [/([\w-]+)(\s?=\s?)(")([^"]*<%.*?%>[^"]*)/m,
          /([\w-]+)(\s?=\s?)(')([^']*<%.*?%>[^']*)'/m,
          /([\w-]+)(\s?=\s?)()(<%.*?%>)(?:\s|>|\z)/m]

        replace_line = erb_attrs_regexs.inject(line.clone) do |replace_line, regex|

          replace_line = line.scan(regex).inject(replace_line) do |replace_line, match|
            replace_line.sub("#{match[0]}#{match[1]}#{match[2]}#{match[3]}#{match[2]}") { |m| m = " data-erb-#{match[0]}=\"#{CGI.escapeHTML(match[3])}\"" }
          end

          replace_line
        end

        i = -1
        #catch all <% %> inside tags id <p <%= test %>> , not inside attrs
        replace_line.scan(/(<%.*?%>)/m).each do |match|
          replace_line.sub!(match[0]) { |m| m = " data-erb-#{i += 1}=\"#{CGI.escapeHTML(match[0])}\"" }
        end

        source.sub!(line) { |m| m = replace_line }
      end

      #replaces all <% %> not inside opening html tags
      replacements = [ {"<%=" => "<erb loud>"},
                       {"<%"  => "<erb silent>"},
                       {"%>"  => "</erb>"} ]

      replacements.each{ |h| h.each { |replace, with| source.gsub! replace, with } }

      source.scan(/(<erb.*?>)((?:(?!<\/erb>)[\s\S])*)(<\/erb>)/).each do |match|
        source.sub!("#{match[0]}#{match[1]}#{match[2]}") { |m| m = "#{match[0]}#{CGI.escapeHTML(match[1])}#{match[2]}" }
      end

      source
    end

    # undoes ERB markup generated by Deface::Parser::ERB
    #
    def self.undo_erb_markup!(source)
      replacements = [ {"<erb silent>"      => '<%'},
                       {"<erb silent=\"\">" => '<%'},
                       {"<erb loud>"        => '<%='},
                       {"<erb loud=\"\">"   => '<%='},
                       {"</erb>"                => '%>'}]

      replacements.each{ |h| h.each { |replace, with| source.gsub! replace, with } }

      source.scan(/data-erb-(\d+)+=(['"])(.*?)\2/m).each do |match|
        source.gsub!("data-erb-#{match[0]}=#{match[1]}#{match[2]}#{match[1]}") { |m| m = CGI.unescapeHTML(match[2]) }
      end

      source.scan(/data-erb-([\w-]+)+=(["'])(.*?)\2/m).each do |match|
        source.gsub!("data-erb-#{match[0]}=#{match[1]}#{match[2]}#{match[1]}") { |m| "#{match[0]}=#{match[1]}#{CGI.unescapeHTML(match[2])}#{match[1]}" }
      end

      #un-escape changes from Nokogiri and erb-markup!
      source.scan(/(<%.*?)((?:(?!%>)[\s\S])*)(%>)/).each do |match|
        source.gsub!("#{match[0]}#{match[1]}#{match[2]}") { |m| m = "#{match[0]}#{ CGI.unescapeHTML match[1] }#{match[2]}" }
      end

      if RUBY_PLATFORM == 'java'
        #un-escapes changes from Nokogiri under Java, where " are converted to %22 when in an attribute of an element
        #
        source.scan(/(<%.*?)((?:(?!%>)[\s\S])*)(%>)/).each do |match|
          source.gsub!("#{match[0]}#{match[1]}#{match[2]}") { |m| m = "#{match[0]}#{ match[1].gsub('%22', '"') }#{match[2]}" }
        end
      end

      source
    end

    def self.convert(source)
      # Look for # encoding: *. If we find one, we'll encode the
      # String in that encoding, otherwise, we'll use the
      # default external encoding.
      encoding = source.scan(/#{ActionView::Template::Handlers::ERB.const_get(:ENCODING_TAG)}/).first.try(:last) || Encoding.default_external

      # Tag the source with the default external encoding
      # or the encoding specified in the file
      if source.frozen?
        source = source.dup.force_encoding(encoding)
      else
        source.force_encoding(encoding)
      end

      unless source.valid_encoding?
        raise ActionView::WrongEncodingError.new(source, encoding)
      end

      erb_markup!(source)

      if source =~ /<html.*?(?:(?!>)[\s\S])*>/
        Nokogiri::HTML::Document.parse(source)
      elsif source =~ /<body.*?(?:(?!>)[\s\S])*>/
        Nokogiri::HTML::Document.parse(source).css('body').first
      else
        Nokogiri::HTML::DocumentFragment.parse(source)
      end
    end
  end
end