kizashi1122 / mail_address

  # --------------------------------------------------------------------------------------------------

  # This module is ported from Google Closure JavaScript Library

  #  -> https://github.com/google/closure-library/blob/master/closure/goog/format/emailaddress.js

  # --------------------------------------------------------------------------------------------------

  OPENERS_ = '"<(['

  CLOSERS_ = '">)]'

  # SPECIAL_CHARS = '()<>@:\\\".[]'

  ADDRESS_SEPARATORS_ = ',;'

  # CHARS_REQUIRE_QUOTES_ = SPECIAL_CHARS + ADDRESS_SEPARATORS_

  ESCAPED_DOUBLE_QUOTES_ = /\\\"/

  ESCAPED_BACKSLASHES_ = /\\\\/

  QUOTED_REGEX_STR_ = '[+a-zA-Z0-9_.!#$%&\'*\\/=?^`{|}~-]+'

  UNQUOTED_REGEX_STR_ = '"' + QUOTED_REGEX_STR_ + '"'

  LOCAL_PART_REGEXP_STR_ = '(?:' + QUOTED_REGEX_STR_ + '|' + UNQUOTED_REGEX_STR_ + ')'

  DOMAIN_PART_REGEXP_STR_ = '([a-zA-Z0-9-]+\\.)+[a-zA-Z0-9]{2,63}'

  EMAIL_ADDRESS_ = Regexp.new('\\A' + LOCAL_PART_REGEXP_STR_ + '@' + DOMAIN_PART_REGEXP_STR_ + '\\z')

  def self.parse_simple(str)

    result = []

    email = token = ''

    # Remove non-UNIX-style newlines that would otherwise cause getToken_ to

    # choke. Remove multiple consecutive whitespace characters for the same

    # reason.

    str = self.collapse_whitespace(str)

    i = 0

    while (i < str.length)

      token = get_token(str, i)

      if self.is_address_separator(token) || (token == ' ' && self.is_valid(self.parse_internal(email)))

        if !self.is_empty_or_whitespace(email)

          result.push(self.parse_internal(email))

        end

        email = ''

        i += 1

        next

      end

      email << token

      i += token.length

    end

    # Add the final token.

    if (!self.is_empty_or_whitespace(email))

      result.push(self.parse_internal(email))

    end

    return result

  end

  def self.get_token(str, pos)

    ch = str[pos]

    p = OPENERS_.index(ch)

    return ch unless p

    if (self.is_escaped_dbl_quote(str, pos))

      # If an opener is an escaped quote we do not treat it as a real opener

      # and keep accumulating the token.

      return ch

    end

    closer_char = CLOSERS_[p]

    end_pos = str.index(closer_char, pos + 1)

    # If the closer is a quote we go forward skipping escaped quotes until we

    # hit the real closing one.

    while (end_pos && end_pos >= 0 && self.is_escaped_dbl_quote(str, end_pos))

      end_pos = str.index(closer_char, end_pos + 1)

    end

    token = (end_pos && end_pos >= 0) ? str[pos .. end_pos] : ch

    return token

  end

  def self.is_escaped_dbl_quote(str, pos)

    return false if str[pos] != '"'

    slash_count = 0

    for idx in (pos - 1).downto(0)

      break unless str[idx] == '\\'

      slash_count += 1

    end

    (slash_count % 2) != 0

  end

  def self.collapse_whitespace(str)

    str.gsub(/[\s\xc2\xa0]+/, ' ').strip

  end

  def self.is_empty_or_whitespace(str)

    /\A[\s\xc2\xa0]*\z/ =~ str

  end

  def self.is_address_separator(ch)

    ADDRESS_SEPARATORS_.include? ch

  end

  def self.is_valid(address)

    EMAIL_ADDRESS_ =~ address.address

  end

end