'use strict';

// Use direct extract instead of `regenerate` to reduse browserified size
var src_Any = exports.src_Any = require('uc.micro/properties/Any/regex').source;
var src_Cc  = exports.src_Cc = require('uc.micro/categories/Cc/regex').source;
var src_Z   = exports.src_Z  = require('uc.micro/categories/Z/regex').source;
var src_P   = exports.src_P  = require('uc.micro/categories/P/regex').source;

// \p{\Z\P\Cc\CF} (white spaces + control + format + punctuation)
var src_ZPCc = exports.src_ZPCc = [ src_Z, src_P, src_Cc ].join('|');

// \p{\Z\Cc} (white spaces + control)
var src_ZCc = exports.src_ZCc = [ src_Z, src_Cc ].join('|');

// All possible word characters (everything without punctuation, spaces & controls)
// Defined via punctuation & spaces to save space
// Should be something like \p{\L\N\S\M} (\w but without `_`)
var src_pseudo_letter       = '(?:(?!>|<|' + src_ZPCc + ')' + src_Any + ')';
// The same as abothe but without [0-9]
// var src_pseudo_letter_non_d = '(?:(?![0-9]|' + src_ZPCc + ')' + src_Any + ')';

////////////////////////////////////////////////////////////////////////////////

var src_ip4 = exports.src_ip4 =

  '(?:(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\\.){3}(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)';

// Prohibit [@/] in user/pass to avoid wrong domain fetch.
exports.src_auth    = '(?:(?:(?!' + src_ZCc + '|[@/]).)+@)?';

var src_port = exports.src_port =

  '(?::(?:6(?:[0-4]\\d{3}|5(?:[0-4]\\d{2}|5(?:[0-2]\\d|3[0-5])))|[1-5]?\\d{1,4}))?';

var src_host_terminator = exports.src_host_terminator =

  '(?=$|>|<|' + src_ZPCc + ')(?!-|_|:\\d|\\.-|\\.(?!$|' + src_ZPCc + '))';

var src_path = exports.src_path =

  '(?:' +
    '[/?#]' +
      '(?:' +
        '(?!' + src_ZCc + '|[()[\\]{}.,"\'?!\\-<>]).|' +
        '\\[(?:(?!' + src_ZCc + '|\\]).)*\\]|' +
        '\\((?:(?!' + src_ZCc + '|[)]).)*\\)|' +
        '\\{(?:(?!' + src_ZCc + '|[}]).)*\\}|' +
        '\\"(?:(?!' + src_ZCc + '|["]).)+\\"|' +
        "\\'(?:(?!" + src_ZCc + "|[']).)+\\'|" +
        "\\'(?=" + src_pseudo_letter + ').|' +  // allow `I'm_king` if no pair found
        '\\.{2,3}[a-zA-Z0-9%/]|' + // github has ... in commit range links. Restrict to
                                   // - english
                                   // - percent-encoded
                                   // - parts of file path
                                   // until more examples found.
        '\\.(?!' + src_ZCc + '|[.]).|' +
        '\\-(?!--(?:[^-]|$))(?:-*)|' +  // `---` => long dash, terminate
        '\\,(?!' + src_ZCc + ').|' +      // allow `,,,` in paths
        '\\!(?!' + src_ZCc + '|[!]).|' +
        '\\?(?!' + src_ZCc + '|[?]).' +
      ')+' +
    '|\\/' +
  ')?';

var src_email_name = exports.src_email_name =

  '[\\-;:&=\\+\\$,\\"\\.a-zA-Z0-9_]+';

var src_xn = exports.src_xn =

  'xn--[a-z0-9\\-]{1,59}';

// More to read about domain names
// http://serverfault.com/questions/638260/

var src_domain_root = exports.src_domain_root =

  // Allow letters & digits (http://test1)
  '(?:' +
    src_xn +
    '|' +
    src_pseudo_letter + '{1,63}' +
  ')';

var src_domain = exports.src_domain =

  '(?:' +
    src_xn +
    '|' +
    '(?:' + src_pseudo_letter + ')' +
    '|' +
    // don't allow `--` in domain names, because:
    // - that can conflict with markdown &mdash; / &ndash;
    // - nobody use those anyway
    '(?:' + src_pseudo_letter + '(?:-(?!-)|' + src_pseudo_letter + '){0,61}' + src_pseudo_letter + ')' +
  ')';

var src_host = exports.src_host =

  '(?:' +
  // Don't need IP check, because digits are already allowed in normal domain names
  //   src_ip4 +
  // '|' +
    '(?:(?:(?:' + src_domain + ')\\.)*' + src_domain_root + ')' +
  ')';

var tpl_host_fuzzy = exports.tpl_host_fuzzy =

  '(?:' +
    src_ip4 +
  '|' +
    '(?:(?:(?:' + src_domain + ')\\.)+(?:%TLDS%))' +
  ')';

var tpl_host_no_ip_fuzzy = exports.tpl_host_no_ip_fuzzy =

  '(?:(?:(?:' + src_domain + ')\\.)+(?:%TLDS%))';

exports.src_host_strict =

  src_host + src_host_terminator;

var tpl_host_fuzzy_strict = exports.tpl_host_fuzzy_strict =

  tpl_host_fuzzy + src_host_terminator;

exports.src_host_port_strict =

  src_host + src_port + src_host_terminator;

var tpl_host_port_fuzzy_strict = exports.tpl_host_port_fuzzy_strict =

  tpl_host_fuzzy + src_port + src_host_terminator;

var tpl_host_port_no_ip_fuzzy_strict = exports.tpl_host_port_no_ip_fuzzy_strict =

  tpl_host_no_ip_fuzzy + src_port + src_host_terminator;


////////////////////////////////////////////////////////////////////////////////
// Main rules

// Rude test fuzzy links by host, for quick deny
exports.tpl_host_fuzzy_test =

  'localhost|www\\.|\\.\\d{1,3}\\.|(?:\\.(?:%TLDS%)(?:' + src_ZPCc + '|>|$))';

exports.tpl_email_fuzzy =

    '(^|<|>|\\(|' + src_ZCc + ')(' + src_email_name + '@' + tpl_host_fuzzy_strict + ')';

exports.tpl_link_fuzzy =
    // Fuzzy link can't be prepended with .:/\- and non punctuation.
    // but can start with > (markdown blockquote)
    '(^|(?![.:/\\-_@])(?:[$+<=>^`|]|' + src_ZPCc + '))' +
    '((?![$+<=>^`|])' + tpl_host_port_fuzzy_strict + src_path + ')';

exports.tpl_link_no_ip_fuzzy =
    // Fuzzy link can't be prepended with .:/\- and non punctuation.
    // but can start with > (markdown blockquote)
    '(^|(?![.:/\\-_@])(?:[$+<=>^`|]|' + src_ZPCc + '))' +
    '((?![$+<=>^`|])' + tpl_host_port_no_ip_fuzzy_strict + src_path + ')';