index.js

'use strict';

var required = require('requires-port')
  , qs = require('querystringify')
  , slashes = /^[A-Za-z][A-Za-z0-9+-.]*:[\\/]+/
  , protocolre = /^([a-z][a-z0-9.+-]*:)?([\\/]{1,})?([\S\s]*)/i
  , whitespace = '[\\x09\\x0A\\x0B\\x0C\\x0D\\x20\\xA0\\u1680\\u180E\\u2000\\u2001\\u2002\\u2003\\u2004\\u2005\\u2006\\u2007\\u2008\\u2009\\u200A\\u202F\\u205F\\u3000\\u2028\\u2029\\uFEFF]'
  , left = new RegExp('^'+ whitespace +'+');

/**
 * Trim a given string.
 *
 * @param {String} str String to trim.
 * @public
 */
function trimLeft(str) {
  return (str ? str : '').toString().replace(left, '');
}

/**
 * These are the parse rules for the URL parser, it informs the parser
 * about:
 *
 * 0. The char it Needs to parse, if it's a string it should be done using
 *    indexOf, RegExp using exec and NaN means set as current value.
 * 1. The property we should set when parsing this value.
 * 2. Indication if it's backwards or forward parsing, when set as number it's
 *    the value of extra chars that should be split off.
 * 3. Inherit from location if non existing in the parser.
 * 4. `toLowerCase` the resulting value.
 */
var rules = [
  ['#', 'hash'],                        // Extract from the back.
  ['?', 'query'],                       // Extract from the back.
  function sanitize(address) {          // Sanitize what is left of the address
    return address.replace(/\\/g, '/');
  },
  ['/', 'pathname'],                    // Extract from the back.
  ['@', 'auth', 1],                     // Extract from the front.
  [NaN, 'host', undefined, 1, 1],       // Set left over value.
  [/:(\d+)$/, 'port', undefined, 1],    // RegExp the back.
  [NaN, 'hostname', undefined, 1, 1]    // Set left over.
];

/**
 * These properties should not be copied or inherited from. This is only needed
 * for all non blob URL's as a blob URL does not include a hash, only the
 * origin.
 *
 * @type {Object}
 * @private
 */
var ignore = { hash: 1, query: 1 };

/**
 * The location object differs when your code is loaded through a normal page,
 * Worker or through a worker using a blob. And with the blobble begins the
 * trouble as the location object will contain the URL of the blob, not the
 * location of the page where our code is loaded in. The actual origin is
 * encoded in the `pathname` so we can thankfully generate a good "default"
 * location from it so we can generate proper relative URL's again.
 *
 * @param {Object|String} loc Optional default location object.
 * @returns {Object} lolcation object.
 * @public
 */
function lolcation(loc) {
  var globalVar;

  if (typeof window !== 'undefined') globalVar = window;
  else if (typeof global !== 'undefined') globalVar = global;
  else if (typeof self !== 'undefined') globalVar = self;
  else globalVar = {};

  var location = globalVar.location || {};
  loc = loc || location;

  var finaldestination = {}
    , type = typeof loc
    , key;

  if ('blob:' === loc.protocol) {
    finaldestination = new Url(unescape(loc.pathname), {});
  } else if ('string' === type) {
    finaldestination = new Url(loc, {});
    for (key in ignore) delete finaldestination[key];
  } else if ('object' === type) {
    for (key in loc) {
      if (key in ignore) continue;
      finaldestination[key] = loc[key];
    }

    if (finaldestination.slashes === undefined) {
      finaldestination.slashes = slashes.test(loc.href);
    }
  }

  return finaldestination;
}

/**
 * @typedef ProtocolExtract
 * @type Object
 * @property {String} protocol Protocol matched in the URL, in lowercase.
 * @property {Boolean} slashes `true` if protocol is followed by "//", else `false`.
 * @property {String} rest Rest of the URL that is not part of the protocol.
 */

/**
 * Extract protocol information from a URL with/without double slash ("//").
 *
 * @param {String} address URL we want to extract from.
 * @return {ProtocolExtract} Extracted information.
 * @private
 */
function extractProtocol(address) {
  address = trimLeft(address);

  var match = protocolre.exec(address)
    , protocol = match[1] ? match[1].toLowerCase() : ''
    , slashes = !!(match[2] && match[2].length >= 2)
    , rest =  match[2] && match[2].length === 1 ? '/' + match[3] : match[3];

  return {
    protocol: protocol,
    slashes: slashes,
    rest: rest
  };
}

/**
 * Resolve a relative URL pathname against a base URL pathname.
 *
 * @param {String} relative Pathname of the relative URL.
 * @param {String} base Pathname of the base URL.
 * @return {String} Resolved pathname.
 * @private
 */
function resolve(relative, base) {
  if (relative === '') return base;

  var path = (base || '/').split('/').slice(0, -1).concat(relative.split('/'))
    , i = path.length
    , last = path[i - 1]
    , unshift = false
    , up = 0;

  while (i--) {
    if (path[i] === '.') {
      path.splice(i, 1);
    } else if (path[i] === '..') {
      path.splice(i, 1);
      up++;
    } else if (up) {
      if (i === 0) unshift = true;
      path.splice(i, 1);
      up--;
    }
  }

  if (unshift) path.unshift('');
  if (last === '.' || last === '..') path.push('');

  return path.join('/');
}

/**
 * The actual URL instance. Instead of returning an object we've opted-in to
 * create an actual constructor as it's much more memory efficient and
 * faster and it pleases my OCD.
 *
 * It is worth noting that we should not use `URL` as class name to prevent
 * clashes with the global URL instance that got introduced in browsers.
 *
 * @constructor
 * @param {String} address URL we want to parse.
 * @param {Object|String} [location] Location defaults for relative paths.
 * @param {Boolean|Function} [parser] Parser for the query string.
 * @private
 */
function Url(address, location, parser) {
  address = trimLeft(address);

  if (!(this instanceof Url)) {
    return new Url(address, location, parser);
  }

  var relative, extracted, parse, instruction, index, key
    , instructions = rules.slice()
    , type = typeof location
    , url = this
    , i = 0;

  //
  // The following if statements allows this module two have compatibility with
  // 2 different API:
  //
  // 1. Node.js's `url.parse` api which accepts a URL, boolean as arguments
  //    where the boolean indicates that the query string should also be parsed.
  //
  // 2. The `URL` interface of the browser which accepts a URL, object as
  //    arguments. The supplied object will be used as default values / fall-back
  //    for relative paths.
  //
  if ('object' !== type && 'string' !== type) {
    parser = location;
    location = null;
  }

  if (parser && 'function' !== typeof parser) parser = qs.parse;

  location = lolcation(location);

  //
  // Extract protocol information before running the instructions.
  //
  extracted = extractProtocol(address || '');
  relative = !extracted.protocol && !extracted.slashes;
  url.slashes = extracted.slashes || relative && location.slashes;
  url.protocol = extracted.protocol || location.protocol || '';
  address = extracted.rest;

  //
  // When the authority component is absent the URL starts with a path
  // component.
  //
  if (!extracted.slashes || url.protocol === 'file:') {
    instructions[3] = [/(.*)/, 'pathname'];
  }

  for (; i < instructions.length; i++) {
    instruction = instructions[i];

    if (typeof instruction === 'function') {
      address = instruction(address);
      continue;
    }

    parse = instruction[0];
    key = instruction[1];

    if (parse !== parse) {
      url[key] = address;
    } else if ('string' === typeof parse) {
      if (~(index = address.indexOf(parse))) {
        if ('number' === typeof instruction[2]) {
          url[key] = address.slice(0, index);
          address = address.slice(index + instruction[2]);
        } else {
          url[key] = address.slice(index);
          address = address.slice(0, index);
        }
      }
    } else if ((index = parse.exec(address))) {
      url[key] = index[1];
      address = address.slice(0, index.index);
    }

    url[key] = url[key] || (
      relative && instruction[3] ? location[key] || '' : ''
    );

    //
    // Hostname, host and protocol should be lowercased so they can be used to
    // create a proper `origin`.
    //
    if (instruction[4]) url[key] = url[key].toLowerCase();
  }

  //
  // Also parse the supplied query string in to an object. If we're supplied
  // with a custom parser as function use that instead of the default build-in
  // parser.
  //
  if (parser) url.query = parser(url.query);

  //
  // If the URL is relative, resolve the pathname against the base URL.
  //
  if (
      relative
    && location.slashes
    && url.pathname.charAt(0) !== '/'
    && (url.pathname !== '' || location.pathname !== '')
  ) {
    url.pathname = resolve(url.pathname, location.pathname);
  }

  //
  // Default to a / for pathname if none exists. This normalizes the URL
  // to always have a /
  //
  if (
      url.pathname.charAt(0) !== '/'
    && (url.hostname || url.protocol === 'file:')
  ) {
    url.pathname = '/' + url.pathname;
  }

  //
  // We should not add port numbers if they are already the default port number
  // for a given protocol. As the host also contains the port number we're going
  // override it with the hostname which contains no port number.
  //
  if (!required(url.port, url.protocol)) {
    url.host = url.hostname;
    url.port = '';
  }

  //
  // Parse down the `auth` for the username and password.
  //
  url.username = url.password = '';
  if (url.auth) {
    instruction = url.auth.split(':');
    url.username = instruction[0] || '';
    url.password = instruction[1] || '';
  }

  url.origin = url.protocol && url.host && url.protocol !== 'file:'
    ? url.protocol +'//'+ url.host
    : 'null';

  //
  // The href is just the compiled result.
  //
  url.href = url.toString();
}

/**
 * This is convenience method for changing properties in the URL instance to
 * insure that they all propagate correctly.
 *
 * @param {String} part          Property we need to adjust.
 * @param {Mixed} value          The newly assigned value.
 * @param {Boolean|Function} fn  When setting the query, it will be the function
 *                               used to parse the query.
 *                               When setting the protocol, double slash will be
 *                               removed from the final url if it is true.
 * @returns {URL} URL instance for chaining.
 * @public
 */
function set(part, value, fn) {
  var url = this;

  switch (part) {
    case 'query':
      if ('string' === typeof value && value.length) {
        value = (fn || qs.parse)(value);
      }

      url[part] = value;
      break;

    case 'port':
      url[part] = value;

      if (!required(value, url.protocol)) {
        url.host = url.hostname;
        url[part] = '';
      } else if (value) {
        url.host = url.hostname +':'+ value;
      }

      break;

    case 'hostname':
      url[part] = value;

      if (url.port) value += ':'+ url.port;
      url.host = value;
      break;

    case 'host':
      url[part] = value;

      if (/:\d+$/.test(value)) {
        value = value.split(':');
        url.port = value.pop();
        url.hostname = value.join(':');
      } else {
        url.hostname = value;
        url.port = '';
      }

      break;

    case 'protocol':
      url.protocol = value.toLowerCase();
      url.slashes = !fn;
      break;

    case 'pathname':
    case 'hash':
      if (value) {
        var char = part === 'pathname' ? '/' : '#';
        url[part] = value.charAt(0) !== char ? char + value : value;
      } else {
        url[part] = value;
      }
      break;

    default:
      url[part] = value;
  }

  for (var i = 0; i < rules.length; i++) {
    var ins = rules[i];

    if (ins[4]) url[ins[1]] = url[ins[1]].toLowerCase();
  }

  url.origin = url.protocol && url.host && url.protocol !== 'file:'
    ? url.protocol +'//'+ url.host
    : 'null';

  url.href = url.toString();

  return url;
}

/**
 * Transform the properties back in to a valid and full URL string.
 *
 * @param {Function} stringify Optional query stringify function.
 * @returns {String} Compiled version of the URL.
 * @public
 */
function toString(stringify) {
  if (!stringify || 'function' !== typeof stringify) stringify = qs.stringify;

  var query
    , url = this
    , protocol = url.protocol;

  if (protocol && protocol.charAt(protocol.length - 1) !== ':') protocol += ':';

  var result = protocol + (url.slashes || url.protocol === 'file:' ? '//' : '');

  if (url.username) {
    result += url.username;
    if (url.password) result += ':'+ url.password;
    result += '@';
  }

  result += url.host + url.pathname;

  query = 'object' === typeof url.query ? stringify(url.query) : url.query;
  if (query) result += '?' !== query.charAt(0) ? '?'+ query : query;

  if (url.hash) result += url.hash;

  return result;
}

Url.prototype = { set: set, toString: toString };

//
// Expose the URL parser and some additional properties that might be useful for
// others or testing.
//
Url.extractProtocol = extractProtocol;
Url.location = lolcation;
Url.trimLeft = trimLeft;
Url.qs = qs;

module.exports = Url;