resources/src/mediawiki.Uri/Uri.js

   1 /**
   2  * Library for simple URI parsing and manipulation.
   3  *
   4  * Intended to be minimal, but featureful; do not expect full RFC 3986 compliance. The use cases we
   5  * have in mind are constructing 'next page' or 'previous page' URLs, detecting whether we need to
   6  * use cross-domain proxies for an API, constructing simple URL-based API calls, etc. Parsing here
   7  * is regex-based, so may not work on all URIs, but is good enough for most.
   8  *
   9  * You can modify the properties directly, then use the #toString method to extract the full URI
  10  * string again. Example:
  11  *
  12  *     var uri = new mw.Uri( 'http://example.com/mysite/mypage.php?quux=2' );
  13  *
  14  *     if ( uri.host == 'example.com' ) {
  15  *         uri.host = 'foo.example.com';
  16  *         uri.extend( { bar: 1 } );
  17  *
  18  *         $( 'a#id1' ).attr( 'href', uri );
  19  *         // anchor with id 'id1' now links to http://foo.example.com/mysite/mypage.php?bar=1&quux=2
  20  *
  21  *         $( 'a#id2' ).attr( 'href', uri.clone().extend( { bar: 3, pif: 'paf' } ) );
  22  *         // anchor with id 'id2' now links to http://foo.example.com/mysite/mypage.php?bar=3&quux=2&pif=paf
  23  *     }
  24  *
  25  * Given a URI like
  26  * `http://usr:pwd@www.example.com:81/dir/dir.2/index.htm?q1=0&&test1&test2=&test3=value+%28escaped%29&r=1&r=2#top`
  27  * the returned object will have the following properties:
  28  *
  29  *     protocol  'http'
  30  *     user      'usr'
  31  *     password  'pwd'
  32  *     host      'www.example.com'
  33  *     port      '81'
  34  *     path      '/dir/dir.2/index.htm'
  35  *     query     {
  36  *                   q1: '0',
  37  *                   test1: null,
  38  *                   test2: '',
  39  *                   test3: 'value (escaped)'
  40  *                   r: ['1', '2']
  41  *               }
  42  *     fragment  'top'
  43  *
  44  * (N.b., 'password' is technically not allowed for HTTP URIs, but it is possible with other kinds
  45  * of URIs.)
  46  *
  47  * Parsing based on parseUri 1.2.2 (c) Steven Levithan <http://stevenlevithan.com>, MIT License.
  48  * <http://stevenlevithan.com/demo/parseuri/js/>
  49  *
  50  * @class mw.Uri
  51  */
  52
  53 ( function () {
  54         var parser, properties;
  55
  56         /**
  57          * Function that's useful when constructing the URI string -- we frequently encounter the pattern
  58          * of having to add something to the URI as we go, but only if it's present, and to include a
  59          * character before or after if so.
  60          *
  61          * @private
  62          * @static
  63          * @param {string|undefined} pre To prepend
  64          * @param {string} val To include
  65          * @param {string} post To append
  66          * @param {boolean} raw If true, val will not be encoded
  67          * @return {string} Result
  68          */
  69         function cat( pre, val, post, raw ) {
  70                 if ( val === undefined || val === null || val === '' ) {
  71                         return '';
  72                 }
  73
  74                 return pre + ( raw ? val : mw.Uri.encode( val ) ) + post;
  75         }
  76
  77         /**
  78          * Regular expressions to parse many common URIs.
  79          *
  80          * As they are gnarly, they have been moved to separate files to allow us to format them in the
  81          * 'extended' regular expression format (which JavaScript normally doesn't support). The subset of
  82          * features handled is minimal, but just the free whitespace gives us a lot.
  83          *
  84          * @private
  85          * @static
  86          * @property {Object} parser
  87          */
  88         parser = {
  89                 strict: mw.template.get( 'mediawiki.Uri', 'strict.regexp' ).render(),
  90                 loose: mw.template.get( 'mediawiki.Uri', 'loose.regexp' ).render()
  91         };
  92
  93         /**
  94          * The order here matches the order of captured matches in the `parser` property regexes.
  95          *
  96          * @private
  97          * @static
  98          * @property {string[]} properties
  99          */
 100         properties = [
 101                 'protocol',
 102                 'user',
 103                 'password',
 104                 'host',
 105                 'port',
 106                 'path',
 107                 'query',
 108                 'fragment'
 109         ];
 110
 111         /**
 112          * @property {string} protocol For example `http` (always present)
 113          */
 114         /**
 115          * @property {string|undefined} user For example `usr`
 116          */
 117         /**
 118          * @property {string|undefined} password For example `pwd`
 119          */
 120         /**
 121          * @property {string} host For example `www.example.com` (always present)
 122          */
 123         /**
 124          * @property {string|undefined} port For example `81`
 125          */
 126         /**
 127          * @property {string} path For example `/dir/dir.2/index.htm` (always present)
 128          */
 129         /**
 130          * @property {Object} query For example `{ a: '0', b: '', c: 'value' }` (always present)
 131          */
 132         /**
 133          * @property {string|undefined} fragment For example `top`
 134          */
 135
 136         /**
 137          * A factory method to create an mw.Uri class with a default location to resolve relative URLs
 138          * against (including protocol-relative URLs).
 139          *
 140          * @method
 141          * @param {string|Function} documentLocation A full url, or function returning one.
 142          *  If passed a function, the return value may change over time and this will be honoured. (T74334)
 143          * @member mw
 144          * @return {Function} An mw.Uri class constructor
 145          */
 146         mw.UriRelative = function ( documentLocation ) {
 147                 var getDefaultUri = ( function () {
 148                         // Cache
 149                         var href, uri;
 150
 151                         return function () {
 152                                 var hrefCur = typeof documentLocation === 'string' ? documentLocation : documentLocation();
 153                                 if ( href === hrefCur ) {
 154                                         return uri;
 155                                 }
 156                                 href = hrefCur;
 157                                 // eslint-disable-next-line no-use-before-define
 158                                 uri = new Uri( href );
 159                                 return uri;
 160                         };
 161                 }() );
 162
 163                 /**
 164                  * Construct a new URI object. Throws error if arguments are illegal/impossible, or
 165                  * otherwise don't parse.
 166                  *
 167                  * @class mw.Uri
 168                  * @constructor
 169                  * @param {Object|string} [uri] URI string, or an Object with appropriate properties (especially
 170                  *  another URI object to clone). Object must have non-blank `protocol`, `host`, and `path`
 171                  *  properties. If omitted (or set to `undefined`, `null` or empty string), then an object
 172                  *  will be created for the default `uri` of this constructor (`location.href` for mw.Uri,
 173                  *  other values for other instances -- see mw.UriRelative for details).
 174                  * @param {Object|boolean} [options] Object with options, or (backwards compatibility) a boolean
 175                  *  for strictMode
 176                  * @param {boolean} [options.strictMode=false] Trigger strict mode parsing of the url.
 177                  * @param {boolean} [options.overrideKeys=false] Whether to let duplicate query parameters
 178                  *  override each other (`true`) or automagically convert them to an array (`false`).
 179                  * @param {boolean} [options.arrayParams=false] Whether to parse array query parameters (e.g.
 180                  *  `&foo[0]=a&foo[1]=b` or `&foo[]=a&foo[]=b`) or leave them alone. Currently this does not
 181                  *  handle associative or multi-dimensional arrays, but that may be improved in the future.
 182                  *  Implies `overrideKeys: true` (query parameters without `[...]` are not parsed as arrays).
 183                  * @throws {Error} when the query string or fragment contains an unknown % sequence
 184                  */
 185                 function Uri( uri, options ) {
 186                         var prop, hrefCur,
 187                                 hasOptions = ( options !== undefined ),
 188                                 defaultUri = getDefaultUri();
 189
 190                         options = typeof options === 'object' ? options : { strictMode: !!options };
 191                         options = $.extend( {
 192                                 strictMode: false,
 193                                 overrideKeys: false,
 194                                 arrayParams: false
 195                         }, options );
 196
 197                         this.arrayParams = options.arrayParams;
 198
 199                         if ( uri !== undefined && uri !== null && uri !== '' ) {
 200                                 if ( typeof uri === 'string' ) {
 201                                         this.parse( uri, options );
 202                                 } else if ( typeof uri === 'object' ) {
 203                                         // Copy data over from existing URI object
 204                                         for ( prop in uri ) {
 205                                                 // Only copy direct properties, not inherited ones
 206                                                 if ( Object.prototype.hasOwnProperty.call( uri, prop ) ) {
 207                                                         // Deep copy object properties
 208                                                         if ( Array.isArray( uri[ prop ] ) || $.isPlainObject( uri[ prop ] ) ) {
 209                                                                 this[ prop ] = $.extend( true, {}, uri[ prop ] );
 210                                                         } else {
 211                                                                 this[ prop ] = uri[ prop ];
 212                                                         }
 213                                                 }
 214                                         }
 215                                         if ( !this.query ) {
 216                                                 this.query = {};
 217                                         }
 218                                 }
 219                         } else if ( hasOptions ) {
 220                                 // We didn't get a URI in the constructor, but we got options.
 221                                 hrefCur = typeof documentLocation === 'string' ? documentLocation : documentLocation();
 222                                 this.parse( hrefCur, options );
 223                         } else {
 224                                 // We didn't get a URI or options in the constructor, use the default instance.
 225                                 return defaultUri.clone();
 226                         }
 227
 228                         // protocol-relative URLs
 229                         if ( !this.protocol ) {
 230                                 this.protocol = defaultUri.protocol;
 231                         }
 232                         // No host given:
 233                         if ( !this.host ) {
 234                                 this.host = defaultUri.host;
 235                                 // port ?
 236                                 if ( !this.port ) {
 237                                         this.port = defaultUri.port;
 238                                 }
 239                         }
 240                         if ( this.path && this.path[ 0 ] !== '/' ) {
 241                                 // A real relative URL, relative to defaultUri.path. We can't really handle that since we cannot
 242                                 // figure out whether the last path component of defaultUri.path is a directory or a file.
 243                                 throw new Error( 'Bad constructor arguments' );
 244                         }
 245                         if ( !( this.protocol && this.host && this.path ) ) {
 246                                 throw new Error( 'Bad constructor arguments' );
 247                         }
 248                 }
 249
 250                 /**
 251                  * Encode a value for inclusion in a url.
 252                  *
 253                  * Standard encodeURIComponent, with extra stuff to make all browsers work similarly and more
 254                  * compliant with RFC 3986. Similar to rawurlencode from PHP and our JS library
 255                  * mw.util.rawurlencode, except this also replaces spaces with `+`.
 256                  *
 257                  * @static
 258                  * @param {string} s String to encode
 259                  * @return {string} Encoded string for URI
 260                  */
 261                 Uri.encode = function ( s ) {
 262                         return encodeURIComponent( s )
 263                                 .replace( /!/g, '%21' ).replace( /'/g, '%27' ).replace( /\(/g, '%28' )
 264                                 .replace( /\)/g, '%29' ).replace( /\*/g, '%2A' )
 265                                 .replace( /%20/g, '+' );
 266                 };
 267
 268                 /**
 269                  * Decode a url encoded value.
 270                  *
 271                  * Reversed #encode. Standard decodeURIComponent, with addition of replacing
 272                  * `+` with a space.
 273                  *
 274                  * @static
 275                  * @param {string} s String to decode
 276                  * @return {string} Decoded string
 277                  * @throws {Error} when the string contains an unknown % sequence
 278                  */
 279                 Uri.decode = function ( s ) {
 280                         return decodeURIComponent( s.replace( /\+/g, '%20' ) );
 281                 };
 282
 283                 Uri.prototype = {
 284
 285                         /**
 286                          * Parse a string and set our properties accordingly.
 287                          *
 288                          * @private
 289                          * @param {string} str URI, see constructor.
 290                          * @param {Object} options See constructor.
 291                          * @throws {Error} when the query string or fragment contains an unknown % sequence
 292                          */
 293                         parse: function ( str, options ) {
 294                                 var q, matches,
 295                                         uri = this,
 296                                         hasOwn = Object.prototype.hasOwnProperty;
 297
 298                                 // Apply parser regex and set all properties based on the result
 299                                 matches = parser[ options.strictMode ? 'strict' : 'loose' ].exec( str );
 300                                 properties.forEach( function ( property, i ) {
 301                                         uri[ property ] = matches[ i + 1 ];
 302                                 } );
 303
 304                                 // uri.query starts out as the query string; we will parse it into key-val pairs then make
 305                                 // that object the "query" property.
 306                                 // we overwrite query in uri way to make cloning easier, it can use the same list of properties.
 307                                 q = {};
 308                                 // using replace to iterate over a string
 309                                 if ( uri.query ) {
 310                                         uri.query.replace( /(?:^|&)([^&=]*)(?:(=)([^&]*))?/g, function ( match, k, eq, v ) {
 311                                                 var arrayKeyMatch, i;
 312                                                 if ( k ) {
 313                                                         k = Uri.decode( k );
 314                                                         v = ( eq === '' || eq === undefined ) ? null : Uri.decode( v );
 315                                                         arrayKeyMatch = k.match( /^([^[]+)\[(\d*)\]$/ );
 316
 317                                                         // If arrayParams and this parameter name contains an array index...
 318                                                         if ( options.arrayParams && arrayKeyMatch ) {
 319                                                                 // Remove the index from parameter name
 320                                                                 k = arrayKeyMatch[ 1 ];
 321
 322                                                                 // Turn the parameter value into an array (throw away anything else)
 323                                                                 if ( !Array.isArray( q[ k ] ) ) {
 324                                                                         q[ k ] = [];
 325                                                                 }
 326
 327                                                                 i = arrayKeyMatch[ 2 ];
 328                                                                 if ( i === '' ) {
 329                                                                         // If no explicit index, append at the end
 330                                                                         i = q[ k ].length;
 331                                                                 }
 332
 333                                                                 q[ k ][ i ] = v;
 334
 335                                                         // If overrideKeys, always (re)set top level value.
 336                                                         // If not overrideKeys but this key wasn't set before, then we set it as well.
 337                                                         // arrayParams implies overrideKeys (no array handling for non-array params).
 338                                                         } else if ( options.arrayParams || options.overrideKeys || !hasOwn.call( q, k ) ) {
 339                                                                 q[ k ] = v;
 340
 341                                                         // Use arrays if overrideKeys is false and key was already seen before
 342                                                         } else {
 343                                                                 // Once before, still a string, turn into an array
 344                                                                 if ( typeof q[ k ] === 'string' ) {
 345                                                                         q[ k ] = [ q[ k ] ];
 346                                                                 }
 347                                                                 // Add to the array
 348                                                                 if ( Array.isArray( q[ k ] ) ) {
 349                                                                         q[ k ].push( v );
 350                                                                 }
 351                                                         }
 352                                                 }
 353                                         } );
 354                                 }
 355                                 uri.query = q;
 356
 357                                 // Decode uri.fragment, otherwise it gets double-encoded when serializing
 358                                 if ( uri.fragment !== undefined ) {
 359                                         uri.fragment = Uri.decode( uri.fragment );
 360                                 }
 361                         },
 362
 363                         /**
 364                          * Get user and password section of a URI.
 365                          *
 366                          * @return {string}
 367                          */
 368                         getUserInfo: function () {
 369                                 return cat( '', this.user, cat( ':', this.password, '' ) );
 370                         },
 371
 372                         /**
 373                          * Get host and port section of a URI.
 374                          *
 375                          * @return {string}
 376                          */
 377                         getHostPort: function () {
 378                                 return this.host + cat( ':', this.port, '' );
 379                         },
 380
 381                         /**
 382                          * Get the userInfo, host and port section of the URI.
 383                          *
 384                          * In most real-world URLs this is simply the hostname, but the definition of 'authority' section is more general.
 385                          *
 386                          * @return {string}
 387                          */
 388                         getAuthority: function () {
 389                                 return cat( '', this.getUserInfo(), '@' ) + this.getHostPort();
 390                         },
 391
 392                         /**
 393                          * Get the query arguments of the URL, encoded into a string.
 394                          *
 395                          * Does not preserve the original order of arguments passed in the URI. Does handle escaping.
 396                          *
 397                          * @return {string}
 398                          */
 399                         getQueryString: function () {
 400                                 var args = [],
 401                                         arrayParams = this.arrayParams;
 402                                 // eslint-disable-next-line no-jquery/no-each-util
 403                                 $.each( this.query, function ( key, val ) {
 404                                         var k = Uri.encode( key ),
 405                                                 isArrayParam = Array.isArray( val ),
 406                                                 vals = isArrayParam ? val : [ val ];
 407                                         vals.forEach( function ( v, i ) {
 408                                                 var ki = k;
 409                                                 if ( arrayParams && isArrayParam ) {
 410                                                         ki += Uri.encode( '[' + i + ']' );
 411                                                 }
 412                                                 if ( v === null ) {
 413                                                         args.push( ki );
 414                                                 } else if ( k === 'title' ) {
 415                                                         args.push( ki + '=' + mw.util.wikiUrlencode( v ) );
 416                                                 } else {
 417                                                         args.push( ki + '=' + Uri.encode( v ) );
 418                                                 }
 419                                         } );
 420                                 } );
 421                                 return args.join( '&' );
 422                         },
 423
 424                         /**
 425                          * Get everything after the authority section of the URI.
 426                          *
 427                          * @return {string}
 428                          */
 429                         getRelativePath: function () {
 430                                 return this.path + cat( '?', this.getQueryString(), '', true ) + cat( '#', this.fragment, '' );
 431                         },
 432
 433                         /**
 434                          * Get the entire URI string.
 435                          *
 436                          * May not be precisely the same as input due to order of query arguments.
 437                          *
 438                          * @return {string} The URI string
 439                          */
 440                         toString: function () {
 441                                 return this.protocol + '://' + this.getAuthority() + this.getRelativePath();
 442                         },
 443
 444                         /**
 445                          * Clone this URI
 446                          *
 447                          * @return {Object} New URI object with same properties
 448                          */
 449                         clone: function () {
 450                                 return new Uri( this );
 451                         },
 452
 453                         /**
 454                          * Extend the query section of the URI with new parameters.
 455                          *
 456                          * @param {Object} parameters Query parameters to add to ours (or to override ours with) as an
 457                          *  object
 458                          * @return {Object} This URI object
 459                          */
 460                         extend: function ( parameters ) {
 461                                 $.extend( this.query, parameters );
 462                                 return this;
 463                         }
 464                 };
 465
 466                 return Uri;
 467         };
 468
 469         // Default to the current browsing location (for relative URLs).
 470         mw.Uri = mw.UriRelative( function () {
 471                 return location.href;
 472         } );
 473
 474 }() );