resources/src/mediawiki/mediawiki.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 ( mw, $ ) {
  54         /**
  55          * Function that's useful when constructing the URI string -- we frequently encounter the pattern
  56          * of having to add something to the URI as we go, but only if it's present, and to include a
  57          * character before or after if so.
  58          *
  59          * @private
  60          * @static
  61          * @param {string|undefined} pre To prepend
  62          * @param {string} val To include
  63          * @param {string} post To append
  64          * @param {boolean} raw If true, val will not be encoded
  65          * @return {string} Result
  66          */
  67         function cat( pre, val, post, raw ) {
  68                 if ( val === undefined || val === null || val === '' ) {
  69                         return '';
  70                 }
  71
  72                 return pre + ( raw ? val : mw.Uri.encode( val ) ) + post;
  73         }
  74
  75         /**
  76          * Regular expressions to parse many common URIs.
  77          *
  78          * As they are gnarly, they have been moved to separate files to allow us to format them in the
  79          * 'extended' regular expression format (which JavaScript normally doesn't support). The subset of
  80          * features handled is minimal, but just the free whitespace gives us a lot.
  81          *
  82          * @private
  83          * @static
  84          * @property {Object} parser
  85          */
  86         var parser = {
  87                 strict: mw.template.get( 'mediawiki.Uri', 'strict.regexp' ).render(),
  88                 loose: mw.template.get( 'mediawiki.Uri', 'loose.regexp' ).render()
  89         },
  90
  91         /**
  92          * The order here matches the order of captured matches in the `parser` property regexes.
  93          *
  94          * @private
  95          * @static
  96          * @property {Array} properties
  97          */
  98         properties = [
  99                 'protocol',
 100                 'user',
 101                 'password',
 102                 'host',
 103                 'port',
 104                 'path',
 105                 'query',
 106                 'fragment'
 107         ];
 108
 109         /**
 110          * @property {string} protocol For example `http` (always present)
 111          */
 112         /**
 113          * @property {string|undefined} user For example `usr`
 114          */
 115         /**
 116          * @property {string|undefined} password For example `pwd`
 117          */
 118         /**
 119          * @property {string} host For example `www.example.com` (always present)
 120          */
 121         /**
 122          * @property {string|undefined} port For example `81`
 123          */
 124         /**
 125          * @property {string} path For example `/dir/dir.2/index.htm` (always present)
 126          */
 127         /**
 128          * @property {Object} query For example `{ a: '0', b: '', c: 'value' }` (always present)
 129          */
 130         /**
 131          * @property {string|undefined} fragment For example `top`
 132          */
 133
 134         /**
 135          * A factory method to create a Uri class with a default location to resolve relative URLs
 136          * against (including protocol-relative URLs).
 137          *
 138          * @method
 139          * @param {string|Function} documentLocation A full url, or function returning one.
 140          *  If passed a function, the return value may change over time and this will be honoured. (T74334)
 141          * @member mw
 142          */
 143         mw.UriRelative = function ( documentLocation ) {
 144                 var getDefaultUri = ( function () {
 145                         // Cache
 146                         var href, uri;
 147
 148                         return function () {
 149                                 var hrefCur = typeof documentLocation === 'string' ? documentLocation : documentLocation();
 150                                 if ( href === hrefCur ) {
 151                                         return uri;
 152                                 }
 153                                 href = hrefCur;
 154                                 uri = new Uri( href );
 155                                 return uri;
 156                         };
 157                 }() );
 158
 159                 /**
 160                  * Construct a new URI object. Throws error if arguments are illegal/impossible, or
 161                  * otherwise don't parse.
 162                  *
 163                  * @class mw.Uri
 164                  * @constructor
 165                  * @param {Object|string} [uri] URI string, or an Object with appropriate properties (especially
 166                  *  another URI object to clone). Object must have non-blank `protocol`, `host`, and `path`
 167                  *  properties. If omitted (or set to `undefined`, `null` or empty string), then an object
 168                  *  will be created for the default `uri` of this constructor (`location.href` for mw.Uri,
 169                  *  other values for other instances -- see mw.UriRelative for details).
 170                  * @param {Object|boolean} [options] Object with options, or (backwards compatibility) a boolean
 171                  *  for strictMode
 172                  * @param {boolean} [options.strictMode=false] Trigger strict mode parsing of the url.
 173                  * @param {boolean} [options.overrideKeys=false] Whether to let duplicate query parameters
 174                  *  override each other (`true`) or automagically convert them to an array (`false`).
 175                  */
 176                 function Uri( uri, options ) {
 177                         var prop,
 178                                 defaultUri = getDefaultUri();
 179
 180                         options = typeof options === 'object' ? options : { strictMode: !!options };
 181                         options = $.extend( {
 182                                 strictMode: false,
 183                                 overrideKeys: false
 184                         }, options );
 185
 186                         if ( uri !== undefined && uri !== null && uri !== '' ) {
 187                                 if ( typeof uri === 'string' ) {
 188                                         this.parse( uri, options );
 189                                 } else if ( typeof uri === 'object' ) {
 190                                         // Copy data over from existing URI object
 191                                         for ( prop in uri ) {
 192                                                 // Only copy direct properties, not inherited ones
 193                                                 if ( uri.hasOwnProperty( prop ) ) {
 194                                                         // Deep copy object properties
 195                                                         if ( $.isArray( uri[ prop ] ) || $.isPlainObject( uri[ prop ] ) ) {
 196                                                                 this[ prop ] = $.extend( true, {}, uri[ prop ] );
 197                                                         } else {
 198                                                                 this[ prop ] = uri[ prop ];
 199                                                         }
 200                                                 }
 201                                         }
 202                                         if ( !this.query ) {
 203                                                 this.query = {};
 204                                         }
 205                                 }
 206                         } else {
 207                                 // If we didn't get a URI in the constructor, use the default one.
 208                                 return defaultUri.clone();
 209                         }
 210
 211                         // protocol-relative URLs
 212                         if ( !this.protocol ) {
 213                                 this.protocol = defaultUri.protocol;
 214                         }
 215                         // No host given:
 216                         if ( !this.host ) {
 217                                 this.host = defaultUri.host;
 218                                 // port ?
 219                                 if ( !this.port ) {
 220                                         this.port = defaultUri.port;
 221                                 }
 222                         }
 223                         if ( this.path && this.path[ 0 ] !== '/' ) {
 224                                 // A real relative URL, relative to defaultUri.path. We can't really handle that since we cannot
 225                                 // figure out whether the last path component of defaultUri.path is a directory or a file.
 226                                 throw new Error( 'Bad constructor arguments' );
 227                         }
 228                         if ( !( this.protocol && this.host && this.path ) ) {
 229                                 throw new Error( 'Bad constructor arguments' );
 230                         }
 231                 }
 232
 233                 /**
 234                  * Encode a value for inclusion in a url.
 235                  *
 236                  * Standard encodeURIComponent, with extra stuff to make all browsers work similarly and more
 237                  * compliant with RFC 3986. Similar to rawurlencode from PHP and our JS library
 238                  * mw.util.rawurlencode, except this also replaces spaces with `+`.
 239                  *
 240                  * @static
 241                  * @param {string} s String to encode
 242                  * @return {string} Encoded string for URI
 243                  */
 244                 Uri.encode = function ( s ) {
 245                         return encodeURIComponent( s )
 246                                 .replace( /!/g, '%21' ).replace( /'/g, '%27' ).replace( /\(/g, '%28' )
 247                                 .replace( /\)/g, '%29' ).replace( /\*/g, '%2A' )
 248                                 .replace( /%20/g, '+' );
 249                 };
 250
 251                 /**
 252                  * Decode a url encoded value.
 253                  *
 254                  * Reversed #encode. Standard decodeURIComponent, with addition of replacing
 255                  * `+` with a space.
 256                  *
 257                  * @static
 258                  * @param {string} s String to decode
 259                  * @return {string} Decoded string
 260                  */
 261                 Uri.decode = function ( s ) {
 262                         return decodeURIComponent( s.replace( /\+/g, '%20' ) );
 263                 };
 264
 265                 Uri.prototype = {
 266
 267                         /**
 268                          * Parse a string and set our properties accordingly.
 269                          *
 270                          * @private
 271                          * @param {string} str URI, see constructor.
 272                          * @param {Object} options See constructor.
 273                          */
 274                         parse: function ( str, options ) {
 275                                 var q, matches,
 276                                         uri = this,
 277                                         hasOwn = Object.prototype.hasOwnProperty;
 278
 279                                 // Apply parser regex and set all properties based on the result
 280                                 matches = parser[ options.strictMode ? 'strict' : 'loose' ].exec( str );
 281                                 $.each( properties, function ( i, property ) {
 282                                         uri[ property ] = matches[ i + 1 ];
 283                                 } );
 284
 285                                 // uri.query starts out as the query string; we will parse it into key-val pairs then make
 286                                 // that object the "query" property.
 287                                 // we overwrite query in uri way to make cloning easier, it can use the same list of properties.
 288                                 q = {};
 289                                 // using replace to iterate over a string
 290                                 if ( uri.query ) {
 291                                         uri.query.replace( /(?:^|&)([^&=]*)(?:(=)([^&]*))?/g, function ( $0, $1, $2, $3 ) {
 292                                                 var k, v;
 293                                                 if ( $1 ) {
 294                                                         k = Uri.decode( $1 );
 295                                                         v = ( $2 === '' || $2 === undefined ) ? null : Uri.decode( $3 );
 296
 297                                                         // If overrideKeys, always (re)set top level value.
 298                                                         // If not overrideKeys but this key wasn't set before, then we set it as well.
 299                                                         if ( options.overrideKeys || !hasOwn.call( q, k ) ) {
 300                                                                 q[ k ] = v;
 301
 302                                                         // Use arrays if overrideKeys is false and key was already seen before
 303                                                         } else {
 304                                                                 // Once before, still a string, turn into an array
 305                                                                 if ( typeof q[ k ] === 'string' ) {
 306                                                                         q[ k ] = [ q[ k ] ];
 307                                                                 }
 308                                                                 // Add to the array
 309                                                                 if ( $.isArray( q[ k ] ) ) {
 310                                                                         q[ k ].push( v );
 311                                                                 }
 312                                                         }
 313                                                 }
 314                                         } );
 315                                 }
 316                                 uri.query = q;
 317                         },
 318
 319                         /**
 320                          * Get user and password section of a URI.
 321                          *
 322                          * @return {string}
 323                          */
 324                         getUserInfo: function () {
 325                                 return cat( '', this.user, cat( ':', this.password, '' ) );
 326                         },
 327
 328                         /**
 329                          * Get host and port section of a URI.
 330                          *
 331                          * @return {string}
 332                          */
 333                         getHostPort: function () {
 334                                 return this.host + cat( ':', this.port, '' );
 335                         },
 336
 337                         /**
 338                          * Get the userInfo, host and port section of the URI.
 339                          *
 340                          * In most real-world URLs this is simply the hostname, but the definition of 'authority' section is more general.
 341                          *
 342                          * @return {string}
 343                          */
 344                         getAuthority: function () {
 345                                 return cat( '', this.getUserInfo(), '@' ) + this.getHostPort();
 346                         },
 347
 348                         /**
 349                          * Get the query arguments of the URL, encoded into a string.
 350                          *
 351                          * Does not preserve the original order of arguments passed in the URI. Does handle escaping.
 352                          *
 353                          * @return {string}
 354                          */
 355                         getQueryString: function () {
 356                                 var args = [];
 357                                 $.each( this.query, function ( key, val ) {
 358                                         var k = Uri.encode( key ),
 359                                                 vals = $.isArray( val ) ? val : [ val ];
 360                                         $.each( vals, function ( i, v ) {
 361                                                 if ( v === null ) {
 362                                                         args.push( k );
 363                                                 } else if ( k === 'title' ) {
 364                                                         args.push( k + '=' + mw.util.wikiUrlencode( v ) );
 365                                                 } else {
 366                                                         args.push( k + '=' + Uri.encode( v ) );
 367                                                 }
 368                                         } );
 369                                 } );
 370                                 return args.join( '&' );
 371                         },
 372
 373                         /**
 374                          * Get everything after the authority section of the URI.
 375                          *
 376                          * @return {string}
 377                          */
 378                         getRelativePath: function () {
 379                                 return this.path + cat( '?', this.getQueryString(), '', true ) + cat( '#', this.fragment, '' );
 380                         },
 381
 382                         /**
 383                          * Get the entire URI string.
 384                          *
 385                          * May not be precisely the same as input due to order of query arguments.
 386                          *
 387                          * @return {string} The URI string
 388                          */
 389                         toString: function () {
 390                                 return this.protocol + '://' + this.getAuthority() + this.getRelativePath();
 391                         },
 392
 393                         /**
 394                          * Clone this URI
 395                          *
 396                          * @return {Object} New URI object with same properties
 397                          */
 398                         clone: function () {
 399                                 return new Uri( this );
 400                         },
 401
 402                         /**
 403                          * Extend the query section of the URI with new parameters.
 404                          *
 405                          * @param {Object} parameters Query parameters to add to ours (or to override ours with) as an
 406                          *  object
 407                          * @return {Object} This URI object
 408                          */
 409                         extend: function ( parameters ) {
 410                                 $.extend( this.query, parameters );
 411                                 return this;
 412                         }
 413                 };
 414
 415                 return Uri;
 416         };
 417
 418         // Default to the current browsing location (for relative URLs).
 419         mw.Uri = mw.UriRelative( function () {
 420                 return location.href;
 421         } );
 422
 423 }( mediaWiki, jQuery ) );