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 /* eslint-disable no-use-before-define */
  54
  55 ( function ( mw, $ ) {
  56         var parser, properties;
  57
  58         /**
  59          * Function that's useful when constructing the URI string -- we frequently encounter the pattern
  60          * of having to add something to the URI as we go, but only if it's present, and to include a
  61          * character before or after if so.
  62          *
  63          * @private
  64          * @static
  65          * @param {string|undefined} pre To prepend
  66          * @param {string} val To include
  67          * @param {string} post To append
  68          * @param {boolean} raw If true, val will not be encoded
  69          * @return {string} Result
  70          */
  71         function cat( pre, val, post, raw ) {
  72                 if ( val === undefined || val === null || val === '' ) {
  73                         return '';
  74                 }
  75
  76                 return pre + ( raw ? val : mw.Uri.encode( val ) ) + post;
  77         }
  78
  79         /**
  80          * Regular expressions to parse many common URIs.
  81          *
  82          * As they are gnarly, they have been moved to separate files to allow us to format them in the
  83          * 'extended' regular expression format (which JavaScript normally doesn't support). The subset of
  84          * features handled is minimal, but just the free whitespace gives us a lot.
  85          *
  86          * @private
  87          * @static
  88          * @property {Object} parser
  89          */
  90         parser = {
  91                 strict: mw.template.get( 'mediawiki.Uri', 'strict.regexp' ).render(),
  92                 loose: mw.template.get( 'mediawiki.Uri', 'loose.regexp' ).render()
  93         };
  94
  95         /**
  96          * The order here matches the order of captured matches in the `parser` property regexes.
  97          *
  98          * @private
  99          * @static
 100          * @property {Array} properties
 101          */
 102         properties = [
 103                 'protocol',
 104                 'user',
 105                 'password',
 106                 'host',
 107                 'port',
 108                 'path',
 109                 'query',
 110                 'fragment'
 111         ];
 112
 113         /**
 114          * @property {string} protocol For example `http` (always present)
 115          */
 116         /**
 117          * @property {string|undefined} user For example `usr`
 118          */
 119         /**
 120          * @property {string|undefined} password For example `pwd`
 121          */
 122         /**
 123          * @property {string} host For example `www.example.com` (always present)
 124          */
 125         /**
 126          * @property {string|undefined} port For example `81`
 127          */
 128         /**
 129          * @property {string} path For example `/dir/dir.2/index.htm` (always present)
 130          */
 131         /**
 132          * @property {Object} query For example `{ a: '0', b: '', c: 'value' }` (always present)
 133          */
 134         /**
 135          * @property {string|undefined} fragment For example `top`
 136          */
 137
 138         /**
 139          * A factory method to create a Uri class with a default location to resolve relative URLs
 140          * against (including protocol-relative URLs).
 141          *
 142          * @method
 143          * @param {string|Function} documentLocation A full url, or function returning one.
 144          *  If passed a function, the return value may change over time and this will be honoured. (T74334)
 145          * @member mw
 146          * @return {Function} Uri class
 147          */
 148         mw.UriRelative = function ( documentLocation ) {
 149                 var getDefaultUri = ( function () {
 150                         // Cache
 151                         var href, uri;
 152
 153                         return function () {
 154                                 var hrefCur = typeof documentLocation === 'string' ? documentLocation : documentLocation();
 155                                 if ( href === hrefCur ) {
 156                                         return uri;
 157                                 }
 158                                 href = hrefCur;
 159                                 uri = new Uri( href );
 160                                 return uri;
 161                         };
 162                 }() );
 163
 164                 /**
 165                  * Construct a new URI object. Throws error if arguments are illegal/impossible, or
 166                  * otherwise don't parse.
 167                  *
 168                  * @class mw.Uri
 169                  * @constructor
 170                  * @param {Object|string} [uri] URI string, or an Object with appropriate properties (especially
 171                  *  another URI object to clone). Object must have non-blank `protocol`, `host`, and `path`
 172                  *  properties. If omitted (or set to `undefined`, `null` or empty string), then an object
 173                  *  will be created for the default `uri` of this constructor (`location.href` for mw.Uri,
 174                  *  other values for other instances -- see mw.UriRelative for details).
 175                  * @param {Object|boolean} [options] Object with options, or (backwards compatibility) a boolean
 176                  *  for strictMode
 177                  * @param {boolean} [options.strictMode=false] Trigger strict mode parsing of the url.
 178                  * @param {boolean} [options.overrideKeys=false] Whether to let duplicate query parameters
 179                  *  override each other (`true`) or automagically convert them to an array (`false`).
 180                  */
 181                 function Uri( uri, options ) {
 182                         var prop,
 183                                 defaultUri = getDefaultUri();
 184
 185                         options = typeof options === 'object' ? options : { strictMode: !!options };
 186                         options = $.extend( {
 187                                 strictMode: false,
 188                                 overrideKeys: false
 189                         }, options );
 190
 191                         if ( uri !== undefined && uri !== null && uri !== '' ) {
 192                                 if ( typeof uri === 'string' ) {
 193                                         this.parse( uri, options );
 194                                 } else if ( typeof uri === 'object' ) {
 195                                         // Copy data over from existing URI object
 196                                         for ( prop in uri ) {
 197                                                 // Only copy direct properties, not inherited ones
 198                                                 if ( uri.hasOwnProperty( prop ) ) {
 199                                                         // Deep copy object properties
 200                                                         if ( $.isArray( uri[ prop ] ) || $.isPlainObject( uri[ prop ] ) ) {
 201                                                                 this[ prop ] = $.extend( true, {}, uri[ prop ] );
 202                                                         } else {
 203                                                                 this[ prop ] = uri[ prop ];
 204                                                         }
 205                                                 }
 206                                         }
 207                                         if ( !this.query ) {
 208                                                 this.query = {};
 209                                         }
 210                                 }
 211                         } else {
 212                                 // If we didn't get a URI in the constructor, use the default one.
 213                                 return defaultUri.clone();
 214                         }
 215
 216                         // protocol-relative URLs
 217                         if ( !this.protocol ) {
 218                                 this.protocol = defaultUri.protocol;
 219                         }
 220                         // No host given:
 221                         if ( !this.host ) {
 222                                 this.host = defaultUri.host;
 223                                 // port ?
 224                                 if ( !this.port ) {
 225                                         this.port = defaultUri.port;
 226                                 }
 227                         }
 228                         if ( this.path && this.path[ 0 ] !== '/' ) {
 229                                 // A real relative URL, relative to defaultUri.path. We can't really handle that since we cannot
 230                                 // figure out whether the last path component of defaultUri.path is a directory or a file.
 231                                 throw new Error( 'Bad constructor arguments' );
 232                         }
 233                         if ( !( this.protocol && this.host && this.path ) ) {
 234                                 throw new Error( 'Bad constructor arguments' );
 235                         }
 236                 }
 237
 238                 /**
 239                  * Encode a value for inclusion in a url.
 240                  *
 241                  * Standard encodeURIComponent, with extra stuff to make all browsers work similarly and more
 242                  * compliant with RFC 3986. Similar to rawurlencode from PHP and our JS library
 243                  * mw.util.rawurlencode, except this also replaces spaces with `+`.
 244                  *
 245                  * @static
 246                  * @param {string} s String to encode
 247                  * @return {string} Encoded string for URI
 248                  */
 249                 Uri.encode = function ( s ) {
 250                         return encodeURIComponent( s )
 251                                 .replace( /!/g, '%21' ).replace( /'/g, '%27' ).replace( /\(/g, '%28' )
 252                                 .replace( /\)/g, '%29' ).replace( /\*/g, '%2A' )
 253                                 .replace( /%20/g, '+' );
 254                 };
 255
 256                 /**
 257                  * Decode a url encoded value.
 258                  *
 259                  * Reversed #encode. Standard decodeURIComponent, with addition of replacing
 260                  * `+` with a space.
 261                  *
 262                  * @static
 263                  * @param {string} s String to decode
 264                  * @return {string} Decoded string
 265                  */
 266                 Uri.decode = function ( s ) {
 267                         return decodeURIComponent( s.replace( /\+/g, '%20' ) );
 268                 };
 269
 270                 Uri.prototype = {
 271
 272                         /**
 273                          * Parse a string and set our properties accordingly.
 274                          *
 275                          * @private
 276                          * @param {string} str URI, see constructor.
 277                          * @param {Object} options See constructor.
 278                          */
 279                         parse: function ( str, options ) {
 280                                 var q, matches,
 281                                         uri = this,
 282                                         hasOwn = Object.prototype.hasOwnProperty;
 283
 284                                 // Apply parser regex and set all properties based on the result
 285                                 matches = parser[ options.strictMode ? 'strict' : 'loose' ].exec( str );
 286                                 $.each( properties, function ( i, property ) {
 287                                         uri[ property ] = matches[ i + 1 ];
 288                                 } );
 289
 290                                 // uri.query starts out as the query string; we will parse it into key-val pairs then make
 291                                 // that object the "query" property.
 292                                 // we overwrite query in uri way to make cloning easier, it can use the same list of properties.
 293                                 q = {};
 294                                 // using replace to iterate over a string
 295                                 if ( uri.query ) {
 296                                         uri.query.replace( /(?:^|&)([^&=]*)(?:(=)([^&]*))?/g, function ( $0, $1, $2, $3 ) {
 297                                                 var k, v;
 298                                                 if ( $1 ) {
 299                                                         k = Uri.decode( $1 );
 300                                                         v = ( $2 === '' || $2 === undefined ) ? null : Uri.decode( $3 );
 301
 302                                                         // If overrideKeys, always (re)set top level value.
 303                                                         // If not overrideKeys but this key wasn't set before, then we set it as well.
 304                                                         if ( options.overrideKeys || !hasOwn.call( q, k ) ) {
 305                                                                 q[ k ] = v;
 306
 307                                                         // Use arrays if overrideKeys is false and key was already seen before
 308                                                         } else {
 309                                                                 // Once before, still a string, turn into an array
 310                                                                 if ( typeof q[ k ] === 'string' ) {
 311                                                                         q[ k ] = [ q[ k ] ];
 312                                                                 }
 313                                                                 // Add to the array
 314                                                                 if ( $.isArray( q[ k ] ) ) {
 315                                                                         q[ k ].push( v );
 316                                                                 }
 317                                                         }
 318                                                 }
 319                                         } );
 320                                 }
 321                                 uri.query = q;
 322                         },
 323
 324                         /**
 325                          * Get user and password section of a URI.
 326                          *
 327                          * @return {string}
 328                          */
 329                         getUserInfo: function () {
 330                                 return cat( '', this.user, cat( ':', this.password, '' ) );
 331                         },
 332
 333                         /**
 334                          * Get host and port section of a URI.
 335                          *
 336                          * @return {string}
 337                          */
 338                         getHostPort: function () {
 339                                 return this.host + cat( ':', this.port, '' );
 340                         },
 341
 342                         /**
 343                          * Get the userInfo, host and port section of the URI.
 344                          *
 345                          * In most real-world URLs this is simply the hostname, but the definition of 'authority' section is more general.
 346                          *
 347                          * @return {string}
 348                          */
 349                         getAuthority: function () {
 350                                 return cat( '', this.getUserInfo(), '@' ) + this.getHostPort();
 351                         },
 352
 353                         /**
 354                          * Get the query arguments of the URL, encoded into a string.
 355                          *
 356                          * Does not preserve the original order of arguments passed in the URI. Does handle escaping.
 357                          *
 358                          * @return {string}
 359                          */
 360                         getQueryString: function () {
 361                                 var args = [];
 362                                 $.each( this.query, function ( key, val ) {
 363                                         var k = Uri.encode( key ),
 364                                                 vals = $.isArray( val ) ? val : [ val ];
 365                                         $.each( vals, function ( i, v ) {
 366                                                 if ( v === null ) {
 367                                                         args.push( k );
 368                                                 } else if ( k === 'title' ) {
 369                                                         args.push( k + '=' + mw.util.wikiUrlencode( v ) );
 370                                                 } else {
 371                                                         args.push( k + '=' + Uri.encode( v ) );
 372                                                 }
 373                                         } );
 374                                 } );
 375                                 return args.join( '&' );
 376                         },
 377
 378                         /**
 379                          * Get everything after the authority section of the URI.
 380                          *
 381                          * @return {string}
 382                          */
 383                         getRelativePath: function () {
 384                                 return this.path + cat( '?', this.getQueryString(), '', true ) + cat( '#', this.fragment, '' );
 385                         },
 386
 387                         /**
 388                          * Get the entire URI string.
 389                          *
 390                          * May not be precisely the same as input due to order of query arguments.
 391                          *
 392                          * @return {string} The URI string
 393                          */
 394                         toString: function () {
 395                                 return this.protocol + '://' + this.getAuthority() + this.getRelativePath();
 396                         },
 397
 398                         /**
 399                          * Clone this URI
 400                          *
 401                          * @return {Object} New URI object with same properties
 402                          */
 403                         clone: function () {
 404                                 return new Uri( this );
 405                         },
 406
 407                         /**
 408                          * Extend the query section of the URI with new parameters.
 409                          *
 410                          * @param {Object} parameters Query parameters to add to ours (or to override ours with) as an
 411                          *  object
 412                          * @return {Object} This URI object
 413                          */
 414                         extend: function ( parameters ) {
 415                                 $.extend( this.query, parameters );
 416                                 return this;
 417                         }
 418                 };
 419
 420                 return Uri;
 421         };
 422
 423         // Default to the current browsing location (for relative URLs).
 424         mw.Uri = mw.UriRelative( function () {
 425                 return location.href;
 426         } );
 427
 428 }( mediaWiki, jQuery ) );