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                  * @throws {Error} when the query string or fragment contains an unknown % sequence
 180                  */
 181                 function Uri( uri, options ) {
 182                         var prop, hrefCur,
 183                                 hasOptions = ( options !== undefined ),
 184                                 defaultUri = getDefaultUri();
 185
 186                         options = typeof options === 'object' ? options : { strictMode: !!options };
 187                         options = $.extend( {
 188                                 strictMode: false,
 189                                 overrideKeys: false
 190                         }, options );
 191
 192                         if ( uri !== undefined && uri !== null && uri !== '' ) {
 193                                 if ( typeof uri === 'string' ) {
 194                                         this.parse( uri, options );
 195                                 } else if ( typeof uri === 'object' ) {
 196                                         // Copy data over from existing URI object
 197                                         for ( prop in uri ) {
 198                                                 // Only copy direct properties, not inherited ones
 199                                                 if ( Object.prototype.hasOwnProperty.call( uri, prop ) ) {
 200                                                         // Deep copy object properties
 201                                                         if ( Array.isArray( uri[ prop ] ) || $.isPlainObject( uri[ prop ] ) ) {
 202                                                                 this[ prop ] = $.extend( true, {}, uri[ prop ] );
 203                                                         } else {
 204                                                                 this[ prop ] = uri[ prop ];
 205                                                         }
 206                                                 }
 207                                         }
 208                                         if ( !this.query ) {
 209                                                 this.query = {};
 210                                         }
 211                                 }
 212                         } else if ( hasOptions ) {
 213                                 // We didn't get a URI in the constructor, but we got options.
 214                                 hrefCur = typeof documentLocation === 'string' ? documentLocation : documentLocation();
 215                                 this.parse( hrefCur, options );
 216                         } else {
 217                                 // We didn't get a URI or options in the constructor, use the default instance.
 218                                 return defaultUri.clone();
 219                         }
 220
 221                         // protocol-relative URLs
 222                         if ( !this.protocol ) {
 223                                 this.protocol = defaultUri.protocol;
 224                         }
 225                         // No host given:
 226                         if ( !this.host ) {
 227                                 this.host = defaultUri.host;
 228                                 // port ?
 229                                 if ( !this.port ) {
 230                                         this.port = defaultUri.port;
 231                                 }
 232                         }
 233                         if ( this.path && this.path[ 0 ] !== '/' ) {
 234                                 // A real relative URL, relative to defaultUri.path. We can't really handle that since we cannot
 235                                 // figure out whether the last path component of defaultUri.path is a directory or a file.
 236                                 throw new Error( 'Bad constructor arguments' );
 237                         }
 238                         if ( !( this.protocol && this.host && this.path ) ) {
 239                                 throw new Error( 'Bad constructor arguments' );
 240                         }
 241                 }
 242
 243                 /**
 244                  * Encode a value for inclusion in a url.
 245                  *
 246                  * Standard encodeURIComponent, with extra stuff to make all browsers work similarly and more
 247                  * compliant with RFC 3986. Similar to rawurlencode from PHP and our JS library
 248                  * mw.util.rawurlencode, except this also replaces spaces with `+`.
 249                  *
 250                  * @static
 251                  * @param {string} s String to encode
 252                  * @return {string} Encoded string for URI
 253                  */
 254                 Uri.encode = function ( s ) {
 255                         return encodeURIComponent( s )
 256                                 .replace( /!/g, '%21' ).replace( /'/g, '%27' ).replace( /\(/g, '%28' )
 257                                 .replace( /\)/g, '%29' ).replace( /\*/g, '%2A' )
 258                                 .replace( /%20/g, '+' );
 259                 };
 260
 261                 /**
 262                  * Decode a url encoded value.
 263                  *
 264                  * Reversed #encode. Standard decodeURIComponent, with addition of replacing
 265                  * `+` with a space.
 266                  *
 267                  * @static
 268                  * @param {string} s String to decode
 269                  * @return {string} Decoded string
 270                  * @throws {Error} when the string contains an unknown % sequence
 271                  */
 272                 Uri.decode = function ( s ) {
 273                         return decodeURIComponent( s.replace( /\+/g, '%20' ) );
 274                 };
 275
 276                 Uri.prototype = {
 277
 278                         /**
 279                          * Parse a string and set our properties accordingly.
 280                          *
 281                          * @private
 282                          * @param {string} str URI, see constructor.
 283                          * @param {Object} options See constructor.
 284                          * @throws {Error} when the query string or fragment contains an unknown % sequence
 285                          */
 286                         parse: function ( str, options ) {
 287                                 var q, matches,
 288                                         uri = this,
 289                                         hasOwn = Object.prototype.hasOwnProperty;
 290
 291                                 // Apply parser regex and set all properties based on the result
 292                                 matches = parser[ options.strictMode ? 'strict' : 'loose' ].exec( str );
 293                                 properties.forEach( function ( property, i ) {
 294                                         uri[ property ] = matches[ i + 1 ];
 295                                 } );
 296
 297                                 // uri.query starts out as the query string; we will parse it into key-val pairs then make
 298                                 // that object the "query" property.
 299                                 // we overwrite query in uri way to make cloning easier, it can use the same list of properties.
 300                                 q = {};
 301                                 // using replace to iterate over a string
 302                                 if ( uri.query ) {
 303                                         uri.query.replace( /(?:^|&)([^&=]*)(?:(=)([^&]*))?/g, function ( match, k, eq, v ) {
 304                                                 if ( k ) {
 305                                                         k = Uri.decode( k );
 306                                                         v = ( eq === '' || eq === undefined ) ? null : Uri.decode( v );
 307
 308                                                         // If overrideKeys, always (re)set top level value.
 309                                                         // If not overrideKeys but this key wasn't set before, then we set it as well.
 310                                                         if ( options.overrideKeys || !hasOwn.call( q, k ) ) {
 311                                                                 q[ k ] = v;
 312
 313                                                         // Use arrays if overrideKeys is false and key was already seen before
 314                                                         } else {
 315                                                                 // Once before, still a string, turn into an array
 316                                                                 if ( typeof q[ k ] === 'string' ) {
 317                                                                         q[ k ] = [ q[ k ] ];
 318                                                                 }
 319                                                                 // Add to the array
 320                                                                 if ( Array.isArray( q[ k ] ) ) {
 321                                                                         q[ k ].push( v );
 322                                                                 }
 323                                                         }
 324                                                 }
 325                                         } );
 326                                 }
 327                                 uri.query = q;
 328
 329                                 // Decode uri.fragment, otherwise it gets double-encoded when serializing
 330                                 if ( uri.fragment !== undefined ) {
 331                                         uri.fragment = Uri.decode( uri.fragment );
 332                                 }
 333                         },
 334
 335                         /**
 336                          * Get user and password section of a URI.
 337                          *
 338                          * @return {string}
 339                          */
 340                         getUserInfo: function () {
 341                                 return cat( '', this.user, cat( ':', this.password, '' ) );
 342                         },
 343
 344                         /**
 345                          * Get host and port section of a URI.
 346                          *
 347                          * @return {string}
 348                          */
 349                         getHostPort: function () {
 350                                 return this.host + cat( ':', this.port, '' );
 351                         },
 352
 353                         /**
 354                          * Get the userInfo, host and port section of the URI.
 355                          *
 356                          * In most real-world URLs this is simply the hostname, but the definition of 'authority' section is more general.
 357                          *
 358                          * @return {string}
 359                          */
 360                         getAuthority: function () {
 361                                 return cat( '', this.getUserInfo(), '@' ) + this.getHostPort();
 362                         },
 363
 364                         /**
 365                          * Get the query arguments of the URL, encoded into a string.
 366                          *
 367                          * Does not preserve the original order of arguments passed in the URI. Does handle escaping.
 368                          *
 369                          * @return {string}
 370                          */
 371                         getQueryString: function () {
 372                                 var args = [];
 373                                 // eslint-disable-next-line no-jquery/no-each-util
 374                                 $.each( this.query, function ( key, val ) {
 375                                         var k = Uri.encode( key ),
 376                                                 vals = Array.isArray( val ) ? val : [ val ];
 377                                         vals.forEach( function ( v ) {
 378                                                 if ( v === null ) {
 379                                                         args.push( k );
 380                                                 } else if ( k === 'title' ) {
 381                                                         args.push( k + '=' + mw.util.wikiUrlencode( v ) );
 382                                                 } else {
 383                                                         args.push( k + '=' + Uri.encode( v ) );
 384                                                 }
 385                                         } );
 386                                 } );
 387                                 return args.join( '&' );
 388                         },
 389
 390                         /**
 391                          * Get everything after the authority section of the URI.
 392                          *
 393                          * @return {string}
 394                          */
 395                         getRelativePath: function () {
 396                                 return this.path + cat( '?', this.getQueryString(), '', true ) + cat( '#', this.fragment, '' );
 397                         },
 398
 399                         /**
 400                          * Get the entire URI string.
 401                          *
 402                          * May not be precisely the same as input due to order of query arguments.
 403                          *
 404                          * @return {string} The URI string
 405                          */
 406                         toString: function () {
 407                                 return this.protocol + '://' + this.getAuthority() + this.getRelativePath();
 408                         },
 409
 410                         /**
 411                          * Clone this URI
 412                          *
 413                          * @return {Object} New URI object with same properties
 414                          */
 415                         clone: function () {
 416                                 return new Uri( this );
 417                         },
 418
 419                         /**
 420                          * Extend the query section of the URI with new parameters.
 421                          *
 422                          * @param {Object} parameters Query parameters to add to ours (or to override ours with) as an
 423                          *  object
 424                          * @return {Object} This URI object
 425                          */
 426                         extend: function ( parameters ) {
 427                                 $.extend( this.query, parameters );
 428                                 return this;
 429                         }
 430                 };
 431
 432                 return Uri;
 433         };
 434
 435         // Default to the current browsing location (for relative URLs).
 436         mw.Uri = mw.UriRelative( function () {
 437                 return location.href;
 438         } );
 439
 440 }() );