| blob | 556631285983849d7470464cd398039900acbc12 |
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 */
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 return pre + ( raw ? val : mw.Uri.encode( val ) ) + post;
72 }
74 /**
75 * Regular expressions to parse many common URIs.
76 *
77 * @private
78 * @static
79 * @property {Object} parser
80 */
81 var parser = {
82 strict: /^(?:([^:\/?#]+):)?(?:\/\/(?:(?:([^:@\/?#]*)(?::([^:@\/?#]*))?)?@)?([^:\/?#]*)(?::(\d*))?)?((?:[^?#\/]*\/)*[^?#]*)(?:\?([^#]*))?(?:#(.*))?/,
83 loose: /^(?:(?![^:@]+:[^:@\/]*@)([^:\/?#.]+):)?(?:\/\/)?(?:(?:([^:@\/?#]*)(?::([^:@\/?#]*))?)?@)?([^:\/?#]*)(?::(\d*))?((?:\/(?:[^?#](?![^?#\/]*\.[^?#\/.]+(?:[?#]|$)))*\/?)?[^?#\/]*)(?:\?([^#]*))?(?:#(.*))?/
84 },
86 /**
87 * The order here matches the order of captured matches in the `parser` property regexes.
88 *
89 * @private
90 * @static
91 * @property {Array} properties
92 */
93 properties = [
94 'protocol',
95 'user',
96 'password',
97 'host',
98 'port',
99 'path',
100 'query',
101 'fragment'
102 ];
104 /**
105 * @property {string} protocol For example `http` (always present)
106 */
107 /**
108 * @property {string|undefined} user For example `usr`
109 */
110 /**
111 * @property {string|undefined} password For example `pwd`
112 */
113 /**
114 * @property {string} host For example `www.example.com` (always present)
115 */
116 /**
117 * @property {string|undefined} port For example `81`
118 */
119 /**
120 * @property {string} path For example `/dir/dir.2/index.htm` (always present)
121 */
122 /**
123 * @property {Object} query For example `{ a: '0', b: '', c: 'value' }` (always present)
124 */
125 /**
126 * @property {string|undefined} fragment For example `top`
127 */
129 /**
130 * A factory method to create a variation of mw.Uri with a different default location (for
131 * relative URLs, including protocol-relative URLs). Used so the library is still testable &
132 * purely functional.
133 *
134 * @method
135 * @member mw
136 */
137 mw.UriRelative = function ( documentLocation ) {
138 var defaultUri;
140 /**
141 * @class mw.Uri
142 * @constructor
143 *
144 * Construct a new URI object. Throws error if arguments are illegal/impossible, or
145 * otherwise don't parse.
146 *
147 * @param {Object|string} [uri] URI string, or an Object with appropriate properties (especially
148 * another URI object to clone). Object must have non-blank `protocol`, `host`, and `path`
149 * properties. If omitted (or set to `undefined`, `null` or empty string), then an object
150 * will be created for the default `uri` of this constructor (`document.location` for
151 * mw.Uri, other values for other instances -- see mw.UriRelative for details).
152 * @param {Object|boolean} [options] Object with options, or (backwards compatibility) a boolean
153 * for strictMode
154 * @param {boolean} [options.strictMode=false] Trigger strict mode parsing of the url.
155 * @param {boolean} [options.overrideKeys=false] Whether to let duplicate query parameters
156 * override each other (`true`) or automagically convert them to an array (`false`).
157 */
158 function Uri( uri, options ) {
159 options = typeof options === 'object' ? options : { strictMode: !!options };
160 options = $.extend( {
161 strictMode: false,
162 overrideKeys: false
163 }, options );
165 if ( uri !== undefined && uri !== null && uri !== '' ) {
166 if ( typeof uri === 'string' ) {
167 this.parse( uri, options );
168 } else if ( typeof uri === 'object' ) {
169 // Copy data over from existing URI object
170 for ( var prop in uri ) {
171 // Only copy direct properties, not inherited ones
172 if ( uri.hasOwnProperty( prop ) ) {
173 // Deep copy object properties
174 if ( $.isArray( uri[prop] ) || $.isPlainObject( uri[prop] ) ) {
175 this[prop] = $.extend( true, {}, uri[prop] );
176 } else {
177 this[prop] = uri[prop];
178 }
179 }
180 }
181 if ( !this.query ) {
182 this.query = {};
183 }
184 }
185 } else {
186 // If we didn't get a URI in the constructor, use the default one.
187 return defaultUri.clone();
188 }
190 // protocol-relative URLs
191 if ( !this.protocol ) {
192 this.protocol = defaultUri.protocol;
193 }
194 // No host given:
195 if ( !this.host ) {
196 this.host = defaultUri.host;
197 // port ?
198 if ( !this.port ) {
199 this.port = defaultUri.port;
200 }
201 }
202 if ( this.path && this.path.charAt( 0 ) !== '/' ) {
203 // A real relative URL, relative to defaultUri.path. We can't really handle that since we cannot
204 // figure out whether the last path component of defaultUri.path is a directory or a file.
205 throw new Error( 'Bad constructor arguments' );
206 }
207 if ( !( this.protocol && this.host && this.path ) ) {
208 throw new Error( 'Bad constructor arguments' );
209 }
210 }
212 /**
213 * Encode a value for inclusion in a url.
214 *
215 * Standard encodeURIComponent, with extra stuff to make all browsers work similarly and more
216 * compliant with RFC 3986. Similar to rawurlencode from PHP and our JS library
217 * mw.util.rawurlencode, except this also replaces spaces with `+`.
218 *
219 * @static
220 * @param {string} s String to encode
221 * @return {string} Encoded string for URI
222 */
223 Uri.encode = function ( s ) {
224 return encodeURIComponent( s )
225 .replace( /!/g, '%21' ).replace( /'/g, '%27' ).replace( /\(/g, '%28' )
226 .replace( /\)/g, '%29' ).replace( /\*/g, '%2A' )
227 .replace( /%20/g, '+' );
228 };
230 /**
231 * Decode a url encoded value.
232 *
233 * Reversed #encode. Standard decodeURIComponent, with addition of replacing
234 * `+` with a space.
235 *
236 * @static
237 * @param {string} s String to decode
238 * @return {string} Decoded string
239 */
240 Uri.decode = function ( s ) {
241 return decodeURIComponent( s.replace( /\+/g, '%20' ) );
242 };
244 Uri.prototype = {
246 /**
247 * Parse a string and set our properties accordingly.
248 *
249 * @private
250 * @param {string} str URI, see constructor.
251 * @param {Object} options See constructor.
252 */
253 parse: function ( str, options ) {
254 var q, matches,
255 uri = this;
257 // Apply parser regex and set all properties based on the result
258 matches = parser[ options.strictMode ? 'strict' : 'loose' ].exec( str );
259 $.each( properties, function ( i, property ) {
260 uri[ property ] = matches[ i + 1 ];
261 } );
263 // uri.query starts out as the query string; we will parse it into key-val pairs then make
264 // that object the "query" property.
265 // we overwrite query in uri way to make cloning easier, it can use the same list of properties.
266 q = {};
267 // using replace to iterate over a string
268 if ( uri.query ) {
269 uri.query.replace( /(?:^|&)([^&=]*)(?:(=)([^&]*))?/g, function ( $0, $1, $2, $3 ) {
270 var k, v;
271 if ( $1 ) {
272 k = Uri.decode( $1 );
273 v = ( $2 === '' || $2 === undefined ) ? null : Uri.decode( $3 );
275 // If overrideKeys, always (re)set top level value.
276 // If not overrideKeys but this key wasn't set before, then we set it as well.
277 if ( options.overrideKeys || q[ k ] === undefined ) {
278 q[ k ] = v;
280 // Use arrays if overrideKeys is false and key was already seen before
281 } else {
282 // Once before, still a string, turn into an array
283 if ( typeof q[ k ] === 'string' ) {
284 q[ k ] = [ q[ k ] ];
285 }
286 // Add to the array
287 if ( $.isArray( q[ k ] ) ) {
288 q[ k ].push( v );
289 }
290 }
291 }
292 } );
293 }
294 uri.query = q;
295 },
297 /**
298 * Get user and password section of a URI.
299 *
300 * @return {string}
301 */
302 getUserInfo: function () {
303 return cat( '', this.user, cat( ':', this.password, '' ) );
304 },
306 /**
307 * Get host and port section of a URI.
308 *
309 * @return {string}
310 */
311 getHostPort: function () {
312 return this.host + cat( ':', this.port, '' );
313 },
315 /**
316 * Get the userInfo, host and port section of the URI.
317 *
318 * In most real-world URLs this is simply the hostname, but the definition of 'authority' section is more general.
319 *
320 * @return {string}
321 */
322 getAuthority: function () {
323 return cat( '', this.getUserInfo(), '@' ) + this.getHostPort();
324 },
326 /**
327 * Get the query arguments of the URL, encoded into a string.
328 *
329 * Does not preserve the original order of arguments passed in the URI. Does handle escaping.
330 *
331 * @return {string}
332 */
333 getQueryString: function () {
334 var args = [];
335 $.each( this.query, function ( key, val ) {
336 var k = Uri.encode( key ),
337 vals = $.isArray( val ) ? val : [ val ];
338 $.each( vals, function ( i, v ) {
339 if ( v === null ) {
340 args.push( k );
341 } else if ( k === 'title' ) {
342 args.push( k + '=' + mw.util.wikiUrlencode( v ) );
343 } else {
344 args.push( k + '=' + Uri.encode( v ) );
345 }
346 } );
347 } );
348 return args.join( '&' );
349 },
351 /**
352 * Get everything after the authority section of the URI.
353 *
354 * @return {string}
355 */
356 getRelativePath: function () {
357 return this.path + cat( '?', this.getQueryString(), '', true ) + cat( '#', this.fragment, '' );
358 },
360 /**
361 * Get the entire URI string.
362 *
363 * May not be precisely the same as input due to order of query arguments.
364 *
365 * @return {string} The URI string
366 */
367 toString: function () {
368 return this.protocol + '://' + this.getAuthority() + this.getRelativePath();
369 },
371 /**
372 * Clone this URI
373 *
374 * @return {Object} New URI object with same properties
375 */
376 clone: function () {
377 return new Uri( this );
378 },
380 /**
381 * Extend the query section of the URI with new parameters.
382 *
383 * @param {Object} parameters Query parameters to add to ours (or to override ours with) as an
384 * object
385 * @return {Object} This URI object
386 */
387 extend: function ( parameters ) {
388 $.extend( this.query, parameters );
389 return this;
390 }
391 };
393 defaultUri = new Uri( documentLocation );
395 return Uri;
396 };
398 // If we are running in a browser, inject the current document location (for relative URLs).
399 if ( document && document.location && document.location.href ) {
400 mw.Uri = mw.UriRelative( document.location.href );
401 }
403 }( mediaWiki, jQuery ) );