| blob | 643e5c3e49a8268258b825202f43f7d4ce251f22 |
1 /**
2 * Library for simple URI parsing and manipulation. Requires jQuery.
3 *
4 * Do not expect full RFC 3986 compliance. Intended to be minimal, but featureful.
5 * The use cases we have in mind are constructing 'next page' or 'previous page' URLs,
6 * detecting whether we need to use cross-domain proxies for an API, constructing
7 * simple URL-based API calls, etc.
8 *
9 * Intended to compress very well if you use a JS-parsing minifier.
10 *
11 * Dependencies: mw, jQuery
12 *
13 * Example:
14 *
15 * var uri = new mw.Uri( 'http://foo.com/mysite/mypage.php?quux=2' );
16 *
17 * if ( uri.host == 'foo.com' ) {
18 * uri.host = 'www.foo.com';
19 * uri.extend( { bar: 1 } );
20 *
21 * $( 'a#id1' ).attr( 'href', uri );
22 * // anchor with id 'id1' now links to http://foo.com/mysite/mypage.php?bar=1&quux=2
23 *
24 * $( 'a#id2' ).attr( 'href', uri.clone().extend( { bar: 3, pif: 'paf' } ) );
25 * // anchor with id 'id2' now links to http://foo.com/mysite/mypage.php?bar=3&quux=2&pif=paf
26 * }
27 *
28 * Parsing here is regex based, so may not work on all URIs, but is good enough for most.
29 *
30 * Given a URI like
31 * 'http://usr:pwd@www.test.com:81/dir/dir.2/index.htm?q1=0&&test1&test2=&test3=value+%28escaped%29&r=1&r=2#top':
32 * The returned object will have the following properties:
33 *
34 * protocol 'http'
35 * user 'usr'
36 * password 'pwd'
37 * host 'www.test.com'
38 * port '81'
39 * path '/dir/dir.2/index.htm'
40 * query {
41 * q1: 0,
42 * test1: null,
43 * test2: '',
44 * test3: 'value (escaped)'
45 * r: [1, 2]
46 * }
47 * fragment 'top'
48 *
49 * n.b. 'password' is not technically allowed for HTTP URIs, but it is possible with other
50 * sorts of URIs.
51 * You can modify the properties directly. Then use the toString() method to extract the
52 * full URI string again.
53 *
54 * Parsing based on parseUri 1.2.2 (c) Steven Levithan <stevenlevithan.com> MIT License
55 * http://stevenlevithan.com/demo/parseuri/js/
56 *
57 */
59 ( function ( mw, $ ) {
61 /**
62 * Function that's useful when constructing the URI string -- we frequently encounter the pattern of
63 * having to add something to the URI as we go, but only if it's present, and to include a character before or after if so.
64 * @param {string|undefined} pre To prepend.
65 * @param {string} val To include.
66 * @param {string} post To append.
67 * @param {boolean} raw If true, val will not be encoded.
68 * @return {string} Result.
69 */
70 function cat( pre, val, post, raw ) {
71 if ( val === undefined || val === null || val === '' ) {
72 return '';
73 }
74 return pre + ( raw ? val : mw.Uri.encode( val ) ) + post;
75 }
77 // Regular expressions to parse many common URIs.
78 var parser = {
79 strict: /^(?:([^:\/?#]+):)?(?:\/\/(?:(?:([^:@\/?#]*)(?::([^:@\/?#]*))?)?@)?([^:\/?#]*)(?::(\d*))?)?((?:[^?#\/]*\/)*[^?#]*)(?:\?([^#]*))?(?:#(.*))?/,
80 loose: /^(?:(?![^:@]+:[^:@\/]*@)([^:\/?#.]+):)?(?:\/\/)?(?:(?:([^:@\/?#]*)(?::([^:@\/?#]*))?)?@)?([^:\/?#]*)(?::(\d*))?((?:\/(?:[^?#](?![^?#\/]*\.[^?#\/.]+(?:[?#]|$)))*\/?)?[^?#\/]*)(?:\?([^#]*))?(?:#(.*))?/
81 },
83 // The order here matches the order of captured matches in the above parser regexes.
84 properties = [
85 'protocol', // http
86 'user', // usr
87 'password', // pwd
88 'host', // www.test.com
89 'port', // 81
90 'path', // /dir/dir.2/index.htm
91 'query', // q1=0&&test1&test2=value (will become { q1: '0', test1: '', test2: 'value' } )
92 'fragment' // top
93 ];
96 /**
97 * We use a factory to inject a document location, for relative URLs, including protocol-relative URLs.
98 * so the library is still testable & purely functional.
99 */
100 mw.UriRelative = function ( documentLocation ) {
101 var defaultUri;
103 /**
104 * Constructs URI object. Throws error if arguments are illegal/impossible, or otherwise don't parse.
105 * @constructor
106 * @param {Object|string} uri URI string, or an Object with appropriate properties (especially another URI object to clone).
107 * Object must have non-blank 'protocol', 'host', and 'path' properties.
108 * This parameter is optional. If omitted (or set to undefined, null or empty string), then an object will be created
109 * for the default uri of this constructor (e.g. document.location for mw.Uri in MediaWiki core).
110 * @param {Object|boolean} Object with options, or (backwards compatibility) a boolean for strictMode
111 * - {boolean} strictMode Trigger strict mode parsing of the url. Default: false
112 * - {boolean} overrideKeys Wether to let duplicate query parameters override eachother (true) or automagically
113 * convert to an array (false, default).
114 */
115 function Uri( uri, options ) {
116 options = typeof options === 'object' ? options : { strictMode: !!options };
117 options = $.extend( {
118 strictMode: false,
119 overrideKeys: false
120 }, options );
122 if ( uri !== undefined && uri !== null && uri !== '' ) {
123 if ( typeof uri === 'string' ) {
124 this.parse( uri, options );
125 } else if ( typeof uri === 'object' ) {
126 // Copy data over from existing URI object
127 for ( var prop in uri ) {
128 // Only copy direct properties, not inherited ones
129 if ( uri.hasOwnProperty( prop ) ) {
130 // Deep copy object properties
131 if ( $.isArray( uri[prop] ) || $.isPlainObject( uri[prop] ) ) {
132 this[prop] = $.extend( true, {}, uri[prop] );
133 } else {
134 this[prop] = uri[prop];
135 }
136 }
137 }
138 if ( !this.query ) {
139 this.query = {};
140 }
141 }
142 } else {
143 // If we didn't get a URI in the constructor, use the default one.
144 return defaultUri.clone();
145 }
147 // protocol-relative URLs
148 if ( !this.protocol ) {
149 this.protocol = defaultUri.protocol;
150 }
151 // No host given:
152 if ( !this.host ) {
153 this.host = defaultUri.host;
154 // port ?
155 if ( !this.port ) {
156 this.port = defaultUri.port;
157 }
158 }
159 if ( this.path && this.path.charAt( 0 ) !== '/' ) {
160 // A real relative URL, relative to defaultUri.path. We can't really handle that since we cannot
161 // figure out whether the last path component of defaultUri.path is a directory or a file.
162 throw new Error( 'Bad constructor arguments' );
163 }
164 if ( !( this.protocol && this.host && this.path ) ) {
165 throw new Error( 'Bad constructor arguments' );
166 }
167 }
169 /**
170 * Standard encodeURIComponent, with extra stuff to make all browsers work similarly and more compliant with RFC 3986
171 * Similar to rawurlencode from PHP and our JS library mw.util.rawurlencode, but we also replace space with a +
172 * @param {string} s String to encode.
173 * @return {string} Encoded string for URI.
174 */
175 Uri.encode = function ( s ) {
176 return encodeURIComponent( s )
177 .replace( /!/g, '%21').replace( /'/g, '%27').replace( /\(/g, '%28')
178 .replace( /\)/g, '%29').replace( /\*/g, '%2A')
179 .replace( /%20/g, '+' );
180 };
182 /**
183 * Standard decodeURIComponent, with '+' to space.
184 * @param {string} s String encoded for URI.
185 * @return {string} Decoded string.
186 */
187 Uri.decode = function ( s ) {
188 return decodeURIComponent( s.replace( /\+/g, '%20' ) );
189 };
191 Uri.prototype = {
193 /**
194 * Parse a string and set our properties accordingly.
195 * @param {string} str URI
196 * @param {Object} options
197 * @return {boolean} Success.
198 */
199 parse: function ( str, options ) {
200 var q,
201 uri = this,
202 matches = parser[ options.strictMode ? 'strict' : 'loose' ].exec( str );
203 $.each( properties, function ( i, property ) {
204 uri[ property ] = matches[ i+1 ];
205 } );
207 // uri.query starts out as the query string; we will parse it into key-val pairs then make
208 // that object the "query" property.
209 // we overwrite query in uri way to make cloning easier, it can use the same list of properties.
210 q = {};
211 // using replace to iterate over a string
212 if ( uri.query ) {
213 uri.query.replace( /(?:^|&)([^&=]*)(?:(=)([^&]*))?/g, function ($0, $1, $2, $3) {
214 var k, v;
215 if ( $1 ) {
216 k = Uri.decode( $1 );
217 v = ( $2 === '' || $2 === undefined ) ? null : Uri.decode( $3 );
219 // If overrideKeys, always (re)set top level value.
220 // If not overrideKeys but this key wasn't set before, then we set it as well.
221 if ( options.overrideKeys || q[ k ] === undefined ) {
222 q[ k ] = v;
224 // Use arrays if overrideKeys is false and key was already seen before
225 } else {
226 // Once before, still a string, turn into an array
227 if ( typeof q[ k ] === 'string' ) {
228 q[ k ] = [ q[ k ] ];
229 }
230 // Add to the array
231 if ( $.isArray( q[ k ] ) ) {
232 q[ k ].push( v );
233 }
234 }
235 }
236 } );
237 }
238 this.query = q;
239 },
241 /**
242 * Returns user and password portion of a URI.
243 * @return {string}
244 */
245 getUserInfo: function () {
246 return cat( '', this.user, cat( ':', this.password, '' ) );
247 },
249 /**
250 * Gets host and port portion of a URI.
251 * @return {string}
252 */
253 getHostPort: function () {
254 return this.host + cat( ':', this.port, '' );
255 },
257 /**
258 * Returns the userInfo and host and port portion of the URI.
259 * In most real-world URLs, this is simply the hostname, but it is more general.
260 * @return {string}
261 */
262 getAuthority: function () {
263 return cat( '', this.getUserInfo(), '@' ) + this.getHostPort();
264 },
266 /**
267 * Returns the query arguments of the URL, encoded into a string
268 * Does not preserve the order of arguments passed into the URI. Does handle escaping.
269 * @return {string}
270 */
271 getQueryString: function () {
272 var args = [];
273 $.each( this.query, function ( key, val ) {
274 var k = Uri.encode( key ),
275 vals = $.isArray( val ) ? val : [ val ];
276 $.each( vals, function ( i, v ) {
277 if ( v === null ) {
278 args.push( k );
279 } else if ( k === 'title' ) {
280 args.push( k + '=' + mw.util.wikiUrlencode( v ) );
281 } else {
282 args.push( k + '=' + Uri.encode( v ) );
283 }
284 } );
285 } );
286 return args.join( '&' );
287 },
289 /**
290 * Returns everything after the authority section of the URI
291 * @return {string}
292 */
293 getRelativePath: function () {
294 return this.path + cat( '?', this.getQueryString(), '', true ) + cat( '#', this.fragment, '' );
295 },
297 /**
298 * Gets the entire URI string. May not be precisely the same as input due to order of query arguments.
299 * @return {string} The URI string.
300 */
301 toString: function () {
302 return this.protocol + '://' + this.getAuthority() + this.getRelativePath();
303 },
305 /**
306 * Clone this URI
307 * @return {Object} new URI object with same properties
308 */
309 clone: function () {
310 return new Uri( this );
311 },
313 /**
314 * Extend the query -- supply query parameters to override or add to ours
315 * @param {Object} query parameters in key-val form to override or add
316 * @return {Object} this URI object
317 */
318 extend: function ( parameters ) {
319 $.extend( this.query, parameters );
320 return this;
321 }
322 };
324 defaultUri = new Uri( documentLocation );
326 return Uri;
327 };
329 // if we are running in a browser, inject the current document location, for relative URLs
330 if ( document && document.location && document.location.href ) {
331 mw.Uri = mw.UriRelative( document.location.href );
332 }
334 }( mediaWiki, jQuery ) );