2 * Library for simple URI parsing and manipulation. Requires jQuery.
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.
9 * Intended to compress very well if you use a JS-parsing minifier.
11 * Dependencies: mw, jQuery
15 * var uri = new mw.Uri( 'http://foo.com/mysite/mypage.php?quux=2' );
17 * if ( uri.host == 'foo.com' ) {
18 * uri.host = 'www.foo.com';
19 * uri.extend( { bar: 1 } );
21 * $( 'a#id1' ).attr( 'href', uri );
22 * // anchor with id 'id1' now links to http://foo.com/mysite/mypage.php?bar=1&quux=2
24 * $( 'a#id2' ).attr( 'href', uri.clone().extend( { bar: 3, pif: 'paf' } ) );
25 * // anchor with id 'id2' now links to http://foo.com/mysite/mypage.php?bar=3&quux=2&pif=paf
28 * Parsing here is regex based, so may not work on all URIs, but is good enough for most.
31 * 'http://usr:pwd@www.test.com:81/dir/dir.2/index.htm?q1=0&&test1&test2=&test3=value+%28escaped%29&r=1&r=2#top':
32 * The returned object will have the following properties:
39 * path '/dir/dir.2/index.htm'
44 * test3: 'value (escaped)'
49 * n.b. 'password' is not technically allowed for HTTP URIs, but it is possible with other
51 * You can modify the properties directly. Then use the toString() method to extract the
52 * full URI string again.
54 * Parsing based on parseUri 1.2.2 (c) Steven Levithan <stevenlevithan.com> MIT License
55 * http://stevenlevithan.com/demo/parseuri/js/
62 * Function that's useful when constructing the URI string -- we frequently encounter the pattern of
63 * having to add something to the URI as we go, but only if it's present, and to include a character before or after if so.
64 * @param {String} to prepend, if value not empty
65 * @param {String} value to include, if not empty
66 * @param {String} to append, if value not empty
67 * @param {Boolean} raw -- if true, do not URI encode
70 function cat( pre, val, post, raw ) {
71 if ( val === undefined || val === null || val === '' ) {
74 return pre + ( raw ? val : mw.Uri.encode( val ) ) + post;
78 // Regular expressions to parse many common URIs.
80 strict: /^(?:([^:\/?#]+):)?(?:\/\/(?:(?:([^:@]*)(?::([^:@]*))?)?@)?([^:\/?#]*)(?::(\d*))?)?((?:[^?#\/]*\/)*[^?#]*)(?:\?([^#]*))?(?:#(.*))?/,
81 loose: /^(?:(?![^:@]+:[^:@\/]*@)([^:\/?#.]+):)?(?:\/\/)?(?:(?:([^:@]*)(?::([^:@]*))?)?@)?([^:\/?#]*)(?::(\d*))?((?:\/(?:[^?#](?![^?#\/]*\.[^?#\/.]+(?:[?#]|$)))*\/?)?[^?#\/]*)(?:\?([^#]*))?(?:#(.*))?/
84 // The order here matches the order of captured matches in the above parser regexes.
89 'host', // www.test.com
91 'path', // /dir/dir.2/index.htm
92 'query', // q1=0&&test1&test2=value (will become { q1: 0, test1: '', test2: 'value' } )
97 * Constructs URI object. Throws error if arguments are illegal/impossible, or otherwise don't parse.
99 * @param {!Object|String} URI string, or an Object with appropriate properties (especially another URI object to clone). Object must have non-blank 'protocol', 'host', and 'path' properties.
100 * @param {Boolean} strict mode (when parsing a string)
102 mw.Uri = function( uri, strictMode ) {
103 strictMode = !!strictMode;
104 if ( uri !== undefined && uri !== null || uri !== '' ) {
105 if ( typeof uri === 'string' ) {
106 this._parse( uri, strictMode );
107 } else if ( typeof uri === 'object' ) {
109 $.each( properties, function( i, property ) {
110 _this[property] = uri[property];
112 if ( this.query === undefined ) {
117 if ( !( this.protocol && this.host && this.path ) ) {
118 throw new Error( 'Bad constructor arguments' );
123 * Standard encodeURIComponent, with extra stuff to make all browsers work similarly and more compliant with RFC 3986
124 * Similar to rawurlencode from PHP and our JS library mw.util.rawurlencode, but we also replace space with a +
125 * @param {String} string
126 * @return {String} encoded for URI
128 mw.Uri.encode = function( s ) {
129 return encodeURIComponent( s )
130 .replace( /!/g, '%21').replace( /'/g, '%27').replace( /\(/g, '%28')
131 .replace( /\)/g, '%29').replace( /\*/g, '%2A')
132 .replace( /%20/g, '+' );
136 * Standard decodeURIComponent, with '+' to space
137 * @param {String} string encoded for URI
138 * @return {String} decoded string
140 mw.Uri.decode = function( s ) {
141 return decodeURIComponent( s ).replace( /\+/g, ' ' );
147 * Parse a string and set our properties accordingly.
148 * @param {String} URI
149 * @param {Boolean} strictness
150 * @return {Boolean} success
152 _parse: function( str, strictMode ) {
153 var matches = parser[ strictMode ? 'strict' : 'loose' ].exec( str );
155 $.each( properties, function( i, property ) {
156 uri[ property ] = matches[ i+1 ];
159 // uri.query starts out as the query string; we will parse it into key-val pairs then make
160 // that object the "query" property.
161 // we overwrite query in uri way to make cloning easier, it can use the same list of properties.
163 // using replace to iterate over a string
165 uri.query.replace( /(?:^|&)([^&=]*)(?:(=)([^&]*))?/g, function ($0, $1, $2, $3) {
167 var k = mw.Uri.decode( $1 );
168 var v = ( $2 === '' || $2 === undefined ) ? null : mw.Uri.decode( $3 );
169 if ( typeof q[ k ] === 'string' ) {
172 if ( typeof q[ k ] === 'object' ) {
184 * Returns user and password portion of a URI.
187 getUserInfo: function() {
188 return cat( '', this.user, cat( ':', this.password, '' ) );
192 * Gets host and port portion of a URI.
195 getHostPort: function() {
196 return this.host + cat( ':', this.port, '' );
200 * Returns the userInfo and host and port portion of the URI.
201 * In most real-world URLs, this is simply the hostname, but it is more general.
204 getAuthority: function() {
205 return cat( '', this.getUserInfo(), '@' ) + this.getHostPort();
209 * Returns the query arguments of the URL, encoded into a string
210 * Does not preserve the order of arguments passed into the URI. Does handle escaping.
213 getQueryString: function() {
215 $.each( this.query, function( key, val ) {
216 var k = mw.Uri.encode( key );
217 var vals = val === null ? [ null ] : $.makeArray( val );
218 $.each( vals, function( i, v ) {
219 args.push( k + ( v === null ? '' : '=' + mw.Uri.encode( v ) ) );
222 return args.join( '&' );
226 * Returns everything after the authority section of the URI
229 getRelativePath: function() {
230 return this.path + cat( '?', this.getQueryString(), '', true ) + cat( '#', this.fragment, '' );
234 * Gets the entire URI string. May not be precisely the same as input due to order of query arguments.
235 * @return {String} the URI string
237 toString: function() {
238 return this.protocol + '://' + this.getAuthority() + this.getRelativePath();
243 * @return {Object} new URI object with same properties
246 return new mw.Uri( this );
250 * Extend the query -- supply query parameters to override or add to ours
251 * @param {Object} query parameters in key-val form to override or add
252 * @return {Object} this URI object
254 extend: function( parameters ) {
255 $.extend( this.query, parameters );