Merge pull request #148 from LeviticusMB/parseUri-fix

Rewrite parseUri to handle unusual but valid URI characters.

lib/browser.js

@@ -584,20 +584,23 @@ hawk.utils = {
     },
 
     parseUri: function (input) {
+        // From RFC 3986 (except for some groups being non-capturing and two extra groups for 'resource' and 'relative')
+        var uriRegex  = /^(?:([^:\/?#]+):)?(?:\/\/([^\/?#]*))?((([^?#]*)(?:\?([^#]*))?)(?:#(.*))?)/;
+        var uriKeys   = ['source', 'protocol', 'authority', 'resource', 'relative', 'pathname', 'query', 'fragment'];
+        var authRegex = /^(?:(([^:@]*)(?::([^@]*))?)@)?(\[[^\]]*\]|[^:]*)(?::(\d*))?/;
+        var authKeys  = ['authority', 'userInfo', 'user', 'password', 'hostname', 'port'];
+        var uri = {}, i;
 
-        // Based on: parseURI 1.2.2
-        // http://blog.stevenlevithan.com/archives/parseuri
-        // (c) Steven Levithan <stevenlevithan.com>
-        // MIT License
+        var parts = uriRegex.exec(input);
 
-        var keys = ['source', 'protocol', 'authority', 'userInfo', 'user', 'password', 'hostname', 'port', 'resource', 'relative', 'pathname', 'directory', 'file', 'query', 'fragment'];
+        for (i = 0; i < parts.length; ++i) {
+            uri[uriKeys[i]] = parts[i] || '';
+        }
 
-        var uriRegex = /^(?:([^:\/?#]+):)?(?:\/\/((?:(([^:@]*)(?::([^:@]*))?)?@)?([^:\/?#]*)(?::(\d*))?))?(((((?:[^?#\/]*\/)*)([^?#]*))(?:\?([^#]*))?)(?:#(.*))?)/;
-        var uriByNumber = input.match(uriRegex);
-        var uri = {};
+        parts = authRegex.exec(uri['authority']);
 
-        for (var i = 0, il = keys.length; i < il; ++i) {
-            uri[keys[i]] = uriByNumber[i] || '';
+        for (i = 0; i < parts.length; ++i) {
+            uri[authKeys[i]] = parts[i] || '';
         }
 
         if (uri.port === '') {

test/browser.js

@@ -1442,6 +1442,33 @@ describe('Browser', function () {
                 expect(uri.port).to.equal('80');
                 done();
             });
+
+            it('handles unusual characters correctly', function (done) {
+                var parts = {
+                    protocol: 'http+vnd.my-extension',
+                    user:     'user!$&\'()*+,;=%40my-domain.com',
+                    password: 'pass!$&\'()*+,;=%40:word',
+                    hostname: 'foo-bar.com',
+                    port:     '99',
+                    pathname: '/path/%40/!$&\'()*+,;=:@/',
+                    query:    'query%40/!$&\'()*+,;=:@/?',
+                    fragment: 'fragm%40/!$&\'()*+,;=:@/?'
+                };
+
+                parts.userInfo  = parts.user + ':' + parts.password;
+                parts.authority = parts.userInfo + '@' + parts.hostname + ':' + parts.port;
+                parts.relative  = parts.pathname + '?' + parts.query;
+                parts.resource  = parts.relative + '#' + parts.fragment;
+                parts.source    = parts.protocol + '://' + parts.authority + parts.resource;
+
+                var uri = Browser.utils.parseUri(parts.source);
+
+                for (var part in parts) {
+                    expect(uri[part]).to.equal(parts[part]);
+                }
+
+                done();
+            });
         });
 
         var str = 'https://www.google.ca/webhp?sourceid=chrome-instant&ion=1&espv=2&ie=UTF-8#q=url';