There seems to be an issue parsing non-numeric keys in array arguments. Check out the following,

> JSON.stringify(url('?','?t[0]=x&t[1]=y'));
"{"t":["x","y"]}"
> JSON.stringify(url('?','?t[a]=x&t[b]=y'));
"{"t[a]":"x","t[b]":"y"}"

The expected behavior is,

> JSON.stringify(url('?','?t[0]=x&t[1]=y'));
"{"t":["x","y"]}"
> JSON.stringify(url('?','?t[a]=x&t[b]=y'));
"{"t":{"a":"x","b":"y"}}"

You might want to note that your parser is incompatible with Internet Explorer 7 and below. Alternatively you may want to replace arg[0] (line 48 and others) by arg.substring(0, 1) or something.

in Line 23 you're assuming that //hostname equals http://hostname. That is only true if the scheme of the document's resource the script is executed in is http. Should the page be served via https, this little line defeats the purpose of "scheme agnostic URLs".

According to RFC 3986 Section 3.3, this is a a valid path: "/foo/xy://bar" - but url("path", "/foo/xy://bar") === "//bar/" - which is wrong…

I have stopped reviewing the code on line 32 because I found it unbearable to read without "running every single expression in my head".

I believe that documentation example

url('?'); // {query1: 'test', silly: 'willy'}

is missing the field parameter which would be an array of length 3, but with 2 elements. Don't know how to denote that clearly, so I made an issue :)

In the documentation the following statement:
url('path', 'http://www.example.com:8080/some/path');
should return
/some/path

the result I get from version 1.8.0 is:
/some/path/

Please, can you make the behaviour the same as the documentation? The ending / should not be added.

Library from: http://cdnjs.cloudflare.com/ajax/libs/js-url/1.8.0/js-url.min.js

$.url('path','http://vm.sovrhistory.ru/february-v1/')
returns "/february-v1"
expected "/february-v1/"

Given the example -

$.url('domain'); // www.domain.com
$.url('host'); // domain.com

'domain.com' is the domain, and 'www.domain.com' is the host.

It would be very good to have the library work in node. Any plans?

I am doing development with a popular framework KendoUI which uses this format for Single Page Application (SPA) on mobile, eg:

theurl = "http://localhost:8080/index.html#details?id=1";

I really like this library, currently using: window.url( '#details?id', theurl); which returns 1.

It would be great to handle the case window.url( '?id', url); and return 1 in my example or if that conflicts with another case just add this example to the tests.js to save others time trying combinations to get this case working. Cheers!

Input

jsUrl('fileext', 'path/pdfname.suffix.pdf')

Output:

suffix.pdf

Expected:

pdf

Hi Rob
Can you please take a look on this line?
Thanks for your work.

Current Behaviour:

url('email', 'mailto:[email protected]?subject=Hi');  // [email protected]?subject=Hi
url('?subject', 'mailto:[email protected]?subject=Hi');  // undefined

Expected Behaviour:

url('email', 'mailto:[email protected]?subject=Hi');  // [email protected]
url('?subject', 'mailto:[email protected]?subject=Hi');  // Hi

Multidimensional arrays are not parsed correctly,

> JSON.stringify(url('?','?t[0][0]=x'));
"{"t[0]":["x"]}"
> JSON.stringify(url('?','?t[0][0]=x&t[1][0]=y'));
"{"t[0]":["x"],"t[1]":["y"]}"

The expected behavior is as follows,

> JSON.stringify(url('?','?t[0][0]=x'));
"{"t":[["x"]]}"
> JSON.stringify(url('?','?t[0][0]=x&t[1][0]=y'));
"{"t":[["x"],["y"]]}"

It would be nice to include AMD support.

You are using arg[0] to access character of a string.
This returned "" for me in IE8.
I have changed all arg[x] to arg.chatAt[x] and it worked well.

there is a bug on your url.min.js
url('domain') not work (undefined)
the url.js works but url.min.js not work.

version: 2.0.2

case 1, end with no extension or forward slash:

var urlString = http://localhost/news/list
window.url('file', urlString): undefined        //should be "list"
window.url('filename', urlString): undefined    //should be "list"
window.url('-1', urlString): list

case 2, end with dot, no extension:

var urlString = http://localhost/news/list.
window.url('file', urlString): undefined        //should be "list."
window.url('filename', urlString): undefined    //should be "list"
window.url('-1', urlString): list.

case 3, end with forward slash:

var urlString = http://localhost/news/list/
window.url('file', urlString): undefined
window.url('filename', urlString): undefined
window.url('-1', urlString): list               //should be ""

case 4, regular:

var urlString = http://localhost/news/list.html
window.url('file', urlString): list.html
window.url('filename', urlString): list
window.url('-1', urlString): list.html

url('?q', 'http://www.google.com/?q=a+b') returns the raw result a+b, rather than the querystring unencoded a b. That means I have to decode the string--I can't think of any case I'd want the current behavior.

Hello,

Very nice script, but doesn't work in IE 9 due to its dependence on the "split" function.

Cheers,
Jeff.

use this url 'https://test-passport.jindashi.cn/TDOA/page/pfyh/quick.html'
the result domain is hi.cn not jindashi.cn

if ?poo is not exist,url('?poo') value is undefined,not null.

url('?poo'); // null should be -> undefined

example.plus
=> pl.us

Current list of TLDs:
http://data.iana.org/TLD/tlds-alpha-by-domain.txt

.ly is widely used.
.xyz is popular now after www.abc.xyz

Found this issue:

When trying to parse the following url: https://facebook.com\@account.google.com/x

url('domain', 'https://facebook.com\@account.google.com/x') // outputs 'google.com'

This outputs google.com as the domain instead of facebook.com

Not sure if people consider this an issues, however, arrays are apparently padded with nulls for missing values,

> JSON.stringify(url('?','?t[0]=x&t[1]=y'));
"{"t":["x","y"]}"
> JSON.stringify(url('?','?t[0]=x&t[2]=y'));
"{"t":["x",null,"y"]}"

Maybe the better option would be to return objects for sparse arrays, like so,

> JSON.stringify(url('?','?t[0]=x&t[1]=y'));
"{"t":["x","y"]}"
> JSON.stringify(url('?','?t[0]=x&t[2]=y'));
"{"0":"x","2":"y"}"

In your readme link "Download the lastest version of url" https://github.com/websanova/url/tags is broken

Would you be so kind to release a nuget package for .NET?
Thx!

Seriously, this library has really nothing to do with jQuery, it should be plain JavaScript.

The library seems to be performing double decoding when user tries to access param or attr method.

How to reproduce:
I just used chrome console:

$.getScript('/js/purl.js')
var test='http://localhost:8080?id=test%255Ftest'; //where test%255Ftest is "test%5Ftest" encoded
purl(test, true).param('id'); //outputs test_test

The first decoding during parseURI phase produces correct param test%test but then it goes onto create the param list in following line

uri.param['query'] = parseString(uri.attr['query']);

parseString internally again calls decodeURI, %5F is URL encoded '_' so the final result is 'test_test'.

in summary

test%5Ftest -> encode -> test%255Ftest -> decode() Line 40 -> test%5Ftest -> decode Line 131 -> results in test_test.

It would be helpful if this was mentioned in the examples

www.example.com/?prop
url("?prop") -> undefined

www.example.com/
url("?prop") -> null

Also it might make more sense if the first case returned "" instead of undefined

From exemple http://rob:[email protected]/path/index.html?query1=test&silly=willy#test=hash&chucky=cheese, I expected $url('#poo') == null rather than an empty string.

Otherwise, there is no mechanisme to distinguish http://exemple.com#foo and http://exemple.com

The domain-first domain approach is incorrect. For example, goo.gl returns undefined

http://url.spec.whatwg.org/

Might be more useful if you refactored to be a polyfill for the actual URL API spec

Needs tests

I noticed that url('protocol', 'mailto:[email protected]') returns http.

Is there any way to detect for 'non-web' URI schemes? (not sure on the exact nomenclature)

It would be nice to have a bower package for js-url library.

Publishing in bower would make it easier to find and install js-url.js library.

https://spikes.marketing/

its active domain but plugin shows undefined

eg. url#/page1.html?param1=value1

url('?param1'); //undefined
url('#page1.html?param1') // value1

$.url('host', 'http://[email protected]')
// "[email protected]"

Should be example.org

bower is handy...

my url is http://localhost:50001/a/b/Index?reportno=1392
when i use
var reportno = $.url('?reportno');
Firebug error:TypeError: split[i].match is not a function url.js(line 31,col 21)
my jquery version is jquery-1.11.3.min.js

http://localhost:6744/index.html#pagesClient/VedListCompare.html?road=2&&vv=143&relvv=148

url('?road') => undefined

Getting the last part of the url (ie. $.url(-1)) also returns the hash value.
e.g. http://www.domain.com/sub#hash gives "sub#hash" and not "hash".

"dependencies": {
    "grunt-contrib-qunit": "",
    "grunt-contrib-uglify": "",
    "grunt-contrib-jshint": ""
  },

grunt-contrib-* should be put into devDependencies not dependencies

url('domain','http://mall.belle.net.cn/')
now this result is wrong: net.cn

correct result: belle.net.cn