Normalizes URL's to prevent duplicates that point to the same server resource.

var urlCanon = require('url-canon')
urlCanon('https://www.google.com/?b=hey&a=yo&#')
// => returns 'https://www.google.com/?&a=yo&b=hey'

npm install url-canon

• Sorts query parameters into alphabetical order
• Decodes needlessly percent-encoded ASCII
• Removes default ports for http/https
• Lowercases the protocol and hostname
• Removes trailing # and ?
• Uppercases percent encoded bytes

Tests pass in

• IE8+
• Chrome
• Firefox
• Opera

Does not support Node.js (as it uses an anchor element for parsing).

Internationalised URLs such as https://見.香港/ are converted to punycode form in Chrome but not in Firefox and IE. Conversely punycode such as https://xn--nw2a.xn--j6w193g/ is decoded automatically in Firefox and IE to https://見.香港/. It is very hard to standardise these URLs and they have not been handled.