conservative-normalize-url

conservative-normalize-url

Normalize URLs conservatively, taking into account old specs to avoid breaking things.

Background

I created this library, because I was frustrated by the lack of backwards-compatibility of existing URL normalizers. If you're normalizing a URL, it's likely you're dealing with a user-supplied URL. After working on software that handles bookmarks for a long time, I realized quite how diverse those can be. This is what I'm using now.

What does it do?

• path normalization (/foo//bar/../baz => /foo/baz; when the protocol is known)
• url encoding non-ASCII characters and unprintable characters (including space)
• sorting query parameters
• punycode encoding domains
• omitting ports if the port is the default port for the given protocol

What does it not do?

• omitting trailing slashes in paths
• omitting www.
• omitting any query parameters
• adding or changing the protocol
• removing the auth portion of the URL
• breaking URLs with use ; as a query parameter separator

Usage

const normalizeUrl = require('conservative-normalize-url')

normalizeUrl('HtTp://foobar.com/bla bla h/shake?this=öppa&oppa=this') // http://foobar.com/bla%20bla%20h/shake?oppa=this&this=%C3%B6ppa

License

(c) Marcel Klehr MIT License