parshap / html-linkify Goto Github PK

The regex gets stuck in an infinite loop (JavaScript) when you have many trailing dots:

var text = 'http://www.google.com............................................';
linkify(text); // stuck in infinite loop

It is stuck here:

var rLink = /\b((?:https?:\/\/|www\d{0,3}[.]|[a-z0-9.\-]+[.][a-z]{2,4}\/)(?:[^\s()<>]+|\(([^\s()<>]+|(\([^\s()<>]+\)))*\))+(?:\(([^\s()<>]+|(\([^\s()<>]+\)))*\)|[^\s`!()\[\]{};:'".,<>?«»“”‘’]))/gi;

...

while (match = rLink.exec(text)) { // stuck here

E.g. convert example.com but not google.com

You should go to example.com

Learn more about it from this <a href="https://www.google.com/search?q=example.com">Google query</a>

When I run this over a text like <iframe src="https://player.vimeo.com/video/123123" ></iframe> it'll kill the iframe.

https://videomail.io/[email protected]&subject=Question blahblah 

ends up in a weird, broken link where the email address becomes another mailto link, but shouldn't.

Attributes are not being passed to the email render. ie.
{ attributes: { target: '_blank' }

Support with port?
e.g www.google.com:443

Currently the link regex only matches links that begin with "www" or "http://" ("or https://"). It may be warranted to consider other links. Possible approaches would be to look for other subdomains, look for other parts of a URI (e.g., a port), or whitelisting gTLDs. It would be useful to see what expectations are set by Twitter's, Facebook's and others' behaviors.