Regex for URI.withinString extremely slow on certain inputs

[jsha]

I believe URI's regex for extracting URIs suffers from "catastrophic backtracking" (http://www.regular-expressions.info/catastrophic.html). Here's a repro case (runs in Node). I've also been able to reproduce the slowness in Chrome.

var URI = require("URIjs");
var input = "background-image:url('http://example.com/my/long/path/to/an/image.png')";
URI.withinString(input, function(u) { console.log(u); });

[rodneyrehm]

I haven't written that URL-finder RegExp myself, I simply used one submitted to a url-regex-golfing thing we did ages ago.

Maybe Regex 101 can help with debugging. Also see Regexpr visualization

[rodneyrehm]

I wonder if we shouldn't just use /\b(?:([a-z][\w-]+:\/\/)|www\.)[^\s]+/gi to find URLs, possibly containing trailing characters (closing brackets, dots, commas, …), and fix that in the replace() callback in withinString().

[jsha]

Yeah, I think that is probably the best approach.

[rodneyrehm]

The issue is resolved in v1.12.0 see the docs for an enhanced API.