Linkify breaks with URLs surrounded by color codes

[milas]

If you have a URL that's got color codes around it and linkify the output, you'll get a </span> in the link because the regex for URL (specifically the host/path portion) will match any non-whitespace character:

anser/lib/index.js

Lines 196 to 198 in fb58237

	linkify (txt) {
	return txt.replace(/(https?:\/\/[^\s]+)/gm, str => `<a href="${str}">${str}</a>`);
	}

Example

(Same code, just inlined here)

var anser = require("anser")

const link = anser.linkify(anser.ansiToHtml(
    "\u001b[31mhttps://github.com/IonicaBizau/anser\u001b[0m"
))

console.log(link)

RunKit repro (same code as here): https://runkit.com/milas/605a270caf5fed001a4b6e93

Output

<span style="color:rgb(187, 0, 0)"><a href="https://github.com/IonicaBizau/anser</span>">https://github.com/IonicaBizau/anser</span></a>

Solution

If we assume the input has been escaped with escapeForHtml, perhaps linkify's regex could be made more strict to only match on URL safe characters? Just adding < to the not matching set would also probably be sufficient for this case, but there might be other edge cases re: linkifying in general (i.e. not induced by having a surrounding color code) that are still missed.

[IonicaBizau]

Hmm, good catch! I will happily merge a PR fixing this! ❇️

[aaronvg]

any plan to fix this?

[perlpunk]

• Fix linkify to only match valid URI characters #75