Does anyone know why (not just on NGA) occasionally a cut and pasted link sometimes isn't "hyperlinked" right to the end, resulting in the link not working? I had this happen in a post on a thread in Ask a Question, the second link down (see below). The 1.pdf is not in blue so a click results in a Page not Found instead of the article:

http://garden.org/thread/view_...

Special characters like parenthesis can break the script that converts URLs to hyperlinks. In the Garden.org code I look at certain characters (like spaces) to mean that a URL has ended. Until now, a parenthesis was treated as an end of the URL character.

I updated the code now to allow extra characters in URLs and they should work just fine now.

I assume you've taken care of the case where a URL is added to a post in parenthesis, such as "comment (URL)" or "comment (refer to URL)"? I find I often insert URLs that way and was glad Garden.org didn't mess that up (often other sites include the closing parenthesis as part of the URL).

This is the regular expression I'm currently using:

$url = '#(^|[\n (])([\w]+?://([^\s<]+)|(www\.[^\s<]+?\.[^\s<]+))(?<![\.,:])#is';

It does have the unfortunate side effect of including the trailing ) as part of the URL. I'm not sure what the right solution is, given that parenthesis in URLs are valid and legal. How can the system be programmed to know whether you intended that character to be a part of the URL, or rather the end of a sentence?

Joshua, I just updated the regexp again to try to fix that. Parenthesis at the end of a URL should work fine now.

(Testing: http://fshs.org/proceedings-o/...)

Thanks @Dave. I know it can be quite a pain to try to avoid, so all I could think of would be to check if there was a parenthesis just before the URL and pair it to one after the URL.

Otherwise I will fallback to my previous workaround, which was to include a space either side of the URL.

Joshua, I got it fixed, actually. I'm checking to see if the URL ends with a parenthesis and if yes, don't include that last character in the URL. It's working well!! (Example: http://www.garden.org/)