Regular Expression matching either a single special character OR a particular sequence

Question

Not sure if I'm doing this right:

/(https?:\/\/\S+)(\s|(& nbsp;))?/g;

This should match a URL beginning with http(s):// and ending with a space character or a & nbsp;

So the problem is this part:

(\s|(& nbsp;))?

That should mean: match either a white space or a & nbsp; but it doesn't work. It never matches for a & nbsp; and just continues until it finds a white space.

I'm not looking for any other http regexp, I'm not looking for a javascript library solution, I'm happy with this, I just want to figure out that last portion.

Edit: some kind of bug in the code formatting on this site, there isn't a space between & and nbsp; but this site turns it into a space if I get rid of that separating space.

Answer 1

The \S+ bit is greedy, and will match as many non-space characters as possible, including any   that might be there. Change it to it's ungreedy equivalent \S+?, and you'll probably have better luck:

/(https?:\/\/\S+?)(\s| |$)/g;

(Updated because I overlooked the trailing ?.)

Answer 2

The problem is the ? quantifier in (\s|( ))?. Remember that * and ? always succeed because they're allowed to match nothing. In your case, \S+ was gobbling up the non-whitespace, and the optional alternative was matching vacuously.

Change your pattern to /(https?:\/\/\S+)(\s| )/.

As for getting   to appear in your question, you can use "&nbsp" in plain text or ampersand followed by "nbsp;" in backticks.

Yes, \S+ is greedy, but the regular expression engine will backtrack to find a match, which forces the earlier subexpression to give back part of what it matched.

The code below shows that the pattern above is not matching the non-breaking space entity, and produces the following results, where the strings inside the brackets are what the parenthesized subpatterns matched:

• no match: http://www.stackoverflow.com/
• match: http://www.stackoverflow.com/ [http://www.stackoverflow.com/][ ]
• match: http://www.stackoverflow.com/ [http://www.stackoverflow.com/][ ]
• no match: https://www.stackoverflow.com/
• match: https://www.stackoverflow.com/ [https://www.stackoverflow.com/][ ]
• match: https://www.stackoverflow.com/ [https://www.stackoverflow.com/][ ]

Code (but remove the space between the ampersand and "nbsp;"):

function testMatches() {
  var lis = document.getElementsByTagName("li");
  for (i = 0; i < lis.length; i++) {
    var url = lis[i].innerHTML;
    var pattern = /(https?:\/\/\S+)(\s|& nbsp;)/;  // FIXME!
    var m = pattern.exec(url);
    if (m) {
      lis[i].innerHTML = "match: " + url  + " ["
                                   + m[1] + "]["
                                   + m[2] + "]";
    }
    else {
      lis[i].innerHTML = "no match: " + url;
    }
  }
}