Re: PHP Regular expression help

I've seen and used weirder and more obfuscated, but nothing more
complex than this one. I worked it out by hand using the RFCs. It is
probably not complete, but was as close as I could stand to get it.

// returns false if URL contains spaces or illegal characters.
// ignores the http:// (programming should strip or require)
// second parameter is whatever you want to return if $val is empty
function djp_is_valid_url($val,$allow_empty=0)
{
    if ($val)
      {
        $http='http(s)?://'; // http(s) is optional.

        // usernames and passwords, theoretically can be transmitted
thru the url,
        // as username:password-9IKiO1iGCm/QT0dZR+AlfA@xxxxxxxxxxxxxxxx, etc.
        $userpass='[a-z0-9]{1,}:[a-z0-9]{1,}@';

        // domain is alphanumeric with hyphens,
        // any number of iterations, separated by periods.
        // followed by at least two alphanumeric with hyphens,
        // followed by a period, then 2-6 alphas (.com, .edu, .tv,
.net, .museum, et al)
        $domain='((([0-9a-z-]*\.)*)?[0-9a-z-]{2,})+\.[a-z]{2,6}';

        // an ip address is 4 numerics delimited by periods.
        $ip='([0-9]{1,3}.){3}([0-9]{1,3}){1}';

        // and very optional port #,
        $port=':[0-9]{1,5}';

        // with an optional following slash.
        $slash='/?';

        // path must start with a slash, followed by optional tilde,
        // then alphanumeric with underscores, slashes and periods.
        $path='/(~?[0-9a-z/\._-])*';

        // "GET" strings. There's an easier way BEGGING to come out of this....
        $getparams='('.
          // the first get string starts with ?,
          // followed by Alphanumeric,'=',optional Alphanumeric or
%(2hex characters)
            '(\?[_0-9a-z]{1,}=(([0-9a-z+/_*.-]*)|(%[a-f0-9]{2}))*)*'.
          // the second and subsequent get parameters start with &,
          // followed by Alphanumeric,'=',optional Alphanumeric or
%(2hex characters)
            '(\&[_0-9a-z]{1,}=(([0-9a-z+/_*.-]*)|(%[a-f0-9]{2}))*)*'.
        // allow zero or one of the preceding.
          ')?';

        $fragment='(#([0-9a-z_.,]*))?';
        // assemble the parts & check.

$validurl="^(($http)?(($userpass)?($domain)|($ip))($port)?){1}$slash($path$getparams$fragment)?$";
// for extra credit, uncomment the next line and read the output
// print("The regular expression that approximates a valid URL is:\n$validurl");
        if (eregi($validurl,$val))
          {
            return 1;
          }
        else
          {
            return 0;
          }
      }
    else return $allow_empty;
}