Validating A URL

[brendandonhue]

I am looking for a way to validate a URL. I don't need to check that it actually exists as in the example below this- I just need to know that its feasible that its a real URL.
For example, it must begin with either http:// or ftp:// (yes thats ok because I dont want https or mailto or other links), then it must contain a string, then it would have to have a period, followed by at least 2 characters.

I'm guessing I will have to use regular expressions, but I dont know exactly how to go about this.
-Thanks in advance.

[Jona]

http://us2.php.net/manual/en/function.ereg-replace.php has the following code, but it does not work for www. It must have http:// or ftp:// at the beginning.

$text = ereg_replace("[[:alpha:]]+://[^<>[:space:]]+[[:alnum:]/]", "<a href=\"\\0\">\\0</a>", $text); 

Jona

[brendandonhue]

Thanks jono that looks perfect!
But how do I use that to validate the URL now?

Lets say the url is stored in $url, how would I check if it meets those requirements?

if ($url == $text) ??

[Jona]

if(preg_match("[[:alpha:]]+://[^<>[:space:]]+[[:alnum:]/]", $url)){
#Is a valid URL
} else {
#Is not a valid URL
} 

Jona

[brendandonhue]

Awesome

[pyro]

I would recommend something more like this, as the one Jona posted has many obvious flaws:

<?PHP
$url = "http://www.infinitypages.com";

if (preg_match("/^(http(s?):\/\/|ftp:\/\/{1})((\w+\.){1,})\w{2,}$/i", $url)) {
    echo "Valid URL";
}
else {
    echo "Invalid URL";
}

?>

Let me explain it now.. It will match like this:

Must start with either http(with an optional s):// or ftp:// followed by any number of alphanumeric characters and the understrike (this allows for subdomains, etc) followed by an ending that is at least 2 characters long.

Note: if you do not want it to allow https:// just remove the (s?) or if you want to add mailto: just add it in with a new | after the ftp://

[brendandonhue]

Ok I mostly understand it now.

If you dont mind
| is boolean operator for or, is that correct?

Can you explain
((w+.){1,})w{2,}
I dont understand what that does..maybe I should just read the manual for preg_match()

[Jona]

Hey, it's not my RegExp, so don't go saying that I suck at them! I was just giving Brendandonhue a suggestion... I didn't even know if it worked or not. lol I anticipated you'd come in here and tell me that I did something wrong.

I'm not the world's greatest PHP programmer, as can plainly be seen here, so don't expect my suggestions to be solutions all the time.

Jona

[brendandonhue]

Jona-I think he was just trying to reccommend a better method. I didn't see any of the flaws in yours but hey im not the best PHPer either

[pyro]

Sure:

((w+.){1,}) means to match one or more occurence of an alphanumeric (and the _) character followed by a period, one or more time. ie. test. or test.test.test. will both validate. This is so that http://infinitypages.com or http://your.very.own.subdomain.infinitypages.com will both validate (as they are valid addresses)

w{2,}$ means $url must end with two or more alphanumeric and the understrike characters... So, this will NOT allow for http://www.infinitypages.com/home.php... Wasn't sure if you wanted it to.

[pyro]

Originally posted by Jona
... you'd come in here and tell me that I did something wrong.

Sorry. Didn't mean it that way. As brendandonhue said, I just wanted to provide a better alternative, as the other regexp would't really do all that much to validate a URL

[Jona]

Yeah, I know. I wasn't upset or anything--besides, it's your job lol. But like I said, I knew you'd come over here and say something about it. No offense taken whatsoever.

Jona

[pyro]

This regexp is a touch tiddier than my last, and allows for an optional trailing slash:

if (preg_match("/^(http(s?):\\/\\/|ftp:\\/\\/{1})((\w+\.)+)\w{2,}(\/?)$/i", $url)) { 

[brendandonhue]

Ok thanks, got it

[jeffmott]

I built a regular expression for URIs about the same time I made one for e-mails. This one comes from RFC2396 Uniform Resource Identifiers (URI): Generic Syntax. Because it is generic syntax, it does not test for any specific scheme, which is good and bad. Good: allows for the many valid schemes currently in existence, and others to be created later. Bad: a non-valid scheme will not be recognized.

But anyway, it is attached. Since this one is far more complicated and lengthy than the e-mail, I've also attached the outlines for each part. It should help you to break it down, or build another one up if you choose to.

I have not found any bugs, but if anyone else does please tell me.