Validate URL

Simple function won't work

[Birdman]

Hello, this is driving me crazy. I'm just trying to make sure that an url exists. Here is the code:

function valid_url($str)
{
if (@fopen($str,"r"))
return 1;
else
return 0;
}
if (!valid_url($url)){
$error .= "<li>The URL, <em>$url</em>, does not exist.</li>";
}

I have tried using file(), as well. Any help is appreciated!

Birdman

[Birdman]

Well, you won't believe this! It is because of Verisign's new redirect policy. It's returning a page, no matter what url.

[bcolflesh]

Is allow_url_fopen enabled on your server?

If you have PEAR, it has a builtin function for this purpose:

dickmann.homeunix.org/pear/phpdoc/PEAR/Validate/0.1.1/Validate/Validate.html#url

[bcolflesh]

Well, you won't believe this! It is because of Verisign's new redirect policy. It's returning a page, no matter what url.

Aaah - another annoying consequence of their actions - you'll have to return the server headers and parse them.

[Birdman]

bcolflesh, allow_url_fopen is set to 1. I'm not sure if that is correct, but I know it was working before. I'm sure it Verisign that is causing it because I ran a separate test with file(), and it returned the Verisign page(with a bad URL).

[Birdman]

Thanks, I was thinking about searching for a common string in their page and using if(!strpos()), or something like that.

Can you give me an example of what you are suggesting?

Birdman

[bcolflesh]

You're right - until they back down, the server header seems to remain constant for bad requests:

HTTP/1.1 302 Moved Temporarily
Date: Tue, 23 Sep 2003 21:44:38 GMT
Location: [sitefinder.verisign.com...]
Content-Type: text/plain
Transfer-Encoding: Chunked
Connection: Close

You can parse the Location line and assume(?) a bad URL.

[Birdman]

Appreciate it!

I wonder what this will mean to the engines...smells like trouble.

Birdman

[bcolflesh]

Use fsockopen:
us3.php.net/manual/en/function.fsockopen.php

Then fgets:
us3.php.net/manual/en/function.fgets.php

Examples at the bottom of each page.

[Birdman]

Just for the record, this is how I fixed it:

function valid_url($str)
{
if($test = implode("\r",file($str))){
 if (strpos($test,"<title>VeriSign ¦")){
 return 0;
 } else {
 return 1;
 }
} else {
return 0;
}
}

If anyone has a more ellegant solution, I'm all ears ;)

Birdman

ps. That's the beauty of dynamic pages...predictability!

[bcolflesh]

That's cool as long as you don't anticipate any pages that match that title, like:

newsfactor.com/perl/story/7866.html

[Birdman]

I couldn't figure out how to get remote file headers, so I went that way. Could you give me an example of parsing the headers?

[bcolflesh]

Here's an example to retrieve them:

<?php
$fp = fsockopen("www.heruhguuih.com", 80, $errno, $errstr, 30);
fputs($fp, "GET / HTTP/1.1\r\nHost: www.heruhguuih.com\r\n\r\n");
while (!feof($fp)) {
echo nl2br(fgets($fp, 1028));
}
fclose($fp);
?>

[wkitty42]

bcolflesh,

wouldn't HEAD instead of GET be a better option to retrieve just the headers?

[bcolflesh]

HEAD won't retrieve Location in this case.

[wkitty42]

ahhh... and that's what we specifically need to check... gottcha...

[daisho]

you can use cURL if you have it.

[Timotheos]

So Birdman did you make the news [apnews.myway.com]?

[Birdman]

:) No, I never got interviewed, but I feel mr. Fitzpatrick's pain. I spent a good hour or two trying to figure out why something that worked was suddenly broke.

By the way, thanks for the examples. I haven't tried them yet, but will.

Birdman