PHP: How to insert a string into matched regex pattern (adding rel=“no-follow” to anchor links)

Question

I am writing a commenting system for my website, using PHP.

I want to do the following:

1. Detect all external links (ie anchor tags with source NOT containing the string mywebsite.com) in a comment
2. Add the string 'rel="no-follow"' to anchor tags identified in step 1 above.

I have an idea for such a function, but I will need some help from more experienced PHP developers so that I'm sure I'm doing things the right way. This is what my first attempt looks like

<?php

function process_comment($comment)
{

    $external_url_pattern = "href=[^mywebsite.com]"; //this regex is probably wrong (Help!)

    //are there any matches
    $matches = array();
    preg_match_all($external_url_pattern, $comment, $matches);

    foreach($matches as $match)
    {
       // how do we insert the 'rel="no-follow" string ?
    }

}


?>

Would appreciate any comments, pointers and tips in helping me complete this function. Thanks.

Answer 1

Dont know if this will be appropriate, but instead of regex you could do with DOMDocument as well:

$dom = new DOMDocument();
$dom->loadHTML($html);

//Evaluate Anchor tag in HTML
$xpath = new DOMXPath($dom);
$hrefs = $xpath->evaluate("/html/body//a");

for ($i = 0; $i length; $i++) {
        $href = $hrefs->item($i);
        $url = $href->getAttribute('href');

        if($url == "mywebsite.com") {        
             $href->setAttribute("rel", "no-follow");
        }
}

// save html
$html=$dom->saveHTML();

echo $html;

Hope it helps

Answer 2

This is a bit tricky but will do the job.

function process_comment($str)
{

    //parses href attribute values into $match
    if(preg_match_all('/href\=\"(.*)\"/',$str,$match))
    {
        foreach($match[1] as $v)
        {
            //check matched value contains your site as host name
            //if not 
            //adds rel="no-follow" and replaces the link with the attribute
            if(!preg_match('@^(?:http://)?(w+\.)?'.$mysite.'(.*)?@i',$v, $m))
            {
                $rel = $v.'" rel="no-follow';
                $str = str_replace($v,$rel,$str);
            }   
        }
    }

    return $str;
}

process_comment($comment);

You can simply use strstr instead of second preg_match . I used it because I think some urls may contain something like this "http://www.external.com/url.php?v=www.mysite.com"