regexp problem

[dnast]

Hi, I'm having a problem with altering data in which my script automatically turns e-mail addresses and webpages into clickable links mixed with a search function that highlights keywords.

Here's is the auto-link code for email:

$info = preg_replace('/([^\s\"?|<>]+@([^\s\"?|<>]+\.)+?[a-zA-Z]{2,4})/i','<a href="mailto:\\1">\\1</a>',$info);

and web addresses:

$info = preg_replace('/((http|ftp):\\/\\/[a-zA-Z0-9]+(\.[-a-zA-Z0-9]+)*\.[a-zA-Z]{2,4}(\/[^\s<>\[\]\|\"\\\\]+)*)[\s]/i','<a href="\1">\\1</a>',$info);
$info = preg_replace('/[\s]((http|ftp):\\/\\/[a-zA-Z0-9]+(\.[-a-zA-Z0-9]+)*\.[a-zA-Z]{2,4}(\/[^\s<>\[\]\|\"\\\\]+)*)/i','<a href="\1">\\1</a>',$info);
$info = preg_replace('/([\s])(www\.[-a-zA-Z0-9]+\.[a-zA-Z]{2,4})/i','\1<a href="http://\\2">\\2</a>',$info); //This one checks for ww.whatever.dom (no http: etc)

*note - had to change 'www' to 'ww' so the forum script wouldn't convert the URL. Also, some backslashes might be missing since I had to go in and double some of the slashes to get them to display on this board (I might have overlooked one or two). I know for a fact that the regexps all work unless they intercept each other.

and here is what I use to highlight keywords (note: $search can contain multiple keywords delimited by |):

  $info = preg_replace("/(".$search.")/i",'<span style="color: red">\\1</span>',$info);

If a keyword matches part of an auto-link, the code will mess up the anchor tag and create unwanted output. I've tried several alterations, but none of them worked, so I just gave the simplest working code I used.

Also, suggestions for better regexps for the ones shown above are welcome (I just noticed that I could've used [[:alphanum:]] 😉). I just can't solve this one!

Thanks for at least looking into this.

[freebie]

you could probably do this quite easily without Regular Expressions (though using Regular expressions is generally a good thing)

eg.

function createLink($link,$linkType){
    if($linkType=="Email"){
        return "<a href='mailto:$link'>$link</a>";
    }else{
       return "<a href='$link'>$link</a>";
    }
}

[dnast]

Thanks!

But I'm using this code to go through and convert information that was entered by other users automatically. So I'd still need a regular expression to find matches.

It's similar to a message in a forum. A user types in descriptive information or maybe a news story and the script catches textual e-mail and web addresses and automatically converts them when it's outputted. For example, if you type www.php.net on this forum it'll be converted into a hyperlink even if you don't use the [url] tag.

This all works fine and dandy unless you do a search for a keyword that matches part of an e-mail or web address. I think what I need is a way for the regular expression (the search one) to not match (and consequently not convert) if it's inside another tag. I could be wrong, though.

Thanks for at least looking into it, freebie. 🙂