Question

I’m trying to create my own function for convert some links string from bbcode (and from normal string) to the right html.

At the moment I have this one :

$format_search=array(
    '#\[url=(((http|https|ftp)://)[a-zA-Z0-9\-_\./\?=&;\#]+)\](.*?)\[/url\]#i'
    '#(?<![>/"])(((http|https|ftp)://)[a-zA-Z0-9\-_\./\?=&;\#]+)#im',  
);

$format_replace=array(
  '<a class="lforum" href="$1">$4</a>',
  '<a class="lforum" href="$1">$1</a>',    
);

$str=preg_replace($format_search, $format_replace, $str);

it works more or less 🙂

Catching bbcode as [url=link]link_name[/url] is not a problem. The problem is when I try to catch every kind of link on the website. (for example when a user insert http://link.com without any bbcode).

For example [b]http://links[/b] doesnt work… and there are many others scenario to get in consideration!

I don’t know how to create a good function without conflicts! I mean : first parse the string searching the link’s strings as bbcode; than, parse the rest, maybe without replace the previous.

Do you have any suggestions?

Answer 1

Not with that approach. The (?<![>/"]) is what prevents it from working. The purpose of that assertion is to prevent double-linkifying <a>http://example.com</a>. But it also prevents matches on <b>http://example.com.

One workaround would be to alter your output links:

 '<a href="$1" class="lforum">$4</a>',

This would allow to use the class= in the negative assertion:

 (?<![/"]|class="lforum">)http..

So it still matches tags other than <a> links.

---

Another approach would be to pre-convert raw text URLs into BBcode before you convert your BBcode into HTML. Use your existing URL regex for that and prefix it with e.g. (?<![\]=]) and use [url=$1]$1[/url] as output instead.