PHP Regex ignoring nested tags

Hi I'm working on a bug in a CMS and I was hoping someone could give me some help with this messy regex! I need to remove everything inside the {{page? }} tags (where 'page' is a dynamic word), including any nested {{tags}} within them.- except for {{links? }}

In the code below, the regex should remove everything inside the {{homepage? }} tag:

<div id="main">   
    <div id="left">
    {{menu1}}<br />

{{homepage?
    <img src="images/{{timenow}}.gif" width="177" height="217" alt="{{imgname}}"开发者_运维技巧 id="biglogo" />
}}

{{links?
    <b>LINKS</b>
}}
</div>
{{menu2}}
</div>

Here's what I have so far. It's getting stuck as soon as it sees the timenow}}

$result=preg_replace("#\{\{(?!links)\S*?\?.*?}}#s","",$result);

Clarification:

There are no {{page? }} sub tags (all subtags are {{thisformat}} ). In other words something like: {{foo? {{links? bar }} baz }} would never occur.

---

You can do something like: #\{\{ (?!links\b) \w+ \? (?: \{\{\w+}} | [^{}]+ | \{(?!\{) | }(?!}) )* }}#sx

---

If I understand it correctly, there's no need for recursive matching here; the {{page? }} tags may contain simple tags like {{this}}, and that's it. In that case, you just have to watch out for the beginning of a nested tag, so you can match the end of that tag when it shows up, then go on looking for either the end of the enclosing {{page? }} tag or the beginning of another nested tag.

$regex='#
  \{\{ (?!links\?) \w++\?     # page-tag start
  (?:
    (?: (?!\{\{|\}\}) . )++   # normal content
  |
    \{\{                      #
    (?: (?!\}\}) . )*+        # embedded tag
    \}\}                      #
  )*+
  \}\}                        # page-tag end
#sx';

The "normal content" part matches one or more of any character, unless the next character is the beginning of a {{ or }} sequence. Once we've started to match an embedded tag, we use the same technique to gobble up its content.

see it in action at ideone.com

---

This is not possible with regex. Read about the millions of failed attempts to parse nested html/xml with regex.