Botched Code

function clean_up_xml($xml_string='', $array_of_tags)
{
    if (count($array_of_tags) == 0)
    {
        throw new Exception('Second parameter, must be an array of strings');
    }
    foreach ($array_of_tags as $tag)
    {
        // Clean up tags, and trim whitespaces around the tag.
        $needle[]  = "@[\n\r\t ]*(\<[/]?{$tag}\>)[\n\r\t ]*@";
        $replace[]  = '$1';
        // find tags that don't have CDATA, and wrap them with CDATA
        $needle[]  = "@\<{$tag}\>(?!\<\!\[CDATA\[)(.*?)\</{$tag}\>@ms";
        $replace[]  = "<{$tag}><![CDATA[$1]]></{$tag}>";

        $xml_string = preg_replace($needle, $replace, $xml_string);
    }
    return $xml_string;
}

Explanation:

1. Remove all spaces, tabs, \r, and \n characters around the desired tag, replace with just the tag itself
2. Find all occurances of $tag, but not followed by <![CDATA[, and wrap contents in <![CDATA[ ... ]]>.

“XSLT is a pain in the ass”. That was coming out of my mouth… until, you start understanding it. Best tool for the job, XSLT. It’s designed to style XML…

The solution to my problem was easy once you see samples. My problem was XSLT scoping was never discussed in much detail, in any example I saw.

I may eventually write a little code bit, for future reference.