This is necessary to obtain a W3C validation (XHTML1.0 Transitionnal)...
phpinfo's output is declared with that DTD :
- "System ID" has the wrong url to validate : "DTD/xhtml1-transitional.dtd" rather than "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"
- Some module names contains space and the function's output use the name in anchors as ID and NAME. these attributes can't be validated like that (unique name only).

<?php

ob_start ();

ob_start ();                              // Capturing
phpinfo ();                               // phpinfo ()
$info = trim (ob_get_clean ());           // output

// Replace white space in ID and NAME attributes... if exists
$info = preg_replace ('/(id|name)(=["\'][^ "\']+) ([^ "\']*["\'])/i', '$1$2_$3', $info);

$imp = new DOMImplementation ();
$dtd = $imp->createDocumentType (
    'html',
    '-//W3C//DTD XHTML 1.0 Transitional//EN',
    'http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd'
);
$doc = $imp->createDocument (
    'http://www.w3.org/1999/xhtml',
    'html',
    $dtd
);
$doc->encoding = 'utf-8';

$info_doc = new DOMDocument ('1.0', 'utf-8');
/* Parse phpinfo's output
 * operator @ used to avoid messages about undefined entities
 * or use loadHTML instead
 */
@$info_doc->loadXML ($info);

$doc->documentElement->appendChild ( // Adding HEAD element to HTML
    $doc->importNode (
        $info_doc->getElementsByTagName ('head')->item (0),
        true                         // With all the subtree
    )
);
$doc->documentElement->appendChild ( // Adding BODY element to HTML
    $doc->importNode (
        $info_doc->getElementsByTagName ('body')->item (0),
        true                         // With all the subtree
    )
);

// Now you get a clean output and you are able to validate...
/*
echo ($doc->saveXML ());
//      OR
echo ($doc->saveHTML ());
 */

// By that way it's easy to add some style declaration :
$style = $doc->getElementsByTagName ('style')->item (0);
$style->appendChild (
    $doc->createTextNode (
        '/* SOME NEW CSS RULES TO ADD TO THE FUNCTION OUTPUT */'
    )
);

// to add some more informations to display :
$body = $doc->getElementsByTagName ('body')->item (0);
$element = $doc->createElement ('p');
$element->appendChild (
    $doc->createTextNode (
        'SOME NEW CONTENT TO DISPLAY'
    )
);
$body->appendChild ($element);

// to add a new header :
$head = $doc->getElementsByTagName ('head')->item (0);
$meta = $doc->createElement ('meta');
$meta->setAttribute ('name', 'author');
$meta->setAttribute ('content', 'arimbourg at ariworld dot eu');
$head->appendChild ($meta);

// As you wish, take the rest of the output and add it for debugging
$out = ob_get_clean ();

$pre = $doc->createElement ('div'); // or pre
$pre->setAttribute ('style', 'white-space: pre;'); // for a div element, useless with pre
$pre->appendChild ($doc->createTextNode ($out));
$body->appendChild ($pre);

$doc->formatOutput = true; // For a nice indentation
$doc->saveXML ();

?>

All that could be done with only RegExp but I prefer the use of DOM for manipulating documents