function for people whining that preg_match_all is less to type and concluding from it that regex must be better for parsing html than a dom parser

xpath_match_all

<?php
/**
 * Run an XPath query against an HTML string and returns the results
 *
 * @param string $xpath The XPath query to run on $thml
 * @param string $html  The HTML to parse. Uses the previously used string if omitted
 * @return array
 */
function xpath_match_all($query, $html = '')
{
    static $dom;
    static $xpath;
    static $content;

    if (!$dom) {
        $dom = new DOMDocument;
    }

    if ($html !== '') {
        $content = $html;
        libxml_use_internal_errors(true);
        $dom->loadHtml($html);
        $xpath = new DOMXPath($dom);
        libxml_use_internal_errors(false);
    }

    $matches = array(array(), array());
    foreach ($xpath->evaluate($query) as $i => $resultNode) {
        $save = version_compare(PHP_VERSION, '5.3.6', '<') ? 'saveXml' : 'saveHtml';
        $matches[0][] = $dom->$save($resultNode);
        $innerHtml = '';
        if ($resultNode->childNodes) {
            foreach ($resultNode->childNodes as $childNode) {
                $innerHtml .= $dom->$save($childNode);
            }
        }
        $matches[1][] = $innerHtml;
    }
    return $matches;
}

Store the current state of $user_errors in libxml_use_internal_errors() https://gist.github.com/4546595/revisions