Domxpath issue with preceding-sibling

rcgabriel

So I have a chunk of HTML that looks roughly like this:

<table>
  <tr>
    <td>
      <font size="-1">
          Blah blah <b>BLAH1</b> <i>blah2 blah2</i> <br/> 
          <span> BLAH3BLAH3 </span> <span> MOREBLAH4 </span>
     </font>
    </td>
  </tr>
</table>

And I'm trying to use Domxpath() and an XPath expression to get all of the contents of the font node before the tag.

I am very easily able to get all the contents of the font node with //td/font - this just works, and then I just dump the nodes out to a string with the following function:

   ....
        $xpath = new Domxpath($doc);
        $elementNodes = $xpath->query($expression);
        //convert these DOM elements to strings for mixed-mode searching
        $result = array();
        print_r($elementNodes);
        foreach($elementNodes as $elementNode){
            $string = dumpXML($elementNode);
            echo("s:".$string."\n\n");
            $result[] = $string;
        }

  ....

function dumpXML($node){
  $output = '';
  $owner_document = $node->ownerDocument;
  if($node->childNodes) foreach ($node->childNodes as $el){
     if($owner_document) $output .= $owner_document->saveXML($el);
     //else print_r($node);
  } else {
    if($owner_document) $output .= $owner_document->saveXML($node);
  }
  //else print_r($node);
  return $output;
}

I try to do the same thing with //td/font/br/preceding-sibling::* and it gives me "BLAH1" in the above example (i.e. the contents of the first child node of the font node, at least that's how I interpret it).

What I want to get is "Blah blah BLAH1 blah2 blah2" (or rather, the text node, bold node, italic node so that they get iterated over and all output into the string by dumpXML).

If I try //td/font/br/preceding-sibling::text() it will give me "Blah blah", but won't give me the contents of the and chunks.

This is driving me crazy.

I can't just hack in a regexp or something else since this is part of a framework, and everything works perfectly in every other instance - I really just need an XPath expression that will get everything in the font chunk that comes before the br tag. Any help would be greatly appreciated.

EDIT: Nevermind, I figured out that I can get ::nodes() to do roughly what I want. Still not sure why ::* didn't do it though.