I'm using the DOMDocument class to parse a fairly unpredictable string of markup. It's not all that well formed and I need some data from it. Regex's are right out, of course.
So far, I've got this:
$dom = new DOMDocument;
$dom->loadHTML($str);
$contents = $dom->getElementsByTagName('body')->item(0);
echo $dom->saveXML($contents);
Now this gives me:
<body>
<p>What I'm really after</p>
<ul><li>Foo</li><li>Bar</li></ul>
<h6>And so on</h6>
</body>
What really annoys me are those <body> tags. I want them gone. After grazing the web, I've stumbled across the weirdest workarounds. Some more hacky than others, so in the end, I settled for:
echo substr($dom->saveXML($contents), 6, -7);
Still feels hacky to me, but it's the best I could find. Is there a more reliable way of getting the innerHTML of the DOM, starting from a given node, without the corresponding tags actually showing up?
I've seen suggestions using regex's (a no-no IMHO), or even looping through all the children, echoing those that have childNodes of their own, and stringing together those that don't:
if ($contents->hasChildNodes())
{
$children = $contents->getElementsByTagName('*');
foreach($children as $child)
{
if ($child->hasChildNodes() || $child->nodeName === 'br')
{//or isset($standaloneNodes[$child->nodeName])
echo $dom->saveXML($child);
continue;
}
echo '<'.$child->nodeName.'>'.$child->nodeValue.'</'.$child->nodeName.'>';
}
}
But that, to me, seems even more absurd...
See Question&Answers more detail:
os 与恶龙缠斗过久,自身亦成为恶龙；凝视深渊过久,深渊将回以凝视…
