Question

I’m trying to find a regex to remove all html special chars (mostly &, <, >) but keeping the html tags intact.

I’m getting these informations from a database, so I can’t make sure that characters like < and > are replaced with &gt; and &lt;

I could manage to replace & and < it with the following RegEx in PHP:

$Value = preg_replace('/<(?!#?\/?[a-zA-Z0-9]+>)/','',$Value);
$Value = preg_replace('/&(?!#?[a-zA-Z0-9]+;)/','&amp;',$Value);

I have now only troubles to fix the > tags, because I’d have to use lookup-behind, which doesn’t allow non fixed length RegEx.

$Value = preg_replace('/(?<!<[a-zA-Z0-9]+)>/','',$Value);

Any ideas?

Greetings
-Thomas

Answer 1

Use a DOM Parser and apply your replacements to the text nodes only.

$partialId = uniqid();
$dom = new DOMDocument;
$dom->loadHTML(sprintf('<div id="%s">%s</div>', $partialId, $html));
echo $dom->saveHtml($dom->getElementById($partialId));

Just parsing the partial will already turn XML special chars to their respective entities:

<div id="4f02efa1a4e9b">this is my fancy <i>text</i> I love my text lalalal &gt; wow I'm great"</div>

If you are not on PHP 5.3.6 you cannot use saveHTML with a node. See How to get innerHTML of DOMNode? and How to return outer html of DOMDocument? for workarounds.

If you need to work on the text nodes, you can do

$xpath = new DOMXPath($dom);
foreach ($xpath->query('//text()') as $textNode) {
    $textNode->nodeValue = doSomething($textNode->nodeValue);
}

Also see DOMDocument in php for an introduction to how DOM works.