How can I prevent html entities with PHP a DOMDocument::saveHTML()?

Asked 4/4, 2015 at 15:25 Answered 18/6, 2023 at 11:49

Due to custom storage needs (the "why" is not important here, thanks!) I have to save html <a> links in a specific format such as this:

myDOMNode->setAttribute("href", "{{{123456}}}");

Everything works fine until i call saveHTML() on the containing DOMDocument. This kills it, since it encodes { in %7B.

This is a legacy application where href="{{{123456}}}" works as a placeholder. The command-line parser look for this pattern exactly (unencoded) and cannot be changed.

I've no choice but to do it this way.

I cannot htmldecode() the result.

This HTML will never be displayed as this, it is just a storage need.

Thanks for your help!

Note: I've looked around for 2 hours but none of the proposed solution worked for me. For those who will blindly mark the question as duplicate: please comment and let me know.

Lallans answered 4/4, 2015 at 15:25 Comment(8)

Indeed it does encode it. But this doesn't explain what you mean by 'this kills it'. It's a URL, they can be encode, no problem. Just so you can see I'm not kidding, I made an example: ergobase.nl/test25.html See the source code, it's encode, click the link, it works! Wow. So please tell me what you mean by 'this kills it'? – Gardia 4/4, 2015 at 16:40

You're right, the link would work, but this is a legacy application where href="{{{123456}}}" works as a placeholder. The command-line parser look for this pattern exactly (unencoded) and cannot be changed. – Lallans 4/4, 2015 at 22:25

What happens if you use html_entity_decode() after the saveHTML() and before you send it to the command line parser? – Gardia 4/4, 2015 at 22:31

As I said cannot do that (the page has other encoded entities that must be left untouched) – Lallans 5/4, 2015 at 7:31

You said it is a legacy application; what PHP version are you using? – Salesin 7/4, 2015 at 12:7

PHP 5.6.7 (the "legacy" is in the whole context only, not on the application itself) – Lallans 7/4, 2015 at 12:12

I know this is from years ago, but by any chance did you find any solutions for it? sadly this is still an issue today. – Highmuckamuck 13/4, 2023 at 14:18

@MacA. . Try the accepted solution. Back in the day I did it like that. – Lallans 13/4, 2023 at 15:9

As the legacy code is using {{{...}}} as a placeholder, it may be safe to use a somewhat hackish approach with preg_replace_callback. The following will restore the URL encoded placeholders once the HTML is generated:

$src = <<<EOS
<html>
    <body>
        <a href="foo">Bar</a>
   </body>
</html>
EOS;

// Create DOM document
$dom = new DOMDocument();
$dom->loadHTML($src);

// Alter `href` attribute of anchor
$a = $dom->getElementsByTagName('a')
    ->item(0)
    ->setAttribute('href', '{{{123456}}}');

// Callback function to URL decode match
$urldecode = function ($matches) {
    return urldecode($matches[0]);
};

// Turn DOMDocument into HTML string, then restore/urldecode placeholders 
$html = preg_replace_callback(
    '/' . urlencode('{{{') . '\d+' . urlencode('}}}') . '/',
    $urldecode,
    $dom->saveHTML()
);

echo $html, PHP_EOL;

Output (indented for clarity):

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">
<html>
    <body>
        <a href="{{{123456}}}">Bar</a>
    </body>
</html>

Salesin answered 7/4, 2015 at 11:17 Comment(0)

I came across this same issue recently and this was my solution. Little bit hacky, but you can do this,

$customTempAttributeName='vygjhvgjvgkf';

//$node is your a tag DOM node

$newAttr = $dom->createAttribute($customTempAttributeName);
$newAttr->value = "{{your_placeholder}}";
$node->setAttributeNode($newAttr);
$node->removeAttribute('href');

//Then replace custom dom node with href
$finalHTMLString = $dom->saveHTML();
$finalHTMLString = str_replace($customTempAttributeName,'href',$finalHTMLString);

echo $finalHTMLString;

// Please don't forget to upvote my answer ;)

Raquelraquela answered 18/6, 2023 at 11:49 Comment(0)

Recommended topics

Hot tags