改进此正则表达式以防止preg_replace抛出PREG_BACKTRACK_LIMIT_ERROR

Question

I want to remove all the scipt-tags from a HTML-page, except those with the word foo or bar. So I came up with this statement:

$content = preg_replace('#<script((?!foo|bar).)*?</script>#is', '', $content);
echo "Last error: " + preg_last_error();

This works on smaller pages. But now I have a page with 30 big script-tags and it doesn't work. The error I get is: PREG_BACKTRACK_LIMIT_ERROR

So I think I need to improve my regex to prevent this error, because this statement works:

$content = preg_replace('#<script.*?</script>#is', '', $content); 

But this statement is removing all the script-tags, while I want to keep some of them.

There are solution about increasing the pcre.backtrack_limit, but I don't want to go that route. There should be a better solution imho.

The thing is that I don't know how to fix this, because the issue is with the regex as far as I can see.

Could you guide me to make the regex better so this error won't occur?

Answer 1

I would strongly suggest not using regular expressions here, but making use of DOM parsing instead, which is often more appropriate in this kind of scenario:

$doc = new \DOMDocument();
$doc->loadHTML($html, LIBXML_HTML_NODEFDTD);

$xpath = new \DOMXPath($doc);
foreach ($xpath->query('//script[contains(text(), \'foo\') or contains(text(), \'bar\')]') as $script_tag) {
  $script_tag->parentNode->removeChild($script_tag);
}

echo $doc->saveHTML();

If you have more words, you can build your xpath query from an array instead:

$blacklist = ['foo', 'bar', 'apple', 'cold'];

$query = '//script[' . join(' or ', array_map(function($banword) { 
  return "contains(text(), '$banword')"; 
}, $blacklist)) . ']';

foreach ($xpath->query($query) as $script_tag) {
  $script_tag->parentNode->removeChild($script_tag);
}

Demo: https://3v4l.org/dHGDt