- Issue created by @jwag956
HtmlFilter doesn't properly handle   resulting in preg_replace further down the line returning NULL due to invalid UTF-8. (note that I had to leave the ';' off these example so you can see it!)
Start with text with an \  in it - CkEditor likes to insert those.
processFieldValue() will call handleAttributes() which loads the document using Drupal HTML::load - this uses the HTML5 parser to load a dom from a string - and this translates the   into UTf-8 0xc2a0.
parseHtml is then called that calls normalizeText with the UTF-8 string. The call there to preg_replace isn't sent the '/u' modifier, so it doesn't understand the UTF-8 - and proceeds to convert THAT to the unicode REPLACEMENT character (whew!). At this point future calls to preg_replace (with the '/u' modifier) will return NULL with an error code of PREG_BAD_UTF8_ERROR. The tokenizer simplifyText() for example does this.
In normalizeText() add the '/u' modifier to the preg_replace pattern
Active
1.38
Plugins