Same as #7650 (comment), need a little more input.

Looking at the CSS regex, in your example, cases that should not match still do:

Moving the ? quantifier into the capturing group will allow the back reference to accept an empty match:

url\(([\"']?)(?P<url>\s*http?[^)'\"]+)\1\)

Looking at the CSS regex, in your example, cases that should not match still do:
...
Moving the ? quantifier into the capturing group will allow the back reference to accept an empty match:
...

url\(([\"']?)(?P<url>\s*http?[^)'\"]+)\1\)

Ah nice catch, missed that one! Will adapt the PR.

Here's a reproduction zip with all 3 quote styles:

9.5.42-css-quoted-urls-missing-from-privacy-plugin.zip

With the current version, only the unquoted external URL gets fetched.

Let me know if these 2 repro zips are clear enough @squidfunk 🙇

Edit: pushed the new regex and also updated the regex101 link above. Should be ready for another round 🏓

Thanks! I believe we need to allow spaces before and after the URL, as according to syntax level 3, specifically how URL tokens are parsed. There are two ways how URLs are handled: if they contain a string, they're just considered to be normal function tokens, which definitely demands for whitespace after and before. If they contain a verbatim URL, I believe we need to support whitespace as well.

url\(\s*([\"']?)(?P<url>http?[^)'\"]+)\1\s*\)

This now works correctly with the following strings, albeit it consumes the trailing whitespace on the verbatim version, which should not be a problem though:

/* correct */
url("https://example.com/images/myImg.jpg");
url('https://example.com/images/myImg.jpg');
url(  'https://example.com/images/myImg.jpg'  );
url(  "https://example.com/images/myImg.jpg"  );
url(https://example.com/images/myImg.jpg);
url(   https://example.com/images/myImg.jpg   );


/* mismatching */
url('https://example.com/images/myImg.jpg);
url("https://example.com/images/myImg.jpg);
url('https://example.com/images/myImg.jpg");
url(https://example.com/images/myImg.jpg');
url(https://example.com/images/myImg.jpg");

/* non-http links */
url("data:image/jpg;base64,iRxVB0…");
url(myImg.jpg);
url(#IDofSVGpath);

PS: How can I share on regex101? I'm too stupid to find the share button 😅

Perfect, thanks! Added the new pattern.

Hehe, Ctrl+S should save it and give you a modal with the link: https://regex101.com/r/LVJJfK/1https://regex101.com/r/LVJJfK/1. Not sure if it requires starting over when receiving a link though.

I also just checked and as you say at least urlparse seems to be happy with trailing whitespace:

>>> from urllib.parse import urlparse
>>> urlparse("https://example.com/images/myImg.jpg   ")
ParseResult(scheme='https', netloc='example.com', path='/images/myImg.jpg   ', params='', query='', fragment='')

Perfect, thanks for investigating! I think this is safe to merge then 🤟

Released as part of 9.5.43.