HTML filter breaking XHTML 1.0 Strict Compliance by removing closing slash at end of tag

i noticed that is is inpossible to add any self-closing xhtml codes in the proper way so it keeps XHTML Strict Compliance

I have noticed this also. The HTML filter will convert

<img src="path/to/image.jpg" />

in to this

<img src="path/to/image.jpg">

breaking XHTML 1.0 Strict Compliance.

4.6.4 and on has filter_xss which takes special care to keep the XHTML slash.

I missed to change the version info, sorry.

I have this problem with 4.6.4, and setting line 1097 in filter.module,v 1.57.2.10

  // Is there a closing XHTML slash at the end of the attributes?
  $xhtml_slash = preg_match('%\s/\s*$%', $attr) ? '/' : '';

to this

  // Is there a closing XHTML slash at the end of the attributes?
  $xhtml_slash =  '/';

doesn't help.

(I believe $attr should be $attrlist since $attr is never set in function _filter_xss_split and should there not be a space before the closing slash?)

Problem persist with Drupal 4.6.5.

$attr to $attrlist fixes this.

patch attached (patched against 4.6.5).

This works for me now, I must have done something wrong when I did the testing in my last post.

I suggest we also add a space before the closing slash, that is what my patch adds.

I committed a slightly tweaked patch: the code would always add a space after the tag name, even if there were no attributes (e.g. <tag >) and it would not recognize the XHTML slash without the (optional) leading space (e.g. <tag/>).

Thanks.

Looks like the extra space will still be there with that commit Steven.

http://drupal.org/node/27230 is a patch for this -- your conditional is cleaner.

Err, how? This is the code:

  // Is there a closing XHTML slash at the end of the attributes?
  $xhtml_slash = preg_match('%\s?/\s*$%', $attrlist) ? ' /' : '';

  // Clean up attributes
  $attr2 = implode(' ', _filter_xss_attributes($attrlist));
  $attr2 = preg_replace('/[<>]/', '', $attr2);
  $attr2 = strlen($attr2) ? ' '. $attr2 : '';

  return "<$elem$attr2$xhtml_slash>";

$attrlist only contains non-empty strings, without leading or trailing spaces. And $attr2 only gets prefixed with a space if it is non-empty. So, there can only be a space between $elem and $attr2, if there are attributes.

As far as the leading space for the XHTML slash, that is recommended by W3C.

  // Is there a closing XHTML slash at the end of the attributes?
  $xhtml_slash = preg_match('%\s?/\s*$%', $attrlist) ? ' /' : '';
  // Clean up attributes
  $attr2 = implode(' ', _filter_xss_attributes($attrlist));
  $attr2 = preg_replace('/[<>]/', '', $attr2);
  $attr2 = strlen($attr2) ? ' '. $attr2 : '';
  return "<$elem $attr2$xhtml_slash>";

that's the code: last line still have space between $elem and $attr
http://cvs.drupal.org/viewcvs/drupal/drupal/modules/filter.module?r1=1.9...

Has*

Sorry about not being very clear there.

Whoops!
http://cvs.drupal.org/viewcvs/drupal/drupal/modules/filter.module?r1=1.9...
doesn't show the change on that line but it's there, fixed indeed.