I have had this happen to me before too. We need to set up some test cases so the bug can be tracked down.
Anonymous
Just posted a bug about this yesterday in the SF.net tracker, test case included:
http://sourceforge.net/tracker/index.php?func=detail&aid=962432&group_id=51422&atid=463233
This the test case:
<b><i>tag soup!</b></i>test
link
And the result I get:
<b><i>tag soup!</i></b>"http://www.google.com">test link
I fixed my code to handle comments properly — just needed to still build up the $tag after detecting it initially blank (to get the contents out of [2] of the regexp). That cleans up the stack handling a lot, probably speeds up the code for people with a lot of inline comments.
My case was something like
text [more]
[comment]
[comment]
text
[ul]
[li][strong]text[/strong]text[/li]
… a bunch of times
.. then one line missing the [/li]
.. few more closed properly
[/ul]
text
[comment]
this text gets chopped off a bit at the start of the line.
=======
-d
Anonymous
if you fixed it, could you share your code? 🙂 Perhaps the core developers are interested in those fixes… and I’m also interested in using this very cool function in one of my little apps.
This is obviously just a portion of the balanceTags function, and I’m working off an earlier (non-1.2) codebase — though the code is likely near-identical…
} else { // Begin Tag
$tag = strtolower($regex[1]);
$attributes = $regex[2];
if (empty($tag)) {
// echo "re0: $regex[0], re1: $regex[1], re2: $regex[2]\n<br>";
} else {
// echo "TAG: $tag\n<br>";
// Tag Cleaning
// Push if not img or br or hr
if($tag != 'br' && $tag != 'img' && $tag != 'hr') {
$stacksize = array_push ($tagstack, $tag);
}
// Attributes
// $attributes = $regex[2];
if($attributes) {
// fix to avoid CSS defacements
if ($is_comment) {
$attributes = str_replace('style=', 'title=', $attributes);
$attributes = str_replace('class=', 'title=', $attributes);
$attributes = str_replace('id=', 'title=', $attributes);
}
$attributes = ' '.$attributes;
}
}
// always build up new $tag, so we capture comments back...
$tag = '<'.$tag.$attributes.'>';
}
-d
