smart quotes

Hrmm, you could replace them by their HTML entity...

I think the problem is with the browser not decently transmitting it, because indeed, PunBB uses htmlspecialchars();

The difficulty we are having is that there is a php function being applied which is supposed to take care of it. Its trying to work out why it not doing its job which is the difficult thing. I don't see that the browser could be the problem since this character is supposed to be removed before it ever gets to the browser. I'm sure there is a simple answer to this, I just don't know what it is.

This has to be fixed for version 1.3 since it is intended to be capable of being served as application/xhtml+xml and this problem could well cause a page to fail completely. I wonder if we need htmlentities() rather than htmlspecialchars().

I think I prefer the shiflett version but it is nice to know there is an easy way around this. My question is do you run that against the output or do you run it before you save the message in the first place.

What I had in mind is that parser.php is far more critical to performance than the routines run when a message is saved so the less it has to do the better. It just occurred to me that even though its best to save messages in as pure a form as possible if something is obviously crud wouldn't it be more efficient not to save it in the first place.

EDIT: Ignore me, thats only a good idea on a new forum, doesn't help one thats already up and running does it.

All posts, including existing ones go through parser.php before they are sent to the browser. If the function Smarty's posted is added to parser.php then it will deal with existing posts. I would just paste the function into parser.php somewhere near the top and then look for the line which says $text = htmlspecialchars($text) and then put $text = CleanupSmartQuotes($text); immediatelly following it. You can't do any harm by doing that since you are only changing what is sent to the browser i.e. you are not altering the stored messages in any way.

P.S. I think I prefer the function given here http://shiflett.org/archive/165 because it just uses standard quotes.

The ellipsis is chr(133) so you need to add chr(133) to the first array and the numeric reference 8230 to the second array. It shouldn't really cause a problem though because PunBB actually uses it for page numbering and its perfectly valid though PunBB does use the entity reference not the character itself.

Yes there is a number attached to each letter/character.

I encountered the same problem today, fixed it, and posted it in another thread, but I was unaware of this thread.

This problem will fix itself when we switch to UTF-8, am I right?