Correct, I've been keeping these "features" a little bit in the "hidden" side since I still have not found a way to deal properly with them, for example, once ": cool :" has been converted to it turns into an "invalid" html that should be parsed out, what I need to do is find a way to "reverse" and reconvert them for preview in the reply function.

What I will end up doing is to reconvert only those emoticons that the system understands, but then we find the issue with auto-generated images and urls, for example:

[ url ]http://www.google.com/[ /url ] will turn into http://www.google.com/
[ img ]http://www.pyforum.org/pyforum/static/images/pyforum_logo.gif[ /img ] will turn into

However, when converted to HTML the system does not know anymore that they were pyforum-generated (this is mostly JS, or client side), I am still looking for an "elegant" and not-hacky way of dealing with this..

There is also suport for [ b ]Bold[ /b ], [ i ]Italic[ /i ] and [ u ]underline[ /u ] (Which I should remove but what the heck)..

Thanks for the comments richard, you're th first member that has taken a look at the actual code (and talked about it), all comments are appreciated..

Regards,

-- sb