Regex on ;) smilie

Question

A minor inconvenience my users have found is that if they use a smilie such as >_> at the end of parentheses (kind of like this: >_>) then during processing it is run through htmlspecialchars(), making it >_>) - you can see the problem, I think. The ;) at the end is then replaced by the "Wink" smilie.

Can anyone give me a regex that will replace ;) with the smilie, but only if the ; is not the end of an HTML entity? (I'm sure it would involve a lookbehind but I can't seem to understand how to use them >_>)

Thank you!

Answer 1

Handling smileys like ;) is always a bit tricky - the way I would do it is transform it to the "canonical" :wink: before encoding HTML entities, and then changing only canonical-form :{smileyname}: smileys afterwards.

Answer 2

Like this: (?<!&[a-zA-Z0-9]+);\)

The (?>!...) is a zero-width assertion that will only allow the following construct to match text that isn't preceded by the ....