I have a PHP script that mirror my mailing list to web-based forum, in order to make the forum import look as nice as possible, I use regular expression to catch email signatures & style them appropriately. The signature formats that I'm catching with the regex are:
This is my message...
--
My signature
TheDude.
And
This is my message...
---------------
My signature
TheDude.
Right now I'm using this regex:
$message = preg_replace('/\s*(.+)(\s*[
]-{2,}\s+.*)/s', '$1<span class="msg_footer">$2</span>', $message);
It works, my but after some quick tests, I realized that this regex is really slow.
I'm not that good in regex, can someone please take a look at the regex & tell me how to optimize it & make it fast?
You are using regular expressions to handle the whole message, which is bound to be slow. A better alternative would be to use proper programming logic to process the message. For instance, go through the message line by line and test for each line whether it matches your "start of signature" regex. If not, add it to the array or string holding the actual message. If it does match, add the rest of the message to the footer.
You might also want to start from the bottom instead of the top, if you think that your users will use lines matching your regex in the middle of the message.
Assuming that a signature starts with at least two - at the beginning of line and ends with either , or one or more times, try this:
$message = preg_replace(
'/^(-{2,})(?=(?:?
|)+)/m',
'<span class="msg_footer">$1</span>',
$message
);