For phrases, you would have to know the number before they added it, unless you did it with regex, knowing the format of common numbers
I customised StopHumanSpam for myself on one forum to do exactly this, in theory the solution could be applied to spam phrases (I believe it allows regex)
- the code is still in there, I would only need to comment it back in
The sort of numbers I tried to match were mobiles, or long strings of numbers such as these
// 07875007195 should match
// 07875 007195 should match
// 07875 - 007195 should match
// 07875m007195 should not match
// x07875007y should not match
In code, this is the solution I used (don't know how useful it is too you, I could turn this functionality on in StopHumanSpam if needed)
Code:
// years are going to cause an issue, for instance 1977 - 2013
// so first remove exact matches for <not num>[1-9][0-9][0-9][0-9]<not num> (basically, 4 numbers in a row with a space
$message = preg_replace('/[^0-9][1-9][0-9][0-9][0-9][^0-9]/i', '', $message);
// this does allows users to sneak phone numbers in like: 1787 5107 195 (but the phone numbers can not start with 0... which almost all of them do, so thats fine
preg_match('/(?<=[^a-z$£\pL])[0-9]([^a-z$£\pL]*?)[0-9]([^a-z$£\pL]*?)[0-9]([^a-z$£\pL]*?)[0-9]([^a-z$£\pL]*?)[0-9]([^a-z$£\pL]*?)[0-9]([^a-z$£\pL])/i', $message, $matches);
There will always be ways around it (e.g. 0787 5007 195 ), but it's worked pretty well now for years
There's probably a simple regex pattern to match, but it won't necessarily catch the above without catching things like years, so a basic 1 line of regex might not be great
My regex is dire, there are many people with good regex experience
@Mike being one of them, maybe he can provide a solution
