Regex in PHP not working

Question

My regex is:

$regex = '/(?<=Α: )(([\w-\.]+)@((?:[\w]+\.)+)([a-zA-Z]{2,4}))/';

My content among others is:

Q: Email Address 
A: name@example.com

Rad Software Regular Expression Designer says that it should work.

Various online sites return the correct results.

If I remove the (?<=Α: ) lookbehind the regex returns all emails correctly.

When I run it from php it returns no matches.

What's going on?

I've also used the specific type of regex (ie (?<=Email: ) with different content. It works just fine in that case.

Which functions are you using to parse the regex? preg_*, eregi_*? — phpisuber01, Apr 25 '13 at 19:45
The `A` in your regular expression has some kind of diacritical on it, the `A` in the content is a normal letter. — Barmar, Apr 25 '13 at 20:00
@Barmar The diacritical doesn't display in my browsers... it looks like an `A` character. — AbsoluteƵERØ, Apr 25 '13 at 20:30
@AbsoluteƵERØ It's not actually a diacritical. When I look at the letters side by side, this one is just taller. When I paste it into Emacs, it says it has code #x51371. — Barmar, Apr 25 '13 at 20:31
@Barmar Good to know. I should really stop using Windows for coding. — AbsoluteƵERØ, Apr 25 '13 at 20:35

AbsoluteƵERØ · Answer 1 · 2013-04-25T20:20:41.887

This is my newer monster script for verifying whether an e-mail "validates" or not. You can feed it strange things and break it, but in production this handles 99.99999999% of the problems I've encountered. A lot more false positives really from typos.

<?php

$pattern = '!^[^@\s]+@[^.@\s]+\.[^@\s]+$!';

$examples = array(
  'email@email.com',
  'my.email@email.com',
  'e.mail.more@email.co.uk',
  'bad.email@..email.com',
  'bad.email@google',
  '@google.com',
  'my@email@my.com',
  'my email@my.com',
);


foreach($examples as $test_mail){
    if(preg_match($pattern,$test_mail)){
      echo ("$test_mail - passes\n");   
    } else {
      echo ("$test_mail - fails\n");                
    }
}

?>

Output

email@email.com - passes
my.email@email.com - passes
e.mail.more@email.co.uk - passes
bad.email@..email.com - fails
bad.email@google - fails
@google.com - fails
my@email@my.com - fails
my email@my.com - fails

Unless there's a reason for the look-behind, you can match all of the emails in the string with preg_match_all(). Since you're working with a string, you would slightly modify the regex slightly:

$string_only_pattern = '!\s([^@\s]+@[^.@\s]+\.[^@\s]+)\s!s';

$mystring = '
email@email.com - passes
my.email@email.com - passes
e.mail.more@email.co.uk - passes
bad.email@..email.com - fails
bad.email@google - fails
@google.com - fails
my@email@my.com - fails
my email@my.com - fails
';

preg_match_all($string_only_pattern,$mystring,$matches);

print_r ($matches[1]);

Output from string only

Array
(
    [0] => email@email.com
    [1] => my.email@email.com
    [2] => e.mail.more@email.co.uk
    [3] => email@my.com
)

What does this have to do with the question, which is about the `A:` lookbehind? — Barmar, Apr 25 '13 at 20:07
@Barmar I was getting to it. Sorry for the delay... I only write 160wpm. — AbsoluteƵERØ, Apr 25 '13 at 20:22

score 1 · Answer 2 · answered Apr 25 '13 at 20:06

You are not most likely not using DOTALL flag s here which will make DOT match newlines as well in your regex:

$str = <<< EOF
Q: Email Address 
A: name@example.com
EOF;
if (preg_match_all('/(?<=A: )(([\w-\.]+)@((?:[\w]+\.)+)([a-zA-Z]{2,4}))/s', 
                   $str, $arr))
   print_r($arr);

OUTPUT:

Array
(
    [0] => Array
        (
            [0] => name@example.com
        )

    [1] => Array
        (
            [0] => name@example.com
        )

    [2] => Array
        (
            [0] => name
        )

    [3] => Array
        (
            [0] => example.
        )

    [4] => Array
        (
            [0] => com
        )

)

score 0 · Answer 3 · edited May 23 '17 at 10:30

0

Outside of your regex issue itself, you should really consider not trying to write your own e-mail address regex parser. See stackoverflow post: Using a regular expression to validate an email address on why -- upshot: the RFC is long and demanding on your regex abilities.

edited May 23 '17 at 10:30

Community

1
1

answered Apr 25 '13 at 19:47

Yes, I've seen the monster of a regex needed to correctly extract RFC email addresses! – jimmy Apr 25 '13 at 19:50

score 0 · Accepted Answer · answered Apr 25 '13 at 20:12

0

The problem is that your regular expression contains Α, which has an accent over it, but the content contains A, which doesn't. So the lookbehind doesn't match.

I change the regex to:

$regex = '/(?<=A: )(([\w-\.]+)@((?:[\w]+\.)+)([a-zA-Z]{2,4}))/';

and it works.

answered Apr 25 '13 at 20:12

Barmar

741,623
53
500
612

You have got to be kidding me.... An hour spent trying to find out what's wrong and in the end I'm looking for a Greek 'A' when I want an English one... – jimmy Apr 25 '13 at 20:20

score 0 · Answer 5 · answered Apr 25 '13 at 20:39

0

The A char in your subject is the "normal" char with the code 65 (unicode or ascii). But The A you use in the lookbehind of your pattern have the code 913 (unicode). They look similar but are different.

answered Apr 25 '13 at 20:39

Casimir et Hippolyte

88,009
5
94
125

Regex in PHP not working

5 Answers5