Regex: continuous matching positions

[theshz]

Hi, I'd like to find all occurances of a pattern and its start/end matching positions:

# To find all occurance of "bc" preceeded by 3 letters, and followed b
+y 2 letters.
$str = "aaabcabcdebcaabebcb";
while($str =~ /(\w{3}bc\w{2})/g){
    print "$-[1],$+[1],$1\n";
}
__OUTPUT__
0,7,aaabcab
7,14,cdebcaa

It didn't quite do what I want. I'd like it to match 3 times:

__OUTPUT__
0,6,aaabcab
3,9,bcabcde
7,13,cdebcaa

Thanks.

[Tanktalus]

Check out the pos function. For example:

$str = "aaabcabcdebcaabebcb";
while($str =~ /(\w{3}bc\w{2})/g){
    print "$-[1],$+[1],$1\n";
    pos($str) = $-[1]+1;
}

This sets the position to start looking for the next match to 1 past the start point of the previous match. And it gets you the output you want.

[duff]

Use zero-width lookahead and lookbehind:

$str = "aaabcabcdebcaabebcb";
while($str =~ /(?<=\w{3})(bc)(?=\w{2})/g){
    print "$-[1],$+[1],$1\n";
}

update: As my esteemed fellow monk Tankalus reminds me, you'll need to adjust your mechanism for retrieving the exact characters you require.

[japhy]

The problem is that the regex consumes the characters it matches (so the next match happens after those characters). Use a look-aahead to not consume:

while ($str =~ /(?=(\w{3}bc\w{2}))/g) {
  print "$-[1],$+[1],$1\n";
}

Update: extra ')' added to regex (thanks SamCG and prasadbabu!)