wtritchie:

I'm having trouble figuring out just what you want. Given the string 'abccabbcabc', just what output do you need?

Also, can you elaborate on just how you differentiate between the first character and the second character of 'the same character in a string' given that they are the same character? E.g., are you looking for the positions of each second character?

Update: Clarified who 'you' refers to in this reply.

then...

my $str = 'abccabbcabc';

my %seen;
while ($str =~ /(.)/g) {
    printf "character %s found at pos %d, reps %d\n", $1, pos($str), $
+seen{$1}
        if $seen{$1}++;
}
[download]

Yes, but is that what wtritchie really wants? That's what I'm still in the dark about, and likely to remain so until wtritchie throws some light on the subject.

BTW: ikegami's guess (Update: anent UTF-16 or some such) seems like it might be somewhere near the ballpark; ikegami possesses a preternatural ability for making such guesses.

Update: ikegami has withdrawn his first reply in its entirety, but I'd still be willing to bet a doughnut to a million computrons that something like his UTF-16 guess is not far wrong. But wtritchie holds the key to all this, and further wtritchie sayeth not.

my @chars = grep 0==(++$seen{$_})%2, /./sg;
[download]

For example,

$ perl -E'say for grep 0==(++$seen{$_})%2, "abccabbcabc" =~ /./sg;'
c
a
b
b
c
[download]

My guess to your meaning would be:

print $2 while 'abccabbcabc' =~ m[(c).*?(\1)]g;;
c
c
[download]

As for your purpose, that's beyond me. Care to enlighten us?

That would be, in the example provided by salva, 'abccabbcabc', and used by several other respondents:

c   b

which this provides:

#!/usr/bin/perl
use warnings;
use strict;
# 856624

my $str = 'abccabbcabc';
my @chars = split(//, $str);
my $seen = '-';
for my $char(@chars) {
    if ($char =~ $seen ) {
        print $char . "\t|\t"; # skip the tabs and pipe if desired
    } else {
        $seen = $char;
    }
}

=head Output:
c       |       b       |
=cut
[download]

If you're right, then the following would do:

my @chars = /(.)\1/sg;
[download]