use strict; 
use warnings; 
my $X = "a1b2c3d4e5"; # or use File::Slurp
my $s = "(\\w\\d)"; # my pattern match $s
my $m = qr/$s/; # compiled to a regular expression $m

my $counter = 0;

while($X=~s/$m//){
        ++$counter;
        next unless $counter > 32766; # wait for it...
        print "this is the $counter iteration, got $1 \n";
}
[download]

No need to go to those lengths:

$s = '0123456789' x 100000;;

( $m ) = $s =~ m[((?:(?:0123456789){32000}){3})];;

print length $m;;
960000
[download]

But for any given application there's almost certainly a better way of tackling the problem.

---

With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'

Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.

"Science is about questioning the status quo. Questioning authority". I knew I was on the right track :)

In the absence of evidence, opinion is indistinguishable from prejudice.

Hmmm, I thought the OP had problems with 'complex regex recursion limit exceeded'. If he just wanted to match something like (\w\d){32767}, sure.

No, it is worse, it seems its a DNA sequence, and wants Subsequence for Longest_common_subsequence_problem. To cut the problem down, you could use the Needleman–Wunsch_algorithm. Sinds python has some bioinformatic libs, they can be done there. It can be done in Perl, of course, but then the algorithm needs to be coded.