In the previous solution we passed the references to the function but returned a full array, thereby copying all the values. If we care about memory and speed we might eliminiate this copying by returning the reference to the resulting array.

The cost is a small (?) inconvenience as now we have to dereference the resulting array reference in the calling code.

examples/references/add_arrays_nocopy_return.pl
#!/usr/bin/perl
use strict;
use warnings;

my @first   = (2, 3);
my @second  = (7, 8, 5);
my $res_ref = add(\@first, \@second);
print "@{ $res_ref }\n";

sub add {
    my ($one_ref, $two_ref) = @_;

    my @result;
    foreach my $i (0..@{ $one_ref }-1) {
        if (defined $two_ref->[$i]) {
            push @result, $one_ref->[$i] + $two_ref->[$i];
        }
        else {
            push @result, $one_ref->[$i];
        }
    }
    foreach my $i (@{ $one_ref }..@{ $two_ref }-1) {
        push @result, $two_ref->[$i];
    }
    my $ret = \@result;
    return $ret;
}