Hi,

to emphasis the two aspects of the function you could do a grep (removing totally empty list elements) with a map (removing leading and trailing whitespace).

UPDATE: Taste is hardly debatable, but performance often not:

use strict;
use warnings;
use Data::Dumper;
use Benchmark qw(cmpthese);

# Returns non-empty strings with leading and trailing spaces removed.
sub mytrim_nonempties {
    grep { $_ ne '' } map { s/^\s+//; s/\s+$//; $_ } @_;
}

sub trim_nonempties
{
   ## no critic (ProhibitMutatingListFunctions)
   grep { !/^\s*$/ && s/^\s*|\s*$//g } @_;
   ## use critic
}


my @x = ( ' hello ', "\thello again\t", '', " ", " \t", 'jock', "\t ab
+c", "def \t " );

cmpthese(1000000, {
    'myfunc' => sub {
        my @y = mytrim_nonempties(@x);
    },
    'yourfunc' => sub {
        my @y = trim_nonempties(@x);
    },
});
[download]

McA