my @libfiles;
           GetOptions ("library=s" => \@libfiles);
           @libfiles = split(/,/,join(',',@libfiles));
[download]

use Getopt::Long;
my (@foos, @bars, @bazen, @quuxi);
GetOptions (
    'foos=s'  => \@foos,
    'bars=s'  => \@bars,
    'bazen=s' => \@bazen,
    'quuxi=s' => \@quuxi
);
# Obnoxious first cut podded out

=begin comment

@foos  = split /,/, join ',', @foos;
@bars  = split /,/, join ',', @bars;
@bazen = split /,/, join ',', @bazen;
@quuxi = split /,/, join ',', @quuxi;

=end comment
[download]

My first thought was of a subroutine.

=begin comment

sub c_split {
    split /,/, join ',', @_;
}

@foos  = c_split(@foos);
@bars  = c_split(@bars);
@bazen = c_split(@bazen);
@quuxi = c_split(@quuxi);

=end comment
[download]

Still redundant, and more lines of code. There is also the addition of c_split to the namespace.

The next idea was to have c_split act on the argument, passed by reference.

=begin comment

sub c_split (\@) {
    my $aref = shift;
    @$aref = split /,/, join ',', @$aref;
}

c_split(@foos);
c_split(@bars);
c_split(@bazen);
c_split(@quuxi);

=end comment
[download]

That eliminates the assignments, but all the other objections remain.

Next idea was to pass all the arrays by reference and act on them in place.

=begin comment

sub c_split {
    @$_ = split /,/, join ',', @$_ for @_;
}

c_split(\@foos, \@bars, \@bazen, \@quuxi);

=end comment
[download]

Much better, but now there is a named sub defined which is only called once. It's too specialized to be of much general use.

An anonymous sub? You bet! Here's what I ended up with:

(sub { @$_ = split /,/, join ',', @$_ for @_ }) ->
    (\@foos, \@bars, \@bazen, \@quuxi);
[download]

Zero redundancy and no namespace pollution.