Have a look at Unicode::Collate.

#!/usr/bin/perl

use Unicode::Collate;

my $coll = Unicode::Collate->new;

my $e = "\xe9clair";
my @cmp = qw/ aardvark each eden zarkon /;

for my $d ( @cmp ) {
    printf "%10s ccmp %5s = %d\n", $d, $e, $coll->cmp( $d, $e );
}

__END__
  aardvark ccmp éclair = -1
      each ccmp éclair = -1
      eden ccmp éclair = 1
    zarkon ccmp éclair = 1
[download]

Update:: Another approach would be to use Text::Unidecode, and strip off any accents before doing the comparison.