Perhaps something like the following might help you:

my $filename = "anything.ldif";

unless (open (LDIF, $filename)) {
  die "Error: couldn't read from '$filename': $!\n";
}
else {
  my %seenObjects = ();

  local $/ = "\n\n"; # read objects, not lines

  while (<LDIF>) { # $_ has one object now
    chomp($_);
    s/\n //g; # kill continuation lines

    if ($seenObjects{$_}++) {
      # object already seen... write to another filehandle
      print OUTFILE "$_\n\n";

    } # if

  } # while
  close (LDIF);
} # else
[download]

(Code not tested)

If the LDIF file is very big, you could either save the Digest::MD5 of the object:

use Digest::MD5 (); # no namespace pollution
...
if ($seenObjects{ &Digest::MD5::md5_hex($_) }++) {
...
[download]

or sort the LDIF (e.g. with Tie::File) and compare one line with its precessor.

Best regards,
perl -e "s>>*F>e=>y)\*martinF)stronat)=>print,print v8.8.8.32.11.32"