As someone once said, "You can program C in any language."

Here is a more Perl-ish solution:

#!/usr/bin/perl -w

use strict;
use warnings 'all';

my @arr = ( 1, 2, 3, 1 );
my %saw = ( );
my @unique = grep { ! $saw{$_}++ } @arr;

use Data::Dumper;
warn Dumper( \@unique );
[download]

That one line...
my @unique = grep { ! $saw{$_}++ } @arr;
...is where all the action is.

We create a hash (%saw) and the only items that make it into @unique are elements that have not been filtered through grep { $saw{$_}++ }.

Kudos to you though for really thinking through this problem. Just goes to show - There's More Than One Way To Do It!

Your problem is on line 42.

Oh, I'm sorry, was there a question somewhere in there? One which can't be answered by checking the Q&A section? If so, maybe you should make it slightly more explicit.

Update: right, posted in CUFP (thanks for the nudge BrowserUk), so no real need for a question. It's still a bit of waste of bits if you ask me, given the existence of far better solutions (which the OP must be aware of, since he posted a more buggy version of this there only minutes late).