I tried the first solution, and it didn't work properly. I haven't tried the second, but here is a non-regex routine that I wrote to do that:

sub relpath($)
{
    my $file = shift;
    my $newfile = "";
    $file =~ s/\n//g;  # remove new lines
    my @peices = split /\//, $file;
 
    for (my $i = 0; $i <= $#peices; $i++)
    {
        if ($peices[$i + 1] eq "..")
        {
            last;
        }
        if ($peices[$i] ne ".")
        {
            $newfile .= "/$peices[$i]";
        }
    }
    $newfile =~ s/\/\//\//g;
    return $newfile;
}
[download]