Better splitting regexp for test_ordinals

    open(my $fh, '<', $filename);

    while (my $line = <$fh>) {

        my @tokens = split(/( |\t|:)+/, $line);

        my @tokens = split(/(?:\s+|\s*:\s*)/, $line);

        #Check the line looks sane

        if ($#tokens < 8 || $#tokens > 10) {

        if ($#tokens < 4 || $#tokens > 5) {

            print STDERR "Invalid line:\n$line\n";

            $ret = 0;

            last;

        }

        if ($tokens[4] eq "NOEXIST") {

        if ($tokens[2] eq "NOEXIST") {

            #Ignore this line

            next;

        }

        #Some ordinals can be repeated, e.g. if one is VMS and another is !VMS

        $newqual = $tokens[6];

        $newqual = $tokens[3];

        $newqual =~ s/!//g;

        if ($cnt > $tokens[2]

                || ($cnt == $tokens[2] && ($qualifier ne $newqual

        if ($cnt > $tokens[1]

                || ($cnt == $tokens[1] && ($qualifier ne $newqual

                                           || $qualifier eq "FUNCTION"))) {

            print STDERR "Invalid ordinal detected: ".$tokens[2]."\n";

            print STDERR "Invalid ordinal detected: ".$tokens[1]."\n";

            $ret = 0;

            last;

        }

        $cnt = $tokens[2];

        $cnt = $tokens[1];

        $qualifier = $newqual;

        $lastfunc = $tokens[0];

    }