Question

I wrote a perl program which send the updated data from a file to remote server periodically. But now i want it to read it from differnt files and send the updated data such that the reciever should know how to seperate the data from the mixed data. Do i just need to put some kind of delimiter? Is there any standards already there for such things?

#############
#Change parameters
############
$PeerAddr='192.168.0.7';
$PeerPort='7070';
##############
# Import packages
##############
use Text::Diff;
use IO::Socket;
#############
# Define global variables
#############
$lineCount=0;
$loopCount=0;
our $stats2 = 0;
for($count = 0; $count <= 10000; $count++){
        my $data_dir="archive/otat/*dat";
        my $data_file= `ls -t $data_dir | head -1`;
        chomp($changed_data_file);
        print "old data_file is $changed_data_file \n";
        chomp($data_file);
        if($data_file ne $changed_data_file){
                $lineCount2=0;
                $changed_data_file=$data_file;
                print ("String:$data_file :$changed_data_file  are not equal\n");
                }
        while(defined($data_file)){
                print "$data_file \n";
                open (DAT,$data_file) || die("Could not open file! $!");
                @iofile = <DAT>;
                $lineCount = @iofile;
                splice(@diffLines);
                print "printing: $lineCount\n";
                print "printing 2: $lineCount2 \n";
                chomp $lineCount;
                chomp $lineCount2;
                if($lineCount != $lineCount2){
                        $j=0;
                        for($i=$lineCount2;$i <= $lineCount; $i++){
                        $diffLines[$j] = $iofile[$i];
                                $j++;
                }
                        $num=@diffLines;
                        print "count of diff lines:$num\n";
                        $lineCount2 = $lineCount;
                        $loopCount=0;
                }
                if($loopCount>2){
                        $loopCount=0;
                        print "Look for recent file \n";
                        last;
                }
                $loopCount++;
                sleep(5);
############################
                &socket_con(@diffLines);
        }

}
#### Methods/Functions
sub socket_con {
        if ($sock== 0){
                $sock = new IO::Socket::INET (
                                                        PeerAddr => $PeerAddr,
                                                        PeerPort => $PeerPort,
                                                        Proto => 'tcp'
                                                        );
                die "Could not create socket: $!\n" unless $sock;
        }
        print $sock @_;
#close($sock);
}

Answer 1

I’ve used JSON a lot with good results http://metacpan.org/pod/JSON You can store your data in a hash, serialize it, send the text to the client and have it turn the string back into a Perl hash for easy use. For example:

# on the server
use JSON;
...
# store changed lines in a hash
$diffLines->{$data_file}[$j]=$io_file[$i];
...
# Serialize the hash reference into a string which you then send to the client
$diffLinesSerialized = encode_json $diffLines;

# on the clinet
use JSON;
...
# convert received data from serialized string into hash
$diffLines = decode_json $diffLinesSerialized;
# $diffLines is now a has reference which can be accessed like normal
foreach my $data_file (keys %$diffLines) {
  foreach my $line (@{$diffLines->{$data_file}}) {
    ...
  }
}

All this being said, as much as I don’t really like XML from a programming perspective, it is a prevalent standard for this kind of thing. If this is just a specialized, internal tool that won’t grow into something larger it probably doesn’t matter, but if you think this could turn into a more general service, say for non-perl clients, it would be good to consider XML as an option. Programming from a service-oriented perspective can make it easier to grow things down the road.