radagast has asked for the wisdom of the Perl Monks concerning the following question:

I'm working on a script where I create a little server with IO::Socket. It accepts a string from the client side which I split into two, a user and a pass. While debugging and writing these two fields to a logfile, I noticed that if I chomped the $pass variable (at the end of the string initially), I get a ^M at the end but if I don't, the carriage return remains. I am deliberating whether or not to use chop which does remove the ^M but don't want to do so if this is a random event.

Replies are listed 'Best First'.
RE: CHOMP VS. CHOP
by Jonathan (Curate) on Aug 02, 2000 at 13:22 UTC
    Are you connecting to a Unix server from a windoze client? Domestos has a two characters at the end of a line. A line feed and a carriage return. Unix has just the line feed. I wrote a win32/unix client server app recently and didn't have any problems with the conversion. I wouldn't use a chop instead of chomp here but I doubt the carriage return is a random event. I've tried to re-create the problem with the following code on NT and Solaris but I'm afraid it works fine for me?

    Server code on a unix box
    #!/usr/local/bin/perl
use IO::Socket;
print "Waiting...\n";
$socket = IO::Socket::INET->new(LocalPort => 4321, Proto => 'udp');
$socket->recv($text,128);
chomp $text;
print "Got this message: >${text}<\n";
close $socket;

    [download]


    And the client
    #!/usr/local/bin/perl
use IO::Socket;
$x="my string ends here\n";
$sock = new IO::Socket::INET(
        PeerAddr => 'machine_name',
        PeerPort => 4321,
        Proto    => 'udp');

$sock->send($x);

    [download]


    Produces;
    Got this message: >my string ends here<

    In my experience some ftp/udp/tcp clients leave the ^M on and some don't. If you have it check out Russ's node Removing ^M
[reply]
[d/l]
[select]
Re: CHOMP VS. CHOP
by atl (Pilgrim) on Aug 02, 2000 at 12:58 UTC
    Welcome to the world of cross platform programming! :-)

    I don't know if there is a elegant solution to this, but since chop will always take away the last character I don't like it very much. I'd suggest to do something along this:

    # ... read into $line
chomp $line;
$line =~ s/\r$//;
# ... carry on

    [download]

    That should work for any win-unix-combo (client as well as server; code is untestet, though, and might contain some lurking errors ;) )

    Cheers ...

    Andreas

    Update: Think before submitting ... the line about "any win-unix-combo" is BS, I'm sorry. This will work if your client (executing this script) is unix and the server (where you read from) is win or unix.

[reply]
[d/l]
[select]
RE: CHOMP VS. CHOP
by pschoonveld (Pilgrim) on Aug 02, 2000 at 15:16 UTC
    chomp, by definition, removes the contents of $/ from the end of a string. Perhaps, if you know that your client will be windows, you should set $/ to \n\r (or is it \r\n, I can never remember). What it comes down to is that windows uses a \r\n combination as it's end of lines, and *nix appropriately uses \n. The sketchy, but easy way to handle this is to do a regex replacement (s/to replace/replacement/). But, that could cripple you should you want multi line data to stick around. My preference would be to remove any instance of \r (your ^M) and chomp when necesary.
[reply]
      It's \r\n, fo' sho'. Think about it - CR then LF makes up the mystical CRLF in DOSland, and you only need LF in UNIXland.

      Aren't mnemonics great?

[reply]
Re: CHOMP VS. CHOP
by amelinda (Friar) on Oct 13, 2000 at 01:26 UTC
    I just recently had a hairy issue where someone, in their CGI script, had a 
    chop; chop; 
$_ = "$_\n";

    [download]
    which made file uploads from browsers in Windows DTRT, but totally broke my UNIX file-upload thingy. Part of the solution involved changing those three lines to read: 
    chomp; chomp;
$_ = "$_\n";

    [download]
    You might try that here.
[reply]
[d/l]
[select]
RE: CHOMP VS. CHOP
by lindex (Friar) on Aug 02, 2000 at 20:41 UTC
    When ever I have had to deal with those pesky ^M's
    I have found the line  s/^M/\n/g; is always nice :)



    lindex 
    /****************************/
  jason@gost.net, wh@ckz.org
    http://jason.gost.net
/*****************************/

    [download]
[reply]
[d/l]
[select]
      That would change all your ^M^J into ^J^J i.e. every newline has just become two newlines. Fine if that's what you're trying to achieve, if not surely 
      s/^M//g;

      [download]
      would be better.

      Nuance

[reply]
[d/l]

Back to Seekers of Perl Wisdom