Getting rid of $/ on linux

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

in the sample

 
my $cmd = 'ps efax 2>/dev/null|grep firstbo|grep -v grep|wc -l';

my $stat = $ssh->exec($cmd);
    print "Before chomp: " , length($cmd);
        print "|".$stat."|";
        chop $stat;
        print "|".$stat."|";
print "After chomp: " , length($cmd);
[download]

i need to get rid of the new line which is the result of "wc -l" but neither chomp() nor chop() is working. Can anyone please help me Thanks, -Kamal. Javafan, i apologize for not providing a working script, this is my first time posting.

#!/usr/bin/perl
use Net::SSH::Expect;
use YAML;
use Data::Dumper;
use warnings;
use strict;
use diagnostics;

#####################
# process the input
#####################
my $in = YAML::LoadFile('test_input.yml');

my $version   = $in->{test_instance}{version};
my $build     = $in->{test_instance}{build};
my $host_data = $in->{host}[0];
my $modelnum  = $host_data->{model_number};
my $company   = $host_data->{company};
my $address   = $host_data->{address};

#print Dumper($address), "\n";

my $ssh = Net::SSH::Expect->new(
    ssh_option => "-v -v -v",
    #ssh_option => "-v",
    #host       => $address,
    host => "baseline-dc1",
    #host => "upgrade-dc1",
    #host => "upgrade-3d1",
    #host  => "baseline-3d1",
    #host => "batterup-dc2",
    user     => 'admin',
    #password => 'OPTIONAL',
    raw_pty  => 1,
    #timeout => 2,
    log_stdout     => 1,
        ssh_option => '-o StrictHostKeyChecking=no '
      . '-o UserKnownHostsFile=/dev/null '
      . '-o LogLevel=ERROR'
);
$ssh->run_ssh();
#$ssh->exec("stty raw -echo");
        $ssh->send("sudo -i");
        $ssh->waitfor('Password:', 3);
        $ssh->send("password");


    my $catschedular = $ssh->exec("cat /sys/block/sda/queue/scheduler"
+);
    my $echoconf = $ssh->exec("echo '/usr/lib/bigloo/3.1a' >> /etc/ld.
+so.conf");

my $stat = $ssh->exec("ps efax 2>/dev/null|grep firstbo|grep -v grep|w
+c -l | perl -pe 'chomp'");
open OUT, "> /tmp/STAT.out";
print OUT $stat;
close OUT;
print "stat: |" . $stat . "|\n";
chomp $stat;
my $newstat;
$newstat = substr($stat, 0, 1);
$stat = $newstat;
print "new stat: |" . $newstat . "|\n";
print "old stat: |" . $stat . "|\n";
[download]

result:

0root@baseline-dc1:~# stat: |0root@baseline-dc1:~# |
new stat: |0|
old stat: |0|
before...|0|after...|0|bat@batbot:~/tests/Install_ISO$ more /tmp/STAT.
+out 
0root@baseline-dc1:~#
[download]

as you can see that the last line is 0root@baseline-dc1:~# whereas it needs to be just a 0 hope this helps. i really appreciate ALL your comments and help.

Comment on Getting rid of $/ on linux Select or Download Code

Replies are listed 'Best First'.
Re: Getting rid of $/ on linux by ikegami (Patriarch) on Jul 13, 2010 at 00:43 UTC
If there's a bug, it's more likely to be yours than in Perl. You're printing the length of the wrong variable.	[reply]
Re: Getting rid of $/ on linux by jwkrahn (Abbot) on Jul 13, 2010 at 01:07 UTC
This may work better: `open CMD, '-\|', 'ps', 'efax' or die "Cannot open pipe from ps: $!"; my $stat; while ( <CMD> ) { $stat++ if /firstbo/; } close CMD or warn $! ? "Error closing ps pipe: $!" : "Exit status $? from ps";` [download]	[reply] [d/l]
Re: Getting rid of $/ on linux by graff (Chancellor) on Jul 13, 2010 at 08:45 UTC
Here's a variant on jwkrahn's approach that's closer to what seems to be the real goal of the OP code: `my $ps_output = $ssh->exec( 'ps efax 2>/dev/null' ); my @stat = ( $ps_output =~ /firstbo./g ); my $stat = scalar @stat;` [download] Counting matches locally is just easier. Note that the regex match here uses "firstbo." to make sure it matches "firstbo" only once per line (just like the remote "grep" command would have done).	[reply] [d/l]
Re: Getting rid of $/ on linux by rowdog (Curate) on Jul 13, 2010 at 01:05 UTC
This is just a guess but all this mucking around with line endings makes me think you might want to look over what Net::SSH::Expect has to say about psuedo-terminals.	[reply]
Re: Getting rid of $/ on linux by Fletch (Bishop) on Jul 13, 2010 at 13:35 UTC
Unrelated to your perl problems that others have already pointed out, but here's a handy little trick: Even if you do use `grep` on the remote end there's no need to run it twice. I don't remember where I found this originally, but if you include a character class with a single member you don't need to use the second `-v` invocation to weed out your grep process from `ps` results. E.g. for your example you'd do `grep '[f]'irstbo`; the brackets keep it from matching against the `grep` process in the results (since `[f]irstbo` can't match aganst `grep [f]irstbo` but does match the process you're looking for). The cake is a lie. The cake is a lie. The cake is a lie.	[reply] [d/l] [select]
Re: Getting rid of $/ on linux by JavaFan (Canon) on Jul 13, 2010 at 08:37 UTC
You claim it's not working. But you don't say why. Your code isn't cut-paste-runnable, and you couldn't be bothered to include your output. What is the output, and what do you expect the output to be?	[reply]