I thought that it’d be easy to break up a httpd.conf file up into multiple files per VirtualHost using a single line of perl. Turns out, I was most mistaken. Here is the one loooong line of perl to perform the explosion.

perl -ne ' if (/<Virtualhost (\*:80|[0-9.:]+)>/ .. /< \/VirtualHost>/) { $data.=$_ }; if (/ServerName/) { chomp; $sn=$_; $sn=~s/\s*ServerName\s*([^\s]+)\s*/$1/; } if (/<\/VirtualHost>/) { open(outfile, “>>$sn.conf”); print outfile $data; close outfile; $data=”" } ‘ < httpd.conf

People keep hammering my sshd with false login requests. I wrote this script which I call nunar_reaper.pl that retaliates against the stupidity in favor of a tarpit. The infamous Dave Dellanave helped out with this one. I still need to fend off imaginary user name attacks, but that’s a little harder.

#!/usr/bin/perl
open(TAIL, "tail -f /var/log/secure|");
while() {
  if(/Failed password for root/) {
  ($ip) = $_ =~ /(\d+\.\d+\.\d+\.\d+)/;
   system "iptables -A INPUT -i eth0 -s $ip  -j DROP"
  }
}

I managed to lop-off the unnecessary space between -e and the expression, yielding a 24-character masterpiece.

    000000000111111111122222222223333
    123456789012345678901234567890123
%1: awk '{printf ""%s"\n",$1};'
%2: perl -ane ' print ""$F[0]"\n" '
%3: perl -pe 's/(\S*).*/"$1"/'
%4: perl -ape 's/.*/"$F[0]"/'
%5: perl -ape 's^.*^"$F[0]"^'
%6: perl -ape's#.*#"$F[0]"#'

Not to be outdone, I immediately came back with a shorter version:

        00000000011111111122222222222
        12345678901234567890123456789
tony%   perl -ape 's/.*/"$F[0]"/'
bumper% perl -pe 's/(\S*).*/"$1"/'

OK, I managed to knock a character off by combining split and print mode. We’re now down to a 25-character perl solution.

Bumper managed to one-up the awk code with this one.

00000000011111111122222222222
12345678901234567890123456789
perl -pe 's/(\S*).*/"$1"/'
awk '{printf ""%s"\n",$1};'

Perl put’s the smackdown on awk by three whole characters. Bumper noted that Perl’s binary is one character longer than awk. So, he thinks we claimed that we should some how normalize this by allowing them to be called as one character aliases. But, awk came before Perl, so the three-character binary name pool wasn’t nearly as exhausted as it is these days. So, I claim that it has to use standard invocation to be a winner.

I was trying to get awk to print the first column of a file and put double quotes around each item. I finally figured out how to do it after breaking out my underused Oreilly “Awk Programming” book. In case you wondered, awk is most efficient at displaying individual whitespace delimited columns from arbitrary files. The syntax is quite simple and easy to remember.

% seq 1 9 | xargs -n 3
1 2 3
4 5 6
7 8 9
% !! | awk ' { print $1 } '
1
4
7

The syntax for slapping double quotes around the colums is a little trickier:

% !-1 | awk ' { printf ""%s"\n", $1 }; '
"1"
"4"
"7"

Little did I know, perl can do the identical thing with the little-known -ane tag.

% !-2 | perl -ane ' print ""$F[0]"\n" '
"1"
"4"
"7"

So, what language requires less typing (unnecessary spaces removed)?

0000000001111111111222222222233
1234567890123456789012345678901
awk '{printf ""%s"\n",$1};'
perl -ane 'print ""$F[0]"\n"'

Awk narrowly edges-out Perl 29 to 31!!!!