This section has been unusually silent: no new obfuscated code for seventy-seven days. It is thus fitting that we contemplate together the words of the silent monk who demonstrated us the virtue of patience.

use warnings; use strict; use Math::BigInt;
our $scripture = q{

The First Letter of Erudil to the Monasterians
[download]

1  Erudil,  a  hacker  of   Perl,   To
Vroom,   the  saints,   and  all   the
dwellers   of   the    Monastery:    2
Mercy,  peace   and   love   be  yours
in  abundance. 3 It  is  now  a little
more  than two years  since  I  joined
this community of believers  in  Perl,
and my appreciation  of this place and
its custodians and  members  continues
to grow.  4  Brothers  and sisters,  I
beseech   you  to   continue  fighting
the good fight,  showing patience  and
understanding  to those  who  come  to
the gates  of our  blessed  Monastery,
bereft  of   clues  and   ignorant  of
our ways.  5  Do not rebuke them  with
harsh   words   such  as  "RTFM",  but
rather lead  them  gently  - with URLs
-  so  that they may  learn  wisdom. 6
Suffer  the newbies to  come  unto us,
instructing them  of  the  many truths
contained   within   the  good   book.
7  For who  among us  has  never  been
in  a state  of  error or ignorance? 8
Consider carefully  the  following:  9
Blessed  are they that "use  strict;",
for  their   typos  shall   be  caught
before trouble arises. 10  Blessed are
they that  use "-w",  for  Perl itself
shall  show  them  the error of  their
ways. 11 Blessed are  they  that  "use
CGI;", for their parameters  shall  be
passed  without error. 12  Blessed are
they  that use  CPAN, for  exceedingly
large shall be the  number  of  wheels    
that they do not re-invent.
[download]

13  Blessed  are  they that understand
regexs,  for  they  shall   waste   no
time   trying   to   parse  HTML.   14
Blessed  are you, when  "they"  insult
you and  ridicule you,  and  speak all
manner  of   evil   against   you  for
using Perl,  for  great  will  be your
reward  when  your project is finished
in  less  time,  and with less effort.
15  If there are those  who refuse  to
accept  our wonderful message, do  not
allow   them  to  draw   you  into   a
fruitless  flame   war,  and   thereby
bring  disrepute  upon  yourselves and
our   Monastery.   16  Remember  well:
"What should it profit  a man,  if  he
should  win a flame war, yet lose  his
cool?".   17  If  this   is  an   area
of  weakness   for  you,   as   it  is
for   me,   then   I  would  encourage
you,   my   fellow  Monasterians,   to
consider  a   vow  of   silence  as  I
have done,  speaking rarely  and  with
words  carefully  chosen to edify  and
inform. 18  Brothers  and  sisters,  I
commend  you once again for  restoring
my faith  in  the idea  of  a  viable,
on-line  community.  And  now,  19 May
your   syntax    always   be   correct
and your  algorithms efficient. 20 May
your  references point  to  good data,
and  may  you  grow in  the  knowledge
and  understanding   of  our   beloved
language, Perl. 21 Amen and amen.
[download]

};
my $r = Math::BigInt->bzero;
for (split /^/, $scripture) {
    my($k, $n, $b, $v, $h, $p) = (-1, (0)x5);
    for (unpack "(A)*", $_) { 
        if (ord) { 
            if ($p) {
                $n++; 
                eval {
                    if ($p % 2) { 
                        $b *= $n / ++$k;
                    } else { 
                        $v += $b; 
                        $b *= $n / ($n - $k);
                    } 
                } or $b = $p % 2;
            }
            $p = 0;
        } else {
            $p++;
        }
    }
    eval { 
        $r *= $b * ($n - $k) / ($k + 1);
        $r += $v;
    };
}
while ($r) {
    print+(" ", "a" .. "z")[($r->bdiv(23))[1]];
}
print "\n";
__END__
[download]

Update 2010-12-19: tried to fix motivational sentence at the beginning to be more meaningful. Not sure it worked.

Perl5porters wants us to believe that they care a lots about backwards compatibility. This is, however, not true, as shown by how many of my obfuscations doesn't work anymore with recent perl versions, or they give warnings even when you don't enable warnings.

This node tries to be a comprehensive listing of which of my obfus still work and which don't. I'll try to update this node with newer breakage in the future.

• LZ77 compression works (as of perl 5.16.1).
• One-liner japh doesn't work: perl 5.10.0 breaks it with an unnecessary change that doesn't really fix a bug. It could be fixed somehow, for some of the similar obfus I link to still work.
• Japh algebra works (as of 5.16.1).
• Fun with duff's device and AUTOLOAD doesn't really count, because it was already sometimes broken when I wrote it, but it definitely doesn't work anywhere on perl 5.10.0 or later.
• Charnames japh gives an unnecessary default warning on 5.12.0 or later. (It still gives the right result as of 5.16.1.)
• Indigo japh doesn't work in perl 5.12.0 or later, because split in void context no longer fills @_. As a hotfix, prepend @_ = (that makes it work as of 5.16.1, though since perl 5.12 it also gives a default warning about pop _), but that ruins the aesthetics of the obfu. The second obfu in that same node works (as of 5.13.6).
• Ode for getprotobyname still works, and so does the variant in Re (Obfu generator for): Ode for getprotobyname (as of 5.16.1).
• Re: Pattern matching was already broken when I wrote it because it only worked on 32 bit perls – that's my fault, it wouldn't have been hard to make it work on 64 bit perls too.
• Fast Japh works (as of 5.16.1).
• Digisig doesn't work starting from 5.20, because the old feature it relied on got removed from perl.
• Re: a poem still works (a of 5.16.1).
• Ode for the CORE:: package works (as of 5.16.1), but starting from 5.12.0, it gives a default warnings that reveals something about the obfu. (It probably won't work starting from 5.20.0)
• Faster japh works fine (as of 5.16.1).
• Japh algebra revisited works (as of 5.16.1), but starting from 5.12.0 it gives an annoying default warning.
• Twin-lines japh doesn't work and is completely unfixable: perl 5.13.6 fixes just the bug this uses. (It still works with at least 5.13.4.)
• All of the ones in Re: Fibo Golf challenge on 3 monkeys still work as of ruby 1.9.2.
• JAPH to commemorate the release of Perl 5.10.0 still works (as of 5.16.1).
• All the ones in Short quines still work (as of 5.16.1), but variant e gives a default warning starting from 5.12.0. What's the deal with all these default warnings lately? Isn't the whole point of -w that the developer enables it to see what's wrong with the code but users of my code don't need to see it and worry about it?
• A feature of use still works (as of 5.16.1).
• General quine still works (as of 5.16.1).
• Linux console japh still works (as of 5.16.1), provided you can find a linux console and stand far enough from your huge TFT that you can read the half feet high letters. In particular, while many Linux distributions set up the console as framebuffer text mode (which means the graphics card is set to graphical mode and the Linux kernel emulates a text console on it), but this obfu still works in such cases.
• Re^2: The Oldest Plays the Piano still works (as of 5.16.1). The variants in Re^3: The Oldest Plays the Piano also work.
• It's not my obfu, but only a slight change in Dollar Plus, so it's no surprise that Re: Dollar Plus still works (as of 5.16.1).
• Both variants in Japh algebra, GF(128) edition still work (as of 5.16.1).
• News of current disasters hidden in 1st Monasterians still works (as of 5.16.1).
• \c*\c!\c0\c( still works (as of perl 5.16.1).
• Just another unpack hacker still works (as of perl 5.16.1).

This trend is also not specific to the obfu I write. The following section is of course just a selection, but you monks are encouraged to post tales about your broken code in replies: then maybe p5p will listen to us and stop these incompatible changes if they hear enough people complain. (Yes, I'm not completely serious, but this is the obfu section so I don't have to say that, right?))

• Finite Automata Question Generator needed a one-character fix because of an incompatible change in perl 5.8.1.
• One of the most brilliant obfus ever, can someone help me?, is unique in that it's broken by a removed diagnostic in perl 5.10.0.
• Re: perl v/s C also gives a default warning now, but still works.

Did you know you could evaluate a polynomial over GF(128) in perl with pack and unpack and tr? Here's how.

sub h {
        $_ = $_[0] = pack b208, 0 . unpack b362, $_[0];
        tr/\0-\c?/\0/;
        tr/\0/\377/c;
        $_
}
for (split //, "k6sNP2B}({ambrusLB%Ox)Z]n0*zf\0I3") {
        $y = $r;
        $v = join $r = '', a .. z;
        $r ^= h($r) & "\217" x 26 ^ h($v) & $y for 0 .. 6;
        $r ^= $_ x 26;
}
print $r;
[download]

This code is based on the same idea as Dollar Plus, but written in a different style, so some of the credit goes to martin.

Update: if you prefer one of these unreadable blocks of code, you can of course write this that way too (Update: reformatted blob from four lines to three.)

sub h($){($_=$_[0]=pack b208,0 .unpack b362,$_[0])=~tr/\0-\c?/\0/;tr
/\0/\377/c;$_}do{$y=$r;$v=join$r='',a..z;$r^=h$r&"\217"x26^h$v&$y for
0..6;$r^=$_ x26}for"k6sNP2B}({ambrusLB%Ox)Z]n0*zf\0I3"=~/./g;print$r
[download]

Update: removed lots of spaces from end of lines of formatted code.

$;^=$;;$:='+';${::}{$:}=$_,for{},[];$==65;@+=map{$=<<=1;
$=^=131,if+128&$=;chr$=}0..1<<8;@+{@+}=(0..126)x$=;print
map{$:=$+[$+{$:}+$+{z}];$/=$;;$/=$;eq$/?$_:$+[$+{$/}+$+{
$:}]^$_,for@_;$/}@_=split//,'C|fqF[Ja-Dh{wzZf8t})1N@C:';
[download]

My second contribution here does not like strictures, I admit, but mostly for aesthetical reasons. With a rather simple change it would. And ambrus is right: This kind is somewhat less like minigolf.

$%=2,$"=6;$;=ord"\c?";print+map{$%*="@_[4,4,
]";$=-=$=;$=*=$%%=$;,$=+=ord,$=%=$;for@_;chr
$=}@_=split?,*?,q=|f475!%L(wz)Hmqkcz?ar~x<m=
[download]

This one won't work correctly on an EBCDIC machine, I am afraid. But you can turn on strictures and warnings if you like.

These scripts are the same forwards and backwards, excluding whitespace characters and newlines.

Making a JAPH palindrome was fairly easy using comments.

#!/usr/bin/perl
#__ATAD__
print 'Just another Perl hacker,'#;
#',rekcah lreP rehtona tsuJ' tnirp
__DATA__#
lrep/nib/rsu/!#
[download]

It took me a bit of thinking to make one without the crutch of comment lines.

#!/usr/bin/perl
open fi, ">&STDOUT";
use constant X__ATAD__;
print fi 'Just another Perl hacker,';
',rekcah lreP rehtona tsuJ' if tnirp
;__DATA__ X tnatsnoc esu
;"TUODTS&>" ,if nepo
lrep/nib/rsu/!#
[download]

Now, trying to create a JAPH palindrome without the use of comments or DATA? :shudder:

EDIT:Fixed based on blokhead's advice/corrections.

@a=split//,"Z~R_'QVj!^VL_.E]`OVj-]`O^AaZj>]LH]`OWx";map{$_=(ord$_)-oct
+ 36;
unless($_>oct 11){$_-=3;$_="0$_"};$b.=$_}@a;$b=scalar reverse$b;while(
+$b){
$y=chop$b;$y.=chop$b;$y.=chop$b;push@t,$y}map{print chr($_- hex hex d9
+)}@t
[download]

$l  = 2/77;
$|  = 2;
@F  = (
    '000*000*000*0*000*00*',
    '0*00*00*0*0*00*00*00*0*0*',  
    '00*0*00*00*00*0000*00*',
);
for (@F) {
    $.  = 0;
    while ($g = length) {
        select $i, $j, $k, $l;
        $s  =  $_;
        $s  =~ s/0/ /g;
        $m  =  unpack 'b*', pack 'v', $.;
        $m  =~ s/0/ /g;
        $m  =~ s/\d/*/g;
        print "\r", ' 'x$g, $s, $m, ' 'x2;
        $.  += chop() ? 2/2 : $.;
    };
    print "\n";
};
[download]