The Perl Toolchain Summit needs more sponsors. If your company depends on Perl, please support this very important event.
========================================
How to get started with One_Penguin-pre2
========================================

Installing Penguin
------------------
  Before any code will be excuted,  you will need the following:

	perl 5.002
	safe 2.03
	IO 1.02
	PGP version something.

  Perl 5.002 will need to built and installed on the system,  as well as the
other two modules.
  You should ftp Safe and IO from the following location:
    
   coriolan.amicus.com/pub/PATCHED_MODULES

These versions have been slightly modified from the originals in order to
help Penguin along...
  This is not the only magic nessercary - Perl5 comes with Safe 1.0 installed,
and you have to get rid of it before you can do anything useful.  You will 
need to delete the following directories:

   ${PERL5LIB}/${ARCH}/5.002/auto/Safe/ and
   ${PERL5LIB}/auto/Safe/

as well as the file ${PERL5LIB}/Safe.pm in order for Safe-2.03 to work. 
This is a rather large hassle,  but it is believed that it will be fixed in
the next release of Perl (note: it wasn't)


Configuring Penguin
-------------------
  Now you have the modules you need, installed in the correct places and
all,  you have to build Penguin.  Go to the top of the Penguin source tree,
and type 'perl Makefile.PL'.  This will create a standard Makefile for your 
system.  Once everything has finished,  you can type 'make' to build Penguin,  
and then 'make install' (You probably need to be root to do this,  as it 
installs it inside your ${PERL5LIB} directory).

The next step along the way is creating a rightsfile.  This is the file
containing the Safe opcodes for each user.  You can start your rightsfile by
typeing 'touch ~/.rightsfile'.
  You must now edit the rightsfile,  it looks intimidating,  but really 
isn't that complicated.  The format is:

   [User Name <email@address.foo>]
   opcodes

To decide what opcodes to give someone,  please read Safe.pod
this will tell you which each opcode means,  and what functions they
allow access to.  
  If you're not sure,  give the user the opcode :default. For instance, if you 
were adding Map Gallo (Felix Gallo's Cat) to your rightsfile,  you could do it 
like this:

   [Map Gallo <mapthecat@amicus.com>]
   :default
 
If you wish to comment your rightsfile,  you can do so with a '#' preceding
the commented line.


Executing a Penguin Application
-------------------------------
  You should be now just about ready to run an application.  To do so,
enter the bin directory underneath the Penguin root.  To try it out,  you 
need a server to use and some working perl code (Write some,  and save it in 
a file called 'test.pl').  Try coriolan.amicus.com (Felix Gallo's system).  
You can achive this by typing:

   client -h coriolan.amicus.com -f test.pl -v

This basically says run the test.pl [-f test.pl] script on 
coriolan.amicus.com [-h coriolan.amicus.com],  and display the results 
nicely [-v, for verbosity].
  The client will disable the screen echo on your machine,  and ask you to
enter your PGP pass phrase.  As soon as you do this,  Penguin signs the
code,  and puts it in a frame.  If Penguin can connect to coriolan.amicus.com,  
it will toss this frame across the network.  The Penguin server unwraps the code 
and reads who sent the code (it uses the PGP signature to verify this).
It then searches in the remote .rightsfile to see what rights your name has.  
Chances are,  it won't find your name, so it will give you default rights.  
After unwrapping the code, it creates a compartment,  where it can safely 
execute your script.  If it finds unsafe operations (ie,  those not specified 
in the rightsfile) then it will return an error message,  otherwise it 
will simply return a 1 (True).  If you want some specific information
returned,  you can always use the perl function 'return' in your scripts. 
  If you got this far, congratulations, you just successfully ran a Penguin
Application.


Starting up your own Penguind
-----------------------------
  Penguind is the counterpart to 'client'.  It receives the information
across the network,  and sets up the secure execution compartment for the
remote code. 
  To run penguind simply type 'penguind' from the command line.  This will
start the penguind server with the default options (No verbosity,  port
8118,  and Interactive).  I suggest you start penguind with the following:

   penguind -v

The -v switch turns verbosity on,  and lets you see what exactly people are
running on your machine.
  After you start penguind,  the program asks you to type in your PGP pass
phrase (You won't see what you are typing,  the same as when you type your
login password).  Penguind will then appear to do nothing, until a client
trys to attach.
  If and when a client attaches,  penguind will print the hostname and port
number that the client is coming from,  and then also print the code it
tries to execute.


Notes
-----
Most of this text was put together this evening (so the it may read a
little strange in places),  and much of it consists of long-winded
ways of saying what has been said before. 
Hopefully this will help people run penguin properly,  the first time
through.  I guess I'll have to completely revamp this text when
One_Penguin-pre3 and finally One_Penguin come along, but it's a start.

Regards,
James.

--
jduncan@hawk.igs.net