Sure.  Message taken on board - once you have the script then it can be reverse 
engineered.  (anyone tried using stunnix to see if it adds *any* obfuscation at all 
once Deparse has had a go at it?)

However, if you package up stuff with the activestate thing then the perl script file 
is at least moderately well hidden, for example there is only one useful reference on 
google on how to reverse engineer it, and the method appears to be inapplicable for 
the latest version.  (It is certainly beyond anyone who is not at least extremely keen 
to look inside)

There is clearly no need to go overboard with XXX bit encryption algorithms, since the 
key is always going to be wired into the file, even XOR will be sufficient if the file 
is compressed.  I was just thinking one could then start to add thin veneers to 
increase security, for example if the key is known or in a fixed place then a simple 
perl script to whack off the header, and grab the key is trivial to write.  However, 
if we split the key up and place it in a slightly different location each time, then 
our perl script has now got substantially more complicated.  I'm sure that we can 
dream up other incremental improvements

There are many hackers quite easily capable of reverse engineering this, and I have no 
intention of aiming at them, however, if one can prevent a customer opening it up in 
winzip then this is a step forward... (if you see what I mean)

Clearly this all depends on NOT having someone who thinks it is useful to pop a 
decompile script on CPAN.  This is clearly feasible and I would not suggest that we 
even try to compete, however, my opinion is that this is not a creative and helpful 
use of peoples time. 

Anyway, not meant to take the focus away from all the other excellent features that 
PAR already has.  

Thanks again for PAR!

Ed W

-----Original Message-----
From: Nicholas Clark [mailto:[EMAIL PROTECTED]
Sent: 29 July 2003 11:00
To: Edward Wildgoose
Cc: [EMAIL PROTECTED]
Subject: Re: PAR 0.70 Released.


On Tue, Jul 29, 2003 at 10:46:13AM +0100, Edward Wildgoose wrote:

> No doubt a decrypt can easily be written, but with luck those people
> with the ability to do so will not actually circulate an example script
> (giving those without, the leg up) and so the integrity will be
> sufficient for many peoples requirements.

Source code recovery has been in core since 5.005_03. The version in 5.8.0
is getting very good:

http://search.cpan.org/author/JHI/perl-5.8.0/ext/B/B/Deparse.pm

Obfuscators (such as Stunnix) will help, in that people have to know to go
to CPAN to find help:

http://search.cpan.org/author/JJORE/B-Deobfuscate-0.10/lib/B/Deobfuscate.pod

Encrypting with a key that the script has to get from somewhere else (such
as a user typing in a startup passphrase) would work.

Nicholas Clark

Reply via email to