john-users - Re: [offlist] Re: inconsistent number of cracked hashes in dynamic

Openwall wordlists collection for password cracking (20+ languages)

[<prev] [next>] [day] [month] [year] [list]

Date: Sun, 8 Apr 2012 16:49:14 -0400
From:  <jfoug@....net>
To: arcctgx <arcctgx@...zta.onet.pl>
Cc: john-users@...ts.openwall.com
Subject: Re: [offlist] Re: inconsistent number of cracked
 hashes in dynamic_6

I have sent a reply to the email list, and a 'conversion' program was sent there also.

The binary salts are causing the problem.   I would strongly recommend any hash that uses binary salt to use the 

$HEX$  salt method.  You 'could' get away with only changing the salts that actually require the $HEX$ type, but that would mean having to 'know' the underlying methods being used by JtR.  This 'can' change over time.

If you convert to using the $HEX$ type salts, then all ambiguity goes away.  NOTE, this likely was not introduced in the older john you have (the md5_gen(6) version).  The $HEX$ will allow any character to be part of the salt.  A : char, a \r \n \t \ even a NULL byte, i.e. true binary data.

Jim.

CC john-users

---- arcctgx <arcctgx@...zta.onet.pl> wrote: 
> > Please provide me with the input files (and whatever dictionary(s) used), and the exact command switches.
> 
> I'm sending you two files: hashes.txt and dictionary.txt. In 
> john-1.7.8-jumbo7 this wordlist cracked 247 hashes from hashes.txt.
> 
> The exact command line I used for john-1.7.9-jumbo5 was:
> john -for=dynamic_6 -w=dictionary.txt hashes.txt
> 
> I'm afraid I don't have anything more to add at this point, everything 
> else was described in my first email to the list. If you have any 
> specific questions, just ask and I'll answer ASAP.
> 
> Thank you for your time!
> 
> Best regards,
> arcctgx