john-users - starting and 'running' the md5-gen format

[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Date: Sun, 13 Mar 2011 15:48:20 -0500
From: "JimF" <jfoug@....net>
To: <john-users@...ts.openwall.com>
Subject: starting and 'running' the md5-gen format

I have been making some signficant changes to the md5-gen format.  It has 
been noted that the current 'syntax' is diffcult to write (due to shell 
quoting), and may be confusing in the near future, as more hash types get 
added to it (such as SHA1, which will be added soon). This post is to hear 
from users.  To find out what format syntax would like best for of this 
multi-format format.

The current method is to:   -format=md5-gen -subformat=md5_gen(number)  And 
in the john.pot file, the lines are listed as:  md5_gen(number)hash:pass

With changes coming soon, this can simply be done 
as -format=md5_gen(number).  The -subformat will no longer be needed.  The 
lines in the john.pot file, still store the same as today.

Now, the problems that have been reported are this:

1. the ( and ) are not nice characters, due to requiring shell quoting for 
most systems.

2. the 'md5' will soon become confusing.   if -format=md5_gen(21) is used, 
does this appear logical, if the format is sha1($p.$s)  ??

So, I ask the users as a whole, what should the selection syntax be for this 
format.

It has been proposed to be -format=generic_number   This can be done, but it 
will cause any pre-built hash files to be 'fixed'.  Also, the lines in 
john.pot will not work.   Now, it could be that we 
allow -format=generic_number to tell john which format to use, but still use 
the md5_gen(#) signature within the input files, and within the john.pot 
file.    This gets past the shell quoting (but would cause anyone that had 
script files written, to possibly have to edit them).  It would allow 
existing input files, and john.pot file to be used.

Even if we go with the -generic_number, it will likely be that the input 
file signature (and john.pot signature) will be generic(number)   This is 
since the hash will but up against signature.

Well, hopefully, I have not been too confusing in asking this.  I hope to 
hear what others think would be best, in going forward.

Jim.