Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sun, 24 Jul 2011 17:23:21 -0500
From: "JFoug" <>
To: <>
Subject: Re: md5_gen(17)

I can do split, the then there has to be logic placed in numerous places. 
prepare FIXES the ciphertext string, thus there are no additional changes 
needed, such as salt parsing, bin parsing, etc.

In a simple format, I really do no care how people do it.  In a format which 
has arbitrary salts, arbitrary other stuff, multiple binary formats, etc, 
etc, etc, about the only way to get anything done, other than continuing to 
write tons of fundamental initialzation code, is to get the 'original' 
ciphertext into a single specified order format.  That is what prepare can 
do, but split can not.    Prepare gets a crack at the text, even prior to 
valid.  It is a fix it one time, and be done with it function.  I know you 
have been pushing for split, and yes it 'can' handle some polymorphism 
within a format, when you get to a truely polymorphic format, its 
limitations are very glaring.  In md5_gen, there is quite a lot of state 
data which is NOT available in split.  That data is part of the runtime of 
the format.  It has to be accessed.


----- Original Message ----- 
From: "Solar Designer" <>
To: <>
Sent: Sunday, July 24, 2011 4:44 PM
Subject: Re: [john-dev] md5_gen(17)

> On Sun, Jul 24, 2011 at 04:36:28PM -0500, JFoug wrote:
>> I can 'fix' this up in prepare() in md5_gen.  Not a big deal.
> As I had suggested before, shouldn't we stick to using split() for
> fixups like this, not prepare()?
> Alexander 

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.