Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Date: Wed, 29 Sep 2010 17:23:34 -0700
From: Corbin Simpson <mostawesomedude@...il.com>
To: john-users@...ts.openwall.com
Subject: 2ch/4ch-style tripcodes?

I'm trying to generate tripcodes, which are little slices of crypt()
used to identify posters on 2ch, 4chan, and other image boards in that
same style. The actual format is pretty simple; a small substitution
is done to the key to obtain the salt, another substitution is done to
remove invalid characters, and then the entire thing is run through
crypt().

Trying to do this "by hand" is pretty easy. I can get around 160K
trips/s with glibc's crypt() and 250K trips/s with OpenSSL's
DES_fcrypt(). Pretty decent. However, after reading some literature,
I'm fairly convinced that my Core i3 is probably capable of squeezing
out somewhere between 750K-1.5M trips/s, with the proper crypt()
implementation.

Since John doesn't export a dynamic-linkable version of crypt(), but
instead has its own internal formats tables, I decided to try to write
a new hash format that would handle the small alterations needed to
crypt and work within John's framework. I have come up against a minor
roadblock, though. John's format hooks expect the salt to be part of
the ciphertext, not part of the plaintext, and the external mode hooks
don't let me specify salt, although they do permit munging the
plaintext.

I kind of doubt anybody in the community's worked on this specific
problem, but is there any advice out there for me?

~ C.

-- 
When the facts change, I change my mind. What do you do, sir? ~ Keynes

Corbin Simpson
<MostAwesomeDude@...il.com>

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.