[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <021201cd4e47$096ac850$1c4058f0$@net>
Date: Tue, 19 Jun 2012 13:12:06 -0500
From: "jfoug" <jfoug@....net>
To: <john-dev@...ts.openwall.com>
Subject: RE: wordlist dupe suppression
I really dislike the worddupe code, being default. Yes, I can see certain
benefit from it, but making it default, .....
>From: Solar Designer [mailto:solar@...nwall.com]
>
>With just 3 lines in a wordlist:
>
>---
>password
>123456
>123
>---
>
>I am getting:
>
>0:00:00:00 loading wordfile w into memory (20 bytes, max_size=83886080)
>0:00:00:00 wordfile had 4 lines and required 32 bytes for index.
>0:00:00:00 4 lines, hash size 256, temporarily allocating 1056 bytes for
>dupe suppression
>
>0:00:00:00 Suppressed 1 duplicate lines and/or comments from wordlist.
>
>BTW, somehow there are empty lines around the "Suppressed ..." log file
>line.
>
>This feels buggy.
I think the buggyness is likely an off by one in the initial computation of
line count. I believe that if the file does not end in \n, the line counts
are correct. If it ends in a \n, then it is 1 larger. This was a work
around for a core if there was no trailing \n line (IIRC). The array of
line pointers was not large enough. I thought the line count was corrected,
after the true load (not the line count check). It looks like the line
count check is being used within this part of logging, and to determine if a
dupe was found or not.
Jim.
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.
