[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 15 Mar 2011 03:40:12 +0300
From: Solar Designer <solar@...nwall.com>
To: john-users@...ts.openwall.com
Subject: Re: unique
On Mon, Feb 28, 2011 at 09:23:20AM -0600, jfoug wrote:
> Unique is nice, but you HAVE to bump up the params.h values a lot to make it
> work, on anything other than a tiny file.
You mean, to make it work "fast". Because it does work on huge files
even with default settings, just slowly.
Also, I bumped up the defaults between 1.7.4 and 1.7.4.2. Starting with
that version, unique uses about 70 MB of RAM by default (previously, it
would use about 9 MB by default). With these settings, unique is
reasonably usable on wordlists of up to a few hundred megabytes. So
maybe your comment applies to a pre-1.7.4.2 version.
That said, I think a future version should have the memory buffer size
tunable from the command-line or/and in john.conf. This has been on my
to-do for ages.
> I also hacked on this and added some things (but never released this code).
> I added a few extra command params.
>
> -v 'verbose' Lists some stuff. If not verbose, it at least outputs # of
> lines read, and # output.
OK.
> -inp=file Allows file to be read, vs only reading from stdin.
What for? Is it somehow more useful on Windows (IIRC, you're on Windows).
> --2nd_uniq_file=file Keeps any lines found in this file, from ending up
> in your output.
> --2nd_uniq_file_only=file Similar to prior, but assumes our input file is
> already unique, so does not unique against self, only against the 2nd file.
We need to come up with self-explanatory names for these options.
> Guess I should get these changes to Solar at some time, lol. They do aide
> in wordlist maintainance.
Sure, please contribute - preferably upload a patch to the wiki.
Thanks,
Alexander
Powered by blists - more mailing lists
Powered by Openwall GNU/*/Linux -
Powered by OpenVZ
