Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Mon, 4 Feb 2008 21:36:16 +0300
From: Solar Designer <>
Subject: Re: faster hash file loading

On Mon, Feb 04, 2008 at 06:17:29PM +0000, helleye wrote:
> if we look @ ldr_load_pw_line
> if ((current_pw = db->password_hash[pw_hash]))
> if (current_pw) continue; 
> looks like this one eliminate dupes


> but it might cause square square complexity 

Yes, it might (worst case), but in practice it doesn't unless the input
file is specially crafted to cause hash collisions with JtR or a custom
*_fmt.c fails to provide a reasonable methods.binary_hash[2]() function.

> so if i can promise there arent dupes in the file
> maybe add a flag -nodupes to the args and then to bypass this check ?

Feel free to experiment with this if you like, but to me this is not
worth the confusion caused by having yet another command line option.

That said, maybe - just maybe - there should be some way to control
whether and when dupes are eliminated, for reasons other than speeding
up the loader.

> reducing complexity to linear

It's very close to being linear anyway, except in special cases
mentioned above.

> saw it does it in a big for (index = 0; index < count; index++) loop
> i hope that i can just bypass this check and it wont hurt nothing else

Yes, you can drop this piece of code, and doing so should not hurt
anything else.

Alexander Peslyak <solar at>
GPG key ID: 5B341F15  fp: B3FB 63F4 D7A3 BCCC 6F6E  FC55 A2FC 027C 5B34 1F15 - bringing security into open computing environments

To unsubscribe, e-mail and reply
to the automated confirmation request that will be sent to you.

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.