Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 08 Nov 2011 00:16:18 +0100
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: post 1.7.8-jumbo-7 changes summary

The original version of the patch segfaulted for me at dynamic_8

The second version has problems with the thin formats:

$ ../run/john -test=0 | grep FAIL
Benchmarking: PHPS -- md5(md5($pass).$salt) [SSE2 10x4x3 (intr)]...
FAILED (get_hash[0](0))
Benchmarking: MediaWiki -- md5($s.'-'.md5($p)) [SSE2 10x4x3 (intr)]...
FAILED (get_hash[0](0))
Benchmarking: PHPass MD5 [SSE2 2x4x3 (intr)]... FAILED (get_hash[0](0))
Benchmarking: Raw MD5 [SSE2 10x4x3 (intr)]... FAILED (get_hash[1](0))
Benchmarking: md5(unicode($p)) [SSE2 10x4x3 (intr)]... FAILED
(get_hash[0](0))
5 out of 104 tests have FAILED

magnum


2011-11-08 00:05, jfoug wrote:
> The original patch had problems. The problem was likely only going to be a
> performance hit, but a performance hit, none the less.  Now, it properly
> does init() once, and make sure it DOES do this init().  There is a -v2 of
> the patch. That version is a replacement for the original simple patch.  
> 
> Jim.
> 
>> From: jfoug [mailto:jfoug@....net]
>>
>> Note, there is a new 'fix' for dynamic.  Under certain situations,
>> init()
>> was not functioning.  I had logic that did a short circuit of init(), if
>> the
>> type did not change.  I removed that so that init will always be run.
>>
>> This showed up in the last format (highest number, i.e. 29), and would
>> happen if the format was being 'forced' to 29.  I think it would also
>> have
>> happened, if forced to the highest number within the john.conf 'script'
>> loaded items also, if that one was force set.
>>
>> 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.