Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Thu, 30 May 2013 10:23:04 +0200
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: 333 (half-badass number?)

On 30 May, 2013, at 9:26 , Frank Dittrich <frank_dittrich@...mail.com> wrote:
> On 05/30/2013 12:41 AM, magnum wrote:
>> In bleeding, there is limited wildcard and class support since some time, for example:
>> 
>> ../run/john -test -format:cpu
>> ../run/john -t -form:gpu
>> ../run/john -t -form:cuda
>> ../run/john -t -form:opencl
> 
> Is there an easy way to test whether --format=cpu|gpu|opencl|cuda is
> supported or not? (So that I can provide bash completion if it is
> supported, but don't suggest completions that aren't supported?
> Should I just check --list=build-info and enable it if the format
> interface version is >= 11 and we have a jumbo version?

Actually it also works in unstable. Unstable's bad behaviour with GPU when running --test was the initial reason I wrote it - now you can run "--t=0 -form:cpu" to quickly self-test all CPU formats in a GPU build whereas just -t=0 ends up in fireworks.

Not sure what to test for, I never thought of that. Perhaps I should add a cryptic hint to usage blob:

--format=NAME             force hash type NAME: AFS LM RAR5 agilekeychain

--format=NAME[*]          force hash type NAME: AFS LM RAR5 agilekeychain

...and if the latter is seen, there is also a --list=format-classes (to be implemented).


>> ../run/john -t -form:hmac-*
>> ../run/john -t -form:krb5*
> 
> Currently bash completion does this:
> $ ./john --format=hmac-
> hmac-md5     hmac-sha224  hmac-sha384
> hmac-sha1    hmac-sha256  hmac-sha512
> 
> Should I add a hmac-* (of course, only if there already are multiple
> possible completions)?
> Should this also depend on jumbo + format version >= 11?

That would be nice. Same test as the classes.


>> We could add support for leading wildcard too, like -format:*sha256. I saw no need for it at the time.
> 
> --format=*256*
> 
> I think I'll not support that in bash completion, even if you decide to
> add support for it.

Sounds reasonable :-)

magnum

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.