```Message-ID: <BLU0-SMTP2804DE644A0B8D708E65534FDE10@phx.gbl>
Date: Mon, 25 Jun 2012 14:26:51 +0200
From: Frank Dittrich <frank_dittrich@...mail.com>
To: john-dev@...ts.openwall.com
Subject: Re: Markov: allow start and/or end to be specified in
percent

On 06/25/2012 02:15 PM, Frank Dittrich wrote:
> I'd really love to get this into jumbo-6.
>
> The code got a little bit uglier, but that can be fixed with a cleanup
> patch later on, moving the new logic into a separate function.
>
> This patch allows to specify start and/or end in Markov mode as percent
> values.
>
> (For now, I only allow integer values (0-99 for start, 0-100 for end,
> where 0 is treated like 100.)
>
> \$ ./john --stdout --markov:200:1%:2%:12 |head -n 1
> End: 2% converted to 5187128
> Start: 1% converted to 2593564
> MKV start (stats=\$JOHN/stats, lvl=200 len=12 pwd=2593564)
> cholplet
>
> \$ ./john --stdout --markov:200:2%:4%:12 |head -n 1
> End: 4% converted to 10374256
> Start: 2% converted to 5187128
> MKV start (stats=\$JOHN/stats, lvl=200 len=12 pwd=5187128)
> coratudo
>
> Allowing floating point values will probably have to be postponed until
> after jumbo 6.

Or should this idea be dropped, and I should allow

--markov=[MODE:]LEVEL:START/END[:LEN]
--markov=[MODE:]LEVEL:START:END[:LEN]

In tghe first case, START could mean an individual TASK out of END total
tasks, and I could compute the correct start offset and end offset like
I currently do with the % values.

This would make distribution of --markov sessions across n nodes easier,
because n can be adjusted to the number of nodes.

On the other hand, the % approach allows to define smaller packages for
slower nodes and larger packages for faster nodes.

Frank
```

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.