Thanks javiergn, I guess I'm starting to understand how I could work without true transactions.
Unfortunately, as is, your code returns some very odd results that might be caused by the fact a single log file records many successive sessions. If I "zoom" to a period of time where I have only one session, the result is OK, apart from some limitation issues:
'stats' command: limit for values of field 'message_text' reached. Some values may have been truncated or ignored.
'stats' command: limit for values of field 'message_type' reached. Some values may have been truncated or ignored.
'stats' command: limit for values of field 'session_time' reached. Some values may have been truncated or ignored.
Ideally, the output would be a plot showing for each session the command line (legend for example) and number-of-cracked-pwd=f(time).
It means I need to be able to declare: while I'm inside an active session, title=my-command-line (or a part of my-command-line, like my wordlist, my passwd file…)
And then timechart count(Cracked) by title.
It's even better if I can plot every timechart with the same origin (hence every ploted session seems to start at the same time, allowing a better graphical comparison between them).
I know I can do this in just minutes with awk and gnuplot, but the result would be very static (and boring).
I've started to configure field extractions, but I'm not finished yet, I need to test every different cases (markov, incremental, wordlist, single, etc.).
... View more