Re: Enhanced nsock logging

[David Fifield <david () bamsoftware com>]

On Sat, Dec 08, 2012 at 03:18:14PM +0100, Henri Doreau wrote:
> Hello,
>
> I've been working on improving the nsock logging abilities recently.
> The initial problem was that nsock loglevel could not be easily set to
> a desired value from command line (the easiest was to patch
> nsock_pool.c and rebuild...) and also that this level could not be
> adjusted at runtime by pressing d/D. Additionally, I had this personal
> frustration with the nsock code which directly printed out the
> messages instead of generating log objects.
>
> The patched version of nsock that you can find in my nsock-logging
> branch[1] exports the following:
>   * four different log levels
>     - NSOCK_LOG_DBG_ALL
>     - NSOCK_LOG_DBG
>     - NSOCK_LOG_INFO
>     - NSOCK_LOG_ERROR
>   * getter/setter to the loglevel of a given nsock_pool
>   * a nsock_log_rec structure type that contains the log message as
> well as some context information (date, line, file, function name...)
>   * a way to set a custom logging function that will receive every
> nsock_log_rec object whose level is >= the one set for the current
> nsock_pool.
>
> With this in hand, I've modified nmap, ncat and nping to fix the
> issues mentioned above. Within nsock, logging is now more consistent
> and cleaner. Nmap has a mapping between its own log level and nsock
> (see nmap_adjust_loglevel() in output.cc) , plus this is adaptive.
> Highest log levels can now be enabled with ncat (was only 0 or 1
> before). Things don't change much for nping.

Thank you Henri, good-looking code as always.

The Windows build doesn't like the variadic macro NSOCK_LOG_WRAP. It
fails to build with error C2010:
        nsock_log.h(64): error C2010: '.' : unexpected in macro formal parameter list
I'm also seeing
        include/nsock.h(142): error C2079: 'nsock_log_rec::time' uses undefined struct 'timeval'

--packet-trace is showing me DEBUG FULL messages, when it shouldn't show
that much detail and didn't before.

David Fifield
_______________________________________________
Sent through the dev mailing list
http://nmap.org/mailman/listinfo/dev
Archived at http://seclists.org/nmap-dev/