Re: Missing nsp_settrace()

[Henri Doreau <henri.doreau () gmail com>]

2013/1/23 David Fifield <david () bamsoftware com>:
> On Tue, Jan 22, 2013 at 05:14:18PM -0600, Daniel Miller wrote:
> > On 01/22/2013 04:59 PM, Gisle Vanem wrote:
> > > This function is in nsock.h and used in 'examples/nsock_telnet.c',
> > > 'examples/nsock_test_timers.c' (and 'ncrack.cc'). Also a comment
> > > in nmap.h refers to it. But the implementation of it is gone.
> > >
> > > My old ncrack repo had it though. What's the story here? Too bad
> > > if tracing is deprecated.
> >
> > This was probably related to Henri's enhanced Nsock logging project
> > (http://seclists.org/nmap-dev/2012/q4/401) that got merged in
> > r30403.
>
> Looking at this now, I think that there should be a default logger
> function that writes to stdout, so that you don't have to make a logger
> function in test programs. Just setting the loglevel should be enough.
> To go with this, nsock_loglevel_t should have a NSOCK_LOG_NONE with the
> value 0, so that by default nothing is logged even though there is a log
> function set.
>
> What do you think, Henri?
>
> David Fifield
Hi,

thanks fixing the nsock examples, I totally forgot them when switching
to the new logging system. Sorry for the nsock header, that was ugly
to forget this declaration there! :)

I didn't add such a NSOCK_LOG_NONE  because errors messages could not
be discarded initially and I wanted to preserve the behavior. I can
add a quiet mode but I wouldn't use it by default. Otherwise nsock
could fail a delivering events without any explanation.

There's a default logger (function nsock_stderr_logger() in
nsock_log.c), but it writes on stderr. Is this a problem? IIRC this is
what old nsock trace did too.

Regards

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