Re: How to start exporting Nessus results to a DB

On Mon, Feb 24, 2003 at 09:51:00AM +0100, Javier Fernandez-Sanguino wrote:
> Re-reading my mail I believe it might not be understood properly. I 
> meant that the code in save_mysql.c which was previously called from 
> nessusd.c/attack.c directly is now called through an abstraction layer 
> that is coded in save_db.c so really:
> 
> nessusd.c --\                      --- save_mysq.{h,c} (only one atm)
>              ---> save_db.{h,c} ---|
> attack.c  --/                      |-- save_pg.{h,c} (planned)

[...]

I will start to have a closer look at the code when I'm done with the
release of Nessus 2.0, but I really don't see any reason why the
database hooks should go in attack.c rather than save_kb.c (which
receives basically everything information-wise). 

The reason I'd prefer to see the code in save_kb is that attack.c is
very complex already, and I'm in the process of simplifying the code as
much as I can (1113 lines in 1.2.x, 887 in 2.0.x and hopefully even less
than that in 2.2.x), so adding stuff in it makes the future merge of
NESSUS_SQL much harder (there will be conflicts if we merge by 2.2).

OTOH, the save_kb stuff is a set of hooks which probably won't change
much over time, because it's supposed to be modular enough and clean.

Maybe I'm wrong though, give me some time and I'll check if what I say
is completely bogus or not.


				-- Renaud