Re: Weird verbosity effect with nse_selectedbyname and post-scanning

[David Fifield <david () bamsoftware com>]

On Wed, Aug 11, 2010 at 02:16:13PM -0400, Patrick Donnelly wrote:
> > > Does someone have a proposed solution to this? Set current to nil after
> > > a round of script scanning? It might still cause a problem when the
> > > engine calls verbosity in the middle of a round. Use a separate
> > > verbosity function for the engine?
> >
> > This also concerns the classic script scanning phase, if a prerule
> > script is selected by name, then we'll have:
> > "NSE: Script scanning 127.0.0.1"
> > (host target is 127.0.0.1)
> >
> > I think that we should go for the simple solution, set current to nil
> > after every thread:close() call. Since current is set in line 746 and
> > should only be valid in the run() function (when scripts are running)
> > and since it's not collected by the garbage, then we should force it
> > to be nil for any *outside* verbosity call especially for the engine.
> >
> > The simple attached patch will clear it, and *perhaps* this will avoid
> > any synchronization problems between referencing pending and running
> > threads etc.
>
> This is a good enough fix in my opinion. You only need to set it to
> nil in one spot though, at the end of the for loop.

Great, go ahead and commit it Djalal.

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