Re: Persistent NSE registry

["Patrick Donnelly" <batrick.donnelly () gmail com>]

On Tue, Nov 18, 2008 at 12:57 PM, David Fifield <david () bamsoftware com> wrote:
> On Tue, Nov 18, 2008 at 01:08:13AM -0700, Patrick Donnelly wrote:
> > o What should happen if a script scan is aborted? Perhaps the Lua
> > State was left corrupt?
>
> That is a concern. Really the only thing I want to persist across script
> scans is the registry. Everything else can be reinitialized.

You either must keep the state or serialize the registry across script
scans. I'm not sure myself what all the ramifications are for keeping
the state for each script_scan.

I'm not sure that there are any initialization concerns. My worry is
handling the cases where an error was caught by the engine and the
scan was aborted. Should anymore scans be run?

> > o These two lines:
> >     lua_settop(L, 0); // safety, is 0 anyway
> >     lua_rawgeti(L, LUA_REGISTRYINDEX, errfunc); // index 1
> > should be executed each time a script_scan is run. It is possible
> > something was left on the stack from previous runs.
>
> I don't understand what the line
>     lua_rawgeti(L, LUA_REGISTRYINDEX, errfunc); // index 1
> does. It pushes error_function on the stack, but then what? It doesn't
> seem to do anything with it. Is this maybe some old code that didn't get
> updated or do I just not understand?

The error function is used for lua_pcall [1] (used later in the code).
The error function executed before the stack is unwound is at index 1
(the final argument to lua_pcall).

[1] http://www.lua.org/manual/5.1/manual.html#lua_pcall

-- 
-Patrick Donnelly

"One of the lessons of history is that nothing is often a good thing
to do and always a clever thing to say."

-Will Durant

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