Status Report #10 of 15

[Dražen Popović <drazen.popovic () fer hr>]

Hi, everyone.
Last week was my "off" week. I used it to finish up with my exams. I'm
glad to say that it all passed well and now I can go back to work. I'm
very enthusiastic about the rest of the summer as now I can concentrate
150% on my work. Still I managed to get some work done:

Accomplishments:
      * Finished implementing structures and arrays, in the mixed
        context. Now you can play around all you want with different
        types and dimension of arrays. You can nest structures, of
        course you can put arrays in them and it will all work just
        fine.
      * Started with the pointer packing and for now I have the
        reference pointers functional. Whats left are the full pointers
        and the unique pointers which will simply be derived from the
        full pointers.

Priorities:
      * Finish pointers.
      * Add union packing, which is fairly simple.
      * Add the string packing, which is similar to array packing with
        few exceptions.
      * Start with unpacking.

Notes:
When I use a phrase like: "finished packing structures and arrays" it
doesn't necessarily mean that it's completely over. As I move on to
implement packing methods for some type, you often must return to
structures and arrays (and other constructed types) as most of NDR types
are differently packed when standalone versus when they are embedded in
a structure, for example.
Furthermore, I try to develop some test cases which will test every
issue I can think off...obviously I cannot cover everything, so what
would be fantastic If I was to make a document describing Windows RPC
programming in NSE. This would provide a good base for all that wish to
collaborate with Ron, me and others on making SMB and MSRPC scripts.
When I started with packing pointers I noticed something, NDR IMHO is
really complex, and it's this way so it provides some optimal
performance. It's somewhat clear to me why perhaps a long type should be
aligned to a longword boundary, but that is where my comprehension of
the whole thing ends. So for now I think it's best to get the NDR over
with without paying too much attention to the actual performance.

Cheers,
Dražen.


-- 
Laboratory for Systems and Signals
Department of Electronic Systems and Information Processing
Faculty of Electrical Engineering and Computing
University of Zagreb
_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev
Archived at http://seclists.org/nmap-dev/