Re: [Ethereal-dev] Re: Kismet protocol dissector

"ronnie sahlberg" <[email protected]> Thu, 13 Jul 2006 09:58:24 +1000
Newsgroups gmane.network.wireshark.devel,gmane.network.ethereal.devel
Message-ID <[email protected]>
On 7/13/06, Krzysztof Burghardt <[email protected]> wrote:
>
> Thanks for your reply.
>
> > 1, please create a WIKI page on wiki.wireshark.org for this protocol.
>
> Done, see it here:
> http://wiki.wireshark.org/Kismet_Client/Server_Protocol
> http://wiki.ethereal.com/Kismet_Client/Server_Protocol


Nice!

However,   port 2501 is registered for the rtsclient protocol   and unless
kismet is the same as rtsclient
it would be incorrect to refer to this as  a well-known port for kismet.


further down on the page the port is referred to as the default port?


if it is a default port  you should add the port as a preference setting
which defaults to 2501  but can be changed by the user.



> 5,   is it possible to change the dissector to be a new style
> > dissector   returning gboolean instead of an oldstyle returning void?
>
> Here is a (small) problem. I changed dissector, so it use gboolean and
> always return true, but I have no idea how to check if it is kismet or
> not. Some suggestions?




Kismet is an ASCII based protocol?
 If so you may check that the first 8 bytes of the tvb (if there are 8 bytes
or more) are actual ascii values  >32 <128  ?


You have a lot of
offset += next_token - line; linelen -= next_token - line; line =
next_token;
can you break these up to one assignment/statement per row  and add a blank
line after each proto_tree_add_text() to separate the dissection block for
each field out from eachothers

_______________________________________________
Wireshark-dev mailing list
Wireshark-dev-IZ8446WsY0/[email protected]
http://www.wireshark.org/mailman/listinfo/wireshark-dev