Checked in, with one hf field that were not used removed.
Can you change the dissectors to be new style dissectors (i.e.
returning int and not being void?)
This will allow them to check that the packet actually looks like JXTA
and if not return 0 and tell ethereal "no this is not my protocol, go
try the next one instead".
I also wanted to point you towards heuristic dissectors but see that
you already asked about it and you are obviously aware then why that
is good change to make :-)
On Fri, 04 Mar 2005 21:38:21 -0800, Michael Duigou <ethereal@xxxxxxxxxx> wrote:
> Yep, the previous patch isn't yet applied and already here's another
> version. Adds fancier dissection of message namespaces and element flags.
>
> I've also got a couple of questions:
>
> For fragmented UDP Messages I can't find a preference to reassemble the
> whole UDP message before it is passed to the dissector. How do I enable
> reassembly of UDP fragments?
>
> Can anyone point me at a good sample for a heuristic tcp dissector? I'd
> like to add dissection of TCP streams, but am wondering how to "claim" a
> tcp stream which begins with JXTA's signature value. I've currently got
> a global preference for port numbers, but JXTA generally doesn't use
> fixed ports so I'd prefer to do it heuristically.
>
> I'm also looking for a sample of dissection of a protocol which is
> running ontop of HTTP.
>
> Mike
>
>
> _______________________________________________
> Ethereal-dev mailing list
> Ethereal-dev@xxxxxxxxxxxx
> http://www.ethereal.com/mailman/listinfo/ethereal-dev
>
>
>
>