I would also like us to consider addressing the following two
possibilities for 1.0:
1) Splitting the libethereal backend and the wiretap library
cleanly free of ethereal so they can be used by other programs.
2) Improving ethereals capacity to scale with very large captures.
Let's face it, the current algorithms and data structures we
use, while fine for small to medium size captures, scale
really badly.
Other Comments inline...
On Fri, 13 Jul 2001, Ronnie Sahlberg wrote:
> Hi list.
>
> Some thoughts of what I would like to see in a future 1.0 release:
> (very incomplete list)
>
> 1, More documentation:
> When defining a preference option, as say the IPv4 reassemble button, one
> should also specify an additional
> parameter which contains a long long text string, or as an alternative a
> pointer into a text file.
> This would be implemented so that in addition to the button in preferences
> which toggles the option on/off
> there would be a HELP button which would display the specified text.
> Would perhaps be easy for the GTK gurus to implement.
>
> 2, Help system, online documentation
> Index, Examples and Tutorial.
>
> 3, More datagram reassembly
> IPX, NBDGM, the three TRANSACTION functions in SMB, fragments in PGM should
> get reassembly support.
>
> 4, Infrastructure for reassembly of streams
> (Needs small changes/additions for packet-tcp and reassembly to fit in
> nicely)
> so we easily can get reassembly for
> NBSS, RPCoverTCP, (SPX?)
>
> 5, TVBuffification of the remaining dissectors.
>
> 6, Check existing dissectors so all fields have a corresponding hf_ entry
> and are searchable.
This is a big task ( but a VERY important one I think ). One of the
big selling points of ethereal to my coworkers is the filtering
capacity. Without the hf_ enteries to tag ALL of the the data there
is not searchability.
>
> (stuff that less people really care about. I would want it though)
> 7, NFS filehandle->filename snooping.
> Implement snooping of NFS packets to determine fhandle->filename mapping of
> nfs files.
> Update the dissector for NFS filehandles to show the filename as well for
> all fhandles where a match is found.
> Perhaps as a generic API so the same code can be reused for
> SMB,AFS,DNS,Ethernet->IP, Ethernet->hostname, ...
>
>
> 8, Informing the web/network-magazines when 1.0 is released. A lot of people
> would use ethereal if only they knew it existed.
> In my opinion ethereal is really today much more competent and useful for
> analyzing captures than almost anything else out there.
> It is an honour to me to be able to contribute to such an extremely
> competent and useful tool.
This is a good idea.
>
> 9, Have even more fun using and hacking ethereal.
>
>
> best regards
> ronnie sahlberg
>
>
>
> _______________________________________________
> Ethereal-dev mailing list
> Ethereal-dev@xxxxxxxxxxxx
> http://www.ethereal.com/mailman/listinfo/ethereal-dev
>