I think we should keep the dissector but either rename it to *_legacy
or something like someone suggested or control it via a preference.
We have similar situations for other protocols already that can be
used to highlught some of the options :
In iSCSI we already have a preference (that defaults to the
current/RFC version) but where we can still switch back to obsolete
pre-rfc versions of the protocol.
in Kerberos5 we automatically support both a special pre-rfc1510
version of the protocol that is used by packetcable? as well as the
rfc version of the protocol.
In that case there is no preference setting since it is trivial to
detect/handle the subtle differences at runtime.
(I think one field that is optional in pre-rfc1510 became a mandatory
field in rfc1510 and beyond.)
On Wed, Dec 18, 2013 at 10:33 AM, Evan Huus <eapache@xxxxxxxxx> wrote:
> This was originally filed as bug 9569. The situation is sufficiently
> unusual that I really don't know what the best solution is, so I
> figured I'd ask for general comments from the list. The company who
> created and used the TPNCP protocol (and submitted the packet-tpncp.c
> dissector) wants to reuse that name for a new, different protocol and
> are asking us to disable the old dissector to avoid conflicts. The bug
> has more detail.
>
> https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=9569
>
> Thoughts?
> ___________________________________________________________________________
> Sent via: Wireshark-dev mailing list <wireshark-dev@xxxxxxxxxxxxx>
> Archives: http://www.wireshark.org/lists/wireshark-dev
> Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
> mailto:wireshark-dev-request@xxxxxxxxxxxxx?subject=unsubscribe