From: "Tomas Kukosa"
Sent: Thursday, March 11, 2004 7:45 PM
Subject: Re: [Ethereal-dev] BER and Kerberos
> Hi,
> the dissect_ber_named_bistring32() could be done. I have no objection
to it.
> I will look at it when I have time.
It would make it easier for handwritten dissectors of BER based protocols
but would not be a make or break feature..
>
> Maybe, my approach is too universal because I make packet-ber changes
> together with the "ASN.1 to Ethereal dissector" compiler.
Very Cool. If that is to surface then the needs of that compiler takes
precedence over the kerberos stuff (which is but one protocol).
> Maybe, I hav not understood all well (as English in not my native
language)
> but the new dissect_ber_octet_string() does not require octet string
containing ASN.1/BER
> encoded data.
Good.
Can you try to update my packet-kerberos.c i sent to the list to use the new
ber api?
I dont really care what the mechanism is as long as I can still easily
dissect the content of an octet string
as a non-ASN.1/BER structure. See the PAC_LOGON_INFO dissection in
packet-kerberos.c
This is a very important part of the new kerberos, the reason why it was
rewritten in the first place.
> I try to keep all functions with common format, i.e. the last parameter
should be
> dissect_ber_...(..., value_type *value). And the best value type for
OCTET/BIT STRING
> seems to be tvbuff_t.
> But I could make version with callback if it help you. E.g.:
As long as the mechanism still exists to allow the NDR encoded
PAC_LOGON_INFO to be passed across to an NDR
based subdissector I dont care what the mechanism is, callback or **tvb
just as long as it will work.
looking forward to seeing the asn2eth compiler when it gets ready for prime
time.
ronnie sahlberg
