I hear you... (I think)... but I'm on a different point... what's the definition
of 'not-accessible', where is it written and in what context? Note that we are
talking about a TRAP pdu - right? A trap is "different" in that it is (at least
I might argue) not ACCESSED by a management application but, rather, delivered -
in fact - unsolicited.
I know I'm probably showing a great deal of SNMP vx, SMI vx, ASN.1 naivety, ...
and I know you are telling me compiler writers have already shipped an
interpretation... still, I'd like to see where 'not-accessible' is actually
defined for SNMP and how it leads to an interpretation that the index to a table
entry cannot be considered "interesting information".
Harry Lewis
IBM Printing Systems
harryl at us.ibm.com
Ira McDonald <imcdonal at sdsp.mc.xerox.com> on 06/02/99 11:54:05 AM
To: Harry Lewis/Boulder/IBM at IBMUS, mike at peer.com
cc: imcdonal at sdsp.mc.xerox.com, jkm at underscore.com, lpyoung at lexmark.com,
pmp at pwg.org
Subject: Re: PMP> prtAlertIndex as defined in
draft-ietf-printmib-mib-info-04.txt
Hi Harry and Mike,
Thanks for explaining the conflict, Mike. Actually, Harry, the
RFC 1759 MIB never did compile under an RFC 190x (1996 SMIv2)
compliant compiler without changing the MAX-ACCESS of 'prtAlertIndex'
to 'read-only' or deleting the object from the Alert trap
bindings (it never should have been there in the first place
- it's value is included in the instance qualifiers of all
the other bound objects).
Now the 1999 version of SMIv2 (RFC 2578) is also more explicit
that index objects (SMI calls them 'auxiliary' objects) should
NOT be accessible. Thus our problem...
Cheers,
- Ira McDonald
High North Inc