> Hi Mike,
>
> I agree that the current I-D breaks SMI, strictly speaking, BUT
> any future I-D MUST be backward compatible with RFC 1759
> 'over-the-wire'. Removing the (redundant) index object from
> the trap binding is NOT backward compatible. It's note
> the IETF's SMI rules that are paramount here. It's the
> PWG's commitment to backward compatible extensions to
> RFC 1759 only in Printer MIB v2.
>
> Cheers,
> - Ira McDonald
I think you are compounding one error with another. Clearly, the bug is
in rfc1579 which included this object in the nofication when it should
not have. Opening up the MAX-ACCESS to the object now permits this
object to be included in get/getnext requests which also changes
operations and affects implementations which did not require support for
get/getnext operations for this object.
May I suggest an alternative. Comment out the object in the
notification and add some text in the notifcation description which
states that for backward compatablilty with Rfc1579, the offending
object MAY be included.
Some thing like:
printerV2Alert NOTIFICATION-TYPE
OBJECTS { --prtAlertIndex, -- prtAlertSeverityLevel, prtAlertGroup,
prtAlertGroupIndex, prtAlertLocation, prtAlertCode }
STATUS current
DESCRIPTION
"This trap is sent whenever a critical event is added to the
prtAlertTable.
RFC1579 incorrectly included prtAlertIndex in the objects list.
It may be included in the varbind list of the corresponding trap
for backward compatability with RFC 1579."
::= { printerV2AlertPrefix 1 }