[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: smilint remembers named number in OID value assignment



David T. Perkins wrote:

> The message below is not correct. That is, it is not a warning
> situtuation and the term "object identifier" is not at all
> appropriate.
> 
> I suggest the folloing text, if you still want to emit this
> bogus warning. 
> 
> So, for 
>   bar OBJECT-TYPE 
>      ....
>      ::= { gar foo(3) }
> and
>   soo OBJECT-TYPE
>      ....
>      ::= { gar foo(4) }
> 
> Emit the following messages:
> ./XXX-MIB:165: warning: local label "foo" used for OID value of
> "soo" has been previously used with a different sub-identifier value
> ./XXX-MIB:40: info: previous definition of `snmpDot3MauMgt'

>From my point of view, the messages currently implemented and quoted below are
more precisely describing the facts: such labels always refer to a complete OID
(derived from a subid and a prefix), hence I don't like the term "used with a
different sub-identifier value". E.g. in the following example, it would be
misleading to refer to a sub-identifier instead of an OID, because the
sub-identifier is the same (3) but a warning should be raised because the OID
is different (gar.1.3 vs. gar.3):

bar OBJECT-TPYE ... := { gar boo(1) foo(3) }
soo OBJECT-TPYE ... := { gar foo(3) }


> But again, the above is NOT illegal and it is fine to have
> the following OID values:
>    ::= { gar foo(1) foo(2) bar(3) bar(4) }
>  ...
>    ::= { gar bar(1) guw(2) foo(3) foo(4) }
>  ...
>    ::= { gar foo(2) }
>  ...
>    ::= { gar bar(2) }
>  etc

Ack. That's why we emit a warning, not an error. We had already this discussion
for several other warnings on the libsmi mailinglist. People, who only want to
know whether a MIB module is formally and syntactically correct (without any
additional "hints"), should use the "-l" option to set the threshold level of
displayed messages.


> On Tue, 30 Nov 2004, [ISO-8859-1] Frank Strauß wrote:
> 
>> Thanks to all who contributed to this issue. I've just comitted a new
>> revision to the libsmi SVN repository that handles OID and identifiers in the
>> "NameAndNumberForm" more acurately, and also emits warnings such as...
>> 
>> ./XXX-MIB:165: warning: object identifier for label `snmpDot3MauMgt' does not
>> match previous label usage
>> ./XXX-MIB:40: info: previous definition of `snmpDot3MauMgt'
>> 
>> ./XXX-MIB:165: warning: label `snmpDot3MauMgtXXX' does not match previous
>> label `snmpDot3MauMgt' for same object identifier
>> ./XXX-MIB:40: info: previous definition of `snmpDot3MauMgt'