RE: Anyone wants to do a double check on an IPCDN mib?

["C. M. Heard" <heard@pobox.com>]

>>>>> On Tue, 28 Jan 2003, Andy Bierman wrote:
Andy> At 01:20 PM 1/28/2003 -0800, C. M. Heard wrote:
Andy> >I'm not going to do a complete review, however, I do have
Andy> >one question:  have we settled whether this is going to
Andy> >be considered legal syntax
Andy> >
Andy> >   OBJECT docsSubMgtCpeIpAddressType
Andy> >       SYNTAX InetAddressType { ipv4(1) }
Andy> >       DESCRIPTION
Andy> >           "An implementation is only required to support IPv4
Andy> >            addresses."
Andy> 
Andy> 
Andy> This has come up in Cisco MIB reviews. MIB writers want to
Andy> do this but end up changing the TC to the base type name
Andy> to make mosy happy.
Andy> 
Andy> Why would it be considered a feature to support a different
Andy> SYNTAX clause in an OBJECT-TYPE macro than is allowed in
Andy> an OBJECT macro?  This seems like a bug in the SMI (or CONF
Andy> or whatever).

I tend to agree with that.

>>>>> On Tue, 28 Jan 2003, Wijnen, Bert (Bert) wrote:
Bert> I have (this was a while ago) checked with the 3 editors
Bert> of SMIv2 (STD 58) documents and back then they basically
Bert> had rough consensus that this was allowed. I must admit that 
Bert> Dave Perkins had some objections... and his SMICng still
Bert> complains about it. For all I remember he promised me to
Bert> fix it, or at least make it controllable with a switch.

I've been told that the current version of SMICng allows
refinement of enum INTEGER and BITS TCs if you specify /c0f.

Bert> smilint accepts it as far as I know.

Indeed it does.

Bert> In any event, this is what I have been telling MIB authors
Bert> for some 2 years now I think.

If the consensus of the MIB doctors is that the guidelines should
say that this usage is OK, then that's what I will put in the
guidelines.  I would like to say something along the lines of this:  
although a strict reading of the SMI documents would lead a person
to conclude that the usage is not legal, the consensus is that the
omission was an oversight and that the usage should be allowed.  We
should not say anything about it being required because a lot of
standard MIBs -- the IF-MIB is one -- do revert to the base type
in refinements.

Mike