Re: Compilation compatibiltiy with draft-ietf-hubmib-rfc3636bis-05

On Tue, 5 Sep 2006, Clay Sikes wrote:
> Wow!  I'm extremely surprised that correcting a typo and breaking
> compilation compatibility is ok. 

Clay,

Here is the relevant excerpt from the MIB review guidelines --
specifically, RFC 4181 Section 4.9:

   RFC 2578 Section 10.2 specifies rules that apply to revisions of
   object definitions.  The following guidelines correct some errors in
   these rules and provide some clarifications:

   - Bullet (1) allows the labels of named numbers and named bits in
     SYNTAX clauses of type enumerated INTEGER or BITS to be changed.
     This can break compilation compatibility, since those labels may be
     used by DEFVAL clauses in modules that import the definitions of
     the affected objects.  Therefore, labels of named numbers and named
     bits MUST NOT be changed when revising IETF MIB modules (except to
     correct typographical errors), and they SHOULD NOT be changed when
     revising enterprise MIB modules.

The only potential compilation problem of which I am aware would
arise when a module containing an AGENT-CAPABILITIES definition
references ifMauAutoNegCapabilityBits in a VARIATION clause and
provides a DEFVAL for it.  Have you actually seen this problem?  Or
is there perhaps another problem of which the MIB Doctors were not
aware when they discussed this part of RFC 4181?

Thanks

Mike Heard


> __________________________________________________________________
> 
> From: Clay Sikes [mailto:[email protected]]
> Sent: Friday, September 01, 2006 17:24
> To: Edward Beili
> Cc: IETF Hub MIB Working Group; C. M. Heard
> Subject: Compilation compatibiltiy with
> draft-ietf-hubmib-rfc3636bis-05
> 
>       Hi Edward,
> 
>       First, thanks for all the editor work you are doing on the
>       MIB modules.
> 
>       I attempted to update an implementation of the MAU-MIB from
>       RFC 2668 to that which is specified in this ID and ran into a
>       compilation compatibility issue.  An update from RFC 2668 to
>       RFC 3636 does not have this compilation compatibility issue.
> 
>       In this ID, the ifMauAutoNegCapabilityBits,
>       ifMauAutoNegCapAdvertisedBits, and
>       ifMauAutoNegCapReceivedBits have their syntax changed to use
>       the IANAifMauAutoNegCapBits TC.  I strongly support this
>       change.  However, RFC 2668 and RFC 3636 have different labels
>       for the following named bits for the
>       ifMauAutoNegCapabilityBits object's syntax:
>        1. Bit 8  -- 2668/3636 bfdxPause(8), this ID bFdxPause(8)
>        2. Bit 9  -- 2668/3636 bfdxAPause(9), this ID bFdxAPause(9)
>        3. Bit 10 -- 2668/3636 bfdxSPause(10), this ID
>           bFdxSPause(10)
>        4. Bit 11 -- 2668/3636 bfdxBPause(11), this ID
>           bFdxBPause(11)
>       Compilation of the ID fails because some these labels changed
>       from 2668/3636 to this ID.
> 
>       Although the RFC 2578 Section 10.2 allows the labels of named
>       bits to change, this is generally not a good idea.  In the
>       case of this ID, "over the wire" operation is not effected,
>       but compilation compatibility has been broken.  Please refer
>       to RFC 4181 Section 4.9 top of page 29 for a discussion on
>       changing labels for named bits.
> 
>       Given the fact that back in RFC 2668
>       ifMauAutoNegCapabilityBits have different labels for bits 8
>       through 11 then ifMauAutoNegCapAdvertisedBits and
>       ifMauAutoNegCapReceivedBits, it would seem like there needs
>       to be a different TC in the proposed IANA-MAU-MIB for
>       ifMauAutoNegCapabilityBits to use then what
>       ifMauAutoNegCapAdvertisedBits and ifMauAutoNegCapReceivedBits
>       use to maintain compilation compatibility.
> 
>       I hope this has not already been discussed.  I did a quick
>       glance in the list and didn't see anything.
> 
>       Thoughts?
> 
>       Best Regards,
>       Clay Sikes