Re: [nfsv4] Verified editorial Errata on 5561 and sesqui-msns

[Chuck Lever <[e-mail]>]

> On Jul 30, 2020, at 10:36 AM, Tom Talpey <[e-mail]> wrote:
> 
> On 7/30/2020 10:25 AM, Magnus Westerlund wrote:
>> Hi Chuck,
>> If Errata 2249 despite its status as a verified editorial errata is not editorial then lets leave it out of these changes. I really only want to propose things that are truly editorial and have the potential for cause issues for an implementor using this specification.
> 
> Totally agree with the goal, but this erratum is 10 years old (!) and
> was filed by someone who well-understands the situation. At a minimum
> it seems prudent to update the text to clarify this, perhaps while
> making a statement of fact that implementations to an earlier version
> might have differing behavior.
> 
> JMHO.

To be clear, my hesitancy is not with the reporter's expertise but with the length of time the error has been left uncorrected. No doubt it is a simple cut-and-paste typo.

However, the error has been uncorrected for 10 years. That opens the window for implementations that stick to the word of the spec rather than its intended meaning. I think we cannot change that language without due diligence or a thorough understanding of the interoperability implications. Certainly well within the purview of rfc5661bis, but I feel it is outside the arena of corrections that can be done without harm just before publication.

I'm open to changing my opinion if someone with experience in that area can convince us the proposed text change is totally harmless.

--
Chuck Lever