Re: [netmod] References Re: RFC 6991

[tom petch <[e-mail]>]

From: Jürgen Schönwälder <[e-mail]>
Sent: 10 January 2022 20:49

On Mon, Jan 10, 2022 at 04:34:41PM +0000, tom petch wrote:
>
> <tp>
> The definition of 'identifier' in RFC7950 is not the same as the definition of 'identifier' in RFC6020 and rfc7991-bis quotes the latter and not the former.  I am not clear what the best way forward is.
>

It helps me and likely also others if we can nail things down and be
explicit. Are you talking about:

RFC 6020:

   ;; An identifier MUST NOT start with (('X'|'x') ('M'|'m') ('L'|'l'))
   identifier          = (ALPHA / "_")
                         *(ALPHA / DIGIT / "_" / "-" / ".")

RFC 7991:

   identifier          = (ALPHA / "_")
                         *(ALPHA / DIGIT / "_" / "-" / ".")

The yang-identifier defined in draft-ietf-netmod-rfc6991-bis-09.txt
did not change and it still references RFC 6020. What exactly is your
concern or proposal? Do you propose to change the definition of
yang-identifier to reference RFC 7950 and to drop the second pattern
(which would be an allowed change according to YANG update rules)? Or
do you propose to add a new definition of a yang-identifier-1.1?

<tp2>
As I recall, the issue came up in YANG 1.1 that XML identifiers were  no longer restricted in their first three letters by their SDO and so the limitation in YANG 1.0 was removed from YANG 1.1 (as you quote above).  It then looks anomalous to me to retain that restriction in the type definition when it has been removed from the language definition.  My preference would be to remove that restriction from the definition of 'identifier' in 6991-bis but I am not sure whether or not there are consequences I am missing, hence my diffidence.

As you say, RFC6020 s.10 would appear to allow it.

Tom Petch

/js

--
Jürgen Schönwälder              Jacobs University Bremen gGmbH
Phone: [phone]         Campus Ring 1 | 28759 Bremen | Germany
Fax:   [phone]         <https://www.jacobs-university.de/>