Hi, William Lupton <wlup...@broadband-forum.org> wrote: > All, > > Here are a couple of questions from the Broadband Forum on RFC > 6020bis-08 and draft-ietf-netconf-yang-library-02. > > Thanks, > William > > -------- > > 1. RFC 6020bis and updating modules > > 6020bis Section 11 (Updating a Module) lists all the ways in which a > definition can be revised, but it's not clear whether this is intended > to be an exhaustive list or just to illustrate the principle of > allowing the value space to be broadened. If the latter then > presumably other changes that adhere to this principle would also be > valid.
The former. > For example, it seems that the following might also be permitted: > > a. Extending a "when" statement so it is true for a wider set of > conditions (example: realising that an RFC 7223 interface object > applies to additional interface types). This is allowed by: o A "when" statement may be removed or its constraint relaxed. > b. Re-writing a "when" statement to use a base identity rather than an > explicit list of identities (with no change to the conditions for > which it is true). Same as above. > c. Converting a leaf node to a choice (with no change to default > behaviour). This is not allowed, but maybe it should. I.e., it should be ok to wrap a node that is not a mandatory node (see terminology) in a choice (+ case). > 2. ietf-yang-library and reporting submodules > > As we understand them, submodules have no impact on a module's > external interface, and (6020 Section 11) "A module may be split into > a set of submodules, or a submodule may be removed, provided the > definitions in the module do not change in any other way than allowed > here". > > draft-ietf-netconf-yang-library-02 says (Section 1) "submodule list: > The name and revision of each submodule used by the module MUST be > identified" but the YANG module's "submodules" container appears not > to be mandatory. A container cannot be mandatory. > Is it intended that the submodules MUST be listed? Yes, that's what the text "The name and revision of each submodule used by the module MUST be identified" says. > If so, what is the rationale for requiring this? The module may have used include w/o revision. Unless the submodule is listed by the server, the client won't know which revision the server uses. /martin _______________________________________________ netmod mailing list netmod@ietf.org https://www.ietf.org/mailman/listinfo/netmod
