Hi Sandy,
firstly, sorry for the chaotic review. I have organized it properly but some formatting was lost, I will know next time.
As for your question, I do not see any node "algorithm-type" of type leaf, only choice. This choice includes 2 empty cases "spf" and "all-path". This YANG construct as it is has no purpose and cannot even be instantiated in YANG data, unless some YANG data nodes were added into the cases by augments. I suppose that is not the intention and you simply wanted to write a configuration node that could be set to either "spf" or "all-path" value. In that case you want to use enumeration like this:
leaf algorithm-type {
type enumeration {
enum spf {
description "The algorithm is SPF.";
}
enum all-path {
description "The algorithm is all-path.";
}
}
description "The possible algorithm types.";
}
If you need the possible "algorithm-type" values to be dynamic - allowed new to be added from other YANG modules, you can even use type "identityref" and define the values as identities.
Regards,
Michal
On Thursday, September 09, 2021 11:09 CEST, <zhang.zheng@xxxxxxxxxx> wrote:
> Hi Michal,
> Thank you again for your review!
> We are do the updating work.
> But we are not sure that we got the meaning of the following comment:
> "algorithm-type - empty cases - redundant on their own, are expected to
> be augmented?"
> The algorithm-type leaf in the model is not empty. Could you please explain this comment more detailedly?
> Thank you very much!
> Best regards,
> Sandy
>
> ------------------原始邮件------------------
> 发件人：MichalVaškoviaDatatracker
> 收件人：yang-doctors@xxxxxxxx;
> 抄送人：draft-ietf-rift-yang.all@xxxxxxxx;last-call@xxxxxxxx;rift@xxxxxxxx;
> 日 期 ：2021年07月08日 18:17
> 主 题 ：[Rift] Yangdoctors last call review of draft-ietf-rift-yang-03
> Reviewer: Michal Vaško
> Review result: Almost Ready
>
> Generally, use references where make sense (features, nodes) and use units
> and/or standard types (ietf-yang-types) for leaves (such as grouping
> neighbor-node/bandwidth). All links are invalid, better to use references
> anyway because the module will be used outside the RFC.
>
> Specific problems:
>
> - description - copyright 2020
> - typedef ieee802-1as-timestamp-type - reference in description, put separately
> - grouping address-families
> - list with a single key can be leaf-list
> - would make sense if meant to be augmented with new nodes
> - grouping node-flag
> - used only once, makes sense if meant to be reused by other modules
> - consider using bits type in the leaf
> - grouping base-node-info/pod - redundant description, use union of number and
> "undefined", or leave out for undefined since it is not mandatory - augment
> rift/rx-lie-multicast-address,tx-lie-multicast-address
> - default value in description - should be defined in YANG
> - consider using
> refine on
> "addresses"
> rx-flood-port - redundant default in description, is obvious in YANG
> algorithm-type - empty cases - redundant on their own, are expected to
> be augmented? HAL - use lowercase
> database/tie/negative_disaggregation_prefixes
> - use hyphen instead of underscore
> - consider abbreviated/shorter node names
> - applicable for the following nodes as well
>
>
> _______________________________________________
> RIFT mailing list
> RIFT@xxxxxxxx
> https://www.ietf.org/mailman/listinfo/rift
--
last-call mailing list
last-call@xxxxxxxx
https://www.ietf.org/mailman/listinfo/last-call
