Re: [netmod] hi Martin, one issue about RFC7950 7.20.3.2. The "deviate" Statement, please help to confirm

[Martin Bjorklund <[e-mail]>]

Hi,

This is an issue that has been reported before.  It is tracked in
https://github.com/netmod-wg/yang-next/issues/14.

This said, I think in this particular case, where you define both the
when expression and the if-feature leaf, I think the correct solution
is to ensure that the when expression can handle the case that the
leaf is not implemented (due to the if-feature).  In this case, the
when expression you have will be "true" if the node ../vrfAny is not
implemented, which is what you want (I assume, since otherwise you
should mark the node with the when expression with the same if-feature
as ../vrfAny).


/martin


"Zhengguangying (Walker)" <[e-mail]> wrote:
> Hi Martin,
> 
> When we define YANG model's deviation file, we find one scenario was
> nor supported but we have one problem depend on it.
> 
> Our scenario:
> In ACl YANG module, there have two leafs:
> acl/aclGroups/aclGroup/aclRuleBas4s/ aclRuleBas4/vrfName
> acl/aclGroups/aclGroup/aclRuleBas4s/ aclRuleBas4/vrfAny
> 
> And the acl/aclGroups/aclGroup/aclRuleBas4s/ aclRuleBas4/vrfName have
> one Constraints: when "not (../vrfAny = 'true')"
> 
> 
> In some product domain, the "vrfAny" does not supported, some they
> deviate the leaf "vrfAny". From the sematic view, the Constraints of
> vrfName when "not (../vrfAny = 'true')" should be deviated as "delete"
> too.
> 
> But, in RFC7950 7.20.3.2.  The "deviate" Statement , "when " is not
> the Substatement of deviate, we can not deviate the when statement,
> and the problem coming, some industry netconf tools compile fail
> because "../vrfAny" does not exist.
> 
> And there have some other scenario about "when" deviate, So, I think
> whether the YANG language should add "when" as the Substatement of
> deviate?
> 
> What's your opinion?
> 
> 
> Thanks & regards
> 
> Walker(guangying zheng)
> 
> 
> [cid:image001.png@01D38B9C.D7570600]
> 
>