Re: [netconf] updates to client/server suite of drafts

[Martin Bjorklund <[e-mail]>]

Kent Watsen <[e-mail]> wrote:
> Hi Martin,
> 
> 
> >>   -  Added 'must' expressions asserting that the 'key-format' leaf
> >>      exists whenever a non-hidden key is specified.
> > 
> > This can be made simpler:
> 
> True.  I got caught up in the following the pattern used for symmetric/private key groupings.
> 
> > OLD:
> > 
> >       <snip/>
> > 
> > Now, since public-key is mandatory, the 'when' expression on
> > public-key-format will always be true (in a valid config).  Hence it
> > isn't needed.  And also, since public-key is mandatory the must on
> > public-key really just says that public-key-format is also mandatory:
> > 
> > NEW:
> > 
> >     <snip/>
> 
> Applied to my local copy.
> 
> 
> 
> 
> >>   -  Added a questionable forward reference to "encrypted-*" leafs in a
> >>      couple 'when' expressions.
> > 
> > Questionable indeed.  I suggest you remove the when expression
> > instead. You have must expressions that says thatt the key-format leaf
> > must exist in the relevant cases anyway.
> 
> Hmmm, but "private-key-format" should NOT be present (I think!) for a "hidden-private-key".  Maybe this?
> 
> OLD:
> 	when "../private-key or ../encrypted-private-key"
> 
> NEW:
> 	when "not(../hidden-private-key)"

Or perhaps use a must in the 'hidden-private-key' instead:

  must 'not(../private-key-format)';



/martin


> 
> 
> 
> BTW, the forward reference was "needed" due to the inability for the consuming module (ietf-keystone) to refine (?) the "when" expression when using the grouping.   Please add a YANG-next issue if you think this should be supported someday.
> 
> 
> Kent // contributor
> 
>