Re: [YANG] meaning of unique in PDUs

Ladislav Lhotka <lhotka@cesnet.cz> wrote:
> Phil Shafer píše v Čt 24. 01. 2008 v 22:11 -0500:
> > >In some cases this mapping is automatic, some cases are defined in the
> > >YANG draft and some are not clear yet (to me at least).
> > 
> > Please let me know where it isn't clear and we can fix the draft.
> > 
> Here is one: what's the interpretation (if any) of the unique statement
> for PDUs?

The draft says:

  The "unique" statement is used to put constraints on valid
  configurations. [...]

  In a valid configuration, the combined values of all the leaf
  instances specified in the string MUST be unique within all list
  entry instances.

Maybe we should be more clear what we mean with "valid configuration"?

> Example:
> 
> list aaa {
>   key "k";
>   unique "b c";
>   leaf k { ... }
>   leaf b { ... }
>   leaf c { ... }
> 
> and assume the existing configuration contains two list items:
> 
> <aaa>
>   <k>k1</k> <b>b1</b> <c>c1</c>
> </aaa>
> <aaa>
>   <k>k2</k> <b>b2</b> <c>c2</c>
> </aaa>  
> 
> 1. Is it possible to send the following fragment in an edit-config
> (merge mode)?
> 
> <aaa>
>   <k>k1</k> <b>b3</b>
> <aaa>
> <aaa>
>   <k>k2</k> <b>b3</b>
> <aaa>
> 
> That is, after merging the resulting configuration will still satisfy
> the statement 'unique "b c"' (values of c are different) but the PDU
> doeasn't.

Absolutely.  Also note that if you're modifying the candidate, it
is also ok to send:

 <aaa>
   <k>k1</k> <b>b3</b> <c>c3</c>
 <aaa>
 <aaa>
   <k>k2</k> <b>b3</b> <c>c3</c>
 <aaa>

When you <validate/> the candidate, it will fail.  It will also fail
to <commit/>.

> 2. Does "unique" and also "key" statements assume any space
> normalisation? For example, are the following keys different?
> 
> <k>k1</k>
> 
> <k>
>   k1
> </k>

This isn't really a "unique" or "key" question; it is a datatype
equestion.  In this case, the YANG built-in type "string" behaves as
"xs:string", so no whitespace normalization is done.

/martin