Re: [YANG] new pyang errors

Andy Bierman wrote:
> Phil Shafer wrote:

 >...

Perhaps you want to revive the debate on the NGO list
about the 'with-defaults' extension?

The consensus was that both views of the configuration database
are useful, and the protocol should support both of them.

Without the 'full' view, how is a manager supposed to know
whether a leaf is really there with the default value, or
not supported by the agent or supported but not used in this instance.

How is a manager supposed to get a complete config diff from
the agent, without analyzing all the schema files between N different
software releases?  A diff (with-defaults='true') will expose silent
default value changes for agent-supplied optional knobs, or agents
that are ignoring the schema and using a different default value.

The solution to this supposed YANG problem is to add the with-defaults='false'
extension to the NETCONF operations to filter out these nodes in
get and copy operations.  Currently, 'state' nodes are auto-filtered
by get-config, but not "default true/false".  It should not be
a big deal to add this as a mandatory filter somehow.

>> Thanks,
>>  Phil
>>
>>
> 
> Andy
> 
> 

Andy

_______________________________________________
YANG mailing list
YANG@ietf.org
https://www1.ietf.org/mailman/listinfo/yang