Re: [YANG] new pyang errors

[Andy Bierman <[e-mail]>]

Phil Shafer wrote:
> Andy Bierman writes:
>> Mandatory means it must be present in a
>> valid config, not that the manager must set some value.
> 
> That's certainly not the way I read mandatory.  Mandatory
> means that it must be set, meaning that some{one,thing}
> must set it.  If there were a reasonable default, you
> shouldn't need to make it mandatory.
> 


I can only find 2 brief paragraphs of text on the mandatory-stmt:

7.6.4.  The leaf's mandatory Statement

    The "mandatory" statement, which is optional, takes as an argument
    the string "true" or "false".  If "mandatory" is "true", the node
    must exist in a valid configuration if its parent node exists.

    If not specified, the default is "false".


7.9.4.  The choice's mandatory Statement

    The "mandatory" statement, which is optional, takes as an argument
    the string "true" or "false".  If "mandatory" is "true", nodes from
    exactly one of the choice's case branches MUST exist in a valid
    configuration.

    If not specified, the default is "false".


The current draft clearly says that 'mandatory true' refers to the
configuration database, not to whether the manager must provide
an explicit value.

I agree with the interpretation in the current draft.

YANG needs to be more consistent wrt/ mixing DM concepts too much.
The data model object definitions (leaf, container, etc.) refer
to the desired/expected state of the configuration database,
not with individual PDUs.

The reason that the data model MUST refer to the configuration
is that schema validation of NETCONF PDUs is mostly meaningless.
Since the PDUs are so flexible, every single element and attribute
would need to be marked optional (minOccurs=0).



> Thanks,
>  Phil
> 

Andy


_______________________________________________
YANG mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/yang