Re: [netmod] yang canonical integer format

[Robert Varga <[e-mail]>]

Hello,

On 03/03/2017 05:52 PM, Kent Watsen wrote:
>>  typedef ether-type {
>>    type uint16 { format "%x"; }
>>  }
>>
>> in plain YANG to make the hexadecimal format the canonical format. Who
>> maintains ideas for the next YANG revision?

This is also useful for other types, for example inet-types:mac-address,
where implementations could use an 'canonical-pattern' type of declaration.

I think this particular idea ties the transport-specific representation
to a type definition -- I mean requiring binary transports to send
uint16s as text is not efficient and they would end up not using the
statement at all. This is also sticky if the transport places additional
restrictions on what is and is not a valid string and reminds me of some
of the 'anyxml' problems...

>>
>> /js
> 
> Ideas for the next revision of YANG are being captured in this tracker:
> 
>   https://github.com/netmod-wg/yang-next/issues
> 

given YANG is an extensible language, would it make sense to deliver
these outside of YANG revisions?

The main reason for doing so is that these, unlike YANG revisions:
- do not change the meta model of the language
- are not mandatory-to-implement

It will also allow for faster delivery to end users.

Regards,
Robert