Deviate should allow to add extension sub-statements

[paulo-g]

Fix for #214

[coveralls]

Coverage remained the same at 64.601% when pulling 9211a6c on paulo-g:master into 38ee1c3 on mbj4668:master.

[mbj4668]

Hi,
There is already support for handling this on a per-extension basis. The idea is that if you have a module with extensions (e.g. ietf-yang-smiv2), you can write a plugin that registers the grammar for the extensions (e.g. smi.py). This plugin can also register keywords that can be deviated, by using the function add_deviation_element() in statements.py.

[paulo-g]

@mbj4668 , understood, but that will limit the possibility to share the model without a companion plugin.

My interpretation from the RFC text was that, by omission, an extension would support to be deviated, and only when providing a dedicated plugin could this change. Which is the opposite interpretation that was implemented.

In any case I agree with the decision.
Thanks!

[mbj4668]

Incorrect reference, should have been 126.

[coveralls]

Coverage decreased (-33.8%) to 30.812% when pulling 9211a6c on paulo-g:master into 38ee1c3 on mbj4668:master.

[coveralls]

Coverage decreased (-33.8%) to 30.752% when pulling 9211a6c on paulo-g:master into 38ee1c3 on mbj4668:master.