Re: [netmod] Y07: do not allow when or if-feature on key

[Andy Bierman <[e-mail]>]

On Fri, Jan 2, 2015 at 8:39 AM, Martin Bjorklund <[e-mail]> wrote:
> Andy Bierman <[e-mail]> wrote:
>> On Fri, Jan 2, 2015 at 6:52 AM, Ladislav Lhotka <[e-mail]> wrote:
>> >
>> >> On 02 Jan 2015, at 15:49, Andy Bierman <[e-mail]> wrote:
>> >>
>> >> Hi,
>> >>
>> >> This would make valid YANG 1.0 modules be rejected.
>> >> I strongly object to this change.
>> >
>> > No 1.0 module will be rejected if the statements are just ignored. It does mean a change in module semantics but then I guess this qualifies as a bug fix.
>> >
>>
>>
>> IMO it is a really bad idea to allow tools to ignore any statements
>> in the base language.  This is only OK for external statements
>> defined with extensions.
>
> We already do this in a couple of instances; e.g. config statement in
> rpc/notification, and default in a key leaf.
>

And now we will change the rules in YANG 1.1 so default-stmt
is not ignored in a leaf?  How is that backward-compatible?

>> It is really confusing to the writer and reader as well.
>>
>> I agree that it is not feasible for a tool to compare XPath statements
>> with different context nodes to see if they always produce the
>> same results.  But it is easy to check if the feature-stmts are the same
>> or less than the parent.
>>
>> When-stmt can be inherited from uses and augment.
>> Not sure if keys can be augmented into a list, but uses can
>> certainly pull in the keys.
>
> [...]
>
>> Note that a when-stmt could appear in the uses-stmt or augment-stmts
>> making the key leaf conditional.  Also, if-feature-stmts could be present
>> as well.
>>
>> I might be in favor of making key leafs that are conditional due to when-stmt
>> invalid and rejected with a compiler error. It's not OK to ignore this
>> condition.
>
> Consider this situation:
>
>   grouping x {
>     leaf a {
>       if-feature "bar";
>       type string;
>     }
>   }
>
>   list y {
>     key "a b";
>     uses x;
>     leaf b {
>       type string;
>     }
>   }
>
> Now, if the server doesn't advertise "bar", what does this mean?
>


As I said before, it is an error for the leaf to have if-feature-stmts
that are not also in the parent node.  Our compiler would flag leaf "a"
as an error.

If an "if-feature bar;" stmt appeared in the parent list y,
then this is not an error at all.


> YANG 1.0 doesn't say.
>
> We can make it illegal in 1.1, or we can say that the if-feature/when
> statements are ignored.
>
> If we make it illegal, it means that there is no way list y can use
> the grouping x.
>

sure it can -- add "if-feature bar;"


> I think Lada's suggestion was to ignore if-feature/when b/c of
> backwards compatibility reasons.
>
> This would be legal in 1.1, if we say that we ignore the if-feature:
>
>   list y {
>     if-feature "bar";
>     key "a b";
>     uses x;
>     leaf b {
>       type string;
>     }
>   }
>
>

This is legal now -- our tools support it and check for key leafs
that are more conditional than their parent. The statements are
validated


> /martin


Andy