Re: [yang-doctors] canonical order question

[Scott Mansfield <[e-mail]>]

Thanks for the context, however I'm talking about body-stmts.

Can body-stmts can be in any order? I don't see anywhere where the ABNF says body-stmts can be in any order.  I see what you mean for import-stmts and linkage-stmts, but those aren't the statements I'm concerned about in my example.

I had the following..
module {
...
   import {
   }
   feature {
   }
   augment {
   }
   notification {
   }
}

That was changed by the --yang-canonical flag to: (notifications were moved above the augment)
module {
...
   import {
   }
   feature {
   }
   notification {
   }
   augment {
   }
}

augment-stmt and notification-stmt are body-stmts.  So I assumed the order in the ABNF in the body-stmts list was the canonical order.  But that is not what pyang --yang-canonical does.

Regards,
-scott.
   

-----Original Message-----
From: Martin Björklund <[e-mail]> 
Sent: Wednesday, November 22, 2023 5:29 AM
To: Scott Mansfield <[e-mail]>
Cc: [e-mail]
Subject: Re: [yang-doctors] canonical order question

Hi,

Scott Mansfield <scott.mansfield=[e-mail]> wrote:
> I looked through the mail archive and the pyang issues list on git, so 
> apologies if I missed this...
> 
> However,
> 
> Clause 14 of RFC 7950 says this...
>    body-stmts          = *(extension-stmt /
>                            feature-stmt /
>                            identity-stmt /
>                            typedef-stmt /
>                            grouping-stmt /
>                            data-def-stmt /
>                            augment-stmt /
>                            rpc-stmt /
>                            notification-stmt /
>                            deviation-stmt)
> 
> 
> But when I use pyang and the -yang-canonical switch, the output I get 
> has the notification statements before the augment statements.  Is 
> this ok?  Am I reading the ABNF incorrectly?

RFC 7950 says:

   The ABNF grammar
   [RFC5234] [RFC7405] defines the canonical order.

And then the grammar has for example:

   linkage-stmts       = ;; these stmts can appear in any order
                         *import-stmt
                         *include-stmt

So, the ABNF says that "import" comes before "include" which means that this is the canonical order.

But for body-stmts, the ABNF allows any order of the child statements.  Hence the canonical order also allows any order.


/martin




> 
> I have attached the before and after for information.
> ieee802-dot1q-rstp.yang is the original which has notifications after 
> the augment statement.  I then ran it though pyang with -f yang 
> -yang-canonical and put the result in ieee802-dot1q-rstp-canon.yang, 
> but now the notification statements were moved before the augment.
> Just trying to understand what the canonical order is.
> 
> Many thanks for the help and experience!
> Regards,
> -scott.
> 
> 
> 
> [http://mediabank.ericsson.net/internet-media/Email_logo_Ericsson.png]
> <http://www.ericsson.com/>
> 
> Scott Mansfield
> Principal Researcher Standardization
> 
> BNEW TS ST STA NAM
> Phone: [phone]
> Mobile: [phone]
> [e-mail]<mailto:[e-mail]>
> 
> Ericsson
> ericsson.com<http://www.ericsson.com/>
> 
> [http://mediabank.ericsson.net/internet-media/Email_Message.gif]<http:
> //www.ericsson.com/current_campaign>
> 
> Our commitment to Technology for
> Good<http://www.ericsson.com/thecompany/sustainability-corporaterespon
> sibility>
> and Diversity and
> Inclusion<http://www.ericsson.com/thecompany/diversity-inclusion>
> contributes to positive change.
> Follow us on: 
> Facebook<https://protect2.fireeye.com/v1/url?k=31323334-501d5122-31327
> 3af-454445555731-a4edbadbd57cb477&q=1&e=fcd2c336-306c-4500-97f4-b1c172
> 77193e&u=https%3A%2F%2Fwww.facebook.com%2Fericsson>
> LinkedIn<https://protect2.fireeye.com/v1/url?k=31323334-501d5122-31327
> 3af-454445555731-25461567501c53b8&q=1&e=fcd2c336-306c-4500-97f4-b1c172
> 77193e&u=https%3A%2F%2Fwww.linkedin.com%2Fcompany%2Fericsson>
> Twitter<https://twitter.com/Ericsson>
> 
> Legal entity:ERICSSON INC., registration number2047104, registered 
> office in A Delaware corporation.
> This communication is confidential. Our email terms:
> www.ericsson.com/en/legal/privacy/email-disclaimer<https://www.ericsso
> n.com/en/legal/privacy/email-disclaimer>
>