Re: [Cbor] draft-ietf-cbor-cddl-control-00 should add CDDL notation for CBOR Sequences

Hi Henk,

The CDDL

my-embedded-cbor-seq = bytes .cborseq my-array 

mathches a CBOR sequence wrapped in a byte string. I.e. "bytes .cborseq [ 1, 2 ]" would match the CBOR encoding 0x420102

All the document I cited use unadorned CBOR Sequences (i.e. not wrapped in a byte string). As far as I know there is no way to write CDDL that matches the unadorned CBOR Sequence 0x0102 ( 1, 2 ). I think this needs to be fixed.

John

-----Original Message-----
From: Henk Birkholz <henk.birkholz@sit.fraunhofer.de>
Date: Thursday, 5 November 2020 at 16:02
To: John Mattsson <john.mattsson@ericsson.com>, "cbor@ietf.org" <cbor@ietf.org>
Subject: Re: [Cbor] draft-ietf-cbor-cddl-control-00 should add CDDL notation for CBOR Sequences

Hi John,

I might be totally missing something obvious here - it's just that I am 
having a hard time visualizing your problem at the moment.

Maybe asking the obvious can help (me getting a grasp of the issue). Why 
does the following recommendation from RFC8764 not help you?

>    my-embedded-cbor-seq = bytes .cborseq my-array
>    my-array = [* my-element]
>    my-element = my-foo / my-bar

Viele Grüße,

Henk

On 05.11.20 15:54, John Mattsson wrote:
> Hi Henk,
> 
> To quote RFC 8764
> 
> 1) "CBOR Sequences are already supported as contents of byte strings using the ".cborseq" control operator"
> 2) "CDDL does not provide for unadorned CBOR Sequences as a top-level subject of a specification"
>   
> All the document I cited do 2) not 1) .cborseq as currently specified is not a solution.
> 
> Cheers,
> John
> 
> -----Original Message-----
> From: Henk Birkholz <henk.birkholz@sit.fraunhofer.de>
> Date: Thursday, 5 November 2020 at 15:29
> To: John Mattsson <john.mattsson@ericsson.com>, "cbor@ietf.org" <cbor@ietf.org>
> Subject: Re: [Cbor] draft-ietf-cbor-cddl-control-00 should add CDDL notation for CBOR Sequences
> 
> Hi John,
> 
> as the control for cborseq is introduced in:
> 
>> https://tools.ietf.org/html/rfc8610#section-3.8.4
> 
> and RFC8742 states that:
> 
>>     Currently, CDDL does not provide for unadorned CBOR Sequences as a
>>     top-level subject of a specification.  For now, the suggestion is to
>>     use an array for the top-level rule, as is used for the ".cborseq"
>>     control operator, and add English text that explains that the
>>     specification is really about a CBOR Sequence with the elements of
>>     the array
> 
> it seems to me that you are asking for a specific CDDL notation that can
> represent a cborseq as a top-level subject. Why is using an array here
> not good enough in your case?
> 
> Viele Grüße,
> 
> Henk
> 
> On 05.11.20 13:49, John Mattsson wrote:
>> Hi,
>>
>> I the most important missing piece in RFC 8610 is the lack of CDDL for CBOR Sequences (RFC 8742) and I think draft-ietf-cbor-cddl-control would be a good place to add CDDL for that.
>>
>> CBOR sequences has already been standardized in RFC 8742. CBOR sequences are used quite heavily in IETF documents such as RFC 8769, draft-ietf-6tisch-minimal-security, draft-ietf-lake-edhoc, draft-palombini-core-oscore-edhoc, draft-mattsson-cose-cbor-cert-compress, etc.
>>
>> I don't understand all the complexities of CDDL formalism, but this seems like a quite easy thing to solve. I do not care exactly which notation is used, but could we please just agree on something and put in a draft.
>>
>> The notation could be something like:
>>
>> - reuse the CDDL notation for group ( ... )
>> - reuse the Diagnostic Notation for CBOR sequences << ... >>
>> - Some modification of the CDDL array notation ] ... [
>> - Something looking like symbol swearing #$%@#$% ... #$%@#$%
>> ...
>>
>> Cheers,
>> John
>>
>>
>>
>>
>>
>>
>>
>> _______________________________________________
>> CBOR mailing list
>> CBOR@ietf.org
>> https://www.ietf.org/mailman/listinfo/cbor
>>
> 