Re: [MMUSIC] [secdir] Review of draft-ietf-mmusic-sctp-sdp-22

[Paul Kyzivat <[e-mail]>]

On 2/10/17 9:15 AM, Christer Holmberg wrote:

>> Section 5.2:
>>
>> 	Leading zeroes MUST NOT be used.
>>
>> What is the problem with leading zeros? What can go wrong when these are
>> used? Should
>> an implementor be warned about something specific?
>
> I don¹t know. The same statement exists for a number of SDP attributes,
> but I didn¹t find any justification.

This is standard practice in RFC4566. It has an ABNF rule:

    integer =             POS-DIGIT *DIGIT

that makes it invalid to insert leading zeros. I don't *know*, but 
speculate, that this might be tied to the C language convention where a 
leading zero means octal, and not wanting to introduce that confusion.

But the sctp port value isn't defined using the 'integer' rule. It has 
its own definition:

             sctp-port-value = 1*5<DIGIT defined in RFC4566>

To be consistent, perhaps that should be changed to:

             sctp-port-value = <POS-DIGIT defined in RFC4566>
                               *4<DIGIT defined in RFC4566>

	Thanks,
	Paul