Re: [netmod] AD review of draft-ietf-netmod-system-mgmt

[Randy Presuhn <[e-mail]>]

Hi -

>From: Ladislav Lhotka <[e-mail]>
>Sent: Dec 13, 2013 11:58 AM
>To: Randy Presuhn <[e-mail]>
>Cc: Martin Björklund <[e-mail]>, [e-mail]
>Subject: Re: [netmod] AD review of draft-ietf-netmod-system-mgmt
>
>
>On 13 Dec 2013, at 20:34, Randy Presuhn <[e-mail]> wrote:
>
>> Hi -
>> 
>>> From: Martin Bjorklund <[e-mail]>
>>> Sent: Dec 12, 2013 11:22 PM
>>> To: [e-mail]
>>> Cc: [e-mail]
>>> Subject: Re: [netmod] AD review of draft-ietf-netmod-system-mgmt
>>> 
>>> Randy Presuhn <[e-mail]> wrote:
>>>> Hi -
>>>> 
>>>>> From: Juergen Schoenwaelder <[e-mail]>
>>>>> Sent: Dec 12, 2013 3:52 AM
>>>>> To: Randy Presuhn <[e-mail]>, [e-mail]
>>>>> Cc: [e-mail]
>>>>> Subject: Re: [netmod] AD review of draft-ietf-netmod-system-mgmt
>>>>> 
>>>>> On Thu, Dec 12, 2013 at 12:42:57AM -0800, Randy Presuhn wrote:
>>>>>> 
>>>>>>>   o  An implementation MUST allow any legal "string" (YANG string).
>>>>>> 
>>>>>> There are good reasons to restrict formatting and control characters -
>>>>>> I'll assume YANG strings do this already.  If not, that's another long
>>>>>> discussion.
>>>>> 
>>>>> RFC 6020 says:
>>>>> 
>>>>>  The string built-in type represents human-readable strings in YANG.
>>>>>  Legal characters are tab, carriage return, line feed, and the legal
>>>>>  characters of Unicode and ISO/IEC 10646 [ISO.10646]:
>>>>> 
>>>>>    ;; any Unicode character, excluding the surrogate blocks,
>>>>>    ;; FFFE, and FFFF.
>>>>>    string = *char
>>>>>    char = %x9 / %xA / %xD / %x20-D7FF / %xE000-FFFD /
>>>>>           %x10000-10FFFF
>>>>> 
>>>>> And so far, we have used these strings without further restrictions in
>>>>> data models when there was something to be named.
>>>> 
>>>> Very odd.  It restricts the C0 controls, but permits the C1
>>>> control codes?  I hope someone thought that through carefully.
>>> 
>>> It is what XML Schema 1.0 says is a string.  See
>>> http://www.w3.org/TR/xmlschema-2/#string and the reference to
>>> http://www.w3.org/TR/2000/WD-xml-2e-20000814#NT-Char
>> 
>> Let me be more direct.  To exclude, for example,
>> VT (vertical tab) but at the same time to include
>> RI (reverse line feed) in the set of permitted
>> characters seems ill-advised, to say the least.
>
>That was the (somewhat arbitrary) decision taken for XML 1.0.
>And since it is the serialization format for NETCONF, it was
>quite logical to adopt it for YANG, too.

This has nothing to do with serialization.  This is
a question of which code points are permitted within
strings intended for human consumption.

>> I hope the rationale for including the C1 controls
>> (%x80-9F) while quite correctly severely limiting
>> the C0 controls is documented somewhere.
>
>Note that XML 1.1 allows #x1..#x1f, represented as
>character references.
...

That's beside the point.  Permitting a reverse line feed
(or any of the other C1 controls) in sysContact, while
prohibiting a vertical tab, makes no sense at all.  See
http://en.wikipedia.org/wiki/C0_and_C1_control_codes#C1_set
for an overview of this rogues' gallery of characters,
and consider whether it's really a good idea to permit
them for use in labeling interfaces, etc.


Note that we took a somewhat different course in defining
SnmpAdminString in RFC 3411, and this results in a technical
incompatibility.  RFC 3411 addresses the issue of C0 and C1
with a mere recommendation: "The use of control codes should
be avoided."  What this means is that it is perfectly legal,
though operationally not recommended, for a MIB object with
syntax SnmpAdminString (not further constrained by the object-
type's description) to contain code points *prohibited*
by the RFC 6020 string type.

Randy