Re: [netmod] syslog-model-17 shepherd writeup issues -references

[Kent Watsen <[e-mail]>]

Good suggestion, Rob.  This might also have come up in the SecDir review.

Separately, Juergen had a suggestion to add an "observation" (since a "warning" seems too strong) that use of POSIX regex precludes UTF8 support.  Makes sense?  FWIW, the module is fine as is, since the regex is under a feature statement, which allows a future effort add another regex-feature if ever needed.

Clyde, since these are not technical changes, and assuming that there is no objection, you can make these two tweaks as well.  Otherwise, I can add notes about these in the shepherd writeup.

Kent

--

Hi Kent, Clyde,

Does the "pattern-match" leaf need to be explicitly pulled out in 
security considerations?  Allowing a client to provide an arbitrary 
regex could potentially cause a regex engine to overflow its stack and 
crash.

An example of an regex overflow is described here: 
http://www.regular-expressions.info/catastrophic.html

Thanks,
Rob


On 13/09/2017 18:08, Kent Watsen wrote:
> Hi Tom,
>
> Thanks.  The fix I'm looking for is for the 'pattern-match' leaf
> to have a 'reference' statement to Std-1003.1-2008, and for S4.1
> to also list Std-1003.1-2008 as a draft-level reference.
>
> I was going to point out the typo "the the" as well, but figured
> that the RFC Editor would get it.
>
> K. // shepherd
>
>
> --
>
> Kent
>
> You flag Std-1003.1-2008 as listed as a normative reference but not used
> anywhere in the document.  In the Descriptions, but not in the s.4.1
> references, I see
>
> This leaf describes a Posix 1003.2 regular expression ...
>
> twice, which may, or may not, relate to this issue.
>
> Back in July, clyde said
> "I will insert a normative reference to POSIX 1003.2 in the next
> revision of the draft."
>
> In a similar vein, RFC6991 appears in a reference statement but nowhere
> else.
>
> As you point out, RFC6021 is referenced but is obsoleted by RFC6991 so
> should not be.
>
> And in a slightly different vein,
>
>     registry [RFC7895]/>.  Following the format in [RFC7950]/>, the the
>
> looks odd for plain text and for the repetition of 'the'..
>
> Tom Petch
>
> ----- Original Message -----
> From: "Kent Watsen" <[e-mail]>
> To: <[e-mail]>
> Cc: <[e-mail]>
> Sent: Tuesday, September 12, 2017 10:50 PM
> Subject: [netmod] syslog-model-17 shepherd writeup issues
>
>
>> Clyde, all,
>>
>> In reviewing the draft for Shepherd writeup, I found the following
> issues that I think need to be addressed before the document can be sent
> to Benoit for AD review:
>>
>> 1. Idnits found the following:
>>
>>    Summary: 3 errors (**), 0 flaws (~~), 3 warnings (==), 1 comment
> (--).
>>      ** There are 2 instances of too long lines in the document, the
> longest one
>>           being 3 characters in excess of 72.
>>
>>      ** Obsolete normative reference: RFC 6021 (Obsoleted by RFC 6991)
>>
>>      ** Downref: Normative reference to an Historic RFC: RFC 6587
>>
>>      == Missing Reference: 'RFC5425' is mentioned on line 359, but not
> defined
>>           '[RFC5425], [RFC5426], [RFC6587], and [RFC5848]....'
>>
>>       == Unused Reference: 'RFC7895' is defined on line 1406, but no
> explicit
>>            reference was found in the text
>>            '[RFC7895]  Bierman, A., Bjorklund, M., and K. Watsen, "YANG
> Module L...'
>>       == Unused Reference: 'RFC6242' is defined on line 1435, but no
> explicit
>>            reference was found in the text
>>            '[RFC6242]  Wasserman, M., "Using the NETCONF Protocol over
> Secure Sh...'
>>
>> 2. `rfcstrip` extracted "ietf-syslog.yang",  which is missing
> "@yyyy-mm-dd" in its name
>> 3.  neither `pyang` nor `yanglint` found any errors with
> ietf-syslog.yang.    pyang says
>>        for vendor-syslog-types-example: statement "identity" must have
> a "description"
>>        substatement.
>>
>> 4. testing the examples in the draft against yanglint:
>>        - for both examples: Missing element's "namespace". (/config)
>>        - just removing the "<config>" element envelop resolves this
> error.
>> 5. the 2nd example uses IP address "2001:db8:a0b:12f0::1", but this
> SHOULD be a
>>       domain name (e.g., foo.example.com)
>>
>> 6. in the YANG module, anywhere you have an RFC listed in a
> 'description' statement,
>>       there should be a 'reference' statement for that RFC.
>>
>> 7. in the tree diagram, the leafrefs no longer indicate what they
> point at, they now all
>>       just say "leafref".  Did you do this on purpose, or are you using
> a different tree
>>       output generator from -15?
>>
>> 8. RFC6536 is listed as a normative reference, but it probably should
> be informative.
>> 9. Std-1003.1-2008 is listed as a normative reference, but it is not
> used anywhere in the document.
>> 10. RFC6242 is listed as an informative reference, but it is not used
> anywhere in the document.
>> 11. the document fails to declare its normative references to
> ietf-keystore and ietf-tls-client-server.
>>          Note: you manually entered the "[RFC yyyy], and [RFC xxxx]"
> references…
>> 12.  The IANA considerations section seems asymmetric.  Either put
> both registry insertions into
>>          subsections, or keep them both at the top-level…
>>
>> 13. reviewing the final document against my original YD review, I have
> the following responses.  Let's be sure to close out these items as
> well.  Ref: https://mailarchive.ietf.org/arch/msg/netmod/10lo41Ud4A3ZN11
> s-0gOfCe8NSE
>> 1. ok
>> 2. better
>> 3. should be: s/the message/these messages/  [RFC Editor might've
> caught this]
>> 4. better
>> 5. still feel the same way, but no biggee
>> 6. better, but from 8174, you should add the part "when, and only
> when, they appear in all capitals, as shown here."
>> 7. fixed
>> 8. fixed
>> 9. you did what I asked, but the result still isn't satisfying...
>> 10. some improvements made in this area, but my ask wasn't among them
>> 11. better
>> 12. better, but I think the 4th line should be indented too, right?
>> 13. better, but I wish you called S1.3 "Tree Diagram Notation"
>> 14. fixed
>> 15. fixed
>> 16. fixed
>> 17. fine
>> 18. still a weird line brake here.  try putting the quoted string on
> the next line.
>> 19. fixed
>> 20. fixed
>> 21. not fixed (re: yang-security-guidelines)
>> 22. fine
>>
>>
>> PS: please also be sure to follow-up with Benoit on his AD review.
>>
>> Thanks,
>> Kent  // shepherd & yang doctor
>>
>>
>>
>> _______________________________________________
>> netmod mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/netmod
>>
>
>
> _______________________________________________
> netmod mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/netmod