Re: [sip-clf] draft-ietf-sipclf-format-00 comments

Anders - 

Inline..
On Mar 11, 2011, at 6:54 PM, Anders Nygren wrote:

> Hi
> In chapter 4.1, the text
>  "The <IndexPointers> portion of the SIP CLF record is a 64-byte header
>   that indicates meta-data about the record."
> is repeated before and after the figure 3.

[GS] Fixed.
> 
> chapter 4.3 figure 5 shows a 0x0A after the TLVs. But I think that
> 0x0A is not part of the
> optional parameters portion but is the termination of the record, and
> should be included
> even if there are no optional parameters in the record.

[GS] In principle I agree with you. For the sake of graphical continuity, I'll need to get creative so that I don't need to represent the logical structure of the SIP CLF record as:

        <IndexPointers>
        <MandatoryFields>
        <OptionalFields>
	0x0A

I'll kick it around in my head as to how best represent that.  Great catch.

Regards,

Gonzalo

> 
> /Anders
> 
> 
> On Fri, Mar 11, 2011 at 5:07 PM, Gonzalo Salgueiro <gsalguei@cisco.com> wrote:
>> Thanks Anders. This reflected an earlier intent for microsecond accuracy
>> that has since been rescinded. I have amended the text.
>> Regards,
>> Gonzalo
>> On Mar 11, 2011, at 5:43 PM, Anders Nygren wrote:
>> 
>> One more thing,
>> page 9,
>> The text says that fractional seconds is 6 bytes but figure 4 shows 3 bytes.
>> 
>> /Anders
>> 
>> On Fri, Mar 11, 2011 at 1:38 PM, Gonzalo Salgueiro <gsalguei@cisco.com>
>> wrote:
>> 
>> Anders -
>> 
>> Thanks for providing your feedback, it is much appreciated as we work
>> through developing this format.
>> 
>> I'm very glad to hear you are implementing this already and would be willing
>> to follow up with you offline to get some data points on how you decided to
>> do this, how difficult it was, and how useful you found it, what you would
>> change, etc.
>> 
>> Comments inline...
>> 
>> On Mar 11, 2011, at 1:01 PM, Anders Nygren wrote:
>> 
>> Hi
>> 
>> I am not sure what is the correct way for commenting on the SIP-CLF spec,
>> 
>> so I will do it here until someone tells me that there is a better
>> 
>> place to do it.
>> 
>> [GS] This is indeed the place.
>> 
>> I am trying to implement a log reader and writer according to
>> 
>> draft-ietf-sipclf-format-00 and have found a few small errors.
>> 
>> page 8:
>> 
>>   "Record Length (6 bytes):  Hexadecimal encoded total length of this
>> 
>>    log record, including "Flags" and "Record Length" fields, and
>> 
>>    terminating line-feed."
>> 
>> Should that include the version field as well?
>> 
>> [GS] Yes this does include the Version field. It is the length of the ENTIRE
>> record including all of <IndexPointers>, <MandatoryFields>, <OptionalFields>
>> 
>> The list of items in the text you have selected is not meant to be  complete
>> listing of all fields, rather a representative list indicating that
>> everything is included. If you think there is value in specifically
>> mentioning the Version field, then I will modify the text as follows:
>> 
>> Hexadecimal encoded total length of this log record, including "Version",
>> "Record Length", "Flags" fields, and terminating line-feed.
>> 
>> 
>> 
>> page 12:
>> 
>>   "Length Field (2 bytes):  Indicates the length of the value coded in
>> 
>>    this TLV, hexadecimal encoded.  This length does NOT include the
>> 
>>    TLV header."
>> 
>> The length field is shown as 4 bytes in figure 5.
>> 
>> Thanks. Great catch.  This is indeed wrong. The figure is correct. I will
>> change the text to 4 bytes.
>> 
>> Warm Regards,
>> 
>> Gonzalo
>> 
>> /Anders Nygren
>> 
>> _______________________________________________
>> 
>> sip-clf mailing list
>> 
>> sip-clf@ietf.org
>> 
>> https://www.ietf.org/mailman/listinfo/sip-clf
>> 
>> 
>> 
>> 