Re: [core] Spencer Dawkins' Discuss on draft-ietf-core-http-mapping-14: (with DISCUSS and COMMENT)

[Jaime Jiménez <[e-mail]>]

Hi all,

Just to catch up on this. From the mails, to my understanding most of the comments have currently been addressed. As we are approaching the 31st Oct maybe we could do a quick recap of outstanding items. 

Much of Stephen’s comments have already been addressed by addressing Kathleen’s. If Stephen feels comfortable with the changes he can change the ballot or comment on the new version. The remaining ballot is Ben’s, that would require some changes to the phrasing to make sure the document remains informational in format.

Ciao,
- - Jaime Jiménez

> On 03 Oct 2016, at 21:15, Fossati, Thomas (Nokia - GB) <[e-mail]> wrote:
> 
> Hi,
> 
> We have just posted -15 which should clear most of the comments and discuss points.
> 
> In particular, we addressed:
> - Kathleen's DISCUSS and COMMENTs;
> - Stephen's COMMENTs;
> - Suresh's DISCUSS; and 
> - Spencer's DISCUSS and COMMENTs.
> 
> The diff is at https://www.ietf.org/rfcdiff?url1=draft-ietf-core-http-mapping-14&url2=draft-ietf-core-http-mapping-15 <https://www.ietf.org/rfcdiff?url1=draft-ietf-core-http-mapping-14&url2=draft-ietf-core-http-mapping-15> — please check if I forgot anything.
> 
> This should leave Stephen's DISCUSS as the only remaining item, which we plan to address in the next few days.
> 
> Thanks very much to everyone involved in the discussion so far.  Much appreciated.
> 
> Cheers, t.
> 
> On 03/10/2016 07:19, "Alexey Melnikov" <[e-mail] <mailto:[e-mail]>> wrote:
>> Hi Thomas,
>> 
>>> On 2 Oct 2016, at 14:23, Fossati, Thomas (Nokia - GB) <[e-mail] <mailto:[e-mail]>> wrote:
>>> Hi Alexey,
>>>> On 30/09/2016 18:49, "Alexey Melnikov" <[e-mail] <mailto:[e-mail]>> wrote:
>>>> On 30 Sep 2016, at 02:51, Fossati, Thomas (Nokia - GB)
>>>> <[e-mail] <mailto:[e-mail]>> wrote:
>>>>>> Can the HTTP client tell whether the result used coaps://?
>>>>> No, if it's using the simple form and there is no OOB agreement.  So, I
>>>>> guess the question is: do you think it'd be better to spell this out
>>>>> explicitly?  If so, we could say that a Target CoAP URI without the
>>>>> scheme
>>>>> is not allowed unless there is an OOB agreement between client and HC
>>>>> proxy (client MUST NOT emit / proxy MUST return 4xx on receipt).  Or do
>>>>> you have something different in mind?
>>>> An alternative is to always require coaps in this case. Might not work
>>>> for your type of document though.
>>> Yep.  But, in this case, I think it's better to fail early and hard (with
>>> a 4xx on the HTTP side of the proxy), because we can safely assume that
>>> the Target URI is syntactically incorrect.  This way we also avoid
>>> injecting a (possibly useless) request in to the constrained network --
>>> and returning a slightly confusing 504 in case the endpoint doesn't reply
>>> back if it's not listening for coaps requests.
>> 
>> Fair enough.