Re: [core] Responses to NON with 4.02

[Achim Kraus <[e-mail]>]

Hi Christian,

 > what were the complaints you got about the behavior about?

(Hope, I understand that ...)

3 years ago, a colleague was working on the the option validation and
found, that this seems to be wired in the specification.

Last year the Eclipse/Californium project got a similar request about that,

https://github.com/eclipse/californium/issues/1288

and so I decided to ask the list about the intention behind that.

AFAIK, there is no real issue with that. Basically the main failure is
sending the "bad option", not the way, the other peer responds. So the
behavior seems to be more important for developing and debugging then
for real production. And during developing it should be not too hard, to
switch to CON to see, which error response your get back.

best regards
Achim Kraus

Am 17.02.21 um 09:27 schrieb Christian Amsüss:
> Hi Achim,
>
>> https://mailarchive.ietf.org/arch/msg/core/wd4LleVnnvt2WOXj28ouNZcCTXU/
>
> thanks, that's indeed essentially the same question, I just failed to
> find it.
>
> I concur with Klaus that it's weird and belongs to the request/response
> layer.
>
> Just to ensure I didn't miss anything too relevant in my (so far, to be
> continued) "I just pretent to be a proxy" excuse, what were the
> complaints you got about the behavior about? For whatever gets upset by
> your original implementation would probably also get upset by the
> presence of proxies.
>
> BR
> c
>