Re: The new version of draft-yevstifeyev-http-headers-not-recognized (Last Call)


I'll now explain that. Let A is the host-under-debugging (HUD). There is 
another host, B, which is debugger. Let there is some new HTTP server 
software on A. B's sending requests with different methods, headers and 
entity to test A for appropriate processing of headers. Let B is sending 
the request with header X-Foo, which is to be used or already used 
widely among the Internet. But A does not recognize it. But it is not 
easy for developers to get known this. If discussed header is used, B 
testers can simply analyze A's answer (as it is to be done at all) to 
find out the problem and fix it. That is debugging purpose, IMO.

Hope I explained everything clearly.

Any suggestions are welcome.

All the best,
Mykyta Yevstifeyev

16.12.2010 17:46, Daniel Stenberg wrote:
> On Thu, 16 Dec 2010, Mykyta Yevstifeyev wrote:
>> Daniel,
>> You may find some related discussions on
>> dated from Monday, 22 November 2010 and Tuesday, 23 November 2010
> Thanks.
> The only tiny piece of explanation I can find there is 
> where you say "But aren't debugging purposes a weighty argument?".
> Does this mean that debugging HTTP clients(?) and servers(?) is your 
> primary purpose for this feature? In what way does this help you debug?
> I've been developing (and debugging) HTTP clients for over a decade 
> and I've not missed this header. I'm genuinely trying to understand 
> the use case and what benefit this header is to bring for that 
> purpose, so that I can read the draft with that in mind.
> So please, even if this is repeating something you've expressed 
> before, can you enlighten us (again) exactly in what ways you intend 
> to use this header/feature?
> (If someone else have any good suggestions, that'll do fine as well!)

Received on Thursday, 16 December 2010 16:00:26 UTC