Re: Support for gzip at the server #424 (Consensus Call)

On 2014-03-18 19:13, Bjoern Hoehrmann wrote:
> * Julian Reschke wrote:
>> On 2014-03-18 18:04, Roberto Peon wrote:
>>> Note that this is true regardless of whether or not compression is in
>>> use-- anything doing chunked entity-body suffers from this when
>>> gatewaying to a 1.0 server.
>>> -=R
>>> ...
>>
>> Ok. So support for chunked already is mandatory -- why can't we add gzip?
>
> What I understand so far is that flawlessly implemented support for
> chunked requests is not universally deployed, and buffering requests
> in intermediaries is not a viable option in any and all cases without
> exception. What is mandated is not relevant to those concerns.

I understand the concerns.

However, just because there are concerns doesn't mean we can't mandate 
it. If we all agree that client->server gzip is desirable, then the spec 
should be clear about that. Be it in non-normative prose, with a 
"SHOULD" or even a "MUST" (which would be similar to the "MUST support 
chunked encoding" in 1.1).

What I hear is that just because there are concerns we'll say nothing. 
My take is that the situation will not improve unless we say *something*.

Best regards, Julian

Received on Tuesday, 18 March 2014 18:25:43 UTC