Re: PUSH Clarifications

In order to justify making push extensible, we ought to require someone to
put a semi-reasonable use case on the table.
 On Aug 6, 2013 2:37 PM, "Julian Reschke" <julian.reschke@gmx.de> wrote:

> On 2013-08-06 23:03, James M Snell wrote:
>
>> On Tue, Aug 6, 2013 at 1:57 PM, Julian Reschke <julian.reschke@gmx.de>
>> wrote:
>> [snip]
>>
>>>
>>>>> Wrong.
>>>>>
>>>>> <http://greenbytes.de/tech/**webdav/draft-ietf-httpbis-**
>>>>> method-registrations-11.html#**updated.registry.contents<http://greenbytes.de/tech/webdav/draft-ietf-httpbis-method-registrations-11.html#updated.registry.contents>
>>>>> >
>>>>>
>>>>>
>>>> Well... yes, there are the PROPFIND, SEARCH and REPORT methods which
>>>> are both safe and idempotent... however, given that PUSH_PROMISE gives
>>>> us no means of sending an implied payload along with the implied
>>>>
>>>
>>>
>>> Yes, that's what needs to get fixed.
>>>
>>>
>> I hear you but I disagree ;-)
>>
>> Perhaps there is a non-theoretical, somewhat-plausible use-case that
>> cannot be addressed using GET/HEAD that I'm missing here? If not, lets
>> draw a line in the sand with GET/HEAD and move on.
>>
>> - James
>>
>
> We need to decide whether PUSH is limited to GET (/HEAD), or extensible.
> If it is supposed to work with other methods (and that's what the rough
> consensus in the interim meeting (*)) was, then we should prepare for safe
> methods with payloads no. If we do not, we'll need to rev the base protocol
> which seems to be a very bad idea.
>
> Best regards, Julian
>
> (*) to be confirmed on the mailing list
>
>