ietf-nntp Concerning Streaming

[Stan Barber]

Clive D.W. Feather wrote:

> Stan Barber said:
> 
>>>I don't see why LIST EXTENSIONS can't be streamable.  Yes, it would be kind
>>>of dumb to go ahead and depend on behavior the server hasn't confirmed yet,
>>>but that's also true of anything that changes the group context or article
>>>pointer, which is, well, most commands.
>>>
>>I think most readers keep up with group context and article poster. 
>>Given that readers can actually check for extensions without having to 
>>try them first, I didn't want to presume anything about the readers 
>>keeping up with what extensions were available between invocations of 
>>LIST EXTENSIONS. More discussion on this are welcome.
>>
> 
> I don't see that this is anything to do with keeping up with context;
> that's a client issue. Rather, streaming (or pipelining, as I think we
> agreed to call it) is to do with whether the server is entitled to throw
> away text that the client sent. MODE READER has good reasons for this; does
> anything else ?



We didn't agree to call it pipelining. See my other note on that.

The maintenance of state by the client and the expectation of that 
client that the server will stop steaming are interconnected. You can't 
think about one without thinking about the other.