XBATCH Command (Was: Re: New NNTP extensions in INN 1.5?)

[Jonathan Grobe]

On Mon, 15 Jul 1996, James A. Brister wrote:

> inn 1.5 will be adding the XBATCH command which lets the server
> take in compressed rnews-ready batches. A rough description is:
> 
> 	XBATCH nnn
> 
> 	The XBATCH command instructs the server that the client has an
> 	batch of articles to be sent and the batch is nnn bytes long. 
> 	If the server wishes the batch to be
> 	sent then it will return a response instructing the client to do
> 	so. If the server doesn't want the batch to be sent (if the server
> 	is a slave server, for example), then a response indicating that
> 	will be sent.
> 
> 	If transmission of the batch is requested, then the client should
> 	send the batch in binary mode--that is no LF->CR-LF translation
> 	must be done and no terminating 'CR-LF.CR-LF' is used to to
> 	terminate the batch.
> 
> 	The format of the batch is identical to what rnews(1) expects to
> 	process.
> 
> 	Responses:
> 
> 	239 batch transferred successfully
> 	339 send batch to be transferred
> 	436 xbatch failed: reason
> 	501 xbatch: invalid or missing size.
> 
> inn 1.5 rnews adds recognition of gzip compression to batches.
> 
If James is willing to provide a precise description (instead of this
rough one) I believe it is reasonable that it be added to the August
draft.

I have no idea of when INN 1.5 is going to come out. If it will be a 
while, I believe this command should be discussed now to see if this
list thinks any changes should be made. For example:

1. I think the "X" should be dropped, the name changed to BATCH. RFC977
said:

   There is no prohibition against additional commands being added;
   however, it is recommended that any such unspecified command begin
   with the letter "X" to avoid conflict with later revisions of this
   specification.

But if this is going to be in the standard, then this suggests it
should not start with X. 

2. Should there be some kind of optional parameter... to allow
different types of batching procedures to be added to the standard
later?

Jonathan Grobe <grobe at netins.net>