Re: [Qemu-devel] [Nbd] [PATCH 3/1] doc: Propose Structured Replies extension

[Wouter Verhelst <w@uter.be>]

On Tue, Mar 29, 2016 at 08:51:57PM +0200, Wouter Verhelst wrote:
> On Tue, Mar 29, 2016 at 12:23:31PM -0600, Eric Blake wrote:
> > Unfortunately, I chose the design of 0 or more structured replies
> > followed by a normal reply, so that the normal reply is a reliable
> > indicator that the read is complete (whether successful or not); and the
> > whole goal of the extension is to avoid sending any data payload on a
> > normal reply.  I'm not sure how to send the offset in the normal reply
> > without violating the premise that a normal reply has no payload.
> 
> Oh. I thought you meant for the concluding message to also be a
> structured reply with the length field be zero, but you mean for it to
> be a non-structured reply message? If so, you should clarify that a bit
> more (this wasn't clear to me)...

Also, I'm not convinced that's a very good approach, since it also
requires analyzers to have more context than just requiring a single
final "empty" structured reply message.

-- 
< ron> I mean, the main *practical* problem with C++, is there's like a dozen
       people in the world who think they really understand all of its rules,
       and pretty much all of them are just lying to themselves too.
 -- #debian-devel, OFTC, 2016-02-12