On Mon, Feb 20, 2023 at 03:22:30PM +0000, Pieter Wuille via bitcoin-dev wrote:
> On Sunday, February 19th, 2023 at 6:56 PM, Anthony Towns 
> <a...@erisian.com.au> wrote:
> > On Fri, Feb 17, 2023 at 10:13:05PM +0000, Pieter Wuille via bitcoin-dev 
> > wrote:
> > > > I think it's probably less complex to close some of the doors?
> > > > 2) are short ids available/meaningful to send prior to VERACK being
> > > > completed?
> > > > Ah, I hadn't considered this nuance. If we don't care about them being 
> > > > available before VERACK negotiation, then it may be possible to 
> > > > introduce a way to negotiate a different short id mapping table without 
> > > > needing a mechanism for re-negotiating.
> > I think you still need/want two negotiation steps -- once to tell each
> > other what tables you know about, once to choose a mutually recognised
> > table and specify any additions.
> Right, I wasn't talking about how many steps/messages the negotiation takes. 
> I just meant that if all negotiation of the mapping table happens just once 
> (before VERACK) and that negotiation itself happens without use of short 
> commands, then there is no need for re-negotiating short commands after they 
> are already in use. Nothing concrete, but I can imagine that that may 
> simplify some implementations.

Yeah; I was just thinking of the fact that currently the negotiation is:

 * send your VERSION message
 * see what their VERSION message is

 * announce a bunch of features, depending on the version (or service
 * send the VERACK (and GETADDR and final ALERT)

 * wait for their announcements and VERACK
 * negotiation is finished; we know everything; we're ready to go

which only gets you two steps if you send the short id stuff as part of
the VERSION message. Obviously you could just add an extra phase either
just before or just after the VERACK, though.

I suppose being able to choose your own short id mapping from day 0 would
mean that every bip324 node could use a single short id mapping for all
outgoing messages, which might also make implementation marginally easier
(no need to use one table for modern nodes, but also support the original
table for old bip324 implementations)...

bitcoin-dev mailing list

Reply via email to