Re: Corrected documentation of data type for the logical replication message formats.

Peter Smith <smithpb2250(at)gmail(dot)com> writes:
> I agree. The specified value looks better when it comes first, as you did it.

Actually, it looks to me like we don't have to resolve the question of
which should come first, because I don't see any cases where it's
useful to have both. I don't agree with appending "uint8" to those
field descriptions, because it's adding no information, especially
when the high bit couldn't be set anyway.

At some point it might be useful to add UInt<n> to the set of base
data types, and then go through all the message types and decide
which fields we think are unsigned. But that is not this patch,
and there would be questions about whether it constituted a protocol
break.

I noticed also that having to add "(Oid)" was sort of self-inflicted
damage, because the field descriptions were using the very vague
term "ID", when they could have said "OID" and been clear. I left
the "(Oid)" additions in place but also changed the text.

Pushed with those changes. I couldn't resist copy-editing the section
intro, too.

regards, tom lane