On Mon, Sep 11, 2023 at 02:42:16PM +0900, bt23nguyent wrote:
> A minor issue with the description here is that while the description for
> the new flush argument in pg_logical_emit_message() with bytea type is
> clearly declared, there is no description of flush argument in the former
> pg_logical_emit_message() with text type at all.
Indeed, I forgot to update the first function signature. Fixed in the
attached.
> On a side note, could you also include a bit more information that "flush is
> set to false by default" in the document as well? It could be helpful for
> the users.
With the function signature saying that, that did not seem stricly
necessary to me, but no objections to add a few words about that.
I'll need a bit more input from Fujii-san before doing anything about
his comments, still it looks like a doc issue to me that may need a
backpatch to clarify how the non-transactional case behaves.
Attaching a v4 with the two doc changes, fow now.
--
Michael