On Thu, Aug 17, 2017 at 1:28 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> -1 ... every datatype I/O function is entitled to assume it's being
> invoked inside a transaction. I do not think we should break that
> on a case-by-case basis. So using timestamptz_in directly in xlog.c
> was a bad idea, and we need to rethink that.
[evil mode]
Why not just calling SetCurrentStatementStartTimestamp() before
parsing the timestamp? A bit hacky, but I think that it would work.
[/evil mode]
> Not sure offhand how to mechanize that given the current code
> structure. We don't want to duplicate all the infrastructure
> used by timestamptz_in, certainly ... but how much of that
> code is depending in some way on being inside a transaction?
> I doubt that this issue with "now" is the only soft spot.
tomorrow, today, etc are other things.
--
Michael