Peter Geoghegan <[email protected]> writes: > On Fri, Jan 29, 2021 at 6:44 PM Tom Lane <[email protected]> wrote: >> Pointer width is interesting, but really it's a solved problem >> compared to these.
> What about USE_FLOAT8_BYVAL?
That's an annoyance, sure, but I don't recall many recent bugs
related to violations of that coding rule. As long as you don't
violate the Datum abstraction it's pretty safe.
regards, tom lane
