Peter Eisentraut <peter.eisentraut@2ndquadrant.com> writes:
> On 2019-11-01 15:41, Robert Haas wrote:
>> On a related note, why do we store typbyval in the catalog anyway
>> instead of inferring it from typlen and maybe typalign? It seems like
>> a bad idea to record on disk the way we pass around values in memory,
>> because it means that a change to how values are passed around in
>> memory has ramifications for on-disk compatibility.
> This sounds interesting. It would remove a pg_upgrade hazard (in the
> long run).
> There is some backward compatibility to be concerned about.
Yeah. The point here is that typbyval specifies what the C functions
concerned with the datatype are expecting. We can't just up and say
"we're going to decide that for you".
I do get the point that supporting two different typbyval options
for float8 and related types is a nontrivial cost.
regards, tom lane