Re: Is float8 a reference type?

2017-09-23 5:10 GMT+02:00 Paul A Jungwirth <pj(at)illuminatedcomputing(dot)com>:

> On Fri, Sep 22, 2017 at 8:05 PM, Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com>
> wrote:
> > yes, it is 8 bytes on 64-bit.
>
> Thanks!
>
> > I don't think so it is good idea to write 64bit only extensions.
>
> I agree, but how about this?:
>
> if (FLOAT8PASSBYVAL) {
> datums = (Datum *)floats;
> } else {
> datums = palloc0(arrlen * sizeof(Datum));
> for (i = 0; i < arrlen; i++) {
> datums[i] = Float8GetDatum(floats[i]);
> }
> }
>

it can work.

You have to solve deallocation in only one path. palloc0 is not necessary
in this case.

>
> Thanks,
> Paul
>