On Tue, 10 May 2022 at 21:45, John Naylor <john.naylor@enterprisedb.com> wrote:
>
> On Tue, May 10, 2022 at 3:46 PM David Rowley <dgrowleyml@gmail.com> wrote:
> > What do you think?
>
> +1
Thanks for having a look.
I pushed this after I made a small adjustment to #ifdef out the
qsort_tuple_signed sort specialization in 32-bit builds. On testing a
32-bit build with the patch I proposed, I was getting a warning about
that function being unused.
I found the 32-bit postgres binary is now an entire 4376 bytes smaller!
David