I finally managed to perform this test on sparc v9 machine which is 64
bit big-endian architecture. I run pgbench script (see previous message)
with default_seed=123 on both x86-64 and sparc machines and visualized
the results. You can find them in the attached chart. Both images showed
the same distribution. So endianness isn't a concern here.
Agree, pushed.
But I have a notice about number of arguments. Seems, special values for hash
and greatest/least functions is not actually needed. If we split nargs option to
n mandatory arguments and n optional ones then special values for that functions
will go away: hash will have 1 mandatory and 1 optional, greatest/least will
have one mandatory and infinite number of optional. Not sure for now about
CASE/WHEN case. But seems it's a deal for separate refactoring.
--
Teodor Sigaev E-mail: teo...@sigaev.ru
WWW: http://www.sigaev.ru/