Bruce Momjian <bruce(at)momjian(dot)us> writes:
> Is there a reason tablespace.c::set_short_version() uses PG_VERSION, and
> not the simpler PG_MAJORVERSION? initdb.c::get_short_version() does the
> same thing.
Probably that code predates the addition of the separate PG_MAJORVERSION
#define. +1 for simplifying. The change I just had to make in
backend/catalog/Makefile reinforces the thought that MAJORVERSION is
what we should be using in everyplace related to this type of check.
regards, tom lane