Re: lots of unused variable warnings in assert-free builds

Robert Haas <robertmhaas@gmail.com> writes:
> On Tue, Jan 24, 2012 at 12:12 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> Ouch. Is it really true that __attribute__((unused)) disables detection
>> of use of uninitialized variables?

> Oh, I think maybe I am confused.  The downsides of disabling *unused*
> variable detection are obviously much less than the downsides of
> disabling *uninitialized* variable declaration... although neither is
> ideal.

OK.  MHO is that __attribute__((unused)) is probably less annoying than
#ifdef overall.  Also, it occurs to me that an intermediate macro
"PG_USED_FOR_ASSERTS_ONLY" would be a good idea, first because it
documents *why* you want to mark the variable as possibly unused,
and second because changing the macro definition would provide an easy way
to check for totally-unused variables, in case we wanted to periodically
make such checks.

This is all modulo the question of what pgindent will do with it,
which I would still like to see tested before we commit to a method.
        regards, tom lane