Andres Freund <andres@anarazel.de> writes:
> On 2015-08-12 18:52:59 -0400, Tom Lane wrote:
>> Looks OK to me, except I wonder why you did this
>>
>> #define TRIGGER_FIRED_FOR_ROW(event) \
>> - ((event) & TRIGGER_EVENT_ROW)
>> + (((event) & TRIGGER_EVENT_ROW) == TRIGGER_EVENT_ROW)
>>
>> rather than != 0. That way doesn't look either more efficient or
>> more readable.
> Purely consistency with the surrounding code. I was on the fence about
> that one...
The adjacent code is doing something different than a bit-test, though:
it's checking whether multibit fields have particular values.
regards, tom lane