Amit Kapila <amit.kapila16@gmail.com> writes:
> Attached should fix the reported warnings but as I am not getting
> those warnings on my system so can't confirm but otherwise, the patch
> works as expected. I am planning to push this unless you or Peter have
> any comments.
Personally I'd get rid of the option for ReplicationSlotNameForTablesync
to allocate the result buffer. There's only one caller using that,
so it's saving no code to have ReplicationSlotNameForTablesync do the
alloc rather than that one caller; and it seems mighty ugly/nonorthogonal
to have effectively two different APIs in one function.
regards, tom lane