Jacob Champion <jchampion(at)timescale(dot)com> writes:
> The guc_strdup() approach really reduces the amount of code, so that's
> what I did in v3. I'm not following why we need to return NULL on
> failure, though -- both palloc() and guc_malloc() ERROR on failure, so
> is it okay to keep those semantics the same?
guc_malloc's behavior varies depending on elevel. It's *not*
equivalent to palloc.
regards, tom lane