I wrote:
> ... It doesn't take much to make one nontrivial either. If
> memory serves, malloc() can trash errno on some platforms such as macOS,
> so even just a palloc creates a hazard of a hard-to-reproduce problem.
After digging around in the archives, the closest thing that we seem to
know for certain is that some versions of free() can trash errno, cf
https://www.postgresql.org/message-id/flat/E1UcmpR-0004nn-2i%40wrigleys.postgresql.org
as a result of possibly calling madvise() which might or might not
succeed. But in the light of that knowledge, how much do you want
to bet that malloc() can't change errno? And there's the possibility
that a called function does both a palloc and a pfree ...
regards, tom lane