Peter Eisentraut <peter(dot)eisentraut(at)2ndquadrant(dot)com> writes:
> Committed with that addition.
Some of the buildfarm doesn't like this. I realized that the common
denominator was EXEC_BACKEND: if you build with -DEXEC_BACKEND the
failure is completely reproducible. Presumably something is thinking
that it will inherit some variable from the postmaster via fork.
I don't know if that's a new bug or was just exposed by a new test.
regards, tom lane