Daniel Gustafsson <daniel(at)yesql(dot)se> writes:
> Doh, of course. The attached is a quick (lightly make check tested) take on
> allowing 0, but I'm not sure that's what we want?
Actually, wait a second. At least some of these are not dealing
with kernel FDs but with our "virtual FD" abstraction. For those,
zero is indeed an invalid value. There might well be some "<= 0"
versus "< 0" errors in this area, but it requires closer inspection.
regards, tom lane