By: Tom Sydney Kerckhove <syd@cs-syd.eu>
warp: say plainly why EBADF is the whole test The comment for the quiet-exit guard was doing too much work in one sentence and @Vlix could not follow it. What it needs to say is only that 'close' swaps the descriptor for -1 first, so every accept() after a deliberate shutdown fails with EBADF and nothing else, which is what makes EBADF sufficient on its own.