Ken Brown wrote:
> Did you perhaps get the condition (socket_to_use < 0) backwards?  If
> socket_to_use < 0, then it is not external_sock_fd that has been closed.

Yes, thanks, I think you're right. I installed the attached on master to fix that.