Recreating the socket is trivial. Recreating all the stuff that the various IO:* modules attach to the glob is far less simple.

True, but who cares. He has a freshly accepted socket, so all he needs is

sub new_inet_from_fd {
    my ($fd, %arg) = @_;
    my $timeout  = delete $arg{Timeout};
    my $blocking = exists $arg{Blocking} ? delete $arg{Blocking} : 1;

    my $sock = IO::Socket::INET->IO::Handle::new_from_fd($fd, '+<');

    ${*$sock}{'io_socket_timeout'} = $timeout;

    ${*$sock}{io_sock_nonblocking} = !$blocking;
    $sock->blocking($blocking);

    $sock->autoflush(1);

    return $sock;
}
[download]

It produces a socket whose operations don't timeout (unless one is specified), whose operations do block (unless otherwise specified), and that does autoflush. Everything in IO::Handle other than autoflush is at its default.

I'm not sure where you get all the stuff about blocking and timeouts from? It doesn't appear to be derived from anything the OP posted.

If he is only ever going to need simple IO::Sockets, he might get away with it, but if he needs any of the stuff that layers atop IO::Socket, he'll find the limitations for sure.

---

Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.

"Science is about questioning the status quo. Questioning authority".

In the absence of evidence, opinion is indistinguishable from prejudice.

RIP PCW It is as I've been saying!(Audio until 20090817)