see also:
https://github.com/rejectedsoftware/vibe.d/issues/1431 api to find an
available port
On Fri, Mar 18, 2016 at 2:50 AM, Marc Schütz via Digitalmars-d-learn <
digitalmars-d-learn@puremagic.com> wrote:
> Looking at an strace of nmap, it seems it opens a bunch of sockets, puts
> them into non-blocking mode, calls connect on them (which will return
> EINPROGRESS), and then uses select(2) to wait for them (in a loop, until
> all have either been accepted or rejected). select(2) accepts a timeout
> value, so you can determine how long you want to wait.
>
> Here's an excerpt:
>
> ...
> socket(PF_INET, SOCK_STREAM, IPPROTO_TCP) = 50
> fcntl(50, F_GETFL) = 0x2 (flags O_RDWR)
> fcntl(50, F_SETFL, O_RDWR|O_NONBLOCK) = 0
> setsockopt(50, SOL_SOCKET, SO_LINGER, {onoff=1, linger=0}, 8) = 0
> setsockopt(50, SOL_SOCKET, SO_BINDTODEVICE, [0], 4) = -1 EPERM (Operation
> not permitted)
> setsockopt(50, SOL_IP, IP_TTL, [-1], 4) = 0
> connect(50, {sa_family=AF_INET, sin_port=htons(32778),
> sin_addr=inet_addr("127.0.0.1")}, 16) = -1 EINPROGRESS (Operation now in
> progress)
> socket(PF_INET, SOCK_STREAM, IPPROTO_TCP) = 51
> fcntl(51, F_GETFL) = 0x2 (flags O_RDWR)
> fcntl(51, F_SETFL, O_RDWR|O_NONBLOCK) = 0
> setsockopt(51, SOL_SOCKET, SO_LINGER, {onoff=1, linger=0}, 8) = 0
> setsockopt(51, SOL_SOCKET, SO_BINDTODEVICE, [0], 4) = -1 EPERM (Operation
> not permitted)
> setsockopt(51, SOL_IP, IP_TTL, [-1], 4) = 0
> connect(51, {sa_family=AF_INET, sin_port=htons(1029),
> sin_addr=inet_addr("127.0.0.1")}, 16) = -1 EINPROGRESS (Operation now in
> progress)
> socket(PF_INET, SOCK_STREAM, IPPROTO_TCP) = 52
> fcntl(52, F_GETFL) = 0x2 (flags O_RDWR)
> fcntl(52, F_SETFL, O_RDWR|O_NONBLOCK) = 0
> setsockopt(52, SOL_SOCKET, SO_LINGER, {onoff=1, linger=0}, 8) = 0
> setsockopt(52, SOL_SOCKET, SO_BINDTODEVICE, [0], 4) = -1 EPERM (Operation
> not permitted)
> setsockopt(52, SOL_IP, IP_TTL, [-1], 4) = 0
> connect(52, {sa_family=AF_INET, sin_port=htons(2013),
> sin_addr=inet_addr("127.0.0.1")}, 16) = -1 EINPROGRESS (Operation now in
> progress)
> select(53, [3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
> 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
> 51 52], [3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
> 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51
> 52], [3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
> 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51
> 52], {0, 0}) = 100 (in [3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
> 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
> 47 48 49 50 51 52], out [3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
> 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
> 47 48 49 50 51 52], left {0, 0})
> ...
>
> I'm pretty sure the setsockopt() calls aren't essential.
>
