Non-blocking fhandler_socket_wsock::send_internal ?

Non-blocking fhandler_socket_wsock::send_internal ?

[Ford, Brian]

I'm trying to debug an application that occasionally takes an excessive amount of time in the sendto call of a non-blocking UDP socket and I noticed the following code that I didn't understand but looked slightly suspicious:
1423       while (res && (WSAGetLastError () == WSAEWOULDBLOCK)
1424              && !(res = wait_for_events (FD_WRITE | FD_CLOSE, wait_flags)));

Could you explain why I shouldn't be concerned about a non-blocking socket apparently calling a function labeled wait_for_events before returning :-)?

Thank you for your time!

[https://elearning.flightsafety.com/media/wysiwyg/SIGNATURE_BANNER/FlightSafety_65th_logo.jpg]<https://elearning.flightsafety.com/fsi_65_redirect>

Brian Ford  *  Systems Architect
FVS  *  FVS Engineering  *  5695 Campus Parkway  *  Hazelwood, MO, 63042  *  United States
Tel: 314.551.8460  *  Cell: 314.3276088
Brian.Ford@FlightSafety.com<mailto:Brian.Ford@FlightSafety.com>  *  flightsafety.com<https://www.flightsafety.com/>  *  A Berkshire Hathaway company
Privacy Notice<https://www.flightsafety.com/fs_privacynotice.php>

RE: Non-blocking fhandler_socket_wsock::send_internal ?

[Ford, Brian]

Nevermind, I see in wait_for_events how it is supposed to work.  Sorry for the noise.  I'd be glad to hear any reasons you might expect a non-blocking UDP socket sendto to block for 50-100 ms though.  Thanks.

-----Original Message-----
From: Cygwin-developers [mailto:cygwin-developers-bounces@cygwin.com] On Behalf Of Ford, Brian via Cygwin-developers
Sent: Monday, March 30, 2020 3:45 PM
To: cygwin-developers@cygwin.com
Subject: Non-blocking fhandler_socket_wsock::send_internal ?

EXTERNAL EMAIL: STOP and verify sender address, links, and attachments before taking action.


I'm trying to debug an application that occasionally takes an excessive amount of time in the sendto call of a non-blocking UDP socket and I noticed the following code that I didn't understand but looked slightly suspicious:
1423       while (res && (WSAGetLastError () == WSAEWOULDBLOCK)
1424              && !(res = wait_for_events (FD_WRITE | FD_CLOSE, wait_flags)));

Could you explain why I shouldn't be concerned about a non-blocking socket apparently calling a function labeled wait_for_events before returning :-)?

Thank you for your time!

[https://elearning.flightsafety.com/media/wysiwyg/SIGNATURE_BANNER/FlightSafety_65th_logo.jpg]<https://elearning.flightsafety.com/fsi_65_redirect>

Brian Ford  *  Systems Architect
FVS  *  FVS Engineering  *  5695 Campus Parkway  *  Hazelwood, MO, 63042  *  United States
Tel: 314.551.8460  *  Cell: 314.3276088
Brian.Ford@FlightSafety.com<mailto:Brian.Ford@FlightSafety.com>  *  flightsafety.com<https://www.flightsafety.com/>  *  A Berkshire Hathaway company Privacy Notice<https://www.flightsafety.com/fs_privacynotice.php>