John M McIntosh wrote:
Do you have a test case you can recreate?
No, not easily. The things we did on the local machine worked (trivial little tests just to see if there is an obvious problem or not).
On destroy, if the code thinks the socket is valid it should close() after setting linger to zerot, but if it gets an EINTR? Anything in the system/console logs, if the close failed with an unhandled condition it should have printed the error to the logs with the output "closeConnection"
Thanks, we'll try again and check to see if there is any output.
Cheers, - Andreas