Thanks for the article link. I was going to write a blog article describing
this problem in more detail. :)
Sent from my iPhone
On Mar 15, 2016, at 2:11 PM, Jason E. Aten <j.e.aten@xxxxxxxxx> wrote:
The end-to-end arguments suggests that the sending application should wait
for an ack from the receiving application.
There's no other way to know for sure that it's gotten there.
https://en.wikipedia.org/wiki/End-to-end_principle
On Tue, Mar 15, 2016 at 7:33 AM, Fernando Testa <fernando@xxxxxxxxxx> wrote:
Hello all,
I'm new to nanomsg, and very pleased to play with the library.
My app has a parent that receives data from a child. After the child has
sent some data, it sends a end of transmission (EOT) special message that
breaks the loop in the parent.
If the child dies straight away, sometimes the EOT never reaches the parent.
Right now, I've put a sleep some milliseconds before the
I couldn't find a proper way to block until all messages are sent.
Sockets already have NN_LINGER timeout to -1, but that does not appear to
fix if the child just exit.
int timeout = -1; //infinite
int rc = nn_setsockopt(socket_, NN_SOL_SOCKET, NN_LINGER, &timeout,
sizeof(timeout));
Any suggestions?
FT
