I can read from documentation:
"A client-streaming RPC is also similar to our simple example, except the client
sends a stream of requests to the server instead of a single request. The server
sends back a single response, typically but not necessarily after it has
received all the client’s requests, along with its status details and optional
trailing metadata."
Usually the client does:
- create channel
- creaet stub
- perform the writer = stub->foo(context, reply);
- loop on:
writer->Write(request)
- writer->Done()
- writer->Finish()
- At this point I can read the reply
now my question is how can I read the reply before the loop is over? Some time the server indeed stops performing the read loop notifying that there is no need for more requests.