-
…he message. This was neccessary because after I activated optimization in gcc the testclient was always only reading 2 bytes in the initial successfull read.
-
…around 1.1 GBytes on a loopback device which sounds not to bad, but is only about 1/4 of the throughput iperf shows me... but this might be related to nc... I will see this when I have written my own client.
-
…oints to read or write when they are already in fin state.
-
…sed. In fact this works now with edge triggered epoll
-
…d it makes things unnessecary complicated.
-
…Point which was neccessary... I have to add it again.