v0.0.7 Update: Adding Keepalive Timeout to TCP RPC Server

Version v0.0.7 builds on v0.0.6 by introducing a keepalive timeout setting for the TCP RPC server, allowing the server to automatically close idle client connections after a configurable period, thereby reducing performance overhead, saving resources, and preventing memory leaks.

When the server closes a client connection, subsequent calls from the client’s call() method will trigger an automatic reconnection, which is transparent to the user, so developers do not need to handle the closed‑connection case manually.

The default keepalive timeout is set to 75 seconds, matching Nginx’s default, but it can be customized via the setKeepaliveTimeout() method as shown in the example below.

from agileutil.rpc.server import TcpRpcServer

def sayHello(name):
    return 'hello ' + name

s = TcpRpcServer('0.0.0.0', 9988)
s.setKeepaliveTimeout(10)  # server will close idle connections after 10 seconds
s.regist(sayHello)
s.serve()

Client usage example demonstrates creating a TcpRpcClient , making a call, sleeping longer than the timeout, and then calling again, which causes the client to automatically reconnect.

from agileutil.rpc.client import TcpRpcClient
import time

cli = TcpRpcClient('127.0.0.1', 9988, timeout = 2)
resp = cli.call(func = 'sayHello', args=('zhangsan',))
print('resp', resp)
time.sleep(15)
resp = cli.call(func = 'sayHello', args=('zhangsan',))  # client will auto‑reconnect
print('resp', resp)