折腾：
【部分解决】Python中实现多线程或多进程中的单例singleton
期间，对于之前的ThreadSafeSingleton，虽然多线程thread测试是OK的，但是对于gunicorn的多worker是无效的。
gunicorn的多worker，目前的理解，应该属于多进程process。
所以也想想办法Mac本地能去调试多进程的情况，看看ThreadSafeSingleton是否有效。
python multiple process singleton
tendo/singleton.py at master · pycontribs/tendo
好像是支持多thread？
multiprocessing.Manager Python Example
Singleton Objects in Python – Michael Goerz
python singleton into multiprocessing – Stack Overflow
-》用shared memory去在多线程之间共享数据：
17.2. multiprocessing — Process-based parallelism — Python 3.7.0 documentation
python – Make Singleton class in Multiprocessing – Stack Overflow
python – State shared in multiprocessing when using singleton classes – Stack Overflow
16.6. multiprocessing — Process-based “threading” interface — Python 2.7.15 documentation
process – Sharing a complex object between Python processes? – Stack Overflow
17.2. multiprocessing — Process-based parallelism — Python 3.7.0 documentation
而对于如果真是多个process之间数据无法共享单例，那么暂时就放弃这条路，
暂时不去考虑用：shared memory，mmap等其他方式实现了，暂时还是考虑用：
【已解决】用gunicorn的gevent解决之前多worker多Process线程的单例的数据共享
的办法。
此处，先还是去参考：
https://github.com/pycontribs/tendo/blob/master/tendo/singleton.py
和：
17.2.1.4. Synchronization between processes
创建多个process去测试看看再说
期间对于Process类，有哪些函数和属性，参考：
17.2. multiprocessing — Process-based parallelism — Python 3.7.0 documentation
其中就有我要的：pid
"pid
Return the process ID. Before the process is spawned, this will be None."
以及：
https://docs.python.org/3/library/multiprocessing.html#multiprocessing.current_process
"multiprocessing.current_process()
Return the Process object corresponding to the current process.
An analogue of threading.current_thread()."
【总结】
最后用代码：
log.info("========== test multiple process singleton ==========")
import multiprocessing
def singleProcessDo(cur_num):
log.info("———- singleProcessDo ———-")
cur_process = multiprocessing.current_process()
curProcessTTokenSingleton = MsTtsTokenSingleton()
log.info("curProcessTTokenSingleton=%s", curProcessTTokenSingleton)
log.info("[%d] name=%s, pid=%s, process=%s", cur_num, cur_process.name, cur_process.pid, cur_process)
max_process_num = 5
for idx in range(max_process_num):
cur_num = idx + 1
each_process_name = "P%s" % cur_num
cur_process = multiprocessing.Process(target=singleProcessDo, name=each_process_name, args=(cur_num, ))
log.info("[%d] name=%s, process=%s", cur_num, each_process_name, cur_process)
cur_process.start()
测试结果竟然是：
多Process进程中，对应的单例，也是工作的：
【后记】
但是gunicorn中的多个worker中单例，就是不工作的：还是会初始化多个实例的。
最终发现是另外的原因：