我正在尝试编写一个 python 程序来测试用 C 编写的服务器。 python 程序使用subprocess
模块启动编译的服务器:
pid = subprocess.Popen(args.server_file_path).pid
这工作正常,但是如果 python 程序由于错误而意外终止,则生成的进程将继续运行。我需要一种方法来确保如果 python 程序意外退出,服务器进程也会被终止。
更多细节:
- 仅限 Linux 或 OSX 操作系统
- 服务器代码不能以任何方式修改
我正在尝试编写一个 python 程序来测试用 C 编写的服务器。 python 程序使用subprocess
模块启动编译的服务器:
pid = subprocess.Popen(args.server_file_path).pid
这工作正常,但是如果 python 程序由于错误而意外终止,则生成的进程将继续运行。我需要一种方法来确保如果 python 程序意外退出,服务器进程也会被终止。
更多细节:
I would atexit.register
a function to terminate the process:
import atexit
process = subprocess.Popen(args.server_file_path)
atexit.register(process.terminate)
pid = process.pid
Or maybe:
import atexit
process = subprocess.Popen(args.server_file_path)
@atexit.register
def kill_process():
try:
process.terminate()
except OSError:
pass #ignore the error. The OSError doesn't seem to be documented(?)
#as such, it *might* be better to process.poll() and check for
#`None` (meaning the process is still running), but that
#introduces a race condition. I'm not sure which is better,
#hopefully someone that knows more about this than I do can
#comment.
pid = process.pid
Note that this doesn't help you if you do something nasty to cause python to die in a non-graceful way (e.g. via os._exit
or if you cause a SegmentationFault
or BusError
)