The Python docs about asyncio - Subprocess say:
using System.ComponentModel; namespace WpfApplication1.ViewModels { public abstract class ViewModel : INotifyPropertyChanged { public event PropertyChangedEventHandler PropertyChanged; protected string _title; protected string Title { get { return _title; } set { _title = value; OnPropertyChanged(nameof(Title)); } } protected void OnPropertyChanged(string propName) { PropertyChanged?.Invoke(this, new PropertyChangedEventArgs(propName)); } } }
和communicate()
方法不使用超时参数:使用wait()
函数
使用wait_for()
在communicate()
上施加超时非常容易,但我找不到从中断的wait_for()
调用中检索部分结果的方法,以及后续调用communicate()
也不会返回丢失的部分。
示例脚本:
communicate()
示例脚本的输出:
#! /usr/bin/env python3
import asyncio
async def communicate_short(loop):
p = await asyncio.create_subprocess_exec('ping', '127.0.0.1', '-n', '4', stdout=asyncio.subprocess.PIPE)
# For Linux: use '-c' instead of '-n'
try:
# 2 seconds timeout
res = await asyncio.wait_for(p.communicate(), 2)
except asyncio.TimeoutError as e:
# After timeout happens:
# How do I get the subprocess's STDOUT up to this point?
try:
print(res[0].decode('utf-8'))
# Will raise NameError since the communicate() call did not complete
except NameError as e:
print('NameError: %s' % e)
res = await p.communicate()
print(res[0].decode('utf-8'))
# Only prints the later half of ping's STDOUT
if __name__ == '__main__':
loop = asyncio.ProactorEventLoop()
asyncio.set_event_loop(loop)
# For Linux: just do loop = asyncio.get_event_loop()
loop.run_until_complete(communicate_short(loop))
请注意,仅打印最后2个数据包。关于前2个数据包的输出会丢失。
那么,如何在超时发生之前从子进程获取输出?
编辑:更确切地说,理想情况下我正在寻找的东西:
NameError: local variable 'res' referenced before assignment
Reply from 127.0.0.1: bytes=32 time<1ms TTL=128
Reply from 127.0.0.1: bytes=32 time<1ms TTL=128
Ping statistics for 127.0.0.1:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 0ms, Maximum = 0ms, Average = 0ms
做什么,即异步写入子进程的STDIN并读取其STDOUT和STDERR,而不会出现死锁(that the docs ominously warn about);
具有可配置的总超时,因此当子进程终止或达到超时时,返回到目前为止收到的STDOUT和STDERR。
看起来这样的东西还不存在,而且必须实现它。
答案 0 :(得分:1)
答案 1 :(得分:0)
对于问题的第二部分,“如何在超时发生之前从子进程获取输出?”我建议使用不取消任务的asyncio.wait()
( p.communicate()
)代替asyncio.wait_for()
(取消任务):
task = asyncio.Task(p.communicate())
done, pending = await asyncio.wait([task], timeout=2)
if pending:
print("timeout!", task._state)
res = await task # Note: It is OK to await a task more than once
print(res[0].decode())
关于“检索部分结果”,我建议不要使用调用communicate()
的{{1}}并使用其他方法:
stdout.read()
组合这两个解决方案(并使用import asyncio
async def ping(loop, host):
p = await asyncio.create_subprocess_exec(
'ping', host, '-c', '4',
stdout=asyncio.subprocess.PIPE, loop=loop)
async for line in p.stdout:
print(host, "==>", line.decode(), end="")
print(host, "done")
if __name__ == '__main__':
loop = loop = asyncio.get_event_loop()
asyncio.set_event_loop(loop)
tasks = [
ping(loop, '8.8.8.8'),
ping(loop, '127.0.0.1'),
ping(loop, 'example.com'),
]
loop.run_until_complete(asyncio.wait(tasks))
loop.close()
而不是更冷的异步)给出:
readline()
注意超时(1秒)是每行。
答案 2 :(得分:0)
如果在超时后杀死子进程,则可以获得如下的部分输出:
future = asyncio.ensure_future(p.communicate())
done, pending = await asyncio.wait([future], timeout=2)
if pending:
# timeout
if p.returncode is None:
# kill the subprocess, then `await future` will return soon
try:
p.kill()
except ProcessLookupError:
pass
output, err = await future
print(output.decode('utf-8'))