如何在pytest同个测试用例中结合使用patch和capsys？

1. 报错根因 #

@patch装饰器生成的mock对象会作为位置参数注入测试函数，所有注入的mock参数必须排在pytest内置fixture（比如capsys）的后面，否则pytest会把靠前的mock参数识别为不存在的fixture，触发你遇到的报错。

2. 正确的测试写法 #

首先修正参数顺序，同时注意print函数输出默认会携带换行符\n，断言输出时需要匹配该字符：

from unittest.mock import patch

# 失败场景测试
@patch('test_file.P_open', return_value=1)
def test_mark_config_deployed_failed(capsys, mock_popen):
    expected_output = "Error, config could not be marked\nSuccess, config was marked\n"
    device = 'test-device'
    mark_config(device)
    out, err = capsys.readouterr()
    assert out == expected_output
    # 可选：验证mock被正确调用
    mock_popen.assert_called_once_with(['command','device','args'])

# 成功场景测试
@patch('test_file.P_open', return_value=0)
def test_mark_config_deployed_success(capsys, mock_popen):
    expected_output = "Success, config was marked\n"
    device = 'test-device'
    mark_config(device)
    out, err = capsys.readouterr()
    assert out == expected_output

3. 额外注意事项 #

你当前的mark_config函数存在逻辑缺陷：无论P_open返回成功还是失败，都会打印Success, config was marked，需要补充else分支调整逻辑：

def mark_config(device):
    args = ['command','device','args']
    result = P_open(args)
    if result != 0:
        print('Error, config could not be marked')
    else: # 补充else分支避免成功、失败场景都输出成功日志
        print('Success, config was marked')

修正逻辑后，失败场景的断言输出可以去掉对应的成功日志内容。

内容的提问来源于stack exchange，提问作者Hugo Reyes