Python subprocess自定义返回码问题：无法获取预期错误返回码

Let's break down why you're getting a return code of 0 instead of the expected 2, and how to fix it quickly.

The Root Problem #

Your test.py script calls main() and gets a return value of 2, but Python doesn't automatically use that return value as the script's exit code. By default, Python scripts exit with code 0 unless you explicitly tell it otherwise. The return statement in your main() function just passes a value back to the caller (the if __name__ == '__main__': block), but that block doesn't do anything with it—so the script exits with the default 0.

The Fix: Explicitly Set the Exit Code #

You need to import the sys module and use sys.exit() to pass the main() return value as the script's exit code. Here's the updated test.py:

import sys

RETCODE_OK = 0
RETCODE_THING_1_FAILED = 1
RETCODE_THING_2_FAILED = 2

def main():
    return RETCODE_THING_2_FAILED

if __name__ == '__main__':
    sys.exit(main())

Verify the Fix #

Now when you run your subprocess code, you'll get the expected return code:

import subprocess
proc = subprocess.Popen('python test.py', shell=True)
proc.communicate()
print(proc.returncode)  # Outputs 2, as desired

A Safer Alternative to shell=True #

Using shell=True can introduce security risks (like shell injection) if your command includes untrusted input. A better practice is to pass the command as a list of arguments:

proc = subprocess.Popen(['python', 'test.py'])

This works the same way but avoids unnecessary shell overhead and security pitfalls.

Key Takeaways #

• Python scripts don't use function return values as exit codes by default—you need sys.exit() for that.
• Always make sure your script explicitly sets its exit code if you're relying on it for subprocess error handling.
• Prefer list-based command arguments over shell=True for safer, more predictable behavior.

内容的提问来源于stack exchange，提问作者Ben Caine