python os.popen 提示 UnicodeDecodeError: ‘gbk‘ codec can‘t decode byte illegal multibyte sequence

原有代码如下：

def run_silently(cmd: str) -> str:
    """返回系统命令的执行结果"""
    with os.popen(cmd) as fp:
        return fp.read().strip()

在ubuntu里执行，一点问题也没有，去到Windows就报错了

经试验：`print(run_silently('echo 拟好'))` 能正常运行， `print(run_silently('git log'))`则报错如下

$ python change_version.py
Traceback (most recent call last):
  File "change_version.py", line 79, in <module>
    main()
  File "change_version.py", line 61, in main
    print(run_silently('git log'))
  File "change_version.py", line 31, in run_silently
    return fp.read().strip()
UnicodeDecodeError: 'gbk' codec can't decode byte 0x92 in position 592: illegal multibyte sequence

原因：编码问题

解决：修改代码如下

def run_silently(cmd: str) -> str:
    """返回系统命令的执行结果"""
    with os.popen(cmd) as fp:
        bf = fp._stream.buffer.read()
    try:
        return bf.decode().strip()
    except UnicodeDecodeError:
        return bf.decode('gbk').strip()

==== Updated At: 2023-08-22

由于0xab和0xa0这两个字符用gbk和utf8解码时分别报错了，所以修改成如下（同时支持Python2/Python3）：

# coding:utf8
import os


def run_silently(
    cmd,  # type: str
    errors="ignore",  # type: str
):  # type: (...) -> str
    """返回系统命令的执行结果"""
    with os.popen(cmd) as fp:
        if not hasattr(fp, "_stream"):  # Compare with python2
            return fp.read().strip()
        bf = fp._stream.buffer.read().strip()
    try:
        return bf.decode("utf8", errors)
    except UnicodeDecodeError:
        return bf.decode("gbk", errors)


# 测试:
#print(run_silently("echo {}".format(b"\xa0\xab" + u"您好".encode("utf8"))))

注：decode函数的errors参数默认时strict，会严格比对字符集，一般场景下用ignore就可以了