I always prefer to add the universal_newlines=True argument to the subprocess calls, so they return proper strings. This seems preferable than faffing about with manual decoding and making the assumption that the system encoding is UTF-8. And fewer bytestrings infecting your tests.
I always prefer to add the universal_ newlines= True argument to the subprocess calls, so they return proper strings. This seems preferable than faffing about with manual decoding and making the assumption that the system encoding is UTF-8. And fewer bytestrings infecting your tests.