Character encoding issue with subprocess module running netsh command

*this question is on windows only. When attempting to retrieve all WiFi passcodes and names stored, I ran into a error running a netsh command due to a mis-encoding of a character through the subprocess module.

Context:

When i run netsh wlan show profiles name=*, I notice that only the first result displays its key content. So, to capture all keys, I had to run the command once for every profile name (that I got through netsh wlan show profiles). The problem is, using subprocess.run for each of these this is incredibly slow, so I decided to use one process through subprocess.Popen, and just write to it's stdin.

Testing code:

import subprocess
process = subprocess.Popen(
    "cmd",  
    stdin=subprocess.PIPE,
    stdout=subprocess.PIPE,
    stderr=subprocess.PIPE,
    text=True,
    encoding='utf-8',
    creationflags=subprocess.CREATE_NO_WINDOW)

process.stdin.write(f'netsh wlan show profiles name="Leonard’s iPhone iP15" key=clear\n')
process.stdin.flush()
process.stdin.write(f'echo test’test\n')
process.stdin.flush()
print(process.communicate())

running it has the result:

C:>netsh wlan show profiles name="Leonard’s iPhone iP15" key=clear
Profile "LeonardΓÇÖs iPhone iP15" is not found on the system.

C:>echo test’test
test’test

as you can see, the ’ has been read as ΓÇÖ in the netsh command result, but correctly in echo and command input area.

This error does not occur when running netsh wlan show profiles name="Leonard’s iPhone iP15" key=clear directly in CMD, as i get the successful result: Profile Leonard’s iPhone iP15 on interface Wi-Fi: ...

How can iI have the command correctly passed to netsh? Is there an easier way to do the same thing?

Additional Details:

I chose encoding='utf-8' because i had been having problems reading the ’ symbol from process.stdout without directly specifying that I wanted to use utf-8. I have tried other encodings such as cp437 (from this post), and cp1252 (from locale.getpreferredencoding) but these led to more encoding/decoding problems than with utf-8.

You face a mojibake case:

print("test’test".encode( 'utf-8').decode( 'cp437'))   ### testΓÇÖtest

This comes from .stdin.write as default cmd code page is OEMCP in Windows; please check it yourself using

REG QUERY "HKLM\SYSTEM\CurrentControlSet\Control\Nls\CodePage" -v *CP|find "REG_SZ"

The following script should work:

import subprocess
cmd_cmds = '&'.join([
        'echo test’test',
        'echo "Leonard’s iPhone iP15"',
        'netsh wlan show profiles name="Leonard’s iPhone iP15"'
    ])

process = subprocess.Popen(
    f'cmd /C {cmd_cmds}',
    # shell = True,
    stdin=subprocess.PIPE,
    stdout=subprocess.PIPE,
    stderr=subprocess.PIPE,
    text=True,
    encoding='utf-8',
    creationflags=subprocess.CREATE_NO_WINDOW,
    process_group=None
    )

for xx in process.communicate():
    print(xx)

Output: D:\Pythons\SO\79317942.py

test’test
"Leonard’s iPhone iP15"
Profile "Leonard’s iPhone iP15" is not found on the system.