SSDP suffers from multibyte string decode problem

[alanmimms]

It seems there are several other problems with multibyte string decoding and encoding in this package. Let me add my $0.0000000000002 😃.

Some device in my network has an SSDP response that blows up ssdp.py discovery sock.recv().decode("utf-8") with UnicodeDecodeError: 'utf-8' codec can't decode byte 0x90 in position 5: invalid start byte. Since I cannot control all devices and since this is really a robustness issue, I tried replacing the bad UTF-8 with the default replacement by changing the code to

                    response = sock.recv(1024).decode("utf-8", "replace")

This solves my problem by allowing the discovery to continue even after a response from such an evil device is found. It seems harmless to do this.

[nielstron]

Does the current policy of just excepting a UnicodeDecodeError and ignoring the whole device solve your problem too?

[alanmimms]

Sorry I have no information on "current policy" since my solution worked for me and I have been using it. Did someone change code to implement the policy you mention? If so I could probably update and answer your question.

[nielstron]

Yes, currently any decode error are just excepted and the whole string discarded. If your solution works though, maybe someone else will emerge with a similar problem.