Lack of trailing slash creates invalid GET call
If you use a line like ws = EchoClient('wss://myserver.org:443') you get an invalid GET request:
0000 47 45 54 20 20 48 54 54 50 2f 31 2e 31 0d 0a 48 GET HTTP/1.1..H
Note the lack of a / there between the GET and HTTP/1.1.
Using ws = EchoClient('wss://myserver.org:443/') works as expected.
Lack of a trailing slash when requesting a site's root really ought to be handled properly.
I agree this should be transparent to the client but currently the implementation doesn't make it easy.
Could anyone tell me how to reproduce this? Tracing the traffic by ngrep shows that it is always GET / HTTP/1.1 whenever there is a trailing slash.
On the other hand, self.resource, which is the header element between GET and HTTP 1.1, is / in both case.
Or maybe the bug is due to cURL?