Hey,
BadStatusLine triggered when protocol version is in lowercase. 
I've encountered a server that answers "Http/1.0  404 Not Found\r\n" instead of "HTTP/1.0  404 Not Found\r\n"

## Expected Result

Requests understanding the status line.

## Actual Result

Requests is closing the connection.

## Reproduction Steps
### Setup a server that answers the line above
bash: ```while 1;do echo "Http/1.0  404 Not Found\r\n" | sudo nc -lnvp 80; done```
### get the server
```python
import requests
req = req = requests.get("http://127.0.0.1/", verify=False, allow_redir=False )
```

## problem location
Look at line 287 of http/client.py
the word "HTTP" should be matched in lowercase too.
```python
if not version.startswith("HTTP/"):```

Regards.

requests is a third-party library, and this isn't the right place to report issues with it. It looks like the requests issue tracker is at https://github.com/psf/requests/issues

If you can duplicate this problem with only using the python standard library, then please let us know and we'll take a look. I'd have to do some research through the standards to determine if the problem is really with your server, though.

Hi,

Here I'm using requests to show the behavior because it relies on python's http lib, and it is faster/simplier to use. The Exception "BadStatusLine" is a part or the http/client.py library.

As per the RFC2616 section 6.1 https://tools.ietf.org/html/rfc2616#section-6.1, there's nothing specifying that the HTTP verb must be uppercase, I think it's more a matter of "common sense". I might have missed something though!

Here is the poc for this error with http.client only (for the server: use the same nc command as my first message): ```python
import http.client
>>> h1 = http.client.HTTPConnection("127.0.0.1:80")
>>> h1.request("GET", "/")
>>> r1 = h1.getresponse()
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib/python3.8/http/client.py", line 1347, in getresponse
    response.begin()
  File "/usr/lib/python3.8/http/client.py", line 307, in begin
    version, status, reason = self._read_status()
  File "/usr/lib/python3.8/http/client.py", line 289, in _read_status
    raise BadStatusLine(line)
http.client.BadStatusLine: Http/1.0  404 Not Found
```

Thanks for the reproducer and the research!

https://tools.ietf.org/html/rfc2616#section-3.1 says the result header is "HTTP", and doesn't say anything else is acceptable. I'd be interested in what other frameworks (probably in other languages) support. I'll do some research.

https://tools.ietf.org/html/rfc2616#section-3.1 defines HTTP version indicator as

   HTTP-Version   = "HTTP" "/" 1*DIGIT "." 1*DIGIT

so the check

   version.startswith("HTTP/")

is correct.

Hi @christian.heimes,
Thank you for your research too. We've also discovered that this check is correct, but this check is very strict and blocking (error raised, stopping the connection), we should maybe be more "laxist" and allow the  lowercase version ? As they do in the others libs ? I've nerver encountered this error with urllib for instance. 

The server that answered this HTTP response line is a clone of the "spip" framework used in many websites. This is clearly a human behavior, but this http.client error could be a bit more "intelligent" I guess.

urllib is a high level API on top of http.client. It wraps and uses http.client internally. urllib does not accept invalid protocol identifiers either:

>>> urllib.request.urlopen('http://localhost:8080')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib64/python3.8/urllib/request.py", line 222, in urlopen
    return opener.open(url, data, timeout)
  File "/usr/lib64/python3.8/urllib/request.py", line 525, in open
    response = self._open(req, data)
  File "/usr/lib64/python3.8/urllib/request.py", line 542, in _open
    result = self._call_chain(self.handle_open, protocol, protocol +
  File "/usr/lib64/python3.8/urllib/request.py", line 502, in _call_chain
    result = func(*args)
  File "/usr/lib64/python3.8/urllib/request.py", line 1379, in http_open
    return self.do_open(http.client.HTTPConnection, req)
  File "/usr/lib64/python3.8/urllib/request.py", line 1354, in do_open
    r = h.getresponse()
  File "/usr/lib64/python3.8/http/client.py", line 1347, in getresponse
    response.begin()
  File "/usr/lib64/python3.8/http/client.py", line 307, in begin
    version, status, reason = self._read_status()
  File "/usr/lib64/python3.8/http/client.py", line 289, in _read_status
    raise BadStatusLine(line)
http.client.BadStatusLine: Http/1.1 200 OK


curl is more forgiving but does not recognize "Http/1.1" as a valid HTTP/1.1 header either. Instead it assumes that any non-valid header means HTTP/1.0.

$ curl -v http://localhost:8080
*   Trying ::1:8080...
* connect to ::1 port 8080 failed: Connection refused
*   Trying 127.0.0.1:8080...
* Connected to localhost (127.0.0.1) port 8080 (#0)
> GET / HTTP/1.1
> Host: localhost:8080
> User-Agent: curl/7.69.1
> Accept: */*
> 
* HTTP 1.0, assume close after body
< Http/1.1 200 OK
< Server: BaseHTTP/0.6 Python/3.8.6
< Date: Mon, 23 Nov 2020 09:10:38 GMT
< Content-Type: text/plain
< 
* Closing connection 0


I'm against changing the behavior of http.client.

Alright, that was a bad idea.