CODEWITHSUNDEEP
pythonpython-3.xhttp
CforLinux
CforLinux

Reputation: 309

python http client stuck on 100 continue

I have a simple http server in python that implement PUT using 100 continue:

class TestHandler(SimpleHTTPRequestHandler):
    def do_PUT(self):
        length = int(self.headers.get('Content-Length'))
        self.send_response_only(100)
        self.end_headers()
        data = self.rfile.read(length)
        res = manipulate(data)
        new_length = len(res)
        self.send_response(200)
        self.send_header("Content-Length", new_length)
        self.end_headers()
        self.wfile.write(res)

server = HTTPServer(("localhost", 8080), TestHandler)
server.serve_forever()

I try to connect to the server using this client:

def send_put(data):
    c = HTTPConnection('localhost', 8080)
    c.request('PUT', 'http://localhost:8080/', headers={'Content-Length': len(data), 'Expect': '100-continue'})
    r = c.getresponse()
    if 100 != r.status:
        return
    c.request('PUT', 'http://localhost:8080/', body=data)
    r = c.getresponse()
    print(r.read())

but the code always get stuck on the first 'getresponse' even though I can see the 100-continue response on wireshark, what am I doing wrong here? Is python http even support 100-continue?

EDIT: after looking at some of python http code I found why the getresponse is stuck; python's http just ignores the 100-continue and waits for the next response that never comes(from python3.4/http/client.py):

# read until we get a non-100 response
while True:
    version, status, reason = self._read_status()
    if status != CONTINUE:
        break
    # skip the header from the 100 response
    while True:
        skip = self.fp.readline(_MAXLINE + 1)
        if len(skip) > _MAXLINE:
            raise LineTooLong("header line")
        skip = skip.strip()
        if not skip:
            break
        if self.debuglevel > 0:
            print("header:", skip)

Upvotes: 3

Views: 3450

Answers (1)

Thomas
Thomas

Reputation: 181998

I ran into this as well; it's a nine year old Python issue. I came up with the following rather gross "just get it to run" workaround, which seems to work in my case (Python 3.5, HTTPS only):

class ContinueHTTPResponse(http.client.HTTPResponse):
    def _read_status(self, *args, **kwargs):
        version, status, reason = super()._read_status(*args, **kwargs)
        if status == 100:
            status = 199
        return version, status, reason

    def begin(self, *args, **kwargs):
        super().begin(*args, **kwargs)
        if self.status == 199:
            self.status = 100

    def _check_close(self, *args, **kwargs):
        return super()._check_close(*args, **kwargs) and self.status != 100


class ContinueHTTPSConnection(http.client.HTTPSConnection):
    response_class = ContinueHTTPResponse

    def getresponse(self, *args, **kwargs):
        logging.debug('running getresponse')
        response = super().getresponse(*args, **kwargs)
        if response.status == 100:
            setattr(self, '_HTTPConnection__state', http.client._CS_REQ_SENT)
            setattr(self, '_HTTPConnection__response', None)
        return response

I'm using it somewhat like this:

conn = ContinueHTTPSConnection(host)
conn.request(...)
resp = conn.getresponse()

if resp.status == http.client.CONTINUE:
    resp.read()
    conn.send(body)
    resp = conn.getresponse()

# do something with resp if you want...

Caveat: super hacky. Probably full of bugs. Use at your own risk.

Upvotes: 2

Related Questions