Broken POST in httplib

Question

I'm having trouble with a POST using httplib. Here is the code:

import base64
import urllib
import httplib
http = urllib3.PoolManager()
head = {"Authorization":"Basic %s" % base64.encodestring("foo:bar")}
fields = {"token":"088cfe772ce0b7760186fe4762843a11"}
conn = httplib.HTTPSConnection("foundation.iplantc.org")
conn.set_debuglevel(2)
conn.request('POST', '/auth-v1/renew', urllib.urlencode(fields), head)
print conn.getresponse().read()
conn.close()

The POST that comes out is correct. I know I started a telnet session and typing it in worked fine. Here it is:

'POST /auth-v1/renew HTTP/1.1\r\nHost: foundation.iplantc.org\r\nAccept-Encoding: identity\r\nContent-Length: 38\r\nAuthorization: Basic YXRlcnJlbDpvTnl12aesf==\n\r\n\r\ntoken=088cfe772ce0b7760186fe4762843a11'

But the response from the server is "token not found" when the python script sends it. BTW this does work fine with urllib3 (urllib2 shows the same error), which uses an multipart encoding, but I want to know what is going wrong with the above. I would rather not depend on yet another 3rd party package.

Answer 1

httplib doesn't automatically add a Content-Type header, you have to add it yourself. (urllib2 automatically adds application/x-www-form-urlencoded as Content-Type).

But what's probably throwing the server off is the additional '\n' after your authorization header, introduced by base64.encodestring. You should better use base64.urlsafe_b64encode instead.

• aterrel: I did play around with the Content-Type but was misspelling it as Content-type... The base64.urlsafe_b64encode didn't change anything.