SF bug 874842 and patch 997626: httplib bugs

Hack httplib to work with broken Akamai proxies.
Make sure that httplib doesn't add extract Accept-Encoding or
Content-Length headers if the client has already set them.

diff --git a/Lib/httplib.py b/Lib/httplib.py
index 5133c8dbe51c39e12d57e58f1be1cbf683566838..c0d372ff7eb21a9f0726c59a9a49acdb16bc69a5 100644 (file)
--- a/Lib/httplib.py
+++ b/Lib/httplib.py
@@ -344,6 +344,7 @@ class HTTPResponse:
             self.will_close = 1
 
     def _check_close(self):
+        conn = self.msg.getheader('connection')
         if self.version == 11:
             # An HTTP/1.1 proxy is assumed to stay open unless
             # explicitly closed.
@@ -352,13 +353,18 @@ class HTTPResponse:
                 return True
             return False
 
-        # An HTTP/1.0 response with a Connection header is probably
-        # the result of a confused proxy.  Ignore it.
+        # Some HTTP/1.0 implementations have support for persistent
+        # connections, using rules different than HTTP/1.1.
 
         # For older HTTP, Keep-Alive indiciates persistent connection.
         if self.msg.getheader('keep-alive'):
             return False
 
+        # At least Akamai returns a "Connection: Keep-Alive" header,
+        # which was supposed to be sent by the client.
+        if conn and "keep-alive" in conn.lower():
+            return False
+
         # Proxy-Connection is a netscape hack.
         pconn = self.msg.getheader('proxy-connection')
         if pconn and "keep-alive" in pconn.lower():
@@ -381,6 +387,8 @@ class HTTPResponse:
         #          called, meaning self.isclosed() is meaningful.
         return self.fp is None
 
+    # XXX It would be nice to have readline and __iter__ for this, too.
+
     def read(self, amt=None):
         if self.fp is None:
             return ''
@@ -728,15 +736,17 @@ class HTTPConnection:
             self._send_request(method, url, body, headers)
 
     def _send_request(self, method, url, body, headers):
-        # If headers already contains a host header, then define the
-        # optional skip_host argument to putrequest().  The check is
-        # harder because field names are case insensitive.
-        if 'host' in [k.lower() for k in headers]:
-            self.putrequest(method, url, skip_host=1)
-        else:
-            self.putrequest(method, url)
+        # honour explicitly requested Host: and Accept-Encoding headers
+        header_names = dict.fromkeys([k.lower() for k in headers])
+        skips = {}
+        if 'host' in header_names:
+            skips['skip_host'] = 1
+        if 'accept-encoding' in header_names:
+            skips['skip_accept_encoding'] = 1
 
-        if body:
+        self.putrequest(method, url, **skips)
+
+        if body and ('content-length' not in header_names):
             self.putheader('Content-Length', str(len(body)))
         for hdr, value in headers.iteritems():
             self.putheader(hdr, value)

diff --git a/Lib/test/test_httplib.py b/Lib/test/test_httplib.py
index c57793ded2f14b276c5f32826ac9d2d45770bcbe..5f252bb52f2ceafc8643d78d2b7ba12e7b6f7e6c 100644 (file)
--- a/Lib/test/test_httplib.py
+++ b/Lib/test/test_httplib.py
@@ -2,13 +2,18 @@ import httplib
 import StringIO
 import sys
 
-from test.test_support import verify,verbose
+from unittest import TestCase
+
+from test import test_support
 
 class FakeSocket:
     def __init__(self, text, fileclass=StringIO.StringIO):
         self.text = text
         self.fileclass = fileclass
 
+    def sendall(self, data):
+        self.data = data
+
     def makefile(self, mode, bufsize=None):
         if mode != 'r' and mode != 'rb':
             raise httplib.UnimplementedFileMode()
@@ -32,6 +37,39 @@ class NoEOFStringIO(StringIO.StringIO):
             raise AssertionError('caller tried to read past EOF')
         return data
 
+
+class HeaderTests(TestCase):
+    def test_auto_headers(self):
+        # Some headers are added automatically, but should not be added by
+        # .request() if they are explicitly set.
+
+        import httplib
+
+        class HeaderCountingBuffer(list):
+            def __init__(self):
+                self.count = {}
+            def append(self, item):
+                kv = item.split(':')
+                if len(kv) > 1:
+                    # item is a 'Key: Value' header string
+                    lcKey = kv[0].lower()
+                    self.count.setdefault(lcKey, 0)
+                    self.count[lcKey] += 1
+                list.append(self, item)
+
+        for explicit_header in True, False:
+            for header in 'Content-length', 'Host', 'Accept-encoding':
+                conn = httplib.HTTPConnection('example.com')
+                conn.sock = FakeSocket('blahblahblah')
+                conn._buffer = HeaderCountingBuffer()
+
+                body = 'spamspamspam'
+                headers = {}
+                if explicit_header:
+                    headers[header] = str(len(body))
+                conn.request('POST', '/', body, headers)
+                self.assertEqual(conn._buffer.count[header.lower()], 1)
+
 # Collect output to a buffer so that we don't have to cope with line-ending
 # issues across platforms.  Specifically, the headers will have \r\n pairs
 # and some platforms will strip them from the output file.
@@ -110,4 +148,9 @@ def _test():
         raise AssertionError, "Did not expect response from HEAD request"
     resp.close()
 
+
+def test_main(verbose=None):
+    tests = [HeaderTests,]
+    test_support.run_unittest(*tests)
+
 test()