Fixed #18972 -- Refactored bundled wsgi server's chunking algorithm.

Thanks to amosonn at yahoo.com for the report, @doda for the initial patch and
@datagrok for the revamped logic and test case.

django/core/servers/basehttp.py

@@ -9,7 +9,7 @@ been reviewed for security issues. DON'T USE IT FOR PRODUCTION USE!
 
 from __future__ import unicode_literals
 
-import os
+from io import BytesIO
 import socket
 import sys
 import traceback
@@ -26,7 +26,13 @@ from django.core.wsgi import get_wsgi_application
 from django.utils.module_loading import import_by_path
 from django.utils import six
 
-__all__ = ['WSGIServer', 'WSGIRequestHandler']
+__all__ = ('WSGIServer', 'WSGIRequestHandler', 'MAX_SOCKET_CHUNK_SIZE')
+
+
+# If data is too large, socket will choke, so write chunks no larger than 32MB
+# at a time. The rationale behind the 32MB can be found on Django's Trac:
+# https://code.djangoproject.com/ticket/5596#comment:4
+MAX_SOCKET_CHUNK_SIZE = 32 * 1024 * 1024  # 32 MB
 
 
 def get_internal_wsgi_application():
@@ -78,19 +84,9 @@ class ServerHandler(simple_server.ServerHandler, object):
             self.bytes_sent += len(data)
 
         # XXX check Content-Length and truncate if too many bytes written?
-
-        # If data is too large, socket will choke, so write chunks no larger
-        # than 32MB at a time.
-        length = len(data)
-        if length > 33554432:
-            offset = 0
-            while offset < length:
-                chunk_size = min(33554432, length)
-                self._write(data[offset:offset+chunk_size])
-                self._flush()
-                offset += chunk_size
-        else:
-            self._write(data)
+        data = BytesIO(data)
+        for chunk in iter(lambda: data.read(MAX_SOCKET_CHUNK_SIZE), b''):
+            self._write(chunk)
             self._flush()
 
     def error_output(self, environ, start_response):

tests/builtin_server/tests.py

@@ -2,22 +2,18 @@ from __future__ import unicode_literals
 
 from io import BytesIO
 
-from django.core.servers.basehttp import ServerHandler
+from django.core.servers.basehttp import ServerHandler, MAX_SOCKET_CHUNK_SIZE
 from django.utils.unittest import TestCase
 
-#
-# Tests for #9659: wsgi.file_wrapper in the builtin server.
-# We need to mock a couple of handlers and keep track of what
-# gets called when using a couple kinds of WSGI apps.
-#
 
 class DummyHandler(object):
-    def log_request(*args, **kwargs):
+    def log_request(self, *args, **kwargs):
         pass
 
+
 class FileWrapperHandler(ServerHandler):
     def __init__(self, *args, **kwargs):
-        ServerHandler.__init__(self, *args, **kwargs)
+        super(FileWrapperHandler, self).__init__(*args, **kwargs)
         self.request_handler = DummyHandler()
         self._used_sendfile = False
 
@@ -25,17 +21,24 @@ class FileWrapperHandler(ServerHandler):
         self._used_sendfile = True
         return True
 
+
 def wsgi_app(environ, start_response):
     start_response(str('200 OK'), [(str('Content-Type'), str('text/plain'))])
     return [b'Hello World!']
 
+
 def wsgi_app_file_wrapper(environ, start_response):
     start_response(str('200 OK'), [(str('Content-Type'), str('text/plain'))])
     return environ['wsgi.file_wrapper'](BytesIO(b'foo'))
 
+
 class WSGIFileWrapperTests(TestCase):
     """
     Test that the wsgi.file_wrapper works for the builting server.
+
+    Tests for #9659: wsgi.file_wrapper in the builtin server.
+    We need to mock a couple of handlers and keep track of what
+    gets called when using a couple kinds of WSGI apps.
     """
 
     def test_file_wrapper_uses_sendfile(self):
@@ -53,3 +56,47 @@ class WSGIFileWrapperTests(TestCase):
         self.assertFalse(handler._used_sendfile)
         self.assertEqual(handler.stdout.getvalue().splitlines()[-1], b'Hello World!')
         self.assertEqual(handler.stderr.getvalue(), b'')
+
+
+class WriteChunkCounterHandler(ServerHandler):
+    """
+    Server handler that counts the number of chunks written after headers were
+    sent. Used to make sure large response body chunking works properly.
+    """
+
+    def __init__(self, *args, **kwargs):
+        super(WriteChunkCounterHandler, self).__init__(*args, **kwargs)
+        self.request_handler = DummyHandler()
+        self.headers_written = False
+        self.write_chunk_counter = 0
+
+    def send_headers(self):
+        super(WriteChunkCounterHandler, self).send_headers()
+        self.headers_written = True
+
+    def _write(self, data):
+        if self.headers_written:
+            self.write_chunk_counter += 1
+        self.stdout.write(data)
+
+
+def send_big_data_app(environ, start_response):
+    start_response(str('200 OK'), [(str('Content-Type'), str('text/plain'))])
+    # Return a blob of data that is 1.5 times the maximum chunk size.
+    return [b'x' * (MAX_SOCKET_CHUNK_SIZE + MAX_SOCKET_CHUNK_SIZE // 2)]
+
+
+class ServerHandlerChunksProperly(TestCase):
+    """
+    Test that the ServerHandler chunks data properly.
+
+    Tests for #18972: The logic that performs the math to break data into
+    32MB (MAX_SOCKET_CHUNK_SIZE) chunks was flawed, BUT it didn't actually
+    cause any problems.
+    """
+
+    def test_chunked_data(self):
+        env = {'SERVER_PROTOCOL': 'HTTP/1.0'}
+        handler = WriteChunkCounterHandler(None, BytesIO(), BytesIO(), env)
+        handler.run(send_big_data_app)
+        self.assertEqual(handler.write_chunk_counter, 2)