1
0
mirror of https://github.com/django/django.git synced 2024-12-23 09:36:06 +00:00
django/tests/builtin_server/tests.py
Nick Pope b47f2f5b90 Fixed #33865 -- Optimized LimitedStream wrapper.
The current implementation of LimitedStream is slow because .read()
performs an extra copy into a buffer and .readline() performs two
extra copies. The stream being wrapped is already typically a BytesIO
object so this is unnecessary.

This implementation has largely been untouched for 12 years and,
inspired by a simpler implementation in werkzeug, it was possible to
achieve the following performance improvement:

LimitedStream.read() (single line):
  Mean +- std dev: [bench_limitedstream-main] 286 ns +- 6 ns
  -> [bench_limitedstream-patch] 227 ns +- 6 ns: 1.26x faster
LimitedStream.readline() (single line):
  Mean +- std dev: [bench_limitedstream-main] 507 ns +- 11 ns
  -> [bench_limitedstream-patch] 232 ns +- 8 ns: 2.18x faster
LimitedStream.read(8192) (single line):
  Mean +- std dev: [bench_limitedstream-main] 360 ns +- 8 ns
  -> [bench_limitedstream-patch] 297 ns +- 6 ns: 1.21x faster
LimitedStream.readline(8192) (single line):
  Mean +- std dev: [bench_limitedstream-main] 602 ns +- 10 ns
  -> [bench_limitedstream-patch] 305 ns +- 10 ns: 1.98x faster
LimitedStream.read() (multiple lines):
  Mean +- std dev: [bench_limitedstream-main] 290 ns +- 5 ns
  -> [bench_limitedstream-patch] 236 ns +- 6 ns: 1.23x faster
LimitedStream.readline() (multiple lines):
  Mean +- std dev: [bench_limitedstream-main] 517 ns +- 19 ns
  -> [bench_limitedstream-patch] 239 ns +- 7 ns: 2.16x faster
LimitedStream.read(8192) (multiple lines):
  Mean +- std dev: [bench_limitedstream-main] 363 ns +- 8 ns
  -> [bench_limitedstream-patch] 311 ns +- 11 ns: 1.17x faster
LimitedStream.readline(8192) (multiple lines):
  Mean +- std dev: [bench_limitedstream-main] 601 ns +- 12 ns
  -> [bench_limitedstream-patch] 308 ns +- 7 ns: 1.95x faster

Geometric mean: 1.59x faster
2023-01-05 19:26:56 +01:00

171 lines
6.1 KiB
Python

import sys
import traceback
from io import BytesIO
from unittest import TestCase, mock
from wsgiref import simple_server
from django.core.servers.basehttp import get_internal_wsgi_application
from django.core.signals import request_finished
from django.test import RequestFactory, override_settings
from .views import FILE_RESPONSE_HOLDER
# 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 in #5596#comment:4.
MAX_SOCKET_CHUNK_SIZE = 32 * 1024 * 1024 # 32 MB
class ServerHandler(simple_server.ServerHandler):
error_status = "500 INTERNAL SERVER ERROR"
def write(self, data):
"""'write()' callable as specified by PEP 3333"""
assert isinstance(data, bytes), "write() argument must be bytestring"
if not self.status:
raise AssertionError("write() before start_response()")
elif not self.headers_sent:
# Before the first output, send the stored headers
self.bytes_sent = len(data) # make sure we know content-length
self.send_headers()
else:
self.bytes_sent += len(data)
# XXX check Content-Length and truncate if too many bytes written?
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):
super().error_output(environ, start_response)
return ["\n".join(traceback.format_exception(*sys.exc_info()))]
class DummyHandler:
def log_request(self, *args, **kwargs):
pass
class FileWrapperHandler(ServerHandler):
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.request_handler = DummyHandler()
self._used_sendfile = False
def sendfile(self):
self._used_sendfile = True
return True
def wsgi_app(environ, start_response):
start_response("200 OK", [("Content-Type", "text/plain")])
return [b"Hello World!"]
def wsgi_app_file_wrapper(environ, start_response):
start_response("200 OK", [("Content-Type", "text/plain")])
return environ["wsgi.file_wrapper"](BytesIO(b"foo"))
class WSGIFileWrapperTests(TestCase):
"""
The wsgi.file_wrapper works for the builtin 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):
env = {"SERVER_PROTOCOL": "HTTP/1.0"}
handler = FileWrapperHandler(BytesIO(), BytesIO(), BytesIO(), env)
handler.run(wsgi_app_file_wrapper)
self.assertTrue(handler._used_sendfile)
self.assertEqual(handler.stdout.getvalue(), b"")
self.assertEqual(handler.stderr.getvalue(), b"")
def test_file_wrapper_no_sendfile(self):
env = {"SERVER_PROTOCOL": "HTTP/1.0"}
handler = FileWrapperHandler(BytesIO(), BytesIO(), BytesIO(), env)
handler.run(wsgi_app)
self.assertFalse(handler._used_sendfile)
self.assertEqual(handler.stdout.getvalue().splitlines()[-1], b"Hello World!")
self.assertEqual(handler.stderr.getvalue(), b"")
@override_settings(ROOT_URLCONF="builtin_server.urls")
def test_file_response_closing(self):
"""
View returning a FileResponse properly closes the file and http
response when file_wrapper is used.
"""
env = RequestFactory().get("/fileresponse/").environ
handler = FileWrapperHandler(BytesIO(), BytesIO(), BytesIO(), env)
handler.run(get_internal_wsgi_application())
# Sendfile is used only when file_wrapper has been used.
self.assertTrue(handler._used_sendfile)
# Fetch the original response object.
self.assertIn("response", FILE_RESPONSE_HOLDER)
response = FILE_RESPONSE_HOLDER["response"]
# The response and file buffers are closed.
self.assertIs(response.closed, True)
buf1, buf2 = FILE_RESPONSE_HOLDER["buffers"]
self.assertIs(buf1.closed, True)
self.assertIs(buf2.closed, True)
FILE_RESPONSE_HOLDER.clear()
@override_settings(ROOT_URLCONF="builtin_server.urls")
def test_file_response_call_request_finished(self):
env = RequestFactory().get("/fileresponse/").environ
handler = FileWrapperHandler(BytesIO(), BytesIO(), BytesIO(), env)
with mock.MagicMock() as signal_handler:
request_finished.connect(signal_handler)
handler.run(get_internal_wsgi_application())
self.assertEqual(signal_handler.call_count, 1)
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().__init__(*args, **kwargs)
self.request_handler = DummyHandler()
self.headers_written = False
self.write_chunk_counter = 0
def send_headers(self):
super().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("200 OK", [("Content-Type", "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):
"""
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)