python / python/cpython

`BaseHTTPRequestHandler` hides `TimeoutError` from `handle_error()`

Open
#99,777 5 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

stdlib type-bug
Dominant language
Python
Stars
77.2k
Forks
36k
PR merge metrics
PR metrics pending

Description

Bug report

If anything in your method handlers (e.g. do_GET()) ends up throwing a TimeoutError, then that unhandled exception is never forwarded to handle_error(), and hence you do not get a chance to do whatever reporting you have set up there.

The cause of this is an undocumented timeout handling that was added in bpo-6267. It blindly assumes that any TimeoutError must come from the request socket and that it was requested and hence expected. That is not the case for any mildly complex request handler that might communicate with other services.

The timeline for this issue is:

?.? to 2.6: No exceptions are hidden from handle_error()
2.7 to 3.2: recv() and send() timeouts are hidden. Other timeouts, e.g. connect(), are not.
3.3 to 3.9: Same situation, but by pure luck since socket.timeout was not converted to TimeoutError yet
3.10: All timeouts are hidden, not just for recv() and send()

Your environment

  • CPython versions tested on: 2.7.18, 3.2.3, 3.5.10, 3.10.7
  • Operating system and architecture: Debian 7 x86_64 and Fedora 35 x86_64

Test case

This test case should give the output:

Got error!
Got error!

But instead gives:

127.0.0.1 - - [25/Nov/2022 13:48:49] Request timed out: TimeoutError(110, 'Connection timed out')
127.0.0.1 - - [25/Nov/2022 13:48:50] Request timed out: TimeoutError('timed out')

For 3.9 or older you get a mix.

Code: (change simulate if you want to have the actual socket code throw the exceptions)

#!/usr/bin/python3

import os
import errno
import socket
import threading
from time import sleep
try:
	from http.server import HTTPServer, BaseHTTPRequestHandler
except ImportError:
	from BaseHTTPServer import HTTPServer, BaseHTTPRequestHandler

simulate = True

class ConnectRequestHandler(BaseHTTPRequestHandler):
	def do_GET(self):
		if simulate:
			raise socket.error(errno.ETIMEDOUT, "Timeout")
		else:
			socket.create_connection(("1.2.3.4", 9000))

class RecvRequestHandler(BaseHTTPRequestHandler):
	def do_GET(self):
		if simulate:
			raise socket.timeout(errno.ETIMEDOUT, "Timeout")
		else:
			(sock, dummy) = socket.socketpair()
			sock.settimeout(1)
			while sock.recv(1024):
				pass

class Server(HTTPServer):
	def handle_error(self, request, client_address):
		print("Got error!")

def run_connect_server():
	s = Server(("127.0.0.1", 8998),
	           ConnectRequestHandler)
	s.handle_request()
	
def run_recv_server():
	s = Server(("127.0.0.1", 8999),
	           RecvRequestHandler)
	s.handle_request()

threading.Thread(target=run_connect_server).start()
threading.Thread(target=run_recv_server).start()
sleep(0.5)

s = socket.create_connection(("127.0.0.1", 8998))
s.send(b'GET / HTTP/1.0\r\n\r\n')
while s.recv(1024):
	pass
s.close()

s = socket.create_connection(("127.0.0.1", 8999))
s.send(b'GET / HTTP/1.0\r\n\r\n')
while s.recv(1024):
	pass
s.close()

sleep(0.5)
Linked PRs
  • gh-99806

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with BaseHTTPRequestHandler and the request-handling path described in the issue, then reproduce the provided test case to observe which TimeoutError exceptions reach handle_error(). Done means handler-raised timeouts are reported through handle_error() without losing the existing request-socket timeout behavior.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
backend, networking
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.