waitress: python-waitress: request processing race condition in HTTP pipelining with invalid first request

A flaw was found in the Waitress WSGI server for Python. A remote client can send a request that is exactly recvbytes, which defaults to 8192 long, followed by a secondary request using HTTP pipelining. When request lookahead is disabled default, Waitress won't read any more requests, and when th...