import guv
guv.monkey_patch()
import guv.wsgi
import logger
logger.configure()
def app(environ, start_response):
"""
This is very basic WSGI app useful for testing the performance of guv and guv.wsgi without
the overhead of a framework such as Flask. However, it can just as easily be any other WSGI app
callable object, such as a Flask or Bottle app.
"""
status = '200 OK'
output = [b'Hello World!']
content_length = str(len(b''.join(output)))
response_headers = [('Content-type', 'text/plain'),
('Content-Length', content_length)]
start_response(status, response_headers)
return output
if __name__ == '__main__':
server_sock = guv.listen(('0.0.0.0', 8001))
guv.wsgi.serve(server_sock, app)
used to serve valid HTTP (as far as ``wrk`` is concerned) to benchmark concurrency and requests/sec. Three basic client handlers are provided: - :func:`handle_http_10` acts as an HTTP 1.0 server which sends a static message and closes the connection (HTTP header ``Connection: close``, which is default for HTTP 1.0). - :func:`handle_http_11` acts as an HTTP 1.1 server which sends a static message, but keeps the connection alive (HTTP header ``Connection: keep-alive``, which is default for HTTP 1.1). - :func:`handle_http` is a slightly more complex client handler which actually reads the client's request and decides to either close or keep-alive the connection based on the HTTP version and what the client wants. If the connection is to be kept alive, this handler cooperatively yields control to other greenlets after every request, which significantly improves request/response latency (as reported by wrk). """ import guv guv.monkey_patch() import guv.server import guv.hubs import guv.greenio from guv import gyield from guv.support import PYPY import logging import logger logger.configure() log = logging.getLogger() if PYPY:
def patch(self):
guv.monkey_patch(os=False)
patch_sendfile()
