Question

This is a test application:

#!/usr/bin/env python

from flask import Flask
from time import sleep

application = Flask(__name__)
application.debug = True

@application.route('/a')
@application.route('/b')
@application.route('/c')
def a():
    sleep(10)
    return 'Hello world.'

if __name__ == '__main__':
    application.run()

This application is deployed on Apache:

WSGIDaemonProcess Test processes=2 threads=15 display-name=%{GROUP}
WSGIProcessGroup Test

if you

• request /a at 00:00
• request /b at 00:01
• request /c at 00:02

, you will

• receive response from /a at 00:10
• receive response from /b at 00:11
• receive response from /c at 00:12

But if you

• request /a at 00:00
• request /a at 00:01
• request /a at 00:02

, you will

• receive response from /a at 00:10
• receive response from /a at 00:20
• receive response from /a at 00:30

So I guess that every request for one URL is processed in one thread. Now I intend to develop a long-pulling server, I think I need to dispatch every request to a independent thread to avoid block subsequent requests. What should I do?

Answer 1

This problem has been solved. My guess is incorrect, the truth is:
In Chrome browser, if two requests composed of identical protocol, hostname, port and path, the later request will not send until the prior request close.
The blocked thing was not server but browser.