Working with a global singleton in Flask (WSGI), do I have to worry about race conditions?

Question

The hello world demo for Flask is:

from flask import Flask
app = Flask(__name__)

@app.route("/")
def hello():
    return "Hello World!"

if __name__ == "__main__":
    app.run()

What if I modified this like so:

from flask import Flask
app = Flask(__name__)

a = 1
b = 2
c = 3

@app.route("/")
def hello():
    a += 1
    b += a
    c += b
    return "Hello World!"

if __name__ == "__main__":
    app.run()

I understand WSGI application might have multiple threads. The hello function could be running on multiple threads at the same time, and then we'd have a race condition. Is this correct? If the above code is not thread safe, what can I do to make it thread safe?

Avoiding globals is a possible solution, but can you always avoid globals? What if I want something like a python object cache?

Answer 1

You could use a lock:

from threading import Lock
from flask import Flask
app = Flask(__name__)

a = 1
b = 2
c = 3
lock = Lock()

@app.route("/")
def hello():
    with lock:
        a += 1
        b += a
        c += b
    return "Hello World!"

if __name__ == "__main__":
    app.run()

Answer 2

You could try the Local class from werkzeug. Here's some info about it: Context Locals

Example:

from flask import Flask
from werkzeug.local import Local
app = Flask(__name__)
loc = Local()
loc.a = 1
loc.b = 2
loc.c = 3

@app.route("/")
def hello():
    loc.a += 1
    loc.b += loc.a
    loc.c += loc.b
    return "Hello World!"

if __name__ == "__main__":
    app.run()

Answer 3

You might take a look at the g object that you can import directly from flask, keeps an object globally for that request. If you're using an event driven WSGI server (tornado, gevent, etc) you shouldn't have any issues.