Python全局变量未在线程中更新

Question

I am writing a code to use scale to update a global value VOL. The initial value of VOL can be passed to function "check_keys", however, when the scale changes its value, VOL remains unchanged in the thread using function "check_keys". Where was I doing wrong?

#!/usr/bin/python
from Xlib.display import Display
from multiprocessing import Process
import threading
import os
import time
import gtk

VOL=1
RUNNING=0

class StatusIcon:
    def __init__(self):
        global RUNNING
        global VOL

        self.statusicon = gtk.StatusIcon()
        self.statusicon.set_from_stock(gtk.STOCK_HOME)
        #self.statusicon.connect("popup-menu", self.right_click_event)

        self.win_build()

        RUNNING=1
        pp = Process(target=self.check_keys)
        pp.start()      

    def quit_program(self, widget):
        global RUNNING
        RUNNING=0
        gtk.main_quit

    def scale_moved(self, widget):
        global VOL
        #print self
        #print VOL
        VOL = self.scale.get_value()
        #print VOL

    def check_keys(self):
        global VOL
        global RUNNING
        disp = Display()
        hold = 0
        samp_inv=0.02
        while RUNNING:
            print "RUNNING"
            print RUNNING
            print "VOL"
            print VOL
            time.sleep(0.2)
    def win_build(self):
        global VOL

        self.window = gtk.Window()
        self.window.set_default_size(400, -1)

        adjustment = gtk.Adjustment(VOL, 0, 150, 5, 10, 0)
        self.scale = gtk.HScale(adjustment)
        self.scale.set_digits(0)
        self.scale.set_update_policy(gtk.UPDATE_DELAYED)

        #VOL=1.
        self.visible=1
        self.window.connect("destroy", lambda w: w.hide_all())
        self.scale.connect("value-changed", self.scale_moved)

        self.window.add(self.scale)
        self.window.set_title("Typewriter Sound Maker")

        self.window.set_position(gtk.WIN_POS_CENTER_ALWAYS)

        self.window.show_all()

StatusIcon()
gtk.main()

Answer 1

I think the main problem here is: You are using multiprocessing and not threading.

When you use multiprocessing, python will actually fork a new Process, when that happens it can't access the memory of the other process anymore.

You have 2 Options here:

• Select some mechanism to share memory between process, you can find those here: http://docs.python.org/2/library/multiprocessing.html

• or just swap out multiprocessing with threading. http://docs.python.org/2/library/threading.html#module-threading

The down side is that it's not really parallel, because Python is doing the context switch, so one part of your app will be blocked. Anyway, I don't think that concerns you here, because you aren't looking for performance benefits.