正确使用threading.RLock

问题描述

我正在用Python创建一个实用程序，该实用程序在启动时会在单独线程上从文件中读取数据，以便可以加载其余的GUI组件.数据被存储到list中，然后附加到组合框.我将如何锁定list，以便其他任何方法都无法在def read_employees(self, read_file):方法正在使用列表的同时调用该列表.

I'm creating a utility in Python that reads data in from a file on startup on a separate thread so that the rest of the GUI components can load. The data gets stored into a list and then appended to a combobox. How would I lock the list so that no other method can call the list at the same time it's being used by the def read_employees(self, read_file): method.

这是我能想到的最好的尝试.

This is the best attempt I can come up with.

#left out imports

class MyDialog(wx.Frame):

    def __init__(self, parent, title):
        self.no_resize = wx.DEFAULT_FRAME_STYLE & ~ (wx.RESIZE_BORDER | wx.MAXIMIZE_BOX)
        wx.Frame.__init__(self, parent, title=title, size=(500, 450),style = self.no_resize)

        self.lock = threading.RLock()
        self.empList = []


    def read_employees(self, read_file):

        with open(read_file) as f_obj:
            employees = json.load(f_obj)

        with self.lock:
            self.empList = [empEmail for empEmail in employees.keys()]
            wx.CallAfter(self.emp_selection.Append, self.empList)


    def start_read_thread(self):
        filename = 'employee.json'
        with concurrent.futures.ThreadPoolExecutor(max_workers=1) as executor:
            executor.submit(self.read_employees, filename)

app = wx.App(False)
frame = MyDialog(None, "Crystal Rose")
app.MainLoop()

在这里使用RLock是否合适?

推荐答案

我不知道您在应用程序中还有什么其他工作，但是我建议您看一下wx.CallAfter函数.它是线程安全的，可用于发送消息或发布事件.

I don't know what else you have going on in the app, but I'd recommend taking a look at the wx.CallAfter function. It is thread-safe and can be used to send messages or post events.

import wx
from wx.lib.pubsub import Publisher
import json
from threading import Thread


def update_employee_list(read_file):
    with open(read_file) as f_obj:
        employee_list = json.load(f_obj) # this line should release the GIL so it continues other threads
    # next line sends a thread-safe message to the main event thread
    wx.CallAfter(Publisher().sendMesage, 'updateEmployeeList', employee_list)


class MyDialog(wx.Frame):
    def __init__(self, parent, title):
        self.no_resize = wx.DEFAULT_FRAME_STYLE & ~ (wx.RESIZE_BORDER | wx.MAXIMIZE_BOX)
        wx.Frame.__init__(self, parent, title=title, size=(500, 450),style = self.no_resize)
        self.empList = []
        # subscribe our function to be called when 'updateEmployeeList' messages are received
        Publisher().subscribe(self.updateDisplay, 'updateEmployeeList')

    def updateDisplay(self, employee_list):
        # this assignment should be atomic and thread-safe
        self.empList = employee_list
        # wxPython GUI runs in a single thread, so this is a blocking call
        # if you have many many list items, you may want to modify this method
        # to add one employee at a time to the list to keep it non-blocking.
        self.emp_selection.Append(employee_list)

    def start_read_thread(self):
        filename = 'employee.json'
        t = Thread(target= update_employee_list, args=(filename, ))
        t.start()  # this starts the thread and immediately continues this thread's execution

更新:

使用with ThreadPoolExecutor块是因为代码等效于:

Using a with ThreadPoolExecutor blocks because the code is equivalent to:

executor = ThreadPoolExecutor(max_workers=1)
executor.submit(worker_func, args)
executor.shutdown(wait=True)  # <--- wait=True causes Executor to block until all threads complete

您仍然可以按以下方式使用ThreadPoolExecutor，而无需使用with块.因为你只是:

You could still use the ThreadPoolExecutor as follows, without the with block. Because you're only :

executor = ThreadPoolExecutor(max_workers=1)
executor.submit(worker_func, args)
executor.shutdown(wait=False)  # <--- threads will still complete, but execution of this thread continues immediately

有关并发期货和执行人的更多信息，请查看此处的记录文件.

For more about concurrent futures and Executors, see here for documention.

这篇关于正确使用threading.RLock的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！