The code below is seen at this SO.

std::weak_ptr<int> g_w;

void f3()
{
    std::shared_ptr<int>l_s3 = g_w.lock(); //2. here will read g_w
    if (l_s3)
    {
        ;/.....
    }
}

void f4()
{
    std::shared_ptr<int> p_s = std::make_shared<int>(1);
    g_w = p_s;

    std::thread th(f3);
    th.detach();
    // 1. p_s destory will motify g_w (write g_w)
}

Although, the same object of weak_ptr(i.e. g_w) is read and written by two different threads, updating g_w in f4() is before the constructation of thread object (i.e. std::thread th(f3);) in which reading g_w via g_w.lock() is invoked. The two operations could not be run at the same time. So I think the code snippet is correct and no data race would occur.

I have read the comments of the said SO again and again but I really can’t understand, say.

Updating g_w in f4() and at the same time reading g_w via g_w.lock() in f3() is undefined behaviour!