你不用问了,我都回复了,要是想认证,请下载编译器

穆扬,

when the place is allocated, then you can use it, otherwise you can not use it.

for example:

int * a = new int[10];

after last statement, you can use a[0] to a[9], but you can not use a[10], why? because, this place is not allocated, so it is not allowed.

now let's talk about this write syntax (new int[10]) +1, such write syntax is ugly, make something confuse, but it is valid, let us look at a democode:

#include <iostream>
using namespace std;

int main()
{
int * a = (new int[5]) + 1;


for(int i = 0; i<5; i++)
{
*(a+i-1) = i;
cout<<a[i-1]<<" ";
}
return 0;
}

穆扬,

I think direct here answer is better than private mail, because everyone here can also take a look,

When you have any question, just ask, I will do my try to answer you.

now to your question above,

int *pi;
double *pd;
……
pi=pd；
在c++的编译器里是非法的还是仅仅给个警告

my answer is: it is not valid, that mean, you will get an error in compile time. The rease for that is, they have not same type. pi is from type pointer integer, but pd is from type pointer double, so pi = pd is error. you should use type cast, such as pi = (int *)pd; then it is ok.

穆扬,

you are welcome, anytime you can ask me question. I would like to answer you so long as I can.

Till now, we have just spoken about how to use new. Now I have a question for you, I can of course answer you, but I hope you can find the answer by yourself and tell me in some day later. The question is: why in c++ will new in stead of malloc used? or in other words, what is the advantage of using new? Think it over.

It is the time for you to learn c++ You need a good book, my suggestion for you is C++ Primer Plus Fifth Edition

Here I want tell you, it is very difficult to learn C++ for C programmer. Why difficult? because the mode of thinkting is different. I took one year to change my thoughmode from c to c++. And after learning java, I began to understood what is OOP. You will learn programming technique, but it is not all, the most import is thought about struct, about model, about object.

Now it is not the time to tell you too more, it is the time for you to start your way.

some declaretion about the democode:

#include <iostream>

using namespace std;


int main()
{
int * a = (new int[5]) + 1; // usually we write code like this int * a = new int[5];
// here new int[5]; is place allocation, for example the first address is 3000
// then (new int[5]) + 1; is address 3004, why 3004 not 3001, because every int
// need 4 byte, so 3000 + 4 = 3004

// with this for loop with this pointer array initialized
for(int i = 0; i<5; i++)
{
*(a+i-1) = i; // because a is not the first address, but the second, so we need minus 1, so that a - 1 is the first address
cout<<a[i-1]<<" ";
}

delete [] a; // it is very import, when you use new, then not forget use delete to frei the allocation.
// when allocation for an array, then delete [] name, otherwise delete name;
return 0;
}

(new int[5]) 开辟的内存会被初始化吗？还是垃圾值？
我感觉您的例子无法说明我那个问题
而且这样是否存在所谓的“指针泄露”问题呢？

To your question, you are right, before return 0; we should insert a statement, delete [] a; otherwise there is space lack problem.