On Feb 25, 2004, at 12:27 PM, voodoo4 wrote:
I downloaded yesterday the boost libraries (mainly for the memory libs) and to tell the truth i'm impressed with pool and smart pointers. Especially the boost smart pointers will save me from months of coding! However i have a question about smart pointers.
Let's say i have these objects(the code is just an example):
class Slave { Slave(); ~Slave(); };
class Master { Master(); Master(int size); ~Master();
Slave *sl; };
I want "sl" to be an array of Slaves on the free store but the sized will be determined in the constructor.
With normal pointers i would write:
Master::Master(int val) { sl=new Slave[val]; }
Master::~Master() { delete sl; }
How can i achieve the same with boost smart pointers?
I have no idea (see below).
I've read the manual three or four times but i still don't get how to work this out. Does this have something to do with the empty ptr? I believe solution is in front of my eyes but i'm too blind(as usual) to see it...
I don't think you're using the right tool for the job. From what I know, smart pointers should be used when you have one object, here you (potentially) have many. I think what you really want to do here is to use a container class like stl::list or stl::vector to contain your slave objects. If you still want to use smart pointers, I guess you could use a list of smart pointers, but there's really no point for that if your slave class is properly coded (for instance, has a ctor, copy ctor, virtual dtor, and operator=). --MP