Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.c++ > #7831 > unrolled thread

how to create an array of objects initializing each object?

Back to article view | Back to comp.lang.c++

Contents

  how to create an array of objects initializing each object? vaysagekv <kvvaysage@gmail.com> - 2011-07-15 17:42 +0530
    Re: how to create an array of objects initializing each object? Victor Bazarov <v.bazarov@comcast.invalid> - 2011-07-15 08:21 -0400
      Re: how to create an array of objects initializing each object? Paul N <gw7rib@aol.com> - 2011-07-16 10:58 -0700
        Re: how to create an array of objects initializing each object? Victor Bazarov <v.bazarov@comcast.invalid> - 2011-07-16 16:13 -0400
      Re: how to create an array of objects initializing each object? Virchanza <virtual@lavabit.com> - 2011-07-17 14:50 -0700
    Re: how to create an array of objects initializing each object? Yakov Gerlovin <yakov.gerlovin@gmail.com> - 2011-07-18 04:24 -0700
      Re: how to create an array of objects initializing each object? Victor Bazarov <v.bazarov@comcast.invalid> - 2011-07-18 08:59 -0400
        Re: how to create an array of objects initializing each object? Yakov Gerlovin <yakov.gerlovin@gmail.com> - 2011-07-18 08:17 -0700
          Re: how to create an array of objects initializing each object? Yakov Gerlovin <yakov.gerlovin@gmail.com> - 2011-07-18 08:51 -0700
          Re: how to create an array of objects initializing each object? Victor Bazarov <v.bazarov@comcast.invalid> - 2011-07-18 13:34 -0400
            Re: how to create an array of objects initializing each object? Leigh Johnston <leigh@i42.co.uk> - 2011-07-18 19:00 +0100
              Re: how to create an array of objects initializing each object? Yakov Gerlovin <yakov.gerlovin@gmail.com> - 2011-07-19 06:56 -0700
                Re: how to create an array of objects initializing each object? Victor Bazarov <v.bazarov@comcast.invalid> - 2011-07-19 11:08 -0400

#7831 — how to create an array of objects initializing each object?

hi,
    how to create an array of objects initializing each object?
thanks in advance

[toc] | [next] | [standalone]

#7834

On 7/15/2011 8:12 AM, vaysagekv wrote:
> how to create an array of objects initializing each object?

You can initialize them if you provide a brace-enclosed initialiser list 
for the array.  You can't do it with a dynamic array, however.

V
-- 
I do not respond to top-posted replies, please don't ask

[toc] | [prev] | [next] | [standalone]

#7883

On Jul 15, 1:21 pm, Victor Bazarov <v.baza...@comcast.invalid> wrote:
> On 7/15/2011 8:12 AM, vaysagekv wrote:
>
> > how to create an array of objects initializing each object?

See FAQ 10.5.

> You can initialize them if you provide a brace-enclosed initialiser list
> for the array.  You can't do it with a dynamic array, however.

Well, if you use a dynamic array everything does get initialised with
the default constructor. Which may be what the OP wanted. If you want
the elements of the array to be different, you can do this by mesing
about with a static member or a global.

[toc] | [prev] | [next] | [standalone]

#7884

On 7/16/2011 1:58 PM, Paul N wrote:
> On Jul 15, 1:21 pm, Victor Bazarov<v.baza...@comcast.invalid>  wrote:
>> On 7/15/2011 8:12 AM, vaysagekv wrote:
>>
>>> how to create an array of objects initializing each object?
>
> See FAQ 10.5.
>
>> You can initialize them if you provide a brace-enclosed initialiser list
>> for the array.  You can't do it with a dynamic array, however.
>
> Well, if you use a dynamic array everything does get initialised with
> the default constructor.

Actually, that's not entirely accurate.  If there is the initializer in 
the form "()", it's default-initialized, if there isn't, an array of 
built-in types will go *uninitialized*.

 > Which may be what the OP wanted. If you want
> the elements of the array to be different, you can do this by mesing
> about with a static member or a global.

V
-- 
I do not respond to top-posted replies, please don't ask

[toc] | [prev] | [next] | [standalone]

#7928

On Jul 15, 1:21 pm, Victor Bazarov <v.baza...@comcast.invalid> wrote:

> You can initialize them if you provide a brace-enclosed initialiser list
> for the array.  You can't do it with a dynamic array, however.


If you want to go to extremes:

    char unsigned *const cu_pointer = new char
unsigned[ sizeof(MyClass) * N ];

    MyClass *const p = reinterpret_cast<MyClass*>(cu_pointer);

    ::new( p ) MyClass(1,2,3);
    ::new( p+1 ) MyClass(4,5,6);
    ::new( p+2 ) MyClass(7,8,9);

    // and then at the end:

    p[2].~MyClass();
    p[1].~MyClass();
    p->~MyClass();

    delete [] cu_pointer;

[toc] | [prev] | [next] | [standalone]

#7931

*If you have tr1*

#include <array>
#include <iostream>

using namespace std;

struct MyClass
{
   MyClass(int i){
      cout << "MyClass(int " << i << ")" << endl;
   }
   MyClass(const char* s) {
      cout << "MyClass(const char* " << s << ")" << endl;
   }
   MyClass(int i,int j, int k)	{
      cout << "MyClass(this=" << this << " i=" << i << " j=" << j << "
k="<< k << ")" << endl;
   }
};

void main()
{
   array<MyClass,4> arr = { 1, 2, "str", MyClass(1,2,3) };
   cout << "Address of elem at pos 3 is " << &(arr[3]) << endl;
}

[toc] | [prev] | [next] | [standalone]

#7934

On 7/18/2011 7:24 AM, Yakov Gerlovin wrote:
> *If you have tr1*
>
> #include<array>
> #include<iostream>
>
> using namespace std;
>
> struct MyClass
> {
>     MyClass(int i){
>        cout<<  "MyClass(int "<<  i<<  ")"<<  endl;
>     }
>     MyClass(const char* s) {
>        cout<<  "MyClass(const char* "<<  s<<  ")"<<  endl;
>     }
>     MyClass(int i,int j, int k)	{
>        cout<<  "MyClass(this="<<  this<<  " i="<<  i<<  " j="<<  j<<  "
> k="<<  k<<  ")"<<  endl;
>     }
> };
>
> void main()

int main()

> {
>     array<MyClass,4>  arr = { 1, 2, "str", MyClass(1,2,3) };
>     cout<<  "Address of elem at pos 3 is "<<  &(arr[3])<<  endl;
> }

And... isn't this still copy-initialization (and not direct 
initialization) of the array elements?

V
-- 
I do not respond to top-posted replies, please don't ask

[toc] | [prev] | [next] | [standalone]

#7935

> And... isn't this still copy-initialization (and not direct
> initialization) of the array elements?

I'm not sure what exactly do you mean by "direct initialization".
If it is compile time initialization, then naturally it is not direct
initialization.

If it means, that the object are created in-place, without creation of
temp objects and then copy construction, then it is a direct
initialization. At least with g++ 4.3.2. ( I'm not sure whether
standard requires a temp objects in this case or it allows in-place
object creation)

[toc] | [prev] | [next] | [standalone]

#7937

For PODs it is possible to initialize at compile time:

struct POD
{
   int i;
   const char* s;
};

array<POD,2> arr2 = {  1, "str1", 2,"str2" };

(like Victor said in his first post)

[toc] | [prev] | [next] | [standalone]

#7940

On 7/18/2011 11:17 AM, Yakov Gerlovin wrote:
>> And... isn't this still copy-initialization (and not direct
>> initialization) of the array elements?
>
> I'm not sure what exactly do you mean by "direct initialization".
> If it is compile time initialization, then naturally it is not direct
> initialization.
>
> If it means, that the object are created in-place, without creation of
> temp objects and then copy construction, then it is a direct
> initialization. At least with g++ 4.3.2. ( I'm not sure whether
> standard requires a temp objects in this case or it allows in-place
> object creation)

I mean, if you define a copy-ctor in your class (instead of letting the 
compiler define one), you might see whether it's called at all during 
initialization of the array.

V
-- 
I do not respond to top-posted replies, please don't ask

[toc] | [prev] | [next] | [standalone]

#7942

On 18/07/2011 18:34, Victor Bazarov wrote:
> On 7/18/2011 11:17 AM, Yakov Gerlovin wrote:
>>> And... isn't this still copy-initialization (and not direct
>>> initialization) of the array elements?
>>
>> I'm not sure what exactly do you mean by "direct initialization".
>> If it is compile time initialization, then naturally it is not direct
>> initialization.
>>
>> If it means, that the object are created in-place, without creation of
>> temp objects and then copy construction, then it is a direct
>> initialization. At least with g++ 4.3.2. ( I'm not sure whether
>> standard requires a temp objects in this case or it allows in-place
>> object creation)
>
> I mean, if you define a copy-ctor in your class (instead of letting the
> compiler define one), you might see whether it's called at all during
> initialization of the array.

Well you can do that if you make the copy ctor private and see if you 
get a compilation error message (you should as it is 
copy-initialization).  Just adding a non-private copy ctor does not 
prove anything though as an implementation is free to elide the copy as 
an optimization.

/Leigh

[toc] | [prev] | [next] | [standalone]

#7970

Leigh is right, g++ requires a public copy constructor, but does not
actually call it

I made a copy constructor with some prints and it had not been called.
Making it private gives a compiler error.

The question is how can optimizer elide function with side effects
(prints)?

[toc] | [prev] | [next] | [standalone]

#7971

On 7/19/2011 9:56 AM, Yakov Gerlovin wrote:
> Leigh is right, g++ requires a public copy constructor, but does not
> actually call it
>
> I made a copy constructor with some prints and it had not been called.
> Making it private gives a compiler error.
>
> The question is how can optimizer elide function with side effects
> (prints)?

Elision of calling the copy constructor is explicitly allowed by the 
Standard regardless of its side effects.

V
-- 
I do not respond to top-posted replies, please don't ask

[toc] | [prev] | [standalone]

Back to top | Article view | comp.lang.c++

csiph-web