Array Variable initialization

toxcct

CPallini wrote:

Am i wrong?

you are, or the compiler is ! the standard says that such a construction (when initializing a variable at the same time than declaration) initializes every element of the array to their default value. I'm even pretty sure Nemanja Trifunovic quoted me somewhere about that...

[VisualCalc][Binary Guide][CommDialogs] | [Forums Guidelines]

David Crow

CPallini wrote:

(it initializes only the first array element).

Not according to the STOSx instructions. Of course, I only use it to initialize POD types to 0. Otherwise, I'd use memset().

"Love people and use things, not love things and use people." - Unknown

"The brick walls are there for a reason...to stop the people who don't want it badly enough." - Randy Pausch

CPallini

toxcct wrote:

you are, or the compiler is !

Maybe the latter: the program:

#include <iostream>
using namespace std;

void main()
{
const int N = 10;
int a[N]={7};

for (int i=0; i<N;i++)
{
cout << a[i] << endl;
}
}

the output:

7
0
0
0
0
0
0
0
0
0

The system: Visual C++ 2008 Express Edition running on Win XP. Eventually YOU may be wrong! ;P (actually I'm quite confident you're right, but don't use it unless you really want surprises!) [added] Actually I think VS2008 cannot be so out-of-the-standard. I suppose the standard establishing that, whenever the initialization list is too short, the remaining array items are default-initialized (to 0 if integers). [/added] :)

If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler. -- Alfonso the Wise, 13th Century King of Castile.
This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong. -- Iain Clarke
[My articles]

modified on Tuesday, September 9, 2008 11:53 AM

CPallini

Sorry no STOS here:

...
const int N = 10;
int a[N]={7};
00401003 xor eax,eax
00401005 push esi
00401006 mov dword ptr [esp+8],7
0040100E mov dword ptr [esp+0Ch],eax
00401012 mov dword ptr [esp+10h],eax
00401016 mov dword ptr [esp+14h],eax
0040101A mov dword ptr [esp+18h],eax
0040101E mov dword ptr [esp+1Ch],eax
00401022 mov dword ptr [esp+20h],eax
00401026 mov dword ptr [esp+24h],eax
0040102A mov dword ptr [esp+28h],eax
0040102E mov dword ptr [esp+2Ch],eax
...

(Visual C++ 2008 Express Edition, Win XP). :)

If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler. -- Alfonso the Wise, 13th Century King of Castile.
This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong. -- Iain Clarke
[My articles]

SandipG · modified on Tuesday, September 9, 2008 11:53 AM

Same output with Visual C++ 6.0, WinXP SP2 :) Surprising..:confused:

Regards, Sandip.

David Crow

CPallini wrote:

00401003 xor eax,eax

Look two lines up from this one and you should find:

rep stosd

"Love people and use things, not love things and use people." - Unknown

"The brick walls are there for a reason...to stop the people who don't want it badly enough." - Randy Pausch

CPallini

No luck (there isn't such insruction). On the other hand, the output of this program [^] confirms my assumption. See also Sandip's post [^]. I should admit I was very surprised by such a behaviour. :)

If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler. -- Alfonso the Wise, 13th Century King of Castile.
This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong. -- Iain Clarke
[My articles]

toxcct

SandipG :) wrote:

Same output with Visual C++ 6.0

Visual C++ is not what a decent programmer call a standard compliant compiler, Sir !

[VisualCalc][Binary Guide][CommDialogs] | [Forums Guidelines]

toxcct

and what about {0} instead of {7}, and in Release Mode, not in Debug Mode ?

[VisualCalc][Binary Guide][CommDialogs] | [Forums Guidelines]

CPallini

See my added remark here [^]. BTW my tests were of course performed both in Debug and the Release mode. ;P :)

If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler. -- Alfonso the Wise, 13th Century King of Castile.
This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong. -- Iain Clarke
[My articles]

David Crow

toxcct wrote:

...and what about {0} instead of {7}...

0 works fine.

toxcct wrote:

...and in Release Mode, not in Debug Mode ?

Same results for both 0 and 7. This is why I only do it when setting things to 0. I use memset() otherwise.

"Love people and use things, not love things and use people." - Unknown

"The brick walls are there for a reason...to stop the people who don't want it badly enough." - Randy Pausch

CPallini

IMHO such a discrepancy would be too gross, even for VC6. :)

If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler. -- Alfonso the Wise, 13th Century King of Castile.
This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong. -- Iain Clarke
[My articles]

toxcct

yes, probably, but it's not *THE* compiler to test the standard ! ;) ;P

[VisualCalc][Binary Guide][CommDialogs] | [Forums Guidelines]

CPallini

DavidCrow wrote:

0 works fine.

It is just a side-effect of default (int) initialization: you're actually initializing only the first array item. Try the following code:

#include <iostream>
using namespace std;

struct MyStruct
{
MyStruct():_i(-1), _j(0),_k(-1){ }
MyStruct(int a):_i(a), _j(a), _k(a){ }
int _i,_j,_k;
};

void main()
{
int i;
MyStruct a[5] = {0};
for (i=0; i<5; i++)
{
cout << i << ") {" << a[i]._i << ", " << a[i]._j <<", " << a[i]._k << "}" << endl;
}
}

DavidCrow wrote:

This is why I only do it when setting things to 0. I use memset() otherwise.

The above is a wise approach. :)

If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler. -- Alfonso the Wise, 13th Century King of Castile.
This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong. -- Iain Clarke
[My articles]

David Crow

For structs, I'd use memset(). I only use 0 for POD types.

"Love people and use things, not love things and use people." - Unknown

"The brick walls are there for a reason...to stop the people who don't want it badly enough." - Randy Pausch

CPallini

DavidCrow wrote:

For structs, I'd use memset(). I only use 0 for POD types.

That's good. :)

If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler. -- Alfonso the Wise, 13th Century King of Castile.
This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong. -- Iain Clarke
[My articles]