|
Because the math operations performed by your code result in a NaN?
You cast an integer with at least the relevant most signficant bits set to float or double?
You have a buffer overflow writing such a value to the memory of a float or double?
Without seeing code, we can't tell you for sure.
|
|
|
|
|
2 most likely reasons:
Invalid math e.g. division by zero, or use of uninitialized variables. Without seeing your code, and input, we can't know.
|
|
|
|
|
ahem... Division by zero in general produces a (signed) Inf, not a NaN.
Software rusts. Simon Stephenson, ca 1994. So does this signature. me, 2012
|
|
|
|
|
Because what ever you are doing, the result is not a number.
"One man's wage rise is another man's price increase." - Harold Wilson
"Fireproof doesn't mean the fire will never come. It means when the fire comes that you will be able to withstand it." - Michael Simmons
"You can easily judge the character of a man by how he treats those who can do nothing for him." - James D. Miles
|
|
|
|
|
|
Hi
I have some questions regarding storage allocations. I know there is default heap value every exe or DLL gets. You can make that bigger or smaller /HEAP
I am guessing this is where the "new" operator gets storage.
using GetProcessHeap gets storage from this heap
If I don't want to bump heads with that I do a HeapAlloc initially the three paramters all zeros
But I "THINK" even if I do a HeapAlloc for 20 byes it really gets 1K
if I see that during my processing I need more I do a HeapRealloc using the same pointer for return value and 1st parameter I get more storage
I don't know if this better than using new
would anyone have any input ?
|
|
|
|
|
|
ForNow wrote: But I "THINK" even if I do a HeapAlloc for 20 byes it really gets 1K
I don't know if this better than using new
There is a very specific reason for using HeapAlloc and nothing you posted suggests you need that.
And not a good idea to attempt to circumvent C++ compiler allocations unless you have identified a specific need and you understand exactly how allocations work.
If you have some limited need or just want to mess around then you can override allocations for a specific class (one class only) and control it in fine detail. And do so knowing that you are not as likely to mess up the rest of the application doing it.
Alternatively find several open source allocators and heaps and examine how they are implemented first to understand what happens with heaps.
Keep in mind that if you do attempt to control your own allocations you can completely destroy the application (runtime) because you can end up overwriting memory that has nothing to do with what you think it should. So do it very carefully.
|
|
|
|
|
The run time has it own heap I think the default heap
When I create my heap I’m segragting the allocations
|
|
|
|
|
ForNow wrote: The run time has it own heap I think the default heap
Modern OSes generally have at least two allocators. So in C++ the OS provides space to the app and then C++ itself manages that heap with its own allocator. Same is true for Java and C#.
ForNow wrote: When I create my heap I’m segragting the allocations
Not sure that what that means.
An heap already represents a segregation. The C++ heap (every one I have looked at going back decades) dynamically manages blocks that it requested from the OS. Then on top of that it lays down a simple (or even complex) traditional heap. Then when you use 'new' (presuming you do not otherwise provide an allocator) it uses that existing API to request an amount of space appropriate for usage. Generally the allocation is exact to the structure.
The structure might have padding but that is intrinsic to the structure and not the allocation.
|
|
|
|
|
New works well when you now Exactly howvmuch storage you want I.e a object or class
In my case I am capturing user input and not quite sure thus HeapReAlloc
|
|
|
|
|
ForNow wrote: New works well when you now Exactly howvmuch storage you want I.e a object or class
That is not true.
For an object in C++ you must ALWAYS allocate at least as much space is needed for the object. You can allocate more but it will be wasted.
One can force C++ to use space that is less but that would be nothing but bad programming.
ForNow wrote: In my case I am capturing user input and not quite sure thus HeapReAlloc
All that means is that you are dynamically allocating the space. Which is something that C++ allows directly. You do not need to use a system API method to do that.
As I previously pointed out HeapReAlloc has a specific purpose. And what you just described here is an inappropriate use of that. And one that could lead to resource (memory) starvation if you are not careful how you use it.
|
|
|
|
|
My Addstring works fine its in OinitDialog
except when I make it ownerdraw then I get an exception however even out it I mean Addstring
DrawItem never gets control
|
|
|
|
|
It would be helpful to know which kind of exception and where it occurs.
However, has your owner drawn list box also the LBS_HASSTRINGS style?
If not, that is probably the reason:
If the list box has an owner-drawn style but not the LBS_HASSTRINGS style, this parameter is stored as item data instead of a string. You can send the LB_GETITEMDATA and LB_SETITEMDATA messages to retrieve or modify the item data.
|
|
|
|
|
I have that "LBS_HASSSTRINGS" actually just stepping thru the code on ClistBox::AddString the VS debugger
comes up with a DialogBox "BreakPoint hit" a message indicating some sort of exception
I am wondering if I cannt do AddString in OnitDialog for a OwneDraw litbox ?
|
|
|
|
|
Calling AddString() itself should not raise an exception (provided that the list box window exists) because that just sends the LB_ADDSTRING message. I guess it is happening later when other list box operations are performed.
|
|
|
|
|
I made a custom DDX to create the list box
DDX_TextVAL(pDX, IDC_STORAGE_AREAS, storage_area);
HWND tempwin = ::GetDlgItem(pDX->m_pDlgWnd->m_hWnd, nIDC);
LPTSTR temptr = (LPTSTR)new char[10];
::GetClassName(tempwin, temptr, 10);
if (strcmp((char *)temptr, "Edit") == 0)
value.Attach(pDX->PrepareEditCtrl(nIDC));
else
value.Attach(pDX->PrepareCtrl(nIDC));
delete temptr;
CListBox storage_area
|
|
|
|
|
ForNow wrote: I made a custom DDX to create the list box
And what for?
IMHO, it is not a good idea...
|
|
|
|
|
|
When I step thru the debugger on ClistBox::AddString it (VS debugger) just says breakpoint hit seems like for OwnerDraw I cann't do addstring in OinitDialog
Notsure
|
|
|
|
|
With that little information it's anyone's guess what is going on. I have used AddString in the OnInitDialog method, in normal and owner drawn list boxes.
|
|
|
|
|
Ok thanks that’s helpful I’ll do some digging
|
|
|
|
|
Consider the following code
"main.cpp"
#include <stdio.h>
#include "../include/ESSolver.h"
int mu;
int lambda ;
int main(void)
{
mu = 5;
lambda = 10;
ESSolver es;
es.Init();
es.Optimize();
return 0;
}
"ESSolver.h"
extern int mu;
extern int lambda ;
class ESSolver
{
public:
const static int nVar = 4;
ESSolver();
~ESSolver(void);
void Init();
void Optimize();
int rand_r(int, int);
private:
double (*muPop)[nVar];
double *muSigma;
double (*lambdaPop)[nVar];
double *lambdaSigma;
double (*mulambdaPop)[nVar];
double *mulambdaSigma;
};
"ESSolver.cpp"
#include <iostream>
#include <cmath>
#include <vector>
#include <algorithm>
#include <queue>
#include "../include/ESSolver.h"
using namespace std;
ESSolver::ESSolver()
{
double _muPop [mu][nVar];
muPop = _muPop;
muSigma = new double[mu];
double _lambdaPop [lambda][nVar];
lambdaPop = _lambdaPop;
lambdaSigma = new double[lambda];
double _mulambdaPop [mu+lambda][nVar];
mulambdaPop = _mulambdaPop;
mulambdaSigma = new double[mu+lambda];
return;
}
ESSolver::~ESSolver(void)
{
if (muSigma) delete muSigma;
if (lambdaSigma) delete lambdaSigma;
if (mulambdaSigma) delete mulambdaSigma;
return;
}
void ESSolver::Init()
{
for (int i = 0; i < mu; i++)
{
for (int j = 0; j < nVar; j++)
{
muPop[i][j] = 50;
}
}
cout << "\n----- muPop -----\n";
for (int i = 0; i < mu; i++)
{
for (int j = 0; j < nVar; j++)
{
cout << muPop[i][j] << " ";
}
cout << endl;
}
}
void ESSolver::Optimize()
{
cout << "\n----- muPop -----\n";
for (int i = 0; i < mu; i++)
{
for (int j = 0; j < nVar; j++)
{
cout << muPop[i][j] << " ";
}
cout << endl;
}
}
On running the following code I am getting Output
[^]
As you can see from the image, Matrix is not getting initialized. Can anybody tell me why.
modified 29-Jan-21 21:01pm.
|
|
|
|
|
Quote: double _muPop [mu][nVar];
muPop = _muPop;
_mPop is a temporary array (a local variable), your muPop pointer is goingo to point to garbage.
|
|
|
|
|
Thanks 
modified 29-Jan-21 21:01pm.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
