Stack overflow exception before entering main()

Question

when compiling (and executing) this c++ -code with vs2010, I get a stack overflow exception just before it could even write a glimpse of "Start" to the console.

All the header-files that I use are included in stdafx.h, but that's obviously not the problem here (same problem when including the header-files directly).

Stack trace is the following:

>   msvcr100d.dll!__set_flsgetvalue()  Zeile 145 + 0xc Bytes    C
    msvcr100d.dll!_getptd_noexit()  Zeile 500   C
    msvcr100d.dll!_getptd()  Zeile 523 + 0x5 Bytes  C
    msvcr100d.dll!_LocaleUpdate::_LocaleUpdate(localeinfo_struct * plocinfo)  Zeile 243 + 0x5 Bytes C++
    003efe3c()  
    TerrainGenerator.exe!pre_cpp_init()  Zeile 298 + 0x21 Bytes C

Any help would be greatly appreciated, I'm an absolute beginner in c++, as you may see in the code-style, I tried exactly EVERYTHING to get rid of this annoying problem, such as not even declaring ONE additional function and so on... Please help me, and don't hesitate to ask if you need additional info. Thanks in advance.

#include "stdafx.h"

int main(int argc, char* argv[])
{
    puts("Start");

    const int res = 4096;
    srand(time(NULL));

    uint16_t data[(res+1)*(res+1)];

    uint16_t variance = 2000;
    int seed = 1337;

    data[0] = ((seed>>1)+(int)(rand()/(RAND_MAX*(seed+(seed>>1)))));
    data[res] = ((seed>>1)+(int)(rand()/(RAND_MAX*(seed+(seed>>1)))));
    data[res*res] = ((seed>>1)+(int)(rand()/(RAND_MAX*(seed+(seed>>1)))));
    data[res*(res+1)] = ((seed>>1)+(int)(rand()/(RAND_MAX*(seed+(seed>>1)))));

    int count = 0;
    for(int size=res;size>=1;size=size>>1,variance=variance>>1)
    {
        count++;
        for(int x=size;x<res;x+=size<<1) 
        {
            for(int y=size;y<res;y+=size<<1) 
            {
                data[x*res+y] = data[(x-size)*res+(y-size)] + data[(x-size)*res+(y+size)] + data[(x+size)*res+(y-size)] + data[(x+size)*res+(y+size)];
                data[x*res+y] /= 4;
                data[x*res+y] += (variance>data[x*res+y])?(-(data[x*res+y]^2/variance)+(int)(rand()/(RAND_MAX*(data[x*res+y]^2/variance)))):(-variance+(int)(rand()/(RAND_MAX*variance*2)));
            }
        }

        for(int x=0;x<res;x+=size)
        {
            for(int y=0;y<res;y+=size)
            {
                if(x!=res-1)
                {
                    data[(x+(size>>1))*res+y] = data[x*res+y] + data[(x+size)*res+y] + ((y!=res-1)?data[(x+(size>>1))*res+(y+(size>>1))]:0) + ((y!=0)?data[(x+(size>>1))*res+(y-(size>>1))]:0);
                    data[(x+(size>>1))*res+y] /= (y!=res-1&&y!=0)?4:3;
                    data[(x+(size>>1))*res+y] += (variance>data[(x+(size>>1))*res+y])?((-(data[(x+(size>>1))*res+y]^2/variance)+(int)(rand()/(RAND_MAX*(data[(x+(size>>1))*res+y]^2/variance))))):(-variance+(int)(rand()/(RAND_MAX*variance*2)));
                }
                if(y!=res-1)
                {
                    data[x*res+(y+(size>>1))] = data[x*res+y] + data[x*res+(y+size)] + ((x!=res-1)?data[(x+(size>>1))*res+(y+(size>>1))]:0) + ((x!=0)?data[(x-(size>>1))*res+(y+(size>>1))]:0);
                    data[x*res+(y+(size>>1))] /= (x!=res-1&&x!=0)?4:3;
                    data[x*res+(y+(size>>1))] += (variance>data[x*res+(y+(size>>1))])?((-(data[x*res+(y+(size>>1))]^2/variance)+(int)(rand()/(RAND_MAX*(data[x*res+(y+(size>>1))]^2/variance))))):(-variance+(int)(rand()/(RAND_MAX*variance*2)));
                }
            }
        }
    }

    size_t Count = res*res;
    uint16_t* Block = data;
    char* Path = "export.raw";

    if(!Block) return false;
    FILE * filePointer = NULL;
    errno_t error = fopen_s(&filePointer, Path, "wb");
    if(error) return false;
    fwrite(Block, sizeof(uint16_t), Count, filePointer);
    fclose(filePointer);
    return true;

    return 0;
}

@user384706: How would that improve the situation?

Benjamin Lindley
– Benjamin Lindley

2012-04-28 19:03:07 +00:00
Commented Apr 28, 2012 at 19:03 — Benjamin Lindley
– Benjamin Lindley, Commented Apr 28, 2012 at 19:03
Looks like a Stack Overflow !

Paul R
– Paul R

2012-04-28 19:03:28 +00:00
Commented Apr 28, 2012 at 19:03 — Paul R
– Paul R, Commented Apr 28, 2012 at 19:03

MiMo · Accepted Answer · 2012-04-28 19:08:39Z

12

The problem is

uint16_t data[(res+1)*(res+1)];

it is a local array of 16 MB+ that is too big to fit in the stack (usually 1 MB) - you have to create it in the heap with a new statement:

uint16_t* data = new uint16_t[(res+1)*(res+1)];

and then remember to de-allocate it with a delete statement when you don't need it any more:

delete[] data;

edited Apr 28, 2012 at 19:08

answered Apr 28, 2012 at 18:59

MiMo

12k1 gold badge35 silver badges48 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

Steve Fallows Over a year ago

Or preferably use a std::vector<uint16_t>.

MiMo Over a year ago

@SteveFallows: as a novice maybe getting a feel for how new and delete work could be good before jumping in the template libraries - not sure about it though, I started myself with C/C++ way before the template libraries were available.

Matthieu M. Over a year ago

@MiMo: I guess it is important to know how it works under the hood, but I would recommend starting with safe constructs and moving to the deeps afterwards. Even if it feels magic to begin with, using RAII is a good habit to take, so let's drill it into beginners early.

Steve Fallows Over a year ago

@MiMo: I started with assembly back in the dark ages, so I totally agree with understanding what's under the hood. Which is better to learn first, I'm not totally sure either - and may vary by individual.

Collectives™ on Stack Overflow

Stack overflow exception before entering main()

1 Answer 1

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related