[SOLVED] Loading objects from Blender by Qt OpenGL

[8Observer8]

Hello

I have crash in the construction Scene::Scene() https://github.com/8Observer8/ObjLoader_v1_0_0

You should run my application -> Click "Load" -> Select the file "Plane.observerFormat"

I think that I cannot fill std::vector like this:

@
#include <iostream>
#include <vector>

void fill( std::vector<float> &vec1, std::vector<float> &vec2 );
void show( const std::vector<float> &vec );

int main()
{
std::vector<float> vec1;
std::vector<float> vec2;
fill( vec1, vec2 );

show( vec1 );
std::cout << "\n";
show( vec2 );

return 0;

}

void fill( std::vector<float> &vec1 , std::vector<float> &vec2 )
{
vec1.push_back( -0.5f );
vec1.push_back( 0.5f );
vec1.push_back( 1.5f );

vec2.push_back( 0.5f );
vec2.push_back( 0.5f );
vec2.push_back( 1.5f );

}

void show( const std::vector<float> &vec )
{
for( unsigned int i = 0; i < vec.size(); ++i )
{
std::cout << vec[i] << "/n";
}
}
@

[Chris Kawa]

This works fine: http://ideone.com/aXGV7Z

But your real code is different and has an error:

void Object::initColors( int numOfVertices ) {
    colors.clear();
    colors.resize( numOfVertices * 3 );

    for ( unsigned int i = 0, j = 0; i < numOfVertices; ++i ) {
        colors[j++] = 1.0f;
        colors[j++] = 0.0f;
        colors[j++] = 0.0f;

        colors[j++] = 0.0f;
        colors[j++] = 1.0f;
        colors[j++] = 0.0f;

        colors[j++] = 0.0f;
        colors[j++] = 0.0f;
        colors[j++] = 1.0f;
    }
}

Size of the vector is 3*N. i iterates from 0 to N and for each i you increment j 9 times, which means j reaches 9 * N, which is waaaay beyond the end of the vector.

[8Observer8]

Thank you very much. It works fine. Now I wrote like this:

@for ( unsigned int i = 0, j = 0; i < numOfVertices / 3; ++i )@

Source: https://github.com/8Observer8/CubeFromBlender_v1.0.1
Picture: http://i9.pixs.ru/storage/2/1/3/SquareFrom_3854842_15269213.png

The file "Cube.observerFormat" must be lie in the "currentPath" folder

You can load objects from Blender like me

Copy this script: https://yadi.sk/d/hvmi_wpxdXG5Y
to C:\Program Files (x86)\Blender Foundation\Blender\2.72\scripts\addons

Open Blender -> Press Ctrl+Alt+U -> Select "Addons" -> Select "Objects" -> Check "Object: SupPupExporter" -> Close the window "Blender User Preferences" -> Save Project (.blend file) -> Select "Object Mode" -> Click "Object" -> Select "SupPupExporter" -> You can find ".observerFormat" file ( it is where ".blend" lie)

[Chris Kawa]

I would argue that this is a lot easier to reason about:

void Object::initColors( int numOfVertices ) {
    int size = numOfVertices * 3;
    colors.resize(size); //no need to clear if resizing anyway

    for (int i = 0; i < size; ) {
        colors[i++] = 1.0f;
        colors[i++] = 0.0f;
        colors[i++] = 0.0f;

        colors[i++] = 0.0f;
        colors[i++] = 1.0f;
        colors[i++] = 0.0f;

        colors[i++] = 0.0f;
        colors[i++] = 0.0f;
        colors[i++] = 1.0f;
    }
}

or if you're on c++11:

void Object::initColors( int numOfVertices ) {
  auto size = numOfVertices * 3;
  colors.resize(size);

  std::vector<float> rgb = {1.0f,0.0f,0.0f, 0.0f,1.0f,0.0f, 0.0f,0.0f,1.0f};
  auto i = 0;
  std::generate(colors.begin(), colors.end(), [&]{return rgb[i++ % rgb.size()];});
}

[wickwire]

I was able to get the example running on Windows 8.1 by changing the code a little:

Cube.cpp

@//#include <GL/gl.h>
#include <qopengl.h>@

The Plane.observerFormat model didn't get drawn (and no error messages) but the Cube and some others already generated by me on Blender do work.

Nice job!