Question

Alright, I’m trying to recreate the old classic, Missile Command, using OpenGL in C++. This is my first foray into OpenGL, although I feel fairly comfortable with C++ at this point.

I figured my first task was to figure out how to move 2d objects around the screen, seemed like it would be fairly simple. I created two quick method calls to make either triangles or quads:

void makeTriangle(color3f theColor, vertex2f &p1, vertex2f &p2, vertex2f &p3,
                  int &xOffset, int &yOffset)
{
    //a triangle
    glBegin(GL_POLYGON);
        glColor3f(theColor.red, theColor.green, theColor.blue);
        glVertex2f(p1.x, p1.y);
        glVertex2f(p2.x, p2.y);
        glVertex2f(p3.x, p3.y);
    glEnd();
}

void makeQuad(color3f theColor, vertex2f &p1, vertex2f &p2, vertex2f &p3,
              vertex2f &p4, int &xOffset, int &yOffset)
{
    //a rectangle
    glBegin(GL_POLYGON);
        glColor3f(theColor.red, theColor.green, theColor.blue);
        glVertex2f(p1.x, p1.y);
        glVertex2f(p2.x, p2.y);
        glVertex2f(p3.x, p3.y);
        glVertex2f(p4.x, p4.y);
    glEnd();
}

color3f and vertex2f are simple classes:

class vertex2f
{
public:
    float x, y;

    vertex2f(float a, float b){x=a; y=b;}
};

class color3f
{
public:
    float red, green, blue;

    color3f(float a, float b, float c){red=a; green=b; blue=c;}
};

And here is my main file:

#include <iostream>
#include "Shapes.hpp"

using namespace std;

int xOffset = 0, yOffset = 0;
bool done = false;

void keyboard(unsigned char key, int x, int y)
{
    if( key == 'q' || key == 'Q')
        {
            exit(0);
            done = true;
        }

        if( key == 'a' )
            xOffset = -10;

        if( key == 'd' )
            xOffset = 10;

        if( key == 's' )
            yOffset = -10;

        if( key == 'w' )
            yOffset = 10;

}


void init(void)
{
    //Set color of display window to white
    glClearColor(0.0f, 0.0f, 0.0f, 0.0f);

    //Set parameters for world-coordiante clipping window
    glMatrixMode(GL_PROJECTION);
    gluOrtho2D(-400.0,400.0,-300.0,300.0);

}


void display(void)
{
    glClear(GL_COLOR_BUFFER_BIT);

    color3f aGreen(0.0, 1.0, 0.0);
    vertex2f pa(-400,-200);
    vertex2f pb(-400,-300);
    vertex2f pc(400,-300);
    vertex2f pd(400,-200);

    makeQuad(aGreen,pa,pb,pc,pd,xOffset,yOffset);

    color3f aRed(1.0, 0.0, 0.0);
    vertex2f p1(-50.0,-25.0);
    vertex2f p2(50.0,-25.0);
    vertex2f p3(0.0,50.0);

    makeTriangle(aRed,p1,p2,p3,xOffset,yOffset);

    glFlush();
}


int main(int argc, char** argv)
{
    // Create Window.
    glutInit(&argc, argv);
    glutInitDisplayMode(GLUT_SINGLE | GLUT_RGB);
    glutInitWindowSize(WINDOW_WIDTH, WINDOW_HEIGHT);
    glutCreateWindow("test");

    // Some initialization.
    init();

    while(!done)
    {
    //display functions
    glutDisplayFunc(display);
    glutKeyboardFunc(keyboard);

    // Start event loop.
    glutMainLoop();
    }

    return 0;
}

A quad is defined as the “background” for the time being and consists of just a green rectangle along the bottom of the screen. The red triangle is the “object” that I wish to move. On a keypress, an offset is saved in the direction indicated.

I’ve tried using glTranslatef(xOffset,yOffset,0); but the problem with that is that it moves both elements on the screen and not just the red triangle. I attempted to put the whole call to draw the triangle between a push and pop matrix operation:

PushMatrix();
glTranslatef(xOffset,yOffset,0);
glBegin(GL_POLYGON);
    glColor3f(theColor.red, theColor.green, theColor.blue);
    glVertex2f(p1.x, p1.y);
    glVertex2f(p2.x, p2.y);
    glVertex2f(p3.x, p3.y);
glEnd();
PopMatrix();

As far as I can tell, that destroys any changes that the translation was doing beforehand.
I’ve also tried just changing the values of the x and y coordinates before calling the draw, but that just causes a brief flicker before leaving the triangle in its original position:

p1.x += xOffset;
p2.x += xOffset;
p3.x += xOffset;

p1.y += yOffset;
p2.y += yOffset;
p3.y += yOffset;

There has to be a nice simple way of doing this, and I’m just overlooking it. Could someone offer a suggestion please?

EDIT:

My actual problem was that I was never refreshing the screen after an initial draw. What I needed was to specify an idle function inside my main loop:

glutIdleFunc(IdleFunc);

Where the actual IdleFunc looks like:

GLvoid IdleFunc(GLvoid)
{
    glutPostRedisplay();
}

Instead of using glFlush() inside my draw function, I should have been using glutSwapBuffers(). By doing that, the code I had first come up with:

p1.x += xOffset;
p2.x += xOffset;
p3.x += xOffset;

p1.y += yOffset;
p2.y += yOffset;
p3.y += yOffset;

Works fine for my purposes. I didn’t have a need to translate the matrix, I just needed to draw the element in a different position from one scene to the next.

Answer 1

GL_MODELVIEW is what you need.

From the OpenGL FAQ, 2.1: http://www.opengl.org/resources/faq/technical/gettingstarted.htm

program_entrypoint
{
    // Determine which depth or pixel format should be used.
    // Create a window with the desired format.
    // Create a rendering context and make it current with the window.
    // Set up initial OpenGL state.
    // Set up callback routines for window resize and window refresh.
}

handle_resize
{
    glViewport(...);
    glMatrixMode(GL_PROJECTION);
    glLoadIdentity();
    // Set projection transform with glOrtho, glFrustum, gluOrtho2D, gluPerspective, etc.
}

handle_refresh
{
    glClear(...);

    glMatrixMode(GL_MODELVIEW);
    glLoadIdentity();
    // Set view transform with gluLookAt or equivalent

    // For each object (i) in the scene that needs to be rendered:
        // Push relevant stacks, e.g., glPushMatrix, glPushAttrib.
        // Set OpenGL state specific to object (i).
        // Set model transform for object (i) using glTranslatef, glScalef, glRotatef, and/or equivalent.
        // Issue rendering commands for object (i).
        // Pop relevant stacks, (e.g., glPopMatrix, glPopAttrib.)
    // End for loop.

    // Swap buffers.
}