I am having a very tough time sorting out this strange clipping bug in my app.
Basically, for some reason, OpenGL is clipping (using the scissor test) my call to glClear(), but not the rendering I do afterwards.
The real problem, however, is that the problem goes away when I resize my window. I can guarantee that resizing the window doesn't change anything in my app or run any code. It is very stange. Worse still, simply putting
glDisable(GL_SCISSOR_TEST);
glDisable(GL_SCISSOR_TEST);
where I need to disable the scissor test, instead of having just one call to glDisable() solves the problem. So does removing the code all together (the scissor test is already disabled in this test case, but the code is there for when it wasn't left to disabled in previous code). It even solves the problem to put:
glEnable(GL_SCISSOR_TEST);
glDisable(GL_SCISSOR_TEST);
There are only two explanations I can think of. Either I am somehow calling UB (which I doubt, because opengl doesn't have UB AFAIK), or there is an implementation bug, because calling glDisable() twice with the same parameter consecutively SHOULD be the same as calling it once... if I'm not mistaken.
JUST incase it is of interest, here is the function for which the problem is happening:
void gle::Renderer::setup3DCamera(gle::CameraNode& cam, gle::Colour bkcol,
int clrmask, int skymode, gle::Texture* skytex, bool uselight) {
// Viewport
Rectangle wr(cam.getViewport()?*cam.getViewport():Rectangle(0,0,1,1));
if (cam.isRatioViewport()||(!cam.getViewport())) {
if (i_frameBind==NULL)
wr.scale(selectedWindow->getWidth(),selectedWindow->getHeight());
else wr.scale(i_frameBind->getWidth(),i_frameBind->getHeight());
}
gle::Rectangle_t<int> iport; iport.set(wr);
int winHei;
if (i_frameBind==NULL)
winHei = selectedWindow->getHeight();
else
winHei = i_frameBind->getHeight();
glViewport(iport.x1(),winHei-iport.y2(),iport.wid(),iport.hei());
// Viewport Clipping
if (cam.isClipping()) {
/* This is never executed in the test case */
glEnable(GL_SCISSOR_TEST);
glScissor(iport.x1(),winHei-iport.y2(),iport.wid(),iport.hei());
} else {
/* This is where I disable the scissor test */
glDisable(GL_SCISSOR_TEST);
glDisable(GL_SCISSOR_TEST);
}
float w=wr.wid()/2, h=wr.hei()/2;
// Projection
glMatrixMode(GL_PROJECTION);
glLoadIdentity();
Projection proj = cam.getProjection();
gluPerspective(proj.fov,proj.aspect*(w/h),proj.cnear,proj.cfar);
// Camera
glMatrixMode(GL_MODELVIEW);
glLoadIdentity();
float m[] = { 1,0,0,0, 0,0,-1,0, 0,1,0,0, 0,0,0,1 };
glMultMatrixf(m);
static gle::Mesh *skyBox = NULL;
// Screen Clearing
switch (clrmask&GLE_CLR_COLOUR&0x00F?skymode:GLE_SKYNONE) {
case GLE_SKYNONE:
clear(clrmask&(~GLE_CLR_COLOUR)); break;
case GLE_SKYCOLOUR:
clearColour(clrmask,bkcol); break;
case GLE_SKYBOX:
glDisable(GL_DEPTH_TEST);
if (!(clrmask&GLE_CLR_DEPTH&0x00F)) glDepthMask(0);
float m = (cam.getProjection().cnear+cam.getProjection().cfar)/2.0f;
if (skyBox==NULL) skyBox = gle::createStockMesh(GLE_MESHSKYBOX,GLE_WHITE,0,m);
glEnable(GL_TEXTURE_2D);
glDisable(GL_CULL_FACE);
skytex->flush();
glBindTexture(GL_TEXTURE_2D,skytex->getID());
glDisable(GL_LIGHTING);
glPushMatrix();
float m3[16];
Orientation::matrixSet(m3,cam.pos().getMatrix(GLE_ROTMATRIX));
Orientation::matrixTranspose(m3);
glMultMatrixf(m3);
if (i_reflectionOn) glMultMatrixf(Orientation::matrixGet3x3(i_reflectionTransform));
renderMesh(*skyBox,NULL,1);
glPopMatrix();
glDisable(GL_TEXTURE_2D);
if (clrmask&GLE_CLR_DEPTH) glClear(GL_DEPTH_BUFFER_BIT);
else glDepthMask(1);
glAble(GL_DEPTH_TEST,depthmode!=GLE_ALWAYS);
break;
}
// Camera
glMultMatrixf(cam.getAbsInverseMatrix());
if (i_reflectionOn) glMultMatrixf(i_reflectionTransform);
// Lighting
i_lightOn = uselight;
glAble(GL_LIGHTING,i_lightOn);
}
This looks like a driver bug to me. However, there are two cases where this may actually be a bug in your code.
First, you might be in the middle of the glBegin() / glEnd() block when calling that glDisable(), causing some error and also ending the block, effectively making the second call to glDisable() legit and effective. Note that this is a dumb example with glBegin() / glEnd(), it could be pretty much any case of OpenGL error being caught. Insert glGetError() calls throughout your code to be sure. My guess is the first call to glDisable() generates GL_INVALID_OPERATION.
Second, you are not scissor testing, but you are still calling glViewport() with the same values. This would have the opposite effect (not clip glClear() and clip drawing) on NVIDIA, but it might very well be it does the opposite on some other driver / GL implementation.
Related
A Project I am working on involves me using glScissor, in some cases i need to perform a scissor on an area twice (or more), with the goal of only rendering what is within both scissor boxes.
The issue im running into is that the second scissor box just overrides the previous one, meaning only the last box set is used instead of both.
I have tried existing solutions such as setting scissor1, push matrix, enable scissor_test, set scissor2, disable scissor_test, popmatrix, disable scissor_test. As proposed here: glScissor() call inside another glScissor()
I could not get these to produce any difference, I had also tried glPushAttrib instead of matrix but still no difference.
Here is an example program I wrote for scissor testing, its compiled by g++ and uses freeglut, the scissoring takes place in display():
/*
Compile: g++ .\scissor.cpp -lglu32 -lfreeglut -lopengl32
*/
#include <GL/gl.h>//standard from mingw, already in glut.h - header library
#include <GL/glu.h>//standard from mingw, already in glut.h - utility library
#include <GL/glut.h>//glut/freeglut - more utilities, utility tool kit
void display();
void reshape(int, int);
void timer(int);
void init(){
glClearColor(0, 0, 0, 1);
}
int main(int argc, char **argv){
glutInit(&argc, argv);//init glut
glutInitDisplayMode(GLUT_RGB | GLUT_DOUBLE);//init display mode, add double buffer mode
//init window
glutInitWindowPosition(200, 100);//if not specified, it will display in a random spot
glutInitWindowSize(500, 500);//size
//create window
glutCreateWindow("Window 1");
//give glut a function pointer so it can call that function later
glutDisplayFunc(display);
glutReshapeFunc(reshape);
glutTimerFunc(0, timer, 0);//call certain function after a specified amount of time
init();
glutMainLoop();//once this loop runs your program has started running, when the loop ends the program terminates
}
float xPos = -10;
int state = 1;//1 = right, -1 = left
//our rendering happens here
void display(){
//clear previous frame
glClear(GL_COLOR_BUFFER_BIT);//pass in flag of frame buffer
//draw next frame below
glLoadIdentity();//reset rotations, transformations, ect. (resets coordinate system)
//we are using a model view matrix by default
//TEST
glEnable(GL_SCISSOR_TEST);
glScissor(0, 0, 100, 1000);
glPushMatrix();
glEnable(GL_SCISSOR_TEST);
glScissor(50, 0, 1000, 1000);
//assuming both scissors intersect, we should only see the square between 50 and 100 pixels
//draw
glBegin(GL_QUADS);//every set of 3 verticies is a triangle
//GL_TRIANGLES = 3 points
//GL_QUADS = 4 points
//GL_POLYGON = any amount of points
glVertex2f(xPos, 1);//the 2 is the amount of args we pass in, the f means theyr floats
glVertex2f(xPos, -1);
glVertex2f(xPos+2, -1);
glVertex2f(xPos+2, 1);
glEnd();//tell opengl your done drawing verticies
glDisable(GL_SCISSOR_TEST);
glPopMatrix();
glDisable(GL_SCISSOR_TEST);
//display frame buffer on screen
//glFlush();
glutSwapBuffers();//if double buffering, call swap buffers instead of flush
}
//gets called when window is reshaped
void reshape(int width, int hight){
//set viewport and projection
//viewport is a rectangle where everything is drawn, like its the window
glViewport(0, 0, width, hight);
//matrix modes: there is model view and projection, projection has depth
glMatrixMode(GL_PROJECTION);
glLoadIdentity();//reset current matrix after changing matrix mode
gluOrtho2D(-10, 10, -10, 10);//specify 2d projection, set opengl's coordinate system
glMatrixMode(GL_MODELVIEW);//change back to model view
}
//this like makes a loop
void timer(int a){
glutPostRedisplay();//opengl will call the display function the next time it gets the chance
glutTimerFunc(1000/60, timer, 0);
//update positions and stuff
//this can be done here or in the display function
switch(state){
case 1:
if(xPos < 8)
xPos += 0.15;
else
state = -1;
break;
case -1:
if(xPos > -10)
xPos -= 0.15;
else
state = 1;
break;
}
}
I tried following example solutions, such as push/pop matrix/attrib, but couldnt get anything to work
There is no first or second scissor box. There is just the scissor box. You can change the scissor box and that change will affect subsequent rendering. But at any one time, there is only one.
What you want is to use the stencil buffer to discard fragments outside of an area defined by rendering certain values into the stencil buffer.
I have an interesting bug that has been "bugging" me for a few days now.
I am currently using OpenGL to draw text on a screen. I am utilizing the OGLFT library to assist the drawing. This library actually uses the freetype2 library. I am actually not doing anything special with the text. I am only looking for monochromatic text.
Anyways, after implementing the library, I noticed that the text is only drawn correct when I have glStipple enabled. I believe that there is some interference issue between the OGLFT library and what I am enabling.
I was wondering if there is anyone out there with some experience on using the OGLFT library. I am posting a minimalist example of my code to demonstrate what is going on:
(Please note that there are some variables that are used to st the zoom factor of my glCanvas and the position of the camera and that this is only for 2D)
double _zoomX = 1.0;
double _zoomY = 1.0;
double _cameraX = 0;
double _cameraY = 0;
/* This function gets called everytime a draw routine is needed */
void modelDefinition::onPaintCanvas(wxPaintEvent &event)
{
wxGLCanvas::SetCurrent(*_geometryContext);// This will make sure the the openGL commands are routed to the wxGLCanvas object
wxPaintDC dc(this);// This is required for drawing
glMatrixMode(GL_MODELVIEW);
glClear(GL_COLOR_BUFFER_BIT);
updateProjection();
OGLFT::Monochrome *testface = new OGLFT::Monochrome( "/usr/share/fonts/truetype/dejavu/DejaVuSansMono.ttf", 8);
testface->draw(0, 0, "test");
glEnable(GL_LINE_STIPPLE);// WHen I comment out this line, the text is unable to be drawn
glLineStipple(1, 0b0001100011000110);
glBegin(GL_LINES);
glVertex2d(_startPoint.x, _startPoint.y);
glVertex2d(_endPoint.x, _endPoint.y);
glEnd();
glDisable(GL_LINE_STIPPLE);
SwapBuffers();
}
void modelDefinition::updateProjection()
{
// First, load the projection matrix and reset the view to a default view
glMatrixMode(GL_PROJECTION);
glLoadIdentity();
glOrtho(-_zoomX, _zoomX, -_zoomY, _zoomY, -1.0, 1.0);
//Reset to modelview matrix
glMatrixMode(GL_MODELVIEW);
glLoadIdentity();
glViewport(0, 0, (double)this->GetSize().GetWidth(), (double)this->GetSize().GetHeight());
/* This section will handle the translation (panning) and scaled (zooming).
* Needs to be called each time a draw occurs in order to update the placement of all the components */
if(_zoomX < 1e-9 || _zoomY < 1e-9)
{
_zoomX = 1e-9;
_zoomY = _zoomX;
}
if(_zoomX > 1e6 || _zoomY > 1e6)
{
_zoomX = 1e6;
_zoomY = _zoomX;
}
glTranslated(-_cameraX, -_cameraY, 0.0);
}
Also one thing to note is that the code below the glEnable(GL_LINE_STIPPLE); is required. It is as if the glStipple needs to be drawn correctly for the text to be displayed correctly.
Looking through your code, I believe that your intention is to render it as a greyscale? If so, then you can simply use the OGLFT::Grayscale *testface = new OGLFT::Grayscale( "/usr/share/fonts/truetype/dejavu/DejaVuSansMono.ttf", 8);
This will get what you need without having to worry about the issue that you posted. In fact, I recommend doing it this way too.
I'm trying to apply transformation to a specific part of my drawing that will only come in effect when a key is pressed.
I'm using glPushMatrix() then I do my transformation, then glPopMatrix().
What happens when I press my key is that the transformation is done only once, then no matter how many times I press the key again nothing happens.
Here's a snippet from my code:
void drawShapes(){
glClear(GL_COLOR_BUFFER_BIT);
//setting my variables
glPushMatrix();
if (translateFlag) //this is set to true in the keyboard function
{
x += 10;
glTranslated(x, 0, 0);
translateFlag = false;
}
drawCircle();
glPopMatrix();
//more drawings
}
when I remove the pushMatrix and popMatrix it works but it applies the transformation to all of my shapes, which is not what I want.
Any help would be appreciated..
[Update]
I've tried to run my code on 4 other computers (one macbook Air, one macbook Pro -an exact copy of mine- and 2 iMacs and) and it doesn't even run on any of them, could this mean the problem lies in my own macbook?!
Also I copied code that runs perfectly fine on the macbook air, rotating parts and everything and when I run it on mine I got the same disappointing result as my own code.
Here's the heypasteit code for my code 31JL
possible problems:
void drawShapes(){
glClear(GL_COLOR_BUFFER_BIT);
//setting my variables
// maybe you have forgot to set matrix mode correctly. make the projection and modelview
matrices correct.
glPushMatrix();
if (translateFlag) //this is set to true in the keyboard function
//probably your key hit function is not working properly (I can't say without looking at full code)
{
x += 10;
glTranslated(x, 0, 0);
translateFlag = false;
}
drawCircle();
glPopMatrix();
//more drawings
// you are using glFlush() ????? try swapbuffers. exact function depends
on the API you are using (GLUT \ GLFW etc.)
}
I am using the GL_SELECT method to achieve mouse selection in OpenGL using JOGL Java Library.
I know the method is deprecated and such, but it is a simple school assignment and this should do it.
However I am having some trouble: even though something is rendered in GL_SELECT mode, glRenderMode(GL_RENDER) returns zero hits. The problem is deterministic, but I don't see a kind of pattern; for example, if I have a sphere in the center, it works if I click on its upper part, but not on its lower part. For a cube, it only won't work on one specific face. For a rectangle it works alright.
I have tested commenting out the glRenderMode(GL_SELECT) to check if something was indeed being rendered and yes, I could see the shape, but even so glRenderMode(GL_RENDER) gave me zero.
EDIT: I have also tested removing the call to gluPickMatrix() and the glRenderMode(GL_SELECT), which gave me exactly the same as the normal (non-picking) render, so the projection and model view matrixes are set up correctly I think.
So, I don't think I am rendering incorrectly in select mode. What can be going on?
EDIT: maybe this could be a hardware problem, as the method is deprecated. Is that possible?
Thanks in advance.
// Get required information
point.y = getHeight() - point.y;
gl.glGetIntegerv(GL2.GL_VIEWPORT, view, 0);
// Setup OpenGL for selection
gl.glSelectBuffer(64, buffer);
gl.glRenderMode(GL2.GL_SELECT);
gl.glInitNames();
// Setup projection matrix
gl.glMatrixMode(GL2.GL_PROJECTION);
gl.glPushMatrix();
gl.glLoadIdentity();
Util.glu.gluPickMatrix( point.x, point.y, 5.0, 5.0, view, 0 );
Util.glu.gluPerspective(camera.getFieldOfView(), getWidth() * 1.0 / getHeight(),
camera.getCloseDistance(), camera.getFarDistance() );
// Setup model view matrix for rendering
gl.glMatrixMode(GL2.GL_MODELVIEW);
camera.setView(gl); // Set to model view and use glLookAt
gl.glClear( GL2.GL_COLOR_BUFFER_BIT | GL2.GL_DEPTH_BUFFER_BIT );
// Render objects
for(int i = 0; i < shapeList.size(); i++) {
gl.glPushName(i);
// Execute transformations for translation/rotation/scale and render shape
shapeList.get(i).display(gl, false);
gl.glPopName();
}
// Process hits
hits = gl.glRenderMode(GL2.GL_RENDER);
System.out.println("Hits = " + hits);
// ... Process hits here ...
// Reset matrixes
gl.glMatrixMode(GL2.GL_PROJECTION);
gl.glPopMatrix();
gl.glMatrixMode(GL2.GL_MODELVIEW);
gl.glLoadIdentity();
camera.setView function:
public void setView( GL2 gl ) {
gl.glMatrixMode(GL2.GL_MODELVIEW);
gl.glLoadIdentity();
Util.glu.gluLookAt( eye[Axis.X], eye[Axis.Y], eye[Axis.Z],
target[Axis.X], target[Axis.Y], target[Axis.Z],
up[Axis.X], up[Axis.Y], up[Axis.Z] );
}
I am working on a small animation of a sinewave being generated by a moving circle similar to this gif.
However I'm having a weird issue with using glutWireTorus for the circle. Even when I comment out all the other drawing code and leave the torus moving along the z axis alone, I get flashing lines on the inside of it:
This is the display function where everything is drawn. I use the display function as my idle function, also, I checked if that was the issue by creating an idle function and putting my update code there but it still persisted.
void display(void)
{
glClearColor(0.f,0.f,0.f,0.f);
glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT); // clear the color buffer
and the depth buffer
glLoadIdentity();
glTranslatef(0.0f,0.0f,-5.0f); //Push everything back 5 units into the scene,
otherwise we won't see the primitive
camera();
//yPrev = myCircle.y;
glPushMatrix();
glTranslatef(0.0f,0.0f,t);
glutWireTorus(0.02,1,100,100);
glPopMatrix();
DrawAxes();
glutSwapBuffers(); //swap the buffers
glFlush();//Flush the OpenGL buffers to the window
t-=0.002;
if (t < -T)
{
t = 0;
}
}
Uh, interesting, I run into the same problem.
Here is the picture using parameter
glutWireTorus(/*innerRadius*/0.3, /*outerRadius*/0.5, /*nsides*/32, /*rings*/32);
another picture using parameter
glutWireTorus(/*innerRadius*/0.3, /*outerRadius*/0.5, /*nsides*/64, /*rings*/128);
We don't know why without glutWireTorus function implementation, but we can implements our own torus, or maybe someone can explain it?