understanding OpenGL ES crash causes and ways to debug one - c++

Leaning how to use OpenGL ES under cocos2d-x.
Help me to understand while the below function crashes with EXC_BAD_ACCESS on noise->visit(); which probably means that my OpenGL state is screwed-up by the time the command is executed.
My method:
Sprite* GameLevelLayer::spriteWithColor(Color4F bgColor, float textureWidth, float textureHeight)
{
// 1: Create new RenderTexture
auto rt = RenderTexture::create(textureWidth, textureHeight);
// 2: Call RenderTexture:begin
rt->beginWithClear(bgColor.r, bgColor.g, bgColor.b, bgColor.a);
// 3: Draw into the texture
//// gradient
gradient_command.init(rt->getGlobalZOrder());
gradient_command.func = std::bind(&GameLevelLayer::onDrawGradient, this, textureWidth, textureHeight);
auto renderer = Director::getInstance()->getRenderer();
renderer->addCommand(&gradient_command);
//// noise cloud
BlendFunc blendFunc;
blendFunc.src = GL_DST_COLOR;
blendFunc.dst = GL_ZERO;
auto noise = Sprite::create("Noise.png");
noise->setBlendFunc(blendFunc);
noise->setPosition(Vec2(textureWidth / 2, textureHeight / 2));
// XXX
noise->visit();
// 4: Call RenderTexture:end
rt->end();
// 5: Create a new Sprite from the texture
return Sprite::createWithTexture(rt->getSprite()->getTexture());
}
and the gradient:
void GameLevelLayer::onDrawGradient(float textureWidth, float textureHeight)
{
setGLProgram(ShaderCache::getInstance()->getGLProgram(GLProgram::SHADER_NAME_POSITION_COLOR));
getGLProgram()->use();
getGLProgram()->setUniformsForBuiltins();
float gradientAlpha = 0.8f;
std::vector<Vec2> vertices;
std::vector<Color4F> colors;
vertices.push_back(Vec2(0, 0));
colors.push_back(Color4F(0.0f, 0.0f, 0.0f, 0.0f ));
vertices.push_back(Vec2(textureWidth, 0));
colors.push_back(Color4F(0.0f, 0.0f, 0.0f, 0.0f ));
vertices.push_back(Vec2(0, textureHeight));
colors.push_back(Color4F(0.0f, 0.0f, 0.0f, gradientAlpha ));
vertices.push_back(Vec2(textureWidth, textureHeight));
colors.push_back(Color4F(0.0f, 0.0f, 0.0f, gradientAlpha ));
GL::enableVertexAttribs(GL::VERTEX_ATTRIB_FLAG_POSITION | GL::VERTEX_ATTRIB_FLAG_COLOR);
glVertexAttribPointer(GLProgram::VERTEX_ATTRIB_POSITION, 2, GL_FLOAT, GL_FALSE, 0, vertices.data());
glVertexAttribPointer(GLProgram::VERTEX_ATTRIB_COLOR, 4, GL_FLOAT, GL_FALSE, 0, colors.data());
glBlendFunc(CC_BLEND_SRC, CC_BLEND_DST);
glDrawArrays(GL_TRIANGLE_STRIP, 0, (GLsizei)vertices.size());
}
Is this because I am using parameters of a wrong size, somewhere? How does one go about debugging something like this?

Related

OpenGL doesn't draw my Quad. Everything seems to be fine. No errors from OpenGL

I'm trying to make a system that allows you to type in a position and scale and it will create a vector that automatically generates all the vertices. The problem is when I try to draw my object it just won't show up. I have used OpenGL's built-in debugging system but it didn't say anything was wrong. So then I tried to manually debug myself but everything seemed to draw just fine.
Renderer::createQuad() method:
Shape Renderer::createQuad(glm::vec2 position, glm::vec2 scale, Shader shader, Texture texture)
{
float x = position.x;
float y = position.y;
float width = scale.x;
float height = scale.y;
std::vector<float> vertices =
{
x+width, y+height, 1.0f, 1.0f, // TR
x+width, y-height, 1.0f, 0.0f, // BR
x-width, y-height, 0.0f, 0.0f, // BL
x-width, y+height, 0.0f, 1.0f // TL
};
std::vector<uint32_t> indices =
{
0, 1, 3,
1, 2, 3
};
m_lenVertices = vertices.size();
m_lenIndices = indices.size();
// these Create methods should be fine as OpenGL does not give me any error
// also I have another function that requires you to pass in the vertex data and indices that works just fine
// I bind the thing I am creating
createVAO();
createVBO(vertices);
createEBO(indices);
createTexture(texture);
createShader(shader.getVertexShader(), shader.getFragmentShader());
Shape shape;
glm::mat4 model(1.0f);
glUniformMatrix4fv(glGetUniformLocation(m_shader, "model"), 1, GL_FALSE, glm::value_ptr(model));
shape.setShader(m_shader);
shape.setVAO(m_VAO);
shape.setTexture(m_texture);
shape.setPosition(position);
return shape;
}
Renderer::draw() method:
void Renderer::draw(Shape shape)
{
if (!m_usingIndices)
{
// Unbinds any other shapes
glBindVertexArray(0);
glUseProgram(0);
shape.bindShader();
shape.bindVAO();
shape.bindTexture();
glDrawArrays(GL_TRIANGLES, 0, m_lenVertices);
}
else
{
// Unbinds any other shapes
glBindVertexArray(0);
glUseProgram(0);
shape.bindShader();
shape.bindVAO();
shape.bindTexture();
glDrawElements(GL_TRIANGLES, m_lenIndices, GL_UNSIGNED_INT, 0);
}
}
Projection matrix:
glm::mat4 m_projectionMat = glm::ortho(-Window::getWidth(), Window::getWidth(), -Window::getHeight(), Window::getHeight, 0.1f, 100.0f);
Creating then rendering the Quad:
// Creates the VBO, VAO, EBO, etc.
quad = renderer.createQuad(glm::vec2(500.0f, 500.0f), glm::vec2(200.0F, 200.0f), LoadFile::loadShader("Res/Shader/VertShader.glsl", "Res/Shader/FragShader.glsl"), LoadFile::loadTexture("Res/Textures/Lake.jpg"));
// In the main game loop we render the quad
quad.setCamera(camera); // Sets the View and Projection matrix for the quad
renderer.draw(quad);
Output:
Output of the code before

openGL es 2.0 TextureCoordinates parameter ios

I have to crop a video.
So i modify the TextureCoordinates.
The code is as below.
const GLfloat squareTextureCoordinates[] = {
0.22f , 1.0f,
0.22f, 0.0f,
0.78f, 1.0f,
0.78f, 0.0f,
};
glVertexAttribPointer(positionAttribute, 2, GL_FLOAT, 0, 0, squareVertices);
glVertexAttribPointer(inputTextureCoordinateAttribute, 2, GL_FLOAT, 0, 0, squareTextureCoordinates);
This code is work.
However,i need to support many size of iphone .(example:iphone6 have different size to iphone5)
So i use variable rather than const in TextureCoordinates.
And the new code is as below;
const float cropX=(float)(SCREEN_HEIGHT-SCREEN_WIDTH)/(float)(SCREEN_HEIGHT*2);
const float cropXR=1- (float)(SCREEN_HEIGHT-SCREEN_WIDTH)/(float)(SCREEN_HEIGHT*2);
const GLfloat squareTextureCoordinates[] = {
cropX , 1.0f,
cropX, 0.0f,
cropXR, 1.0f,
cropXR, 0.0f,
};
glVertexAttribPointer(positionAttribute, 2, GL_FLOAT, 0, 0, squareVertices);
glVertexAttribPointer(inputTextureCoordinateAttribute, 2, GL_FLOAT, 0, 0, squareTextureCoordinates);
However, it doesn't work.
The screen is display nothing.
i cant figure out why i can't use const variable for TextureCoordinates.
I think you could test the second part of your code by changing the const values to the constants you used in the first part:
const float cropX = 0.22f;
const float cropXR = 0.78f;
const GLfloat squareTextureCoordinates[] = {
cropX , 1.0f,
cropX, 0.0f,
cropXR, 1.0f,
cropXR, 0.0f,
};
glVertexAttribPointer(positionAttribute, 2, GL_FLOAT, 0, 0, squareVertices);
glVertexAttribPointer(inputTextureCoordinateAttribute, 2, GL_FLOAT, 0, 0, squareTextureCoordinates);
If this works (it should), you will see you can use const variables as a texture coordinate. However, I reckon the wrong part of the code is this:
const float cropX=(float)(SCREEN_HEIGHT-SCREEN_WIDTH)/(float)(SCREEN_HEIGHT*2);
const float cropXR=1- (float)(SCREEN_HEIGHT-SCREEN_WIDTH)/(float)(SCREEN_HEIGHT*2);
In texture coordinates, correct values are between 0 and 1. If width is bigger than height, this code won't work. In the specific case of an iPhone 5 and 6, which have the same proportion, same texture coordinates should work if the view that renders OpenGL code was resized (if you use autolayout for that view, you will probably solve the problem automatically).

DirectX 11 LINELIST_TOPOLOGY uncompleted 2D rectangle at left-top coordinates

I'm doing something with DirectX 11 and came to a rectagle drawing (empty, non-colored), seemed simple for me at start (linelist_topology, 8 indices) but when I have it on the screen I see that my rectangle is kinda incomleted at left-top coordinate, there is a point of a background color there, the code is not complicated at all, vertices are 2D space:
SIMPLEVERTEX gvFrameVertices[4]=
{XMFLOAT3(0.0f,0.0f,1.0f),XMFLOAT2(0.0f, 0.0f),
XMFLOAT3(1.0f, 0.0f, 1.0f), XMFLOAT2(1.0f, 0.0f),
XMFLOAT3(1.0f, -1.0f, 1.0f), XMFLOAT2(1.0f, 1.0f),
XMFLOAT3(0.0f, -1.0f, 1.0f), XMFLOAT2(0.0f, 1.0f)};
indices:
WORD gvRectangularIndices[8] = { 0, 1, 1, 2, 2, 3, 3, 0 };
Shader just returns given color in constant buffer:
float4 PS_PANEL(PS_INPUT input) : SV_Target
{
return fontcolor;
}
Function code itself:
VOID rectangle(INT _left, INT _top, INT _width, INT _height, XMFLOAT4 _color)
{
XMMATRIX scale;
XMMATRIX translate;
XMMATRIX world;
scale = XMMatrixScaling( _width, _height, 1.0f );
translate = XMMatrixTranslation(_left, gvHeight - _top, 1.0f);
world = scale * translate;
gvConstantBufferData.world = XMMatrixTranspose(world);
gvConstantBufferData.index = 1.0f;
gvConstantBufferData.color = _color;
gvContext->PSSetShader(gvPanelPixelshader, NULL, 0);
gvContext->UpdateSubresource(gvConstantBuffer, 0, NULL, &gvConstantBufferData, 0, 0 );
gvContext->IASetIndexBuffer(gvLinelistIndexBuffer, DXGI_FORMAT_R16_UINT, 0);
gvContext->IASetPrimitiveTopology( D3D11_PRIMITIVE_TOPOLOGY_LINELIST );
gvContext->DrawIndexed(8, 0, 0);
gvContext->IASetIndexBuffer(gvTriangleslistIndexBuffer, DXGI_FORMAT_R16_UINT, 0);
gvContext->IASetPrimitiveTopology( D3D11_PRIMITIVE_TOPOLOGY_TRIANGLELIST );
};
gvHeight - _top - I'm using orthographic matrix for projection, so the coordinate center is at left-bottom, that why need to substract for proper Y coordinate.
gvOrthographicProjection = XMMatrixOrthographicOffCenterLH( 0.0f, gvWidth, 0.0f, gvHeight, 0.01f, 100.0f );
Do you have any idea what can cause this pointal incompletness of a rectangle in my case or I need to supply more code info (don't really want to link lines of the whole initializations cause they seem very obvious and simple for me, done for /at c++ and directx amateur level :)
Thank you:)

glUseProgram affecting more than just the VAO

I have successfully created a VAO which produces a triangle which can then be rotated with the mouse (with help from shaders).
My problem comes when I try to draw something else using the standard 'glBegin()' and 'glEnd()' functions. It draws successfully, but now, when I try to rotate the triangle the new drawing also rotates.
I know the problem is somehow fixed using the glUseProgram() function, but I'm not entirely sure why or where it should be added.
Here is my code (I've added it all but the main area of focus should be the display() and init() functions:
#include <GL/glew/glew.h>
#include <GL/freeglut.h>
#include <CoreStructures\CoreStructures.h>
#include <iostream>
#include "texture_loader.h"
#include "shader_setup.h"
using namespace std;
using namespace CoreStructures;
float theta = 0.0f;
bool mDown = false;
int mouse_x, mouse_y;
GLuint myShaderProgram;
GLuint locT; // location of "T" uniform variable in myShaderProgram
GLuint locR; // location of "R" uniform variable in myShaderProgram
GLuint sunPosVBO, sunColourVBO, sunIndicesVBO, sunVAO;
// Packed vertex arrays for the star object
// 1) Position Array - Store vertices as (x,y) pairs
static GLfloat sunVertices [] = {
-0.1f, 0.7f,
0.1f, 0.7f,
0.0f, 0.55f
};
// 2) Colour Array - Store RGB values as unsigned bytes
static GLubyte sunColors [] = {
255, 0, 0, 255,
255, 255, 0, 255,
0, 255, 0, 255
};
// 4) Index Array - Store indices to star vertices - this determines the order the vertices are to be processed
static GLubyte sunVertexIndices [] = {0, 1, 2};
void setupSunVAO(void) {
glGenVertexArrays(1, &sunVAO);
glBindVertexArray(sunVAO);
// copy star vertex position data to VBO
glGenBuffers(1, &sunPosVBO);
glBindBuffer(GL_ARRAY_BUFFER, sunPosVBO);
glBufferData(GL_ARRAY_BUFFER, sizeof(sunVertices), sunVertices, GL_STATIC_DRAW);
glVertexAttribPointer(0, 2, GL_FLOAT, GL_FALSE, 0, (const GLvoid*)0);
// copy star vertex colour data to VBO
glGenBuffers(1, &sunColourVBO);
glBindBuffer(GL_ARRAY_BUFFER, sunColourVBO);
glBufferData(GL_ARRAY_BUFFER, sizeof(sunColors), sunColors, GL_STATIC_DRAW);
glVertexAttribPointer(1, 4, GL_UNSIGNED_BYTE, GL_TRUE, 0, (const GLvoid*)0);
// enable position, colour buffer inputs
glEnableVertexAttribArray(0);
glEnableVertexAttribArray(1);
// setup star vertex index array
glGenBuffers(1, &sunIndicesVBO);
glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, sunIndicesVBO);
glBufferData(GL_ELEMENT_ARRAY_BUFFER, sizeof(sunVertexIndices), sunVertexIndices, GL_STATIC_DRAW);
glBindVertexArray(0);
}
void report_version(void) {
int majorVersion, minorVersion;
glGetIntegerv(GL_MAJOR_VERSION, &majorVersion);
glGetIntegerv(GL_MINOR_VERSION, &minorVersion);
cout << "OpenGL version " << majorVersion << "." << minorVersion << "\n\n";
}
void init(void) {
// initialise glew library
GLenum err = glewInit();
// ensure glew was initialised successfully before proceeding
if (err==GLEW_OK)
cout << "GLEW initialised okay\n";
else
cout << "GLEW could not be initialised\n";
report_version();
glClearColor(0.0, 0.0, 0.0, 0.0);
//
// setup "sun" VBO and VAO object
//
setupSunVAO();
//
// load shader program
//
myShaderProgram = setupShaders(string("Resources\\Shaders\\basic_vertex_shader.txt"), string("Resources\\Shaders\\basic_fragment_shader.txt"));
// get the index / location of the uniform variables "T" and "R" in shader program "myShaderProgram"
locT = glGetUniformLocation(myShaderProgram, "T");
locR = glGetUniformLocation(myShaderProgram, "R");
// "plug-in" shader into GPU pipeline
glUseProgram(myShaderProgram); // we're in the driving seat!!!!! Our shaders now intercept and process our vertices as part of the GPU rendering pipeline (as shown in the lecture notes)
}
// Example rendering functions - draw objects in local, or modelling coordinates
void drawSun(void) {
glBindVertexArray(sunVAO);
glDrawElements(GL_TRIANGLE_STRIP, 3, GL_UNSIGNED_BYTE, (GLvoid*)0);
}
void drawShape()
{
glColor3f(0.0f, 0.6f, 0.2f);
glBegin(GL_POLYGON);
glVertex2f(-1.0f, -1.0f); // Left
glVertex2f(-1.0f, -0.1f);
glVertex2f(-0.9f, -0.05f);
glVertex2f(-0.55f, -0.045f);
glVertex2f(-0.49f, -0.06f);
glVertex2f(-0.4f, -0.055f);
glVertex2f(-0.2f, -0.052f);
glVertex2f(0.0f, -0.02f); // Middle
glVertex2f(0.3f, -0.085f);
glVertex2f(0.5f, -0.08f);
glVertex2f(0.8f, -0.088f);
glVertex2f(1.0f, -0.1f);
glVertex2f(1.0f, -1.0f); // Right
glEnd();
}
//
//
void drawScene()
{
drawSun();
drawShape();
}
void display(void) {
glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);
// Setup translation matrix and store in T. Pass this over the the shader with the function glUniformMatrix4fv
GUMatrix4 T = GUMatrix4::translationMatrix(0.01f, 0.01f, 0.0f);
glUniformMatrix4fv(locT, 1, GL_FALSE, (GLfloat*)&T);
// Setup rotation matrix and store in R. Pass this over the the shader with the function glUniformMatrix4fv
GUMatrix4 R = GUMatrix4::rotationMatrix(0.0f, 0.0f, theta);
glUniformMatrix4fv(locR, 1, GL_FALSE, (GLfloat*)&R);
// Draw the scene (the above transformations will be applied to each vertex in the vertex shader)
drawScene();
glutSwapBuffers();
}
void mouseButtonDown(int button_id, int state, int x, int y) {
if (button_id==GLUT_LEFT_BUTTON) {
if (state==GLUT_DOWN) {
mouse_x = x;
mouse_y = y;
mDown = true;
} else if (state == GLUT_UP) {
mDown = false;
}
}
}
void mouseMove(int x, int y) {
if (mDown) {
int dx = x - mouse_x;
int dy = y - mouse_y;
float delta_theta = (float)dy * (3.142f * 0.01f);
theta += delta_theta;
mouse_x = x;
mouse_y = y;
glutPostRedisplay();
}
}
void keyDown(unsigned char key, int x, int y) {
if (key=='r') {
theta = 0.0f;
glutPostRedisplay();
}
}
int main(int argc, char **argv) {
glutInit(&argc, argv);
initCOM();
glutInitContextVersion(3, 3);
glutInitContextProfile (GLUT_COMPATIBILITY_PROFILE);
glutInitDisplayMode(GLUT_RGBA | GLUT_DEPTH | GLUT_DOUBLE);
glutInitWindowSize(800, 800);
glutInitWindowPosition(0, 0);
glutCreateWindow("Combining Transforms");
glutDisplayFunc(display);
glutKeyboardFunc(keyDown);
glutMouseFunc(mouseButtonDown);
glutMotionFunc(mouseMove);
init();
glutSetOption(GLUT_ACTION_ON_WINDOW_CLOSE, GLUT_ACTION_GLUTMAINLOOP_RETURNS);
glutMainLoop();
shutdownCOM();
return 0;
}
EDIT
I have an array of x,y vertices and am trying to draw them alongside the above code. For some reason this seems to take vertex data from the sunVAO.
Is there some kind of cache that needs to be cleared? I've searched google and I can't seem to find anyone else who has conflicting VAO and vertex arrays.
(Also, I have checked my code and the vertex data supplied in the array of vertices is correct, they're just not displayed correctly.)
Code:
static GLfloat bottomMarkerVertices[] = {
-0.045f, -0.75f,
0.045f, -0.75f,
-0.07f, -1.0f,
0.07f, -1.0f
};
glVertexPointer(2, GL_FLOAT, 0, bottomMarkerVertices);
glDrawArrays(GL_TRIANGLE_STRIP, 0, 4);
note: vertex arrays have been enabled.
Assuming you're defining your coordinates in normalized device space (suggested by the apparent absence of a projection matrix), the rendering loop needs to look a little like this:
void drawScene()
{
//update shader parameters for the sun shader if necessary
drawSun();
glUseProgram(0);
// at this point, the PROJECTION and MODELVIEW matrices are both the identity
// so the shape is expected to be in NDCs and is not to be transformed
// at all
drawShape();
glUseProgram(progForSun);
}
Note that I don't advise to mix legacy and modern OpenGL like that. The results of vertex processing triggered by drawShape() are only defined because you're using a compatibility profile context.
The two elements of your scene move together because they are both using the same transformation matrices, specificed by these lines:
// Setup translation matrix and store in T. Pass this over the the shader with the function glUniformMatrix4fv
GUMatrix4 T = GUMatrix4::translationMatrix(0.01f, 0.01f, 0.0f);
glUniformMatrix4fv(locT, 1, GL_FALSE, (GLfloat*)&T);
// Setup rotation matrix and store in R. Pass this over the the shader with the function glUniformMatrix4fv
GUMatrix4 R = GUMatrix4::rotationMatrix(0.0f, 0.0f, theta);
glUniformMatrix4fv(locR, 1, GL_FALSE, (GLfloat*)&R);
If you want drawShape() not to move with the mouse, you need to reset locR with a fixed theta value before you call it.
drawSun();
GUMatrix4 R = GUMatrix4::rotationMatrix(0.0f, 0.0f, 0.0f);
glUniformMatrix4fv(locR, 1, GL_FALSE, (GLfloat*)&R);
drawShape();

Fill in a polygon with texture/color using CCRenderTexture

I have a concave polygon. The first thing I do is to triangulate it and store all triangle points in a vertex array.
I am trying to fill that polygon with texture or color but none of the two solutions work.
Here is the code when trying to get a texture with one color:
CCTexture2D* TextureTest::textureColorWithVertices(CCPoint* vertices,
int count,
CCRect spriteRect)
{
CCRenderTexture *renderTexture = CCRenderTexture::create(spriteRect.size.width, spriteRect.size.height);
renderTexture->beginWithClear(1, 1, 1, 0);
mShaderProgram->use();
mShaderProgram->setUniformForModelViewProjectionMatrix();
ccVertex2F* verts = new ccVertex2F[count];
for( int i=0;i<count;i++) {
verts[i].x = vertices[i].x + spriteRect.size.width / 2;//*CC_CONTENT_SCALE_FACTOR();
verts[i].y = vertices[i].y + spriteRect.size.height / 2;//*CC_CONTENT_SCALE_FACTOR();
}
ccColor4F* colors = new ccColor4F[count];
ccColor4F red = ccc4f(1, 0, 0, 1);
for (int i = 0; i < count; i++)
{
colors[i] = red;
}
glDisable(GL_TEXTURE_2D);
mShaderProgram->setUniformLocationWith4f(mColorLocation, 1.0f, 1.0f, 1.0f, 1.0f);
glVertexAttribPointer(kCCVertexAttrib_Position, 2, GL_FLOAT, GL_FALSE, 0, verts);
glVertexAttribPointer(kCCVertexAttrib_Color, 4, GL_FLOAT, GL_FALSE, 0, colors);
glDrawArrays(GL_TRIANGLES, 0, count);
glEnable(GL_TEXTURE_2D);
CC_INCREMENT_GL_DRAWS(1);
delete[] colors;
delete[] verts;
renderTexture->end();
return renderTexture->getSprite()->getTexture();
}
And the other one with a 256*256 seamless texture:
CCTexture2D* TextureTest::textureWithVertices(CCPoint* vertices,
int count,
CCTexture2D* baseTexture,
CCRect spriteRect)
{
CCRenderTexture *renderTexture = CCRenderTexture::create(spriteRect.size.width, spriteRect.size.height);
renderTexture->beginWithClear(1, 1, 1, 0);
mShaderProgram->use();
mShaderProgram->setUniformForModelViewProjectionMatrix();
ccVertex2F* verts = new ccVertex2F[count];
for( int i=0;i<count;i++)
{
verts[i].x = vertices[i].x + spriteRect.size.width / 2;//*CC_CONTENT_SCALE_FACTOR();
verts[i].y = vertices[i].y + spriteRect.size.height / 2;//*CC_CONTENT_SCALE_FACTOR();
}
ccTex2F* uvs = new ccTex2F[count];
float baseTexturePixelsWide = baseTexture->getPixelsWide();
float baseTexturePixelsHigh = baseTexture->getPixelsHigh();
for( int i=0;i<count;i++)
{
uvs[i].u = ((spriteRect.size.width / 2 + vertices[i].x) / baseTexturePixelsWide) * CC_CONTENT_SCALE_FACTOR();
uvs[i].v = ((spriteRect.size.height / 2 - vertices[i].y) / baseTexturePixelsHigh) * CC_CONTENT_SCALE_FACTOR();
}
ccGLBindTexture2D( baseTexture->getName() );
mShaderProgram->setUniformLocationWith4f(mColorLocation, 1.0f, 1.0f, 1.0f, 1.0f);
glVertexAttribPointer(kCCVertexAttrib_TexCoords, 2, GL_FLOAT, GL_FALSE, 0, uvs);
glVertexAttribPointer(kCCVertexAttrib_Position, 2, GL_FLOAT, GL_FALSE, 0, verts);
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_REPEAT);
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_REPEAT);
glDrawArrays(GL_TRIANGLES, 0, count);
CC_INCREMENT_GL_DRAWS(1);
delete[] uvs;
delete[] verts;
renderTexture->end();
return renderTexture->getSprite()->getTexture();
}
Finally I call these functions in this piece of code, rockPolygon is the original sprite:
//Get the seamless square texture
CCSpriteBatchNode* rockBatch = CCSpriteBatchNode::create("rock_small.png", kDefaultSpriteBatchCapacity);
//Triangulate the old sprite
vector<CCPoint> triangles = triangulateSprite(rockPolygon);
//CCTexture2D* rockTexture = textureColorWithVertices(&triangles[0], triangles.size(), rockPolygon->getOriginalRect());
CCTexture2D* rockTexture = textureWithVertices(&triangles[0], triangles.size(), rockBatch->getTexture(), rockPolygon->getOriginalRect());
CCSprite* genSprite = new CCSprite();
genSprite->initWithTexture(rockTexture);
genSprite->setPosition(ccp(200,200));
addChild(genSprite); //add generated sprite to main layer
I checked the triangulation, it works fine. I also remove all code between renderTexture->beginWithClear(1, 1, 1, 0); and renderTexture->end(), set the alpha channel to 1 beginWithClear(1, 1, 1, 1) and a white square is correctly displaying. So I guess the problem comes from the opengl calls but I can't find what is wrong. Do you have any idea on how to solve this issue please?
You're already on cocos2d 2.x. Your triangulation to convex shapes works. Then you can draw the shapes with CCDrawNode, no need to write your own render code.