neercs: use gl_VertexID instead of sending a whole new vertex array.

neercs/video/text-render.cpp

@@ -44,18 +44,15 @@ void TextRender::Init()
                          ivec2(256, 256), ivec2(1));
 
     m_shader = Shader::Create(lolfx_text);
-    m_coord = m_shader->GetAttribLocation("in_Position",
-                                          VertexUsage::Position, 0);
     m_color = m_shader->GetAttribLocation("in_Attr",
                                           VertexUsage::Color, 0);
     m_char = m_shader->GetAttribLocation("in_Char",
                                          VertexUsage::Color, 1);
     m_texture = m_shader->GetUniformLocation("u_Texture");
     m_transform = m_shader->GetUniformLocation("u_Transform");
-    m_pointsize = m_shader->GetUniformLocation("u_PointSize");
+    m_datasize = m_shader->GetUniformLocation("u_DataSize");
     m_vdecl
-      = new VertexDeclaration(VertexStream<vec2>(VertexUsage::Position),
-                              VertexStream<uint32_t>(VertexUsage::Color),
+      = new VertexDeclaration(VertexStream<uint32_t>(VertexUsage::Color),
                               VertexStream<uint32_t>(VertexUsage::Color));
 
     CreateBuffers();
@@ -63,13 +60,6 @@ void TextRender::Init()
 
 void TextRender::CreateBuffers()
 {
-    m_vbo1 = new VertexBuffer(m_cells * sizeof(vec2));
-    vec2 *vertices = (vec2 *)m_vbo1->Lock(0, 0);
-    for (int j = 0; j < m_canvas_size.y; j++)
-    for (int i = 0; i < m_canvas_size.x; i++)
-        vertices[j * m_canvas_size.x + i] = vec2(i, j);
-    m_vbo1->Unlock();
-
     m_vbo2 = new VertexBuffer(m_cells * sizeof(int32_t));
     m_vbo3 = new VertexBuffer(m_cells * sizeof(int32_t));
 
@@ -83,7 +73,6 @@ void TextRender::Render()
                        caca_get_canvas_height(m_caca));
     if (current_size != m_canvas_size)
     {
-        delete m_vbo1;
         delete m_vbo2;
         delete m_vbo3;
         delete m_fbo;
@@ -140,8 +129,8 @@ void TextRender::Render()
     m_font->Bind();
     m_shader->SetUniform(m_texture, 0);
     m_shader->SetUniform(m_transform, xform);
-    m_shader->SetUniform(m_pointsize, (float)max(m_font_size.x, m_font_size.y));
-    m_vdecl->SetStream(m_vbo1, m_coord);
+    m_shader->SetUniform(m_datasize, vec2(m_canvas_size.x,
+                                          max(m_font_size.x, m_font_size.y)));
     m_vdecl->SetStream(m_vbo2, m_color);
     m_vdecl->SetStream(m_vbo3, m_char);
     m_vdecl->Bind();

neercs/video/text-render.h

@@ -22,10 +22,10 @@ private:
     TileSet *m_font;
 
     Shader *m_shader;
-    ShaderAttrib m_coord, m_color, m_char;
-    ShaderUniform m_texture, m_transform, m_pointsize;
+    ShaderAttrib m_color, m_char;
+    ShaderUniform m_texture, m_transform, m_datasize;
     VertexDeclaration *m_vdecl;
-    VertexBuffer *m_vbo1, *m_vbo2, *m_vbo3;
+    VertexBuffer *m_vbo2, *m_vbo3;
 
     FrameBuffer *m_fbo;
 };

neercs/video/text.lolfx

@@ -4,7 +4,6 @@
 
 #define HAVE_UINT 1
 
-attribute vec2 in_Position;
 #if HAVE_UINT
 attribute uint in_Char, in_Attr;
 #else
@@ -15,7 +14,7 @@ varying vec4 pass_Foreground;
 varying vec4 pass_Background;
 varying vec2 pass_UV;
 
-uniform float u_PointSize;
+uniform vec2 u_DataSize;
 uniform mat4 u_Transform;
 
 void main()
@@ -57,9 +56,11 @@ void main()
     if (F + G + H < 0.01) E = 1.0;
 
     // This only works with glEnable(GL_VERTEX_PROGRAM_POINT_SIZE);
-    gl_PointSize = u_PointSize;
+    gl_PointSize = u_DataSize[1];
 
-    gl_Position = u_Transform * vec4(in_Position, 0.0, 1.0);
+    vec2 coord = vec2(gl_VertexID % int(u_DataSize[0]),
+                      gl_VertexID / int(u_DataSize[0]));
+    gl_Position = u_Transform * vec4(coord, 0.0, 1.0);
 }
 
 [frag.glsl]