- Start work on drawlist

- Fix issue in texloader
- add ivec2 to tecloader / screen
- add draw func for iron
- add bufCfg in 3 variants to c3d
- add poc for c3d_permutation

CMakeLists.txt

@@ -19,7 +19,8 @@ add_library(${PROJECT_NAME} STATIC
     source/utils.cpp
     source/c3d.cpp
     source/ctru.cpp
-    source/iron.cpp
+    source/iron/iron.cpp
+    source/iron/drawlist.cpp
     source/maths/mat.cpp
 )
 target_include_directories(${PROJECT_NAME} PUBLIC include)

include/amethyst/c3d.hpp

@@ -5,6 +5,7 @@
 
 #include <amethyst/asset.hpp>
 #include <amethyst/maths/mat.hpp>
+#include <amethyst/maths/vec.hpp>
 #include <amethyst/types.hpp>
 #include <string>
 
@@ -26,6 +27,7 @@ class c3d {
 
     int width() const { return m_width; }
     int height() const { return m_height; }
+    ivec2 size() const { return ivec2(m_width, m_height); }
     void clear() { C3D_RenderTargetClear(m_target, C3D_CLEAR_ALL, 0, 0); }
     void startDraw() { C3D_FrameDrawOn(m_target); }
 
@@ -83,5 +85,57 @@ class c3d {
                          GPU_Primitive_t prim = GPU_TRIANGLES);
   static void drawElements(int count, const void* idx_ptr, int type = GPU_SHORT,
                            GPU_Primitive_t prim = GPU_TRIANGLES);
+  static void depthTest(bool on, GPU_TESTFUNC func = GPU_GREATER,
+                        GPU_WRITEMASK mask = GPU_WRITE_ALL);
+  static void disableScissor();
+  static void enableScissor(const ivec4 rect);
+  /**
+   * Buf cfg die permutation at runtime berechnet
+   */
+  static void bufCfg(void* ptr, int stride, int shader_attribs);
+  /**
+   * Klassische config bei der man selber die permutation eintragen muss
+   */
+  static void bufCfg(void* ptr, int stride, int shader_attribs,
+                     u64 permutation);
+  /**
+   * Hacky funktion um die permutation automatisch at compile time zu berechnen,
+   * falls diese immer gleich bleibt.
+   * In der <> steht die anzahl der shader input werte
+   * usage: c3d::bufCfg<3>(data, sizeof(vertex));
+   */
+  template <int attribs>
+  constexpr static void bufCfg(void* ptr, int stride) {
+    auto buf = C3D_GetBufInfo();
+    BufInfo_Init(buf);
+    constexpr int pm = permutation(attribs);
+    BufInfo_Add(buf, ptr, stride, attribs, pm);
+  }
+
+  static int drawcalls() { return m_drawcalls; }
+
+ private:
+  static int m_drawcalls;
+  static int m__dc__;
+
+  /**
+   * Funktion die anhand von **ac** genau den permu station wert für BufInfo
+   * ausrechnet wie z.B
+   * ```
+   * ac = 3 -> 0x210
+   * ac = 4 -> 0x3210
+   * ```
+   */
+  constexpr static u64 permutation(int ac) {
+    u64 ret = 0;
+    if (ac < 1 || ac > 15) {
+      throw std::runtime_error("[amy] " + std::to_string(ac) +
+                               " is out of range (1...15)!");
+    }
+    for (int i = 0; i < ac; i++) {
+      ret = (ret << 4) | (ac - 1 - i);
+    }
+    return ret;
+  }
 };
 }  // namespace amy

include/amethyst/iron.hpp

@@ -26,6 +26,7 @@ class iron {
     u32 color = 0;
   };
   class command {
+   public:
     command() = default;
     using ref = up<command>;
     command& add(const u16& idx) {
@@ -58,7 +59,7 @@ class iron {
 
     void merge(drawlist* list);
     command::ref newCommand();
-    void push(commad* cmd);
+    void push(command ::ref cmd);
     void clear();
 
     void drawSolid();
@@ -74,7 +75,7 @@ class iron {
   static void init();
   static void newFrame();
   static void drawOn(c3d::screen* screen);
-  static void draw(const std::vector<command>& data);
+  static void draw(const std::vector<command::ref>& data);
 
  private:
   static void setupShader();

include/amethyst/texture.hpp

@@ -4,6 +4,7 @@
 
 #include <amethyst/asset.hpp>
 #include <amethyst/image.hpp>
+#include <amethyst/types.hpp>
 
 namespace amy {
 class texture : public asset {
@@ -14,17 +15,20 @@ class texture : public asset {
   void load(cstr& path);
   void unload();
 
-  int w() const { return m_w; }
-  int& w() { return m_w; }
-  int h() const { return m_h; }
-  int& h() { return m_h; }
+  int w() const { return m_size.x; }
+  int& w() { return m_size.x; }
+  int h() const { return m_size.y; }
+  int& h() { return m_size.y; }
+  ivec2 size() const { return m_size; }
+  ivec2& size() { return m_size; }
 
-  C3D_Tex* getTex() { return m_loaded ? &m_tex : nullptr; }
+  C3D_Tex* ptr() { return m_loaded ? &m_tex : nullptr; }
+
+  void bind(int reg = 0);
 
  private:
   C3D_Tex m_tex;
-  int m_w = 0;
-  int m_h = 0;
+  ivec2 m_size;
   bool m_loaded = false;
 };
 }  // namespace amy

include/amethyst/types.hpp

@@ -1,7 +1,10 @@
 #pragma once
 
+#include <amethyst/maths/vec.hpp>
+#include <cinttypes>
 #include <functional>
 #include <map>
+#include <memory>
 #include <stdexcept>
 #include <string>
 #include <vector>

poc/CMakeLists.txt

@@ -0,0 +1,3 @@
+cmake_minimum_required(VERSION 3.22)
+
+add_subdirectory(c3d_permutation)

poc/README.md

@@ -0,0 +1,3 @@
+# Proof of concepts
+
+This folder contains some proof of concept tests.

poc/c3d_permutation/CMakeLists.txt

@@ -0,0 +1,8 @@
+cmake_minimum_required(VERSION 3.22)
+
+project(c3d-permutation)
+
+set(CMAKE_CXX_STANDARD 20)
+set(CMAKE_CXX_STANDARD_REQUIRED true)
+
+add_executable(c3d-permutation source/main.cpp)

poc/c3d_permutation/README.md

@@ -0,0 +1,10 @@
+# Citro3D Shader Buf info Permutation calculation
+
+```cpp
+// Why do 
+BufInfo_Add(buf, data, sizeof(data[0]), 3, 0x210);
+BufInfo_Add(buf, data, sizeof(data[0]), 4, 0x3210);
+// if we can do 
+BufInfo_Add(buf, data, sizeof(data[0]), 3, permutation(3));
+BufInfo_Add(buf, data, sizeof(data[0]), 3, permutation(4));
+```

poc/c3d_permutation/source/main.cpp

@@ -0,0 +1,30 @@
+#include <format>
+#include <iostream>
+#include <stdexcept>
+
+using u64 = unsigned long long;
+
+constexpr u64 permutation(int ac) {
+  u64 ret = 0;
+  if (ac < 1 || ac > 15) {
+    throw std::runtime_error("[amy] " + std::to_string(ac) +
+                             " is out of range (1...15)!");
+  }
+  for (int i = 0; i < ac; i++) {
+    ret = (ret << 4) | (ac - 1 - i);
+  }
+  return ret;
+}
+
+int main(int argc, char** argv) {
+  int v = 3;
+  if (argc == 2) {
+    v = std::stoi(argv[1]);
+  } else {
+    std::cout << "No input provided! using example..." << std::endl;
+  }
+  std::cout << std::format("{} -> {:#x} ({})", v, permutation(v),
+                           permutation(v))
+            << std::endl;
+  return 0;
+}

source/c3d.cpp

@@ -12,6 +12,8 @@ const auto DISPLAY_TRANSFER_FLAGS =
     GX_TRANSFER_OUT_FORMAT(GX_TRANSFER_FMT_RGB8) |
     GX_TRANSFER_SCALING(GX_TRANSFER_SCALE_NO);
 C3D_TexEnv* c3d::frag::m_env = nullptr;
+int c3d::m_drawcalls = 0;
+int c3d::m__dc__ = 0;
 
 void c3d::init() { C3D_Init(C3D_DEFAULT_CMDBUF_SIZE); }
 
@@ -19,9 +21,13 @@ void c3d::deinit() { C3D_Fini(); }
 
 void c3d::startFrame(bool sync) {
   C3D_FrameBegin(sync ? C3D_FRAME_SYNCDRAW : C3D_FRAME_NONBLOCK);
+  m__dc__ = 0;
 }
 
-void c3d::endFrame() { C3D_FrameEnd(0); }
+void c3d::endFrame() {
+  C3D_FrameEnd(0);
+  m_drawcalls = m__dc__;
+}
 
 c3d::screen* c3d::createScreen(gfxScreen_t screen, gfx3dSide_t side) {
   auto t = C3D_RenderTargetCreate(240, screen == GFX_TOP ? 400 : 320,
@@ -39,7 +45,8 @@ c3d::shader::~shader() {}
 void c3d::shader::load(const std::string& path) {
   auto code = utils::loadFile2Mem(path);
   if (!code.size()) {
-    throw std::runtime_error("[amy] shader: unable to load " + path);
+    throw std::runtime_error(
+        std::format("[amy] unsable to load shader ({})", path));
   }
   load(code);
 }
@@ -58,6 +65,10 @@ void c3d::shader::compile(const std::string& code) {
 }
 
 void c3d::shader::use() {
+  C3D_BindProgram(&m_program);
+  // for some reason i need both ???
+  // code works perfectly without C3D_BindProgram
+  // but nor withour shaderProgramUse ...
   shaderProgramUse(&m_program);
   C3D_SetAttrInfo(&m_info);
 }
@@ -92,10 +103,32 @@ void c3d::frag::edit(int id) {
 
 void c3d::drawArrays(int start, int count, GPU_Primitive_t prim) {
   C3D_DrawArrays(prim, start, count);
+  m__dc__++;
 }
 
 void c3d::drawElements(int count, const void* idx_ptr, int type,
                        GPU_Primitive_t prim) {
   C3D_DrawElements(prim, count, type, idx_ptr);
+  m__dc__++;
+}
+
+void c3d::depthTest(bool on, GPU_TESTFUNC func, GPU_WRITEMASK mask) {
+  C3D_DepthTest(on, func, mask);
+}
+
+void c3d::disableScissor() { C3D_SetScissor(GPU_SCISSOR_DISABLE, 0, 0, 0, 0); }
+
+void c3d::enableScissor(const ivec4 rect) {
+  C3D_SetScissor(GPU_SCISSOR_NORMAL, rect.x, rect.y, rect.z, rect.w);
+}
+
+void c3d::bufCfg(void* ptr, int stride, int shader_attribs, u64 permutation) {
+  auto buf = C3D_GetBufInfo();
+  BufInfo_Init(buf);
+  BufInfo_Add(buf, ptr, stride, shader_attribs, permutation);
+}
+
+void c3d::bufCfg(void* ptr, int stride, int shader_attribs) {
+  bufCfg(ptr, stride, shader_attribs, permutation(shader_attribs));
 }
 }  // namespace amy

source/iron/drawlist.cpp

@@ -0,0 +1,5 @@
+#include <amethyst/iron.hpp>
+
+namespace amy {
+void iron::drawlist::merge(iron::drawlist* list) {}
+}  // namespace amy

source/iron/iron.cpp

@@ -60,6 +60,41 @@ void iron::drawOn(c3d::screen* screen) {
   m_shader->setMat4(uLocProj, m_mtx);
 }
 
+void iron::draw(const std::vector<iron::command::ref>& data) {
+  // disable depthtest cause we have no z buffer
+  c3d::depthTest(false);
+  fragConfig();
+  size_t i = 0;
+  while (i < data.size()) {
+    texture* tex = data[i]->tex;
+    if (!tex) {
+      i++;
+      continue;
+    }
+    auto scissorOn = data[i]->scissorOn;
+    auto scissor = data[i]->scissorRect;
+    auto start = i;
+    // Loop until a statgechange and copy all data into vertex/index buf
+    while (i < data.size() && scissorOn == data[i]->scissorOn &&
+           scissor == data[i]->scissorRect && tex == data[i]->tex) {
+      auto c = data[i].get();
+      for (int j = 0; j < c->indexBuf.size(); j++) {
+        m_ibuf[m_idx++] = m_vtx + c->indexBuf[i];
+      }
+      for (int j = 0; j < c->vertexBuf.size(); j++) {
+        m_vbuf[m_vtx++] = c->vertexBuf[i];
+      }
+      i++;
+    }
+    ///// SCISSOR LOGIC BEG /////
+    ///// SCISSOR LOGIC END /////
+    tex->bind();
+    c3d::bufCfg<3>(m_vbuf.data(), sizeof(vertex));
+    c3d::drawElements(i - start, m_ibuf.data() + start);
+  }
+  c3d::depthTest(true);
+}
+
 void iron::setupShader() {
   m_shader = new c3d::shader();
   m_shader->compile(__ironshader__);

source/texture.cpp

@@ -51,24 +51,24 @@ void texture::load(cstr& path) {
   }
 
   int bpp = img.bpp();
-  m_w = img.width();
-  if (utils::isSingleBitNum(m_w)) {
-    m_w = utils::nextPow2(m_w);
+  m_size.x = img.width();
+  if (utils::isSingleBitNum(m_size.x)) {
+    m_size.x = utils::nextPow2(m_size.x);
   }
-  m_h = img.width();
-  if (utils::isSingleBitNum(m_h)) {
-    m_h = utils::nextPow2(m_h);
+  m_size.y = img.height();
+  if (utils::isSingleBitNum(m_size.y)) {
+    m_size.y = utils::nextPow2(m_size.y);
   }
   auto filter = GPU_NEAREST;
   auto format = image2TexFmt(img.fmt());
-  C3D_TexInit(&m_tex, (u16)m_w, (u16)m_h, format);
+  C3D_TexInit(&m_tex, (u16)m_size.x, (u16)m_size.y, format);
   C3D_TexSetFilter(&m_tex, filter, filter);
   // Using std::fill_n instead cause i hate this error lines
   // under the memset func in my editor
   std::fill_n((unsigned char*)m_tex.data, m_tex.size, 0);
   for (int x = 0; x < img.width(); x++) {
     for (int y = 0; y < img.height(); y++) {
-      int dst_pos = tile3dsTex(x, y, m_w) * bpp;
+      int dst_pos = tile3dsTex(x, y, m_size.x) * bpp;
       int src_pos = (y * img.width() + x) * bpp;
       /// Best idea i had
       for (int i = 0; i < bpp; i++) {
@@ -81,4 +81,6 @@ void texture::load(cstr& path) {
   C3D_TexSetWrap(&m_tex, GPU_REPEAT, GPU_REPEAT);
   m_loaded = true;
 }
+
+void texture::bind(int reg) { C3D_TexBind(reg, &m_tex); }
 }  // namespace amy