Merge branch 'main' into develop

* main: (91 commits)
  igl | opengl | fix glPushDebugGroup() & glDebugMessageInsert() miss the last letter in the label. (facebook#142)
  Fix shell/windows/opengles:appWindows build
  Fix inconsistent formats between VkImageView and VkImage
  igl | vulkan - fix usage of createWithExportMemory
  igl - fix cmake for hw buffers support
  The DeviceFeatureLimits::BufferAlignment is not always equals 16 byte on OpenGL backend. (facebook#113)
  Suppress swiftshader's TSAN false-positives
  Remove OpenGL dependency from Vulkan tests
  Disable debug names on Android due to emulator crash
  igl | vulkan - scope guard for locked hw buffer
  igl | vulkan - external hw buffer support facebook#2
  igl | vulkan - external hw buffer support facebook#1
  igl - native hw buffer desc helper
  igl | vulkan - native hw buffer
  igl | vulkan - hw native buffer external memory support
  igl - native hw buffer interface
  add multiviewmultisample devicefeature check
  Assert even if we don't terminate to allow debug break on validation error
  Fixed shell tests compilation
  igl | shell | Demonstrate both YUV formats `NV12` and `420p`.
  ...

CMakeLists.txt

@@ -54,6 +54,9 @@ if(ANDROID)
   set(IGL_WITH_OPENGL OFF)
   set(IGL_WITH_VULKAN ON)
   set(IGL_WITH_WEBGL OFF)
+  # disable for all targets due to warnings in third-party code
+  add_definitions(-Wno-nullability-completeness)
+  add_definitions(-Wno-deprecated-volatile)
 endif()
 
 if(EMSCRIPTEN)
@@ -263,7 +266,7 @@ if(IGL_WITH_SAMPLES)
       add_subdirectory(samples/desktop)
       igl_set_folder(glfw "third-party/GLFW3")
     endif()
-    endif()
+  endif()
 endif()
 
 if (IGL_WITH_VULKAN OR IGL_WITH_IGLU OR IGL_WITH_SAMPLES)

IGLU/CMakeLists.txt

@@ -32,6 +32,7 @@ add_iglu_module(managedUniformBuffer)
 add_iglu_module(sentinel)
 add_iglu_module(simple_renderer)
 add_iglu_module(state_pool)
+add_iglu_module(shaderCross)
 add_iglu_module(texture_accessor)
 add_iglu_module(texture_loader)
 add_iglu_module(uniform)
@@ -63,3 +64,28 @@ if(UNIX)
     target_compile_options(IGLUimgui PUBLIC "-Wno-volatile")
   endif()
 endif()
+
+# SPIRV-Cross
+# cmake-format: off
+set(SPIRV_CROSS_SHARED         OFF CACHE BOOL "")
+set(SPIRV_CROSS_STATIC         ON  CACHE BOOL "")
+set(SPIRV_CROSS_CLI            OFF CACHE BOOL "")
+set(SPIRV_CROSS_ENABLE_TESTS   OFF CACHE BOOL "")
+set(SPIRV_CROSS_ENABLE_GLSL    ON  CACHE BOOL "")
+set(SPIRV_CROSS_ENABLE_HLSL    OFF CACHE BOOL "")
+set(SPIRV_CROSS_ENABLE_MSL     ON  CACHE BOOL "")
+set(SPIRV_CROSS_ENABLE_CPP     OFF CACHE BOOL "")
+set(SPIRV_CROSS_ENABLE_REFLECT OFF CACHE BOOL "")
+set(SPIRV_CROSS_SKIP_INSTALL   ON  CACHE BOOL "")
+
+add_subdirectory(${IGL_ROOT_DIR}/third-party/deps/src/SPIRV-Cross "SPIRV-Cross")
+
+igl_set_folder(spirv-cross-core "third-party/spirv-cross/spirv-cross-core")
+igl_set_folder(spirv-cross-glsl "third-party/spirv-cross/spirv-cross-glsl")
+igl_set_folder(spirv-cross-msl  "third-party/spirv-cross/spirv-cross-msl")
+igl_set_folder(spirv-cross-util "third-party/spirv-cross/spirv-cross-util")
+igl_set_folder(spirv-cross-c    "third-party/spirv-cross/spirv-cross-c")
+# cmake-format: on
+
+target_include_directories(IGLUshaderCross PRIVATE "${IGL_ROOT_DIR}/third-party/deps/src/SPIRV-Cross" "${IGL_ROOT_DIR}/third-party/deps/src/glslang")
+target_link_libraries(IGLUshaderCross PRIVATE IGLGlslang IGLUmanagedUniformBuffer spirv-cross-core spirv-cross-glsl spirv-cross-msl spirv-cross-util spirv-cross-c)

IGLU/imgui/InputListener.cpp

@@ -21,8 +21,7 @@
 // is one frame old. This can be a source of problems if we have multiple input listeners and
 // depending on how they process inputs.
 
-namespace iglu {
-namespace imgui {
+namespace iglu::imgui {
 
 InputListener::InputListener(ImGuiContext* context) {
   _context = context;
@@ -67,5 +66,4 @@ void InputListener::makeCurrentContext() const {
   ImGui::SetCurrentContext(_context);
 }
 
-} // namespace imgui
-} // namespace iglu
+} // namespace iglu::imgui

IGLU/imgui/InputListener.h

@@ -13,8 +13,7 @@
 
 #include "imgui.h"
 
-namespace iglu {
-namespace imgui {
+namespace iglu::imgui {
 
 class InputListener : public igl::shell::IMouseListener, public igl::shell::ITouchListener {
  public:
@@ -33,5 +32,4 @@ class InputListener : public igl::shell::IMouseListener, public igl::shell::ITou
   void makeCurrentContext() const;
 };
 
-} // namespace imgui
-} // namespace iglu
+} // namespace iglu::imgui

IGLU/imgui/Session.cpp

@@ -13,8 +13,7 @@
 #include <IGLU/simple_renderer/Material.h>
 #include <igl/ShaderCreator.h>
 
-namespace iglu {
-namespace imgui {
+namespace iglu::imgui {
 
 /* internal renderer -- based on imgui_impl_metal.mm */
 
@@ -156,8 +155,8 @@ static std::unique_ptr<igl::IShaderStages> getShaderStagesForBackend(igl::IDevic
   }
   case igl::BackendType::OpenGL: {
     auto shaderVersion = device.getShaderVersion();
-    std::string vertexStr = getOpenGLVertexShaderSource(shaderVersion);
-    std::string fragmentStr = getOpenGLFragmentShaderSource(shaderVersion);
+    const std::string vertexStr = getOpenGLVertexShaderSource(shaderVersion);
+    const std::string fragmentStr = getOpenGLFragmentShaderSource(shaderVersion);
     return igl::ShaderStagesCreator::fromModuleStringInput(
         device, vertexStr.c_str(), "main", "", fragmentStr.c_str(), "main", "", &result);
     break;
@@ -203,7 +202,7 @@ struct DrawableData {
 
 class Session::Renderer {
  public:
-  Renderer(igl::IDevice& device);
+  explicit Renderer(igl::IDevice& device);
   ~Renderer();
 
   void newFrame(const igl::FramebufferDesc& desc);
@@ -275,7 +274,7 @@ Session::Renderer::Renderer(igl::IDevice& device) {
 }
 
 Session::Renderer::~Renderer() {
-  ImGuiIO& io = ImGui::GetIO();
+  const ImGuiIO& io = ImGui::GetIO();
   _fontTexture = nullptr;
   io.Fonts->TexID = nullptr;
 }
@@ -299,15 +298,15 @@ void Session::Renderer::renderDrawData(igl::IDevice& device,
                                        ImDrawData* drawData) {
   // Avoid rendering when minimized, scale coordinates for retina displays (screen coordinates !=
   // framebuffer coordinates)
-  int fb_width = (int)(drawData->DisplaySize.x * drawData->FramebufferScale.x);
-  int fb_height = (int)(drawData->DisplaySize.y * drawData->FramebufferScale.y);
+  const int fb_width = (int)(drawData->DisplaySize.x * drawData->FramebufferScale.x);
+  const int fb_height = (int)(drawData->DisplaySize.y * drawData->FramebufferScale.y);
   if (fb_width <= 0 || fb_height <= 0 || drawData->CmdListsCount == 0) {
     return;
   }
 
   cmdEncoder.pushDebugGroupLabel("ImGui Rendering", igl::Color(0, 1, 0));
 
-  igl::Viewport viewport = {
+  const igl::Viewport viewport = {
       /*.x = */ 0.0,
       /*.y = */ 0.0,
       /*.width = */ (drawData->DisplaySize.x * drawData->FramebufferScale.x),
@@ -319,10 +318,10 @@ void Session::Renderer::renderDrawData(igl::IDevice& device,
   float4x4 orthoProjection{};
 
   { // setup projection matrix
-    float L = drawData->DisplayPos.x;
-    float R = drawData->DisplayPos.x + drawData->DisplaySize.x;
-    float T = drawData->DisplayPos.y;
-    float B = drawData->DisplayPos.y + drawData->DisplaySize.y;
+    const float L = drawData->DisplayPos.x;
+    const float R = drawData->DisplayPos.x + drawData->DisplaySize.x;
+    const float T = drawData->DisplayPos.y;
+    const float B = drawData->DisplayPos.y + drawData->DisplaySize.y;
     orthoProjection.columns[0] = float4{2.0f / (R - L), 0.0f, 0.0f, 0.0f};
     orthoProjection.columns[1] = float4{0.0f, 2.0f / (T - B), 0.0f, 0.0f};
     orthoProjection.columns[2] = float4{0.0f, 0.0f, -1.0f, 0.0f};
@@ -333,8 +332,8 @@ void Session::Renderer::renderDrawData(igl::IDevice& device,
     }
   }
 
-  ImVec2 clip_off = drawData->DisplayPos; // (0,0) unless using multi-viewports
-  ImVec2 clip_scale =
+  const ImVec2 clip_off = drawData->DisplayPos; // (0,0) unless using multi-viewports
+  const ImVec2 clip_scale =
       drawData->FramebufferScale; // (1,1) unless using retina display which are often (2,2)
 
   // Since vertex buffers are updated every frame, we must use triple buffering for Metal to work
@@ -352,7 +351,7 @@ void Session::Renderer::renderDrawData(igl::IDevice& device,
     if (n >= curFrameDrawables.size()) {
       curFrameDrawables.emplace_back(device, _vertexInputState, _material);
     }
-    DrawableData& drawableData = curFrameDrawables[n];
+    const DrawableData& drawableData = curFrameDrawables[n];
 
     // Upload vertex/index buffers
     drawableData.vertexData->vertexBuffer().upload(
@@ -421,7 +420,7 @@ void Session::Renderer::renderDrawData(igl::IDevice& device,
 Session::Session(igl::IDevice& device,
                  igl::shell::InputDispatcher& inputDispatcher,
                  bool needInitializeSession /* = true */) :
-  _inputDispatcher(inputDispatcher), _isInitialized(false) {
+  _inputDispatcher(inputDispatcher) {
   _context = ImGui::CreateContext();
   makeCurrentContext();
 
@@ -481,5 +480,4 @@ void Session::makeCurrentContext() const {
   ImGui::SetCurrentContext(_context);
 }
 
-} // namespace imgui
-} // namespace iglu
+} // namespace iglu::imgui

IGLU/imgui/Session.h

@@ -18,8 +18,7 @@
 
 #include "imgui.h"
 
-namespace iglu {
-namespace imgui {
+namespace iglu::imgui {
 
 class Session {
  public:
@@ -40,10 +39,9 @@ class Session {
   std::shared_ptr<InputListener> _inputListener;
   ImGuiContext* _context;
   std::unique_ptr<Renderer> _renderer;
-  bool _isInitialized;
+  bool _isInitialized = false;
 
   void makeCurrentContext() const;
 };
 
-} // namespace imgui
-} // namespace iglu
+} // namespace iglu::imgui

IGLU/managedUniformBuffer/ManagedUniformBuffer.cpp

@@ -208,7 +208,7 @@ bool ManagedUniformBuffer::updateData(const char* name, const void* data, size_t
       // This could mean the user knows only a portion of the uniform data needs updating
       // However, if dataSize is larger than or equal to what we expect for this uniform, we will
       // only copy data up to the expected data size for this uniform
-      size_t uniformDataSize = getUniformDataSizeInternal(uniform);
+      const size_t uniformDataSize = getUniformDataSizeInternal(uniform);
       if (dataSize > uniformDataSize) {
         dataSize = uniformDataSize;
 #if IGL_DEBUG
@@ -236,9 +236,9 @@ size_t ManagedUniformBuffer::getUniformDataSize(const char* name) {
 }
 
 size_t ManagedUniformBuffer::getUniformDataSizeInternal(igl::UniformDesc& uniform) {
-  size_t uniformDataSize = uniform.elementStride != 0
-                               ? uniform.numElements * uniform.elementStride
-                               : uniform.numElements * igl::sizeForUniformType(uniform.type);
+  const size_t uniformDataSize = uniform.elementStride != 0
+                                     ? uniform.numElements * uniform.elementStride
+                                     : uniform.numElements * igl::sizeForUniformType(uniform.type);
   return uniformDataSize;
 }
 

IGLU/simdtypes/SimdTypes.h

@@ -16,8 +16,7 @@
 
 /// Polyfill to make it so we don't have to have ifdefs in code
 
-namespace iglu {
-namespace simdtypes {
+namespace iglu::simdtypes {
 
 // Use Apple-provided simd if available.
 #if defined(__APPLE__)
@@ -259,5 +258,4 @@ struct float2x2 {
 
 #endif // defined(__APPLE__)
 
-} // namespace simdtypes
-} // namespace iglu
+} // namespace iglu::simdtypes

IGLU/simdtypes/SimdUtilities.h

@@ -10,8 +10,7 @@
 #include <IGLU/simdtypes/SimdTypes.h>
 #include <cmath>
 
-namespace iglu {
-namespace simdtypes {
+namespace iglu::simdtypes {
 
 #if defined(__APPLE__)
 
@@ -54,7 +53,7 @@ inline float1 fract(float1 x) {
 }
 
 inline float4x4 inverse(const float4x4& m) {
-  float4x4 result;
+  float4x4 result{};
 
   result.columns[0][0] = m.columns[1][1] * m.columns[2][2] * m.columns[3][3] -
                          m.columns[1][1] * m.columns[2][3] * m.columns[3][2] -
@@ -177,9 +176,9 @@ inline float4x4 inverse(const float4x4& m) {
 
   det = 1.0f / det;
 
-  for (int i = 0; i < 4; i++) {
+  for (auto& column : result.columns) {
     for (int j = 0; j < 4; j++) {
-      result.columns[i][j] *= det;
+      column[j] *= det;
     }
   }
 
@@ -202,13 +201,12 @@ inline float4 multiply(const float4x4& m, const float4& v) {
 
 // result = m1 * m2
 inline float4x4 multiply(const float4x4& m1, const float4x4& m2) {
-  float4x4 result;
+  float4x4 result{};
   result.columns[0] = multiply(m1, m2.columns[0]);
   result.columns[1] = multiply(m1, m2.columns[1]);
   result.columns[2] = multiply(m1, m2.columns[2]);
   result.columns[3] = multiply(m1, m2.columns[3]);
   return result;
 }
 #endif
-} // namespace simdtypes
-} // namespace iglu
+} // namespace iglu::simdtypes

IGLU/simple_renderer/Drawable.cpp

@@ -11,8 +11,7 @@
 
 #include <utility>
 
-namespace iglu {
-namespace drawable {
+namespace iglu::drawable {
 
 Drawable::Drawable(std::shared_ptr<vertexdata::VertexData> vertexData,
                    std::shared_ptr<material::Material> material) :
@@ -24,7 +23,7 @@ void Drawable::draw(igl::IDevice& device,
                     size_t pushConstantsDataSize,
                     const void* pushConstantsData) {
   // Assumption: _vertexData and _material are immutable
-  size_t pipelineDescHash = std::hash<igl::RenderPipelineDesc>()(pipelineDesc);
+  const size_t pipelineDescHash = std::hash<igl::RenderPipelineDesc>()(pipelineDesc);
   if (!_pipelineState || pipelineDescHash != _lastPipelineDescHash) {
     igl::RenderPipelineDesc mutablePipelineDesc = pipelineDesc;
     _vertexData->populatePipelineDescriptor(mutablePipelineDesc);
@@ -45,5 +44,4 @@ void Drawable::draw(igl::IDevice& device,
   _vertexData->draw(commandEncoder);
 }
 
-} // namespace drawable
-} // namespace iglu
+} // namespace iglu::drawable

IGLU/simple_renderer/Drawable.h

@@ -13,8 +13,7 @@
 #include <IGLU/simple_renderer/VertexData.h>
 #include <memory>
 
-namespace iglu {
-namespace drawable {
+namespace iglu::drawable {
 
 /// A drawable aggregates all the data and configurations for a single draw call.
 ///
@@ -44,5 +43,4 @@ class Drawable final {
   size_t _lastPipelineDescHash = 0;
 };
 
-} // namespace drawable
-} // namespace iglu
+} // namespace iglu::drawable

IGLU/simple_renderer/ForwardRenderPass.cpp

@@ -11,11 +11,10 @@
 
 #include <utility>
 
-namespace iglu {
-namespace renderpass {
+namespace iglu::renderpass {
 
 ForwardRenderPass::ForwardRenderPass(igl::IDevice& device) {
-  igl::CommandQueueDesc desc;
+  const igl::CommandQueueDesc desc{};
   _commandQueue = device.createCommandQueue(desc, nullptr);
   _backendType = device.getBackendType();
 }
@@ -45,7 +44,7 @@ void ForwardRenderPass::begin(std::shared_ptr<igl::IFramebuffer> target,
   const igl::RenderPassDesc* finalDesc = renderPassDescOverride ? renderPassDescOverride
                                                                 : &defaultRenderPassDesc;
 
-  igl::CommandBufferDesc cbDesc;
+  const igl::CommandBufferDesc cbDesc;
   _commandBuffer = _commandQueue->createCommandBuffer(cbDesc, nullptr);
   _commandEncoder =
       _commandBuffer->createRenderCommandEncoder(*finalDesc, _framebuffer, {}, nullptr);
@@ -98,5 +97,4 @@ igl::IRenderCommandEncoder& ForwardRenderPass::activeCommandEncoder() {
   return *_commandEncoder;
 }
 
-} // namespace renderpass
-} // namespace iglu
+} // namespace iglu::renderpass

IGLU/simple_renderer/ForwardRenderPass.h

@@ -15,8 +15,7 @@
 #include <string>
 #include <vector>
 
-namespace iglu {
-namespace renderpass {
+namespace iglu::renderpass {
 
 /// A simple "render pass" abstraction that hides low level graphics API details
 /// like command queue, command encoder, render pipeline state and presentation.
@@ -62,5 +61,4 @@ class ForwardRenderPass final {
   std::unique_ptr<igl::IRenderCommandEncoder> _commandEncoder;
 };
 
-} // namespace renderpass
-} // namespace iglu
+} // namespace iglu::renderpass