Simplifica Makefile a 3 targets delegats a CMake amb release a dist/

- Removed voxel_descent.frag.glsl (incompatible with OpenGL - multiple overflow issues)
- Added cube_lines.frag.glsl by Danil (raytraced refraction cube with self-intersections)
- Updated README.md:
  - Corrected executable name (shadertoy_sdl3 → shadertoy)
  - Added features section (FPS counter, VSync, metadata, feedback system)
  - Updated keyboard shortcuts (F3, F4, arrows)
  - Added Shadertoy compatibility guide
  - Updated usage examples with correct paths
  - Added shader conversion guide from Shadertoy

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

.gitignore

@@ -5,18 +5,24 @@
 /CMakeFiles/
 /CMakeCache.txt
 /cmake_install.cmake
-/Makefile
 /*.cmake
 /*.obj
 /*.o
 /*.lo
 /*.a
 /*.lib
-/*.dll
+*.pdb
-/*.so
+
-/*.dylib
+# Binarios compilados
-/*.exe
+*.dll
-/*.pdb
+*.so
+*.dylib
+*.exe
+shadertoy
+Shadertoy
+
+# Permitir DLLs de distribucion en release/
+!release/**/*.dll
 
 # Generated by CMake/IDE
 /compile_commands.json
@@ -29,9 +35,16 @@ third_party/**/build/
 third_party/**/CMakeFiles/
 third_party/**/CMakeCache.txt
 
+# Release artifacts
+dist/
+*.zip
+*.tar.gz
+*.dmg
+
 # IDEs and editors
 .vscode/
 .idea/
+.claude/
 *.suo
 *.user
 *.userprefs
@@ -62,6 +75,3 @@ coverage.info
 venv/
 env/
 __pycache__/
-
-# Avoid committing local SDL runtime if you copy it next to the exe
-SDL3.dll

CLAUDE.md

@@ -0,0 +1,290 @@
+# CLAUDE.md
+
+This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
+
+## Project Overview
+
+This is a cross-platform Shadertoy-like fragment shader viewer built with SDL3. It supports two rendering backends:
+- **SDL3 GPU** (default): renders via Vulkan on Linux/Windows and Metal on macOS.
+- **OpenGL 3.3** (fallback): the original backend, used automatically if SDL3 GPU init fails.
+
+Selection at launch:
+```bash
+./shadertoy --backend=gpu     # force SDL3 GPU
+./shadertoy --backend=opengl  # force OpenGL
+./shadertoy --backend=auto    # default: GPU first, fall back to OpenGL on failure
+```
+
+## Build and Development Commands
+
+### Building (CMake - Development)
+```bash
+mkdir build && cd build
+cmake ..
+cmake --build . --config Release
+```
+
+### Building (Makefile - Release Packages)
+```bash
+make windows_release   # Creates .zip with DLLs
+make macos_release     # Creates .dmg with app bundle
+make linux_release     # Creates .tar.gz
+make show_version      # Display build version (YYYY-MM-DD format)
+```
+
+Platform-specific debug builds:
+```bash
+make windows_debug
+make macos_debug
+make linux_debug
+```
+
+### Compiling Shaders to SPIR-V
+The SDL3 GPU backend reads compiled SPIR-V (`.frag.spv`) at runtime. Run this after touching any `.vk.glsl` file:
+```bash
+cmake --build build --target compile_shaders
+```
+Requires `glslc` (Debian/Ubuntu: `apt install glslang-tools`; macOS: `brew install glslang`).
+
+### Running the Application
+```bash
+./shadertoy [SHADER_NAME_OR_PATH] [-F|--fullscreen] [--backend=auto|gpu|opengl]
+
+# Examples:
+./shadertoy test                            # auto backend, data/shaders/test/
+./shadertoy --backend=gpu seascape          # force SDL3 GPU (Vulkan/Metal)
+./shadertoy -F data/shaders/fractal_pyramid # explicit folder path, fullscreen
+```
+
+**Runtime Controls:**
+- `ESC` - Exit
+- `F3` - Toggle fullscreen
+- `LEFT/RIGHT ARROW` - Cycle through shaders in directory
+
+## Architecture
+
+### Layered design
+- `source/main.cpp` — SDL3 window, event loop, shader-list scanning, dispatch to a backend.
+- `source/rendering/shader_backend.hpp` — abstract `IShaderBackend` interface + `ShaderMetadata` / `ShaderUniforms` / `ShaderProgramSpec` shared types. Two factories: `makeOpenGLBackend()`, `makeSdl3GpuBackend()`.
+- `source/rendering/opengl_shader_backend.{hpp,cpp}` — OpenGL 3.3 implementation. Loads `<folder>/<name>.gl.glsl` at runtime.
+- `source/rendering/sdl3gpu/sdl3gpu_shader_backend.{hpp,cpp}` — SDL3 GPU implementation. Loads `<folder>/<name>.frag.spv` (Linux/Windows) or `<folder>/<name>.frag.msl` (macOS); shares one passthrough vertex shader from `shaders/_common/passthrough.vert.{spv,msl}`.
+- `source/rendering/sdl3gpu/shader_factory.hpp` — small helper that loads a shader binary/source from disk and creates an `SDL_GPUShader`.
+
+### Shader folder layout
+One folder per shader, under `data/shaders/`:
+```
+data/shaders/<name>/
+    <name>.gl.glsl    # OpenGL GLSL 3.30 (handwritten)
+    <name>.vk.glsl    # Vulkan GLSL 4.50 (handwritten)
+    <name>.frag.msl   # Metal Shading Language (handwritten)
+    <name>.frag.spv   # generated by `make compile_shaders` from .vk.glsl
+    meta.txt          # Name / Author / iChannelN
+data/shaders/_common/
+    passthrough.vk.glsl    # shared fullscreen-triangle vertex (Vulkan)
+    passthrough.vert.spv   # generated
+    passthrough.vert.msl   # handwritten Metal
+```
+Folders starting with `_` or `.` are skipped by the scanner. The `meta.txt` parser is in `Rendering::parseMetaFile`.
+
+### Backend selection logic (`main.cpp`)
+1. Parse `--backend=...` flag (default `auto`).
+2. If not OpenGL: create a window with no flags, instantiate `Sdl3GpuShaderBackend`. If `init()` fails AND choice was `auto`, destroy the window and continue. If choice was `gpu`, exit with error.
+3. Otherwise: create a window with `SDL_WINDOW_OPENGL` and instantiate `OpenGLShaderBackend`.
+
+### Global State (main.cpp)
+```cpp
+shader_list_           // Vector<ShaderEntry { folder, base_name }>
+current_shader_index_  // Active shader in rotation
+shader_start_ticks_    // Base time for iTime uniform calculation
+window_                // SDL3 window pointer
+backend_               // unique_ptr<IShaderBackend>
+shaders_directory_     // Shaders root path (resolved at startup)
+```
+
+### Dependencies
+- **SDL3** ≥ 3.2 — windowing, events, GPU API (`SDL_gpu.h`)
+- **GLAD** — OpenGL 3.3 loader (used only by the OpenGL backend, statically linked via `source/external/glad/`)
+- **C++17 stdlib** — filesystem, fstream, vector, algorithms
+- Build-time tool: `glslc` (only required to regenerate `.frag.spv`)
+
+### Platform-Specific Code
+Uses preprocessor defines (`WINDOWS_BUILD`, `MACOS_BUILD`, `LINUX_BUILD`) for:
+- `getExecutableDirectory()` - Windows API, mach-o dyld, or /proc/self/exe
+- `getResourcesDirectory()` - Special macOS app bundle handling (Contents/Resources)
+- Linking flags - Windows uses static linking, macOS links OpenGL framework
+
+## Shader System
+
+### Authoring a shader (manual steps)
+For each new shader, three handwritten files live in `data/shaders/<name>/`:
+1. **`<name>.gl.glsl`** — OpenGL GLSL 3.30 (the original Shadertoy-style format described below).
+2. **`<name>.vk.glsl`** — Vulkan GLSL 4.50 with explicit `layout(set=3, binding=0) uniform ShadertoyUBO { float iTime; vec2 iResolution; } u;` and `layout(set=2, binding=N) uniform sampler2D` for any texture channel.
+3. **`<name>.frag.msl`** — Metal Shading Language. Entry point must be named `<name>_fs`. Uniforms come in via `[[buffer(0)]]`; samplers via `[[texture(N)]] [[sampler(N)]]`.
+
+After editing any `.vk.glsl`, rebuild SPIR-V: `cmake --build build --target compile_shaders`.
+
+The MSL is **not** auto-generated — write it by hand following the same convention as `aee_2026/source/core/rendering/sdl3gpu/msl/*.msl.h`.
+
+### Shadertoy → Vulkan/Metal porting cheatsheet
+- `iResolution` is `vec2` here (no `.xy`). Reconstruct vec3 manually if needed.
+- Replace `texture(iChannel0, ...)` calls — no texture channels currently wired through the GPU backend (planned).
+- The shared vertex stage emits `vUV` already in Shadertoy convention `(0,0)` bottom-left → `(1,1)` top-right; multiply by `iResolution` to get `fragCoord`.
+- The Y axis is flipped in NDC inside the shared vertex shader so Vulkan/Metal render right-side-up like OpenGL.
+
+### Shader Format (GLSL 3.3 Core)
+All shaders must follow this structure:
+
+```glsl
+#version 330 core
+precision highp float;
+
+out vec4 FragColor;
+in vec2 vUV;              // Normalized [0,1] coordinates from vertex shader
+uniform vec2 iResolution; // Window resolution in pixels
+uniform float iTime;      // Time since shader loaded (seconds)
+
+// Shadertoy-style entry point
+void mainImage(out vec4 fragColor, in vec2 fragCoord) {
+    // fragCoord is in pixel coordinates
+    // Your shader code here
+}
+
+// Wrapper converts vUV to pixel coordinates
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}
+```
+
+### Adding New Shaders
+1. Create `.glsl` file in `data/shaders/<name>/` directory (use `.frag.glsl` convention)
+2. Follow required format above
+3. File automatically appears in runtime shader rotation (arrow keys to navigate)
+4. **No code changes required** - directory is scanned at startup
+
+### Shader Loading Pipeline
+1. Directory scan on startup (`scanShaderDirectory()`)
+2. Sort alphabetically
+3. Load fragment shader source from disk
+4. Compile vertex shader (fixed, embedded in main.cpp)
+5. Compile fragment shader with error logging
+6. Link program with error handling
+7. Update uniforms each frame (iResolution, iTime)
+
+### Supported Shadertoy Features
+- ✅ `iTime` - Time uniform
+- ✅ `iResolution` - Window resolution (vec2, not vec3)
+- ✅ `mainImage()` function signature
+- ❌ `iMouse` - Not implemented
+- ❌ `iChannel0-3` - No texture channels (multi-pass not supported)
+- ❌ `iFrame`, `iTimeDelta`, `iDate` - Not implemented
+
+### Converting from Shadertoy
+When porting Shadertoy shaders:
+1. Copy the `mainImage()` function as-is
+2. Add standard header (see format above)
+3. Add wrapper `main()` function
+4. Remove texture channel references (iChannel0-3)
+5. Change `iResolution.xy` to `iResolution` (this project uses vec2)
+
+### Common Conversion Issues and Solutions
+
+Based on experience converting complex shaders like ddla_light_tunnel:
+
+#### iResolution vec3 vs vec2
+- **Shadertoy:** `iResolution` is `vec3(width, height, width/height)`
+- **This project:** `iResolution` is `vec2(width, height)`
+- **Solution:** Create vec3 manually: `vec3 r = vec3(iResolution.xy, iResolution.x/iResolution.y);`
+- **Then use:** `r.xy` for resolution, `r.y` for height (as in original)
+
+#### Uninitialized Variables
+- **Problem:** Shadertoy code may have `vec3 rgb;` without initialization
+- **Shadertoy behavior:** Likely initializes to `vec3(0.0)` (black)
+- **This project:** Uninitialized variables contain undefined values (often causes black screen or wrong colors)
+- **Solution:** Always initialize: `vec3 rgb = vec3(0.0);`
+- **Wrong approach:** Don't use random noise unless shader explicitly uses iChannel texture for noise
+
+#### Low mix() Factors Are Intentional
+- **Example:** `rgb = mix(rgb, calculated_color, 0.01);` means 1% new color, 99% existing
+- **Don't change these factors** - they create subtle effects intentionally
+- **If output is black:** Problem is likely the base value (rgb), not the mix factor
+
+#### iChannel Textures
+- **Shadertoy shows iChannel0-3** in UI, but shader may not use them
+- **If iChannel3 has RGB noise but code doesn't reference it:** The noise is not actually used
+- **Check code for `texture(iChannelN, ...)` calls** - if none exist, ignore the iChannel setup
+- **Procedural noise replacement:** Only if shader explicitly samples the texture
+
+#### Color Swapping Issues
+- **If colors are completely wrong:** Don't randomly swap color variables
+- **First verify:** Code matches original Shadertoy exactly (except required GLSL changes)
+- **Common mistake:** Changing color applications that were correct in original
+- **Debug approach:** Revert to exact original code, only modify for GLSL 3.3 compatibility
+
+#### Division by Small Values - Color Overflow Artifacts
+- **Problem:** Division by values near zero causes extreme values → color overflow artifacts (green points, strange colors)
+- **Example:** `.01*vec4(6,2,1,0)/length(u*sin(iTime))` can divide by ~0.0 when near center and sin≈0
+- **Symptoms:** Bright white center that pulses to green, or random color artifacts in specific areas
+- **Root cause:** Even with tonemap (tanh/clamp), extreme intermediate values cause precision issues or saturation
+- **Solution - Double Protection:**
+  1. **Add epsilon to denominator:** `max(denominator, 0.001)` prevents division by exact zero
+  2. **Clamp the result:** `min(result, vec4(50.0))` prevents extreme accumulation
+  3. **Only clamp problematic terms:** Don't clamp everything or scene becomes dim
+- **Example fix:**
+  ```glsl
+  // Original (causes overflow):
+  o += .01*vec4(6,2,1,0)/length(u*sin(t+t+t)) + 1./s * length(u);
+
+  // Fixed (prevents overflow while maintaining brightness):
+  vec4 brightTerm = min(.01*vec4(6,2,1,0)/max(length(u*sin(t+t+t)), 0.001), vec4(50.0));
+  o += brightTerm + 1./s * length(u);
+  ```
+- **Key values:** epsilon ~0.001 (not too small, not too large), clamp ~50.0 (allows brightness without explosion)
+- **Why this works in Shadertoy:** WebGL/browsers may handle float overflow differently than native OpenGL drivers
+
+#### Debugging Black/Wrong Output
+1. **Check compilation errors first** - shader must compile without errors
+2. **Initialize all variables** - especially vec3 colors
+3. **Verify vec3 iResolution handling** - create vec3 from vec2 if needed
+4. **Don't modify mix factors** - keep original values
+5. **Compare with original code** - ensure logic is identical
+6. **Test progressive changes** - add header, test; add wrapper, test; etc.
+7. **Check for division by small values** - if you see color artifacts (especially green), look for divisions that can approach zero
+
+## Build System Details
+
+### Version Numbering
+Releases use build date format: `shadertoy-YYYY-MM-DD-{platform}` (auto-generated by Makefile)
+
+### Release Artifacts
+- **Windows:** `.zip` with `shadertoy.exe` + `SDL3.dll` + shaders
+- **macOS:** `.dmg` with app bundle containing embedded SDL3.framework (arm64 only)
+- **Linux:** `.tar.gz` with binary + shaders
+
+### Resource Bundling
+- Shaders copied to release packages
+- LICENSE and README.md included
+- Platform-specific dependencies bundled (DLLs on Windows, frameworks on macOS)
+
+## Important Notes
+
+### Vertex Shader
+The vertex shader is hardcoded in `main.cpp` and creates a fullscreen quad. It:
+- Takes `vec2 aPos` (location 0) in NDC space [-1, 1]
+- Outputs `vec2 vUV` normalized to [0, 1]
+- No transformation matrices needed
+
+### Shader Hot-Reloading
+Currently **not implemented**. Shader changes require application restart. The architecture would support adding this via file watching.
+
+### Multi-Pass Rendering
+Single-pass only. To add multi-pass (BufferA/B/C like Shadertoy):
+- Create FBOs and textures per buffer
+- Render buffers in dependency order
+- Pass textures as `iChannel0-3` uniforms
+- Use ping-pong for feedback loops
+
+### OpenGL Context
+Created via SDL3 with core profile (no deprecated functions). Context version: 3.3 core.

CMakeLists.txt

@@ -1,143 +1,120 @@
-cmake_minimum_required(VERSION 3.14)
+# CMakeLists.txt
-project(shadertoy_sdl3 LANGUAGES C CXX)
 
+cmake_minimum_required(VERSION 3.10)
+project(shadertoy VERSION 1.00)
+
+# Estàndard de C++
 set(CMAKE_CXX_STANDARD 17)
-set(CMAKE_CXX_STANDARD_REQUIRED ON)
+set(CMAKE_CXX_STANDARD_REQUIRED True)
 
-# Default to Release for single-config generators when none specified
+# Política CMP0072: usar GLVND (libOpenGL/libGLX) en lloc de la libGL clàssica.
-if(NOT CMAKE_BUILD_TYPE AND NOT CMAKE_CONFIGURATION_TYPES)
+cmake_policy(SET CMP0072 NEW)
-  set(CMAKE_BUILD_TYPE Release CACHE STRING "Build type" FORCE)
+set(OpenGL_GL_PREFERENCE GLVND)
-endif()
 
-# Paths
+# Opció per a empaquetat de bundle macOS (.app). El Makefile la passa amb -DMACOS_BUNDLE=ON.
-set(GLAD_SRC "${CMAKE_CURRENT_SOURCE_DIR}/third_party/glad/src/glad.c")
+option(MACOS_BUNDLE "Build for macOS .app bundle (rpath into ../Frameworks)" OFF)
-if(NOT EXISTS "${GLAD_SRC}")
-  message(FATAL_ERROR "glad.c no encontrado en: ${GLAD_SRC}")
-endif()
 
-# glad library
+# --- LISTA EXPLÍCITA DE FUENTES ---
-add_library(glad_src STATIC "${GLAD_SRC}")
+set(APP_SOURCES
-target_include_directories(glad_src PUBLIC
+    source/main.cpp
-  ${CMAKE_CURRENT_SOURCE_DIR}/third_party/glad/include
+    source/rendering/shader_backend.cpp
+    source/rendering/opengl_shader_backend.cpp
+    source/rendering/sdl3gpu/sdl3gpu_shader_backend.cpp
+    source/audio/jail_audio.cpp
 )
-set_target_properties(glad_src PROPERTIES LINKER_LANGUAGE C)
 
-# Executable
+# Fuentes de librerías de terceros
-add_executable(shadertoy_sdl3
+set(EXTERNAL_SOURCES
-  src/main.cpp
+    source/external/glad/src/glad.c
+    source/external/stb_vorbis_impl.cpp
 )
-target_include_directories(shadertoy_sdl3 PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/third_party/glad/include)
 
-# SDL3
+# Configuració de SDL3
-find_package(SDL3 REQUIRED)
+find_package(SDL3 REQUIRED CONFIG REQUIRED COMPONENTS SDL3)
-target_link_libraries(shadertoy_sdl3 PRIVATE glad_src SDL3::SDL3)
+message(STATUS "SDL3 trobat: ${SDL3_INCLUDE_DIRS}")
 
-# Platform-specific flags and linkers (try to mimic your Makefile)
+# --- COMPILACIÓ DE SHADERS (Vulkan SPIR-V) ---
-if(WIN32)
+# Es fa a cada build: compile_shaders.cmake fa check de timestamps, no
-  target_compile_definitions(shadertoy_sdl3 PRIVATE WINDOWS_BUILD)
+# recompila el que ja està al dia.
-  if(MSVC)
+find_program(GLSLC_EXE NAMES glslc)
-    # MSVC: reasonable defaults; adjust as needed
+if(GLSLC_EXE)
-    target_compile_options(shadertoy_sdl3 PRIVATE /W3 /permissive-)
+    message(STATUS "glslc trobat: ${GLSLC_EXE}")
-    # Link to system libraries (MSVC uses system libs automatically for system includes)
+    add_custom_target(compile_shaders ALL
-    target_link_libraries(shadertoy_sdl3 PRIVATE opengl32)
+        COMMAND ${CMAKE_COMMAND}
-  else()
+            -D GLSLC=${GLSLC_EXE}
-    # MinGW / GNU on Windows
+            -D SHADERS_DIR=${CMAKE_SOURCE_DIR}/data/shaders
-    target_compile_options(shadertoy_sdl3 PRIVATE -std=c++17 -Wall -Os -ffunction-sections -fdata-sections)
+            -P ${CMAKE_SOURCE_DIR}/tools/shaders/compile_shaders.cmake
-    target_link_options(shadertoy_sdl3 PRIVATE "-Wl,--gc-sections" "-Wl,-Bstatic" "-Wl,-Bdynamic")
+        WORKING_DIRECTORY ${CMAKE_SOURCE_DIR}
-    target_link_libraries(shadertoy_sdl3 PRIVATE -lmingw32 -lws2_32 -lSDL3 -lopengl32)
+        COMMENT "Compiling .vk.glsl shaders to SPIR-V")
-  endif()
 else()
-  # POSIX platforms
+    message(STATUS "glslc no trobat — el target compile_shaders no estarà disponible")
-  target_compile_options(shadertoy_sdl3 PRIVATE -std=c++17 -Wall -Os)
+endif()
-  if(APPLE)
+
-    find_library(COCOA_LIBRARY Cocoa REQUIRED)
+# --- RECURSO DE WINDOWS (icona) ---
-    find_library(IOKIT_LIBRARY IOKit REQUIRED)
+# A Windows compilem release/windows/shadertoy.rc amb windres via CMake.
-    find_library(CORE_VIDEO_LIBRARY CoreVideo REQUIRED)
+# El .rc fa referència a icon.ico per nom; afegim --include-dir perquè el
-    target_link_libraries(shadertoy_sdl3 PRIVATE ${COCOA_LIBRARY} ${IOKIT_LIBRARY} ${CORE_VIDEO_LIBRARY})
+# trobi a release/icons/.
-    target_compile_options(shadertoy_sdl3 PRIVATE -arch arm64)
+if(WIN32)
-  else()
+    enable_language(RC)
+    set(WIN_RC_FILE ${CMAKE_SOURCE_DIR}/release/windows/shadertoy.rc)
+    set_source_files_properties(${WIN_RC_FILE} PROPERTIES
+        COMPILE_FLAGS "--include-dir=${CMAKE_SOURCE_DIR}/release/icons")
+    list(APPEND APP_SOURCES ${WIN_RC_FILE})
+endif()
+
+# --- AÑADIR EJECUTABLE ---
+add_executable(${PROJECT_NAME} ${APP_SOURCES} ${EXTERNAL_SOURCES})
+
+# --- DIRECTORIOS DE INCLUSIÓN ---
+target_include_directories(${PROJECT_NAME} PUBLIC
+    "${CMAKE_SOURCE_DIR}/source"
+    "${CMAKE_SOURCE_DIR}/source/external/glad/include"
+    "${CMAKE_SOURCE_DIR}/source/external"
+)
+
+# Enllaçar la llibreria SDL3
+target_link_libraries(${PROJECT_NAME} PRIVATE SDL3::SDL3)
+
+
+# --- CONFIGURACIÓ PLATAFORMES I COMPILADOR  ---
+# Flags de compilació
+target_compile_options(${PROJECT_NAME} PRIVATE -Wall)
+target_compile_options(${PROJECT_NAME} PRIVATE $<$<CONFIG:RELEASE>:-Os -ffunction-sections -fdata-sections>)
+
+# Definir _DEBUG en mode Debug
+target_compile_definitions(${PROJECT_NAME} PRIVATE $<$<CONFIG:DEBUG>:_DEBUG>)
+
+
+# Configuració específica per a cada plataforma
+if(WIN32)
+    target_compile_definitions(${PROJECT_NAME} PRIVATE WINDOWS_BUILD)
+    target_link_libraries(${PROJECT_NAME} PRIVATE ws2_32 mingw32 opengl32)
+elseif(APPLE)
+    target_compile_definitions(${PROJECT_NAME} PRIVATE MACOS_BUILD)
+    target_compile_options(${PROJECT_NAME} PRIVATE -Wno-deprecated)
+    if(NOT CMAKE_OSX_ARCHITECTURES)
+        set(CMAKE_OSX_ARCHITECTURES "arm64")
+    endif()
+    if(MACOS_BUNDLE)
+        target_compile_definitions(${PROJECT_NAME} PRIVATE MACOS_BUNDLE)
+        target_link_options(${PROJECT_NAME} PRIVATE
+            -rpath @executable_path/../Frameworks/
+        )
+    endif()
+elseif(UNIX AND NOT APPLE)
+    target_compile_definitions(${PROJECT_NAME} PRIVATE LINUX_BUILD)
+endif()
+
+# Configuració comuna per a OpenGL
+if(NOT WIN32)
     find_package(OpenGL REQUIRED)
-    target_link_libraries(shadertoy_sdl3 PRIVATE OpenGL::GL)
+    if(OPENGL_FOUND)
-  endif()
+        message(STATUS "OpenGL trobat: ${OPENGL_LIBRARIES}")
+        target_link_libraries(${PROJECT_NAME} PRIVATE ${OPENGL_LIBRARIES})
+    else()
+        message(FATAL_ERROR "OpenGL no trobat")
+    endif()
 endif()
 
-# Ensure RUNTIME output directory for single-config builds (Release/Debug)
+# El binari surt a l'arrel del projecte (com fins ara).
-set_target_properties(shadertoy_sdl3 PROPERTIES
+set_target_properties(${PROJECT_NAME} PROPERTIES RUNTIME_OUTPUT_DIRECTORY ${CMAKE_SOURCE_DIR})
-  RUNTIME_OUTPUT_DIRECTORY "${CMAKE_BINARY_DIR}/bin"
-  RUNTIME_OUTPUT_DIRECTORY_RELEASE "${CMAKE_BINARY_DIR}/bin/Release"
-  RUNTIME_OUTPUT_DIRECTORY_DEBUG "${CMAKE_BINARY_DIR}/bin/Debug"
-)
-
-if(WIN32)
-  set(RELEASE_DIR "${CMAKE_BINARY_DIR}/release_package")
-  set(WINDOWS_ZIP "${CMAKE_BINARY_DIR}/${PROJECT_NAME}-${CMAKE_SYSTEM_PROCESSOR}-win.zip")
-  set(DLL_SOURCE_DIR "${CMAKE_SOURCE_DIR}/release")   # coloca tus SDL3.dll aquí
-
-  # PowerShell script que realizará las copias y el ZIP de forma robusta.
-  set(PS_SCRIPT "${CMAKE_BINARY_DIR}/package_windows.ps1")
-  file(WRITE "${PS_SCRIPT}" 
-"param(\n"
-"  [string]\$srcExe,\n"
-"  [string]\$releaseDir,\n"
-"  [string]\$readme,\n"
-"  [string]\$license,\n"
-"  [string]\$shadersDir,\n"
-"  [string]\$dllDir,\n"
-"  [string]\$outZip\n"
-")\n"
-"\n"
-"# Remove old\n"
-"if (Test-Path \$releaseDir) { Remove-Item -Recurse -Force \$releaseDir }\n"
-"New-Item -ItemType Directory -Path \$releaseDir | Out-Null\n"
-"\n"
-"# Copy exe\n"
-"Write-Host 'Copying executable...'\n"
-"Copy-Item -Path \$srcExe -Destination (Join-Path \$releaseDir (Split-Path \$srcExe -Leaf)) -Force\n"
-"\n"
-"# Copy README / LICENSE if exist\n"
-"if (Test-Path \$readme) { Copy-Item -Path \$readme -Destination (Join-Path \$releaseDir 'README.md') -Force }\n"
-"if (Test-Path \$license) { Copy-Item -Path \$license -Destination (Join-Path \$releaseDir 'LICENSE') -Force }\n"
-"\n"
-"# Copy shaders if present\n"
-"if (Test-Path \$shadersDir) { Copy-Item -Path (Join-Path \$shadersDir '*') -Destination (Join-Path \$releaseDir 'shaders') -Recurse -Force }\n"
-"\n"
-"# Copy DLLs if present\n"
-"if (Test-Path \$dllDir) { Copy-Item -Path (Join-Path \$dllDir '*') -Destination (Join-Path \$releaseDir 'release') -Recurse -Force }\n"
-"\n"
-"# Create ZIP\n"
-"Write-Host 'Creating zip: ' \$outZip\n"
-"if (Test-Path \$outZip) { Remove-Item -Force \$outZip }\n"
-"Compress-Archive -Path (Join-Path \$releaseDir '*') -DestinationPath \$outZip\n"
-"Write-Host 'Package created: ' \$outZip\n"
-)
-
-  add_custom_target(package_windows
-    COMMAND "${CMAKE_COMMAND}" --build "${CMAKE_BINARY_DIR}" --config Release
-    COMMAND powershell.exe -NoProfile -ExecutionPolicy Bypass -File "${PS_SCRIPT}"
-      -srcExe "$<TARGET_FILE:shadertoy_sdl3>"
-      -releaseDir "${RELEASE_DIR}"
-      -readme "${CMAKE_SOURCE_DIR}/README.md"
-      -license "${CMAKE_SOURCE_DIR}/LICENSE"
-      -shadersDir "${CMAKE_SOURCE_DIR}/shaders"
-      -dllDir "${DLL_SOURCE_DIR}"
-      -outZip "${WINDOWS_ZIP}"
-    COMMENT "Building Release and creating Windows package"
-    BYPRODUCTS "${WINDOWS_ZIP}"
-  )
-
-endif()
-
-# Helpful messages
-message(STATUS "CMake generator: ${CMAKE_GENERATOR}")
-message(STATUS "Build type: ${CMAKE_BUILD_TYPE}")
-message(STATUS "Binary output (runtime): ${CMAKE_BINARY_DIR}/bin")
-
-# Developer target: build Release and then package on Windows
-add_custom_target(build_release
-  COMMAND ${CMAKE_COMMAND} --build "${CMAKE_BINARY_DIR}" --config Release
-  COMMENT "Building Release configuration"
-)
-
-if(WIN32)
-  add_dependencies(package_windows build_release)
-endif()

LICENSE

@@ -1,6 +1,6 @@
 Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License
 
-Copyright (c) 2025 Coffee Crisis Arcade Edition
+Copyright (c) 2025 Shadertoy
 
 This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
 

Makefile

@@ -0,0 +1,189 @@
+# ==============================================================================
+# Makefile per a Shadertoy
+#
+# Tres targets:
+#   make            → Release (cmake)
+#   make debug      → Debug (cmake)
+#   make release    → Release + empaquetat per a distribució a dist/
+#
+# El SO es detecta automàticament. La compilació sempre passa per CMake.
+# ==============================================================================
+
+PROJECT  := shadertoy
+APP_NAME := Shadertoy
+DIR_ROOT := $(dir $(abspath $(MAKEFILE_LIST)))
+BUILDDIR := build
+DIST_DIR := dist
+RELEASE_FOLDER := $(DIST_DIR)/_tmp
+
+# ==============================================================================
+# DETECCIÓ DE PLATAFORMA + GENERADOR CMAKE + JOBS
+# ==============================================================================
+ifeq ($(OS),Windows_NT)
+    # MSYS2/Git Bash usa /dev/null, cmd.exe pur usa NUL.
+    ifneq ($(MSYSTEM),)
+        NULDEV := /dev/null
+    else
+        NULDEV := NUL
+    endif
+    JOBS      ?= $(NUMBER_OF_PROCESSORS)
+    HAS_NINJA := $(shell ninja --version 2>$(NULDEV))
+    ifneq ($(HAS_NINJA),)
+        CMAKE_GEN := -G "Ninja"
+    else
+        CMAKE_GEN := -G "MinGW Makefiles"
+    endif
+    UNAME_S   := Windows
+    SHELL     := cmd.exe
+else
+    NULDEV    := /dev/null
+    JOBS      ?= $(shell nproc 2>/dev/null || echo 4)
+    CMAKE_GEN :=
+    UNAME_S   := $(shell uname -s)
+endif
+
+# ==============================================================================
+# VERSIÓ (data, format YYYY.MM.DD)
+# ==============================================================================
+ifeq ($(OS),Windows_NT)
+    VERSION := $(shell powershell -Command "Get-Date -Format 'yyyy.MM.dd'")
+else
+    VERSION := $(shell date +%Y.%m.%d)
+endif
+
+# ==============================================================================
+# NOMS DELS ARTEFACTES DE RELEASE
+# ==============================================================================
+WINDOWS_RELEASE             := $(DIST_DIR)/$(PROJECT)-$(VERSION)-win32-x64.zip
+MACOS_APPLE_SILICON_RELEASE := $(DIST_DIR)/$(PROJECT)-$(VERSION)-macos-apple-silicon.dmg
+LINUX_RELEASE               := $(DIST_DIR)/$(PROJECT)-$(VERSION)-linux.tar.gz
+
+.PHONY: all debug release clean help _windows_release _macos_release _linux_release
+
+# ==============================================================================
+# COMPILACIÓ (delegada a CMake)
+# ==============================================================================
+all:
+	@cmake $(CMAKE_GEN) -S . -B $(BUILDDIR) -DCMAKE_BUILD_TYPE=Release
+	@cmake --build $(BUILDDIR) -j$(JOBS)
+
+debug:
+	@cmake $(CMAKE_GEN) -S . -B $(BUILDDIR) -DCMAKE_BUILD_TYPE=Debug
+	@cmake --build $(BUILDDIR) -j$(JOBS)
+
+# ==============================================================================
+# RELEASE (auto-dispatch per SO)
+# ==============================================================================
+release:
+ifeq ($(OS),Windows_NT)
+	@"$(MAKE)" _windows_release
+else
+ifeq ($(UNAME_S),Darwin)
+	@$(MAKE) _macos_release
+else
+	@$(MAKE) _linux_release
+endif
+endif
+
+clean:
+	@rm -rf $(BUILDDIR) $(DIST_DIR)
+
+# ==============================================================================
+# RELEASE LINUX
+# ==============================================================================
+_linux_release:
+	@echo "Creant release per a Linux - Versió: $(VERSION)"
+	@cmake $(CMAKE_GEN) -S . -B $(BUILDDIR) -DCMAKE_BUILD_TYPE=Release
+	@cmake --build $(BUILDDIR) -j$(JOBS)
+	@mkdir -p $(DIST_DIR)
+	@rm -rf $(RELEASE_FOLDER)
+	@mkdir -p $(RELEASE_FOLDER)
+	@cp -R data $(RELEASE_FOLDER)/
+	@cp LICENSE $(RELEASE_FOLDER)/
+	@cp README.md $(RELEASE_FOLDER)/
+	@cp $(PROJECT) $(RELEASE_FOLDER)/
+	@strip -s -R .comment -R .gnu.version $(RELEASE_FOLDER)/$(PROJECT) --strip-unneeded
+	@rm -f $(LINUX_RELEASE)
+	@tar -czf $(LINUX_RELEASE) -C $(RELEASE_FOLDER) .
+	@echo "Release creat: $(LINUX_RELEASE)"
+	@rm -rf $(RELEASE_FOLDER)
+
+# ==============================================================================
+# RELEASE MACOS (Apple Silicon, .dmg amb .app bundle)
+# ==============================================================================
+_macos_release:
+	@echo "Creant release per a macOS - Versió: $(VERSION)"
+	@command -v create-dmg >/dev/null 2>&1 || { \
+		echo "Falta create-dmg. Instal·la amb: brew install create-dmg"; \
+		exit 1; \
+	}
+	@cmake -S . -B $(BUILDDIR) -DCMAKE_BUILD_TYPE=Release -DMACOS_BUNDLE=ON -DCMAKE_OSX_ARCHITECTURES=arm64 -DCMAKE_OSX_DEPLOYMENT_TARGET=11.0
+	@cmake --build $(BUILDDIR) -j$(JOBS)
+	@mkdir -p $(DIST_DIR)
+	@rm -rf $(RELEASE_FOLDER)
+	@rm -f tmp.dmg "$(MACOS_APPLE_SILICON_RELEASE)" $(DIST_DIR)/rw.*
+	@mkdir -p "$(RELEASE_FOLDER)/$(APP_NAME).app/Contents/Frameworks"
+	@mkdir -p "$(RELEASE_FOLDER)/$(APP_NAME).app/Contents/MacOS"
+	@mkdir -p "$(RELEASE_FOLDER)/$(APP_NAME).app/Contents/Resources"
+	@cp -R data "$(RELEASE_FOLDER)/$(APP_NAME).app/Contents/Resources/"
+	@cp -R release/macos/frameworks/SDL3.xcframework/macos-arm64_x86_64/SDL3.framework "$(RELEASE_FOLDER)/$(APP_NAME).app/Contents/Frameworks/"
+	@cp release/icons/icon.icns "$(RELEASE_FOLDER)/$(APP_NAME).app/Contents/Resources/"
+	@cp release/macos/Info.plist "$(RELEASE_FOLDER)/$(APP_NAME).app/Contents/"
+	@cp $(PROJECT) "$(RELEASE_FOLDER)/$(APP_NAME).app/Contents/MacOS/$(PROJECT)"
+	@cp LICENSE "$(RELEASE_FOLDER)/"
+	@cp README.md "$(RELEASE_FOLDER)/"
+	@RAW_VERSION=$$(echo "$(VERSION)" | sed 's/^v//'); \
+		sed -i '' '/<key>CFBundleShortVersionString<\/key>/{n;s|<string>.*</string>|<string>'"$$RAW_VERSION"'</string>|;}' "$(RELEASE_FOLDER)/$(APP_NAME).app/Contents/Info.plist"; \
+		sed -i '' '/<key>CFBundleVersion<\/key>/{n;s|<string>.*</string>|<string>'"$$RAW_VERSION"'</string>|;}' "$(RELEASE_FOLDER)/$(APP_NAME).app/Contents/Info.plist"
+	@codesign --deep --force --sign - --timestamp=none "$(RELEASE_FOLDER)/$(APP_NAME).app"
+	@create-dmg \
+		--volname "$(APP_NAME)" \
+		--window-pos 200 120 \
+		--window-size 720 300 \
+		--icon-size 96 \
+		--text-size 12 \
+		--icon "$(APP_NAME).app" 278 102 \
+		--icon "LICENSE" 441 102 \
+		--icon "README.md" 604 102 \
+		--app-drop-link 115 102 \
+		--hide-extension "$(APP_NAME).app" \
+		"$(MACOS_APPLE_SILICON_RELEASE)" \
+		"$(RELEASE_FOLDER)" || true
+	@echo "Release creat: $(MACOS_APPLE_SILICON_RELEASE)"
+	@rm -rf $(RELEASE_FOLDER)
+	@rm -f $(DIST_DIR)/rw.*
+
+# ==============================================================================
+# RELEASE WINDOWS (.zip amb DLLs)
+# ==============================================================================
+_windows_release:
+	@echo off
+	@echo Creant release per a Windows - Versió: $(VERSION)
+	@cmake $(CMAKE_GEN) -S . -B $(BUILDDIR) -DCMAKE_BUILD_TYPE=Release
+	@cmake --build $(BUILDDIR) -j$(JOBS)
+	@powershell -Command "if (-not (Test-Path '$(DIST_DIR)')) {New-Item '$(DIST_DIR)' -ItemType Directory | Out-Null}"
+	@powershell -Command "if (Test-Path '$(RELEASE_FOLDER)') {Remove-Item '$(RELEASE_FOLDER)' -Recurse -Force}"
+	@powershell -Command "New-Item '$(RELEASE_FOLDER)' -ItemType Directory | Out-Null"
+	@powershell -Command "Copy-Item -Path 'data' -Destination '$(RELEASE_FOLDER)' -Recurse"
+	@powershell -Command "Copy-Item 'LICENSE' -Destination '$(RELEASE_FOLDER)'"
+	@powershell -Command "Copy-Item 'README.md' -Destination '$(RELEASE_FOLDER)'"
+	@powershell -Command "Copy-Item 'release\windows\dll\*.dll' -Destination '$(RELEASE_FOLDER)'"
+	@powershell -Command "Copy-Item -Path '$(PROJECT).exe' -Destination '$(RELEASE_FOLDER)/$(APP_NAME).exe'"
+	@strip -s -R .comment -R .gnu.version "$(RELEASE_FOLDER)/$(APP_NAME).exe" --strip-unneeded
+	@powershell -Command "if (Test-Path '$(WINDOWS_RELEASE)') {Remove-Item '$(WINDOWS_RELEASE)'}"
+	@powershell -Command "Compress-Archive -Path '$(RELEASE_FOLDER)/*' -DestinationPath '$(WINDOWS_RELEASE)'"
+	@echo Release creat: $(WINDOWS_RELEASE)
+	@powershell -Command "Remove-Item '$(RELEASE_FOLDER)' -Recurse -Force"
+
+# ==============================================================================
+# AJUDA
+# ==============================================================================
+help:
+	@echo "Makefile per a $(PROJECT)"
+	@echo ""
+	@echo "  make            Compilar Release (cmake)"
+	@echo "  make debug      Compilar Debug (cmake)"
+	@echo "  make release    Compilar + empaquetar per a distribució a $(DIST_DIR)/"
+	@echo "  make clean      Esborrar $(BUILDDIR)/ i $(DIST_DIR)/"
+	@echo ""
+	@echo "  Versió actual: $(VERSION)"

README.md

@@ -11,17 +11,27 @@ Proyecto minimal para ejecutar fragment shaders tipo Shadertoy usando SDL3 + GLA
 
 
 
+## Características
+
+- **Visualizador de shaders fragment** compatible con formato Shadertoy
+- **Contador de FPS** en barra de título
+- **Toggle VSync** con tecla F4
+- **Cambio de shaders en runtime** con flechas ←/→
+- **Sistema de metadata** en shaders (Name/Author automático en título)
+- **Sistema de feedback** para shaders que requieren frame anterior (opcional)
+- **Soporte de audio** con jail_audio (música de fondo)
+
+
+
 ## Estructura
 
-- src/ — código fuente C++ (incluye main.cpp).
+- src/ — código fuente C++ (incluye main.cpp)
-
+- shaders/ — shaders fragment (.frag.glsl) que se cargan en runtime
-- shaders/ — shaders fragment (.frag.glsl) que se cargan en runtime.
+- data/ — recursos (música, texturas, etc.)
-
+- third_party/glad/ — glad.c + headers
-- third_party/glad/ — glad.c + headers.
+- third_party/ — jail_audio y otras dependencias
-
+- CMakeLists.txt — configuración de build
-- CMakeLists.txt — configuración de build.
+- Makefile — builds de release para Windows/macOS/Linux
-
-- .gitignore — recomendado.
 
   
 
@@ -63,55 +73,74 @@ brew install cmake sdl3
 
 ## Uso
 
-Por defecto el ejecutable carga shaders/test.frag.glsl.
+Ejecutar con un shader específico:
+
+    ./shadertoy shaders/fractal_pyramid.frag.glsl
 
 
 
-Ejecutar en ventana:
+Ejecutar en fullscreen:
 
-    ./shadertoy_sdl3 ../shaders/test.frag.glsl
+    ./shadertoy -F shaders/seascape.frag.glsl
-
+    # o
-  
+    ./shadertoy --fullscreen shaders/seascape.frag.glsl
-
-Ejecutar en fullscreen nativo:
-
-    ./shadertoy_sdl3 -F ../shaders/test.frag.glsl
 
 
 
 Si ejecutas desde la raíz del repo:
 
-    ./build/shadertoy_sdl3 shaders/frag_tile_iter.frag.glsl
+    ./build/shadertoy shaders/water.glsl
 
 
 
-Atajos en ejecución
+### Atajos de teclado
 
-- Escape — salir.
+- **ESC** — Salir de la aplicación
-
+- **F3** — Toggle fullscreen
-- F11 — alternar fullscreen desktop.
+- **F4** — Toggle VSync (ON/OFF)
+- **← / →** — Cambiar al shader anterior/siguiente en la carpeta shaders/
 
   
 
 ## Formato de shader esperado
 
-- #version 330 core
+### Header obligatorio:
+```glsl
+// Name: Nombre del shader
+// Author: Autor
+#version 330 core
+precision highp float;
 
-- Debe declarar:
+out vec4 FragColor;
+in vec2 vUV;
+uniform vec2 iResolution;
+uniform float iTime;
+```
 
-      uniform vec2 iResolution;
+### Función mainImage (estilo Shadertoy):
+```glsl
+void mainImage(out vec4 fragColor, in vec2 fragCoord) {
+    // Tu código aquí
+    // fragCoord está en píxeles
+    vec2 uv = fragCoord / iResolution.xy;
+    fragColor = vec4(uv, 0.5, 1.0);
+}
+```
 
-      uniform float iTime;
+### Wrapper main():
+```glsl
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}
+```
 
-      in vec2 vUV;
+### Metadata opcional:
-
+- `// Name: Nombre` - Aparece en barra de título
-      out vec4 FragColor;
+- `// Author: Autor` - Aparece en barra de título
-
+- `// iChannel3: self` - Habilita feedback (frame anterior)
-- Función de entrada esperada (opcional, el main llama a mainImage):
-
-      void mainImage(out vec4 fragColor, in vec2 fragCoord);
-
-- main() debe convertir vUV a fragCoord y llamar a mainImage.
 
   
 
@@ -133,15 +162,33 @@ Si obtienes "Failed to load fragment shader file", ejecuta desde la carpeta buil
 
   
 
-## Multi-pass / Buffers (nota)
+## Compatibilidad con Shadertoy
 
-Ejemplo actual es single-pass. Para portar Shadertoy con BufferA/B/C necesitas:
+### Características soportadas:
+✅ `iTime` - Tiempo en segundos
+✅ `iResolution` - Resolución de ventana (vec2)
+✅ `mainImage()` - Función de entrada estándar
+✅ Self-feedback - Frame anterior con `// iChannel3: self`
 
-- crear FBOs y textures por buffer
+### No soportado actualmente:
+❌ `iMouse` - Interacción con ratón
+❌ `iChannel0-2` - Texturas externas
+❌ Multi-pass completo (BufferA/B/C/D)
+❌ `iFrame`, `iTimeDelta`, `iDate`
 
-- renderizar buffers en orden y pasar las texturas como iChannelN a pases posteriores
+### Diferencias importantes:
+- **`iResolution`**: En Shadertoy es `vec3(width, height, width/height)`, aquí es `vec2(width, height)`
+  - Solución: `vec3 r = vec3(iResolution.xy, iResolution.x/iResolution.y);`
+- **Inicialización de variables**: OpenGL nativo requiere inicializar variables explícitamente
+- **División por valores pequeños**: Puede causar overflow - usar `max(divisor, epsilon)`
 
-- evitar leer y escribir la misma textura (usar ping-pong si hace falta)
+### Conversión de shaders de Shadertoy:
+1. Copiar función `mainImage()` tal cual
+2. Añadir header estándar (ver arriba)
+3. Añadir wrapper `main()`
+4. Eliminar referencias a `iChannel0-3` si no se usan
+5. Adaptar `iResolution` de vec3 a vec2 si es necesario
+6. Inicializar variables: `vec3 col = vec3(0.0);` en lugar de `vec3 col;`
 
   
 

data/music/_readme_and_license.txt

@@ -0,0 +1,41 @@
+Pack downloaded from Freesound
+----------------------------------------
+
+"Ambient pads"
+
+This Pack of sounds contains sounds by the following user:
+ - Mat397 ( https://freesound.org/people/Mat397/ )
+
+You can find this pack online at: https://freesound.org/people/Mat397/packs/27416/
+
+
+Licenses in this Pack (see below for individual sound licenses)
+---------------------------------------------------------------
+
+Creative Commons 0: http://creativecommons.org/publicdomain/zero/1.0/
+Attribution 3.0: http://creativecommons.org/licenses/by/3.0/
+
+
+Sounds in this Pack
+-------------------
+
+  * 618042__mat397__bad-electric-dark-pad.wav.wav
+    * url: https://freesound.org/s/618042/
+    * license: Creative Commons 0
+  * 618041__mat397__mangle-dark-pad.wav.wav
+    * url: https://freesound.org/s/618041/
+    * license: Creative Commons 0
+  * 486083__mat397__world-of-ants-pad.wav.wav
+    * url: https://freesound.org/s/486083/
+    * license: Attribution 3.0
+  * 485079__mat397__melancholic-flutes-pad.wav.wav
+    * url: https://freesound.org/s/485079/
+    * license: Attribution 3.0
+  * 485078__mat397__polyflute-pad.wav.wav
+    * url: https://freesound.org/s/485078/
+    * license: Attribution 3.0
+  * 485077__mat397__confused-voices.wav.wav
+    * url: https://freesound.org/s/485077/
+    * license: Attribution 3.0
+
+

data/shaders/_common/passthrough.vert.msl

@@ -0,0 +1,19 @@
+#include <metal_stdlib>
+using namespace metal;
+
+// Shared fullscreen-triangle vertex shader for the SDL3 GPU backend.
+// Emits uv in Shadertoy convention: (0,0) bottom-left, (1,1) top-right.
+
+struct PassthroughVOut {
+    float4 pos [[position]];
+    float2 uv;
+};
+
+vertex PassthroughVOut passthrough_vs(uint vid [[vertex_id]]) {
+    const float2 positions[3] = { {-1.0, -1.0}, {3.0, -1.0}, {-1.0, 3.0} };
+    PassthroughVOut out;
+    float2 pos = positions[vid];
+    out.uv = pos * 0.5 + 0.5;
+    out.pos = float4(pos, 0.0, 1.0);
+    return out;
+}

data/shaders/_common/passthrough.vk.glsl

@@ -0,0 +1,17 @@
+#version 450
+
+// Shared fullscreen-triangle vertex shader for the SDL3 GPU backend.
+// Emits vUV in Shadertoy convention: (0,0) bottom-left, (1,1) top-right.
+
+layout(location = 0) out vec2 vUV;
+
+void main() {
+    const vec2 positions[3] = vec2[3](
+        vec2(-1.0, -1.0),
+        vec2( 3.0, -1.0),
+        vec2(-1.0,  3.0)
+    );
+    vec2 pos = positions[gl_VertexIndex];
+    vUV = pos * 0.5 + 0.5;
+    gl_Position = vec4(pos, 0.0, 1.0);
+}

data/shaders/cineshader_lava/cineshader_lava.frag.msl

@@ -0,0 +1,75 @@
+#include <metal_stdlib>
+using namespace metal;
+
+// Cineshader Lava — edankwan
+// MSL port of cineshader_lava.vk.glsl.
+
+struct ShadertoyUBO {
+    float iTime;
+    float2 iResolution;
+};
+
+struct PassthroughVOut {
+    float4 pos [[position]];
+    float2 uv;
+};
+
+static float opSmoothUnion(float d1, float d2, float k) {
+    float h = clamp(0.5 + 0.5 * (d2 - d1) / k, 0.0, 1.0);
+    return mix(d2, d1, h) - k * h * (1.0 - h);
+}
+
+static float sdSphere(float3 p, float s) {
+    return length(p) - s;
+}
+
+static float map(float3 p, float iTime) {
+    float d = 2.0;
+    for (int i = 0; i < 16; i++) {
+        float fi = float(i);
+        float time = iTime * (fract(fi * 412.531 + 0.513) - 0.5) * 2.0;
+        d = opSmoothUnion(
+            sdSphere(p + sin(time + fi * float3(52.5126, 64.62744, 632.25)) * float3(2.0, 2.0, 0.8),
+                     mix(0.5, 1.0, fract(fi * 412.531 + 0.5124))),
+            d,
+            0.4
+        );
+    }
+    return d;
+}
+
+static float3 calcNormal(float3 p, float iTime) {
+    const float h = 1e-5;
+    const float2 k = float2(1, -1);
+    return normalize(k.xyy * map(p + k.xyy * h, iTime) +
+                     k.yyx * map(p + k.yyx * h, iTime) +
+                     k.yxy * map(p + k.yxy * h, iTime) +
+                     k.xxx * map(p + k.xxx * h, iTime));
+}
+
+fragment float4 cineshader_lava_fs(PassthroughVOut in [[stage_in]],
+                                    constant ShadertoyUBO& U [[buffer(0)]]) {
+    float2 fragCoord = in.uv * U.iResolution;
+    float2 uv = fragCoord / U.iResolution;
+
+    float3 rayOri = float3((uv - 0.5) * float2(U.iResolution.x / U.iResolution.y, 1.0) * 6.0, 3.0);
+    float3 rayDir = float3(0.0, 0.0, -1.0);
+
+    float depth = 0.0;
+    float3 p = float3(0.0);
+
+    for (int i = 0; i < 64; i++) {
+        p = rayOri + rayDir * depth;
+        float dist = map(p, U.iTime);
+        depth += dist;
+        if (dist < 1e-6) { break; }
+    }
+
+    depth = min(6.0, depth);
+    float3 n = calcNormal(p, U.iTime);
+    float b = max(0.0, dot(n, float3(0.577)));
+    float3 col = (0.5 + 0.5 * cos((b + U.iTime * 3.0) + uv.xyx * 2.0 + float3(0, 2, 4))) * (0.85 + b * 0.35);
+    col *= exp(-depth * 0.15);
+
+    return float4(col, 1.0 - (depth - 0.5) / 2.0);
+}

data/shaders/cineshader_lava/cineshader_lava.gl.glsl

@@ -0,0 +1,83 @@
+// Name: Cineshader Lava
+// Author: edankwan
+// URL: https://www.shadertoy.com/view/3sySRK
+#version 330 core
+precision highp float;
+
+out vec4 FragColor;
+in vec2 vUV;
+uniform vec2 iResolution;
+uniform float iTime;
+
+float opSmoothUnion( float d1, float d2, float k )
+{
+    float h = clamp( 0.5 + 0.5*(d2-d1)/k, 0.0, 1.0 );
+    return mix( d2, d1, h ) - k*h*(1.0-h);
+}
+
+float sdSphere( vec3 p, float s )
+{
+  return length(p)-s;
+}
+
+float map(vec3 p)
+{
+    float d = 2.0;
+    for (int i = 0; i < 16; i++) {
+        float fi = float(i);
+        float time = iTime * (fract(fi * 412.531 + 0.513) - 0.5) * 2.0;
+        d = opSmoothUnion(
+            sdSphere(p + sin(time + fi * vec3(52.5126, 64.62744, 632.25)) * vec3(2.0, 2.0, 0.8), mix(0.5, 1.0, fract(fi * 412.531 + 0.5124))),
+            d,
+            0.4
+        );
+    }
+    return d;
+}
+
+vec3 calcNormal( in vec3 p )
+{
+    const float h = 1e-5; // or some other value
+    const vec2 k = vec2(1,-1);
+    return normalize( k.xyy*map( p + k.xyy*h ) +
+                      k.yyx*map( p + k.yyx*h ) +
+                      k.yxy*map( p + k.yxy*h ) +
+                      k.xxx*map( p + k.xxx*h ) );
+}
+
+void mainImage( out vec4 fragColor, in vec2 fragCoord )
+{
+    vec2 uv = fragCoord/iResolution.xy;
+
+    // screen size is 6m x 6m
+    vec3 rayOri = vec3((uv - 0.5) * vec2(iResolution.x/iResolution.y, 1.0) * 6.0, 3.0);
+    vec3 rayDir = vec3(0.0, 0.0, -1.0);
+
+    float depth = 0.0;
+    vec3 p;
+
+    for(int i = 0; i < 64; i++) {
+        p = rayOri + rayDir * depth;
+        float dist = map(p);
+        depth += dist;
+        if (dist < 1e-6) {
+            break;
+        }
+    }
+
+    depth = min(6.0, depth);
+    vec3 n = calcNormal(p);
+    float b = max(0.0, dot(n, vec3(0.577)));
+    vec3 col = (0.5 + 0.5 * cos((b + iTime * 3.0) + uv.xyx * 2.0 + vec3(0,2,4))) * (0.85 + b * 0.35);
+    col *= exp( -depth * 0.15 );
+
+    // maximum thickness is 2m in alpha channel
+    fragColor = vec4(col, 1.0 - (depth - 0.5) / 2.0);
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

data/shaders/cineshader_lava/cineshader_lava.vk.glsl

@@ -0,0 +1,77 @@
+#version 450
+
+// Name: Cineshader Lava
+// Author: edankwan
+// URL: https://www.shadertoy.com/view/3sySRK
+
+layout(location = 0) in  vec2 vUV;
+layout(location = 0) out vec4 FragColor;
+
+layout(set = 3, binding = 0) uniform ShadertoyUBO {
+    float iTime;
+    vec2  iResolution;
+};
+
+float opSmoothUnion(float d1, float d2, float k) {
+    float h = clamp(0.5 + 0.5 * (d2 - d1) / k, 0.0, 1.0);
+    return mix(d2, d1, h) - k * h * (1.0 - h);
+}
+
+float sdSphere(vec3 p, float s) {
+    return length(p) - s;
+}
+
+float map(vec3 p) {
+    float d = 2.0;
+    for (int i = 0; i < 16; i++) {
+        float fi = float(i);
+        float time = iTime * (fract(fi * 412.531 + 0.513) - 0.5) * 2.0;
+        d = opSmoothUnion(
+            sdSphere(p + sin(time + fi * vec3(52.5126, 64.62744, 632.25)) * vec3(2.0, 2.0, 0.8), mix(0.5, 1.0, fract(fi * 412.531 + 0.5124))),
+            d,
+            0.4
+        );
+    }
+    return d;
+}
+
+vec3 calcNormal(in vec3 p) {
+    const float h = 1e-5;
+    const vec2 k = vec2(1, -1);
+    return normalize(k.xyy * map(p + k.xyy * h) +
+                     k.yyx * map(p + k.yyx * h) +
+                     k.yxy * map(p + k.yxy * h) +
+                     k.xxx * map(p + k.xxx * h));
+}
+
+void mainImage(out vec4 fragColor, in vec2 fragCoord) {
+    vec2 uv = fragCoord / iResolution.xy;
+
+    vec3 rayOri = vec3((uv - 0.5) * vec2(iResolution.x / iResolution.y, 1.0) * 6.0, 3.0);
+    vec3 rayDir = vec3(0.0, 0.0, -1.0);
+
+    float depth = 0.0;
+    vec3 p;
+
+    for (int i = 0; i < 64; i++) {
+        p = rayOri + rayDir * depth;
+        float dist = map(p);
+        depth += dist;
+        if (dist < 1e-6) { break; }
+    }
+
+    depth = min(6.0, depth);
+    vec3 n = calcNormal(p);
+    float b = max(0.0, dot(n, vec3(0.577)));
+    vec3 col = (0.5 + 0.5 * cos((b + iTime * 3.0) + uv.xyx * 2.0 + vec3(0, 2, 4))) * (0.85 + b * 0.35);
+    col *= exp(-depth * 0.15);
+
+    fragColor = vec4(col, 1.0 - (depth - 0.5) / 2.0);
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

data/shaders/cineshader_lava/meta.txt

@@ -0,0 +1,2 @@
+Name: Cineshader Lava
+Author: edankwan

data/shaders/creation/creation.frag.msl

@@ -0,0 +1,42 @@
+#include <metal_stdlib>
+using namespace metal;
+
+// Creation by Silexars — Danguafer
+// MSL port of creation.vk.glsl.
+
+struct ShadertoyUBO {
+    float iTime;
+    float2 iResolution;
+};
+
+struct PassthroughVOut {
+    float4 pos [[position]];
+    float2 uv;
+};
+
+fragment float4 creation_fs(PassthroughVOut in [[stage_in]],
+                            constant ShadertoyUBO& U [[buffer(0)]]) {
+    float2 fragCoord = in.uv * U.iResolution;
+    float t = U.iTime;
+    float2 r = U.iResolution;
+
+    float3 c = float3(0.0);
+    float l = 0.0;
+    float z = t;
+    for (int i = 0; i < 3; i++) {
+        float2 p = fragCoord / r;
+        float2 uv = p;
+        p -= 0.5;
+        p.x *= r.x / r.y;
+        z += 0.07;
+        l = length(p);
+        float invl = (l > 0.0) ? (1.0 / l) : 0.0;
+        uv += p * invl * (sin(z) + 1.0) * abs(sin(l * 9.0 - z - z));
+        float2 m = fract(uv) - 0.5;
+        float denom = length(m);
+        if (denom < 1e-6) denom = 1e-6;
+        c[i] = 0.01 / denom;
+    }
+    float L = (l > 0.0) ? l : 1.0;
+    return float4(c / L, t);
+}

data/shaders/creation/creation.gl.glsl

@@ -1,3 +1,6 @@
+// Name: Creation by Silexars
+// Author: Danguafer
+// URL: https://www.shadertoy.com/view/XsXXDn
 #version 330 core
 precision highp float;
 

data/shaders/creation/creation.vk.glsl

@@ -0,0 +1,45 @@
+#version 450
+
+// Name: Creation by Silexars
+// Author: Danguafer
+// URL: https://www.shadertoy.com/view/XsXXDn
+
+layout(location = 0) in  vec2 vUV;
+layout(location = 0) out vec4 FragColor;
+
+layout(set = 3, binding = 0) uniform ShadertoyUBO {
+    float iTime;
+    vec2  iResolution;
+};
+
+#define t iTime
+#define r iResolution.xy
+
+void mainImage(out vec4 fragColor, in vec2 fragCoord) {
+    vec3 c = vec3(0.0);
+    float l;
+    float z = t;
+    for (int i = 0; i < 3; i++) {
+        vec2 uv, p = fragCoord.xy / r;
+        uv = p;
+        p -= 0.5;
+        p.x *= r.x / r.y;
+        z += 0.07;
+        l = length(p);
+        float invl = (l > 0.0) ? (1.0 / l) : 0.0;
+        uv += p * invl * (sin(z) + 1.0) * abs(sin(l * 9.0 - z - z));
+        vec2 m = mod(uv, 1.0) - 0.5;
+        float denom = length(m);
+        if (denom < 1e-6) denom = 1e-6;
+        c[i] = 0.01 / denom;
+    }
+    float L = (l > 0.0) ? l : 1.0;
+    fragColor = vec4(c / L, t);
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

data/shaders/creation/meta.txt

@@ -0,0 +1,2 @@
+Name: Creation by Silexars
+Author: Danguafer

data/shaders/cube_lines/cube_lines.frag.msl

@@ -0,0 +1,455 @@
+#include <metal_stdlib>
+using namespace metal;
+
+// Cube lines — Danil
+// MSL port of cube_lines.vk.glsl. The original Shadertoy is ~780 lines and
+// uses every GLSL trick (mat3 chains, arrays, fwidth-based AA, swap macros).
+//
+// Notes:
+//   - fcos() uses fwidth() like the GLSL original to soften high-frequency
+//     terms. The GLSL fallback path that divides by iResolution.y when
+//     fwidth returns 0/NaN/Inf is omitted here; the main fwidth-based
+//     branch covers virtually all fragments and the fallback is rarely
+//     hit in practice.
+//   - Optional Shadertoy defines (ANIM_SHAPE, ANIM_COLOR, ROTATION_SPEED,
+//     CAMERA_*, USE_COLOR, AA_*, ONLY_BOX, etc.) are left at their defaults
+//     just like the .vk.glsl ships them.
+
+struct ShadertoyUBO {
+    float iTime;
+    float2 iResolution;
+};
+
+struct PassthroughVOut {
+    float4 pos [[position]];
+    float2 uv;
+};
+
+constant float ROTATION_SPEED = 0.8999;
+constant float tshift = 53.0;
+constant float FDIST = 0.7;
+constant float PI = 3.1415926;
+constant float3 BOXDIMS = float3(0.75, 0.75, 1.25);
+constant float IOR = 1.33;
+
+static float3x3 rotx(float a) {
+    float s = sin(a), c = cos(a);
+    return float3x3(float3(1.0, 0.0, 0.0), float3(0.0, c, s), float3(0.0, -s, c));
+}
+static float3x3 roty(float a) {
+    float s = sin(a), c = cos(a);
+    return float3x3(float3(c, 0.0, s), float3(0.0, 1.0, 0.0), float3(-s, 0.0, c));
+}
+static float3x3 rotz(float a) {
+    float s = sin(a), c = cos(a);
+    return float3x3(float3(c, s, 0.0), float3(-s, c, 0.0), float3(0.0, 0.0, 1.0));
+}
+
+static float3 fcos(float3 x) {
+    float3 w = fwidth(x);
+    return cos(x) * smoothstep(float3(3.14 * 2.0), float3(0.0), w);
+}
+
+static float3 getColor(float3 p) {
+    p = abs(p);
+    p *= 1.25;
+    p = 0.5 * p / dot(p, p);
+
+    float t = 0.13 * length(p);
+    float3 col = float3(0.3, 0.4, 0.5);
+    col += 0.12 * fcos(6.28318 * t * 1.0 + float3(0.0, 0.8, 1.1));
+    col += 0.11 * fcos(6.28318 * t * 3.1 + float3(0.3, 0.4, 0.1));
+    col += 0.10 * fcos(6.28318 * t * 5.1 + float3(0.1, 0.7, 1.1));
+    col += 0.10 * fcos(6.28318 * t * 17.1 + float3(0.2, 0.6, 0.7));
+    col += 0.10 * fcos(6.28318 * t * 31.1 + float3(0.1, 0.6, 0.7));
+    col += 0.10 * fcos(6.28318 * t * 65.1 + float3(0.0, 0.5, 0.8));
+    col += 0.10 * fcos(6.28318 * t * 115.1 + float3(0.1, 0.4, 0.7));
+    col += 0.10 * fcos(6.28318 * t * 265.1 + float3(1.1, 1.4, 2.7));
+    col = clamp(col, 0.0, 1.0);
+    return col;
+}
+
+static void calcColor(float3 ro, float3 rd, float3 nor, float d, float len, int idx,
+                      bool si, float td,
+                      thread float4& colx, thread float4& colsi) {
+    float3 pos = ro + rd * d;
+    float a = 1.0 - smoothstep(len - 0.15 * 0.5, len + 0.00001, length(pos));
+    float3 col = getColor(pos);
+    colx = float4(col, a);
+    if (si) {
+        pos = ro + rd * td;
+        float ta = 1.0 - smoothstep(len - 0.15 * 0.5, len + 0.00001, length(pos));
+        col = getColor(pos);
+        colsi = float4(col, ta);
+    }
+}
+
+static bool iBilinearPatch(float3 ro, float3 rd, float4 ps, float4 ph, float sz,
+                           thread float& t, thread float3& norm,
+                           thread bool& si, thread float& tsi, thread float3& normsi,
+                           thread float& fade, thread float& fadesi) {
+    float3 va = float3(0.0, 0.0, ph.x + ph.w - ph.y - ph.z);
+    float3 vb = float3(0.0, ps.w - ps.y, ph.z - ph.x);
+    float3 vc = float3(ps.z - ps.x, 0.0, ph.y - ph.x);
+    float3 vd = float3(ps.xy, ph.x);
+    t = -1.0;
+    tsi = -1.0;
+    si = false;
+    fade = 1.0;
+    fadesi = 1.0;
+    norm = float3(0.0, 1.0, 0.0);
+    normsi = float3(0.0, 1.0, 0.0);
+
+    float tmp = 1.0 / (vb.y * vc.x);
+    float a = 0.0, b = 0.0, c = 0.0;
+    float d = va.z * tmp;
+    float e = 0.0, f = 0.0;
+    float g = (vc.z * vb.y - vd.y * va.z) * tmp;
+    float h = (vb.z * vc.x - va.z * vd.x) * tmp;
+    float i = -1.0;
+    float j = (vd.x * vd.y * va.z + vd.z * vb.y * vc.x) * tmp - (vd.y * vb.z * vc.x + vd.x * vc.z * vb.y) * tmp;
+
+    float p = dot(float3(a, b, c), rd.xzy * rd.xzy) + dot(float3(d, e, f), rd.xzy * rd.zyx);
+    float q = dot(float3(2.0, 2.0, 2.0) * ro.xzy * rd.xyz, float3(a, b, c)) + dot(ro.xzz * rd.zxy, float3(d, d, e)) +
+              dot(ro.yyx * rd.zxy, float3(e, f, f)) + dot(float3(g, h, i), rd.xzy);
+    float r = dot(float3(a, b, c), ro.xzy * ro.xzy) + dot(float3(d, e, f), ro.xzy * ro.zyx) + dot(float3(g, h, i), ro.xzy) + j;
+
+    if (abs(p) < 0.000001) {
+        float tt = -r / q;
+        if (tt <= 0.0) return false;
+        t = tt;
+        float3 pos = ro + t * rd;
+        if (length(pos) > sz) return false;
+        float3 grad = float3(2.0) * pos.xzy * float3(a, b, c) + pos.zxz * float3(d, d, e) + pos.yyx * float3(f, e, f) + float3(g, h, i);
+        norm = -normalize(grad);
+        return true;
+    } else {
+        float sq = q * q - 4.0 * p * r;
+        if (sq < 0.0) return false;
+        float s = sqrt(sq);
+        float t0 = (-q + s) / (2.0 * p);
+        float t1 = (-q - s) / (2.0 * p);
+        float tt1 = min(t0 < 0.0 ? t1 : t0, t1 < 0.0 ? t0 : t1);
+        float tt2 = max(t0 > 0.0 ? t1 : t0, t1 > 0.0 ? t0 : t1);
+        float tt0 = tt1;
+        if (tt0 <= 0.0) return false;
+        float3 pos = ro + tt0 * rd;
+        bool ru = step(sz, length(pos)) > 0.5;
+        if (ru) {
+            tt0 = tt2;
+            pos = ro + tt0 * rd;
+        }
+        if (tt0 <= 0.0) return false;
+        bool ru2 = step(sz, length(pos)) > 0.5;
+        if (ru2) return false;
+
+        if ((tt2 > 0.0) && (!ru) && !(step(sz, length(ro + tt2 * rd)) > 0.5)) {
+            si = true;
+            fadesi = s;
+            tsi = tt2;
+            float3 tpos = ro + tsi * rd;
+            float3 tgrad = float3(2.0) * tpos.xzy * float3(a, b, c) + tpos.zxz * float3(d, d, e) +
+                           tpos.yyx * float3(f, e, f) + float3(g, h, i);
+            normsi = -normalize(tgrad);
+        }
+
+        fade = s;
+        t = tt0;
+        float3 grad = float3(2.0) * pos.xzy * float3(a, b, c) + pos.zxz * float3(d, d, e) + pos.yyx * float3(f, e, f) + float3(g, h, i);
+        norm = -normalize(grad);
+        return true;
+    }
+}
+
+static float dot2(float3 v) { return dot(v, v); }
+
+static float segShadow(float3 ro, float3 rd, float3 pa, float sh) {
+    float dm = dot(rd.yz, rd.yz);
+    float k1 = (ro.x - pa.x) * dm;
+    float k2 = (ro.x + pa.x) * dm;
+    float2 k5 = (ro.yz + pa.yz) * dm;
+    float k3 = dot(ro.yz + pa.yz, rd.yz);
+    float2 k4 = (pa.yz + pa.yz) * rd.yz;
+    float2 k6 = (pa.yz + pa.yz) * dm;
+
+    for (int i = 0; i < 4; i++) {
+        float2 s = float2(i & 1, i >> 1);
+        float t = dot(s, k4) - k3;
+        if (t > 0.0) {
+            sh = min(sh, dot2(float3(clamp(-rd.x * t, k1, k2), k5 - k6 * s) + rd * t) / (t * t));
+        }
+    }
+    return sh;
+}
+
+static float boxSoftShadow(float3 ro, float3 rd, float3 rad, float sk) {
+    rd += 0.0001 * (1.0 - abs(sign(rd)));
+    float3 rdd = rd;
+    float3 roo = ro;
+
+    float3 m = 1.0 / rdd;
+    float3 n = m * roo;
+    float3 k = abs(m) * rad;
+
+    float3 t1 = -n - k;
+    float3 t2 = -n + k;
+
+    float tN = max(max(t1.x, t1.y), t1.z);
+    float tF = min(min(t2.x, t2.y), t2.z);
+
+    if (tN < tF && tF > 0.0) return 0.0;
+
+    float sh = 1.0;
+    sh = segShadow(roo.xyz, rdd.xyz, rad.xyz, sh);
+    sh = segShadow(roo.yzx, rdd.yzx, rad.yzx, sh);
+    sh = segShadow(roo.zxy, rdd.zxy, rad.zxy, sh);
+    sh = clamp(sk * sqrt(sh), 0.0, 1.0);
+    return sh * sh * (3.0 - 2.0 * sh);
+}
+
+static float boxRay(float3 ro, float3 rd, float3 r, thread float3& nn, bool entering) {
+    rd += 0.0001 * (1.0 - abs(sign(rd)));
+    float3 dr = 1.0 / rd;
+    float3 n = ro * dr;
+    float3 k = r * abs(dr);
+
+    float3 pin = -k - n;
+    float3 pout = k - n;
+    float tin = max(pin.x, max(pin.y, pin.z));
+    float tout = min(pout.x, min(pout.y, pout.z));
+    if (tin > tout) return -1.0;
+    if (entering) {
+        nn = -sign(rd) * step(pin.zxy, pin.xyz) * step(pin.yzx, pin.xyz);
+    } else {
+        nn = sign(rd) * step(pout.xyz, pout.zxy) * step(pout.xyz, pout.yzx);
+    }
+    return entering ? tin : tout;
+}
+
+static float3 bgcol(float3 rd) {
+    return mix(float3(0.01), float3(0.336, 0.458, 0.668), 1.0 - pow(abs(rd.z + 0.25), 1.3));
+}
+
+static float3 background(float3 ro, float3 rd, float3 l_dir, thread float& alpha) {
+    float t = (-BOXDIMS.z - ro.z) / rd.z;
+    alpha = 0.0;
+    float3 bgc = bgcol(rd);
+    if (t < 0.0) return bgc;
+    float2 uv = ro.xy + t * rd.xy;
+    float shad = boxSoftShadow((ro + t * rd), normalize(l_dir + float3(0.0, 0.0, 1.0)) * rotz(PI * 0.65), BOXDIMS, 1.5);
+    float aofac = smoothstep(-0.95, 0.75, length(abs(uv) - min(abs(uv), float2(0.45))));
+    aofac = min(aofac, smoothstep(-0.65, 1.0, shad));
+    float lght = max(dot(normalize(ro + t * rd + float3(0.0, -0.0, -5.0)), normalize(l_dir - float3(0.0, 0.0, 1.0)) * rotz(PI * 0.65)), 0.0);
+    float3 col = mix(float3(0.4), float3(0.71, 0.772, 0.895), lght * lght * aofac + 0.05) * aofac;
+    alpha = 1.0 - smoothstep(7.0, 10.0, length(uv));
+    return mix(col * length(col) * 0.8, bgc, smoothstep(7.0, 10.0, length(uv)));
+}
+
+static float4 insides(float3 ro, float3 rd, float3 nor_c, float3 l_dir, thread float& tout) {
+    tout = -1.0;
+
+    float pi = 3.1415926;
+
+    if (abs(nor_c.x) > 0.5) {
+        rd = rd.xzy * nor_c.x;
+        ro = ro.xzy * nor_c.x;
+    } else if (abs(nor_c.z) > 0.5) {
+        l_dir = l_dir * roty(pi);
+        rd = rd.yxz * nor_c.z;
+        ro = ro.yxz * nor_c.z;
+    } else if (abs(nor_c.y) > 0.5) {
+        l_dir = l_dir * rotz(-pi * 0.5);
+        rd = rd * nor_c.y;
+        ro = ro * nor_c.y;
+    }
+
+    const float curvature = 0.5;
+    float bil_size = 1.0;
+    float4 ps = float4(-bil_size, -bil_size, bil_size, bil_size) * curvature;
+    float4 ph = float4(-bil_size, bil_size, bil_size, -bil_size) * curvature;
+
+    float4 colx[3] = { float4(0.0), float4(0.0), float4(0.0) };
+    float3 dx[3] = { float3(-1.0), float3(-1.0), float3(-1.0) };
+    float4 colxsi[3] = { float4(0.0), float4(0.0), float4(0.0) };
+    int order[3] = { 0, 1, 2 };
+
+    for (int i = 0; i < 3; i++) {
+        if (abs(nor_c.x) > 0.5) {
+            ro = ro * rotz(-pi * (1.0 / 3.0));
+            rd = rd * rotz(-pi * (1.0 / 3.0));
+        } else if (abs(nor_c.z) > 0.5) {
+            ro = ro * rotz(pi * (1.0 / 3.0));
+            rd = rd * rotz(pi * (1.0 / 3.0));
+        } else if (abs(nor_c.y) > 0.5) {
+            ro = ro * rotx(pi * (1.0 / 3.0));
+            rd = rd * rotx(pi * (1.0 / 3.0));
+        }
+        float3 normnew;
+        float tnew;
+        bool si;
+        float tsi;
+        float3 normsi;
+        float fade;
+        float fadesi;
+
+        if (iBilinearPatch(ro, rd, ps, ph, bil_size, tnew, normnew, si, tsi, normsi, fade, fadesi)) {
+            if (tnew > 0.0) {
+                float4 tcol, tcolsi;
+                calcColor(ro, rd, normnew, tnew, bil_size, i, si, tsi, tcol, tcolsi);
+                if (tcol.a > 0.0) {
+                    dx[i] = float3(tnew, float(si), tsi);
+
+                    float dif = clamp(dot(normnew, l_dir), 0.0, 1.0);
+                    float amb = clamp(0.5 + 0.5 * dot(normnew, l_dir), 0.0, 1.0);
+
+                    {
+                        float3 shad = float3(0.32, 0.43, 0.54) * amb + float3(1.0, 0.9, 0.7) * dif;
+                        const float3 tcr = float3(1.0, 0.21, 0.11);
+                        float ta = clamp(length(tcol.rgb), 0.0, 1.0);
+                        tcol = clamp(tcol * tcol * 2.0, 0.0, 1.0);
+                        float4 tvalx = float4(tcol.rgb * shad * 1.4 + 3.0 * (tcr * tcol.rgb) * clamp(1.0 - (amb + dif), 0.0, 1.0), min(tcol.a, ta));
+                        tvalx.rgb = clamp(2.0 * tvalx.rgb * tvalx.rgb, 0.0, 1.0);
+                        tvalx *= min(fade * 5.0, 1.0);
+                        colx[i] = tvalx;
+                    }
+                    if (si) {
+                        dif = clamp(dot(normsi, l_dir), 0.0, 1.0);
+                        amb = clamp(0.5 + 0.5 * dot(normsi, l_dir), 0.0, 1.0);
+                        float3 shad = float3(0.32, 0.43, 0.54) * amb + float3(1.0, 0.9, 0.7) * dif;
+                        const float3 tcr = float3(1.0, 0.21, 0.11);
+                        float ta = clamp(length(tcolsi.rgb), 0.0, 1.0);
+                        tcolsi = clamp(tcolsi * tcolsi * 2.0, 0.0, 1.0);
+                        float4 tvalx = float4(tcolsi.rgb * shad + 3.0 * (tcr * tcolsi.rgb) * clamp(1.0 - (amb + dif), 0.0, 1.0), min(tcolsi.a, ta));
+                        tvalx.rgb = clamp(2.0 * tvalx.rgb * tvalx.rgb, 0.0, 1.0);
+                        tvalx.rgb *= min(fadesi * 5.0, 1.0);
+                        colxsi[i] = tvalx;
+                    }
+                }
+            }
+        }
+    }
+
+    // sort by dx[*].x descending (3 passes of bubble sort like the GLSL)
+    if (dx[0].x < dx[1].x) { float3 tv = dx[0]; dx[0] = dx[1]; dx[1] = tv; int ti = order[0]; order[0] = order[1]; order[1] = ti; }
+    if (dx[1].x < dx[2].x) { float3 tv = dx[1]; dx[1] = dx[2]; dx[2] = tv; int ti = order[1]; order[1] = order[2]; order[2] = ti; }
+    if (dx[0].x < dx[1].x) { float3 tv = dx[0]; dx[0] = dx[1]; dx[1] = tv; int ti = order[0]; order[0] = order[1]; order[1] = ti; }
+
+    tout = max(max(dx[0].x, dx[1].x), dx[2].x);
+
+    float a = 1.0;
+    if (dx[0].y < 0.5) {
+        a = colx[order[0]].a;
+    }
+
+    bool rul[3] = {
+        ((dx[0].y > 0.5) && (dx[1].x <= 0.0)),
+        ((dx[1].y > 0.5) && (dx[0].x > dx[1].z)),
+        ((dx[2].y > 0.5) && (dx[1].x > dx[2].z))
+    };
+    for (int k = 0; k < 3; k++) {
+        if (rul[k]) {
+            float4 tcolxsi = colxsi[order[k]];
+            float4 tcolx = colx[order[k]];
+            float4 tvalx = mix(tcolxsi, tcolx, tcolx.a);
+            float4 tvalx2 = mix(float4(0.0), tvalx, max(tcolx.a, tcolxsi.a));
+            colx[order[k]] = tvalx2;
+        }
+    }
+
+    float a1 = (dx[1].y < 0.5) ? colx[order[1]].a : ((dx[1].z > dx[0].x) ? colx[order[1]].a : 1.0);
+    float a2 = (dx[2].y < 0.5) ? colx[order[2]].a : ((dx[2].z > dx[1].x) ? colx[order[2]].a : 1.0);
+    float3 col = mix(mix(colx[order[0]].rgb, colx[order[1]].rgb, a1), colx[order[2]].rgb, a2);
+    a = max(max(a, a1), a2);
+    return float4(col, a);
+}
+
+fragment float4 cube_lines_fs(PassthroughVOut in [[stage_in]],
+                               constant ShadertoyUBO& U [[buffer(0)]]) {
+    float2 fragCoord = in.uv * U.iResolution;
+    float iTime = U.iTime;
+
+    float3 l_dir = normalize(float3(0.0, 1.0, 0.0));
+    l_dir = l_dir * rotz(0.5);
+    float mouseY = PI * 0.49 - smoothstep(0.0, 8.5, fmod((iTime + tshift) * 0.33, 25.0))
+                            * (1.0 - smoothstep(14.0, 24.0, fmod((iTime + tshift) * 0.33, 25.0))) * 0.55 * PI;
+    float mouseX = -2.0 * PI - 0.25 * (iTime * ROTATION_SPEED + tshift);
+
+    float3 eye = 4.0 * float3(cos(mouseX) * cos(mouseY), sin(mouseX) * cos(mouseY), sin(mouseY));
+    float3 w = normalize(-eye);
+    float3 up = float3(0.0, 0.0, 1.0);
+    float3 u = normalize(cross(w, up));
+    float3 v = cross(u, w);
+
+    float4 tot = float4(0.0);
+    float2 uv = (fragCoord - 0.5 * U.iResolution) / U.iResolution.x;
+    float3 rd = normalize(w * FDIST + uv.x * u + uv.y * v);
+
+    float3 ni;
+    float t = boxRay(eye, rd, BOXDIMS, ni, true);
+    float3 ro = eye + t * rd;
+    float2 coords = ro.xy * ni.z / BOXDIMS.xy + ro.yz * ni.x / BOXDIMS.yz + ro.zx * ni.y / BOXDIMS.zx;
+    float fadeborders = (1.0 - smoothstep(0.915, 1.05, abs(coords.x))) * (1.0 - smoothstep(0.915, 1.05, abs(coords.y)));
+
+    if (t > 0.0) {
+        float3 col = float3(0.0);
+        float R0 = (IOR - 1.0) / (IOR + 1.0);
+        R0 *= R0;
+
+        float2 theta = float2(0.0);
+        float3 n = float3(cos(theta.x) * sin(theta.y), sin(theta.x) * sin(theta.y), cos(theta.y));
+
+        float3 nr = n.zxy * ni.x + n.yzx * ni.y + n.xyz * ni.z;
+        float3 rdr = reflect(rd, nr);
+        float talpha;
+        float3 reflcol = background(ro, rdr, l_dir, talpha);
+
+        float3 rd2 = refract(rd, nr, 1.0 / IOR);
+
+        float accum = 1.0;
+        float3 no2 = ni;
+        float3 ro_refr = ro;
+
+        float4 colo[2] = { float4(0.0), float4(0.0) };
+
+        for (int j = 0; j < 2; j++) {
+            float tb;
+            float2 coords2 = ro_refr.xy * no2.z + ro_refr.yz * no2.x + ro_refr.zx * no2.y;
+            float3 eye2 = float3(coords2, -1.0);
+            float3 rd2trans = rd2.yzx * no2.x + rd2.zxy * no2.y + rd2.xyz * no2.z;
+
+            rd2trans.z = -rd2trans.z;
+            float4 internalcol = insides(eye2, rd2trans, no2, l_dir, tb);
+            if (tb > 0.0) {
+                internalcol.rgb *= accum;
+                colo[j] = internalcol;
+            }
+
+            if ((tb <= 0.0) || (internalcol.a < 1.0)) {
+                float tout = boxRay(ro_refr, rd2, BOXDIMS, no2, false);
+                no2 = n.zyx * no2.x + n.xzy * no2.y + n.yxz * no2.z;
+                float3 rout = ro_refr + tout * rd2;
+                float3 rdout = refract(rd2, -no2, IOR);
+                float fresnel2 = R0 + (1.0 - R0) * pow(1.0 - dot(rdout, no2), 1.3);
+                rd2 = reflect(rd2, -no2);
+
+                ro_refr = rout;
+                ro_refr.z = max(ro_refr.z, -0.999);
+
+                accum *= fresnel2;
+            }
+        }
+        float fresnel = R0 + (1.0 - R0) * pow(1.0 - dot(-rd, nr), 5.0);
+        col = mix(mix(colo[1].rgb * colo[1].a, colo[0].rgb, colo[0].a) * fadeborders, reflcol, pow(fresnel, 1.5));
+        col = clamp(col, 0.0, 1.0);
+
+        float cineshader_alpha = clamp(0.15 * dot(eye, ro), 0.0, 1.0);
+        tot += float4(col, cineshader_alpha);
+    } else {
+        float alpha;
+        tot += float4(background(eye, rd, l_dir, alpha), 0.15);
+    }
+
+    float4 outColor = tot;
+    outColor.rgb = clamp(outColor.rgb, 0.0, 1.0);
+    return outColor;
+}

data/shaders/cube_lines/cube_lines.gl.glsl

@@ -0,0 +1,778 @@
+// Name: Cube lines
+// Author: Danil
+// URL: https://www.shadertoy.com/view/NslGRN
+#version 330 core
+precision highp float;
+
+out vec4 FragColor;
+in vec2 vUV;
+uniform vec2 iResolution;
+uniform float iTime;
+
+// Note: Original shader uses iChannel0 for background blending (optional feature)
+// Since we don't support texture channels, that line is commented out
+// The shader works perfectly without it in default mode
+
+// Created by Danil (2021+) https://cohost.org/arugl
+// License Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License.
+// self https://www.shadertoy.com/view/NslGRN
+
+
+// --defines for "DESKTOP WALLPAPERS" that use this shader--
+// comment or uncomment every define to make it work (add or remove "//" before #define)
+
+
+// this shadertoy use ALPHA, NO_ALPHA set alpha to 1, BG_ALPHA set background as alpha
+// iChannel0 used as background if alpha ignored by wallpaper-app
+//#define NO_ALPHA
+//#define BG_ALPHA
+//#define SHADOW_ALPHA
+//#define ONLY_BOX
+
+
+// save PERFORMANCE by disabling shadow
+//#define NO_SHADOW
+
+
+// static CAMERA position, 0.49 on top, 0.001 horizontal
+//#define CAMERA_POS 0.049
+
+
+// speed of ROTATION
+#define ROTATION_SPEED 0.8999
+
+
+// static SHAPE form, default 0.5
+//#define STATIC_SHAPE 0.15
+
+
+// static SCALE far/close to camera, 2.0 is default, exampe 0.5 or 10.0
+//#define CAMERA_FAR 0.1
+
+
+// ANIMATION shape change
+//#define ANIM_SHAPE
+
+
+// ANIMATION color change
+//#define ANIM_COLOR
+
+
+// custom COLOR, and change those const values
+//#define USE_COLOR
+const vec3 color_blue=vec3(0.5,0.65,0.8);
+const vec3 color_red=vec3(0.99,0.2,0.1);
+
+
+// use 4xAA for cube only (set 2-4-etc level of AA)
+//#define AA_CUBE 4
+
+// use 4xAA for everything - derivative filtering will not be used, look fcos2
+// this is very slow - DO NOT USE
+//#define AA_ALL 4
+
+
+
+// --shader code--
+
+// Layers sorted and support transparency and self-intersection-transparency
+// Antialiasing is only dFd. (with some dFd fixes around edges)
+
+// using iq's intersectors: https://iquilezles.org/articles/intersectors
+// using https://www.shadertoy.com/view/ltKBzG
+// using https://www.shadertoy.com/view/tsVXzh
+// using https://www.shadertoy.com/view/WlffDn
+// using https://www.shadertoy.com/view/WslGz4
+
+#define tshift 53.
+
+// reflect back side
+//#define backside_refl
+
+// Camera with mouse
+// #define MOUSE_control (disabled - no iMouse support)
+
+// min(iFrame,0) does not speedup compilation in ANGLE
+#define ANGLE_loops 0
+
+
+// this shader discover Nvidia bug with arrays https://www.shadertoy.com/view/NslGR4
+// use DEBUG with BUG, BUG trigger that bug and one layer will be white on Nvidia in OpenGL
+//#define DEBUG
+//#define BUG
+
+#define FDIST 0.7
+#define PI 3.1415926
+#define GROUNDSPACING 0.5
+#define GROUNDGRID 0.05
+#define BOXDIMS vec3(0.75, 0.75, 1.25)
+
+#define IOR 1.33
+
+mat3 rotx(float a){float s = sin(a);float c = cos(a);return mat3(vec3(1.0, 0.0, 0.0), vec3(0.0, c, s), vec3(0.0, -s, c));  }
+mat3 roty(float a){float s = sin(a);float c = cos(a);return mat3(vec3(c, 0.0, s), vec3(0.0, 1.0, 0.0), vec3(-s, 0.0, c));}
+mat3 rotz(float a){float s = sin(a);float c = cos(a);return mat3(vec3(c, s, 0.0), vec3(-s, c, 0.0), vec3(0.0, 0.0, 1.0 ));}
+
+vec3 fcos1(vec3 x) {
+    vec3 w = fwidth(x);
+    //if((length(w)==0.))return vec3(0.); // dFd fix2
+    //w*=0.; //test
+    float lw=length(w);
+    if((lw==0.)||isnan(lw)||isinf(lw)){vec3 tc=vec3(0.); for(int i=0;i<8;i++)tc+=cos(x+x*float(i-4)*(0.01*400./iResolution.y));return tc/8.;}
+
+    return cos(x) * smoothstep(3.14 * 2.0, 0.0, w);
+}
+
+vec3 fcos2( vec3 x){return cos(x);}
+vec3 fcos( vec3 x){
+#ifdef AA_ALL
+    return fcos2(x);
+#else
+    return fcos1(x);
+#endif
+}
+
+vec3 getColor(vec3 p)
+{
+    // dFd fix, dFd broken on borders, but it fix only top level dFd, self intersection has border
+    //if (length(p) > 0.99)return vec3(0.);
+    p = abs(p);
+
+    p *= 01.25;
+    p = 0.5 * p / dot(p, p);
+#ifdef ANIM_COLOR
+    p+=0.072*iTime;
+#endif
+
+    float t = (0.13) * length(p);
+    vec3 col = vec3(0.3, 0.4, 0.5);
+    col += 0.12 * fcos(6.28318 * t * 1.0 + vec3(0.0, 0.8, 1.1));
+    col += 0.11 * fcos(6.28318 * t * 3.1 + vec3(0.3, 0.4, 0.1));
+    col += 0.10 * fcos(6.28318 * t * 5.1 + vec3(0.1, 0.7, 1.1));
+    col += 0.10 * fcos(6.28318 * t * 17.1 + vec3(0.2, 0.6, 0.7));
+    col += 0.10 * fcos(6.28318 * t * 31.1 + vec3(0.1, 0.6, 0.7));
+    col += 0.10 * fcos(6.28318 * t * 65.1 + vec3(0.0, 0.5, 0.8));
+    col += 0.10 * fcos(6.28318 * t * 115.1 + vec3(0.1, 0.4, 0.7));
+    col += 0.10 * fcos(6.28318 * t * 265.1 + vec3(1.1, 1.4, 2.7));
+    col = clamp(col, 0., 1.);
+
+    return col;
+}
+
+void calcColor(vec3 ro, vec3 rd, vec3 nor, float d, float len, int idx, bool si, float td, out vec4 colx,
+               out vec4 colsi)
+{
+
+    vec3 pos = (ro + rd * d);
+#ifdef DEBUG
+    float a = 1. - smoothstep(len - 0.15, len + 0.00001, length(pos));
+    if (idx == 0)colx = vec4(1., 0., 0., a);
+    if (idx == 1)colx = vec4(0., 1., 0., a);
+    if (idx == 2)colx = vec4(0., 0., 1., a);
+    if (si)
+    {
+        pos = (ro + rd * td);
+        float ta = 1. - smoothstep(len - 0.15, len + 0.00001, length(pos));
+        if (idx == 0)colsi = vec4(1., 0., 0., ta);
+        if (idx == 1)colsi = vec4(0., 1., 0., ta);
+        if (idx == 2)colsi = vec4(0., 0., 1., ta);
+    }
+#else
+    float a = 1. - smoothstep(len - 0.15*0.5, len + 0.00001, length(pos));
+    //a=1.;
+    vec3 col = getColor(pos);
+    colx = vec4(col, a);
+    if (si)
+    {
+        pos = (ro + rd * td);
+        float ta = 1. - smoothstep(len - 0.15*0.5, len + 0.00001, length(pos));
+        //ta=1.;
+        col = getColor(pos);
+        colsi = vec4(col, ta);
+    }
+#endif
+}
+
+// xSI is self intersect data, fade to fix dFd on edges
+bool iBilinearPatch(in vec3 ro, in vec3 rd, in vec4 ps, in vec4 ph, in float sz, out float t, out vec3 norm,
+                    out bool si, out float tsi, out vec3 normsi, out float fade, out float fadesi)
+{
+    vec3 va = vec3(0.0, 0.0, ph.x + ph.w - ph.y - ph.z);
+    vec3 vb = vec3(0.0, ps.w - ps.y, ph.z - ph.x);
+    vec3 vc = vec3(ps.z - ps.x, 0.0, ph.y - ph.x);
+    vec3 vd = vec3(ps.xy, ph.x);
+    t = -1.;
+    tsi = -1.;
+    si = false;
+    fade = 1.;
+    fadesi = 1.;
+    norm=vec3(0.,1.,0.);normsi=vec3(0.,1.,0.);
+
+    float tmp = 1.0 / (vb.y * vc.x);
+    float a = 0.0;
+    float b = 0.0;
+    float c = 0.0;
+    float d = va.z * tmp;
+    float e = 0.0;
+    float f = 0.0;
+    float g = (vc.z * vb.y - vd.y * va.z) * tmp;
+    float h = (vb.z * vc.x - va.z * vd.x) * tmp;
+    float i = -1.0;
+    float j = (vd.x * vd.y * va.z + vd.z * vb.y * vc.x) * tmp - (vd.y * vb.z * vc.x + vd.x * vc.z * vb.y) * tmp;
+
+    float p = dot(vec3(a, b, c), rd.xzy * rd.xzy) + dot(vec3(d, e, f), rd.xzy * rd.zyx);
+    float q = dot(vec3(2.0, 2.0, 2.0) * ro.xzy * rd.xyz, vec3(a, b, c)) + dot(ro.xzz * rd.zxy, vec3(d, d, e)) +
+              dot(ro.yyx * rd.zxy, vec3(e, f, f)) + dot(vec3(g, h, i), rd.xzy);
+    float r =
+        dot(vec3(a, b, c), ro.xzy * ro.xzy) + dot(vec3(d, e, f), ro.xzy * ro.zyx) + dot(vec3(g, h, i), ro.xzy) + j;
+
+    if (abs(p) < 0.000001)
+    {
+        float tt = -r / q;
+        if (tt <= 0.)
+            return false;
+        t = tt;
+        // normal
+
+        vec3 pos = ro + t * rd;
+        if(length(pos)>sz)return false;
+        vec3 grad =
+            vec3(2.0) * pos.xzy * vec3(a, b, c) + pos.zxz * vec3(d, d, e) + pos.yyx * vec3(f, e, f) + vec3(g, h, i);
+        norm = -normalize(grad);
+        return true;
+    }
+    else
+    {
+        float sq = q * q - 4.0 * p * r;
+        if (sq < 0.0)
+        {
+            return false;
+        }
+        else
+        {
+            float s = sqrt(sq);
+            float t0 = (-q + s) / (2.0 * p);
+            float t1 = (-q - s) / (2.0 * p);
+            float tt1 = min(t0 < 0.0 ? t1 : t0, t1 < 0.0 ? t0 : t1);
+            float tt2 = max(t0 > 0.0 ? t1 : t0, t1 > 0.0 ? t0 : t1);
+            float tt0 = tt1;
+            if (tt0 <= 0.)
+                return false;
+            vec3 pos = ro + tt0 * rd;
+            // black border on end of circle and self intersection with alpha come because dFd
+            // uncomment this to see or rename fcos2 to fcos
+            //sz+=0.3;
+            bool ru = step(sz, length(pos)) > 0.5;
+            if (ru)
+            {
+                tt0 = tt2;
+                pos = ro + tt0 * rd;
+            }
+            if (tt0 <= 0.)
+                return false;
+            bool ru2 = step(sz, length(pos)) > 0.5;
+            if (ru2)
+                return false;
+
+            // self intersect
+            if ((tt2 > 0.) && ((!ru)) && !(step(sz, length(ro + tt2 * rd)) > 0.5))
+            {
+                si = true;
+                fadesi=s;
+                tsi = tt2;
+                vec3 tpos = ro + tsi * rd;
+                // normal
+                vec3 tgrad = vec3(2.0) * tpos.xzy * vec3(a, b, c) + tpos.zxz * vec3(d, d, e) +
+                             tpos.yyx * vec3(f, e, f) + vec3(g, h, i);
+                normsi = -normalize(tgrad);
+            }
+
+            fade=s;
+            t = tt0;
+            // normal
+            vec3 grad =
+                vec3(2.0) * pos.xzy * vec3(a, b, c) + pos.zxz * vec3(d, d, e) + pos.yyx * vec3(f, e, f) + vec3(g, h, i);
+            norm = -normalize(grad);
+
+            return true;
+        }
+    }
+}
+
+float dot2( in vec3 v ) { return dot(v,v); }
+
+float segShadow( in vec3 ro, in vec3 rd, in vec3 pa, float sh )
+{
+    float dm = dot(rd.yz,rd.yz);
+    float k1 = (ro.x-pa.x)*dm;
+    float k2 = (ro.x+pa.x)*dm;
+    vec2  k5 = (ro.yz+pa.yz)*dm;
+    float k3 = dot(ro.yz+pa.yz,rd.yz);
+    vec2  k4 = (pa.yz+pa.yz)*rd.yz;
+    vec2  k6 = (pa.yz+pa.yz)*dm;
+
+    for( int i=0; i<4 + ANGLE_loops; i++ )
+    {
+        vec2  s = vec2(i&1,i>>1);
+        float t = dot(s,k4) - k3;
+
+        if( t>0.0 )
+        sh = min(sh,dot2(vec3(clamp(-rd.x*t,k1,k2),k5-k6*s)+rd*t)/(t*t));
+    }
+    return sh;
+}
+
+float boxSoftShadow( in vec3 ro, in vec3 rd, in vec3 rad, in float sk )
+{
+    rd += 0.0001 * (1.0 - abs(sign(rd)));
+    vec3 rdd = rd;
+    vec3 roo = ro;
+
+    vec3 m = 1.0/rdd;
+    vec3 n = m*roo;
+    vec3 k = abs(m)*rad;
+
+    vec3 t1 = -n - k;
+    vec3 t2 = -n + k;
+
+    float tN = max( max( t1.x, t1.y ), t1.z );
+    float tF = min( min( t2.x, t2.y ), t2.z );
+
+    if( tN<tF && tF>0.0) return 0.0;
+
+    float sh = 1.0;
+    sh = segShadow( roo.xyz, rdd.xyz, rad.xyz, sh );
+    sh = segShadow( roo.yzx, rdd.yzx, rad.yzx, sh );
+    sh = segShadow( roo.zxy, rdd.zxy, rad.zxy, sh );
+    sh = clamp(sk*sqrt(sh),0.0,1.0);
+    return sh*sh*(3.0-2.0*sh);
+}
+
+float box(in vec3 ro, in vec3 rd, in vec3 r, out vec3 nn, bool entering)
+{
+    rd += 0.0001 * (1.0 - abs(sign(rd)));
+    vec3 dr = 1.0 / rd;
+    vec3 n = ro * dr;
+    vec3 k = r * abs(dr);
+
+    vec3 pin = -k - n;
+    vec3 pout = k - n;
+    float tin = max(pin.x, max(pin.y, pin.z));
+    float tout = min(pout.x, min(pout.y, pout.z));
+    if (tin > tout)
+        return -1.;
+    if (entering)
+    {
+        nn = -sign(rd) * step(pin.zxy, pin.xyz) * step(pin.yzx, pin.xyz);
+    }
+    else
+    {
+        nn = sign(rd) * step(pout.xyz, pout.zxy) * step(pout.xyz, pout.yzx);
+    }
+    return entering ? tin : tout;
+}
+
+vec3 bgcol(in vec3 rd)
+{
+    return mix(vec3(0.01), vec3(0.336, 0.458, .668), 1. - pow(abs(rd.z+0.25), 1.3));
+}
+
+vec3 background(in vec3 ro, in vec3 rd , vec3 l_dir, out float alpha)
+{
+#ifdef ONLY_BOX
+alpha=0.;
+return vec3(0.01);
+#endif
+    float t = (-BOXDIMS.z - ro.z) / rd.z;
+    alpha=0.;
+    vec3 bgc = bgcol(rd);
+    if (t < 0.)
+        return bgc;
+    vec2 uv = ro.xy + t * rd.xy;
+#ifdef NO_SHADOW
+    float shad=1.;
+#else
+    float shad = boxSoftShadow((ro + t * rd), normalize(l_dir+vec3(0.,0.,1.))*rotz(PI*0.65) , BOXDIMS, 1.5);
+#endif
+    float aofac = smoothstep(-0.95, .75, length(abs(uv) - min(abs(uv), vec2(0.45))));
+    aofac = min(aofac,smoothstep(-0.65, 1., shad));
+    float lght=max(dot(normalize(ro + t * rd+vec3(0.,-0.,-5.)), normalize(l_dir-vec3(0.,0.,1.))*rotz(PI*0.65)), 0.0);
+    vec3 col = mix(vec3(0.4), vec3(.71,.772,0.895), lght*lght* aofac+ 0.05) * aofac;
+    alpha=1.-smoothstep(7.,10.,length(uv));
+#ifdef SHADOW_ALPHA
+    //alpha=clamp(alpha*max(lght*lght*0.95,(1.-aofac)*1.25),0.,1.);
+    alpha=clamp(alpha*(1.-aofac)*1.25,0.,1.);
+#endif
+    return mix(col*length(col)*0.8,bgc,smoothstep(7.,10.,length(uv)));
+}
+
+#define swap(a,b) tv=a;a=b;b=tv
+
+vec4 insides(vec3 ro, vec3 rd, vec3 nor_c, vec3 l_dir, out float tout)
+{
+    tout = -1.;
+    vec3 trd=rd;
+
+    vec3 col = vec3(0.);
+
+    float pi = 3.1415926;
+
+    if (abs(nor_c.x) > 0.5)
+    {
+        rd = rd.xzy * nor_c.x;
+        ro = ro.xzy * nor_c.x;
+    }
+    else if (abs(nor_c.z) > 0.5)
+    {
+        l_dir *= roty(pi);
+        rd = rd.yxz * nor_c.z;
+        ro = ro.yxz * nor_c.z;
+    }
+    else if (abs(nor_c.y) > 0.5)
+    {
+        l_dir *= rotz(-pi * 0.5);
+        rd = rd * nor_c.y;
+        ro = ro * nor_c.y;
+    }
+
+#ifdef ANIM_SHAPE
+    float curvature = (0.001+1.5-1.5*smoothstep(0.,8.5,mod((iTime+tshift)*0.44,20.))*(1.-smoothstep(10.,18.5,mod((iTime+tshift)*0.44,20.))));
+    // curvature(to not const above) make compilation on Angle 15+ sec
+#else
+#ifdef STATIC_SHAPE
+    const float curvature = STATIC_SHAPE;
+#else
+    const float curvature = .5;
+#endif
+#endif
+    float bil_size = 1.;
+    vec4 ps = vec4(-bil_size, -bil_size, bil_size, bil_size) * curvature;
+    vec4 ph = vec4(-bil_size, bil_size, bil_size, -bil_size) * curvature;
+
+    vec4 [3]colx=vec4[3](vec4(0.),vec4(0.),vec4(0.));
+    vec3 [3]dx=vec3[3](vec3(-1.),vec3(-1.),vec3(-1.));
+    vec4 [3]colxsi=vec4[3](vec4(0.),vec4(0.),vec4(0.));
+    int [3]order=int[3](0,1,2);
+
+    for (int i = 0; i < 3 + ANGLE_loops; i++)
+    {
+        if (abs(nor_c.x) > 0.5)
+        {
+            ro *= rotz(-pi * (1. / float(3)));
+            rd *= rotz(-pi * (1. / float(3)));
+        }
+        else if (abs(nor_c.z) > 0.5)
+        {
+            ro *= rotz(pi * (1. / float(3)));
+            rd *= rotz(pi * (1. / float(3)));
+        }
+        else if (abs(nor_c.y) > 0.5)
+        {
+            ro *= rotx(pi * (1. / float(3)));
+            rd *= rotx(pi * (1. / float(3)));
+        }
+        vec3 normnew;
+        float tnew;
+        bool si;
+        float tsi;
+        vec3 normsi;
+        float fade;
+        float fadesi;
+
+        if (iBilinearPatch(ro, rd, ps, ph, bil_size, tnew, normnew, si, tsi, normsi, fade, fadesi))
+        {
+            if (tnew > 0.)
+            {
+                vec4 tcol, tcolsi;
+                calcColor(ro, rd, normnew, tnew, bil_size, i, si, tsi, tcol, tcolsi);
+                if (tcol.a > 0.0)
+                {
+                    {
+                        vec3 tvalx = vec3(tnew, float(si), tsi);
+                        dx[i]=tvalx;
+                    }
+#ifdef DEBUG
+                    colx[i]=tcol;
+                    if (si)colxsi[i]=tcolsi;
+#else
+
+                    float dif = clamp(dot(normnew, l_dir), 0.0, 1.0);
+                    float amb = clamp(0.5 + 0.5 * dot(normnew, l_dir), 0.0, 1.0);
+
+                    {
+#ifdef USE_COLOR
+                        vec3 shad = 0.57 * color_blue * amb + 1.5*color_blue.bgr * dif;
+                        const vec3 tcr = color_red;
+#else
+                        vec3 shad = vec3(0.32, 0.43, 0.54) * amb + vec3(1.0, 0.9, 0.7) * dif;
+                        const vec3 tcr = vec3(1.,0.21,0.11);
+#endif
+                        float ta = clamp(length(tcol.rgb),0.,1.);
+                        tcol=clamp(tcol*tcol*2.,0.,1.);
+                        vec4 tvalx =
+                            vec4((tcol.rgb*shad*1.4 + 3.*(tcr*tcol.rgb)*clamp(1.-(amb+dif),0.,1.)), min(tcol.a,ta));
+                        tvalx.rgb=clamp(2.*tvalx.rgb*tvalx.rgb,0.,1.);
+                        tvalx*=(min(fade*5.,1.));
+                        colx[i]=tvalx;
+                    }
+                    if (si)
+                    {
+                        dif = clamp(dot(normsi, l_dir), 0.0, 1.0);
+                        amb = clamp(0.5 + 0.5 * dot(normsi, l_dir), 0.0, 1.0);
+                        {
+#ifdef USE_COLOR
+                            vec3 shad = 0.57 * color_blue * amb + 1.5*color_blue.bgr * dif;
+                            const vec3 tcr = color_red;
+#else
+                            vec3 shad = vec3(0.32, 0.43, 0.54) * amb + vec3(1.0, 0.9, 0.7) * dif;
+                            const vec3 tcr = vec3(1.,0.21,0.11);
+#endif
+                            float ta = clamp(length(tcolsi.rgb),0.,1.);
+                            tcolsi=clamp(tcolsi*tcolsi*2.,0.,1.);
+                            vec4 tvalx =
+                                vec4(tcolsi.rgb * shad + 3.*(tcr*tcolsi.rgb)*clamp(1.-(amb+dif),0.,1.), min(tcolsi.a,ta));
+                            tvalx.rgb=clamp(2.*tvalx.rgb*tvalx.rgb,0.,1.);
+                            tvalx.rgb*=(min(fadesi*5.,1.));
+                            colxsi[i]=tvalx;
+                        }
+                    }
+#endif
+                }
+            }
+        }
+    }
+    // transparency logic and layers sorting
+    float a = 1.;
+    if (dx[0].x < dx[1].x){{vec3 swap(dx[0], dx[1]);}{int swap(order[0], order[1]);}}
+    if (dx[1].x < dx[2].x){{vec3 swap(dx[1], dx[2]);}{int swap(order[1], order[2]);}}
+    if (dx[0].x < dx[1].x){{vec3 swap(dx[0], dx[1]);}{int swap(order[0], order[1]);}}
+
+    tout = max(max(dx[0].x, dx[1].x), dx[2].x);
+
+    if (dx[0].y < 0.5)
+    {
+        a=colx[order[0]].a;
+    }
+
+#if !(defined(DEBUG)&&defined(BUG))
+
+    // self intersection
+    bool [3] rul= bool[3](
+        ((dx[0].y > 0.5) && (dx[1].x <= 0.)),
+        ((dx[1].y > 0.5) && (dx[0].x > dx[1].z)),
+        ((dx[2].y > 0.5) && (dx[1].x > dx[2].z))
+    );
+    for(int k=0;k<3;k++){
+        if(rul[k]){
+            vec4 tcolxsi = vec4(0.);
+            tcolxsi=colxsi[order[k]];
+            vec4 tcolx = vec4(0.);
+            tcolx=colx[order[k]];
+
+            vec4 tvalx = mix(tcolxsi, tcolx, tcolx.a);
+            colx[order[k]]=tvalx;
+
+            vec4 tvalx2 = mix(vec4(0.), tvalx, max(tcolx.a, tcolxsi.a));
+            colx[order[k]]=tvalx2;
+        }
+    }
+
+#endif
+
+    float a1 = (dx[1].y < 0.5) ? colx[order[1]].a : ((dx[1].z > dx[0].x) ? colx[order[1]].a : 1.);
+    float a2 = (dx[2].y < 0.5) ? colx[order[2]].a : ((dx[2].z > dx[1].x) ? colx[order[2]].a : 1.);
+    col = mix(mix(colx[order[0]].rgb, colx[order[1]].rgb, a1), colx[order[2]].rgb, a2);
+    a = max(max(a, a1), a2);
+    return vec4(col, a);
+}
+
+void mainImage(out vec4 fragColor, in vec2 fragCoord)
+{
+    float osc = 0.5;
+    vec3 l_dir = normalize(vec3(0., 1., 0.));
+    l_dir *= rotz(0.5);
+    float mouseY = 1.0 * 0.5 * PI;
+    // Note: MOUSE_control disabled (no iMouse support)
+    // #ifdef MOUSE_control
+    // mouseY = (1.0 - 1.15 * iMouse.y / iResolution.y) * 0.5 * PI;
+    // if(iMouse.y < 1.)
+    // #endif
+#ifdef CAMERA_POS
+    mouseY = PI*CAMERA_POS;
+#else
+    mouseY = PI*0.49 - smoothstep(0.,8.5,mod((iTime+tshift)*0.33,25.))*(1.-smoothstep(14.,24.0,mod((iTime+tshift)*0.33,25.))) * 0.55 * PI;
+#endif
+#ifdef ROTATION_SPEED
+    float mouseX = -2.*PI-0.25*(iTime*ROTATION_SPEED+tshift);
+#else
+    float mouseX = -2.*PI-0.25*(iTime+tshift);
+#endif
+    // Note: MOUSE_control disabled (no iMouse support)
+    // #ifdef MOUSE_control
+    // mouseX+=-(iMouse.x / iResolution.x) * 2. * PI;
+    // #endif
+
+#ifdef CAMERA_FAR
+    vec3 eye = (2. + CAMERA_FAR) * vec3(cos(mouseX) * cos(mouseY), sin(mouseX) * cos(mouseY), sin(mouseY));
+#else
+    vec3 eye = 4. * vec3(cos(mouseX) * cos(mouseY), sin(mouseX) * cos(mouseY), sin(mouseY));
+#endif
+    vec3 w = normalize(-eye);
+    vec3 up = vec3(0., 0., 1.);
+    vec3 u = normalize(cross(w, up));
+    vec3 v = cross(u, w);
+
+    vec4 tot=vec4(0.);
+#if defined(AA_CUBE)||defined(AA_ALL)
+#ifdef AA_CUBE
+    const int AA = AA_CUBE;
+#else
+    const int AA = AA_ALL;
+#endif
+    vec3 incol_once=vec3(0.);
+    bool in_once=false;
+    vec4 incolbg_once=vec4(0.);
+    bool bg_in_once=false;
+    vec4 outcolbg_once=vec4(0.);
+    bool bg_out_once=false;
+    for( int mx=0; mx<AA; mx++ )
+    for( int nx=0; nx<AA; nx++ )
+    {
+    vec2 o = vec2(mod(float(mx+AA/2),float(AA)),mod(float(nx+AA/2),float(AA))) / float(AA) - 0.5;
+    vec2 uv = (fragCoord + o - 0.5 * iResolution.xy) / iResolution.x;
+#else
+    vec2 uv = (fragCoord - 0.5 * iResolution.xy) / iResolution.x;
+#endif
+    vec3 rd = normalize(w * FDIST + uv.x * u + uv.y * v);
+
+    vec3 ni;
+    float t = box(eye, rd, BOXDIMS, ni, true);
+    vec3 ro = eye + t * rd;
+    vec2 coords = ro.xy * ni.z/BOXDIMS.xy + ro.yz * ni.x/BOXDIMS.yz + ro.zx * ni.y/BOXDIMS.zx;
+    float fadeborders = (1.-smoothstep(0.915,1.05,abs(coords.x)))*(1.-smoothstep(0.915,1.05,abs(coords.y)));
+
+    if (t > 0.)
+    {
+        float ang = -iTime * 0.33;
+        vec3 col = vec3(0.);
+#ifdef AA_CUBE
+        if(in_once)col=incol_once;
+        else{
+        in_once=true;
+#endif
+        float R0 = (IOR - 1.) / (IOR + 1.);
+        R0 *= R0;
+
+        vec2 theta = vec2(0.);
+        vec3 n = vec3(cos(theta.x) * sin(theta.y), sin(theta.x) * sin(theta.y), cos(theta.y));
+
+        vec3 nr = n.zxy * ni.x + n.yzx * ni.y + n.xyz * ni.z;
+        vec3 rdr = reflect(rd, nr);
+        float talpha;
+        vec3 reflcol = background(ro, rdr, l_dir,talpha);
+
+        vec3 rd2 = refract(rd, nr, 1. / IOR);
+
+        float accum = 1.;
+        vec3 no2 = ni;
+        vec3 ro_refr = ro;
+
+        vec4 [2] colo = vec4[2](vec4(0.),vec4(0.));
+
+        for (int j = 0; j < 2 + ANGLE_loops; j++)
+        {
+            float tb;
+            vec2 coords2 = ro_refr.xy * no2.z + ro_refr.yz * no2.x + ro_refr.zx * no2.y;
+            vec3 eye2 = vec3(coords2, -1.);
+            vec3 rd2trans = rd2.yzx * no2.x + rd2.zxy * no2.y + rd2.xyz * no2.z;
+
+            rd2trans.z = -rd2trans.z;
+            vec4 internalcol = insides(eye2, rd2trans, no2, l_dir, tb);
+            if (tb > 0.)
+            {
+                internalcol.rgb *= accum;
+                colo[j]=internalcol;
+            }
+
+            if ((tb <= 0.) || (internalcol.a < 1.))
+            {
+                float tout = box(ro_refr, rd2, BOXDIMS, no2, false);
+                no2 = n.zyx * no2.x + n.xzy * no2.y + n.yxz * no2.z;
+                vec3 rout = ro_refr + tout * rd2;
+                vec3 rdout = refract(rd2, -no2, IOR);
+                float fresnel2 = R0 + (1. - R0) * pow(1. - dot(rdout, no2), 1.3);
+                rd2 = reflect(rd2, -no2);
+
+#ifdef backside_refl
+                if((dot(rdout, no2))>0.5){fresnel2=1.;}
+#endif
+                ro_refr = rout;
+                ro_refr.z = max(ro_refr.z, -0.999);
+
+                accum *= fresnel2;
+            }
+        }
+        float fresnel = R0 + (1. - R0) * pow(1. - dot(-rd, nr), 5.);
+        col = mix(mix(colo[1].rgb * colo[1].a, colo[0].rgb, colo[0].a)*fadeborders, reflcol, pow(fresnel, 1.5));
+        col=clamp(col,0.,1.);
+#ifdef AA_CUBE
+        }
+        incol_once=col;
+        if(!bg_in_once){
+        bg_in_once=true;
+        float alpha;
+        incolbg_once = vec4(background(eye, rd, l_dir, alpha), 0.15);
+#if defined(BG_ALPHA)||defined(ONLY_BOX)||defined(SHADOW_ALPHA)
+        incolbg_once.w = alpha;
+#endif
+        }
+#endif
+
+        float cineshader_alpha = 0.;
+        cineshader_alpha = clamp(0.15*dot(eye,ro),0.,1.);
+        vec4 tcolx = vec4(col, cineshader_alpha);
+#if defined(BG_ALPHA)||defined(ONLY_BOX)||defined(SHADOW_ALPHA)
+        tcolx.w = 1.;
+#endif
+        tot += tcolx;
+    }
+    else
+    {
+        vec4 tcolx = vec4(0.);
+#ifdef AA_CUBE
+        if(!bg_out_once){
+        bg_out_once=true;
+#endif
+        float alpha;
+        tcolx = vec4(background(eye, rd, l_dir, alpha), 0.15);
+#if defined(BG_ALPHA)||defined(ONLY_BOX)||defined(SHADOW_ALPHA)
+        tcolx.w = alpha;
+#endif
+#ifdef AA_CUBE
+        outcolbg_once=tcolx;
+        }else tcolx=max(outcolbg_once,incolbg_once);
+#endif
+        tot += tcolx;
+    }
+#if defined(AA_CUBE)||defined(AA_ALL)
+    }
+    tot /= float(AA*AA);
+#endif
+    fragColor = tot;
+#ifdef NO_ALPHA
+    fragColor.w = 1.;
+#endif
+    fragColor.rgb=clamp(fragColor.rgb,0.,1.);
+    // Note: iChannel0 line removed (texture channel not supported)
+    // Original line was for optional background blending when BG_ALPHA/ONLY_BOX/SHADOW_ALPHA defined
+    // #if defined(BG_ALPHA)||defined(ONLY_BOX)||defined(SHADOW_ALPHA)
+    // fragColor.rgb=fragColor.rgb*fragColor.w+texture(iChannel0, fragCoord/iResolution.xy).rgb*(1.-fragColor.w);
+    // #endif
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

data/shaders/cube_lines/cube_lines.vk.glsl

@@ -0,0 +1,782 @@
+#version 450
+
+// Name: Cube lines
+// Author: Danil
+// URL: https://www.shadertoy.com/view/NslGRN
+
+layout(location = 0) in  vec2 vUV;
+layout(location = 0) out vec4 FragColor;
+
+layout(set = 3, binding = 0) uniform ShadertoyUBO {
+    float iTime;
+    vec2  iResolution;
+};
+
+
+// Note: Original shader uses iChannel0 for background blending (optional feature)
+// Since we don't support texture channels, that line is commented out
+// The shader works perfectly without it in default mode
+
+// Created by Danil (2021+) https://cohost.org/arugl
+// License Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License.
+// self https://www.shadertoy.com/view/NslGRN
+
+
+// --defines for "DESKTOP WALLPAPERS" that use this shader--
+// comment or uncomment every define to make it work (add or remove "//" before #define)
+
+
+// this shadertoy use ALPHA, NO_ALPHA set alpha to 1, BG_ALPHA set background as alpha
+// iChannel0 used as background if alpha ignored by wallpaper-app
+//#define NO_ALPHA
+//#define BG_ALPHA
+//#define SHADOW_ALPHA
+//#define ONLY_BOX
+
+
+// save PERFORMANCE by disabling shadow
+//#define NO_SHADOW
+
+
+// static CAMERA position, 0.49 on top, 0.001 horizontal
+//#define CAMERA_POS 0.049
+
+
+// speed of ROTATION
+#define ROTATION_SPEED 0.8999
+
+
+// static SHAPE form, default 0.5
+//#define STATIC_SHAPE 0.15
+
+
+// static SCALE far/close to camera, 2.0 is default, exampe 0.5 or 10.0
+//#define CAMERA_FAR 0.1
+
+
+// ANIMATION shape change
+//#define ANIM_SHAPE
+
+
+// ANIMATION color change
+//#define ANIM_COLOR
+
+
+// custom COLOR, and change those const values
+//#define USE_COLOR
+const vec3 color_blue=vec3(0.5,0.65,0.8);
+const vec3 color_red=vec3(0.99,0.2,0.1);
+
+
+// use 4xAA for cube only (set 2-4-etc level of AA)
+//#define AA_CUBE 4
+
+// use 4xAA for everything - derivative filtering will not be used, look fcos2
+// this is very slow - DO NOT USE
+//#define AA_ALL 4
+
+
+
+// --shader code--
+
+// Layers sorted and support transparency and self-intersection-transparency
+// Antialiasing is only dFd. (with some dFd fixes around edges)
+
+// using iq's intersectors: https://iquilezles.org/articles/intersectors
+// using https://www.shadertoy.com/view/ltKBzG
+// using https://www.shadertoy.com/view/tsVXzh
+// using https://www.shadertoy.com/view/WlffDn
+// using https://www.shadertoy.com/view/WslGz4
+
+#define tshift 53.
+
+// reflect back side
+//#define backside_refl
+
+// Camera with mouse
+// #define MOUSE_control (disabled - no iMouse support)
+
+// min(iFrame,0) does not speedup compilation in ANGLE
+#define ANGLE_loops 0
+
+
+// this shader discover Nvidia bug with arrays https://www.shadertoy.com/view/NslGR4
+// use DEBUG with BUG, BUG trigger that bug and one layer will be white on Nvidia in OpenGL
+//#define DEBUG
+//#define BUG
+
+#define FDIST 0.7
+#define PI 3.1415926
+#define GROUNDSPACING 0.5
+#define GROUNDGRID 0.05
+#define BOXDIMS vec3(0.75, 0.75, 1.25)
+
+#define IOR 1.33
+
+mat3 rotx(float a){float s = sin(a);float c = cos(a);return mat3(vec3(1.0, 0.0, 0.0), vec3(0.0, c, s), vec3(0.0, -s, c));  }
+mat3 roty(float a){float s = sin(a);float c = cos(a);return mat3(vec3(c, 0.0, s), vec3(0.0, 1.0, 0.0), vec3(-s, 0.0, c));}
+mat3 rotz(float a){float s = sin(a);float c = cos(a);return mat3(vec3(c, s, 0.0), vec3(-s, c, 0.0), vec3(0.0, 0.0, 1.0 ));}
+
+vec3 fcos1(vec3 x) {
+    vec3 w = fwidth(x);
+    //if((length(w)==0.))return vec3(0.); // dFd fix2
+    //w*=0.; //test
+    float lw=length(w);
+    if((lw==0.)||isnan(lw)||isinf(lw)){vec3 tc=vec3(0.); for(int i=0;i<8;i++)tc+=cos(x+x*float(i-4)*(0.01*400./iResolution.y));return tc/8.;}
+
+    return cos(x) * smoothstep(3.14 * 2.0, 0.0, w);
+}
+
+vec3 fcos2( vec3 x){return cos(x);}
+vec3 fcos( vec3 x){
+#ifdef AA_ALL
+    return fcos2(x);
+#else
+    return fcos1(x);
+#endif
+}
+
+vec3 getColor(vec3 p)
+{
+    // dFd fix, dFd broken on borders, but it fix only top level dFd, self intersection has border
+    //if (length(p) > 0.99)return vec3(0.);
+    p = abs(p);
+
+    p *= 01.25;
+    p = 0.5 * p / dot(p, p);
+#ifdef ANIM_COLOR
+    p+=0.072*iTime;
+#endif
+
+    float t = (0.13) * length(p);
+    vec3 col = vec3(0.3, 0.4, 0.5);
+    col += 0.12 * fcos(6.28318 * t * 1.0 + vec3(0.0, 0.8, 1.1));
+    col += 0.11 * fcos(6.28318 * t * 3.1 + vec3(0.3, 0.4, 0.1));
+    col += 0.10 * fcos(6.28318 * t * 5.1 + vec3(0.1, 0.7, 1.1));
+    col += 0.10 * fcos(6.28318 * t * 17.1 + vec3(0.2, 0.6, 0.7));
+    col += 0.10 * fcos(6.28318 * t * 31.1 + vec3(0.1, 0.6, 0.7));
+    col += 0.10 * fcos(6.28318 * t * 65.1 + vec3(0.0, 0.5, 0.8));
+    col += 0.10 * fcos(6.28318 * t * 115.1 + vec3(0.1, 0.4, 0.7));
+    col += 0.10 * fcos(6.28318 * t * 265.1 + vec3(1.1, 1.4, 2.7));
+    col = clamp(col, 0., 1.);
+
+    return col;
+}
+
+void calcColor(vec3 ro, vec3 rd, vec3 nor, float d, float len, int idx, bool si, float td, out vec4 colx,
+               out vec4 colsi)
+{
+
+    vec3 pos = (ro + rd * d);
+#ifdef DEBUG
+    float a = 1. - smoothstep(len - 0.15, len + 0.00001, length(pos));
+    if (idx == 0)colx = vec4(1., 0., 0., a);
+    if (idx == 1)colx = vec4(0., 1., 0., a);
+    if (idx == 2)colx = vec4(0., 0., 1., a);
+    if (si)
+    {
+        pos = (ro + rd * td);
+        float ta = 1. - smoothstep(len - 0.15, len + 0.00001, length(pos));
+        if (idx == 0)colsi = vec4(1., 0., 0., ta);
+        if (idx == 1)colsi = vec4(0., 1., 0., ta);
+        if (idx == 2)colsi = vec4(0., 0., 1., ta);
+    }
+#else
+    float a = 1. - smoothstep(len - 0.15*0.5, len + 0.00001, length(pos));
+    //a=1.;
+    vec3 col = getColor(pos);
+    colx = vec4(col, a);
+    if (si)
+    {
+        pos = (ro + rd * td);
+        float ta = 1. - smoothstep(len - 0.15*0.5, len + 0.00001, length(pos));
+        //ta=1.;
+        col = getColor(pos);
+        colsi = vec4(col, ta);
+    }
+#endif
+}
+
+// xSI is self intersect data, fade to fix dFd on edges
+bool iBilinearPatch(in vec3 ro, in vec3 rd, in vec4 ps, in vec4 ph, in float sz, out float t, out vec3 norm,
+                    out bool si, out float tsi, out vec3 normsi, out float fade, out float fadesi)
+{
+    vec3 va = vec3(0.0, 0.0, ph.x + ph.w - ph.y - ph.z);
+    vec3 vb = vec3(0.0, ps.w - ps.y, ph.z - ph.x);
+    vec3 vc = vec3(ps.z - ps.x, 0.0, ph.y - ph.x);
+    vec3 vd = vec3(ps.xy, ph.x);
+    t = -1.;
+    tsi = -1.;
+    si = false;
+    fade = 1.;
+    fadesi = 1.;
+    norm=vec3(0.,1.,0.);normsi=vec3(0.,1.,0.);
+
+    float tmp = 1.0 / (vb.y * vc.x);
+    float a = 0.0;
+    float b = 0.0;
+    float c = 0.0;
+    float d = va.z * tmp;
+    float e = 0.0;
+    float f = 0.0;
+    float g = (vc.z * vb.y - vd.y * va.z) * tmp;
+    float h = (vb.z * vc.x - va.z * vd.x) * tmp;
+    float i = -1.0;
+    float j = (vd.x * vd.y * va.z + vd.z * vb.y * vc.x) * tmp - (vd.y * vb.z * vc.x + vd.x * vc.z * vb.y) * tmp;
+
+    float p = dot(vec3(a, b, c), rd.xzy * rd.xzy) + dot(vec3(d, e, f), rd.xzy * rd.zyx);
+    float q = dot(vec3(2.0, 2.0, 2.0) * ro.xzy * rd.xyz, vec3(a, b, c)) + dot(ro.xzz * rd.zxy, vec3(d, d, e)) +
+              dot(ro.yyx * rd.zxy, vec3(e, f, f)) + dot(vec3(g, h, i), rd.xzy);
+    float r =
+        dot(vec3(a, b, c), ro.xzy * ro.xzy) + dot(vec3(d, e, f), ro.xzy * ro.zyx) + dot(vec3(g, h, i), ro.xzy) + j;
+
+    if (abs(p) < 0.000001)
+    {
+        float tt = -r / q;
+        if (tt <= 0.)
+            return false;
+        t = tt;
+        // normal
+
+        vec3 pos = ro + t * rd;
+        if(length(pos)>sz)return false;
+        vec3 grad =
+            vec3(2.0) * pos.xzy * vec3(a, b, c) + pos.zxz * vec3(d, d, e) + pos.yyx * vec3(f, e, f) + vec3(g, h, i);
+        norm = -normalize(grad);
+        return true;
+    }
+    else
+    {
+        float sq = q * q - 4.0 * p * r;
+        if (sq < 0.0)
+        {
+            return false;
+        }
+        else
+        {
+            float s = sqrt(sq);
+            float t0 = (-q + s) / (2.0 * p);
+            float t1 = (-q - s) / (2.0 * p);
+            float tt1 = min(t0 < 0.0 ? t1 : t0, t1 < 0.0 ? t0 : t1);
+            float tt2 = max(t0 > 0.0 ? t1 : t0, t1 > 0.0 ? t0 : t1);
+            float tt0 = tt1;
+            if (tt0 <= 0.)
+                return false;
+            vec3 pos = ro + tt0 * rd;
+            // black border on end of circle and self intersection with alpha come because dFd
+            // uncomment this to see or rename fcos2 to fcos
+            //sz+=0.3;
+            bool ru = step(sz, length(pos)) > 0.5;
+            if (ru)
+            {
+                tt0 = tt2;
+                pos = ro + tt0 * rd;
+            }
+            if (tt0 <= 0.)
+                return false;
+            bool ru2 = step(sz, length(pos)) > 0.5;
+            if (ru2)
+                return false;
+
+            // self intersect
+            if ((tt2 > 0.) && ((!ru)) && !(step(sz, length(ro + tt2 * rd)) > 0.5))
+            {
+                si = true;
+                fadesi=s;
+                tsi = tt2;
+                vec3 tpos = ro + tsi * rd;
+                // normal
+                vec3 tgrad = vec3(2.0) * tpos.xzy * vec3(a, b, c) + tpos.zxz * vec3(d, d, e) +
+                             tpos.yyx * vec3(f, e, f) + vec3(g, h, i);
+                normsi = -normalize(tgrad);
+            }
+
+            fade=s;
+            t = tt0;
+            // normal
+            vec3 grad =
+                vec3(2.0) * pos.xzy * vec3(a, b, c) + pos.zxz * vec3(d, d, e) + pos.yyx * vec3(f, e, f) + vec3(g, h, i);
+            norm = -normalize(grad);
+
+            return true;
+        }
+    }
+}
+
+float dot2( in vec3 v ) { return dot(v,v); }
+
+float segShadow( in vec3 ro, in vec3 rd, in vec3 pa, float sh )
+{
+    float dm = dot(rd.yz,rd.yz);
+    float k1 = (ro.x-pa.x)*dm;
+    float k2 = (ro.x+pa.x)*dm;
+    vec2  k5 = (ro.yz+pa.yz)*dm;
+    float k3 = dot(ro.yz+pa.yz,rd.yz);
+    vec2  k4 = (pa.yz+pa.yz)*rd.yz;
+    vec2  k6 = (pa.yz+pa.yz)*dm;
+
+    for( int i=0; i<4 + ANGLE_loops; i++ )
+    {
+        vec2  s = vec2(i&1,i>>1);
+        float t = dot(s,k4) - k3;
+
+        if( t>0.0 )
+        sh = min(sh,dot2(vec3(clamp(-rd.x*t,k1,k2),k5-k6*s)+rd*t)/(t*t));
+    }
+    return sh;
+}
+
+float boxSoftShadow( in vec3 ro, in vec3 rd, in vec3 rad, in float sk )
+{
+    rd += 0.0001 * (1.0 - abs(sign(rd)));
+    vec3 rdd = rd;
+    vec3 roo = ro;
+
+    vec3 m = 1.0/rdd;
+    vec3 n = m*roo;
+    vec3 k = abs(m)*rad;
+
+    vec3 t1 = -n - k;
+    vec3 t2 = -n + k;
+
+    float tN = max( max( t1.x, t1.y ), t1.z );
+    float tF = min( min( t2.x, t2.y ), t2.z );
+
+    if( tN<tF && tF>0.0) return 0.0;
+
+    float sh = 1.0;
+    sh = segShadow( roo.xyz, rdd.xyz, rad.xyz, sh );
+    sh = segShadow( roo.yzx, rdd.yzx, rad.yzx, sh );
+    sh = segShadow( roo.zxy, rdd.zxy, rad.zxy, sh );
+    sh = clamp(sk*sqrt(sh),0.0,1.0);
+    return sh*sh*(3.0-2.0*sh);
+}
+
+float box(in vec3 ro, in vec3 rd, in vec3 r, out vec3 nn, bool entering)
+{
+    rd += 0.0001 * (1.0 - abs(sign(rd)));
+    vec3 dr = 1.0 / rd;
+    vec3 n = ro * dr;
+    vec3 k = r * abs(dr);
+
+    vec3 pin = -k - n;
+    vec3 pout = k - n;
+    float tin = max(pin.x, max(pin.y, pin.z));
+    float tout = min(pout.x, min(pout.y, pout.z));
+    if (tin > tout)
+        return -1.;
+    if (entering)
+    {
+        nn = -sign(rd) * step(pin.zxy, pin.xyz) * step(pin.yzx, pin.xyz);
+    }
+    else
+    {
+        nn = sign(rd) * step(pout.xyz, pout.zxy) * step(pout.xyz, pout.yzx);
+    }
+    return entering ? tin : tout;
+}
+
+vec3 bgcol(in vec3 rd)
+{
+    return mix(vec3(0.01), vec3(0.336, 0.458, .668), 1. - pow(abs(rd.z+0.25), 1.3));
+}
+
+vec3 background(in vec3 ro, in vec3 rd , vec3 l_dir, out float alpha)
+{
+#ifdef ONLY_BOX
+alpha=0.;
+return vec3(0.01);
+#endif
+    float t = (-BOXDIMS.z - ro.z) / rd.z;
+    alpha=0.;
+    vec3 bgc = bgcol(rd);
+    if (t < 0.)
+        return bgc;
+    vec2 uv = ro.xy + t * rd.xy;
+#ifdef NO_SHADOW
+    float shad=1.;
+#else
+    float shad = boxSoftShadow((ro + t * rd), normalize(l_dir+vec3(0.,0.,1.))*rotz(PI*0.65) , BOXDIMS, 1.5);
+#endif
+    float aofac = smoothstep(-0.95, .75, length(abs(uv) - min(abs(uv), vec2(0.45))));
+    aofac = min(aofac,smoothstep(-0.65, 1., shad));
+    float lght=max(dot(normalize(ro + t * rd+vec3(0.,-0.,-5.)), normalize(l_dir-vec3(0.,0.,1.))*rotz(PI*0.65)), 0.0);
+    vec3 col = mix(vec3(0.4), vec3(.71,.772,0.895), lght*lght* aofac+ 0.05) * aofac;
+    alpha=1.-smoothstep(7.,10.,length(uv));
+#ifdef SHADOW_ALPHA
+    //alpha=clamp(alpha*max(lght*lght*0.95,(1.-aofac)*1.25),0.,1.);
+    alpha=clamp(alpha*(1.-aofac)*1.25,0.,1.);
+#endif
+    return mix(col*length(col)*0.8,bgc,smoothstep(7.,10.,length(uv)));
+}
+
+#define swap(a,b) tv=a;a=b;b=tv
+
+vec4 insides(vec3 ro, vec3 rd, vec3 nor_c, vec3 l_dir, out float tout)
+{
+    tout = -1.;
+    vec3 trd=rd;
+
+    vec3 col = vec3(0.);
+
+    float pi = 3.1415926;
+
+    if (abs(nor_c.x) > 0.5)
+    {
+        rd = rd.xzy * nor_c.x;
+        ro = ro.xzy * nor_c.x;
+    }
+    else if (abs(nor_c.z) > 0.5)
+    {
+        l_dir *= roty(pi);
+        rd = rd.yxz * nor_c.z;
+        ro = ro.yxz * nor_c.z;
+    }
+    else if (abs(nor_c.y) > 0.5)
+    {
+        l_dir *= rotz(-pi * 0.5);
+        rd = rd * nor_c.y;
+        ro = ro * nor_c.y;
+    }
+
+#ifdef ANIM_SHAPE
+    float curvature = (0.001+1.5-1.5*smoothstep(0.,8.5,mod((iTime+tshift)*0.44,20.))*(1.-smoothstep(10.,18.5,mod((iTime+tshift)*0.44,20.))));
+    // curvature(to not const above) make compilation on Angle 15+ sec
+#else
+#ifdef STATIC_SHAPE
+    const float curvature = STATIC_SHAPE;
+#else
+    const float curvature = .5;
+#endif
+#endif
+    float bil_size = 1.;
+    vec4 ps = vec4(-bil_size, -bil_size, bil_size, bil_size) * curvature;
+    vec4 ph = vec4(-bil_size, bil_size, bil_size, -bil_size) * curvature;
+
+    vec4 [3]colx=vec4[3](vec4(0.),vec4(0.),vec4(0.));
+    vec3 [3]dx=vec3[3](vec3(-1.),vec3(-1.),vec3(-1.));
+    vec4 [3]colxsi=vec4[3](vec4(0.),vec4(0.),vec4(0.));
+    int [3]order=int[3](0,1,2);
+
+    for (int i = 0; i < 3 + ANGLE_loops; i++)
+    {
+        if (abs(nor_c.x) > 0.5)
+        {
+            ro *= rotz(-pi * (1. / float(3)));
+            rd *= rotz(-pi * (1. / float(3)));
+        }
+        else if (abs(nor_c.z) > 0.5)
+        {
+            ro *= rotz(pi * (1. / float(3)));
+            rd *= rotz(pi * (1. / float(3)));
+        }
+        else if (abs(nor_c.y) > 0.5)
+        {
+            ro *= rotx(pi * (1. / float(3)));
+            rd *= rotx(pi * (1. / float(3)));
+        }
+        vec3 normnew;
+        float tnew;
+        bool si;
+        float tsi;
+        vec3 normsi;
+        float fade;
+        float fadesi;
+
+        if (iBilinearPatch(ro, rd, ps, ph, bil_size, tnew, normnew, si, tsi, normsi, fade, fadesi))
+        {
+            if (tnew > 0.)
+            {
+                vec4 tcol, tcolsi;
+                calcColor(ro, rd, normnew, tnew, bil_size, i, si, tsi, tcol, tcolsi);
+                if (tcol.a > 0.0)
+                {
+                    {
+                        vec3 tvalx = vec3(tnew, float(si), tsi);
+                        dx[i]=tvalx;
+                    }
+#ifdef DEBUG
+                    colx[i]=tcol;
+                    if (si)colxsi[i]=tcolsi;
+#else
+
+                    float dif = clamp(dot(normnew, l_dir), 0.0, 1.0);
+                    float amb = clamp(0.5 + 0.5 * dot(normnew, l_dir), 0.0, 1.0);
+
+                    {
+#ifdef USE_COLOR
+                        vec3 shad = 0.57 * color_blue * amb + 1.5*color_blue.bgr * dif;
+                        const vec3 tcr = color_red;
+#else
+                        vec3 shad = vec3(0.32, 0.43, 0.54) * amb + vec3(1.0, 0.9, 0.7) * dif;
+                        const vec3 tcr = vec3(1.,0.21,0.11);
+#endif
+                        float ta = clamp(length(tcol.rgb),0.,1.);
+                        tcol=clamp(tcol*tcol*2.,0.,1.);
+                        vec4 tvalx =
+                            vec4((tcol.rgb*shad*1.4 + 3.*(tcr*tcol.rgb)*clamp(1.-(amb+dif),0.,1.)), min(tcol.a,ta));
+                        tvalx.rgb=clamp(2.*tvalx.rgb*tvalx.rgb,0.,1.);
+                        tvalx*=(min(fade*5.,1.));
+                        colx[i]=tvalx;
+                    }
+                    if (si)
+                    {
+                        dif = clamp(dot(normsi, l_dir), 0.0, 1.0);
+                        amb = clamp(0.5 + 0.5 * dot(normsi, l_dir), 0.0, 1.0);
+                        {
+#ifdef USE_COLOR
+                            vec3 shad = 0.57 * color_blue * amb + 1.5*color_blue.bgr * dif;
+                            const vec3 tcr = color_red;
+#else
+                            vec3 shad = vec3(0.32, 0.43, 0.54) * amb + vec3(1.0, 0.9, 0.7) * dif;
+                            const vec3 tcr = vec3(1.,0.21,0.11);
+#endif
+                            float ta = clamp(length(tcolsi.rgb),0.,1.);
+                            tcolsi=clamp(tcolsi*tcolsi*2.,0.,1.);
+                            vec4 tvalx =
+                                vec4(tcolsi.rgb * shad + 3.*(tcr*tcolsi.rgb)*clamp(1.-(amb+dif),0.,1.), min(tcolsi.a,ta));
+                            tvalx.rgb=clamp(2.*tvalx.rgb*tvalx.rgb,0.,1.);
+                            tvalx.rgb*=(min(fadesi*5.,1.));
+                            colxsi[i]=tvalx;
+                        }
+                    }
+#endif
+                }
+            }
+        }
+    }
+    // transparency logic and layers sorting
+    float a = 1.;
+    if (dx[0].x < dx[1].x){{vec3 swap(dx[0], dx[1]);}{int swap(order[0], order[1]);}}
+    if (dx[1].x < dx[2].x){{vec3 swap(dx[1], dx[2]);}{int swap(order[1], order[2]);}}
+    if (dx[0].x < dx[1].x){{vec3 swap(dx[0], dx[1]);}{int swap(order[0], order[1]);}}
+
+    tout = max(max(dx[0].x, dx[1].x), dx[2].x);
+
+    if (dx[0].y < 0.5)
+    {
+        a=colx[order[0]].a;
+    }
+
+#if !(defined(DEBUG)&&defined(BUG))
+
+    // self intersection
+    bool [3] rul= bool[3](
+        ((dx[0].y > 0.5) && (dx[1].x <= 0.)),
+        ((dx[1].y > 0.5) && (dx[0].x > dx[1].z)),
+        ((dx[2].y > 0.5) && (dx[1].x > dx[2].z))
+    );
+    for(int k=0;k<3;k++){
+        if(rul[k]){
+            vec4 tcolxsi = vec4(0.);
+            tcolxsi=colxsi[order[k]];
+            vec4 tcolx = vec4(0.);
+            tcolx=colx[order[k]];
+
+            vec4 tvalx = mix(tcolxsi, tcolx, tcolx.a);
+            colx[order[k]]=tvalx;
+
+            vec4 tvalx2 = mix(vec4(0.), tvalx, max(tcolx.a, tcolxsi.a));
+            colx[order[k]]=tvalx2;
+        }
+    }
+
+#endif
+
+    float a1 = (dx[1].y < 0.5) ? colx[order[1]].a : ((dx[1].z > dx[0].x) ? colx[order[1]].a : 1.);
+    float a2 = (dx[2].y < 0.5) ? colx[order[2]].a : ((dx[2].z > dx[1].x) ? colx[order[2]].a : 1.);
+    col = mix(mix(colx[order[0]].rgb, colx[order[1]].rgb, a1), colx[order[2]].rgb, a2);
+    a = max(max(a, a1), a2);
+    return vec4(col, a);
+}
+
+void mainImage(out vec4 fragColor, in vec2 fragCoord)
+{
+    float osc = 0.5;
+    vec3 l_dir = normalize(vec3(0., 1., 0.));
+    l_dir *= rotz(0.5);
+    float mouseY = 1.0 * 0.5 * PI;
+    // Note: MOUSE_control disabled (no iMouse support)
+    // #ifdef MOUSE_control
+    // mouseY = (1.0 - 1.15 * iMouse.y / iResolution.y) * 0.5 * PI;
+    // if(iMouse.y < 1.)
+    // #endif
+#ifdef CAMERA_POS
+    mouseY = PI*CAMERA_POS;
+#else
+    mouseY = PI*0.49 - smoothstep(0.,8.5,mod((iTime+tshift)*0.33,25.))*(1.-smoothstep(14.,24.0,mod((iTime+tshift)*0.33,25.))) * 0.55 * PI;
+#endif
+#ifdef ROTATION_SPEED
+    float mouseX = -2.*PI-0.25*(iTime*ROTATION_SPEED+tshift);
+#else
+    float mouseX = -2.*PI-0.25*(iTime+tshift);
+#endif
+    // Note: MOUSE_control disabled (no iMouse support)
+    // #ifdef MOUSE_control
+    // mouseX+=-(iMouse.x / iResolution.x) * 2. * PI;
+    // #endif
+
+#ifdef CAMERA_FAR
+    vec3 eye = (2. + CAMERA_FAR) * vec3(cos(mouseX) * cos(mouseY), sin(mouseX) * cos(mouseY), sin(mouseY));
+#else
+    vec3 eye = 4. * vec3(cos(mouseX) * cos(mouseY), sin(mouseX) * cos(mouseY), sin(mouseY));
+#endif
+    vec3 w = normalize(-eye);
+    vec3 up = vec3(0., 0., 1.);
+    vec3 u = normalize(cross(w, up));
+    vec3 v = cross(u, w);
+
+    vec4 tot=vec4(0.);
+#if defined(AA_CUBE)||defined(AA_ALL)
+#ifdef AA_CUBE
+    const int AA = AA_CUBE;
+#else
+    const int AA = AA_ALL;
+#endif
+    vec3 incol_once=vec3(0.);
+    bool in_once=false;
+    vec4 incolbg_once=vec4(0.);
+    bool bg_in_once=false;
+    vec4 outcolbg_once=vec4(0.);
+    bool bg_out_once=false;
+    for( int mx=0; mx<AA; mx++ )
+    for( int nx=0; nx<AA; nx++ )
+    {
+    vec2 o = vec2(mod(float(mx+AA/2),float(AA)),mod(float(nx+AA/2),float(AA))) / float(AA) - 0.5;
+    vec2 uv = (fragCoord + o - 0.5 * iResolution.xy) / iResolution.x;
+#else
+    vec2 uv = (fragCoord - 0.5 * iResolution.xy) / iResolution.x;
+#endif
+    vec3 rd = normalize(w * FDIST + uv.x * u + uv.y * v);
+
+    vec3 ni;
+    float t = box(eye, rd, BOXDIMS, ni, true);
+    vec3 ro = eye + t * rd;
+    vec2 coords = ro.xy * ni.z/BOXDIMS.xy + ro.yz * ni.x/BOXDIMS.yz + ro.zx * ni.y/BOXDIMS.zx;
+    float fadeborders = (1.-smoothstep(0.915,1.05,abs(coords.x)))*(1.-smoothstep(0.915,1.05,abs(coords.y)));
+
+    if (t > 0.)
+    {
+        float ang = -iTime * 0.33;
+        vec3 col = vec3(0.);
+#ifdef AA_CUBE
+        if(in_once)col=incol_once;
+        else{
+        in_once=true;
+#endif
+        float R0 = (IOR - 1.) / (IOR + 1.);
+        R0 *= R0;
+
+        vec2 theta = vec2(0.);
+        vec3 n = vec3(cos(theta.x) * sin(theta.y), sin(theta.x) * sin(theta.y), cos(theta.y));
+
+        vec3 nr = n.zxy * ni.x + n.yzx * ni.y + n.xyz * ni.z;
+        vec3 rdr = reflect(rd, nr);
+        float talpha;
+        vec3 reflcol = background(ro, rdr, l_dir,talpha);
+
+        vec3 rd2 = refract(rd, nr, 1. / IOR);
+
+        float accum = 1.;
+        vec3 no2 = ni;
+        vec3 ro_refr = ro;
+
+        vec4 [2] colo = vec4[2](vec4(0.),vec4(0.));
+
+        for (int j = 0; j < 2 + ANGLE_loops; j++)
+        {
+            float tb;
+            vec2 coords2 = ro_refr.xy * no2.z + ro_refr.yz * no2.x + ro_refr.zx * no2.y;
+            vec3 eye2 = vec3(coords2, -1.);
+            vec3 rd2trans = rd2.yzx * no2.x + rd2.zxy * no2.y + rd2.xyz * no2.z;
+
+            rd2trans.z = -rd2trans.z;
+            vec4 internalcol = insides(eye2, rd2trans, no2, l_dir, tb);
+            if (tb > 0.)
+            {
+                internalcol.rgb *= accum;
+                colo[j]=internalcol;
+            }
+
+            if ((tb <= 0.) || (internalcol.a < 1.))
+            {
+                float tout = box(ro_refr, rd2, BOXDIMS, no2, false);
+                no2 = n.zyx * no2.x + n.xzy * no2.y + n.yxz * no2.z;
+                vec3 rout = ro_refr + tout * rd2;
+                vec3 rdout = refract(rd2, -no2, IOR);
+                float fresnel2 = R0 + (1. - R0) * pow(1. - dot(rdout, no2), 1.3);
+                rd2 = reflect(rd2, -no2);
+
+#ifdef backside_refl
+                if((dot(rdout, no2))>0.5){fresnel2=1.;}
+#endif
+                ro_refr = rout;
+                ro_refr.z = max(ro_refr.z, -0.999);
+
+                accum *= fresnel2;
+            }
+        }
+        float fresnel = R0 + (1. - R0) * pow(1. - dot(-rd, nr), 5.);
+        col = mix(mix(colo[1].rgb * colo[1].a, colo[0].rgb, colo[0].a)*fadeborders, reflcol, pow(fresnel, 1.5));
+        col=clamp(col,0.,1.);
+#ifdef AA_CUBE
+        }
+        incol_once=col;
+        if(!bg_in_once){
+        bg_in_once=true;
+        float alpha;
+        incolbg_once = vec4(background(eye, rd, l_dir, alpha), 0.15);
+#if defined(BG_ALPHA)||defined(ONLY_BOX)||defined(SHADOW_ALPHA)
+        incolbg_once.w = alpha;
+#endif
+        }
+#endif
+
+        float cineshader_alpha = 0.;
+        cineshader_alpha = clamp(0.15*dot(eye,ro),0.,1.);
+        vec4 tcolx = vec4(col, cineshader_alpha);
+#if defined(BG_ALPHA)||defined(ONLY_BOX)||defined(SHADOW_ALPHA)
+        tcolx.w = 1.;
+#endif
+        tot += tcolx;
+    }
+    else
+    {
+        vec4 tcolx = vec4(0.);
+#ifdef AA_CUBE
+        if(!bg_out_once){
+        bg_out_once=true;
+#endif
+        float alpha;
+        tcolx = vec4(background(eye, rd, l_dir, alpha), 0.15);
+#if defined(BG_ALPHA)||defined(ONLY_BOX)||defined(SHADOW_ALPHA)
+        tcolx.w = alpha;
+#endif
+#ifdef AA_CUBE
+        outcolbg_once=tcolx;
+        }else tcolx=max(outcolbg_once,incolbg_once);
+#endif
+        tot += tcolx;
+    }
+#if defined(AA_CUBE)||defined(AA_ALL)
+    }
+    tot /= float(AA*AA);
+#endif
+    fragColor = tot;
+#ifdef NO_ALPHA
+    fragColor.w = 1.;
+#endif
+    fragColor.rgb=clamp(fragColor.rgb,0.,1.);
+    // Note: iChannel0 line removed (texture channel not supported)
+    // Original line was for optional background blending when BG_ALPHA/ONLY_BOX/SHADOW_ALPHA defined
+    // #if defined(BG_ALPHA)||defined(ONLY_BOX)||defined(SHADOW_ALPHA)
+    // fragColor.rgb=fragColor.rgb*fragColor.w+texture(iChannel0, fragCoord/iResolution.xy).rgb*(1.-fragColor.w);
+    // #endif
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

data/shaders/cube_lines/meta.txt

@@ -0,0 +1,2 @@
+Name: Cube lines
+Author: Danil

data/shaders/dbz/dbz.frag.msl

@@ -0,0 +1,127 @@
+#include <metal_stdlib>
+using namespace metal;
+
+// New Leaked 3I/Atlas NASA Footage — msm01
+// MSL port of dbz.vk.glsl.
+
+struct ShadertoyUBO {
+    float iTime;
+    float2 iResolution;
+};
+
+struct PassthroughVOut {
+    float4 pos [[position]];
+    float2 uv;
+};
+
+constant float NBCaps = 3.0;
+
+static float2x2 r2d(float a) {
+    float c = cos(a), s = sin(a);
+    return float2x2(c, s, -s, c);
+}
+
+static float hash12(float2 p) {
+    float3 p3 = fract(float3(p.xyx) * 0.1031);
+    p3 += dot(p3, p3.yzx + 33.33);
+    return fract((p3.x + p3.y) * p3.z);
+}
+
+static float fbm(float2 v_p) {
+    float pvpx = v_p.x;
+    float2 V1 = float2(floor(pvpx));
+    float2 V2 = float2(floor(pvpx + 1.0));
+    return mix(hash12(V1), hash12(V2), smoothstep(0.0, 1.0, fract(pvpx)));
+}
+
+#define S(a,b,c) smoothstep(a,b,c)
+
+fragment float4 dbz_fs(PassthroughVOut in [[stage_in]],
+                       constant ShadertoyUBO& U [[buffer(0)]]) {
+    float2 fragCoord = in.uv * U.iResolution;
+    float2 p = float2((1.0 / U.iResolution.y) * (fragCoord.x - U.iResolution.x / 2.0),
+                      fragCoord.y / U.iResolution.y - 0.5);
+
+    p.x = -p.x;
+    p *= 150.0;
+
+    float4 col = float4(0.05, 0.05, 0.15, 1.0);
+
+    float2 save1 = p;
+
+    p = p * r2d(-0.05);
+    col = mix(col, float4(0.2, 0.3, 0.5, 1.0),
+              smoothstep(75.0, 0.0, abs(p.y - 5.0 * fbm(float2(0.01 * (p.x - 33.333 * U.iTime))) + 3.5)));
+    p = p * r2d(0.05);
+
+    p = p * r2d(-0.05);
+    p *= 0.35;
+    p += float2(-5.0 * U.iTime, 0.0);
+
+    float2 b = fract(5.0 * p);
+    p = floor(5.0 * p);
+    if (fbm(float2(p.x * p.y)) > 0.996) {
+        col += clamp(1.0 - pow(3.0 * length(b + float2(-0.5)), 0.5), 0.0, 1.0);
+    }
+
+    p = save1;
+    float2 save3;
+
+    float Nb_Capsules = clamp(NBCaps, 0.0, 4.0);
+    for (float i = 0.0; i < Nb_Capsules; i++) {
+        p = save1;
+        p = p * r2d(-0.05);
+        p *= 2.5;
+        p *= 1.0 - 0.25 * i;
+        p += float2(150.0 * fbm(float2(0.15 * U.iTime + i * 54.321)) - 75.0,
+                    50.0 * sin(0.25 * U.iTime + i * 54.321) - 25.0);
+        save3 = p;
+
+        p *= 0.04;
+        p.y = abs(p.y);
+
+        if (p.x > 0.0) {
+            col += float4(0.0, 1.0, 0.5, 1.0) * smoothstep(0.2, 0.0, abs(p.y - 0.05 * fbm(float2(1.5 * p.x - 40.0 * U.iTime))) - 0.05) * smoothstep(29.0, 0.0, abs(p.x));
+            col += float4(1.0, 1.0, 1.0, 1.0) * smoothstep(0.1, 0.0, abs(p.y - 0.05 * fbm(float2(1.5 * p.x - 40.0 * U.iTime))) - 0.05) * smoothstep(29.0, 0.0, abs(p.x));
+        }
+
+        p = save3;
+        p.y = abs(p.y);
+        p += float2(-10.0, 0.0);
+        p *= float2(0.75, 1.0);
+
+        col += 0.8 * float4(0.0, 1.0, 0.5, 1.0) * S(20.0, 0.0, length(p) - 25.0 + 7.0 * sin(0.30 * length(p) * atan2(p.y, p.x) + 55.0 * U.iTime));
+        col += 0.8 * float4(1.0, 1.0, 1.0, 1.0) * S(20.0, 0.0, length(p) - 20.0 + 7.0 * sin(0.30 * length(p) * atan2(p.y, p.x) + 55.0 * U.iTime));
+
+        p = save3;
+
+        col = mix(col, float4(1.0), 0.5 * S(10.0, 0.0, length(p + float2(5.0, 0.0)) - 20.0) * abs(sin(50.0 * U.iTime)));
+        col = mix(col, float4(1.0), 0.5 * S(20.0, 0.0, length(p + float2(5.0, 0.0)) - 20.0));
+        col = mix(col, float4(1.0), S(0.01, 0.0, length(p) - 20.0));
+
+        if (length(p) - 20.0 < 0.0) {
+            col = mix(col, float4(0.65, 0.68, 0.68 + 0.1 * (3.0 - i), 1.0), S(0.5, 0.0, length(p - float2(2.0, 0.0)) - 17.0));
+            if (S(0.0, 1.0, length(float2(3.0, 2.0) * p + float2(33.5, 0.0)) - 23.0) > 0.0) {
+                col = mix(col, float4(0.45, 0.55, 0.55 + 0.1 * (3.0 - i), 1.0),
+                          0.75 * S(0.5, 0.0, length(p - float2(2.0, 0.0) + 0.5 * fbm(float2(4.5 * atan2(p.y, p.x)))) - 9.0));
+            }
+
+            col = mix(col, float4(0.0), S(0.2, 0.0, abs(length(p) - 19.9) - 0.20));
+            col = mix(col, float4(0.5, 0.2, 0.3, 1.0)
+                          - S(5.0, 0.0, length(p + float2(-6.0, 15.0)) - 20.0)
+                          - S(0.25, 0.0, abs(length(p + float2(0.0, 3.0)) - 15.0) - 0.4)
+                          - S(0.0, 1.5, p.y - 8.5)
+                          + 0.25 * float4(1.0, 0.5, 0.0, 1.0) * S(10.0, 0.0, abs(p.y))
+                          ,
+                          S(0.5, 0.0, length(float2(3.0, 2.0) * p + float2(35.0, 0.0)) - 19.9));
+
+            col = mix(col, float4(0.0, 0.0, 0.0, 1.0), S(1.0, 0.0, abs(length(float2(3.0, 2.0) * p + float2(35.0, 0.0)) - 19.9) - 0.1));
+            col = mix(col, float4(0.0, 0.0, 0.0, 1.0), S(1.0, 0.0, abs(length(float2(3.0, 2.0) * p + float2(33.5, 0.0)) - 23.0) - 0.1));
+
+            if (p.y > 0.0) col = mix(col, float4(0.0, 0.0, 0.0, 1.0), S(1.0, 0.0, abs(length(float2(3.0, 2.0) * p + float2(29.0, 0.0)) - 30.0) - 0.1));
+            if (p.y < 0.0) col = mix(col, float4(0.0, 0.0, 0.0, 1.0), S(1.0, 0.0, abs(length(float2(3.0, 2.0) * p + float2(-31.0, 0.0)) - 30.0) - 0.1));
+        }
+    }
+
+    return clamp(col, 0.0, 1.0);
+}

data/shaders/dbz/dbz.gl.glsl

@@ -0,0 +1,190 @@
+// Name: New Leaked 3I/Atlas NASA Footage
+// Author: msm01
+// URL: https://www.shadertoy.com/view/3ftcRr
+#version 330 core
+precision highp float;
+
+out vec4 FragColor;
+in vec2 vUV;
+uniform vec2 iResolution;
+uniform float iTime;
+
+// A small improv/fanart from yesterday.
+
+#define s(a,b,c) smoothstep(a,b,c)
+#define PI     3.14159
+#define NBCaps 3.
+
+mat2 r2d( float a ){ float c = cos(a), s = sin(a); return mat2( c, s, -s, c ); }
+float metaDiamond(vec2 p, vec2 pixel, float r){ vec2 d = abs(p-pixel); return r / (d.x + d.y); }
+
+// Dave Hoskins's hash ! Noone can hash hashes like he hashes !
+float hash12(vec2 p)
+{
+      vec3 p3  = fract(vec3(p.xyx) * .1031);
+      p3 += dot(p3, p3.yzx + 33.33);
+      return fract((p3.x + p3.y) * p3.z);
+}
+// Smoothed 1D-noise. Just like Zoltraak : stupid simple. Powerful.
+float fbm(in vec2 v_p)
+{
+      float pvpx = v_p.x;
+      vec2 V1 = vec2(floor(pvpx      ));
+      vec2 V2 = vec2(floor(pvpx + 1.0));
+      return mix(hash12(V1),hash12(V2),smoothstep(0.0,1.0,fract(pvpx)));
+}
+
+void mainImage( out vec4 fragColor, in vec2 fragCoord )
+{
+     // Get Base Coordinates
+     vec2 p = vec2( (1.0/iResolution.y)*(fragCoord.x - iResolution.x/2.0),fragCoord.y / iResolution.y - 0.5);
+
+     // reversed, for accurate Martian perspective... :)
+     p.x=-p.x;
+
+     // Zoom out
+     p*=150.0;
+
+     // Init the Accumulator
+     vec4 col = vec4(0.05,0.05,0.15,1.0);
+
+     // Make up other boxes and save base in them.
+     vec2 save1 = p;
+     vec2 save2 = p;
+
+     // Faint Nebula Background
+
+     // Tilt the camera
+     p*= r2d(-0.05);
+     // Space Background Gradient
+     col = mix(col,vec4(0.2,0.3,0.5,1.0),smoothstep(75.0,0.0,abs(p.y - 5.0*fbm(vec2(0.01*(p.x - 33.333*iTime))) + 3.5)));
+     // Untilt the camera
+     p*= r2d( 0.05);
+
+     // BG Starfield
+
+     // Rotate
+     p*= r2d(-0.05);
+     // Zoom In
+     p*= 0.35;
+     // Scroll Left
+     p+= vec2(-5.0*iTime,0.0);
+
+     // Hack the coords...
+     vec2 b = fract(5.0*p);
+     p = floor(5.0*p);
+     // Draw the stars
+     if( fbm(vec2(p.x*p.y)) > 0.996)col += clamp(1.0-pow(3.0*length(b+vec2(-0.5)),0.5),0.0,1.0);
+
+     // Reload because the coords are all f.. up now !
+     p = save1;
+     // Another Box...
+     vec2 save3;
+
+     // We're going to draw max 4 capsules max.
+     // Yes we could draw more but Earth must survive, man. Have mercy !
+     float Nb_Capsules = clamp(NBCaps,0.0,4.0);
+     for( float i = 0.0;i<Nb_Capsules; i++ )
+     {
+          // Reloooooaaaaad !
+          p = save1;
+          // Tilt as much as we tilted the stars...
+          p*= r2d(-0.05);
+          // Zoom out a bit
+          p*=2.5;
+          // Then zoom in a bit closer every loop.
+          p*=1.0-0.25*i;
+          // Compute Random Coordinates For Each Capsule Position
+          // Move To That Position.
+          p += vec2(150.0*fbm(vec2(0.15*iTime + i*54.321)) - 75.0,  // X varies randomly
+                     50.0*sin(     0.25*iTime + i*54.321)  - 25.0); // Y varies periodically
+          // Save Position
+          save3 = p;
+
+          // Mega Zoom
+          p*=0.04;
+          // (Ox)-axis Symetry for jets
+          p.y = abs(p.y);
+
+          if( p.x > 0.0 )
+          {
+              // Green Jet
+              col += vec4(0.0,1.0,0.5,1.0)*smoothstep(0.2,0.0,abs(p.y - 0.05*fbm(vec2(1.5*p.x - 40.0*iTime)))-0.05)*smoothstep(29.0,0.0,abs(p.x));
+              // White Jet
+              col += vec4(1.0,1.0,1.0,1.0)*smoothstep(0.1,0.0,abs(p.y - 0.05*fbm(vec2(1.5*p.x - 40.0*iTime)))-0.05)*smoothstep(29.0,0.0,abs(p.x));
+          };
+
+          // Reload !
+          p = save3;
+          // (Ox)-axis Symetry for the flames
+          p.y = abs(p.y);
+          // Fine-tuning Flames position
+          p+= vec2(-10.0,0.0);
+          // Fine-tuning Flames Shape
+          p *= vec2(0.75,1.0);
+
+         // Green Flames
+         col += 0.8*vec4(0.0,1.0,0.5,1.0)*s(20.0,0.0,length(p)-25.0+7.0*sin(0.30*length(p)*atan(p.y,p.x) + 55.0*iTime));
+         // White flames
+         col += 0.8*vec4(1.0,1.0,1.0,1.0)*s(20.0,0.0,length(p)-20.0+7.0*sin(0.30*length(p)*atan(p.y,p.x) + 55.0*iTime));
+
+         p = save3;
+
+         // Fat Aura
+         col = mix(col,vec4(1.0),0.5*s(10.0,0.0,length(p + vec2(5.0,0.0))-20.0)*abs(sin(50.0*iTime)));
+         // Less-Fat Aura
+         col = mix(col,vec4(1.0),0.5*s(20.0,0.0,length(p + vec2(5.0,0.0))-20.0));
+         // Frieren : "Aura ? Shader yourself !"
+
+         // The Pod
+
+         // White Disk
+         col = mix(col,vec4(1.0),s(0.01,0.0,length(p)-20.0));
+
+         if( length(p) - 20.0 < 0.0 ) // Basic Masking
+         {
+             // 2D Shading : bluish large shadow
+             col = mix(col,vec4(0.65,0.68,0.68 + 0.1*(3.0-i),1.0),s(0.5,0.0,length(p - vec2(2.0,0.0))-17.0));
+             // 2D Shading : dark small shadow
+             // If Outside Porthole Zone
+             if(s(0.0,1.0,length(vec2(3.0,2.0)*p + vec2(33.5,0.0))-23.0)>0.0)
+             col = mix(col,vec4(0.45,0.55,0.55 + 0.1*(3.0-i),1.0),0.75*s(0.5,0.0,length(p - vec2(2.0,0.0)+ 0.5*fbm(vec2(4.5*atan(p.y,p.x))))-9.0));
+
+             // Small 2D Indentation Details On The Spheres Using A Procedural Texture
+             // NOTE: Original used texture(iChannel0, ...) which is not supported
+             // Texture detail removed - not essential for the effect
+             // vec4 colorCapsule = vec4(hash12(0.0003*p*dot(p,p) + 0.789*i));
+             // if(colorCapsule.x>0.75)if(s(0.0,1.0,length(vec2(3.0,2.0)*p + vec2(33.5,0.0))-23.0)>0.0)col *= vec4(0.25,0.25,0.25,1.0);
+
+             // Bigger Dark Line All Around The Pod
+             col = mix(col,vec4(0.0),s(0.2,0.0,abs(length(p)-19.9)-0.20));
+             // Draw The Porthole :
+             col = mix(col,vec4(0.5,0.2,0.3,1.0)                               // Base Color
+                           -s(5.0,0.0,length(p + vec2(-6.0,15.0))-20.0)        // Main Shadow
+                           -s(0.25,0.0,abs(length(p + vec2(0.0,3.0))-15.0)-0.4)// Vertical Shadow
+                           -s(0.0,1.5,p.y-8.5)                                 // top Shadow
+                           +0.25*vec4(1.0,0.5,0.0,1.0)*s(10.0,0.0,abs(p.y))    // Fake Glass Gradient
+                           ,
+                           s(0.5,0.0,length(vec2(3.0,2.0)*p + vec2(35.0,0.0))-19.9));
+
+             // Porthole Black Rings
+             // Internal
+             col = mix(col,vec4(0.0,0.0,0.0,1.0),s(1.0,0.0,abs(length(vec2(3.0,2.0)*p + vec2(35.0,0.0))-19.9)-0.1));
+             // External
+             col = mix(col,vec4(0.0,0.0,0.0,1.0),s(1.0,0.0,abs(length(vec2(3.0,2.0)*p + vec2(33.5,0.0))-23.0)-0.1));
+
+             // Pod Tennis-Ball Door Line...
+             if(p.y>0.0)col = mix(col,vec4(0.0,0.0,0.0,1.0),s(1.0,0.0,abs(length(vec2(3.0,2.0)*p + vec2( 29.0,0.0))-30.0)-0.1));
+             if(p.y<0.0)col = mix(col,vec4(0.0,0.0,0.0,1.0),s(1.0,0.0,abs(length(vec2(3.0,2.0)*p + vec2(-31.0,0.0))-30.0)-0.1));
+         };
+     };
+   // WAKE UP SHEEPLE !
+   fragColor = clamp(col,0.0,1.0);
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

data/shaders/dbz/dbz.vk.glsl

@@ -0,0 +1,127 @@
+#version 450
+
+// Name: New Leaked 3I/Atlas NASA Footage
+// Author: msm01
+// URL: https://www.shadertoy.com/view/3ftcRr
+
+layout(location = 0) in  vec2 vUV;
+layout(location = 0) out vec4 FragColor;
+
+layout(set = 3, binding = 0) uniform ShadertoyUBO {
+    float iTime;
+    vec2  iResolution;
+};
+
+#define s(a,b,c) smoothstep(a,b,c)
+#define PI     3.14159
+#define NBCaps 3.
+
+mat2 r2d(float a) { float c = cos(a), s = sin(a); return mat2(c, s, -s, c); }
+float metaDiamond(vec2 p, vec2 pixel, float r) { vec2 d = abs(p - pixel); return r / (d.x + d.y); }
+
+float hash12(vec2 p) {
+    vec3 p3 = fract(vec3(p.xyx) * .1031);
+    p3 += dot(p3, p3.yzx + 33.33);
+    return fract((p3.x + p3.y) * p3.z);
+}
+
+float fbm(in vec2 v_p) {
+    float pvpx = v_p.x;
+    vec2 V1 = vec2(floor(pvpx));
+    vec2 V2 = vec2(floor(pvpx + 1.0));
+    return mix(hash12(V1), hash12(V2), smoothstep(0.0, 1.0, fract(pvpx)));
+}
+
+void mainImage(out vec4 fragColor, in vec2 fragCoord) {
+    vec2 p = vec2((1.0 / iResolution.y) * (fragCoord.x - iResolution.x / 2.0),
+                  fragCoord.y / iResolution.y - 0.5);
+
+    p.x = -p.x;
+    p *= 150.0;
+
+    vec4 col = vec4(0.05, 0.05, 0.15, 1.0);
+
+    vec2 save1 = p;
+    vec2 save2 = p;
+
+    p *= r2d(-0.05);
+    col = mix(col, vec4(0.2, 0.3, 0.5, 1.0),
+              smoothstep(75.0, 0.0, abs(p.y - 5.0 * fbm(vec2(0.01 * (p.x - 33.333 * iTime))) + 3.5)));
+    p *= r2d(0.05);
+
+    p *= r2d(-0.05);
+    p *= 0.35;
+    p += vec2(-5.0 * iTime, 0.0);
+
+    vec2 b = fract(5.0 * p);
+    p = floor(5.0 * p);
+    if (fbm(vec2(p.x * p.y)) > 0.996) col += clamp(1.0 - pow(3.0 * length(b + vec2(-0.5)), 0.5), 0.0, 1.0);
+
+    p = save1;
+    vec2 save3;
+
+    float Nb_Capsules = clamp(NBCaps, 0.0, 4.0);
+    for (float i = 0.0; i < Nb_Capsules; i++) {
+        p = save1;
+        p *= r2d(-0.05);
+        p *= 2.5;
+        p *= 1.0 - 0.25 * i;
+        p += vec2(150.0 * fbm(vec2(0.15 * iTime + i * 54.321)) - 75.0,
+                  50.0 * sin(0.25 * iTime + i * 54.321) - 25.0);
+        save3 = p;
+
+        p *= 0.04;
+        p.y = abs(p.y);
+
+        if (p.x > 0.0) {
+            col += vec4(0.0, 1.0, 0.5, 1.0) * smoothstep(0.2, 0.0, abs(p.y - 0.05 * fbm(vec2(1.5 * p.x - 40.0 * iTime))) - 0.05) * smoothstep(29.0, 0.0, abs(p.x));
+            col += vec4(1.0, 1.0, 1.0, 1.0) * smoothstep(0.1, 0.0, abs(p.y - 0.05 * fbm(vec2(1.5 * p.x - 40.0 * iTime))) - 0.05) * smoothstep(29.0, 0.0, abs(p.x));
+        }
+
+        p = save3;
+        p.y = abs(p.y);
+        p += vec2(-10.0, 0.0);
+        p *= vec2(0.75, 1.0);
+
+        col += 0.8 * vec4(0.0, 1.0, 0.5, 1.0) * s(20.0, 0.0, length(p) - 25.0 + 7.0 * sin(0.30 * length(p) * atan(p.y, p.x) + 55.0 * iTime));
+        col += 0.8 * vec4(1.0, 1.0, 1.0, 1.0) * s(20.0, 0.0, length(p) - 20.0 + 7.0 * sin(0.30 * length(p) * atan(p.y, p.x) + 55.0 * iTime));
+
+        p = save3;
+
+        col = mix(col, vec4(1.0), 0.5 * s(10.0, 0.0, length(p + vec2(5.0, 0.0)) - 20.0) * abs(sin(50.0 * iTime)));
+        col = mix(col, vec4(1.0), 0.5 * s(20.0, 0.0, length(p + vec2(5.0, 0.0)) - 20.0));
+
+        col = mix(col, vec4(1.0), s(0.01, 0.0, length(p) - 20.0));
+
+        if (length(p) - 20.0 < 0.0) {
+            col = mix(col, vec4(0.65, 0.68, 0.68 + 0.1 * (3.0 - i), 1.0), s(0.5, 0.0, length(p - vec2(2.0, 0.0)) - 17.0));
+            if (s(0.0, 1.0, length(vec2(3.0, 2.0) * p + vec2(33.5, 0.0)) - 23.0) > 0.0)
+                col = mix(col, vec4(0.45, 0.55, 0.55 + 0.1 * (3.0 - i), 1.0),
+                          0.75 * s(0.5, 0.0, length(p - vec2(2.0, 0.0) + 0.5 * fbm(vec2(4.5 * atan(p.y, p.x)))) - 9.0));
+
+            col = mix(col, vec4(0.0), s(0.2, 0.0, abs(length(p) - 19.9) - 0.20));
+            col = mix(col, vec4(0.5, 0.2, 0.3, 1.0)
+                          - s(5.0, 0.0, length(p + vec2(-6.0, 15.0)) - 20.0)
+                          - s(0.25, 0.0, abs(length(p + vec2(0.0, 3.0)) - 15.0) - 0.4)
+                          - s(0.0, 1.5, p.y - 8.5)
+                          + 0.25 * vec4(1.0, 0.5, 0.0, 1.0) * s(10.0, 0.0, abs(p.y))
+                          ,
+                          s(0.5, 0.0, length(vec2(3.0, 2.0) * p + vec2(35.0, 0.0)) - 19.9));
+
+            col = mix(col, vec4(0.0, 0.0, 0.0, 1.0), s(1.0, 0.0, abs(length(vec2(3.0, 2.0) * p + vec2(35.0, 0.0)) - 19.9) - 0.1));
+            col = mix(col, vec4(0.0, 0.0, 0.0, 1.0), s(1.0, 0.0, abs(length(vec2(3.0, 2.0) * p + vec2(33.5, 0.0)) - 23.0) - 0.1));
+
+            if (p.y > 0.0) col = mix(col, vec4(0.0, 0.0, 0.0, 1.0), s(1.0, 0.0, abs(length(vec2(3.0, 2.0) * p + vec2(29.0, 0.0)) - 30.0) - 0.1));
+            if (p.y < 0.0) col = mix(col, vec4(0.0, 0.0, 0.0, 1.0), s(1.0, 0.0, abs(length(vec2(3.0, 2.0) * p + vec2(-31.0, 0.0)) - 30.0) - 0.1));
+        }
+    }
+
+    fragColor = clamp(col, 0.0, 1.0);
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

data/shaders/dbz/meta.txt

@@ -0,0 +1,2 @@
+Name: New Leaked 3I/Atlas NASA Footage
+Author: msm01

data/shaders/fractal_pyramid/fractal_pyramid.frag.msl

@@ -0,0 +1,67 @@
+#include <metal_stdlib>
+using namespace metal;
+
+// Fractal Pyramid — bradjamesgrant
+// MSL port of fractal_pyramid.vk.glsl.
+
+struct ShadertoyUBO {
+    float iTime;
+    float2 iResolution;
+};
+
+struct PassthroughVOut {
+    float4 pos [[position]];
+    float2 uv;
+};
+
+static float3 palette(float d) {
+    return mix(float3(0.2, 0.7, 0.9), float3(1.0, 0.0, 1.0), d);
+}
+
+static float2 rotate(float2 p, float a) {
+    float c = cos(a);
+    float s = sin(a);
+    return p * float2x2(c, s, -s, c);
+}
+
+static float mapScene(float3 p, float iTime) {
+    for (int i = 0; i < 8; ++i) {
+        float t = iTime * 0.2;
+        p.xz = rotate(p.xz, t);
+        p.xy = rotate(p.xy, t * 1.89);
+        p.xz = abs(p.xz);
+        p.xz -= 0.5;
+    }
+    return dot(sign(p), p) / 5.0;
+}
+
+static float4 rm(float3 ro, float3 rd, float iTime) {
+    float t = 0.0;
+    float3 col = float3(0.0);
+    float d = 1.0;
+    for (int i = 0; i < 64; ++i) {
+        float3 p = ro + rd * t;
+        d = mapScene(p, iTime) * 0.5;
+        if (d < 0.02) break;
+        if (d > 100.0) break;
+        col += palette(length(p) * 0.1) / (400.0 * d);
+        t += d;
+    }
+    return float4(col, 1.0 / (d * 100.0));
+}
+
+fragment float4 fractal_pyramid_fs(PassthroughVOut in [[stage_in]],
+                                    constant ShadertoyUBO& U [[buffer(0)]]) {
+    float2 fragCoord = in.uv * U.iResolution;
+    float2 uv = (fragCoord - (U.iResolution * 0.5)) / U.iResolution.x;
+    float3 ro = float3(0.0, 0.0, -50.0);
+    ro.xz = rotate(ro.xz, U.iTime);
+    float3 cf = normalize(-ro);
+    float3 cs = normalize(cross(cf, float3(0.0, 1.0, 0.0)));
+    float3 cu = normalize(cross(cf, cs));
+
+    float3 uuv = ro + cf * 3.0 + uv.x * cs + uv.y * cu;
+    float3 rd = normalize(uuv - ro);
+
+    return rm(ro, rd, U.iTime);
+}

data/shaders/fractal_pyramid/fractal_pyramid.gl.glsl

@@ -1,3 +1,6 @@
+// Name: Fractal Pyramid
+// Author:  bradjamesgrant
+// URL: https://www.shadertoy.com/view/tsXBzS
 #version 330 core
 precision highp float;
 

data/shaders/fractal_pyramid/fractal_pyramid.vk.glsl

@@ -0,0 +1,71 @@
+#version 450
+
+// Name: Fractal Pyramid
+// Author: bradjamesgrant
+// URL: https://www.shadertoy.com/view/tsXBzS
+
+layout(location = 0) in  vec2 vUV;
+layout(location = 0) out vec4 FragColor;
+
+layout(set = 3, binding = 0) uniform ShadertoyUBO {
+    float iTime;
+    vec2  iResolution;
+};
+
+vec3 palette(float d) {
+    return mix(vec3(0.2, 0.7, 0.9), vec3(1.0, 0.0, 1.0), d);
+}
+
+vec2 rotate(vec2 p, float a) {
+    float c = cos(a);
+    float s = sin(a);
+    return p * mat2(c, s, -s, c);
+}
+
+float mapScene(vec3 p) {
+    for (int i = 0; i < 8; ++i) {
+        float t = iTime * 0.2;
+        p.xz = rotate(p.xz, t);
+        p.xy = rotate(p.xy, t * 1.89);
+        p.xz = abs(p.xz);
+        p.xz -= 0.5;
+    }
+    return dot(sign(p), p) / 5.0;
+}
+
+vec4 rm(vec3 ro, vec3 rd) {
+    float t = 0.0;
+    vec3 col = vec3(0.0);
+    float d = 1.0;
+    for (int i = 0; i < 64; ++i) {
+        vec3 p = ro + rd * t;
+        d = mapScene(p) * 0.5;
+        if (d < 0.02) break;
+        if (d > 100.0) break;
+        col += palette(length(p) * 0.1) / (400.0 * d);
+        t += d;
+    }
+    return vec4(col, 1.0 / (d * 100.0));
+}
+
+void mainImage(out vec4 fragColor, in vec2 fragCoord) {
+    vec2 uv = (fragCoord - (iResolution.xy * 0.5)) / iResolution.x;
+    vec3 ro = vec3(0.0, 0.0, -50.0);
+    ro.xz = rotate(ro.xz, iTime);
+    vec3 cf = normalize(-ro);
+    vec3 cs = normalize(cross(cf, vec3(0.0, 1.0, 0.0)));
+    vec3 cu = normalize(cross(cf, cs));
+
+    vec3 uuv = ro + cf * 3.0 + uv.x * cs + uv.y * cu;
+    vec3 rd = normalize(uuv - ro);
+
+    vec4 col = rm(ro, rd);
+    fragColor = col;
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

data/shaders/fractal_pyramid/meta.txt

@@ -0,0 +1,2 @@
+Name: Fractal Pyramid
+Author: bradjamesgrant

data/shaders/just_another_cube/just_another_cube.frag.msl

@@ -0,0 +1,89 @@
+#include <metal_stdlib>
+using namespace metal;
+
+// Just Another Cube — mrange
+// MSL port of just_another_cube.vk.glsl. The Shadertoy original uses
+// mutable file-scope globals; in MSL we pass them by reference.
+
+struct ShadertoyUBO {
+    float iTime;
+    float2 iResolution;
+};
+
+struct PassthroughVOut {
+    float4 pos [[position]];
+    float2 uv;
+};
+
+constant float M = 1e-3;
+
+static float D(float3 p,
+               thread float2x2& R,
+               thread float& d,
+               thread float& G) {
+    p.xy = p.xy * R;
+    p.xz = p.xz * R;
+
+    float3 S = sin(123.0 * p);
+
+    float shell = abs(length(p) - 0.6);
+    p *= p * p * p;
+    d = pow(dot(p, p), 0.125) - 0.5
+        - pow(1.0 + S.x * S.y * S.z, 8.0) / 1e5;
+
+    G = min(G, max(shell, d));
+
+    return d;
+}
+
+fragment float4 just_another_cube_fs(PassthroughVOut in [[stage_in]],
+                                      constant ShadertoyUBO& U [[buffer(0)]]) {
+    float2 C = in.uv * U.iResolution;
+
+    float2x2 R;
+    float d = 1.0, z = 0.0, G = 9.0;
+
+    float3 r = float3(U.iResolution, U.iResolution.y);
+    float3 I = normalize(float3(C - 0.5 * r.xy, r.y));
+    float3 B = float3(1, 2, 9) * M;
+    float3 p = float3(0.0);
+    float3 O = float3(0.0);
+
+    {
+        float4 cs = cos(0.3 * U.iTime + float4(0, 11, 33, 0));
+        R = float2x2(cs.x, cs.y, cs.z, cs.w);
+    }
+
+    for (; z < 9.0 && d > M; z += D(p, R, d, G)) {
+        p = z * I;
+        p.z -= 2.0;
+    }
+
+    if (z < 9.0) {
+        for (int i = 0; i < 3; ) {
+            r = float3(0.0);
+            r[i] = M;
+            O[i++] = D(p + r, R, d, G) - D(p - r, R, d, G);
+        }
+
+        O = normalize(O);
+        z = 1.0 + dot(O, I);
+
+        r = reflect(I, O);
+
+        C = (p + r * (5.0 - p.y) / abs(r.y)).xz;
+
+        float3 colorTerm;
+        if (r.y > 0.0) {
+            d = sqrt(length(C * C)) + 1.0;
+            colorTerm = 5e2 * smoothstep(5.0, 4.0, d) * d * B;
+        } else {
+            colorTerm = exp(-2.0 * length(C)) * (B / M - 1.0);
+        }
+
+        O = z * z * colorTerm + pow(1.0 + O.y, 5.0) * B;
+    }
+
+    float4 result = sqrt(float4(O + B / G, O.x + B.x / G));
+    return result;
+}

data/shaders/just_another_cube/just_another_cube.gl.glsl

@@ -0,0 +1,159 @@
+// Name: Just Another Cube
+// Author: mrange
+// URL: https://www.shadertoy.com/view/3XdXRr
+#version 330 core
+precision highp float;
+
+out vec4 FragColor;
+in vec2 vUV;
+uniform vec2 iResolution;
+uniform float iTime;
+
+// CC0: Just another cube
+// Glowtracers are great for compact coding, but I wanted to see how much
+// I could squeeze a more normal raymarcher in terms of characters used.
+
+// Twigl: https://twigl.app?ol=true&ss=-OW-y9xgRgWubwKcn0Nd
+
+// == Globals ==
+// Single-letter variable names are used to save characters (code golfing).
+mat2 R;    // A 2D rotation matrix, calculated once per frame in mainImage and used by D.
+float d=1. // Stores the most recent distance to the scene from the ray's position.
+    , z=0. // Stores the total distance traveled along the ray (initialized to avoid undefined behavior)
+    , G=9. // "Glow" variable. Tracks the closest the ray comes to the object (for volumetric glow effect).
+    , M=1e-3
+    ;
+// == Distance Function (SDF - Signed Distance Field) ==
+// This function calculates the shortest distance from a given point 'p' to the scene geometry.
+// A positive result means the point is outside an object, negative is inside, and zero is on the surface.
+// This is the core of "raymarching", as it tells us the largest safe step we can take along a ray.
+float D(vec3 p) {
+  // Apply two rotations to the point's coordinates. This twists the space the object
+  // exists in, making the simple cube shape appear more complex and animated.
+  p.xy *= R;
+  p.xz *= R;
+
+  // Create a higher-frequency version of the coordinate for detailed surface patterns.
+  vec3 S = sin(123.*p);
+
+  // This creates a volumetric glow effect by tracking the minimum distance
+  // to either the existing glow value or a glowing shell around the object.
+  G = min(
+      G
+      // The glowing shell
+    , max(
+        abs(length(p)-.6)
+        // The main object distance calculation:
+        // 1. A superquadric (rounded cube shape) is created using an L8-norm.
+        // The expression `pow(dot(p=p*p*p*p,p),.125)` is a golfed version of
+        // `pow(pow(p.x,8)+pow(p.y,8)+pow(p.z,8), 1./8.)`.
+        // The `- .5` defines the object's size.
+      , d = pow(dot(p*=p*p*p,p),.125) - .5
+        // 2. Surface detail subtraction. This creates small surface variations
+        // using high-frequency sine waves for more appealing reflections.
+        - pow(1.+S.x*S.y*S.z,8.)/1e5
+      )
+    );
+
+  return d;
+}
+
+// == Main Render Function ==
+// This function is called for every pixel on the screen to determine its color.
+// 'o' is the final output color (rgba). 'C' is the input pixel coordinate (xy).
+void mainImage(out vec4 o, vec2 C) {
+  // Single-letter variable names are used to save characters (code golfing).
+  vec3  p  // The current point in 3D space along the ray.
+      , O  // Multi-purpose vector: color accumulator, then normal vector, then final color.
+      , r=vec3(iResolution.xy, iResolution.y)  // 'r' holds screen resolution, later re-used for the epsilon vector and reflection.
+        // 'I' is the Ray Direction vector. It's calculated once per pixel.
+        // This converts the 2D screen coordinate 'C' into a 3D direction, creating the camera perspective.
+      , I=normalize(vec3(C-.5*r.xy, r.y))
+        // Base glow color (dark bluish tint).
+      , B=vec3(1,2,9)*M
+      ;
+
+  // == Raymarching Loop ==
+  // This loop "marches" a ray from the camera out into the scene to find what it hits.
+  // It uses a golfed structure where the body of the loop updates the ray position 'p',
+  // and the "advancement" step moves the ray forward.
+  for(
+        // -- Initializer (runs once before the loop) --
+        // Calculate the rotation matrix for this frame based on time.
+        R = mat2(cos(.3*iTime+vec4(0,11,33,0)))
+    // -- Condition --
+    // Loop while total distance 'z' is less than 9 and we are not yet touching a surface (d > 1e-3).
+    ; z<9. && d > M
+    // -- Advancement --
+    // The ray advances by the safe distance 'd' returned by D(p).
+    // The result of D(p) is also assigned to the global 'd' inside the function.
+    ; z += D(p)
+    )
+    // -- Loop Body --
+    // Calculate the current position 'p' in world space.
+    // The camera starts at (0,0,-2) and points forward.
+    p = z*I
+  , p.z -= 2.
+  ;
+
+  // -- Hit Condition --
+  // If the loop finished because z exceeded the max distance, we hit nothing. Otherwise, we hit the surface.
+  if (z < 9.) {
+    // -- Calculate Surface Normal --
+    // Estimate the gradient ∇D at the hit point 'p' via central differences on the SDF D.
+    // We use ε = 1e-3 and loop over each axis (x, y, z):
+    //   • Zero r, then set r[i] = ε.
+    //   • Compute O[i] = D(p + r) – D(p – r).
+    // After the loop, O holds the unnormalized normal vector.
+    for (
+        int i=0               // axis index: 0→x, 1→y, 2→z (initialized to avoid warnings)
+      ; i < 3
+      ; O[i++] = D(p+r) - D(p-r)
+      )
+        r -= r                // clear r to vec3(0)
+      , r[i] = M              // set only the i-th component
+      ;
+
+    // -- Lighting and Shading --
+    // 'z' is re-purposed to store a fresnel factor (1 - cos(angle)) for edge brightness.
+    // `dot(O, I)` calculates how much the surface faces away from the camera.
+    // O is also normalized here to become a proper normal vector.
+    z = 1.+dot(O = normalize(O),I);
+
+    // 'r' is re-purposed to store the reflection vector.
+    r = reflect(I,O);
+
+    // Calculate a point 'C' along the reflection vector 'r' to sample a background color.
+    // For upward reflections (r.y > 0), this finds the intersection with the plane y=5.
+    C = (p+r*(5.-p.y)/abs(r.y)).xz;
+
+    // Calculate the final color 'O' of the hit point.
+    O =
+        // Multiply by the fresnel factor squared for stronger edge reflections.
+        z*z *
+        // Use a ternary operator to decide the color based on where the reflection ray goes.
+        (
+            // If the reflection vector points upward...
+            r.y>0.
+            // ...sample a procedural "sky" with a radial gradient and blue tint.
+          ? 5e2*smoothstep(5., 4., d = sqrt(length(C*C))+1.)*d*B
+            // ...otherwise, sample a "floor" with a deep blue exponential falloff.
+          : exp(-2.*length(C))*(B/M-1.)
+        )
+      // Add rim lighting (brighter on upward-facing surfaces).
+      + pow(1.+O.y,5.)*B
+      ;
+  }
+
+  // == Tonemapping & Output ==
+  // Apply final effects and map the High Dynamic Range (HDR) color to a displayable range.
+  // Add glow contribution: smaller G values (closer ray passes) create a brighter blue glow.
+  o = sqrt(O+B/G).xyzx;
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

data/shaders/just_another_cube/just_another_cube.vk.glsl

@@ -0,0 +1,88 @@
+#version 450
+
+// Name: Just Another Cube
+// Author: mrange
+// URL: https://www.shadertoy.com/view/3XdXRr
+// CC0: Just another cube
+
+layout(location = 0) in  vec2 vUV;
+layout(location = 0) out vec4 FragColor;
+
+layout(set = 3, binding = 0) uniform ShadertoyUBO {
+    float iTime;
+    vec2  iResolution;
+};
+
+mat2 R;
+float d = 1.;
+float z = 0.;
+float G = 9.;
+float M = 1e-3;
+
+float D(vec3 p) {
+    p.xy *= R;
+    p.xz *= R;
+
+    vec3 S = sin(123. * p);
+
+    G = min(
+        G,
+        max(
+            abs(length(p) - .6),
+            d = pow(dot(p *= p * p * p, p), .125) - .5
+                - pow(1. + S.x * S.y * S.z, 8.) / 1e5
+        )
+    );
+
+    return d;
+}
+
+void mainImage(out vec4 o, vec2 C) {
+    vec3 p,
+         O,
+         r = vec3(iResolution.xy, iResolution.y),
+         I = normalize(vec3(C - .5 * r.xy, r.y)),
+         B = vec3(1, 2, 9) * M;
+
+    for (
+        R = mat2(cos(.3 * iTime + vec4(0, 11, 33, 0)));
+        z < 9. && d > M;
+        z += D(p)
+    )
+        p = z * I,
+        p.z -= 2.;
+
+    O = vec3(0.0);
+
+    if (z < 9.) {
+        for (int i = 0; i < 3; ) {
+            r -= r;
+            r[i] = M;
+            O[i++] = D(p + r) - D(p - r);
+        }
+
+        z = 1. + dot(O = normalize(O), I);
+
+        r = reflect(I, O);
+
+        C = (p + r * (5. - p.y) / abs(r.y)).xz;
+
+        O =
+            z * z *
+            (
+                r.y > 0.
+                ? 5e2 * smoothstep(5., 4., d = sqrt(length(C * C)) + 1.) * d * B
+                : exp(-2. * length(C)) * (B / M - 1.)
+            )
+            + pow(1. + O.y, 5.) * B;
+    }
+
+    o = sqrt(O + B / G).xyzx;
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

data/shaders/just_another_cube/meta.txt

@@ -0,0 +1,2 @@
+Name: Just Another Cube
+Author: mrange

data/shaders/octograms/meta.txt

@@ -0,0 +1,2 @@
+Name: Octograms
+Author: whisky_shusuky

data/shaders/octograms/octograms.frag.msl

@@ -0,0 +1,95 @@
+#include <metal_stdlib>
+using namespace metal;
+
+// Octograms — whisky_shusuky
+// MSL port of octograms.vk.glsl. The Shadertoy original uses a mutable
+// file-scope `gTime`; in MSL we pass it as a parameter through the chain.
+
+struct ShadertoyUBO {
+    float iTime;
+    float2 iResolution;
+};
+
+struct PassthroughVOut {
+    float4 pos [[position]];
+    float2 uv;
+};
+
+static float2x2 rot(float a) {
+    float c = cos(a), s = sin(a);
+    return float2x2(c, s, -s, c);
+}
+
+static float sdBox(float3 p, float3 b) {
+    float3 q = abs(p) - b;
+    return length(max(q, float3(0.0))) + min(max(q.x, max(q.y, q.z)), 0.0);
+}
+
+static float boxGeom(float3 pos, float scale) {
+    pos *= scale;
+    float base = sdBox(pos, float3(0.4, 0.4, 0.1)) / 1.5;
+    pos.xy = (pos.xy * 5.0 - float2(0.0, 3.5)) * rot(0.75);
+    return -base;
+}
+
+static float box_set(float3 pos, float gTime) {
+    float3 pos_origin = pos;
+
+    pos = pos_origin;
+    pos.y += sin(gTime * 0.4) * 2.5;
+    pos.xy = pos.xy * rot(0.8);
+    float box1 = boxGeom(pos, 2.0 - abs(sin(gTime * 0.4)) * 1.5);
+
+    pos = pos_origin;
+    pos.y -= sin(gTime * 0.4) * 2.5;
+    pos.xy = pos.xy * rot(0.8);
+    float box2 = boxGeom(pos, 2.0 - abs(sin(gTime * 0.4)) * 1.5);
+
+    pos = pos_origin;
+    pos.x += sin(gTime * 0.4) * 2.5;
+    pos.xy = pos.xy * rot(0.8);
+    float box3 = boxGeom(pos, 2.0 - abs(sin(gTime * 0.4)) * 1.5);
+
+    pos = pos_origin;
+    pos.x -= sin(gTime * 0.4) * 2.5;
+    pos.xy = pos.xy * rot(0.8);
+    float box4 = boxGeom(pos, 2.0 - abs(sin(gTime * 0.4)) * 1.5);
+
+    pos = pos_origin;
+    pos.xy = pos.xy * rot(0.8);
+    float box5 = boxGeom(pos, 0.5) * 6.0;
+
+    pos = pos_origin;
+    float box6 = boxGeom(pos, 0.5) * 6.0;
+
+    return max(max(max(max(max(box1, box2), box3), box4), box5), box6);
+}
+
+fragment float4 octograms_fs(PassthroughVOut in [[stage_in]],
+                              constant ShadertoyUBO& U [[buffer(0)]]) {
+    float2 fragCoord = in.uv * U.iResolution;
+    float2 p = (fragCoord * 2.0 - U.iResolution) / min(U.iResolution.x, U.iResolution.y);
+
+    float3 ro = float3(0.0, -0.2, U.iTime * 4.0);
+    float3 ray = normalize(float3(p, 1.5));
+    ray.xy = ray.xy * rot(sin(U.iTime * 0.03) * 5.0);
+    ray.yz = ray.yz * rot(sin(U.iTime * 0.05) * 0.2);
+    float t = 0.1;
+    float ac = 0.0;
+
+    for (int i = 0; i < 99; i++) {
+        float3 pos = ro + ray * t;
+        pos = fract((pos - 2.0) / 4.0) * 4.0 - 2.0;
+        float gTime = U.iTime - float(i) * 0.01;
+
+        float d = box_set(pos, gTime);
+        d = max(abs(d), 0.01);
+        ac += exp(-d * 23.0);
+        t += d * 0.55;
+    }
+
+    float3 col = float3(ac * 0.02);
+    col += float3(0.0, 0.2 * abs(sin(U.iTime)), 0.5 + sin(U.iTime) * 0.2);
+
+    return float4(col, 1.0 - t * (0.02 + 0.02 * sin(U.iTime)));
+}

data/shaders/octograms/octograms.gl.glsl

@@ -1,3 +1,6 @@
+// Name: Octograms
+// Author: whisky_shusuky
+// URL: https://www.shadertoy.com/view/tlVGDt
 #version 330 core
 precision highp float;
 

data/shaders/octograms/octograms.vk.glsl

@@ -0,0 +1,109 @@
+#version 450
+
+// Name: Octograms
+// Author: whisky_shusuky
+// URL: https://www.shadertoy.com/view/tlVGDt
+
+layout(location = 0) in  vec2 vUV;
+layout(location = 0) out vec4 FragColor;
+
+layout(set = 3, binding = 0) uniform ShadertoyUBO {
+    float iTime;
+    vec2  iResolution;
+};
+
+float gTime = 0.0;
+const float REPEAT = 5.0;
+
+mat2 rot(float a) {
+    float c = cos(a), s = sin(a);
+    return mat2(c, s, -s, c);
+}
+
+float sdBox(vec3 p, vec3 b) {
+    vec3 q = abs(p) - b;
+    return length(max(q, vec3(0.0))) + min(max(q.x, max(q.y, q.z)), 0.0);
+}
+
+float boxGeom(vec3 pos, float scale) {
+    pos *= scale;
+    float base = sdBox(pos, vec3(.4, .4, .1)) / 1.5;
+    pos.xy *= 5.0;
+    pos.y -= 3.5;
+    pos.xy *= rot(.75);
+    float result = -base;
+    return result;
+}
+
+float box_set(vec3 pos, float iTimeLocal) {
+    vec3 pos_origin = pos;
+    pos = pos_origin;
+    pos.y += sin(gTime * 0.4) * 2.5;
+    pos.xy *= rot(.8);
+    float box1 = boxGeom(pos, 2.0 - abs(sin(gTime * 0.4)) * 1.5);
+
+    pos = pos_origin;
+    pos.y -= sin(gTime * 0.4) * 2.5;
+    pos.xy *= rot(.8);
+    float box2 = boxGeom(pos, 2.0 - abs(sin(gTime * 0.4)) * 1.5);
+
+    pos = pos_origin;
+    pos.x += sin(gTime * 0.4) * 2.5;
+    pos.xy *= rot(.8);
+    float box3 = boxGeom(pos, 2.0 - abs(sin(gTime * 0.4)) * 1.5);
+
+    pos = pos_origin;
+    pos.x -= sin(gTime * 0.4) * 2.5;
+    pos.xy *= rot(.8);
+    float box4 = boxGeom(pos, 2.0 - abs(sin(gTime * 0.4)) * 1.5);
+
+    pos = pos_origin;
+    pos.xy *= rot(.8);
+    float box5 = boxGeom(pos, .5) * 6.0;
+
+    pos = pos_origin;
+    float box6 = boxGeom(pos, .5) * 6.0;
+
+    float result = max(max(max(max(max(box1, box2), box3), box4), box5), box6);
+    return result;
+}
+
+float mapScene(vec3 pos, float iTimeLocal) {
+    return box_set(pos, iTimeLocal);
+}
+
+void mainImage(out vec4 fragColor, in vec2 fragCoord) {
+    vec2 p = (fragCoord.xy * 2.0 - iResolution.xy) / min(iResolution.x, iResolution.y);
+    vec3 ro = vec3(0.0, -0.2, iTime * 4.0);
+    vec3 ray = normalize(vec3(p, 1.5));
+    ray.xy = ray.xy * rot(sin(iTime * .03) * 5.0);
+    ray.yz = ray.yz * rot(sin(iTime * .05) * .2);
+    float t = 0.1;
+    vec3 col = vec3(0.0);
+    float ac = 0.0;
+
+    for (int i = 0; i < 99; i++) {
+        vec3 pos = ro + ray * t;
+        pos = mod(pos - 2.0, 4.0) - 2.0;
+        gTime = iTime - float(i) * 0.01;
+
+        float d = mapScene(pos, iTime);
+
+        d = max(abs(d), 0.01);
+        ac += exp(-d * 23.0);
+
+        t += d * 0.55;
+    }
+
+    col = vec3(ac * 0.02);
+    col += vec3(0.0, 0.2 * abs(sin(iTime)), 0.5 + sin(iTime) * 0.2);
+
+    fragColor = vec4(col, 1.0 - t * (0.02 + 0.02 * sin(iTime)));
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

data/shaders/remember/meta.txt

@@ -0,0 +1,2 @@
+Name: Remember
+Author: diatribes

data/shaders/remember/remember.frag.msl

@@ -0,0 +1,64 @@
+#include <metal_stdlib>
+using namespace metal;
+
+// Remember — diatribes
+// MSL port of remember.vk.glsl.
+
+struct ShadertoyUBO {
+    float iTime;
+    float2 iResolution;
+};
+
+struct PassthroughVOut {
+    float4 pos [[position]];
+    float2 uv;
+};
+
+static float hash12(float2 p) {
+    float3 p3 = fract(float3(p.xyx) * 0.1031);
+    p3 += dot(p3, p3.yzx + 33.33);
+    return fract((p3.x + p3.y) * p3.z);
+}
+
+fragment float4 remember_fs(PassthroughVOut in [[stage_in]],
+                            constant ShadertoyUBO& U [[buffer(0)]]) {
+    float2 fragCoord = in.uv * U.iResolution;
+    float2 u = fragCoord;
+
+    float3 q, p = float3(U.iResolution, U.iResolution.x / U.iResolution.y);
+
+    float i = 0.0, s = 0.0;
+    float d = 0.125 * hash12(u);
+    float t = U.iTime * 0.1;
+
+    u = (u + u - p.xy) / p.y;
+    if (abs(u.y) > 0.8) { return float4(0); }
+
+    float4 o = float4(0.0);
+
+    for (; i < 64.0; i++) {
+
+        q = p = float3(u * d, d + t * 5.0);
+        // mat2(cos(.1*p.z+.1*t+vec4(0,33,11,0))) — column-major: (m00,m10,m01,m11)
+        float4 cs = cos(0.1 * p.z + 0.1 * t + float4(0, 33, 11, 0));
+        float2x2 m = float2x2(cs.x, cs.y, cs.z, cs.w);
+        p.xy = p.xy * m;
+
+        q.xz = cos(q.xz);
+        p.z = cos(p.z);
+
+        for (s = 1.0; s++ < 6.0;) {
+            q += sin(0.6 * t + p.zxy * 0.6);
+            p += sin(t + t + p.yzx * s) * 0.6;
+        }
+
+        s = 0.02 + abs(min(length(p + 3.0 * sin(p.z * 0.5)) - 4.0, length(q - 2.0 * sin(p.z * 0.4)) - 6.0)) * 0.2;
+        d += s;
+
+        float4 brightTerm = min(0.01 * float4(6, 2, 1, 0) / max(length(u * sin(t + t + t)), 0.001), float4(50.0));
+        o += brightTerm + 1.0 / s * length(u);
+    }
+
+    o = tanh(max(o / 6e2 + dot(u, u) * 0.35, 0.0));
+    return o;
+}

data/shaders/remember/remember.gl.glsl

@@ -0,0 +1,73 @@
+// Name: Remember
+// Author: diatribes
+// URL: https://www.shadertoy.com/view/tXSBDK
+#version 330 core
+precision highp float;
+
+out vec4 FragColor;
+in vec2 vUV;
+uniform vec2 iResolution;
+uniform float iTime;
+
+// fuzzy brain
+
+// Hash function to replace iChannel0 texture noise
+float hash12(vec2 p) {
+    vec3 p3 = fract(vec3(p.xyx) * .1031);
+    p3 += dot(p3, p3.yzx + 33.33);
+    return fract((p3.x + p3.y) * p3.z);
+}
+
+void mainImage(out vec4 o, vec2 u) {
+
+    vec3 q,p = vec3(iResolution.xy, iResolution.x / iResolution.y);
+
+    float i = 0.0, s,
+          // start the ray at a small random distance,
+          // this will reduce banding
+          // Replaced texelFetch(iChannel0, ...) with hash function
+          d = .125 * hash12(u),
+          t = iTime * .1;
+
+    // scale coords
+    u = (u+u-p.xy)/p.y;
+    if (abs(u.y) > .8) { o = vec4(0); return; }
+
+    // Initialize output color (out parameter must be initialized before use)
+    o = vec4(0.0);
+
+    for(; i<64.; i++) {
+
+        // shorthand for standard raymarch sample, then move forward:
+        // p = ro + rd * d, p.z + t
+        q = p = vec3(u * d, d + t*5.);
+        p.xy *= mat2(cos(.1*p.z+.1*t+vec4(0,33,11,0)));
+
+        q.xz = cos(q.xz);
+        p.z = cos(p.z) ;
+        // turbulence
+        for (s = 1.; s++ <6.;
+            q += sin(.6*t+p.zxy*.6),
+            p += sin(t+t+p.yzx*s)*.6);
+
+        // distance to spheres
+        d += s = .02 + abs(min(length(p+3.*sin(p.z*.5))-4., length(q-2.*sin(p.z*.4))-6.))*.2;
+
+        // color: 1.+cos so we don't go negative, cos(d+vec4(6,4,2,0)) samples from the palette
+        // divide by s for form and distance
+        // Clamp only the first term to prevent extreme overflow, leave second term free
+        vec4 brightTerm = min(.01*vec4(6,2,1,0)/max(length(u*sin(t+t+t)), 0.001), vec4(50.0));
+        o += brightTerm + 1. / s * length(u);
+
+    }
+
+    // tonemap and divide brightness
+    o = tanh(max(o /6e2 + dot(u,u)*.35, 0.));
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

data/shaders/remember/remember.vk.glsl

@@ -0,0 +1,61 @@
+#version 450
+
+// Name: Remember
+// Author: diatribes
+// URL: https://www.shadertoy.com/view/tXSBDK
+
+layout(location = 0) in  vec2 vUV;
+layout(location = 0) out vec4 FragColor;
+
+layout(set = 3, binding = 0) uniform ShadertoyUBO {
+    float iTime;
+    vec2  iResolution;
+};
+
+// fuzzy brain — Hash function to replace iChannel0 texture noise
+float hash12(vec2 p) {
+    vec3 p3 = fract(vec3(p.xyx) * .1031);
+    p3 += dot(p3, p3.yzx + 33.33);
+    return fract((p3.x + p3.y) * p3.z);
+}
+
+void mainImage(out vec4 o, vec2 u) {
+
+    vec3 q, p = vec3(iResolution.xy, iResolution.x / iResolution.y);
+
+    float i = 0.0, s,
+          d = .125 * hash12(u),
+          t = iTime * .1;
+
+    u = (u + u - p.xy) / p.y;
+    if (abs(u.y) > .8) { o = vec4(0); return; }
+
+    o = vec4(0.0);
+
+    for (; i < 64.; i++) {
+
+        q = p = vec3(u * d, d + t * 5.);
+        p.xy *= mat2(cos(.1 * p.z + .1 * t + vec4(0, 33, 11, 0)));
+
+        q.xz = cos(q.xz);
+        p.z = cos(p.z);
+        for (s = 1.; s++ < 6.;
+            q += sin(.6 * t + p.zxy * .6),
+            p += sin(t + t + p.yzx * s) * .6);
+
+        d += s = .02 + abs(min(length(p + 3. * sin(p.z * .5)) - 4., length(q - 2. * sin(p.z * .4)) - 6.)) * .2;
+
+        vec4 brightTerm = min(.01 * vec4(6, 2, 1, 0) / max(length(u * sin(t + t + t)), 0.001), vec4(50.0));
+        o += brightTerm + 1. / s * length(u);
+
+    }
+
+    o = tanh(max(o / 6e2 + dot(u, u) * .35, 0.));
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

data/shaders/seascape/meta.txt

@@ -0,0 +1,2 @@
+Name: Seascape
+Author: Alexander Alekseev

data/shaders/seascape/seascape.frag.msl

@@ -0,0 +1,181 @@
+#include <metal_stdlib>
+using namespace metal;
+
+// Seascape — Alexander Alekseev (TDM, 2014)
+// MSL port of seascape.vk.glsl. iTime threaded through helpers because
+// MSL does not allow file-scope mutable globals.
+
+struct ShadertoyUBO {
+    float iTime;
+    float2 iResolution;
+};
+
+struct PassthroughVOut {
+    float4 pos [[position]];
+    float2 uv;
+};
+
+constant int NUM_STEPS = 32;
+constant float PI = 3.141592;
+constant float EPSILON = 1e-3;
+
+constant int ITER_GEOMETRY = 3;
+constant int ITER_FRAGMENT = 5;
+constant float SEA_HEIGHT = 0.6;
+constant float SEA_CHOPPY = 4.0;
+constant float SEA_SPEED = 0.8;
+constant float SEA_FREQ = 0.16;
+constant float3 SEA_BASE = float3(0.0, 0.09, 0.18);
+constant float3 SEA_WATER_COLOR = float3(0.8, 0.9, 0.6) * 0.6;
+constant float2x2 octave_m = float2x2(1.6, 1.2, -1.2, 1.6);
+
+static float3x3 fromEuler(float3 ang) {
+    float2 a1 = float2(sin(ang.x), cos(ang.x));
+    float2 a2 = float2(sin(ang.y), cos(ang.y));
+    float2 a3 = float2(sin(ang.z), cos(ang.z));
+    float3x3 m;
+    m[0] = float3(a1.y * a3.y + a1.x * a2.x * a3.x, a1.y * a2.x * a3.x + a3.y * a1.x, -a2.y * a3.x);
+    m[1] = float3(-a2.y * a1.x, a1.y * a2.y, a2.x);
+    m[2] = float3(a3.y * a1.x * a2.x + a1.y * a3.x, a1.x * a3.x - a1.y * a3.y * a2.x, a2.y * a3.y);
+    return m;
+}
+
+static float hash(float2 p) {
+    float h = dot(p, float2(127.1, 311.7));
+    return fract(sin(h) * 43758.5453123);
+}
+
+static float noise(float2 p) {
+    float2 i = floor(p);
+    float2 f = fract(p);
+    float2 u = f * f * (3.0 - 2.0 * f);
+    return -1.0 + 2.0 * mix(mix(hash(i + float2(0.0, 0.0)),
+                                hash(i + float2(1.0, 0.0)), u.x),
+                            mix(hash(i + float2(0.0, 1.0)),
+                                hash(i + float2(1.0, 1.0)), u.x), u.y);
+}
+
+static float diffuseLight(float3 n, float3 l, float p) {
+    return pow(dot(n, l) * 0.4 + 0.6, p);
+}
+
+static float specularLight(float3 n, float3 l, float3 e, float s) {
+    float nrm = (s + 8.0) / (PI * 8.0);
+    return pow(max(dot(reflect(e, n), l), 0.0), s) * nrm;
+}
+
+static float3 getSkyColor(float3 e) {
+    e.y = (max(e.y, 0.0) * 0.8 + 0.2) * 0.8;
+    return float3(pow(1.0 - e.y, 2.0), 1.0 - e.y, 0.6 + (1.0 - e.y) * 0.4) * 1.1;
+}
+
+static float sea_octave(float2 uv, float choppy) {
+    uv += noise(uv);
+    float2 wv = 1.0 - abs(sin(uv));
+    float2 swv = abs(cos(uv));
+    wv = mix(wv, swv, wv);
+    return pow(1.0 - pow(wv.x * wv.y, 0.65), choppy);
+}
+
+static float seaMap(float3 p, float iTime, int iter) {
+    float SEA_TIME = 1.0 + iTime * SEA_SPEED;
+    float freq = SEA_FREQ;
+    float amp = SEA_HEIGHT;
+    float choppy = SEA_CHOPPY;
+    float2 uv = p.xz; uv.x *= 0.75;
+
+    float d, h = 0.0;
+    for (int i = 0; i < iter; i++) {
+        d = sea_octave((uv + SEA_TIME) * freq, choppy);
+        d += sea_octave((uv - SEA_TIME) * freq, choppy);
+        h += d * amp;
+        uv = uv * octave_m;
+        freq *= 1.9;
+        amp *= 0.22;
+        choppy = mix(choppy, 1.0, 0.2);
+    }
+    return p.y - h;
+}
+
+static float3 getSeaColor(float3 p, float3 n, float3 l, float3 eye, float3 dist) {
+    float fresnel = clamp(1.0 - dot(n, -eye), 0.0, 1.0);
+    fresnel = min(fresnel * fresnel * fresnel, 0.5);
+
+    float3 reflected = getSkyColor(reflect(eye, n));
+    float3 refracted = SEA_BASE + diffuseLight(n, l, 80.0) * SEA_WATER_COLOR * 0.12;
+
+    float3 color = mix(refracted, reflected, fresnel);
+
+    float atten = max(1.0 - dot(dist, dist) * 0.001, 0.0);
+    color += SEA_WATER_COLOR * (p.y - SEA_HEIGHT) * 0.18 * atten;
+
+    color += specularLight(n, l, eye, 600.0 * rsqrt(dot(dist, dist)));
+
+    return color;
+}
+
+static float3 getNormal(float3 p, float eps, float iTime) {
+    float3 n;
+    n.y = seaMap(p, iTime, ITER_FRAGMENT);
+    n.x = seaMap(float3(p.x + eps, p.y, p.z), iTime, ITER_FRAGMENT) - n.y;
+    n.z = seaMap(float3(p.x, p.y, p.z + eps), iTime, ITER_FRAGMENT) - n.y;
+    n.y = eps;
+    return normalize(n);
+}
+
+static float heightMapTracing(float3 ori, float3 dir, thread float3& p, float iTime) {
+    float tm = 0.0;
+    float tx = 1000.0;
+    float hx = seaMap(ori + dir * tx, iTime, ITER_GEOMETRY);
+    if (hx > 0.0) {
+        p = ori + dir * tx;
+        return tx;
+    }
+    float hm = seaMap(ori, iTime, ITER_GEOMETRY);
+    for (int i = 0; i < NUM_STEPS; i++) {
+        float tmid = mix(tm, tx, hm / (hm - hx));
+        p = ori + dir * tmid;
+        float hmid = seaMap(p, iTime, ITER_GEOMETRY);
+        if (hmid < 0.0) {
+            tx = tmid;
+            hx = hmid;
+        } else {
+            tm = tmid;
+            hm = hmid;
+        }
+        if (abs(hmid) < EPSILON) break;
+    }
+    return mix(tm, tx, hm / (hm - hx));
+}
+
+static float3 getPixel(float2 coord, float time, float2 iResolution, float iTime) {
+    float2 uv = coord / iResolution;
+    uv = uv * 2.0 - 1.0;
+    uv.x *= iResolution.x / iResolution.y;
+
+    float3 ang = float3(sin(time * 3.0) * 0.1, sin(time) * 0.2 + 0.3, time);
+    float3 ori = float3(0.0, 3.5, time * 5.0);
+    float3 dir = normalize(float3(uv.xy, -2.0));
+    dir.z += length(uv) * 0.14;
+    dir = normalize(dir) * fromEuler(ang);
+
+    float3 p;
+    heightMapTracing(ori, dir, p, iTime);
+    float3 dist = p - ori;
+    float EPSILON_NRM = 0.1 / iResolution.x;
+    float3 n = getNormal(p, dot(dist, dist) * EPSILON_NRM, iTime);
+    float3 light = normalize(float3(0.0, 1.0, 0.8));
+
+    return mix(
+        getSkyColor(dir),
+        getSeaColor(p, n, light, dir, dist),
+        pow(smoothstep(0.0, -0.02, dir.y), 0.2));
+}
+
+fragment float4 seascape_fs(PassthroughVOut in [[stage_in]],
+                            constant ShadertoyUBO& U [[buffer(0)]]) {
+    float2 fragCoord = in.uv * U.iResolution;
+    float time = U.iTime * 0.3;
+    float3 color = getPixel(fragCoord, time, U.iResolution, U.iTime);
+    return float4(pow(color, float3(0.65)), 1.0);
+}

data/shaders/seascape/seascape.gl.glsl

@@ -0,0 +1,223 @@
+// Name: Seascape
+// Author: Alexander Alekseev
+// URL: https://www.shadertoy.com/view/Ms2SD1
+#version 330 core
+precision highp float;
+
+out vec4 FragColor;
+in vec2 vUV;
+uniform vec2 iResolution;
+uniform float iTime;
+
+/*
+ * "Seascape" by Alexander Alekseev aka TDM - 2014
+ * License Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License.
+ * Contact: tdmaav@gmail.com
+ */
+
+const int NUM_STEPS = 32;
+const float PI         = 3.141592;
+const float EPSILON    = 1e-3;
+#define EPSILON_NRM (0.1 / iResolution.x)
+//#define AA
+
+// sea
+const int ITER_GEOMETRY = 3;
+const int ITER_FRAGMENT = 5;
+const float SEA_HEIGHT = 0.6;
+const float SEA_CHOPPY = 4.0;
+const float SEA_SPEED = 0.8;
+const float SEA_FREQ = 0.16;
+const vec3 SEA_BASE = vec3(0.0,0.09,0.18);
+const vec3 SEA_WATER_COLOR = vec3(0.8,0.9,0.6)*0.6;
+#define SEA_TIME (1.0 + iTime * SEA_SPEED)
+const mat2 octave_m = mat2(1.6,1.2,-1.2,1.6);
+
+// math
+mat3 fromEuler(vec3 ang) {
+    vec2 a1 = vec2(sin(ang.x),cos(ang.x));
+    vec2 a2 = vec2(sin(ang.y),cos(ang.y));
+    vec2 a3 = vec2(sin(ang.z),cos(ang.z));
+    mat3 m;
+    m[0] = vec3(a1.y*a3.y+a1.x*a2.x*a3.x,a1.y*a2.x*a3.x+a3.y*a1.x,-a2.y*a3.x);
+    m[1] = vec3(-a2.y*a1.x,a1.y*a2.y,a2.x);
+    m[2] = vec3(a3.y*a1.x*a2.x+a1.y*a3.x,a1.x*a3.x-a1.y*a3.y*a2.x,a2.y*a3.y);
+    return m;
+}
+float hash( vec2 p ) {
+    float h = dot(p,vec2(127.1,311.7));
+    return fract(sin(h)*43758.5453123);
+}
+float noise( in vec2 p ) {
+    vec2 i = floor( p );
+    vec2 f = fract( p );
+    vec2 u = f*f*(3.0-2.0*f);
+    return -1.0+2.0*mix( mix( hash( i + vec2(0.0,0.0) ),
+                     hash( i + vec2(1.0,0.0) ), u.x),
+                mix( hash( i + vec2(0.0,1.0) ),
+                     hash( i + vec2(1.0,1.0) ), u.x), u.y);
+}
+
+// lighting
+float diffuse(vec3 n,vec3 l,float p) {
+    return pow(dot(n,l) * 0.4 + 0.6,p);
+}
+float specular(vec3 n,vec3 l,vec3 e,float s) {
+    float nrm = (s + 8.0) / (PI * 8.0);
+    return pow(max(dot(reflect(e,n),l),0.0),s) * nrm;
+}
+
+// sky
+vec3 getSkyColor(vec3 e) {
+    e.y = (max(e.y,0.0)*0.8+0.2)*0.8;
+    return vec3(pow(1.0-e.y,2.0), 1.0-e.y, 0.6+(1.0-e.y)*0.4) * 1.1;
+}
+
+// sea
+float sea_octave(vec2 uv, float choppy) {
+    uv += noise(uv);
+    vec2 wv = 1.0-abs(sin(uv));
+    vec2 swv = abs(cos(uv));
+    wv = mix(wv,swv,wv);
+    return pow(1.0-pow(wv.x * wv.y,0.65),choppy);
+}
+
+float map(vec3 p) {
+    float freq = SEA_FREQ;
+    float amp = SEA_HEIGHT;
+    float choppy = SEA_CHOPPY;
+    vec2 uv = p.xz; uv.x *= 0.75;
+
+    float d, h = 0.0;
+    for(int i = 0; i < ITER_GEOMETRY; i++) {
+        d = sea_octave((uv+SEA_TIME)*freq,choppy);
+        d += sea_octave((uv-SEA_TIME)*freq,choppy);
+        h += d * amp;
+        uv *= octave_m; freq *= 1.9; amp *= 0.22;
+        choppy = mix(choppy,1.0,0.2);
+    }
+    return p.y - h;
+}
+
+float map_detailed(vec3 p) {
+    float freq = SEA_FREQ;
+    float amp = SEA_HEIGHT;
+    float choppy = SEA_CHOPPY;
+    vec2 uv = p.xz; uv.x *= 0.75;
+
+    float d, h = 0.0;
+    for(int i = 0; i < ITER_FRAGMENT; i++) {
+        d = sea_octave((uv+SEA_TIME)*freq,choppy);
+        d += sea_octave((uv-SEA_TIME)*freq,choppy);
+        h += d * amp;
+        uv *= octave_m; freq *= 1.9; amp *= 0.22;
+        choppy = mix(choppy,1.0,0.2);
+    }
+    return p.y - h;
+}
+
+vec3 getSeaColor(vec3 p, vec3 n, vec3 l, vec3 eye, vec3 dist) {
+    float fresnel = clamp(1.0 - dot(n, -eye), 0.0, 1.0);
+    fresnel = min(fresnel * fresnel * fresnel, 0.5);
+
+    vec3 reflected = getSkyColor(reflect(eye, n));
+    vec3 refracted = SEA_BASE + diffuse(n, l, 80.0) * SEA_WATER_COLOR * 0.12;
+
+    vec3 color = mix(refracted, reflected, fresnel);
+
+    float atten = max(1.0 - dot(dist, dist) * 0.001, 0.0);
+    color += SEA_WATER_COLOR * (p.y - SEA_HEIGHT) * 0.18 * atten;
+
+    color += specular(n, l, eye, 600.0 * inversesqrt(dot(dist,dist)));
+
+    return color;
+}
+
+// tracing
+vec3 getNormal(vec3 p, float eps) {
+    vec3 n;
+    n.y = map_detailed(p);
+    n.x = map_detailed(vec3(p.x+eps,p.y,p.z)) - n.y;
+    n.z = map_detailed(vec3(p.x,p.y,p.z+eps)) - n.y;
+    n.y = eps;
+    return normalize(n);
+}
+
+float heightMapTracing(vec3 ori, vec3 dir, out vec3 p) {
+    float tm = 0.0;
+    float tx = 1000.0;
+    float hx = map(ori + dir * tx);
+    if(hx > 0.0) {
+        p = ori + dir * tx;
+        return tx;
+    }
+    float hm = map(ori);
+    for(int i = 0; i < NUM_STEPS; i++) {
+        float tmid = mix(tm, tx, hm / (hm - hx));
+        p = ori + dir * tmid;
+        float hmid = map(p);
+        if(hmid < 0.0) {
+            tx = tmid;
+            hx = hmid;
+        } else {
+            tm = tmid;
+            hm = hmid;
+        }
+        if(abs(hmid) < EPSILON) break;
+    }
+    return mix(tm, tx, hm / (hm - hx));
+}
+
+vec3 getPixel(in vec2 coord, float time) {
+    vec2 uv = coord / iResolution.xy;
+    uv = uv * 2.0 - 1.0;
+    uv.x *= iResolution.x / iResolution.y;
+
+    // ray
+    vec3 ang = vec3(sin(time*3.0)*0.1,sin(time)*0.2+0.3,time);
+    vec3 ori = vec3(0.0,3.5,time*5.0);
+    vec3 dir = normalize(vec3(uv.xy,-2.0)); dir.z += length(uv) * 0.14;
+    dir = normalize(dir) * fromEuler(ang);
+
+    // tracing
+    vec3 p;
+    heightMapTracing(ori,dir,p);
+    vec3 dist = p - ori;
+    vec3 n = getNormal(p, dot(dist,dist) * EPSILON_NRM);
+    vec3 light = normalize(vec3(0.0,1.0,0.8));
+
+    // color
+    return mix(
+        getSkyColor(dir),
+        getSeaColor(p,n,light,dir,dist),
+        pow(smoothstep(0.0,-0.02,dir.y),0.2));
+}
+
+// main
+void mainImage( out vec4 fragColor, in vec2 fragCoord ) {
+    // Removed mouse interaction (iMouse not available)
+    float time = iTime * 0.3;
+
+#ifdef AA
+    vec3 color = vec3(0.0);
+    for(int i = -1; i <= 1; i++) {
+        for(int j = -1; j <= 1; j++) {
+            vec2 uv = fragCoord+vec2(i,j)/3.0;
+            color += getPixel(uv, time);
+        }
+    }
+    color /= 9.0;
+#else
+    vec3 color = getPixel(fragCoord, time);
+#endif
+
+    // post
+    fragColor = vec4(pow(color,vec3(0.65)), 1.0);
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

data/shaders/seascape/seascape.vk.glsl

@@ -0,0 +1,200 @@
+#version 450
+
+// Name: Seascape
+// Author: Alexander Alekseev
+// URL: https://www.shadertoy.com/view/Ms2SD1
+//
+// "Seascape" by Alexander Alekseev aka TDM - 2014
+// CC BY-NC-SA 3.0 Unported License.
+
+layout(location = 0) in  vec2 vUV;
+layout(location = 0) out vec4 FragColor;
+
+layout(set = 3, binding = 0) uniform ShadertoyUBO {
+    float iTime;
+    vec2  iResolution;
+};
+
+const int NUM_STEPS = 32;
+const float PI = 3.141592;
+const float EPSILON = 1e-3;
+#define EPSILON_NRM (0.1 / iResolution.x)
+
+const int ITER_GEOMETRY = 3;
+const int ITER_FRAGMENT = 5;
+const float SEA_HEIGHT = 0.6;
+const float SEA_CHOPPY = 4.0;
+const float SEA_SPEED = 0.8;
+const float SEA_FREQ = 0.16;
+const vec3 SEA_BASE = vec3(0.0, 0.09, 0.18);
+const vec3 SEA_WATER_COLOR = vec3(0.8, 0.9, 0.6) * 0.6;
+#define SEA_TIME (1.0 + iTime * SEA_SPEED)
+const mat2 octave_m = mat2(1.6, 1.2, -1.2, 1.6);
+
+mat3 fromEuler(vec3 ang) {
+    vec2 a1 = vec2(sin(ang.x), cos(ang.x));
+    vec2 a2 = vec2(sin(ang.y), cos(ang.y));
+    vec2 a3 = vec2(sin(ang.z), cos(ang.z));
+    mat3 m;
+    m[0] = vec3(a1.y * a3.y + a1.x * a2.x * a3.x, a1.y * a2.x * a3.x + a3.y * a1.x, -a2.y * a3.x);
+    m[1] = vec3(-a2.y * a1.x, a1.y * a2.y, a2.x);
+    m[2] = vec3(a3.y * a1.x * a2.x + a1.y * a3.x, a1.x * a3.x - a1.y * a3.y * a2.x, a2.y * a3.y);
+    return m;
+}
+
+float hash(vec2 p) {
+    float h = dot(p, vec2(127.1, 311.7));
+    return fract(sin(h) * 43758.5453123);
+}
+
+float noise(in vec2 p) {
+    vec2 i = floor(p);
+    vec2 f = fract(p);
+    vec2 u = f * f * (3.0 - 2.0 * f);
+    return -1.0 + 2.0 * mix(mix(hash(i + vec2(0.0, 0.0)),
+                                hash(i + vec2(1.0, 0.0)), u.x),
+                            mix(hash(i + vec2(0.0, 1.0)),
+                                hash(i + vec2(1.0, 1.0)), u.x), u.y);
+}
+
+float diffuse(vec3 n, vec3 l, float p) {
+    return pow(dot(n, l) * 0.4 + 0.6, p);
+}
+
+float specular(vec3 n, vec3 l, vec3 e, float s) {
+    float nrm = (s + 8.0) / (PI * 8.0);
+    return pow(max(dot(reflect(e, n), l), 0.0), s) * nrm;
+}
+
+vec3 getSkyColor(vec3 e) {
+    e.y = (max(e.y, 0.0) * 0.8 + 0.2) * 0.8;
+    return vec3(pow(1.0 - e.y, 2.0), 1.0 - e.y, 0.6 + (1.0 - e.y) * 0.4) * 1.1;
+}
+
+float sea_octave(vec2 uv, float choppy) {
+    uv += noise(uv);
+    vec2 wv = 1.0 - abs(sin(uv));
+    vec2 swv = abs(cos(uv));
+    wv = mix(wv, swv, wv);
+    return pow(1.0 - pow(wv.x * wv.y, 0.65), choppy);
+}
+
+float map(vec3 p) {
+    float freq = SEA_FREQ;
+    float amp = SEA_HEIGHT;
+    float choppy = SEA_CHOPPY;
+    vec2 uv = p.xz; uv.x *= 0.75;
+
+    float d, h = 0.0;
+    for (int i = 0; i < ITER_GEOMETRY; i++) {
+        d = sea_octave((uv + SEA_TIME) * freq, choppy);
+        d += sea_octave((uv - SEA_TIME) * freq, choppy);
+        h += d * amp;
+        uv *= octave_m; freq *= 1.9; amp *= 0.22;
+        choppy = mix(choppy, 1.0, 0.2);
+    }
+    return p.y - h;
+}
+
+float map_detailed(vec3 p) {
+    float freq = SEA_FREQ;
+    float amp = SEA_HEIGHT;
+    float choppy = SEA_CHOPPY;
+    vec2 uv = p.xz; uv.x *= 0.75;
+
+    float d, h = 0.0;
+    for (int i = 0; i < ITER_FRAGMENT; i++) {
+        d = sea_octave((uv + SEA_TIME) * freq, choppy);
+        d += sea_octave((uv - SEA_TIME) * freq, choppy);
+        h += d * amp;
+        uv *= octave_m; freq *= 1.9; amp *= 0.22;
+        choppy = mix(choppy, 1.0, 0.2);
+    }
+    return p.y - h;
+}
+
+vec3 getSeaColor(vec3 p, vec3 n, vec3 l, vec3 eye, vec3 dist) {
+    float fresnel = clamp(1.0 - dot(n, -eye), 0.0, 1.0);
+    fresnel = min(fresnel * fresnel * fresnel, 0.5);
+
+    vec3 reflected = getSkyColor(reflect(eye, n));
+    vec3 refracted = SEA_BASE + diffuse(n, l, 80.0) * SEA_WATER_COLOR * 0.12;
+
+    vec3 color = mix(refracted, reflected, fresnel);
+
+    float atten = max(1.0 - dot(dist, dist) * 0.001, 0.0);
+    color += SEA_WATER_COLOR * (p.y - SEA_HEIGHT) * 0.18 * atten;
+
+    color += specular(n, l, eye, 600.0 * inversesqrt(dot(dist, dist)));
+
+    return color;
+}
+
+vec3 getNormal(vec3 p, float eps) {
+    vec3 n;
+    n.y = map_detailed(p);
+    n.x = map_detailed(vec3(p.x + eps, p.y, p.z)) - n.y;
+    n.z = map_detailed(vec3(p.x, p.y, p.z + eps)) - n.y;
+    n.y = eps;
+    return normalize(n);
+}
+
+float heightMapTracing(vec3 ori, vec3 dir, out vec3 p) {
+    float tm = 0.0;
+    float tx = 1000.0;
+    float hx = map(ori + dir * tx);
+    if (hx > 0.0) {
+        p = ori + dir * tx;
+        return tx;
+    }
+    float hm = map(ori);
+    for (int i = 0; i < NUM_STEPS; i++) {
+        float tmid = mix(tm, tx, hm / (hm - hx));
+        p = ori + dir * tmid;
+        float hmid = map(p);
+        if (hmid < 0.0) {
+            tx = tmid;
+            hx = hmid;
+        } else {
+            tm = tmid;
+            hm = hmid;
+        }
+        if (abs(hmid) < EPSILON) break;
+    }
+    return mix(tm, tx, hm / (hm - hx));
+}
+
+vec3 getPixel(in vec2 coord, float time) {
+    vec2 uv = coord / iResolution.xy;
+    uv = uv * 2.0 - 1.0;
+    uv.x *= iResolution.x / iResolution.y;
+
+    vec3 ang = vec3(sin(time * 3.0) * 0.1, sin(time) * 0.2 + 0.3, time);
+    vec3 ori = vec3(0.0, 3.5, time * 5.0);
+    vec3 dir = normalize(vec3(uv.xy, -2.0)); dir.z += length(uv) * 0.14;
+    dir = normalize(dir) * fromEuler(ang);
+
+    vec3 p;
+    heightMapTracing(ori, dir, p);
+    vec3 dist = p - ori;
+    vec3 n = getNormal(p, dot(dist, dist) * EPSILON_NRM);
+    vec3 light = normalize(vec3(0.0, 1.0, 0.8));
+
+    return mix(
+        getSkyColor(dir),
+        getSeaColor(p, n, light, dir, dist),
+        pow(smoothstep(0.0, -0.02, dir.y), 0.2));
+}
+
+void mainImage(out vec4 fragColor, in vec2 fragCoord) {
+    float time = iTime * 0.3;
+    vec3 color = getPixel(fragCoord, time);
+    fragColor = vec4(pow(color, vec3(0.65)), 1.0);
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

data/shaders/shader_art_coding_introduction/meta.txt

@@ -0,0 +1,2 @@
+Name: Shader Art Coding Introduction
+Author: kishimisu

data/shaders/shader_art_coding_introduction/shader_art_coding_introduction.frag.msl

@@ -0,0 +1,47 @@
+#include <metal_stdlib>
+using namespace metal;
+
+// Shader Art Coding Introduction — kishimisu
+// MSL port of shader_art_coding_introduction.vk.glsl.
+
+struct ShadertoyUBO {
+    float iTime;
+    float2 iResolution;
+};
+
+struct PassthroughVOut {
+    float4 pos [[position]];
+    float2 uv;
+};
+
+static float3 palette(float t) {
+    float3 a = float3(0.5, 0.5, 0.5);
+    float3 b = float3(0.5, 0.5, 0.5);
+    float3 c = float3(1.0, 1.0, 1.0);
+    float3 d = float3(0.263, 0.416, 0.557);
+    return a + b * cos(6.28318 * (c * t * d));
+}
+
+fragment float4 shader_art_coding_introduction_fs(PassthroughVOut in [[stage_in]],
+                                                   constant ShadertoyUBO& U [[buffer(0)]]) {
+    float2 fragCoord = in.uv * U.iResolution;
+    float2 uv = (fragCoord * 2.0 - U.iResolution) / U.iResolution.y;
+    float2 uv0 = uv;
+    float3 finalColor = float3(0.0);
+
+    for (float i = 0.0; i < 4.0; i++) {
+        uv = fract(uv * 1.5) - 0.5;
+
+        float d = length(uv) * exp(-length(uv0));
+
+        float3 col = palette(length(uv0) + i * 0.4 + U.iTime * 0.4);
+
+        d = sin(d * 8.0 + U.iTime) / 8.0;
+        d = abs(d);
+        d = pow(0.01 / d, 1.2);
+
+        finalColor += col * d;
+    }
+
+    return float4(finalColor, 1.0);
+}

data/shaders/shader_art_coding_introduction/shader_art_coding_introduction.gl.glsl

@@ -1,3 +1,6 @@
+// Name: Shader Art Coding Introduction
+// Author: kishimisu
+// URL: https://www.shadertoy.com/view/mtyGWy
 #version 330 core
 precision highp float;
 

data/shaders/shader_art_coding_introduction/shader_art_coding_introduction.vk.glsl

@@ -0,0 +1,50 @@
+#version 450
+
+// Name: Shader Art Coding Introduction
+// Author: kishimisu
+// URL: https://www.shadertoy.com/view/mtyGWy
+
+layout(location = 0) in  vec2 vUV;
+layout(location = 0) out vec4 FragColor;
+
+layout(set = 3, binding = 0) uniform ShadertoyUBO {
+    float iTime;
+    vec2  iResolution;
+};
+
+vec3 palette(float t) {
+    vec3 a = vec3(0.5, 0.5, 0.5);
+    vec3 b = vec3(0.5, 0.5, 0.5);
+    vec3 c = vec3(1.0, 1.0, 1.0);
+    vec3 d = vec3(0.263, 0.416, 0.557);
+    return a + b * cos(6.28318 * (c * t * d));
+}
+
+void mainImage(out vec4 fragColor, in vec2 fragCoord) {
+    vec2 uv = (fragCoord * 2.0 - iResolution.xy) / iResolution.y;
+    vec2 uv0 = uv;
+    vec3 finalColor = vec3(0.0);
+
+    for (float i = 0.0; i < 4.0; i++) {
+        uv = fract(uv * 1.5) - 0.5;
+
+        float d = length(uv) * exp(-length(uv0));
+
+        vec3 col = palette(length(uv0) + i * 0.4 + iTime * 0.4);
+
+        d = sin(d * 8.0 + iTime) / 8.0;
+        d = abs(d);
+        d = pow(0.01 / d, 1.2);
+
+        finalColor += col * d;
+    }
+
+    fragColor = vec4(finalColor, 1.0);
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

data/shaders/test/meta.txt

@@ -0,0 +1,2 @@
+Name: Test
+Author: JailDesigner

data/shaders/test/test.frag.msl

@@ -0,0 +1,36 @@
+#include <metal_stdlib>
+using namespace metal;
+
+// Test shader (Metal Shading Language port of test.vk.glsl).
+// Author: JailDesigner
+
+struct ShadertoyUBO {
+    float iTime;
+    float2 iResolution;
+};
+
+struct PassthroughVOut {
+    float4 pos [[position]];
+    float2 uv;
+};
+
+static float3 palette(float t) {
+    float3 a = float3(1.0, 0.5, 0.5);
+    float3 b = float3(1.0, 0.5, 0.5);
+    float3 c = float3(1.0, 1.0, 1.0);
+    float3 d = float3(0.263, 0.416, 0.557);
+    return a + b * cos(6.28318 * (c * t * d));
+}
+
+fragment float4 test_fs(PassthroughVOut in [[stage_in]],
+                        constant ShadertoyUBO& u [[buffer(0)]]) {
+    float2 fragCoord = in.uv * u.iResolution;
+    float2 uv = (fragCoord * 2.0 - u.iResolution) / u.iResolution.y;
+    float d = length(uv);
+    float3 col = palette(d);
+    d = sin(d * 8.0 + u.iTime) / 8.0;
+    d = abs(d);
+    d = 0.02 / d;
+    col *= d;
+    return float4(col, 1.0);
+}

data/shaders/test/test.gl.glsl

@@ -1,3 +1,5 @@
+// Name: Test
+// Author: JailDesigner
 #version 330 core
 out vec4 FragColor;
 in vec2 vUV;

data/shaders/test/test.vk.glsl

@@ -0,0 +1,42 @@
+#version 450
+
+// Name: Test
+// Author: JailDesigner
+//
+// Vulkan port of test.gl.glsl. Bindings follow the SDL3 GPU convention:
+//   set=3, binding=0 — fragment uniform buffer (anonymous block: members
+//   are accessible directly as iTime / iResolution, like Shadertoy).
+
+layout(location = 0) in  vec2 vUV;
+layout(location = 0) out vec4 FragColor;
+
+layout(set = 3, binding = 0) uniform ShadertoyUBO {
+    float iTime;
+    vec2  iResolution;
+};
+
+vec3 palette(float t) {
+    vec3 a = vec3(1.0, 0.5, 0.5);
+    vec3 b = vec3(1.0, 0.5, 0.5);
+    vec3 c = vec3(1.0, 1.0, 1.0);
+    vec3 d = vec3(0.263, 0.416, 0.557);
+    return a + b * cos(6.28318 * (c * t * d));
+}
+
+void mainImage(out vec4 fragColor, in vec2 fragCoord) {
+    vec2 uv = (fragCoord * 2.0 - iResolution) / iResolution.y;
+    float d = length(uv);
+    vec3 col = palette(d);
+    d = sin(d * 8.0 + iTime) / 8.0;
+    d = abs(d);
+    d = 0.02 / d;
+    col *= d;
+    fragColor = vec4(col, 1.0);
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

data/shaders/water/meta.txt

@@ -0,0 +1,2 @@
+Name: Water
+Author: diatribes

data/shaders/water/water.frag.msl

@@ -0,0 +1,40 @@
+#include <metal_stdlib>
+using namespace metal;
+
+// Water — diatribes
+// MSL port of water.vk.glsl. Entry point is named after the shader (water_fs).
+
+struct ShadertoyUBO {
+    float iTime;
+    float2 iResolution;
+};
+
+struct PassthroughVOut {
+    float4 pos [[position]];
+    float2 uv;
+};
+
+fragment float4 water_fs(PassthroughVOut in [[stage_in]],
+                         constant ShadertoyUBO& U [[buffer(0)]]) {
+    float2 fragCoord = in.uv * U.iResolution;
+    float2 u = fragCoord;
+
+    float s = 0.002, i = 0.0, n;
+    float3 r = float3(U.iResolution, U.iResolution.x / U.iResolution.y);
+    float3 p = float3(0);
+    u = (u - r.xy / 2.0) / r.y - 0.3;
+
+    float4 o = float4(0);
+
+    for (; i < 32.0 && s > 0.001; i++) {
+        float4 term = float4(5, 2, 1, 0) / max(length(u - 0.1), 0.001);
+        o += min(term, float4(100.0));
+        p += float3(u * s, s);
+        s = 1.0 + p.y;
+        for (n = 0.01; n < 1.0; n += n) {
+            s += abs(dot(sin(p.z + U.iTime + p / n), float3(1))) * n * 0.1;
+        }
+    }
+    o = tanh(o / 5e2);
+    return o;
+}

data/shaders/water/water.gl.glsl

@@ -0,0 +1,44 @@
+// Name: Water
+// Author: diatribes
+// URL: https://www.shadertoy.com/view/tXjXDy
+#version 330 core
+precision highp float;
+
+out vec4 FragColor;
+in vec2 vUV;
+uniform vec2 iResolution;
+uniform float iTime;
+
+/*
+    -2 by @FabriceNeyret2
+
+    thanks!!  :D
+
+    If it doesn't display correctly, change line 17 "r/r" to "vec3(1)"
+*/
+
+void mainImage( out vec4 o, vec2 u ) {
+    float s=.002, i=0., n;  // FIXED: Initialize i=0
+    vec3 r = vec3(iResolution.xy, iResolution.x/iResolution.y);
+    vec3 p = vec3(0);
+    u = (u-r.xy/2.)/r.y-.3;
+
+    o = vec4(0);  // FIXED: Initialize output to black
+
+    for(; i < 32. && s > .001; i++) {
+        // Clamp only extreme overflow values, let normal brightness through
+        vec4 term = vec4(5,2,1,0)/max(length(u-.1), 0.001);
+        o += min(term, vec4(100.0));
+        for (p += vec3(u*s,s), s = 1. + p.y, n =.01; n < 1.; n+=n) {
+            s += abs(dot(sin(p.z+iTime+p / n), vec3(1))) * n*.1;
+        }
+    }
+    o = tanh(o/5e2);
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

data/shaders/water/water.vk.glsl

@@ -0,0 +1,45 @@
+#version 450
+
+// Name: Water
+// Author: diatribes
+// URL: https://www.shadertoy.com/view/tXjXDy
+
+layout(location = 0) in  vec2 vUV;
+layout(location = 0) out vec4 FragColor;
+
+layout(set = 3, binding = 0) uniform ShadertoyUBO {
+    float iTime;
+    vec2  iResolution;
+};
+
+/*
+    -2 by @FabriceNeyret2
+    thanks!!  :D
+
+    If it doesn't display correctly, change line "r/r" to "vec3(1)"
+*/
+
+void mainImage( out vec4 o, vec2 u ) {
+    float s=.002, i=0., n;
+    vec3 r = vec3(iResolution.xy, iResolution.x/iResolution.y);
+    vec3 p = vec3(0);
+    u = (u-r.xy/2.)/r.y-.3;
+
+    o = vec4(0);
+
+    for(; i < 32. && s > .001; i++) {
+        vec4 term = vec4(5,2,1,0)/max(length(u-.1), 0.001);
+        o += min(term, vec4(100.0));
+        for (p += vec3(u*s,s), s = 1. + p.y, n =.01; n < 1.; n+=n) {
+            s += abs(dot(sin(p.z+iTime+p / n), vec3(1))) * n*.1;
+        }
+    }
+    o = tanh(o/5e2);
+}
+
+void main() {
+    vec2 fragCoordPixels = vUV * iResolution;
+    vec4 outColor;
+    mainImage(outColor, fragCoordPixels);
+    FragColor = outColor;
+}

release/icons/create_icons.py

@@ -0,0 +1,150 @@
+#!/usr/bin/env python3
+
+import os
+import sys
+import shutil
+import subprocess
+from pathlib import Path
+
+
+def check_dependencies():
+    """Verifica que ImageMagick esté instalado"""
+    try:
+        subprocess.run(['magick', '--version'], capture_output=True, check=True)
+    except (subprocess.CalledProcessError, FileNotFoundError):
+        print("Error: ImageMagick no está instalado o no se encuentra en el PATH")
+        print("Instala ImageMagick desde: https://imagemagick.org/script/download.php")
+        sys.exit(1)
+    
+    # Verificar iconutil solo en macOS
+    if sys.platform == 'darwin':
+        try:
+            # iconutil no tiene --version, mejor usar which o probar con -h
+            result = subprocess.run(['which', 'iconutil'], capture_output=True, check=True)
+            if result.returncode == 0:
+                print("✓ iconutil disponible - se crearán archivos .ico e .icns")
+        except (subprocess.CalledProcessError, FileNotFoundError):
+            print("Error: iconutil no está disponible (solo funciona en macOS)")
+            print("Solo se creará el archivo .ico")
+    else:
+        print("ℹ️  Sistema no-macOS detectado - solo se creará archivo .ico")
+
+
+def create_icons(input_file):
+    """Crea archivos .icns e .ico a partir de un PNG"""
+    
+    # Verificar que el archivo existe
+    if not os.path.isfile(input_file):
+        print(f"Error: El archivo {input_file} no existe.")
+        return False
+    
+    # Obtener información del archivo
+    file_path = Path(input_file)
+    file_dir = file_path.parent
+    file_name = file_path.stem  # Nombre sin extensión
+    file_extension = file_path.suffix
+    
+    if file_extension.lower() not in ['.png', '.jpg', '.jpeg', '.bmp', '.tiff']:
+        print(f"Advertencia: {file_extension} puede no ser compatible. Se recomienda usar PNG.")
+    
+    # Crear archivo .ico usando el método simplificado
+    ico_output = file_dir / f"{file_name}.ico"
+    try:
+        print(f"Creando {ico_output}...")
+        subprocess.run([
+            'magick', str(input_file),
+            '-define', 'icon:auto-resize=256,128,64,48,32,16',
+            str(ico_output)
+        ], check=True)
+        print(f"✓ Archivo .ico creado: {ico_output}")
+    except subprocess.CalledProcessError as e:
+        print(f"Error creando archivo .ico: {e}")
+        return False
+    
+    # Crear archivo .icns (solo en macOS)
+    if sys.platform == 'darwin':
+        try:
+            # Crear carpeta temporal para iconset
+            temp_folder = file_dir / "icon.iconset"
+            
+            # Eliminar carpeta temporal si existe
+            if temp_folder.exists():
+                shutil.rmtree(temp_folder)
+            
+            # Crear carpeta temporal
+            temp_folder.mkdir(parents=True)
+            
+            # Definir los tamaños y nombres de archivo para .icns
+            icon_sizes = [
+                (16, "icon_16x16.png"),
+                (32, "icon_16x16@2x.png"),
+                (32, "icon_32x32.png"),
+                (64, "icon_32x32@2x.png"),
+                (128, "icon_128x128.png"),
+                (256, "icon_128x128@2x.png"),
+                (256, "icon_256x256.png"),
+                (512, "icon_256x256@2x.png"),
+                (512, "icon_512x512.png"),
+                (1024, "icon_512x512@2x.png")
+            ]
+            
+            print("Generando imágenes para .icns...")
+            # Crear cada tamaño de imagen
+            for size, output_name in icon_sizes:
+                output_path = temp_folder / output_name
+                subprocess.run([
+                    'magick', str(input_file),
+                    '-resize', f'{size}x{size}',
+                    str(output_path)
+                ], check=True)
+            
+            # Crear archivo .icns usando iconutil
+            icns_output = file_dir / f"{file_name}.icns"
+            print(f"Creando {icns_output}...")
+            subprocess.run([
+                'iconutil', '-c', 'icns',
+                str(temp_folder),
+                '-o', str(icns_output)
+            ], check=True)
+            
+            # Limpiar carpeta temporal
+            if temp_folder.exists():
+                shutil.rmtree(temp_folder)
+            
+            print(f"✓ Archivo .icns creado: {icns_output}")
+            
+        except subprocess.CalledProcessError as e:
+            print(f"Error creando archivo .icns: {e}")
+            # Limpiar carpeta temporal en caso de error
+            if temp_folder.exists():
+                shutil.rmtree(temp_folder)
+            return False
+    else:
+        print("ℹ️  Archivo .icns no creado (solo disponible en macOS)")
+    
+    return True
+
+
+def main():
+    """Función principal"""
+    # Verificar argumentos
+    if len(sys.argv) != 2:
+        print(f"Uso: {sys.argv[0]} ARCHIVO")
+        print("Ejemplo: python3 create_icons.py imagen.png")
+        sys.exit(0)
+    
+    input_file = sys.argv[1]
+    
+    # Verificar dependencias
+    check_dependencies()
+    
+    # Crear iconos
+    if create_icons(input_file):
+        print("\n✅ Proceso completado exitosamente")
+    else:
+        print("\n❌ El proceso falló")
+        sys.exit(1)
+
+
+if __name__ == "__main__":
+    main()

release/macos/Info.plist

@@ -0,0 +1,42 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+<plist version="1.0">
+<dict>
+	<key>CFBundleDevelopmentRegion</key>
+	<string>es</string>
+	<key>CFBundleDisplayName</key>
+	<string>Shadertoy</string>
+	<key>CFBundleExecutable</key>
+	<string>shadertoy</string>
+	<key>CFBundleIconFile</key>
+	<string>icon</string>
+	<key>CFBundleIconName</key>
+	<string>icon</string>
+	<key>CFBundleIdentifier</key>
+	<string>net.jailgames.shadertoy</string>
+	<key>CFBundleInfoDictionaryVersion</key>
+	<string>6.0</string>
+	<key>CFBundleName</key>
+	<string>shadertoy</string>
+	<key>CFBundlePackageType</key>
+	<string>APPL</string>
+	<key>CFBundleShortVersionString</key>
+	<string>1.00</string>
+	<key>CFBundleSignature</key>
+	<string>????</string>
+	<key>CFBundleVersion</key>
+	<string>1.00</string>
+	<key>CSResourcesFileMapped</key>
+	<true/>
+	<key>LSMinimumSystemVersion</key>
+	<string>12.0</string>
+	<key>NSHighResolutionCapable</key>
+	<true/>
+	<key>NSHumanReadableCopyright</key>
+	<string>Copyright 2025 JailDesigner</string>
+	<key>NSPrincipalClass</key>
+	<string>NSApplication</string>
+	<key>SUPublicDSAKeyFile</key>
+	<string>dsa_pub.pem</string>
+</dict>
+</plist>

release/macos/frameworks/SDL3.xcframework/Info.plist

@@ -0,0 +1,90 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+<plist version="1.0">
+<dict>
+	<key>AvailableLibraries</key>
+	<array>
+		<dict>
+			<key>BinaryPath</key>
+			<string>SDL3.framework/Versions/A/SDL3</string>
+			<key>LibraryIdentifier</key>
+			<string>macos-arm64_x86_64</string>
+			<key>LibraryPath</key>
+			<string>SDL3.framework</string>
+			<key>SupportedArchitectures</key>
+			<array>
+				<string>arm64</string>
+				<string>x86_64</string>
+			</array>
+			<key>SupportedPlatform</key>
+			<string>macos</string>
+		</dict>
+		<dict>
+			<key>BinaryPath</key>
+			<string>SDL3.framework/SDL3</string>
+			<key>LibraryIdentifier</key>
+			<string>tvos-arm64</string>
+			<key>LibraryPath</key>
+			<string>SDL3.framework</string>
+			<key>SupportedArchitectures</key>
+			<array>
+				<string>arm64</string>
+			</array>
+			<key>SupportedPlatform</key>
+			<string>tvos</string>
+		</dict>
+		<dict>
+			<key>BinaryPath</key>
+			<string>SDL3.framework/SDL3</string>
+			<key>LibraryIdentifier</key>
+			<string>ios-arm64</string>
+			<key>LibraryPath</key>
+			<string>SDL3.framework</string>
+			<key>SupportedArchitectures</key>
+			<array>
+				<string>arm64</string>
+			</array>
+			<key>SupportedPlatform</key>
+			<string>ios</string>
+		</dict>
+		<dict>
+			<key>BinaryPath</key>
+			<string>SDL3.framework/SDL3</string>
+			<key>LibraryIdentifier</key>
+			<string>ios-arm64_x86_64-simulator</string>
+			<key>LibraryPath</key>
+			<string>SDL3.framework</string>
+			<key>SupportedArchitectures</key>
+			<array>
+				<string>arm64</string>
+				<string>x86_64</string>
+			</array>
+			<key>SupportedPlatform</key>
+			<string>ios</string>
+			<key>SupportedPlatformVariant</key>
+			<string>simulator</string>
+		</dict>
+		<dict>
+			<key>BinaryPath</key>
+			<string>SDL3.framework/SDL3</string>
+			<key>LibraryIdentifier</key>
+			<string>tvos-arm64_x86_64-simulator</string>
+			<key>LibraryPath</key>
+			<string>SDL3.framework</string>
+			<key>SupportedArchitectures</key>
+			<array>
+				<string>arm64</string>
+				<string>x86_64</string>
+			</array>
+			<key>SupportedPlatform</key>
+			<string>tvos</string>
+			<key>SupportedPlatformVariant</key>
+			<string>simulator</string>
+		</dict>
+	</array>
+	<key>CFBundlePackageType</key>
+	<string>XFWK</string>
+	<key>XCFrameworkFormatVersion</key>
+	<string>1.0</string>
+</dict>
+</plist>

release/macos/frameworks/SDL3.xcframework/macos-arm64_x86_64/SDL3.framework/Headers

@@ -0,0 +1 @@
+Versions/Current/Headers

release/macos/frameworks/SDL3.xcframework/macos-arm64_x86_64/SDL3.framework/Resources

@@ -0,0 +1 @@
+Versions/Current/Resources

release/macos/frameworks/SDL3.xcframework/macos-arm64_x86_64/SDL3.framework/SDL3

@@ -0,0 +1 @@
+Versions/Current/SDL3

release/macos/frameworks/SDL3.xcframework/macos-arm64_x86_64/SDL3.framework/Versions/A/Headers/SDL.h

@@ -0,0 +1,90 @@
+/*
+  Simple DirectMedia Layer
+  Copyright (C) 1997-2025 Sam Lantinga <slouken@libsdl.org>
+
+  This software is provided 'as-is', without any express or implied
+  warranty.  In no event will the authors be held liable for any damages
+  arising from the use of this software.
+
+  Permission is granted to anyone to use this software for any purpose,
+  including commercial applications, and to alter it and redistribute it
+  freely, subject to the following restrictions:
+
+  1. The origin of this software must not be misrepresented; you must not
+     claim that you wrote the original software. If you use this software
+     in a product, an acknowledgment in the product documentation would be
+     appreciated but is not required.
+  2. Altered source versions must be plainly marked as such, and must not be
+     misrepresented as being the original software.
+  3. This notice may not be removed or altered from any source distribution.
+*/
+
+/**
+ * Main include header for the SDL library, version 3.2.20
+ *
+ * It is almost always best to include just this one header instead of
+ * picking out individual headers included here. There are exceptions to
+ * this rule--SDL_main.h is special and not included here--but usually
+ * letting SDL.h include the kitchen sink for you is the correct approach.
+ */
+
+#ifndef SDL_h_
+#define SDL_h_
+
+#include <SDL3/SDL_stdinc.h>
+#include <SDL3/SDL_assert.h>
+#include <SDL3/SDL_asyncio.h>
+#include <SDL3/SDL_atomic.h>
+#include <SDL3/SDL_audio.h>
+#include <SDL3/SDL_bits.h>
+#include <SDL3/SDL_blendmode.h>
+#include <SDL3/SDL_camera.h>
+#include <SDL3/SDL_clipboard.h>
+#include <SDL3/SDL_cpuinfo.h>
+#include <SDL3/SDL_dialog.h>
+#include <SDL3/SDL_endian.h>
+#include <SDL3/SDL_error.h>
+#include <SDL3/SDL_events.h>
+#include <SDL3/SDL_filesystem.h>
+#include <SDL3/SDL_gamepad.h>
+#include <SDL3/SDL_gpu.h>
+#include <SDL3/SDL_guid.h>
+#include <SDL3/SDL_haptic.h>
+#include <SDL3/SDL_hidapi.h>
+#include <SDL3/SDL_hints.h>
+#include <SDL3/SDL_init.h>
+#include <SDL3/SDL_iostream.h>
+#include <SDL3/SDL_joystick.h>
+#include <SDL3/SDL_keyboard.h>
+#include <SDL3/SDL_keycode.h>
+#include <SDL3/SDL_loadso.h>
+#include <SDL3/SDL_locale.h>
+#include <SDL3/SDL_log.h>
+#include <SDL3/SDL_messagebox.h>
+#include <SDL3/SDL_metal.h>
+#include <SDL3/SDL_misc.h>
+#include <SDL3/SDL_mouse.h>
+#include <SDL3/SDL_mutex.h>
+#include <SDL3/SDL_pen.h>
+#include <SDL3/SDL_pixels.h>
+#include <SDL3/SDL_platform.h>
+#include <SDL3/SDL_power.h>
+#include <SDL3/SDL_process.h>
+#include <SDL3/SDL_properties.h>
+#include <SDL3/SDL_rect.h>
+#include <SDL3/SDL_render.h>
+#include <SDL3/SDL_scancode.h>
+#include <SDL3/SDL_sensor.h>
+#include <SDL3/SDL_storage.h>
+#include <SDL3/SDL_surface.h>
+#include <SDL3/SDL_system.h>
+#include <SDL3/SDL_thread.h>
+#include <SDL3/SDL_time.h>
+#include <SDL3/SDL_timer.h>
+#include <SDL3/SDL_tray.h>
+#include <SDL3/SDL_touch.h>
+#include <SDL3/SDL_version.h>
+#include <SDL3/SDL_video.h>
+#include <SDL3/SDL_oldnames.h>
+
+#endif /* SDL_h_ */

release/macos/frameworks/SDL3.xcframework/macos-arm64_x86_64/SDL3.framework/Versions/A/Headers/SDL_assert.h

@@ -0,0 +1,662 @@
+/*
+  Simple DirectMedia Layer
+  Copyright (C) 1997-2025 Sam Lantinga <slouken@libsdl.org>
+
+  This software is provided 'as-is', without any express or implied
+  warranty.  In no event will the authors be held liable for any damages
+  arising from the use of this software.
+
+  Permission is granted to anyone to use this software for any purpose,
+  including commercial applications, and to alter it and redistribute it
+  freely, subject to the following restrictions:
+
+  1. The origin of this software must not be misrepresented; you must not
+     claim that you wrote the original software. If you use this software
+     in a product, an acknowledgment in the product documentation would be
+     appreciated but is not required.
+  2. Altered source versions must be plainly marked as such, and must not be
+     misrepresented as being the original software.
+  3. This notice may not be removed or altered from any source distribution.
+*/
+
+/**
+ * # CategoryAssert
+ *
+ * A helpful assertion macro!
+ *
+ * SDL assertions operate like your usual `assert` macro, but with some added
+ * features:
+ *
+ * - It uses a trick with the `sizeof` operator, so disabled assertions
+ *   vaporize out of the compiled code, but variables only referenced in the
+ *   assertion won't trigger compiler warnings about being unused.
+ * - It is safe to use with a dangling-else: `if (x) SDL_assert(y); else
+ *   do_something();`
+ * - It works the same everywhere, instead of counting on various platforms'
+ *   compiler and C runtime to behave.
+ * - It provides multiple levels of assertion (SDL_assert, SDL_assert_release,
+ *   SDL_assert_paranoid) instead of a single all-or-nothing option.
+ * - It offers a variety of responses when an assertion fails (retry, trigger
+ *   the debugger, abort the program, ignore the failure once, ignore it for
+ *   the rest of the program's run).
+ * - It tries to show the user a dialog by default, if possible, but the app
+ *   can provide a callback to handle assertion failures however they like.
+ * - It lets failed assertions be retried. Perhaps you had a network failure
+ *   and just want to retry the test after plugging your network cable back
+ *   in? You can.
+ * - It lets the user ignore an assertion failure, if there's a harmless
+ *   problem that one can continue past.
+ * - It lets the user mark an assertion as ignored for the rest of the
+ *   program's run; if there's a harmless problem that keeps popping up.
+ * - It provides statistics and data on all failed assertions to the app.
+ * - It allows the default assertion handler to be controlled with environment
+ *   variables, in case an automated script needs to control it.
+ * - It can be used as an aid to Clang's static analysis; it will treat SDL
+ *   assertions as universally true (under the assumption that you are serious
+ *   about the asserted claims and that your debug builds will detect when
+ *   these claims were wrong). This can help the analyzer avoid false
+ *   positives.
+ *
+ * To use it: compile a debug build and just sprinkle around tests to check
+ * your code!
+ */
+
+#ifndef SDL_assert_h_
+#define SDL_assert_h_
+
+#include <SDL3/SDL_stdinc.h>
+
+#include <SDL3/SDL_begin_code.h>
+/* Set up for C function definitions, even when using C++ */
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#ifdef SDL_WIKI_DOCUMENTATION_SECTION
+
+/**
+ * The level of assertion aggressiveness.
+ *
+ * This value changes depending on compiler options and other preprocessor
+ * defines.
+ *
+ * It is currently one of the following values, but future SDL releases might
+ * add more:
+ *
+ * - 0: All SDL assertion macros are disabled.
+ * - 1: Release settings: SDL_assert disabled, SDL_assert_release enabled.
+ * - 2: Debug settings: SDL_assert and SDL_assert_release enabled.
+ * - 3: Paranoid settings: All SDL assertion macros enabled, including
+ *   SDL_assert_paranoid.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_ASSERT_LEVEL SomeNumberBasedOnVariousFactors
+
+#elif !defined(SDL_ASSERT_LEVEL)
+#ifdef SDL_DEFAULT_ASSERT_LEVEL
+#define SDL_ASSERT_LEVEL SDL_DEFAULT_ASSERT_LEVEL
+#elif defined(_DEBUG) || defined(DEBUG) || \
+      (defined(__GNUC__) && !defined(__OPTIMIZE__))
+#define SDL_ASSERT_LEVEL 2
+#else
+#define SDL_ASSERT_LEVEL 1
+#endif
+#endif
+
+#ifdef SDL_WIKI_DOCUMENTATION_SECTION
+
+/**
+ * Attempt to tell an attached debugger to pause.
+ *
+ * This allows an app to programmatically halt ("break") the debugger as if it
+ * had hit a breakpoint, allowing the developer to examine program state, etc.
+ *
+ * This is a macro--not a function--so that the debugger breaks on the source
+ * code line that used SDL_TriggerBreakpoint and not in some random guts of
+ * SDL. SDL_assert uses this macro for the same reason.
+ *
+ * If the program is not running under a debugger, SDL_TriggerBreakpoint will
+ * likely terminate the app, possibly without warning. If the current platform
+ * isn't supported, this macro is left undefined.
+ *
+ * \threadsafety It is safe to call this macro from any thread.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_TriggerBreakpoint() TriggerABreakpointInAPlatformSpecificManner
+
+#elif defined(_MSC_VER) && _MSC_VER >= 1310
+    /* Don't include intrin.h here because it contains C++ code */
+    extern void __cdecl __debugbreak(void);
+    #define SDL_TriggerBreakpoint() __debugbreak()
+#elif defined(_MSC_VER) && defined(_M_IX86)
+    #define SDL_TriggerBreakpoint() { _asm { int 0x03 }  }
+#elif defined(ANDROID)
+    #include <assert.h>
+    #define SDL_TriggerBreakpoint() assert(0)
+#elif SDL_HAS_BUILTIN(__builtin_debugtrap)
+    #define SDL_TriggerBreakpoint() __builtin_debugtrap()
+#elif SDL_HAS_BUILTIN(__builtin_trap)
+    #define SDL_TriggerBreakpoint() __builtin_trap()
+#elif (defined(__GNUC__) || defined(__clang__)) && (defined(__i386__) || defined(__x86_64__))
+    #define SDL_TriggerBreakpoint() __asm__ __volatile__ ( "int $3\n\t" )
+#elif (defined(__GNUC__) || defined(__clang__)) && defined(__riscv)
+    #define SDL_TriggerBreakpoint() __asm__ __volatile__ ( "ebreak\n\t" )
+#elif ( defined(SDL_PLATFORM_APPLE) && (defined(__arm64__) || defined(__aarch64__)) )  /* this might work on other ARM targets, but this is a known quantity... */
+    #define SDL_TriggerBreakpoint() __asm__ __volatile__ ( "brk #22\n\t" )
+#elif defined(SDL_PLATFORM_APPLE) && defined(__arm__)
+    #define SDL_TriggerBreakpoint() __asm__ __volatile__ ( "bkpt #22\n\t" )
+#elif defined(_WIN32) && ((defined(__GNUC__) || defined(__clang__)) && (defined(__arm64__) || defined(__aarch64__)) )
+    #define SDL_TriggerBreakpoint() __asm__ __volatile__ ( "brk #0xF000\n\t" )
+#elif defined(__GNUC__) || defined(__clang__)
+    #define SDL_TriggerBreakpoint() __builtin_trap()  /* older gcc may not support SDL_HAS_BUILTIN(__builtin_trap) above */
+#elif defined(__386__) && defined(__WATCOMC__)
+    #define SDL_TriggerBreakpoint() { _asm { int 0x03 } }
+#elif defined(HAVE_SIGNAL_H) && !defined(__WATCOMC__)
+    #include <signal.h>
+    #define SDL_TriggerBreakpoint() raise(SIGTRAP)
+#else
+    /* SDL_TriggerBreakpoint is intentionally left undefined on unknown platforms. */
+#endif
+
+#ifdef SDL_WIKI_DOCUMENTATION_SECTION
+
+/**
+ * A macro that reports the current function being compiled.
+ *
+ * If SDL can't figure how the compiler reports this, it will use "???".
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_FUNCTION __FUNCTION__
+
+#elif defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 supports __func__ as a standard. */
+#   define SDL_FUNCTION __func__
+#elif ((defined(__GNUC__) && (__GNUC__ >= 2)) || defined(_MSC_VER) || defined (__WATCOMC__))
+#   define SDL_FUNCTION __FUNCTION__
+#else
+#   define SDL_FUNCTION "???"
+#endif
+
+/**
+ * A macro that reports the current file being compiled.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_FILE    __FILE__
+
+/**
+ * A macro that reports the current line number of the file being compiled.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_LINE    __LINE__
+
+/*
+sizeof (x) makes the compiler still parse the expression even without
+assertions enabled, so the code is always checked at compile time, but
+doesn't actually generate code for it, so there are no side effects or
+expensive checks at run time, just the constant size of what x WOULD be,
+which presumably gets optimized out as unused.
+This also solves the problem of...
+
+    int somevalue = blah();
+    SDL_assert(somevalue == 1);
+
+...which would cause compiles to complain that somevalue is unused if we
+disable assertions.
+*/
+
+#ifdef SDL_WIKI_DOCUMENTATION_SECTION
+
+/**
+ * A macro for wrapping code in `do {} while (0);` without compiler warnings.
+ *
+ * Visual Studio with really aggressive warnings enabled needs this to avoid
+ * compiler complaints.
+ *
+ * the `do {} while (0);` trick is useful for wrapping code in a macro that
+ * may or may not be a single statement, to avoid various C language
+ * accidents.
+ *
+ * To use:
+ *
+ * ```c
+ * do { SomethingOnce(); } while (SDL_NULL_WHILE_LOOP_CONDITION (0));
+ * ```
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_NULL_WHILE_LOOP_CONDITION (0)
+
+#elif defined(_MSC_VER)  /* Avoid /W4 warnings. */
+/* "while (0,0)" fools Microsoft's compiler's /W4 warning level into thinking
+    this condition isn't constant. And looks like an owl's face! */
+#define SDL_NULL_WHILE_LOOP_CONDITION (0,0)
+#else
+#define SDL_NULL_WHILE_LOOP_CONDITION (0)
+#endif
+
+/**
+ * The macro used when an assertion is disabled.
+ *
+ * This isn't for direct use by apps, but this is the code that is inserted
+ * when an SDL_assert is disabled (perhaps in a release build).
+ *
+ * The code does nothing, but wraps `condition` in a sizeof operator, which
+ * generates no code and has no side effects, but avoid compiler warnings
+ * about unused variables.
+ *
+ * \param condition the condition to assert (but not actually run here).
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_disabled_assert(condition) \
+    do { (void) sizeof ((condition)); } while (SDL_NULL_WHILE_LOOP_CONDITION)
+
+/**
+ * Possible outcomes from a triggered assertion.
+ *
+ * When an enabled assertion triggers, it may call the assertion handler
+ * (possibly one provided by the app via SDL_SetAssertionHandler), which will
+ * return one of these values, possibly after asking the user.
+ *
+ * Then SDL will respond based on this outcome (loop around to retry the
+ * condition, try to break in a debugger, kill the program, or ignore the
+ * problem).
+ *
+ * \since This enum is available since SDL 3.2.0.
+ */
+typedef enum SDL_AssertState
+{
+    SDL_ASSERTION_RETRY,  /**< Retry the assert immediately. */
+    SDL_ASSERTION_BREAK,  /**< Make the debugger trigger a breakpoint. */
+    SDL_ASSERTION_ABORT,  /**< Terminate the program. */
+    SDL_ASSERTION_IGNORE,  /**< Ignore the assert. */
+    SDL_ASSERTION_ALWAYS_IGNORE  /**< Ignore the assert from now on. */
+} SDL_AssertState;
+
+/**
+ * Information about an assertion failure.
+ *
+ * This structure is filled in with information about a triggered assertion,
+ * used by the assertion handler, then added to the assertion report. This is
+ * returned as a linked list from SDL_GetAssertionReport().
+ *
+ * \since This struct is available since SDL 3.2.0.
+ */
+typedef struct SDL_AssertData
+{
+    bool always_ignore;  /**< true if app should always continue when assertion is triggered. */
+    unsigned int trigger_count; /**< Number of times this assertion has been triggered. */
+    const char *condition;  /**< A string of this assert's test code. */
+    const char *filename;  /**< The source file where this assert lives. */
+    int linenum;  /**< The line in `filename` where this assert lives. */
+    const char *function;  /**< The name of the function where this assert lives. */
+    const struct SDL_AssertData *next;  /**< next item in the linked list. */
+} SDL_AssertData;
+
+/**
+ * Never call this directly.
+ *
+ * Use the SDL_assert macros instead.
+ *
+ * \param data assert data structure.
+ * \param func function name.
+ * \param file file name.
+ * \param line line number.
+ * \returns assert state.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ */
+extern SDL_DECLSPEC SDL_AssertState SDLCALL SDL_ReportAssertion(SDL_AssertData *data,
+                                                            const char *func,
+                                                            const char *file, int line) SDL_ANALYZER_NORETURN;
+
+
+#ifdef SDL_WIKI_DOCUMENTATION_SECTION
+
+/**
+ * The macro used when an assertion triggers a breakpoint.
+ *
+ * This isn't for direct use by apps; use SDL_assert or SDL_TriggerBreakpoint
+ * instead.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_AssertBreakpoint() SDL_TriggerBreakpoint()
+
+#elif !defined(SDL_AssertBreakpoint)
+#  if defined(ANDROID) && defined(assert)
+     /* Define this as empty in case assert() is defined as SDL_assert */
+#    define SDL_AssertBreakpoint()
+#  else
+#    define SDL_AssertBreakpoint() SDL_TriggerBreakpoint()
+#  endif
+#endif /* !SDL_AssertBreakpoint */
+
+/**
+ * The macro used when an assertion is enabled.
+ *
+ * This isn't for direct use by apps, but this is the code that is inserted
+ * when an SDL_assert is enabled.
+ *
+ * The `do {} while(0)` avoids dangling else problems:
+ *
+ * ```c
+ * if (x) SDL_assert(y); else blah();
+ * ```
+ *
+ * ... without the do/while, the "else" could attach to this macro's "if". We
+ * try to handle just the minimum we need here in a macro...the loop, the
+ * static vars, and break points. The heavy lifting is handled in
+ * SDL_ReportAssertion().
+ *
+ * \param condition the condition to assert.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_enabled_assert(condition) \
+    do { \
+        while ( !(condition) ) { \
+            static struct SDL_AssertData sdl_assert_data = { false, 0, #condition, NULL, 0, NULL, NULL }; \
+            const SDL_AssertState sdl_assert_state = SDL_ReportAssertion(&sdl_assert_data, SDL_FUNCTION, SDL_FILE, SDL_LINE); \
+            if (sdl_assert_state == SDL_ASSERTION_RETRY) { \
+                continue; /* go again. */ \
+            } else if (sdl_assert_state == SDL_ASSERTION_BREAK) { \
+                SDL_AssertBreakpoint(); \
+            } \
+            break; /* not retrying. */ \
+        } \
+    } while (SDL_NULL_WHILE_LOOP_CONDITION)
+
+#ifdef SDL_WIKI_DOCUMENTATION_SECTION
+
+/**
+ * An assertion test that is normally performed only in debug builds.
+ *
+ * This macro is enabled when the SDL_ASSERT_LEVEL is >= 2, otherwise it is
+ * disabled. This is meant to only do these tests in debug builds, so they can
+ * tend to be more expensive, and they are meant to bring everything to a halt
+ * when they fail, with the programmer there to assess the problem.
+ *
+ * In short: you can sprinkle these around liberally and assume they will
+ * evaporate out of the build when building for end-users.
+ *
+ * When assertions are disabled, this wraps `condition` in a `sizeof`
+ * operator, which means any function calls and side effects will not run, but
+ * the compiler will not complain about any otherwise-unused variables that
+ * are only referenced in the assertion.
+ *
+ * One can set the environment variable "SDL_ASSERT" to one of several strings
+ * ("abort", "break", "retry", "ignore", "always_ignore") to force a default
+ * behavior, which may be desirable for automation purposes. If your platform
+ * requires GUI interfaces to happen on the main thread but you're debugging
+ * an assertion in a background thread, it might be desirable to set this to
+ * "break" so that your debugger takes control as soon as assert is triggered,
+ * instead of risking a bad UI interaction (deadlock, etc) in the application.
+ *
+ * \param condition boolean value to test.
+ *
+ * \threadsafety It is safe to call this macro from any thread.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_assert(condition) if (assertion_enabled && (condition)) { trigger_assertion; }
+
+/**
+ * An assertion test that is performed even in release builds.
+ *
+ * This macro is enabled when the SDL_ASSERT_LEVEL is >= 1, otherwise it is
+ * disabled. This is meant to be for tests that are cheap to make and
+ * extremely unlikely to fail; generally it is frowned upon to have an
+ * assertion failure in a release build, so these assertions generally need to
+ * be of more than life-and-death importance if there's a chance they might
+ * trigger. You should almost always consider handling these cases more
+ * gracefully than an assert allows.
+ *
+ * When assertions are disabled, this wraps `condition` in a `sizeof`
+ * operator, which means any function calls and side effects will not run, but
+ * the compiler will not complain about any otherwise-unused variables that
+ * are only referenced in the assertion.
+ *
+ * One can set the environment variable "SDL_ASSERT" to one of several strings
+ * ("abort", "break", "retry", "ignore", "always_ignore") to force a default
+ * behavior, which may be desirable for automation purposes. If your platform
+ * requires GUI interfaces to happen on the main thread but you're debugging
+ * an assertion in a background thread, it might be desirable to set this to
+ * "break" so that your debugger takes control as soon as assert is triggered,
+ * instead of risking a bad UI interaction (deadlock, etc) in the application.
+ * *
+ *
+ * \param condition boolean value to test.
+ *
+ * \threadsafety It is safe to call this macro from any thread.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_assert_release(condition) SDL_disabled_assert(condition)
+
+/**
+ * An assertion test that is performed only when built with paranoid settings.
+ *
+ * This macro is enabled when the SDL_ASSERT_LEVEL is >= 3, otherwise it is
+ * disabled. This is a higher level than both release and debug, so these
+ * tests are meant to be expensive and only run when specifically looking for
+ * extremely unexpected failure cases in a special build.
+ *
+ * When assertions are disabled, this wraps `condition` in a `sizeof`
+ * operator, which means any function calls and side effects will not run, but
+ * the compiler will not complain about any otherwise-unused variables that
+ * are only referenced in the assertion.
+ *
+ * One can set the environment variable "SDL_ASSERT" to one of several strings
+ * ("abort", "break", "retry", "ignore", "always_ignore") to force a default
+ * behavior, which may be desirable for automation purposes. If your platform
+ * requires GUI interfaces to happen on the main thread but you're debugging
+ * an assertion in a background thread, it might be desirable to set this to
+ * "break" so that your debugger takes control as soon as assert is triggered,
+ * instead of risking a bad UI interaction (deadlock, etc) in the application.
+ *
+ * \param condition boolean value to test.
+ *
+ * \threadsafety It is safe to call this macro from any thread.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_assert_paranoid(condition) SDL_disabled_assert(condition)
+
+/* Enable various levels of assertions. */
+#elif SDL_ASSERT_LEVEL == 0   /* assertions disabled */
+#   define SDL_assert(condition) SDL_disabled_assert(condition)
+#   define SDL_assert_release(condition) SDL_disabled_assert(condition)
+#   define SDL_assert_paranoid(condition) SDL_disabled_assert(condition)
+#elif SDL_ASSERT_LEVEL == 1  /* release settings. */
+#   define SDL_assert(condition) SDL_disabled_assert(condition)
+#   define SDL_assert_release(condition) SDL_enabled_assert(condition)
+#   define SDL_assert_paranoid(condition) SDL_disabled_assert(condition)
+#elif SDL_ASSERT_LEVEL == 2  /* debug settings. */
+#   define SDL_assert(condition) SDL_enabled_assert(condition)
+#   define SDL_assert_release(condition) SDL_enabled_assert(condition)
+#   define SDL_assert_paranoid(condition) SDL_disabled_assert(condition)
+#elif SDL_ASSERT_LEVEL == 3  /* paranoid settings. */
+#   define SDL_assert(condition) SDL_enabled_assert(condition)
+#   define SDL_assert_release(condition) SDL_enabled_assert(condition)
+#   define SDL_assert_paranoid(condition) SDL_enabled_assert(condition)
+#else
+#   error Unknown assertion level.
+#endif
+
+/**
+ * An assertion test that is always performed.
+ *
+ * This macro is always enabled no matter what SDL_ASSERT_LEVEL is set to. You
+ * almost never want to use this, as it could trigger on an end-user's system,
+ * crashing your program.
+ *
+ * One can set the environment variable "SDL_ASSERT" to one of several strings
+ * ("abort", "break", "retry", "ignore", "always_ignore") to force a default
+ * behavior, which may be desirable for automation purposes. If your platform
+ * requires GUI interfaces to happen on the main thread but you're debugging
+ * an assertion in a background thread, it might be desirable to set this to
+ * "break" so that your debugger takes control as soon as assert is triggered,
+ * instead of risking a bad UI interaction (deadlock, etc) in the application.
+ *
+ * \param condition boolean value to test.
+ *
+ * \threadsafety It is safe to call this macro from any thread.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_assert_always(condition) SDL_enabled_assert(condition)
+
+
+/**
+ * A callback that fires when an SDL assertion fails.
+ *
+ * \param data a pointer to the SDL_AssertData structure corresponding to the
+ *             current assertion.
+ * \param userdata what was passed as `userdata` to SDL_SetAssertionHandler().
+ * \returns an SDL_AssertState value indicating how to handle the failure.
+ *
+ * \threadsafety This callback may be called from any thread that triggers an
+ *               assert at any time.
+ *
+ * \since This datatype is available since SDL 3.2.0.
+ */
+typedef SDL_AssertState (SDLCALL *SDL_AssertionHandler)(
+                                 const SDL_AssertData *data, void *userdata);
+
+/**
+ * Set an application-defined assertion handler.
+ *
+ * This function allows an application to show its own assertion UI and/or
+ * force the response to an assertion failure. If the application doesn't
+ * provide this, SDL will try to do the right thing, popping up a
+ * system-specific GUI dialog, and probably minimizing any fullscreen windows.
+ *
+ * This callback may fire from any thread, but it runs wrapped in a mutex, so
+ * it will only fire from one thread at a time.
+ *
+ * This callback is NOT reset to SDL's internal handler upon SDL_Quit()!
+ *
+ * \param handler the SDL_AssertionHandler function to call when an assertion
+ *                fails or NULL for the default handler.
+ * \param userdata a pointer that is passed to `handler`.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetAssertionHandler
+ */
+extern SDL_DECLSPEC void SDLCALL SDL_SetAssertionHandler(
+                                            SDL_AssertionHandler handler,
+                                            void *userdata);
+
+/**
+ * Get the default assertion handler.
+ *
+ * This returns the function pointer that is called by default when an
+ * assertion is triggered. This is an internal function provided by SDL, that
+ * is used for assertions when SDL_SetAssertionHandler() hasn't been used to
+ * provide a different function.
+ *
+ * \returns the default SDL_AssertionHandler that is called when an assert
+ *          triggers.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetAssertionHandler
+ */
+extern SDL_DECLSPEC SDL_AssertionHandler SDLCALL SDL_GetDefaultAssertionHandler(void);
+
+/**
+ * Get the current assertion handler.
+ *
+ * This returns the function pointer that is called when an assertion is
+ * triggered. This is either the value last passed to
+ * SDL_SetAssertionHandler(), or if no application-specified function is set,
+ * is equivalent to calling SDL_GetDefaultAssertionHandler().
+ *
+ * The parameter `puserdata` is a pointer to a void*, which will store the
+ * "userdata" pointer that was passed to SDL_SetAssertionHandler(). This value
+ * will always be NULL for the default handler. If you don't care about this
+ * data, it is safe to pass a NULL pointer to this function to ignore it.
+ *
+ * \param puserdata pointer which is filled with the "userdata" pointer that
+ *                  was passed to SDL_SetAssertionHandler().
+ * \returns the SDL_AssertionHandler that is called when an assert triggers.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_SetAssertionHandler
+ */
+extern SDL_DECLSPEC SDL_AssertionHandler SDLCALL SDL_GetAssertionHandler(void **puserdata);
+
+/**
+ * Get a list of all assertion failures.
+ *
+ * This function gets all assertions triggered since the last call to
+ * SDL_ResetAssertionReport(), or the start of the program.
+ *
+ * The proper way to examine this data looks something like this:
+ *
+ * ```c
+ * const SDL_AssertData *item = SDL_GetAssertionReport();
+ * while (item) {
+ *    printf("'%s', %s (%s:%d), triggered %u times, always ignore: %s.\\n",
+ *           item->condition, item->function, item->filename,
+ *           item->linenum, item->trigger_count,
+ *           item->always_ignore ? "yes" : "no");
+ *    item = item->next;
+ * }
+ * ```
+ *
+ * \returns a list of all failed assertions or NULL if the list is empty. This
+ *          memory should not be modified or freed by the application. This
+ *          pointer remains valid until the next call to SDL_Quit() or
+ *          SDL_ResetAssertionReport().
+ *
+ * \threadsafety This function is not thread safe. Other threads calling
+ *               SDL_ResetAssertionReport() simultaneously, may render the
+ *               returned pointer invalid.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_ResetAssertionReport
+ */
+extern SDL_DECLSPEC const SDL_AssertData * SDLCALL SDL_GetAssertionReport(void);
+
+/**
+ * Clear the list of all assertion failures.
+ *
+ * This function will clear the list of all assertions triggered up to that
+ * point. Immediately following this call, SDL_GetAssertionReport will return
+ * no items. In addition, any previously-triggered assertions will be reset to
+ * a trigger_count of zero, and their always_ignore state will be false.
+ *
+ * \threadsafety This function is not thread safe. Other threads triggering an
+ *               assertion, or simultaneously calling this function may cause
+ *               memory leaks or crashes.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetAssertionReport
+ */
+extern SDL_DECLSPEC void SDLCALL SDL_ResetAssertionReport(void);
+
+/* Ends C function definitions when using C++ */
+#ifdef __cplusplus
+}
+#endif
+#include <SDL3/SDL_close_code.h>
+
+#endif /* SDL_assert_h_ */

release/macos/frameworks/SDL3.xcframework/macos-arm64_x86_64/SDL3.framework/Versions/A/Headers/SDL_asyncio.h

@@ -0,0 +1,546 @@
+/*
+  Simple DirectMedia Layer
+  Copyright (C) 1997-2025 Sam Lantinga <slouken@libsdl.org>
+
+  This software is provided 'as-is', without any express or implied
+  warranty.  In no event will the authors be held liable for any damages
+  arising from the use of this software.
+
+  Permission is granted to anyone to use this software for any purpose,
+  including commercial applications, and to alter it and redistribute it
+  freely, subject to the following restrictions:
+
+  1. The origin of this software must not be misrepresented; you must not
+     claim that you wrote the original software. If you use this software
+     in a product, an acknowledgment in the product documentation would be
+     appreciated but is not required.
+  2. Altered source versions must be plainly marked as such, and must not be
+     misrepresented as being the original software.
+  3. This notice may not be removed or altered from any source distribution.
+*/
+
+/* WIKI CATEGORY: AsyncIO */
+
+/**
+ * # CategoryAsyncIO
+ *
+ * SDL offers a way to perform I/O asynchronously. This allows an app to read
+ * or write files without waiting for data to actually transfer; the functions
+ * that request I/O never block while the request is fulfilled.
+ *
+ * Instead, the data moves in the background and the app can check for results
+ * at their leisure.
+ *
+ * This is more complicated than just reading and writing files in a
+ * synchronous way, but it can allow for more efficiency, and never having
+ * framerate drops as the hard drive catches up, etc.
+ *
+ * The general usage pattern for async I/O is:
+ *
+ * - Create one or more SDL_AsyncIOQueue objects.
+ * - Open files with SDL_AsyncIOFromFile.
+ * - Start I/O tasks to the files with SDL_ReadAsyncIO or SDL_WriteAsyncIO,
+ *   putting those tasks into one of the queues.
+ * - Later on, use SDL_GetAsyncIOResult on a queue to see if any task is
+ *   finished without blocking. Tasks might finish in any order with success
+ *   or failure.
+ * - When all your tasks are done, close the file with SDL_CloseAsyncIO. This
+ *   also generates a task, since it might flush data to disk!
+ *
+ * This all works, without blocking, in a single thread, but one can also wait
+ * on a queue in a background thread, sleeping until new results have arrived:
+ *
+ * - Call SDL_WaitAsyncIOResult from one or more threads to efficiently block
+ *   until new tasks complete.
+ * - When shutting down, call SDL_SignalAsyncIOQueue to unblock any sleeping
+ *   threads despite there being no new tasks completed.
+ *
+ * And, of course, to match the synchronous SDL_LoadFile, we offer
+ * SDL_LoadFileAsync as a convenience function. This will handle allocating a
+ * buffer, slurping in the file data, and null-terminating it; you still check
+ * for results later.
+ *
+ * Behind the scenes, SDL will use newer, efficient APIs on platforms that
+ * support them: Linux's io_uring and Windows 11's IoRing, for example. If
+ * those technologies aren't available, SDL will offload the work to a thread
+ * pool that will manage otherwise-synchronous loads without blocking the app.
+ *
+ * ## Best Practices
+ *
+ * Simple non-blocking I/O--for an app that just wants to pick up data
+ * whenever it's ready without losing framerate waiting on disks to spin--can
+ * use whatever pattern works well for the program. In this case, simply call
+ * SDL_ReadAsyncIO, or maybe SDL_LoadFileAsync, as needed. Once a frame, call
+ * SDL_GetAsyncIOResult to check for any completed tasks and deal with the
+ * data as it arrives.
+ *
+ * If two separate pieces of the same program need their own I/O, it is legal
+ * for each to create their own queue. This will prevent either piece from
+ * accidentally consuming the other's completed tasks. Each queue does require
+ * some amount of resources, but it is not an overwhelming cost. Do not make a
+ * queue for each task, however. It is better to put many tasks into a single
+ * queue. They will be reported in order of completion, not in the order they
+ * were submitted, so it doesn't generally matter what order tasks are
+ * started.
+ *
+ * One async I/O queue can be shared by multiple threads, or one thread can
+ * have more than one queue, but the most efficient way--if ruthless
+ * efficiency is the goal--is to have one queue per thread, with multiple
+ * threads working in parallel, and attempt to keep each queue loaded with
+ * tasks that are both started by and consumed by the same thread. On modern
+ * platforms that can use newer interfaces, this can keep data flowing as
+ * efficiently as possible all the way from storage hardware to the app, with
+ * no contention between threads for access to the same queue.
+ *
+ * Written data is not guaranteed to make it to physical media by the time a
+ * closing task is completed, unless SDL_CloseAsyncIO is called with its
+ * `flush` parameter set to true, which is to say that a successful result
+ * here can still result in lost data during an unfortunately-timed power
+ * outage if not flushed. However, flushing will take longer and may be
+ * unnecessary, depending on the app's needs.
+ */
+
+#ifndef SDL_asyncio_h_
+#define SDL_asyncio_h_
+
+#include <SDL3/SDL_stdinc.h>
+
+#include <SDL3/SDL_begin_code.h>
+/* Set up for C function definitions, even when using C++ */
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/**
+ * The asynchronous I/O operation structure.
+ *
+ * This operates as an opaque handle. One can then request read or write
+ * operations on it.
+ *
+ * \since This struct is available since SDL 3.2.0.
+ *
+ * \sa SDL_AsyncIOFromFile
+ */
+typedef struct SDL_AsyncIO SDL_AsyncIO;
+
+/**
+ * Types of asynchronous I/O tasks.
+ *
+ * \since This enum is available since SDL 3.2.0.
+ */
+typedef enum SDL_AsyncIOTaskType
+{
+    SDL_ASYNCIO_TASK_READ,   /**< A read operation. */
+    SDL_ASYNCIO_TASK_WRITE,  /**< A write operation. */
+    SDL_ASYNCIO_TASK_CLOSE   /**< A close operation. */
+} SDL_AsyncIOTaskType;
+
+/**
+ * Possible outcomes of an asynchronous I/O task.
+ *
+ * \since This enum is available since SDL 3.2.0.
+ */
+typedef enum SDL_AsyncIOResult
+{
+    SDL_ASYNCIO_COMPLETE,  /**< request was completed without error */
+    SDL_ASYNCIO_FAILURE,   /**< request failed for some reason; check SDL_GetError()! */
+    SDL_ASYNCIO_CANCELED   /**< request was canceled before completing. */
+} SDL_AsyncIOResult;
+
+/**
+ * Information about a completed asynchronous I/O request.
+ *
+ * \since This struct is available since SDL 3.2.0.
+ */
+typedef struct SDL_AsyncIOOutcome
+{
+    SDL_AsyncIO *asyncio;   /**< what generated this task. This pointer will be invalid if it was closed! */
+    SDL_AsyncIOTaskType type;  /**< What sort of task was this? Read, write, etc? */
+    SDL_AsyncIOResult result;  /**< the result of the work (success, failure, cancellation). */
+    void *buffer;  /**< buffer where data was read/written. */
+    Uint64 offset;  /**< offset in the SDL_AsyncIO where data was read/written. */
+    Uint64 bytes_requested;  /**< number of bytes the task was to read/write. */
+    Uint64 bytes_transferred;  /**< actual number of bytes that were read/written. */
+    void *userdata;    /**< pointer provided by the app when starting the task */
+} SDL_AsyncIOOutcome;
+
+/**
+ * A queue of completed asynchronous I/O tasks.
+ *
+ * When starting an asynchronous operation, you specify a queue for the new
+ * task. A queue can be asked later if any tasks in it have completed,
+ * allowing an app to manage multiple pending tasks in one place, in whatever
+ * order they complete.
+ *
+ * \since This struct is available since SDL 3.2.0.
+ *
+ * \sa SDL_CreateAsyncIOQueue
+ * \sa SDL_ReadAsyncIO
+ * \sa SDL_WriteAsyncIO
+ * \sa SDL_GetAsyncIOResult
+ * \sa SDL_WaitAsyncIOResult
+ */
+typedef struct SDL_AsyncIOQueue SDL_AsyncIOQueue;
+
+/**
+ * Use this function to create a new SDL_AsyncIO object for reading from
+ * and/or writing to a named file.
+ *
+ * The `mode` string understands the following values:
+ *
+ * - "r": Open a file for reading only. It must exist.
+ * - "w": Open a file for writing only. It will create missing files or
+ *   truncate existing ones.
+ * - "r+": Open a file for update both reading and writing. The file must
+ *   exist.
+ * - "w+": Create an empty file for both reading and writing. If a file with
+ *   the same name already exists its content is erased and the file is
+ *   treated as a new empty file.
+ *
+ * There is no "b" mode, as there is only "binary" style I/O, and no "a" mode
+ * for appending, since you specify the position when starting a task.
+ *
+ * This function supports Unicode filenames, but they must be encoded in UTF-8
+ * format, regardless of the underlying operating system.
+ *
+ * This call is _not_ asynchronous; it will open the file before returning,
+ * under the assumption that doing so is generally a fast operation. Future
+ * reads and writes to the opened file will be async, however.
+ *
+ * \param file a UTF-8 string representing the filename to open.
+ * \param mode an ASCII string representing the mode to be used for opening
+ *             the file.
+ * \returns a pointer to the SDL_AsyncIO structure that is created or NULL on
+ *          failure; call SDL_GetError() for more information.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_CloseAsyncIO
+ * \sa SDL_ReadAsyncIO
+ * \sa SDL_WriteAsyncIO
+ */
+extern SDL_DECLSPEC SDL_AsyncIO * SDLCALL SDL_AsyncIOFromFile(const char *file, const char *mode);
+
+/**
+ * Use this function to get the size of the data stream in an SDL_AsyncIO.
+ *
+ * This call is _not_ asynchronous; it assumes that obtaining this info is a
+ * non-blocking operation in most reasonable cases.
+ *
+ * \param asyncio the SDL_AsyncIO to get the size of the data stream from.
+ * \returns the size of the data stream in the SDL_IOStream on success or a
+ *          negative error code on failure; call SDL_GetError() for more
+ *          information.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ */
+extern SDL_DECLSPEC Sint64 SDLCALL SDL_GetAsyncIOSize(SDL_AsyncIO *asyncio);
+
+/**
+ * Start an async read.
+ *
+ * This function reads up to `size` bytes from `offset` position in the data
+ * source to the area pointed at by `ptr`. This function may read less bytes
+ * than requested.
+ *
+ * This function returns as quickly as possible; it does not wait for the read
+ * to complete. On a successful return, this work will continue in the
+ * background. If the work begins, even failure is asynchronous: a failing
+ * return value from this function only means the work couldn't start at all.
+ *
+ * `ptr` must remain available until the work is done, and may be accessed by
+ * the system at any time until then. Do not allocate it on the stack, as this
+ * might take longer than the life of the calling function to complete!
+ *
+ * An SDL_AsyncIOQueue must be specified. The newly-created task will be added
+ * to it when it completes its work.
+ *
+ * \param asyncio a pointer to an SDL_AsyncIO structure.
+ * \param ptr a pointer to a buffer to read data into.
+ * \param offset the position to start reading in the data source.
+ * \param size the number of bytes to read from the data source.
+ * \param queue a queue to add the new SDL_AsyncIO to.
+ * \param userdata an app-defined pointer that will be provided with the task
+ *                 results.
+ * \returns true on success or false on failure; call SDL_GetError() for more
+ *          information.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_WriteAsyncIO
+ * \sa SDL_CreateAsyncIOQueue
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_ReadAsyncIO(SDL_AsyncIO *asyncio, void *ptr, Uint64 offset, Uint64 size, SDL_AsyncIOQueue *queue, void *userdata);
+
+/**
+ * Start an async write.
+ *
+ * This function writes `size` bytes from `offset` position in the data source
+ * to the area pointed at by `ptr`.
+ *
+ * This function returns as quickly as possible; it does not wait for the
+ * write to complete. On a successful return, this work will continue in the
+ * background. If the work begins, even failure is asynchronous: a failing
+ * return value from this function only means the work couldn't start at all.
+ *
+ * `ptr` must remain available until the work is done, and may be accessed by
+ * the system at any time until then. Do not allocate it on the stack, as this
+ * might take longer than the life of the calling function to complete!
+ *
+ * An SDL_AsyncIOQueue must be specified. The newly-created task will be added
+ * to it when it completes its work.
+ *
+ * \param asyncio a pointer to an SDL_AsyncIO structure.
+ * \param ptr a pointer to a buffer to write data from.
+ * \param offset the position to start writing to the data source.
+ * \param size the number of bytes to write to the data source.
+ * \param queue a queue to add the new SDL_AsyncIO to.
+ * \param userdata an app-defined pointer that will be provided with the task
+ *                 results.
+ * \returns true on success or false on failure; call SDL_GetError() for more
+ *          information.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_ReadAsyncIO
+ * \sa SDL_CreateAsyncIOQueue
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_WriteAsyncIO(SDL_AsyncIO *asyncio, void *ptr, Uint64 offset, Uint64 size, SDL_AsyncIOQueue *queue, void *userdata);
+
+/**
+ * Close and free any allocated resources for an async I/O object.
+ *
+ * Closing a file is _also_ an asynchronous task! If a write failure were to
+ * happen during the closing process, for example, the task results will
+ * report it as usual.
+ *
+ * Closing a file that has been written to does not guarantee the data has
+ * made it to physical media; it may remain in the operating system's file
+ * cache, for later writing to disk. This means that a successfully-closed
+ * file can be lost if the system crashes or loses power in this small window.
+ * To prevent this, call this function with the `flush` parameter set to true.
+ * This will make the operation take longer, and perhaps increase system load
+ * in general, but a successful result guarantees that the data has made it to
+ * physical storage. Don't use this for temporary files, caches, and
+ * unimportant data, and definitely use it for crucial irreplaceable files,
+ * like game saves.
+ *
+ * This function guarantees that the close will happen after any other pending
+ * tasks to `asyncio`, so it's safe to open a file, start several operations,
+ * close the file immediately, then check for all results later. This function
+ * will not block until the tasks have completed.
+ *
+ * Once this function returns true, `asyncio` is no longer valid, regardless
+ * of any future outcomes. Any completed tasks might still contain this
+ * pointer in their SDL_AsyncIOOutcome data, in case the app was using this
+ * value to track information, but it should not be used again.
+ *
+ * If this function returns false, the close wasn't started at all, and it's
+ * safe to attempt to close again later.
+ *
+ * An SDL_AsyncIOQueue must be specified. The newly-created task will be added
+ * to it when it completes its work.
+ *
+ * \param asyncio a pointer to an SDL_AsyncIO structure to close.
+ * \param flush true if data should sync to disk before the task completes.
+ * \param queue a queue to add the new SDL_AsyncIO to.
+ * \param userdata an app-defined pointer that will be provided with the task
+ *                 results.
+ * \returns true on success or false on failure; call SDL_GetError() for more
+ *          information.
+ *
+ * \threadsafety It is safe to call this function from any thread, but two
+ *               threads should not attempt to close the same object.
+ *
+ * \since This function is available since SDL 3.2.0.
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_CloseAsyncIO(SDL_AsyncIO *asyncio, bool flush, SDL_AsyncIOQueue *queue, void *userdata);
+
+/**
+ * Create a task queue for tracking multiple I/O operations.
+ *
+ * Async I/O operations are assigned to a queue when started. The queue can be
+ * checked for completed tasks thereafter.
+ *
+ * \returns a new task queue object or NULL if there was an error; call
+ *          SDL_GetError() for more information.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_DestroyAsyncIOQueue
+ * \sa SDL_GetAsyncIOResult
+ * \sa SDL_WaitAsyncIOResult
+ */
+extern SDL_DECLSPEC SDL_AsyncIOQueue * SDLCALL SDL_CreateAsyncIOQueue(void);
+
+/**
+ * Destroy a previously-created async I/O task queue.
+ *
+ * If there are still tasks pending for this queue, this call will block until
+ * those tasks are finished. All those tasks will be deallocated. Their
+ * results will be lost to the app.
+ *
+ * Any pending reads from SDL_LoadFileAsync() that are still in this queue
+ * will have their buffers deallocated by this function, to prevent a memory
+ * leak.
+ *
+ * Once this function is called, the queue is no longer valid and should not
+ * be used, including by other threads that might access it while destruction
+ * is blocking on pending tasks.
+ *
+ * Do not destroy a queue that still has threads waiting on it through
+ * SDL_WaitAsyncIOResult(). You can call SDL_SignalAsyncIOQueue() first to
+ * unblock those threads, and take measures (such as SDL_WaitThread()) to make
+ * sure they have finished their wait and won't wait on the queue again.
+ *
+ * \param queue the task queue to destroy.
+ *
+ * \threadsafety It is safe to call this function from any thread, so long as
+ *               no other thread is waiting on the queue with
+ *               SDL_WaitAsyncIOResult.
+ *
+ * \since This function is available since SDL 3.2.0.
+ */
+extern SDL_DECLSPEC void SDLCALL SDL_DestroyAsyncIOQueue(SDL_AsyncIOQueue *queue);
+
+/**
+ * Query an async I/O task queue for completed tasks.
+ *
+ * If a task assigned to this queue has finished, this will return true and
+ * fill in `outcome` with the details of the task. If no task in the queue has
+ * finished, this function will return false. This function does not block.
+ *
+ * If a task has completed, this function will free its resources and the task
+ * pointer will no longer be valid. The task will be removed from the queue.
+ *
+ * It is safe for multiple threads to call this function on the same queue at
+ * once; a completed task will only go to one of the threads.
+ *
+ * \param queue the async I/O task queue to query.
+ * \param outcome details of a finished task will be written here. May not be
+ *                NULL.
+ * \returns true if a task has completed, false otherwise.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_WaitAsyncIOResult
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_GetAsyncIOResult(SDL_AsyncIOQueue *queue, SDL_AsyncIOOutcome *outcome);
+
+/**
+ * Block until an async I/O task queue has a completed task.
+ *
+ * This function puts the calling thread to sleep until there a task assigned
+ * to the queue that has finished.
+ *
+ * If a task assigned to the queue has finished, this will return true and
+ * fill in `outcome` with the details of the task. If no task in the queue has
+ * finished, this function will return false.
+ *
+ * If a task has completed, this function will free its resources and the task
+ * pointer will no longer be valid. The task will be removed from the queue.
+ *
+ * It is safe for multiple threads to call this function on the same queue at
+ * once; a completed task will only go to one of the threads.
+ *
+ * Note that by the nature of various platforms, more than one waiting thread
+ * may wake to handle a single task, but only one will obtain it, so
+ * `timeoutMS` is a _maximum_ wait time, and this function may return false
+ * sooner.
+ *
+ * This function may return false if there was a system error, the OS
+ * inadvertently awoke multiple threads, or if SDL_SignalAsyncIOQueue() was
+ * called to wake up all waiting threads without a finished task.
+ *
+ * A timeout can be used to specify a maximum wait time, but rather than
+ * polling, it is possible to have a timeout of -1 to wait forever, and use
+ * SDL_SignalAsyncIOQueue() to wake up the waiting threads later.
+ *
+ * \param queue the async I/O task queue to wait on.
+ * \param outcome details of a finished task will be written here. May not be
+ *                NULL.
+ * \param timeoutMS the maximum time to wait, in milliseconds, or -1 to wait
+ *                  indefinitely.
+ * \returns true if task has completed, false otherwise.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_SignalAsyncIOQueue
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_WaitAsyncIOResult(SDL_AsyncIOQueue *queue, SDL_AsyncIOOutcome *outcome, Sint32 timeoutMS);
+
+/**
+ * Wake up any threads that are blocking in SDL_WaitAsyncIOResult().
+ *
+ * This will unblock any threads that are sleeping in a call to
+ * SDL_WaitAsyncIOResult for the specified queue, and cause them to return
+ * from that function.
+ *
+ * This can be useful when destroying a queue to make sure nothing is touching
+ * it indefinitely. In this case, once this call completes, the caller should
+ * take measures to make sure any previously-blocked threads have returned
+ * from their wait and will not touch the queue again (perhaps by setting a
+ * flag to tell the threads to terminate and then using SDL_WaitThread() to
+ * make sure they've done so).
+ *
+ * \param queue the async I/O task queue to signal.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_WaitAsyncIOResult
+ */
+extern SDL_DECLSPEC void SDLCALL SDL_SignalAsyncIOQueue(SDL_AsyncIOQueue *queue);
+
+/**
+ * Load all the data from a file path, asynchronously.
+ *
+ * This function returns as quickly as possible; it does not wait for the read
+ * to complete. On a successful return, this work will continue in the
+ * background. If the work begins, even failure is asynchronous: a failing
+ * return value from this function only means the work couldn't start at all.
+ *
+ * The data is allocated with a zero byte at the end (null terminated) for
+ * convenience. This extra byte is not included in SDL_AsyncIOOutcome's
+ * bytes_transferred value.
+ *
+ * This function will allocate the buffer to contain the file. It must be
+ * deallocated by calling SDL_free() on SDL_AsyncIOOutcome's buffer field
+ * after completion.
+ *
+ * An SDL_AsyncIOQueue must be specified. The newly-created task will be added
+ * to it when it completes its work.
+ *
+ * \param file the path to read all available data from.
+ * \param queue a queue to add the new SDL_AsyncIO to.
+ * \param userdata an app-defined pointer that will be provided with the task
+ *                 results.
+ * \returns true on success or false on failure; call SDL_GetError() for more
+ *          information.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_LoadFile_IO
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_LoadFileAsync(const char *file, SDL_AsyncIOQueue *queue, void *userdata);
+
+/* Ends C function definitions when using C++ */
+#ifdef __cplusplus
+}
+#endif
+#include <SDL3/SDL_close_code.h>
+
+#endif /* SDL_asyncio_h_ */

release/macos/frameworks/SDL3.xcframework/macos-arm64_x86_64/SDL3.framework/Versions/A/Headers/SDL_atomic.h

@@ -0,0 +1,664 @@
+/*
+  Simple DirectMedia Layer
+  Copyright (C) 1997-2025 Sam Lantinga <slouken@libsdl.org>
+
+  This software is provided 'as-is', without any express or implied
+  warranty.  In no event will the authors be held liable for any damages
+  arising from the use of this software.
+
+  Permission is granted to anyone to use this software for any purpose,
+  including commercial applications, and to alter it and redistribute it
+  freely, subject to the following restrictions:
+
+  1. The origin of this software must not be misrepresented; you must not
+     claim that you wrote the original software. If you use this software
+     in a product, an acknowledgment in the product documentation would be
+     appreciated but is not required.
+  2. Altered source versions must be plainly marked as such, and must not be
+     misrepresented as being the original software.
+  3. This notice may not be removed or altered from any source distribution.
+*/
+
+/**
+ * # CategoryAtomic
+ *
+ * Atomic operations.
+ *
+ * IMPORTANT: If you are not an expert in concurrent lockless programming, you
+ * should not be using any functions in this file. You should be protecting
+ * your data structures with full mutexes instead.
+ *
+ * ***Seriously, here be dragons!***
+ *
+ * You can find out a little more about lockless programming and the subtle
+ * issues that can arise here:
+ * https://learn.microsoft.com/en-us/windows/win32/dxtecharts/lockless-programming
+ *
+ * There's also lots of good information here:
+ *
+ * - https://www.1024cores.net/home/lock-free-algorithms
+ * - https://preshing.com/
+ *
+ * These operations may or may not actually be implemented using processor
+ * specific atomic operations. When possible they are implemented as true
+ * processor specific atomic operations. When that is not possible the are
+ * implemented using locks that *do* use the available atomic operations.
+ *
+ * All of the atomic operations that modify memory are full memory barriers.
+ */
+
+#ifndef SDL_atomic_h_
+#define SDL_atomic_h_
+
+#include <SDL3/SDL_stdinc.h>
+#include <SDL3/SDL_platform_defines.h>
+
+#include <SDL3/SDL_begin_code.h>
+
+/* Set up for C function definitions, even when using C++ */
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/**
+ * An atomic spinlock.
+ *
+ * The atomic locks are efficient spinlocks using CPU instructions, but are
+ * vulnerable to starvation and can spin forever if a thread holding a lock
+ * has been terminated. For this reason you should minimize the code executed
+ * inside an atomic lock and never do expensive things like API or system
+ * calls while holding them.
+ *
+ * They are also vulnerable to starvation if the thread holding the lock is
+ * lower priority than other threads and doesn't get scheduled. In general you
+ * should use mutexes instead, since they have better performance and
+ * contention behavior.
+ *
+ * The atomic locks are not safe to lock recursively.
+ *
+ * Porting Note: The spin lock functions and type are required and can not be
+ * emulated because they are used in the atomic emulation code.
+ */
+typedef int SDL_SpinLock;
+
+/**
+ * Try to lock a spin lock by setting it to a non-zero value.
+ *
+ * ***Please note that spinlocks are dangerous if you don't know what you're
+ * doing. Please be careful using any sort of spinlock!***
+ *
+ * \param lock a pointer to a lock variable.
+ * \returns true if the lock succeeded, false if the lock is already held.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_LockSpinlock
+ * \sa SDL_UnlockSpinlock
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_TryLockSpinlock(SDL_SpinLock *lock);
+
+/**
+ * Lock a spin lock by setting it to a non-zero value.
+ *
+ * ***Please note that spinlocks are dangerous if you don't know what you're
+ * doing. Please be careful using any sort of spinlock!***
+ *
+ * \param lock a pointer to a lock variable.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_TryLockSpinlock
+ * \sa SDL_UnlockSpinlock
+ */
+extern SDL_DECLSPEC void SDLCALL SDL_LockSpinlock(SDL_SpinLock *lock);
+
+/**
+ * Unlock a spin lock by setting it to 0.
+ *
+ * Always returns immediately.
+ *
+ * ***Please note that spinlocks are dangerous if you don't know what you're
+ * doing. Please be careful using any sort of spinlock!***
+ *
+ * \param lock a pointer to a lock variable.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_LockSpinlock
+ * \sa SDL_TryLockSpinlock
+ */
+extern SDL_DECLSPEC void SDLCALL SDL_UnlockSpinlock(SDL_SpinLock *lock);
+
+
+#ifdef SDL_WIKI_DOCUMENTATION_SECTION
+
+/**
+ * Mark a compiler barrier.
+ *
+ * A compiler barrier prevents the compiler from reordering reads and writes
+ * to globally visible variables across the call.
+ *
+ * This macro only prevents the compiler from reordering reads and writes, it
+ * does not prevent the CPU from reordering reads and writes. However, all of
+ * the atomic operations that modify memory are full memory barriers.
+ *
+ * \threadsafety Obviously this macro is safe to use from any thread at any
+ *               time, but if you find yourself needing this, you are probably
+ *               dealing with some very sensitive code; be careful!
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_CompilerBarrier() DoCompilerSpecificReadWriteBarrier()
+
+#elif defined(_MSC_VER) && (_MSC_VER > 1200) && !defined(__clang__)
+void _ReadWriteBarrier(void);
+#pragma intrinsic(_ReadWriteBarrier)
+#define SDL_CompilerBarrier()   _ReadWriteBarrier()
+#elif (defined(__GNUC__) && !defined(SDL_PLATFORM_EMSCRIPTEN)) || (defined(__SUNPRO_C) && (__SUNPRO_C >= 0x5120))
+/* This is correct for all CPUs when using GCC or Solaris Studio 12.1+. */
+#define SDL_CompilerBarrier()   __asm__ __volatile__ ("" : : : "memory")
+#elif defined(__WATCOMC__)
+extern __inline void SDL_CompilerBarrier(void);
+#pragma aux SDL_CompilerBarrier = "" parm [] modify exact [];
+#else
+#define SDL_CompilerBarrier()   \
+{ SDL_SpinLock _tmp = 0; SDL_LockSpinlock(&_tmp); SDL_UnlockSpinlock(&_tmp); }
+#endif
+
+/**
+ * Insert a memory release barrier (function version).
+ *
+ * Please refer to SDL_MemoryBarrierRelease for details. This is a function
+ * version, which might be useful if you need to use this functionality from a
+ * scripting language, etc. Also, some of the macro versions call this
+ * function behind the scenes, where more heavy lifting can happen inside of
+ * SDL. Generally, though, an app written in C/C++/etc should use the macro
+ * version, as it will be more efficient.
+ *
+ * \threadsafety Obviously this function is safe to use from any thread at any
+ *               time, but if you find yourself needing this, you are probably
+ *               dealing with some very sensitive code; be careful!
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_MemoryBarrierRelease
+ */
+extern SDL_DECLSPEC void SDLCALL SDL_MemoryBarrierReleaseFunction(void);
+
+/**
+ * Insert a memory acquire barrier (function version).
+ *
+ * Please refer to SDL_MemoryBarrierRelease for details. This is a function
+ * version, which might be useful if you need to use this functionality from a
+ * scripting language, etc. Also, some of the macro versions call this
+ * function behind the scenes, where more heavy lifting can happen inside of
+ * SDL. Generally, though, an app written in C/C++/etc should use the macro
+ * version, as it will be more efficient.
+ *
+ * \threadsafety Obviously this function is safe to use from any thread at any
+ *               time, but if you find yourself needing this, you are probably
+ *               dealing with some very sensitive code; be careful!
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_MemoryBarrierAcquire
+ */
+extern SDL_DECLSPEC void SDLCALL SDL_MemoryBarrierAcquireFunction(void);
+
+
+#ifdef SDL_WIKI_DOCUMENTATION_SECTION
+
+/**
+ * Insert a memory release barrier (macro version).
+ *
+ * Memory barriers are designed to prevent reads and writes from being
+ * reordered by the compiler and being seen out of order on multi-core CPUs.
+ *
+ * A typical pattern would be for thread A to write some data and a flag, and
+ * for thread B to read the flag and get the data. In this case you would
+ * insert a release barrier between writing the data and the flag,
+ * guaranteeing that the data write completes no later than the flag is
+ * written, and you would insert an acquire barrier between reading the flag
+ * and reading the data, to ensure that all the reads associated with the flag
+ * have completed.
+ *
+ * In this pattern you should always see a release barrier paired with an
+ * acquire barrier and you should gate the data reads/writes with a single
+ * flag variable.
+ *
+ * For more information on these semantics, take a look at the blog post:
+ * http://preshing.com/20120913/acquire-and-release-semantics
+ *
+ * This is the macro version of this functionality; if possible, SDL will use
+ * compiler intrinsics or inline assembly, but some platforms might need to
+ * call the function version of this, SDL_MemoryBarrierReleaseFunction to do
+ * the heavy lifting. Apps that can use the macro should favor it over the
+ * function.
+ *
+ * \threadsafety Obviously this macro is safe to use from any thread at any
+ *               time, but if you find yourself needing this, you are probably
+ *               dealing with some very sensitive code; be careful!
+ *
+ * \since This macro is available since SDL 3.2.0.
+ *
+ * \sa SDL_MemoryBarrierAcquire
+ * \sa SDL_MemoryBarrierReleaseFunction
+ */
+#define SDL_MemoryBarrierRelease() SDL_MemoryBarrierReleaseFunction()
+
+/**
+ * Insert a memory acquire barrier (macro version).
+ *
+ * Please see SDL_MemoryBarrierRelease for the details on what memory barriers
+ * are and when to use them.
+ *
+ * This is the macro version of this functionality; if possible, SDL will use
+ * compiler intrinsics or inline assembly, but some platforms might need to
+ * call the function version of this, SDL_MemoryBarrierAcquireFunction, to do
+ * the heavy lifting. Apps that can use the macro should favor it over the
+ * function.
+ *
+ * \threadsafety Obviously this macro is safe to use from any thread at any
+ *               time, but if you find yourself needing this, you are probably
+ *               dealing with some very sensitive code; be careful!
+ *
+ * \since This macro is available since SDL 3.2.0.
+ *
+ * \sa SDL_MemoryBarrierRelease
+ * \sa SDL_MemoryBarrierAcquireFunction
+ */
+#define SDL_MemoryBarrierAcquire() SDL_MemoryBarrierAcquireFunction()
+
+#elif defined(__GNUC__) && (defined(__powerpc__) || defined(__ppc__))
+#define SDL_MemoryBarrierRelease()   __asm__ __volatile__ ("lwsync" : : : "memory")
+#define SDL_MemoryBarrierAcquire()   __asm__ __volatile__ ("lwsync" : : : "memory")
+#elif defined(__GNUC__) && defined(__aarch64__)
+#define SDL_MemoryBarrierRelease()   __asm__ __volatile__ ("dmb ish" : : : "memory")
+#define SDL_MemoryBarrierAcquire()   __asm__ __volatile__ ("dmb ish" : : : "memory")
+#elif defined(__GNUC__) && defined(__arm__)
+#if 0 /* defined(SDL_PLATFORM_LINUX) || defined(SDL_PLATFORM_ANDROID) */
+/* Information from:
+   https://chromium.googlesource.com/chromium/chromium/+/trunk/base/atomicops_internals_arm_gcc.h#19
+
+   The Linux kernel provides a helper function which provides the right code for a memory barrier,
+   hard-coded at address 0xffff0fa0
+*/
+typedef void (*SDL_KernelMemoryBarrierFunc)();
+#define SDL_MemoryBarrierRelease()  ((SDL_KernelMemoryBarrierFunc)0xffff0fa0)()
+#define SDL_MemoryBarrierAcquire()  ((SDL_KernelMemoryBarrierFunc)0xffff0fa0)()
+#else
+#if defined(__ARM_ARCH_7__) || defined(__ARM_ARCH_7A__) || defined(__ARM_ARCH_7EM__) || defined(__ARM_ARCH_7R__) || defined(__ARM_ARCH_7M__) || defined(__ARM_ARCH_7S__) || defined(__ARM_ARCH_8A__)
+#define SDL_MemoryBarrierRelease()   __asm__ __volatile__ ("dmb ish" : : : "memory")
+#define SDL_MemoryBarrierAcquire()   __asm__ __volatile__ ("dmb ish" : : : "memory")
+#elif defined(__ARM_ARCH_6__) || defined(__ARM_ARCH_6J__) || defined(__ARM_ARCH_6K__) || defined(__ARM_ARCH_6T2__) || defined(__ARM_ARCH_6Z__) || defined(__ARM_ARCH_6ZK__)
+#ifdef __thumb__
+/* The mcr instruction isn't available in thumb mode, use real functions */
+#define SDL_MEMORY_BARRIER_USES_FUNCTION
+#define SDL_MemoryBarrierRelease()   SDL_MemoryBarrierReleaseFunction()
+#define SDL_MemoryBarrierAcquire()   SDL_MemoryBarrierAcquireFunction()
+#else
+#define SDL_MemoryBarrierRelease()   __asm__ __volatile__ ("mcr p15, 0, %0, c7, c10, 5" : : "r"(0) : "memory")
+#define SDL_MemoryBarrierAcquire()   __asm__ __volatile__ ("mcr p15, 0, %0, c7, c10, 5" : : "r"(0) : "memory")
+#endif /* __thumb__ */
+#else
+#define SDL_MemoryBarrierRelease()   __asm__ __volatile__ ("" : : : "memory")
+#define SDL_MemoryBarrierAcquire()   __asm__ __volatile__ ("" : : : "memory")
+#endif /* SDL_PLATFORM_LINUX || SDL_PLATFORM_ANDROID */
+#endif /* __GNUC__ && __arm__ */
+#else
+#if (defined(__SUNPRO_C) && (__SUNPRO_C >= 0x5120))
+/* This is correct for all CPUs on Solaris when using Solaris Studio 12.1+. */
+#include <mbarrier.h>
+#define SDL_MemoryBarrierRelease()  __machine_rel_barrier()
+#define SDL_MemoryBarrierAcquire()  __machine_acq_barrier()
+#else
+/* This is correct for the x86 and x64 CPUs, and we'll expand this over time. */
+#define SDL_MemoryBarrierRelease()  SDL_CompilerBarrier()
+#define SDL_MemoryBarrierAcquire()  SDL_CompilerBarrier()
+#endif
+#endif
+
+/* "REP NOP" is PAUSE, coded for tools that don't know it by that name. */
+#ifdef SDL_WIKI_DOCUMENTATION_SECTION
+
+/**
+ * A macro to insert a CPU-specific "pause" instruction into the program.
+ *
+ * This can be useful in busy-wait loops, as it serves as a hint to the CPU as
+ * to the program's intent; some CPUs can use this to do more efficient
+ * processing. On some platforms, this doesn't do anything, so using this
+ * macro might just be a harmless no-op.
+ *
+ * Note that if you are busy-waiting, there are often more-efficient
+ * approaches with other synchronization primitives: mutexes, semaphores,
+ * condition variables, etc.
+ *
+ * \threadsafety This macro is safe to use from any thread.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_CPUPauseInstruction() DoACPUPauseInACompilerAndArchitectureSpecificWay
+
+#elif (defined(__GNUC__) || defined(__clang__)) && (defined(__i386__) || defined(__x86_64__))
+    #define SDL_CPUPauseInstruction() __asm__ __volatile__("pause\n")  /* Some assemblers can't do REP NOP, so go with PAUSE. */
+#elif (defined(__arm__) && defined(__ARM_ARCH) && __ARM_ARCH >= 7) || defined(__aarch64__)
+    #define SDL_CPUPauseInstruction() __asm__ __volatile__("yield" ::: "memory")
+#elif (defined(__powerpc__) || defined(__powerpc64__))
+    #define SDL_CPUPauseInstruction() __asm__ __volatile__("or 27,27,27");
+#elif (defined(__riscv) && __riscv_xlen == 64)
+    #define SDL_CPUPauseInstruction() __asm__ __volatile__(".insn i 0x0F, 0, x0, x0, 0x010");
+#elif defined(_MSC_VER) && (defined(_M_IX86) || defined(_M_X64))
+    #define SDL_CPUPauseInstruction() _mm_pause()  /* this is actually "rep nop" and not a SIMD instruction. No inline asm in MSVC x86-64! */
+#elif defined(_MSC_VER) && (defined(_M_ARM) || defined(_M_ARM64))
+    #define SDL_CPUPauseInstruction() __yield()
+#elif defined(__WATCOMC__) && defined(__386__)
+    extern __inline void SDL_CPUPauseInstruction(void);
+    #pragma aux SDL_CPUPauseInstruction = ".686p" ".xmm2" "pause"
+#else
+    #define SDL_CPUPauseInstruction()
+#endif
+
+
+/**
+ * A type representing an atomic integer value.
+ *
+ * This can be used to manage a value that is synchronized across multiple
+ * CPUs without a race condition; when an app sets a value with
+ * SDL_SetAtomicInt all other threads, regardless of the CPU it is running on,
+ * will see that value when retrieved with SDL_GetAtomicInt, regardless of CPU
+ * caches, etc.
+ *
+ * This is also useful for atomic compare-and-swap operations: a thread can
+ * change the value as long as its current value matches expectations. When
+ * done in a loop, one can guarantee data consistency across threads without a
+ * lock (but the usual warnings apply: if you don't know what you're doing, or
+ * you don't do it carefully, you can confidently cause any number of
+ * disasters with this, so in most cases, you _should_ use a mutex instead of
+ * this!).
+ *
+ * This is a struct so people don't accidentally use numeric operations on it
+ * directly. You have to use SDL atomic functions.
+ *
+ * \since This struct is available since SDL 3.2.0.
+ *
+ * \sa SDL_CompareAndSwapAtomicInt
+ * \sa SDL_GetAtomicInt
+ * \sa SDL_SetAtomicInt
+ * \sa SDL_AddAtomicInt
+ */
+typedef struct SDL_AtomicInt { int value; } SDL_AtomicInt;
+
+/**
+ * Set an atomic variable to a new value if it is currently an old value.
+ *
+ * ***Note: If you don't know what this function is for, you shouldn't use
+ * it!***
+ *
+ * \param a a pointer to an SDL_AtomicInt variable to be modified.
+ * \param oldval the old value.
+ * \param newval the new value.
+ * \returns true if the atomic variable was set, false otherwise.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetAtomicInt
+ * \sa SDL_SetAtomicInt
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_CompareAndSwapAtomicInt(SDL_AtomicInt *a, int oldval, int newval);
+
+/**
+ * Set an atomic variable to a value.
+ *
+ * This function also acts as a full memory barrier.
+ *
+ * ***Note: If you don't know what this function is for, you shouldn't use
+ * it!***
+ *
+ * \param a a pointer to an SDL_AtomicInt variable to be modified.
+ * \param v the desired value.
+ * \returns the previous value of the atomic variable.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetAtomicInt
+ */
+extern SDL_DECLSPEC int SDLCALL SDL_SetAtomicInt(SDL_AtomicInt *a, int v);
+
+/**
+ * Get the value of an atomic variable.
+ *
+ * ***Note: If you don't know what this function is for, you shouldn't use
+ * it!***
+ *
+ * \param a a pointer to an SDL_AtomicInt variable.
+ * \returns the current value of an atomic variable.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_SetAtomicInt
+ */
+extern SDL_DECLSPEC int SDLCALL SDL_GetAtomicInt(SDL_AtomicInt *a);
+
+/**
+ * Add to an atomic variable.
+ *
+ * This function also acts as a full memory barrier.
+ *
+ * ***Note: If you don't know what this function is for, you shouldn't use
+ * it!***
+ *
+ * \param a a pointer to an SDL_AtomicInt variable to be modified.
+ * \param v the desired value to add.
+ * \returns the previous value of the atomic variable.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_AtomicDecRef
+ * \sa SDL_AtomicIncRef
+ */
+extern SDL_DECLSPEC int SDLCALL SDL_AddAtomicInt(SDL_AtomicInt *a, int v);
+
+#ifndef SDL_AtomicIncRef
+
+/**
+ * Increment an atomic variable used as a reference count.
+ *
+ * ***Note: If you don't know what this macro is for, you shouldn't use it!***
+ *
+ * \param a a pointer to an SDL_AtomicInt to increment.
+ * \returns the previous value of the atomic variable.
+ *
+ * \threadsafety It is safe to call this macro from any thread.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ *
+ * \sa SDL_AtomicDecRef
+ */
+#define SDL_AtomicIncRef(a)    SDL_AddAtomicInt(a, 1)
+#endif
+
+#ifndef SDL_AtomicDecRef
+
+/**
+ * Decrement an atomic variable used as a reference count.
+ *
+ * ***Note: If you don't know what this macro is for, you shouldn't use it!***
+ *
+ * \param a a pointer to an SDL_AtomicInt to decrement.
+ * \returns true if the variable reached zero after decrementing, false
+ *          otherwise.
+ *
+ * \threadsafety It is safe to call this macro from any thread.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ *
+ * \sa SDL_AtomicIncRef
+ */
+#define SDL_AtomicDecRef(a)    (SDL_AddAtomicInt(a, -1) == 1)
+#endif
+
+/**
+ * A type representing an atomic unsigned 32-bit value.
+ *
+ * This can be used to manage a value that is synchronized across multiple
+ * CPUs without a race condition; when an app sets a value with
+ * SDL_SetAtomicU32 all other threads, regardless of the CPU it is running on,
+ * will see that value when retrieved with SDL_GetAtomicU32, regardless of CPU
+ * caches, etc.
+ *
+ * This is also useful for atomic compare-and-swap operations: a thread can
+ * change the value as long as its current value matches expectations. When
+ * done in a loop, one can guarantee data consistency across threads without a
+ * lock (but the usual warnings apply: if you don't know what you're doing, or
+ * you don't do it carefully, you can confidently cause any number of
+ * disasters with this, so in most cases, you _should_ use a mutex instead of
+ * this!).
+ *
+ * This is a struct so people don't accidentally use numeric operations on it
+ * directly. You have to use SDL atomic functions.
+ *
+ * \since This struct is available since SDL 3.2.0.
+ *
+ * \sa SDL_CompareAndSwapAtomicU32
+ * \sa SDL_GetAtomicU32
+ * \sa SDL_SetAtomicU32
+ */
+typedef struct SDL_AtomicU32 { Uint32 value; } SDL_AtomicU32;
+
+/**
+ * Set an atomic variable to a new value if it is currently an old value.
+ *
+ * ***Note: If you don't know what this function is for, you shouldn't use
+ * it!***
+ *
+ * \param a a pointer to an SDL_AtomicU32 variable to be modified.
+ * \param oldval the old value.
+ * \param newval the new value.
+ * \returns true if the atomic variable was set, false otherwise.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetAtomicU32
+ * \sa SDL_SetAtomicU32
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_CompareAndSwapAtomicU32(SDL_AtomicU32 *a, Uint32 oldval, Uint32 newval);
+
+/**
+ * Set an atomic variable to a value.
+ *
+ * This function also acts as a full memory barrier.
+ *
+ * ***Note: If you don't know what this function is for, you shouldn't use
+ * it!***
+ *
+ * \param a a pointer to an SDL_AtomicU32 variable to be modified.
+ * \param v the desired value.
+ * \returns the previous value of the atomic variable.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetAtomicU32
+ */
+extern SDL_DECLSPEC Uint32 SDLCALL SDL_SetAtomicU32(SDL_AtomicU32 *a, Uint32 v);
+
+/**
+ * Get the value of an atomic variable.
+ *
+ * ***Note: If you don't know what this function is for, you shouldn't use
+ * it!***
+ *
+ * \param a a pointer to an SDL_AtomicU32 variable.
+ * \returns the current value of an atomic variable.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_SetAtomicU32
+ */
+extern SDL_DECLSPEC Uint32 SDLCALL SDL_GetAtomicU32(SDL_AtomicU32 *a);
+
+/**
+ * Set a pointer to a new value if it is currently an old value.
+ *
+ * ***Note: If you don't know what this function is for, you shouldn't use
+ * it!***
+ *
+ * \param a a pointer to a pointer.
+ * \param oldval the old pointer value.
+ * \param newval the new pointer value.
+ * \returns true if the pointer was set, false otherwise.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_CompareAndSwapAtomicInt
+ * \sa SDL_GetAtomicPointer
+ * \sa SDL_SetAtomicPointer
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_CompareAndSwapAtomicPointer(void **a, void *oldval, void *newval);
+
+/**
+ * Set a pointer to a value atomically.
+ *
+ * ***Note: If you don't know what this function is for, you shouldn't use
+ * it!***
+ *
+ * \param a a pointer to a pointer.
+ * \param v the desired pointer value.
+ * \returns the previous value of the pointer.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_CompareAndSwapAtomicPointer
+ * \sa SDL_GetAtomicPointer
+ */
+extern SDL_DECLSPEC void * SDLCALL SDL_SetAtomicPointer(void **a, void *v);
+
+/**
+ * Get the value of a pointer atomically.
+ *
+ * ***Note: If you don't know what this function is for, you shouldn't use
+ * it!***
+ *
+ * \param a a pointer to a pointer.
+ * \returns the current value of a pointer.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_CompareAndSwapAtomicPointer
+ * \sa SDL_SetAtomicPointer
+ */
+extern SDL_DECLSPEC void * SDLCALL SDL_GetAtomicPointer(void **a);
+
+/* Ends C function definitions when using C++ */
+#ifdef __cplusplus
+}
+#endif
+
+#include <SDL3/SDL_close_code.h>
+
+#endif /* SDL_atomic_h_ */

release/macos/frameworks/SDL3.xcframework/macos-arm64_x86_64/SDL3.framework/Versions/A/Headers/SDL_begin_code.h

@@ -0,0 +1,486 @@
+/*
+  Simple DirectMedia Layer
+  Copyright (C) 1997-2025 Sam Lantinga <slouken@libsdl.org>
+
+  This software is provided 'as-is', without any express or implied
+  warranty.  In no event will the authors be held liable for any damages
+  arising from the use of this software.
+
+  Permission is granted to anyone to use this software for any purpose,
+  including commercial applications, and to alter it and redistribute it
+  freely, subject to the following restrictions:
+
+  1. The origin of this software must not be misrepresented; you must not
+     claim that you wrote the original software. If you use this software
+     in a product, an acknowledgment in the product documentation would be
+     appreciated but is not required.
+  2. Altered source versions must be plainly marked as such, and must not be
+     misrepresented as being the original software.
+  3. This notice may not be removed or altered from any source distribution.
+*/
+
+/* WIKI CATEGORY: BeginCode */
+
+/**
+ * # CategoryBeginCode
+ *
+ * `SDL_begin_code.h` sets things up for C dynamic library function
+ * definitions, static inlined functions, and structures aligned at 4-byte
+ * alignment. If you don't like ugly C preprocessor code, don't look at this
+ * file. :)
+ *
+ * SDL's headers use this; applications generally should not include this
+ * header directly.
+ */
+
+/* This shouldn't be nested -- included it around code only. */
+#ifdef SDL_begin_code_h
+#error Nested inclusion of SDL_begin_code.h
+#endif
+#define SDL_begin_code_h
+
+#ifdef SDL_WIKI_DOCUMENTATION_SECTION
+
+/**
+ * A macro to tag a symbol as deprecated.
+ *
+ * A function is marked deprecated by adding this macro to its declaration:
+ *
+ * ```c
+ * extern SDL_DEPRECATED int ThisFunctionWasABadIdea(void);
+ * ```
+ *
+ * Compilers with deprecation support can give a warning when a deprecated
+ * function is used. This symbol may be used in SDL's headers, but apps are
+ * welcome to use it for their own interfaces as well.
+ *
+ * SDL, on occasion, might deprecate a function for various reasons. However,
+ * SDL never removes symbols before major versions, so deprecated interfaces
+ * in SDL3 will remain available until SDL4, where it would be expected an app
+ * would have to take steps to migrate anyhow.
+ *
+ * On compilers without a deprecation mechanism, this is defined to nothing,
+ * and using a deprecated function will not generate a warning.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_DEPRECATED __attribute__((deprecated))
+
+/**
+ * A macro to tag a symbol as a public API.
+ *
+ * SDL uses this macro for all its public functions. On some targets, it is
+ * used to signal to the compiler that this function needs to be exported from
+ * a shared library, but it might have other side effects.
+ *
+ * This symbol is used in SDL's headers, but apps and other libraries are
+ * welcome to use it for their own interfaces as well.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_DECLSPEC __attribute__ ((visibility("default")))
+
+/**
+ * A macro to set a function's calling conventions.
+ *
+ * SDL uses this macro for all its public functions, and any callbacks it
+ * defines. This macro guarantees that calling conventions match between SDL
+ * and the app, even if the two were built with different compilers or
+ * optimization settings.
+ *
+ * When writing a callback function, it is very important for it to be
+ * correctly tagged with SDLCALL, as mismatched calling conventions can cause
+ * strange behaviors and can be difficult to diagnose. Plus, on many
+ * platforms, SDLCALL is defined to nothing, so compilers won't be able to
+ * warn that the tag is missing.
+ *
+ * This symbol is used in SDL's headers, but apps and other libraries are
+ * welcome to use it for their own interfaces as well.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDLCALL __cdecl
+
+/**
+ * A macro to request a function be inlined.
+ *
+ * This is a hint to the compiler to inline a function. The compiler is free
+ * to ignore this request. On compilers without inline support, this is
+ * defined to nothing.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_INLINE __inline
+
+/**
+ * A macro to demand a function be inlined.
+ *
+ * This is a command to the compiler to inline a function. SDL uses this macro
+ * in its public headers for a handful of simple functions. On compilers
+ * without forceinline support, this is defined to `static SDL_INLINE`, which
+ * is often good enough.
+ *
+ * This symbol is used in SDL's headers, but apps and other libraries are
+ * welcome to use it for their own interfaces as well.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_FORCE_INLINE __forceinline
+
+/**
+ * A macro to tag a function as never-returning.
+ *
+ * This is a hint to the compiler that a function does not return. An example
+ * of a function like this is the C runtime's exit() function.
+ *
+ * This hint can lead to code optimizations, and help analyzers understand
+ * code flow better. On compilers without noreturn support, this is defined to
+ * nothing.
+ *
+ * This symbol is used in SDL's headers, but apps and other libraries are
+ * welcome to use it for their own interfaces as well.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_NORETURN __attribute__((noreturn))
+
+/**
+ * A macro to tag a function as never-returning (for analysis purposes).
+ *
+ * This is almost identical to SDL_NORETURN, except functions marked with this
+ * _can_ actually return. The difference is that this isn't used for code
+ * generation, but rather static analyzers use this information to assume
+ * truths about program state and available code paths. Specifically, this tag
+ * is useful for writing an assertion mechanism. Indeed, SDL_assert uses this
+ * tag behind the scenes. Generally, apps that don't understand the specific
+ * use-case for this tag should avoid using it directly.
+ *
+ * On compilers without analyzer_noreturn support, this is defined to nothing.
+ *
+ * This symbol is used in SDL's headers, but apps and other libraries are
+ * welcome to use it for their own interfaces as well.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_ANALYZER_NORETURN __attribute__((analyzer_noreturn))
+
+
+/**
+ * A macro to signal that a case statement without a `break` is intentional.
+ *
+ * C compilers have gotten more aggressive about warning when a switch's
+ * `case` block does not end with a `break` or other flow control statement,
+ * flowing into the next case's code, as this is a common accident that leads
+ * to strange bugs. But sometimes falling through to the next case is the
+ * correct and desired behavior. This symbol lets an app communicate this
+ * intention to the compiler, so it doesn't generate a warning.
+ *
+ * It is used like this:
+ *
+ * ```c
+ * switch (x) {
+ *     case 1:
+ *         DoSomethingOnlyForOne();
+ *         SDL_FALLTHROUGH;  // tell the compiler this was intentional.
+ *     case 2:
+ *         DoSomethingForOneAndTwo();
+ *         break;
+ * }
+ * ```
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_FALLTHROUGH [[fallthrough]]
+
+/**
+ * A macro to tag a function's return value as critical.
+ *
+ * This is a hint to the compiler that a function's return value should not be
+ * ignored.
+ *
+ * If an NODISCARD function's return value is thrown away (the function is
+ * called as if it returns `void`), the compiler will issue a warning.
+ *
+ * While it's generally good practice to check return values for errors, often
+ * times legitimate programs do not for good reasons. Be careful about what
+ * functions are tagged as NODISCARD. It operates best when used on a function
+ * that's failure is surprising and catastrophic; a good example would be a
+ * program that checks the return values of all its file write function calls
+ * but not the call to close the file, which it assumes incorrectly never
+ * fails.
+ *
+ * Function callers that want to throw away a NODISCARD return value can call
+ * the function with a `(void)` cast, which informs the compiler the act is
+ * intentional.
+ *
+ * On compilers without nodiscard support, this is defined to nothing.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_NODISCARD [[nodiscard]]
+
+/**
+ * A macro to tag a function as an allocator.
+ *
+ * This is a hint to the compiler that a function is an allocator, like
+ * malloc(), with certain rules. A description of how GCC treats this hint is
+ * here:
+ *
+ * https://gcc.gnu.org/onlinedocs/gcc/Common-Function-Attributes.html#index-malloc-function-attribute
+ *
+ * On compilers without allocator tag support, this is defined to nothing.
+ *
+ * Most apps don't need to, and should not, use this directly.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_MALLOC __declspec(allocator) __desclspec(restrict)
+
+/**
+ * A macro to tag a function as returning a certain allocation.
+ *
+ * This is a hint to the compiler that a function allocates and returns a
+ * specific amount of memory based on one of its arguments. For example, the C
+ * runtime's malloc() function could use this macro with an argument of 1
+ * (first argument to malloc is the size of the allocation).
+ *
+ * On compilers without alloc_size support, this is defined to nothing.
+ *
+ * Most apps don't need to, and should not, use this directly.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_ALLOC_SIZE(p) __attribute__((alloc_size(p)))
+
+/**
+ * A macro to tag a pointer variable, to help with pointer aliasing.
+ *
+ * A good explanation of the restrict keyword is here:
+ *
+ * https://en.wikipedia.org/wiki/Restrict
+ *
+ * On compilers without restrict support, this is defined to nothing.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_RESTRICT __restrict__
+
+/**
+ * Check if the compiler supports a given builtin functionality.
+ *
+ * This allows preprocessor checks for things that otherwise might fail to
+ * compile.
+ *
+ * Supported by virtually all clang versions and more-recent GCCs. Use this
+ * instead of checking the clang version if possible.
+ *
+ * On compilers without has_builtin support, this is defined to 0 (always
+ * false).
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_HAS_BUILTIN(x) __has_builtin(x)
+
+/* end of wiki documentation section. */
+#endif
+
+#ifndef SDL_HAS_BUILTIN
+#ifdef __has_builtin
+#define SDL_HAS_BUILTIN(x) __has_builtin(x)
+#else
+#define SDL_HAS_BUILTIN(x) 0
+#endif
+#endif
+
+#ifndef SDL_DEPRECATED
+#  if defined(__GNUC__) && (__GNUC__ >= 4)  /* technically, this arrived in gcc 3.1, but oh well. */
+#    define SDL_DEPRECATED __attribute__((deprecated))
+#  elif defined(_MSC_VER)
+#    define SDL_DEPRECATED __declspec(deprecated)
+#  else
+#    define SDL_DEPRECATED
+#  endif
+#endif
+
+#ifndef SDL_UNUSED
+#  ifdef __GNUC__
+#    define SDL_UNUSED __attribute__((unused))
+#  else
+#    define SDL_UNUSED
+#  endif
+#endif
+
+/* Some compilers use a special export keyword */
+#ifndef SDL_DECLSPEC
+# if defined(SDL_PLATFORM_WINDOWS)
+#  ifdef DLL_EXPORT
+#   define SDL_DECLSPEC __declspec(dllexport)
+#  else
+#   define SDL_DECLSPEC
+#  endif
+# else
+#  if defined(__GNUC__) && __GNUC__ >= 4
+#   define SDL_DECLSPEC __attribute__ ((visibility("default")))
+#  else
+#   define SDL_DECLSPEC
+#  endif
+# endif
+#endif
+
+/* By default SDL uses the C calling convention */
+#ifndef SDLCALL
+#if defined(SDL_PLATFORM_WINDOWS) && !defined(__GNUC__)
+#define SDLCALL __cdecl
+#else
+#define SDLCALL
+#endif
+#endif /* SDLCALL */
+
+/* Force structure packing at 4 byte alignment.
+   This is necessary if the header is included in code which has structure
+   packing set to an alternate value, say for loading structures from disk.
+   The packing is reset to the previous value in SDL_close_code.h
+ */
+#if defined(_MSC_VER) || defined(__MWERKS__) || defined(__BORLANDC__)
+#ifdef _MSC_VER
+#pragma warning(disable: 4103)
+#endif
+#ifdef __clang__
+#pragma clang diagnostic ignored "-Wpragma-pack"
+#endif
+#ifdef __BORLANDC__
+#pragma nopackwarning
+#endif
+#ifdef _WIN64
+/* Use 8-byte alignment on 64-bit architectures, so pointers are aligned */
+#pragma pack(push,8)
+#else
+#pragma pack(push,4)
+#endif
+#endif /* Compiler needs structure packing set */
+
+#ifndef SDL_INLINE
+#ifdef __GNUC__
+#define SDL_INLINE __inline__
+#elif defined(_MSC_VER) || defined(__BORLANDC__) || \
+      defined(__DMC__) || defined(__SC__) || \
+      defined(__WATCOMC__) || defined(__LCC__) || \
+      defined(__DECC) || defined(__CC_ARM)
+#define SDL_INLINE __inline
+#ifndef __inline__
+#define __inline__ __inline
+#endif
+#else
+#define SDL_INLINE inline
+#ifndef __inline__
+#define __inline__ inline
+#endif
+#endif
+#endif /* SDL_INLINE not defined */
+
+#ifndef SDL_FORCE_INLINE
+#ifdef _MSC_VER
+#define SDL_FORCE_INLINE __forceinline
+#elif ( (defined(__GNUC__) && (__GNUC__ >= 4)) || defined(__clang__) )
+#define SDL_FORCE_INLINE __attribute__((always_inline)) static __inline__
+#else
+#define SDL_FORCE_INLINE static SDL_INLINE
+#endif
+#endif /* SDL_FORCE_INLINE not defined */
+
+#ifndef SDL_NORETURN
+#ifdef __GNUC__
+#define SDL_NORETURN __attribute__((noreturn))
+#elif defined(_MSC_VER)
+#define SDL_NORETURN __declspec(noreturn)
+#else
+#define SDL_NORETURN
+#endif
+#endif /* SDL_NORETURN not defined */
+
+#ifdef __clang__
+#if __has_feature(attribute_analyzer_noreturn)
+#define SDL_ANALYZER_NORETURN __attribute__((analyzer_noreturn))
+#endif
+#endif
+
+#ifndef SDL_ANALYZER_NORETURN
+#define SDL_ANALYZER_NORETURN
+#endif
+
+/* Apparently this is needed by several Windows compilers */
+#ifndef __MACH__
+#ifndef NULL
+#ifdef __cplusplus
+#define NULL 0
+#else
+#define NULL ((void *)0)
+#endif
+#endif /* NULL */
+#endif /* ! macOS - breaks precompiled headers */
+
+#ifndef SDL_FALLTHROUGH
+#if (defined(__cplusplus) && __cplusplus >= 201703L) || \
+    (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 202000L)
+#define SDL_FALLTHROUGH [[fallthrough]]
+#else
+#if defined(__has_attribute) && !defined(__SUNPRO_C) && !defined(__SUNPRO_CC)
+#define SDL_HAS_FALLTHROUGH __has_attribute(__fallthrough__)
+#else
+#define SDL_HAS_FALLTHROUGH 0
+#endif /* __has_attribute */
+#if SDL_HAS_FALLTHROUGH && \
+   ((defined(__GNUC__) && __GNUC__ >= 7) || \
+    (defined(__clang_major__) && __clang_major__ >= 10))
+#define SDL_FALLTHROUGH __attribute__((__fallthrough__))
+#else
+#define SDL_FALLTHROUGH do {} while (0) /* fallthrough */
+#endif /* SDL_HAS_FALLTHROUGH */
+#undef SDL_HAS_FALLTHROUGH
+#endif /* C++17 or C2x */
+#endif /* SDL_FALLTHROUGH not defined */
+
+#ifndef SDL_NODISCARD
+#if (defined(__cplusplus) && __cplusplus >= 201703L) || \
+    (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 202311L)
+#define SDL_NODISCARD [[nodiscard]]
+#elif ( (defined(__GNUC__) && (__GNUC__ >= 4)) || defined(__clang__) )
+#define SDL_NODISCARD __attribute__((warn_unused_result))
+#elif defined(_MSC_VER) && (_MSC_VER >= 1700)
+#define SDL_NODISCARD _Check_return_
+#else
+#define SDL_NODISCARD
+#endif /* C++17 or C23 */
+#endif /* SDL_NODISCARD not defined */
+
+#ifndef SDL_MALLOC
+#if defined(__GNUC__) && (__GNUC__ >= 3)
+#define SDL_MALLOC __attribute__((malloc))
+/** FIXME
+#elif defined(_MSC_VER)
+#define SDL_MALLOC __declspec(allocator) __desclspec(restrict)
+**/
+#else
+#define SDL_MALLOC
+#endif
+#endif /* SDL_MALLOC not defined */
+
+#ifndef SDL_ALLOC_SIZE
+#if (defined(__clang__) && __clang_major__ >= 4) || (defined(__GNUC__) && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 3)))
+#define SDL_ALLOC_SIZE(p) __attribute__((alloc_size(p)))
+#elif defined(_MSC_VER)
+#define SDL_ALLOC_SIZE(p)
+#else
+#define SDL_ALLOC_SIZE(p)
+#endif
+#endif /* SDL_ALLOC_SIZE not defined */
+
+#ifndef SDL_ALLOC_SIZE2
+#if (defined(__clang__) && __clang_major__ >= 4) || (defined(__GNUC__) && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 3)))
+#define SDL_ALLOC_SIZE2(p1, p2) __attribute__((alloc_size(p1, p2)))
+#elif defined(_MSC_VER)
+#define SDL_ALLOC_SIZE2(p1, p2)
+#else
+#define SDL_ALLOC_SIZE2(p1, p2)
+#endif
+#endif /* SDL_ALLOC_SIZE2 not defined */

release/macos/frameworks/SDL3.xcframework/macos-arm64_x86_64/SDL3.framework/Versions/A/Headers/SDL_bits.h

@@ -0,0 +1,147 @@
+/*
+  Simple DirectMedia Layer
+  Copyright (C) 1997-2025 Sam Lantinga <slouken@libsdl.org>
+
+  This software is provided 'as-is', without any express or implied
+  warranty.  In no event will the authors be held liable for any damages
+  arising from the use of this software.
+
+  Permission is granted to anyone to use this software for any purpose,
+  including commercial applications, and to alter it and redistribute it
+  freely, subject to the following restrictions:
+
+  1. The origin of this software must not be misrepresented; you must not
+     claim that you wrote the original software. If you use this software
+     in a product, an acknowledgment in the product documentation would be
+     appreciated but is not required.
+  2. Altered source versions must be plainly marked as such, and must not be
+     misrepresented as being the original software.
+  3. This notice may not be removed or altered from any source distribution.
+*/
+
+/**
+ * # CategoryBits
+ *
+ * Functions for fiddling with bits and bitmasks.
+ */
+
+#ifndef SDL_bits_h_
+#define SDL_bits_h_
+
+#include <SDL3/SDL_stdinc.h>
+
+#include <SDL3/SDL_begin_code.h>
+/* Set up for C function definitions, even when using C++ */
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#if defined(__WATCOMC__) && defined(__386__)
+extern __inline int _SDL_bsr_watcom(Uint32);
+#pragma aux _SDL_bsr_watcom = \
+    "bsr eax, eax" \
+    parm [eax] nomemory \
+    value [eax] \
+    modify exact [eax] nomemory;
+#endif
+
+/**
+ * Get the index of the most significant (set) bit in a 32-bit number.
+ *
+ * Result is undefined when called with 0. This operation can also be stated
+ * as "count leading zeroes" and "log base 2".
+ *
+ * Note that this is a forced-inline function in a header, and not a public
+ * API function available in the SDL library (which is to say, the code is
+ * embedded in the calling program and the linker and dynamic loader will not
+ * be able to find this function inside SDL itself).
+ *
+ * \param x the 32-bit value to examine.
+ * \returns the index of the most significant bit, or -1 if the value is 0.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ */
+SDL_FORCE_INLINE int SDL_MostSignificantBitIndex32(Uint32 x)
+{
+#if defined(__GNUC__) && (__GNUC__ >= 4 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 4))
+    /* Count Leading Zeroes builtin in GCC.
+     * http://gcc.gnu.org/onlinedocs/gcc-4.3.4/gcc/Other-Builtins.html
+     */
+    if (x == 0) {
+        return -1;
+    }
+    return 31 - __builtin_clz(x);
+#elif defined(__WATCOMC__) && defined(__386__)
+    if (x == 0) {
+        return -1;
+    }
+    return _SDL_bsr_watcom(x);
+#elif defined(_MSC_VER) && _MSC_VER >= 1400
+    unsigned long index;
+    if (_BitScanReverse(&index, x)) {
+        return (int)index;
+    }
+    return -1;
+#else
+    /* Based off of Bit Twiddling Hacks by Sean Eron Anderson
+     * <seander@cs.stanford.edu>, released in the public domain.
+     * http://graphics.stanford.edu/~seander/bithacks.html#IntegerLog
+     */
+    const Uint32 b[] = {0x2, 0xC, 0xF0, 0xFF00, 0xFFFF0000};
+    const int    S[] = {1, 2, 4, 8, 16};
+
+    int msbIndex = 0;
+    int i;
+
+    if (x == 0) {
+        return -1;
+    }
+
+    for (i = 4; i >= 0; i--)
+    {
+        if (x & b[i])
+        {
+            x >>= S[i];
+            msbIndex |= S[i];
+        }
+    }
+
+    return msbIndex;
+#endif
+}
+
+/**
+ * Determine if a unsigned 32-bit value has exactly one bit set.
+ *
+ * If there are no bits set (`x` is zero), or more than one bit set, this
+ * returns false. If any one bit is exclusively set, this returns true.
+ *
+ * Note that this is a forced-inline function in a header, and not a public
+ * API function available in the SDL library (which is to say, the code is
+ * embedded in the calling program and the linker and dynamic loader will not
+ * be able to find this function inside SDL itself).
+ *
+ * \param x the 32-bit value to examine.
+ * \returns true if exactly one bit is set in `x`, false otherwise.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ */
+SDL_FORCE_INLINE bool SDL_HasExactlyOneBitSet32(Uint32 x)
+{
+    if (x && !(x & (x - 1))) {
+        return true;
+    }
+    return false;
+}
+
+/* Ends C function definitions when using C++ */
+#ifdef __cplusplus
+}
+#endif
+#include <SDL3/SDL_close_code.h>
+
+#endif /* SDL_bits_h_ */

release/macos/frameworks/SDL3.xcframework/macos-arm64_x86_64/SDL3.framework/Versions/A/Headers/SDL_blendmode.h

@@ -0,0 +1,202 @@
+/*
+  Simple DirectMedia Layer
+  Copyright (C) 1997-2025 Sam Lantinga <slouken@libsdl.org>
+
+  This software is provided 'as-is', without any express or implied
+  warranty.  In no event will the authors be held liable for any damages
+  arising from the use of this software.
+
+  Permission is granted to anyone to use this software for any purpose,
+  including commercial applications, and to alter it and redistribute it
+  freely, subject to the following restrictions:
+
+  1. The origin of this software must not be misrepresented; you must not
+     claim that you wrote the original software. If you use this software
+     in a product, an acknowledgment in the product documentation would be
+     appreciated but is not required.
+  2. Altered source versions must be plainly marked as such, and must not be
+     misrepresented as being the original software.
+  3. This notice may not be removed or altered from any source distribution.
+*/
+
+/**
+ * # CategoryBlendmode
+ *
+ * Blend modes decide how two colors will mix together. There are both
+ * standard modes for basic needs and a means to create custom modes,
+ * dictating what sort of math to do on what color components.
+ */
+
+#ifndef SDL_blendmode_h_
+#define SDL_blendmode_h_
+
+#include <SDL3/SDL_stdinc.h>
+
+#include <SDL3/SDL_begin_code.h>
+/* Set up for C function definitions, even when using C++ */
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/**
+ * A set of blend modes used in drawing operations.
+ *
+ * These predefined blend modes are supported everywhere.
+ *
+ * Additional values may be obtained from SDL_ComposeCustomBlendMode.
+ *
+ * \since This datatype is available since SDL 3.2.0.
+ *
+ * \sa SDL_ComposeCustomBlendMode
+ */
+typedef Uint32 SDL_BlendMode;
+
+#define SDL_BLENDMODE_NONE                  0x00000000u /**< no blending: dstRGBA = srcRGBA */
+#define SDL_BLENDMODE_BLEND                 0x00000001u /**< alpha blending: dstRGB = (srcRGB * srcA) + (dstRGB * (1-srcA)), dstA = srcA + (dstA * (1-srcA)) */
+#define SDL_BLENDMODE_BLEND_PREMULTIPLIED   0x00000010u /**< pre-multiplied alpha blending: dstRGBA = srcRGBA + (dstRGBA * (1-srcA)) */
+#define SDL_BLENDMODE_ADD                   0x00000002u /**< additive blending: dstRGB = (srcRGB * srcA) + dstRGB, dstA = dstA */
+#define SDL_BLENDMODE_ADD_PREMULTIPLIED     0x00000020u /**< pre-multiplied additive blending: dstRGB = srcRGB + dstRGB, dstA = dstA */
+#define SDL_BLENDMODE_MOD                   0x00000004u /**< color modulate: dstRGB = srcRGB * dstRGB, dstA = dstA */
+#define SDL_BLENDMODE_MUL                   0x00000008u /**< color multiply: dstRGB = (srcRGB * dstRGB) + (dstRGB * (1-srcA)), dstA = dstA */
+#define SDL_BLENDMODE_INVALID               0x7FFFFFFFu
+
+/**
+ * The blend operation used when combining source and destination pixel
+ * components.
+ *
+ * \since This enum is available since SDL 3.2.0.
+ */
+typedef enum SDL_BlendOperation
+{
+    SDL_BLENDOPERATION_ADD              = 0x1,  /**< dst + src: supported by all renderers */
+    SDL_BLENDOPERATION_SUBTRACT         = 0x2,  /**< src - dst : supported by D3D, OpenGL, OpenGLES, and Vulkan */
+    SDL_BLENDOPERATION_REV_SUBTRACT     = 0x3,  /**< dst - src : supported by D3D, OpenGL, OpenGLES, and Vulkan */
+    SDL_BLENDOPERATION_MINIMUM          = 0x4,  /**< min(dst, src) : supported by D3D, OpenGL, OpenGLES, and Vulkan */
+    SDL_BLENDOPERATION_MAXIMUM          = 0x5   /**< max(dst, src) : supported by D3D, OpenGL, OpenGLES, and Vulkan */
+} SDL_BlendOperation;
+
+/**
+ * The normalized factor used to multiply pixel components.
+ *
+ * The blend factors are multiplied with the pixels from a drawing operation
+ * (src) and the pixels from the render target (dst) before the blend
+ * operation. The comma-separated factors listed above are always applied in
+ * the component order red, green, blue, and alpha.
+ *
+ * \since This enum is available since SDL 3.2.0.
+ */
+typedef enum SDL_BlendFactor
+{
+    SDL_BLENDFACTOR_ZERO                = 0x1,  /**< 0, 0, 0, 0 */
+    SDL_BLENDFACTOR_ONE                 = 0x2,  /**< 1, 1, 1, 1 */
+    SDL_BLENDFACTOR_SRC_COLOR           = 0x3,  /**< srcR, srcG, srcB, srcA */
+    SDL_BLENDFACTOR_ONE_MINUS_SRC_COLOR = 0x4,  /**< 1-srcR, 1-srcG, 1-srcB, 1-srcA */
+    SDL_BLENDFACTOR_SRC_ALPHA           = 0x5,  /**< srcA, srcA, srcA, srcA */
+    SDL_BLENDFACTOR_ONE_MINUS_SRC_ALPHA = 0x6,  /**< 1-srcA, 1-srcA, 1-srcA, 1-srcA */
+    SDL_BLENDFACTOR_DST_COLOR           = 0x7,  /**< dstR, dstG, dstB, dstA */
+    SDL_BLENDFACTOR_ONE_MINUS_DST_COLOR = 0x8,  /**< 1-dstR, 1-dstG, 1-dstB, 1-dstA */
+    SDL_BLENDFACTOR_DST_ALPHA           = 0x9,  /**< dstA, dstA, dstA, dstA */
+    SDL_BLENDFACTOR_ONE_MINUS_DST_ALPHA = 0xA   /**< 1-dstA, 1-dstA, 1-dstA, 1-dstA */
+} SDL_BlendFactor;
+
+/**
+ * Compose a custom blend mode for renderers.
+ *
+ * The functions SDL_SetRenderDrawBlendMode and SDL_SetTextureBlendMode accept
+ * the SDL_BlendMode returned by this function if the renderer supports it.
+ *
+ * A blend mode controls how the pixels from a drawing operation (source) get
+ * combined with the pixels from the render target (destination). First, the
+ * components of the source and destination pixels get multiplied with their
+ * blend factors. Then, the blend operation takes the two products and
+ * calculates the result that will get stored in the render target.
+ *
+ * Expressed in pseudocode, it would look like this:
+ *
+ * ```c
+ * dstRGB = colorOperation(srcRGB * srcColorFactor, dstRGB * dstColorFactor);
+ * dstA = alphaOperation(srcA * srcAlphaFactor, dstA * dstAlphaFactor);
+ * ```
+ *
+ * Where the functions `colorOperation(src, dst)` and `alphaOperation(src,
+ * dst)` can return one of the following:
+ *
+ * - `src + dst`
+ * - `src - dst`
+ * - `dst - src`
+ * - `min(src, dst)`
+ * - `max(src, dst)`
+ *
+ * The red, green, and blue components are always multiplied with the first,
+ * second, and third components of the SDL_BlendFactor, respectively. The
+ * fourth component is not used.
+ *
+ * The alpha component is always multiplied with the fourth component of the
+ * SDL_BlendFactor. The other components are not used in the alpha
+ * calculation.
+ *
+ * Support for these blend modes varies for each renderer. To check if a
+ * specific SDL_BlendMode is supported, create a renderer and pass it to
+ * either SDL_SetRenderDrawBlendMode or SDL_SetTextureBlendMode. They will
+ * return with an error if the blend mode is not supported.
+ *
+ * This list describes the support of custom blend modes for each renderer.
+ * All renderers support the four blend modes listed in the SDL_BlendMode
+ * enumeration.
+ *
+ * - **direct3d**: Supports all operations with all factors. However, some
+ *   factors produce unexpected results with `SDL_BLENDOPERATION_MINIMUM` and
+ *   `SDL_BLENDOPERATION_MAXIMUM`.
+ * - **direct3d11**: Same as Direct3D 9.
+ * - **opengl**: Supports the `SDL_BLENDOPERATION_ADD` operation with all
+ *   factors. OpenGL versions 1.1, 1.2, and 1.3 do not work correctly here.
+ * - **opengles2**: Supports the `SDL_BLENDOPERATION_ADD`,
+ *   `SDL_BLENDOPERATION_SUBTRACT`, `SDL_BLENDOPERATION_REV_SUBTRACT`
+ *   operations with all factors.
+ * - **psp**: No custom blend mode support.
+ * - **software**: No custom blend mode support.
+ *
+ * Some renderers do not provide an alpha component for the default render
+ * target. The `SDL_BLENDFACTOR_DST_ALPHA` and
+ * `SDL_BLENDFACTOR_ONE_MINUS_DST_ALPHA` factors do not have an effect in this
+ * case.
+ *
+ * \param srcColorFactor the SDL_BlendFactor applied to the red, green, and
+ *                       blue components of the source pixels.
+ * \param dstColorFactor the SDL_BlendFactor applied to the red, green, and
+ *                       blue components of the destination pixels.
+ * \param colorOperation the SDL_BlendOperation used to combine the red,
+ *                       green, and blue components of the source and
+ *                       destination pixels.
+ * \param srcAlphaFactor the SDL_BlendFactor applied to the alpha component of
+ *                       the source pixels.
+ * \param dstAlphaFactor the SDL_BlendFactor applied to the alpha component of
+ *                       the destination pixels.
+ * \param alphaOperation the SDL_BlendOperation used to combine the alpha
+ *                       component of the source and destination pixels.
+ * \returns an SDL_BlendMode that represents the chosen factors and
+ *          operations.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_SetRenderDrawBlendMode
+ * \sa SDL_GetRenderDrawBlendMode
+ * \sa SDL_SetTextureBlendMode
+ * \sa SDL_GetTextureBlendMode
+ */
+extern SDL_DECLSPEC SDL_BlendMode SDLCALL SDL_ComposeCustomBlendMode(SDL_BlendFactor srcColorFactor,
+                                                                 SDL_BlendFactor dstColorFactor,
+                                                                 SDL_BlendOperation colorOperation,
+                                                                 SDL_BlendFactor srcAlphaFactor,
+                                                                 SDL_BlendFactor dstAlphaFactor,
+                                                                 SDL_BlendOperation alphaOperation);
+
+/* Ends C function definitions when using C++ */
+#ifdef __cplusplus
+}
+#endif
+#include <SDL3/SDL_close_code.h>
+
+#endif /* SDL_blendmode_h_ */

release/macos/frameworks/SDL3.xcframework/macos-arm64_x86_64/SDL3.framework/Versions/A/Headers/SDL_camera.h

@@ -0,0 +1,519 @@
+/*
+  Simple DirectMedia Layer
+  Copyright (C) 1997-2025 Sam Lantinga <slouken@libsdl.org>
+
+  This software is provided 'as-is', without any express or implied
+  warranty.  In no event will the authors be held liable for any damages
+  arising from the use of this software.
+
+  Permission is granted to anyone to use this software for any purpose,
+  including commercial applications, and to alter it and redistribute it
+  freely, subject to the following restrictions:
+
+  1. The origin of this software must not be misrepresented; you must not
+     claim that you wrote the original software. If you use this software
+     in a product, an acknowledgment in the product documentation would be
+     appreciated but is not required.
+  2. Altered source versions must be plainly marked as such, and must not be
+     misrepresented as being the original software.
+  3. This notice may not be removed or altered from any source distribution.
+*/
+
+/**
+ * # CategoryCamera
+ *
+ * Video capture for the SDL library.
+ *
+ * This API lets apps read input from video sources, like webcams. Camera
+ * devices can be enumerated, queried, and opened. Once opened, it will
+ * provide SDL_Surface objects as new frames of video come in. These surfaces
+ * can be uploaded to an SDL_Texture or processed as pixels in memory.
+ *
+ * Several platforms will alert the user if an app tries to access a camera,
+ * and some will present a UI asking the user if your application should be
+ * allowed to obtain images at all, which they can deny. A successfully opened
+ * camera will not provide images until permission is granted. Applications,
+ * after opening a camera device, can see if they were granted access by
+ * either polling with the SDL_GetCameraPermissionState() function, or waiting
+ * for an SDL_EVENT_CAMERA_DEVICE_APPROVED or SDL_EVENT_CAMERA_DEVICE_DENIED
+ * event. Platforms that don't have any user approval process will report
+ * approval immediately.
+ *
+ * Note that SDL cameras only provide video as individual frames; they will
+ * not provide full-motion video encoded in a movie file format, although an
+ * app is free to encode the acquired frames into any format it likes. It also
+ * does not provide audio from the camera hardware through this API; not only
+ * do many webcams not have microphones at all, many people--from streamers to
+ * people on Zoom calls--will want to use a separate microphone regardless of
+ * the camera. In any case, recorded audio will be available through SDL's
+ * audio API no matter what hardware provides the microphone.
+ *
+ * ## Camera gotchas
+ *
+ * Consumer-level camera hardware tends to take a little while to warm up,
+ * once the device has been opened. Generally most camera apps have some sort
+ * of UI to take a picture (a button to snap a pic while a preview is showing,
+ * some sort of multi-second countdown for the user to pose, like a photo
+ * booth), which puts control in the users' hands, or they are intended to
+ * stay on for long times (Pokemon Go, etc).
+ *
+ * It's not uncommon that a newly-opened camera will provide a couple of
+ * completely black frames, maybe followed by some under-exposed images. If
+ * taking a single frame automatically, or recording video from a camera's
+ * input without the user initiating it from a preview, it could be wise to
+ * drop the first several frames (if not the first several _seconds_ worth of
+ * frames!) before using images from a camera.
+ */
+
+#ifndef SDL_camera_h_
+#define SDL_camera_h_
+
+#include <SDL3/SDL_stdinc.h>
+#include <SDL3/SDL_error.h>
+#include <SDL3/SDL_pixels.h>
+#include <SDL3/SDL_properties.h>
+#include <SDL3/SDL_surface.h>
+
+#include <SDL3/SDL_begin_code.h>
+/* Set up for C function definitions, even when using C++ */
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/**
+ * This is a unique ID for a camera device for the time it is connected to the
+ * system, and is never reused for the lifetime of the application.
+ *
+ * If the device is disconnected and reconnected, it will get a new ID.
+ *
+ * The value 0 is an invalid ID.
+ *
+ * \since This datatype is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetCameras
+ */
+typedef Uint32 SDL_CameraID;
+
+/**
+ * The opaque structure used to identify an opened SDL camera.
+ *
+ * \since This struct is available since SDL 3.2.0.
+ */
+typedef struct SDL_Camera SDL_Camera;
+
+/**
+ * The details of an output format for a camera device.
+ *
+ * Cameras often support multiple formats; each one will be encapsulated in
+ * this struct.
+ *
+ * \since This struct is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetCameraSupportedFormats
+ * \sa SDL_GetCameraFormat
+ */
+typedef struct SDL_CameraSpec
+{
+    SDL_PixelFormat format;     /**< Frame format */
+    SDL_Colorspace colorspace;  /**< Frame colorspace */
+    int width;                  /**< Frame width */
+    int height;                 /**< Frame height */
+    int framerate_numerator;     /**< Frame rate numerator ((num / denom) == FPS, (denom / num) == duration in seconds) */
+    int framerate_denominator;   /**< Frame rate demoninator ((num / denom) == FPS, (denom / num) == duration in seconds) */
+} SDL_CameraSpec;
+
+/**
+ * The position of camera in relation to system device.
+ *
+ * \since This enum is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetCameraPosition
+ */
+typedef enum SDL_CameraPosition
+{
+    SDL_CAMERA_POSITION_UNKNOWN,
+    SDL_CAMERA_POSITION_FRONT_FACING,
+    SDL_CAMERA_POSITION_BACK_FACING
+} SDL_CameraPosition;
+
+
+/**
+ * Use this function to get the number of built-in camera drivers.
+ *
+ * This function returns a hardcoded number. This never returns a negative
+ * value; if there are no drivers compiled into this build of SDL, this
+ * function returns zero. The presence of a driver in this list does not mean
+ * it will function, it just means SDL is capable of interacting with that
+ * interface. For example, a build of SDL might have v4l2 support, but if
+ * there's no kernel support available, SDL's v4l2 driver would fail if used.
+ *
+ * By default, SDL tries all drivers, in its preferred order, until one is
+ * found to be usable.
+ *
+ * \returns the number of built-in camera drivers.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetCameraDriver
+ */
+extern SDL_DECLSPEC int SDLCALL SDL_GetNumCameraDrivers(void);
+
+/**
+ * Use this function to get the name of a built in camera driver.
+ *
+ * The list of camera drivers is given in the order that they are normally
+ * initialized by default; the drivers that seem more reasonable to choose
+ * first (as far as the SDL developers believe) are earlier in the list.
+ *
+ * The names of drivers are all simple, low-ASCII identifiers, like "v4l2",
+ * "coremedia" or "android". These never have Unicode characters, and are not
+ * meant to be proper names.
+ *
+ * \param index the index of the camera driver; the value ranges from 0 to
+ *              SDL_GetNumCameraDrivers() - 1.
+ * \returns the name of the camera driver at the requested index, or NULL if
+ *          an invalid index was specified.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetNumCameraDrivers
+ */
+extern SDL_DECLSPEC const char * SDLCALL SDL_GetCameraDriver(int index);
+
+/**
+ * Get the name of the current camera driver.
+ *
+ * The names of drivers are all simple, low-ASCII identifiers, like "v4l2",
+ * "coremedia" or "android". These never have Unicode characters, and are not
+ * meant to be proper names.
+ *
+ * \returns the name of the current camera driver or NULL if no driver has
+ *          been initialized.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ */
+extern SDL_DECLSPEC const char * SDLCALL SDL_GetCurrentCameraDriver(void);
+
+/**
+ * Get a list of currently connected camera devices.
+ *
+ * \param count a pointer filled in with the number of cameras returned, may
+ *              be NULL.
+ * \returns a 0 terminated array of camera instance IDs or NULL on failure;
+ *          call SDL_GetError() for more information. This should be freed
+ *          with SDL_free() when it is no longer needed.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_OpenCamera
+ */
+extern SDL_DECLSPEC SDL_CameraID * SDLCALL SDL_GetCameras(int *count);
+
+/**
+ * Get the list of native formats/sizes a camera supports.
+ *
+ * This returns a list of all formats and frame sizes that a specific camera
+ * can offer. This is useful if your app can accept a variety of image formats
+ * and sizes and so want to find the optimal spec that doesn't require
+ * conversion.
+ *
+ * This function isn't strictly required; if you call SDL_OpenCamera with a
+ * NULL spec, SDL will choose a native format for you, and if you instead
+ * specify a desired format, it will transparently convert to the requested
+ * format on your behalf.
+ *
+ * If `count` is not NULL, it will be filled with the number of elements in
+ * the returned array.
+ *
+ * Note that it's legal for a camera to supply an empty list. This is what
+ * will happen on Emscripten builds, since that platform won't tell _anything_
+ * about available cameras until you've opened one, and won't even tell if
+ * there _is_ a camera until the user has given you permission to check
+ * through a scary warning popup.
+ *
+ * \param instance_id the camera device instance ID.
+ * \param count a pointer filled in with the number of elements in the list,
+ *              may be NULL.
+ * \returns a NULL terminated array of pointers to SDL_CameraSpec or NULL on
+ *          failure; call SDL_GetError() for more information. This is a
+ *          single allocation that should be freed with SDL_free() when it is
+ *          no longer needed.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetCameras
+ * \sa SDL_OpenCamera
+ */
+extern SDL_DECLSPEC SDL_CameraSpec ** SDLCALL SDL_GetCameraSupportedFormats(SDL_CameraID instance_id, int *count);
+
+/**
+ * Get the human-readable device name for a camera.
+ *
+ * \param instance_id the camera device instance ID.
+ * \returns a human-readable device name or NULL on failure; call
+ *          SDL_GetError() for more information.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetCameras
+ */
+extern SDL_DECLSPEC const char * SDLCALL SDL_GetCameraName(SDL_CameraID instance_id);
+
+/**
+ * Get the position of the camera in relation to the system.
+ *
+ * Most platforms will report UNKNOWN, but mobile devices, like phones, can
+ * often make a distinction between cameras on the front of the device (that
+ * points towards the user, for taking "selfies") and cameras on the back (for
+ * filming in the direction the user is facing).
+ *
+ * \param instance_id the camera device instance ID.
+ * \returns the position of the camera on the system hardware.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetCameras
+ */
+extern SDL_DECLSPEC SDL_CameraPosition SDLCALL SDL_GetCameraPosition(SDL_CameraID instance_id);
+
+/**
+ * Open a video recording device (a "camera").
+ *
+ * You can open the device with any reasonable spec, and if the hardware can't
+ * directly support it, it will convert data seamlessly to the requested
+ * format. This might incur overhead, including scaling of image data.
+ *
+ * If you would rather accept whatever format the device offers, you can pass
+ * a NULL spec here and it will choose one for you (and you can use
+ * SDL_Surface's conversion/scaling functions directly if necessary).
+ *
+ * You can call SDL_GetCameraFormat() to get the actual data format if passing
+ * a NULL spec here. You can see the exact specs a device can support without
+ * conversion with SDL_GetCameraSupportedFormats().
+ *
+ * SDL will not attempt to emulate framerate; it will try to set the hardware
+ * to the rate closest to the requested speed, but it won't attempt to limit
+ * or duplicate frames artificially; call SDL_GetCameraFormat() to see the
+ * actual framerate of the opened the device, and check your timestamps if
+ * this is crucial to your app!
+ *
+ * Note that the camera is not usable until the user approves its use! On some
+ * platforms, the operating system will prompt the user to permit access to
+ * the camera, and they can choose Yes or No at that point. Until they do, the
+ * camera will not be usable. The app should either wait for an
+ * SDL_EVENT_CAMERA_DEVICE_APPROVED (or SDL_EVENT_CAMERA_DEVICE_DENIED) event,
+ * or poll SDL_GetCameraPermissionState() occasionally until it returns
+ * non-zero. On platforms that don't require explicit user approval (and
+ * perhaps in places where the user previously permitted access), the approval
+ * event might come immediately, but it might come seconds, minutes, or hours
+ * later!
+ *
+ * \param instance_id the camera device instance ID.
+ * \param spec the desired format for data the device will provide. Can be
+ *             NULL.
+ * \returns an SDL_Camera object or NULL on failure; call SDL_GetError() for
+ *          more information.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetCameras
+ * \sa SDL_GetCameraFormat
+ */
+extern SDL_DECLSPEC SDL_Camera * SDLCALL SDL_OpenCamera(SDL_CameraID instance_id, const SDL_CameraSpec *spec);
+
+/**
+ * Query if camera access has been approved by the user.
+ *
+ * Cameras will not function between when the device is opened by the app and
+ * when the user permits access to the hardware. On some platforms, this
+ * presents as a popup dialog where the user has to explicitly approve access;
+ * on others the approval might be implicit and not alert the user at all.
+ *
+ * This function can be used to check the status of that approval. It will
+ * return 0 if still waiting for user response, 1 if the camera is approved
+ * for use, and -1 if the user denied access.
+ *
+ * Instead of polling with this function, you can wait for a
+ * SDL_EVENT_CAMERA_DEVICE_APPROVED (or SDL_EVENT_CAMERA_DEVICE_DENIED) event
+ * in the standard SDL event loop, which is guaranteed to be sent once when
+ * permission to use the camera is decided.
+ *
+ * If a camera is declined, there's nothing to be done but call
+ * SDL_CloseCamera() to dispose of it.
+ *
+ * \param camera the opened camera device to query.
+ * \returns -1 if user denied access to the camera, 1 if user approved access,
+ *          0 if no decision has been made yet.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_OpenCamera
+ * \sa SDL_CloseCamera
+ */
+extern SDL_DECLSPEC int SDLCALL SDL_GetCameraPermissionState(SDL_Camera *camera);
+
+/**
+ * Get the instance ID of an opened camera.
+ *
+ * \param camera an SDL_Camera to query.
+ * \returns the instance ID of the specified camera on success or 0 on
+ *          failure; call SDL_GetError() for more information.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_OpenCamera
+ */
+extern SDL_DECLSPEC SDL_CameraID SDLCALL SDL_GetCameraID(SDL_Camera *camera);
+
+/**
+ * Get the properties associated with an opened camera.
+ *
+ * \param camera the SDL_Camera obtained from SDL_OpenCamera().
+ * \returns a valid property ID on success or 0 on failure; call
+ *          SDL_GetError() for more information.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ */
+extern SDL_DECLSPEC SDL_PropertiesID SDLCALL SDL_GetCameraProperties(SDL_Camera *camera);
+
+/**
+ * Get the spec that a camera is using when generating images.
+ *
+ * Note that this might not be the native format of the hardware, as SDL might
+ * be converting to this format behind the scenes.
+ *
+ * If the system is waiting for the user to approve access to the camera, as
+ * some platforms require, this will return false, but this isn't necessarily
+ * a fatal error; you should either wait for an
+ * SDL_EVENT_CAMERA_DEVICE_APPROVED (or SDL_EVENT_CAMERA_DEVICE_DENIED) event,
+ * or poll SDL_GetCameraPermissionState() occasionally until it returns
+ * non-zero.
+ *
+ * \param camera opened camera device.
+ * \param spec the SDL_CameraSpec to be initialized by this function.
+ * \returns true on success or false on failure; call SDL_GetError() for more
+ *          information.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_OpenCamera
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_GetCameraFormat(SDL_Camera *camera, SDL_CameraSpec *spec);
+
+/**
+ * Acquire a frame.
+ *
+ * The frame is a memory pointer to the image data, whose size and format are
+ * given by the spec requested when opening the device.
+ *
+ * This is a non blocking API. If there is a frame available, a non-NULL
+ * surface is returned, and timestampNS will be filled with a non-zero value.
+ *
+ * Note that an error case can also return NULL, but a NULL by itself is
+ * normal and just signifies that a new frame is not yet available. Note that
+ * even if a camera device fails outright (a USB camera is unplugged while in
+ * use, etc), SDL will send an event separately to notify the app, but
+ * continue to provide blank frames at ongoing intervals until
+ * SDL_CloseCamera() is called, so real failure here is almost always an out
+ * of memory condition.
+ *
+ * After use, the frame should be released with SDL_ReleaseCameraFrame(). If
+ * you don't do this, the system may stop providing more video!
+ *
+ * Do not call SDL_DestroySurface() on the returned surface! It must be given
+ * back to the camera subsystem with SDL_ReleaseCameraFrame!
+ *
+ * If the system is waiting for the user to approve access to the camera, as
+ * some platforms require, this will return NULL (no frames available); you
+ * should either wait for an SDL_EVENT_CAMERA_DEVICE_APPROVED (or
+ * SDL_EVENT_CAMERA_DEVICE_DENIED) event, or poll
+ * SDL_GetCameraPermissionState() occasionally until it returns non-zero.
+ *
+ * \param camera opened camera device.
+ * \param timestampNS a pointer filled in with the frame's timestamp, or 0 on
+ *                    error. Can be NULL.
+ * \returns a new frame of video on success, NULL if none is currently
+ *          available.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_ReleaseCameraFrame
+ */
+extern SDL_DECLSPEC SDL_Surface * SDLCALL SDL_AcquireCameraFrame(SDL_Camera *camera, Uint64 *timestampNS);
+
+/**
+ * Release a frame of video acquired from a camera.
+ *
+ * Let the back-end re-use the internal buffer for camera.
+ *
+ * This function _must_ be called only on surface objects returned by
+ * SDL_AcquireCameraFrame(). This function should be called as quickly as
+ * possible after acquisition, as SDL keeps a small FIFO queue of surfaces for
+ * video frames; if surfaces aren't released in a timely manner, SDL may drop
+ * upcoming video frames from the camera.
+ *
+ * If the app needs to keep the surface for a significant time, they should
+ * make a copy of it and release the original.
+ *
+ * The app should not use the surface again after calling this function;
+ * assume the surface is freed and the pointer is invalid.
+ *
+ * \param camera opened camera device.
+ * \param frame the video frame surface to release.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_AcquireCameraFrame
+ */
+extern SDL_DECLSPEC void SDLCALL SDL_ReleaseCameraFrame(SDL_Camera *camera, SDL_Surface *frame);
+
+/**
+ * Use this function to shut down camera processing and close the camera
+ * device.
+ *
+ * \param camera opened camera device.
+ *
+ * \threadsafety It is safe to call this function from any thread, but no
+ *               thread may reference `device` once this function is called.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_OpenCamera
+ */
+extern SDL_DECLSPEC void SDLCALL SDL_CloseCamera(SDL_Camera *camera);
+
+/* Ends C function definitions when using C++ */
+#ifdef __cplusplus
+}
+#endif
+#include <SDL3/SDL_close_code.h>
+
+#endif /* SDL_camera_h_ */

release/macos/frameworks/SDL3.xcframework/macos-arm64_x86_64/SDL3.framework/Versions/A/Headers/SDL_clipboard.h

@@ -0,0 +1,331 @@
+/*
+  Simple DirectMedia Layer
+  Copyright (C) 1997-2025 Sam Lantinga <slouken@libsdl.org>
+
+  This software is provided 'as-is', without any express or implied
+  warranty.  In no event will the authors be held liable for any damages
+  arising from the use of this software.
+
+  Permission is granted to anyone to use this software for any purpose,
+  including commercial applications, and to alter it and redistribute it
+  freely, subject to the following restrictions:
+
+  1. The origin of this software must not be misrepresented; you must not
+     claim that you wrote the original software. If you use this software
+     in a product, an acknowledgment in the product documentation would be
+     appreciated but is not required.
+  2. Altered source versions must be plainly marked as such, and must not be
+     misrepresented as being the original software.
+  3. This notice may not be removed or altered from any source distribution.
+*/
+
+/**
+ * # CategoryClipboard
+ *
+ * SDL provides access to the system clipboard, both for reading information
+ * from other processes and publishing information of its own.
+ *
+ * This is not just text! SDL apps can access and publish data by mimetype.
+ *
+ * ## Basic use (text)
+ *
+ * Obtaining and publishing simple text to the system clipboard is as easy as
+ * calling SDL_GetClipboardText() and SDL_SetClipboardText(), respectively.
+ * These deal with C strings in UTF-8 encoding. Data transmission and encoding
+ * conversion is completely managed by SDL.
+ *
+ * ## Clipboard callbacks (data other than text)
+ *
+ * Things get more complicated when the clipboard contains something other
+ * than text. Not only can the system clipboard contain data of any type, in
+ * some cases it can contain the same data in different formats! For example,
+ * an image painting app might let the user copy a graphic to the clipboard,
+ * and offers it in .BMP, .JPG, or .PNG format for other apps to consume.
+ *
+ * Obtaining clipboard data ("pasting") like this is a matter of calling
+ * SDL_GetClipboardData() and telling it the mimetype of the data you want.
+ * But how does one know if that format is available? SDL_HasClipboardData()
+ * can report if a specific mimetype is offered, and
+ * SDL_GetClipboardMimeTypes() can provide the entire list of mimetypes
+ * available, so the app can decide what to do with the data and what formats
+ * it can support.
+ *
+ * Setting the clipboard ("copying") to arbitrary data is done with
+ * SDL_SetClipboardData. The app does not provide the data in this call, but
+ * rather the mimetypes it is willing to provide and a callback function.
+ * During the callback, the app will generate the data. This allows massive
+ * data sets to be provided to the clipboard, without any data being copied
+ * before it is explicitly requested. More specifically, it allows an app to
+ * offer data in multiple formats without providing a copy of all of them
+ * upfront. If the app has an image that it could provide in PNG or JPG
+ * format, it doesn't have to encode it to either of those unless and until
+ * something tries to paste it.
+ *
+ * ## Primary Selection
+ *
+ * The X11 and Wayland video targets have a concept of the "primary selection"
+ * in addition to the usual clipboard. This is generally highlighted (but not
+ * explicitly copied) text from various apps. SDL offers APIs for this through
+ * SDL_GetPrimarySelectionText() and SDL_SetPrimarySelectionText(). SDL offers
+ * these APIs on platforms without this concept, too, but only so far that it
+ * will keep a copy of a string that the app sets for later retrieval; the
+ * operating system will not ever attempt to change the string externally if
+ * it doesn't support a primary selection.
+ */
+
+#ifndef SDL_clipboard_h_
+#define SDL_clipboard_h_
+
+#include <SDL3/SDL_stdinc.h>
+#include <SDL3/SDL_error.h>
+
+#include <SDL3/SDL_begin_code.h>
+/* Set up for C function definitions, even when using C++ */
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/* Function prototypes */
+
+/**
+ * Put UTF-8 text into the clipboard.
+ *
+ * \param text the text to store in the clipboard.
+ * \returns true on success or false on failure; call SDL_GetError() for more
+ *          information.
+ *
+ * \threadsafety This function should only be called on the main thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetClipboardText
+ * \sa SDL_HasClipboardText
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_SetClipboardText(const char *text);
+
+/**
+ * Get UTF-8 text from the clipboard.
+ *
+ * This function returns an empty string if there is not enough memory left
+ * for a copy of the clipboard's content.
+ *
+ * \returns the clipboard text on success or an empty string on failure; call
+ *          SDL_GetError() for more information. This should be freed with
+ *          SDL_free() when it is no longer needed.
+ *
+ * \threadsafety This function should only be called on the main thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_HasClipboardText
+ * \sa SDL_SetClipboardText
+ */
+extern SDL_DECLSPEC char * SDLCALL SDL_GetClipboardText(void);
+
+/**
+ * Query whether the clipboard exists and contains a non-empty text string.
+ *
+ * \returns true if the clipboard has text, or false if it does not.
+ *
+ * \threadsafety This function should only be called on the main thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetClipboardText
+ * \sa SDL_SetClipboardText
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_HasClipboardText(void);
+
+/**
+ * Put UTF-8 text into the primary selection.
+ *
+ * \param text the text to store in the primary selection.
+ * \returns true on success or false on failure; call SDL_GetError() for more
+ *          information.
+ *
+ * \threadsafety This function should only be called on the main thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetPrimarySelectionText
+ * \sa SDL_HasPrimarySelectionText
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_SetPrimarySelectionText(const char *text);
+
+/**
+ * Get UTF-8 text from the primary selection.
+ *
+ * This function returns an empty string if there is not enough memory left
+ * for a copy of the primary selection's content.
+ *
+ * \returns the primary selection text on success or an empty string on
+ *          failure; call SDL_GetError() for more information. This should be
+ *          freed with SDL_free() when it is no longer needed.
+ *
+ * \threadsafety This function should only be called on the main thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_HasPrimarySelectionText
+ * \sa SDL_SetPrimarySelectionText
+ */
+extern SDL_DECLSPEC char * SDLCALL SDL_GetPrimarySelectionText(void);
+
+/**
+ * Query whether the primary selection exists and contains a non-empty text
+ * string.
+ *
+ * \returns true if the primary selection has text, or false if it does not.
+ *
+ * \threadsafety This function should only be called on the main thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_GetPrimarySelectionText
+ * \sa SDL_SetPrimarySelectionText
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_HasPrimarySelectionText(void);
+
+/**
+ * Callback function that will be called when data for the specified mime-type
+ * is requested by the OS.
+ *
+ * The callback function is called with NULL as the mime_type when the
+ * clipboard is cleared or new data is set. The clipboard is automatically
+ * cleared in SDL_Quit().
+ *
+ * \param userdata a pointer to the provided user data.
+ * \param mime_type the requested mime-type.
+ * \param size a pointer filled in with the length of the returned data.
+ * \returns a pointer to the data for the provided mime-type. Returning NULL
+ *          or setting the length to 0 will cause no data to be sent to the
+ *          "receiver". It is up to the receiver to handle this. Essentially
+ *          returning no data is more or less undefined behavior and may cause
+ *          breakage in receiving applications. The returned data will not be
+ *          freed, so it needs to be retained and dealt with internally.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_SetClipboardData
+ */
+typedef const void *(SDLCALL *SDL_ClipboardDataCallback)(void *userdata, const char *mime_type, size_t *size);
+
+/**
+ * Callback function that will be called when the clipboard is cleared, or when new
+ * data is set.
+ *
+ * \param userdata a pointer to the provided user data.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_SetClipboardData
+ */
+typedef void (SDLCALL *SDL_ClipboardCleanupCallback)(void *userdata);
+
+/**
+ * Offer clipboard data to the OS.
+ *
+ * Tell the operating system that the application is offering clipboard data
+ * for each of the provided mime-types. Once another application requests the
+ * data the callback function will be called, allowing it to generate and
+ * respond with the data for the requested mime-type.
+ *
+ * The size of text data does not include any terminator, and the text does
+ * not need to be null-terminated (e.g., you can directly copy a portion of a
+ * document).
+ *
+ * \param callback a function pointer to the function that provides the
+ *                 clipboard data.
+ * \param cleanup a function pointer to the function that cleans up the
+ *                clipboard data.
+ * \param userdata an opaque pointer that will be forwarded to the callbacks.
+ * \param mime_types a list of mime-types that are being offered. SDL copies the given list.
+ * \param num_mime_types the number of mime-types in the mime_types list.
+ * \returns true on success or false on failure; call SDL_GetError() for more
+ *          information.
+ *
+ * \threadsafety This function should only be called on the main thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_ClearClipboardData
+ * \sa SDL_GetClipboardData
+ * \sa SDL_HasClipboardData
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_SetClipboardData(SDL_ClipboardDataCallback callback, SDL_ClipboardCleanupCallback cleanup, void *userdata, const char **mime_types, size_t num_mime_types);
+
+/**
+ * Clear the clipboard data.
+ *
+ * \returns true on success or false on failure; call SDL_GetError() for more
+ *          information.
+ *
+ * \threadsafety This function should only be called on the main thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_SetClipboardData
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_ClearClipboardData(void);
+
+/**
+ * Get the data from the clipboard for a given mime type.
+ *
+ * The size of text data does not include the terminator, but the text is
+ * guaranteed to be null-terminated.
+ *
+ * \param mime_type the mime type to read from the clipboard.
+ * \param size a pointer filled in with the length of the returned data.
+ * \returns the retrieved data buffer or NULL on failure; call SDL_GetError()
+ *          for more information. This should be freed with SDL_free() when it
+ *          is no longer needed.
+ *
+ * \threadsafety This function should only be called on the main thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_HasClipboardData
+ * \sa SDL_SetClipboardData
+ */
+extern SDL_DECLSPEC void * SDLCALL SDL_GetClipboardData(const char *mime_type, size_t *size);
+
+/**
+ * Query whether there is data in the clipboard for the provided mime type.
+ *
+ * \param mime_type the mime type to check for data.
+ * \returns true if data exists in the clipboard for the provided mime type,
+ *          false if it does not.
+ *
+ * \threadsafety This function should only be called on the main thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_SetClipboardData
+ * \sa SDL_GetClipboardData
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_HasClipboardData(const char *mime_type);
+
+/**
+ * Retrieve the list of mime types available in the clipboard.
+ *
+ * \param num_mime_types a pointer filled with the number of mime types, may
+ *                       be NULL.
+ * \returns a null-terminated array of strings with mime types, or NULL on
+ *          failure; call SDL_GetError() for more information. This should be
+ *          freed with SDL_free() when it is no longer needed.
+ *
+ * \threadsafety This function should only be called on the main thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_SetClipboardData
+ */
+extern SDL_DECLSPEC char ** SDLCALL SDL_GetClipboardMimeTypes(size_t *num_mime_types);
+
+/* Ends C function definitions when using C++ */
+#ifdef __cplusplus
+}
+#endif
+#include <SDL3/SDL_close_code.h>
+
+#endif /* SDL_clipboard_h_ */

release/macos/frameworks/SDL3.xcframework/macos-arm64_x86_64/SDL3.framework/Versions/A/Headers/SDL_close_code.h

@@ -0,0 +1,41 @@
+/*
+  Simple DirectMedia Layer
+  Copyright (C) 1997-2025 Sam Lantinga <slouken@libsdl.org>
+
+  This software is provided 'as-is', without any express or implied
+  warranty.  In no event will the authors be held liable for any damages
+  arising from the use of this software.
+
+  Permission is granted to anyone to use this software for any purpose,
+  including commercial applications, and to alter it and redistribute it
+  freely, subject to the following restrictions:
+
+  1. The origin of this software must not be misrepresented; you must not
+     claim that you wrote the original software. If you use this software
+     in a product, an acknowledgment in the product documentation would be
+     appreciated but is not required.
+  2. Altered source versions must be plainly marked as such, and must not be
+     misrepresented as being the original software.
+  3. This notice may not be removed or altered from any source distribution.
+*/
+
+/*
+ * This file reverses the effects of SDL_begin_code.h and should be included
+ * after you finish any function and structure declarations in your headers.
+ *
+ * SDL's headers use this; applications generally should not include this
+ * header directly.
+ */
+
+#ifndef SDL_begin_code_h
+#error SDL_close_code.h included without matching SDL_begin_code.h
+#endif
+#undef SDL_begin_code_h
+
+/* Reset structure packing at previous byte alignment */
+#if defined(_MSC_VER) || defined(__MWERKS__) || defined(__BORLANDC__)
+#ifdef __BORLANDC__
+#pragma nopackwarning
+#endif
+#pragma pack(pop)
+#endif /* Compiler needs structure packing set */

release/macos/frameworks/SDL3.xcframework/macos-arm64_x86_64/SDL3.framework/Versions/A/Headers/SDL_copying.h

@@ -0,0 +1,22 @@
+/*
+  Simple DirectMedia Layer
+  Copyright (C) 1997-2025 Sam Lantinga <slouken@libsdl.org>
+
+  This software is provided 'as-is', without any express or implied
+  warranty.  In no event will the authors be held liable for any damages
+  arising from the use of this software.
+
+  Permission is granted to anyone to use this software for any purpose,
+  including commercial applications, and to alter it and redistribute it
+  freely, subject to the following restrictions:
+
+  1. The origin of this software must not be misrepresented; you must not
+     claim that you wrote the original software. If you use this software
+     in a product, an acknowledgment in the product documentation would be
+     appreciated but is not required.
+  2. Altered source versions must be plainly marked as such, and must not be
+     misrepresented as being the original software.
+  3. This notice may not be removed or altered from any source distribution.
+*/
+
+/* Header file containing SDL's license. */

release/macos/frameworks/SDL3.xcframework/macos-arm64_x86_64/SDL3.framework/Versions/A/Headers/SDL_cpuinfo.h

@@ -0,0 +1,353 @@
+/*
+  Simple DirectMedia Layer
+  Copyright (C) 1997-2025 Sam Lantinga <slouken@libsdl.org>
+
+  This software is provided 'as-is', without any express or implied
+  warranty.  In no event will the authors be held liable for any damages
+  arising from the use of this software.
+
+  Permission is granted to anyone to use this software for any purpose,
+  including commercial applications, and to alter it and redistribute it
+  freely, subject to the following restrictions:
+
+  1. The origin of this software must not be misrepresented; you must not
+     claim that you wrote the original software. If you use this software
+     in a product, an acknowledgment in the product documentation would be
+     appreciated but is not required.
+  2. Altered source versions must be plainly marked as such, and must not be
+     misrepresented as being the original software.
+  3. This notice may not be removed or altered from any source distribution.
+*/
+
+/* WIKI CATEGORY: CPUInfo */
+
+/**
+ * # CategoryCPUInfo
+ *
+ * CPU feature detection for SDL.
+ *
+ * These functions are largely concerned with reporting if the system has
+ * access to various SIMD instruction sets, but also has other important info
+ * to share, such as system RAM size and number of logical CPU cores.
+ *
+ * CPU instruction set checks, like SDL_HasSSE() and SDL_HasNEON(), are
+ * available on all platforms, even if they don't make sense (an ARM processor
+ * will never have SSE and an x86 processor will never have NEON, for example,
+ * but these functions still exist and will simply return false in these
+ * cases).
+ */
+
+#ifndef SDL_cpuinfo_h_
+#define SDL_cpuinfo_h_
+
+#include <SDL3/SDL_stdinc.h>
+
+#include <SDL3/SDL_begin_code.h>
+/* Set up for C function definitions, even when using C++ */
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/**
+ * A guess for the cacheline size used for padding.
+ *
+ * Most x86 processors have a 64 byte cache line. The 64-bit PowerPC
+ * processors have a 128 byte cache line. We use the larger value to be
+ * generally safe.
+ *
+ * \since This macro is available since SDL 3.2.0.
+ */
+#define SDL_CACHELINE_SIZE  128
+
+/**
+ * Get the number of logical CPU cores available.
+ *
+ * \returns the total number of logical CPU cores. On CPUs that include
+ *          technologies such as hyperthreading, the number of logical cores
+ *          may be more than the number of physical cores.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ */
+extern SDL_DECLSPEC int SDLCALL SDL_GetNumLogicalCPUCores(void);
+
+/**
+ * Determine the L1 cache line size of the CPU.
+ *
+ * This is useful for determining multi-threaded structure padding or SIMD
+ * prefetch sizes.
+ *
+ * \returns the L1 cache line size of the CPU, in bytes.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ */
+extern SDL_DECLSPEC int SDLCALL SDL_GetCPUCacheLineSize(void);
+
+/**
+ * Determine whether the CPU has AltiVec features.
+ *
+ * This always returns false on CPUs that aren't using PowerPC instruction
+ * sets.
+ *
+ * \returns true if the CPU has AltiVec features or false if not.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_HasAltiVec(void);
+
+/**
+ * Determine whether the CPU has MMX features.
+ *
+ * This always returns false on CPUs that aren't using Intel instruction sets.
+ *
+ * \returns true if the CPU has MMX features or false if not.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_HasMMX(void);
+
+/**
+ * Determine whether the CPU has SSE features.
+ *
+ * This always returns false on CPUs that aren't using Intel instruction sets.
+ *
+ * \returns true if the CPU has SSE features or false if not.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_HasSSE2
+ * \sa SDL_HasSSE3
+ * \sa SDL_HasSSE41
+ * \sa SDL_HasSSE42
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_HasSSE(void);
+
+/**
+ * Determine whether the CPU has SSE2 features.
+ *
+ * This always returns false on CPUs that aren't using Intel instruction sets.
+ *
+ * \returns true if the CPU has SSE2 features or false if not.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_HasSSE
+ * \sa SDL_HasSSE3
+ * \sa SDL_HasSSE41
+ * \sa SDL_HasSSE42
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_HasSSE2(void);
+
+/**
+ * Determine whether the CPU has SSE3 features.
+ *
+ * This always returns false on CPUs that aren't using Intel instruction sets.
+ *
+ * \returns true if the CPU has SSE3 features or false if not.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_HasSSE
+ * \sa SDL_HasSSE2
+ * \sa SDL_HasSSE41
+ * \sa SDL_HasSSE42
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_HasSSE3(void);
+
+/**
+ * Determine whether the CPU has SSE4.1 features.
+ *
+ * This always returns false on CPUs that aren't using Intel instruction sets.
+ *
+ * \returns true if the CPU has SSE4.1 features or false if not.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_HasSSE
+ * \sa SDL_HasSSE2
+ * \sa SDL_HasSSE3
+ * \sa SDL_HasSSE42
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_HasSSE41(void);
+
+/**
+ * Determine whether the CPU has SSE4.2 features.
+ *
+ * This always returns false on CPUs that aren't using Intel instruction sets.
+ *
+ * \returns true if the CPU has SSE4.2 features or false if not.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_HasSSE
+ * \sa SDL_HasSSE2
+ * \sa SDL_HasSSE3
+ * \sa SDL_HasSSE41
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_HasSSE42(void);
+
+/**
+ * Determine whether the CPU has AVX features.
+ *
+ * This always returns false on CPUs that aren't using Intel instruction sets.
+ *
+ * \returns true if the CPU has AVX features or false if not.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_HasAVX2
+ * \sa SDL_HasAVX512F
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_HasAVX(void);
+
+/**
+ * Determine whether the CPU has AVX2 features.
+ *
+ * This always returns false on CPUs that aren't using Intel instruction sets.
+ *
+ * \returns true if the CPU has AVX2 features or false if not.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_HasAVX
+ * \sa SDL_HasAVX512F
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_HasAVX2(void);
+
+/**
+ * Determine whether the CPU has AVX-512F (foundation) features.
+ *
+ * This always returns false on CPUs that aren't using Intel instruction sets.
+ *
+ * \returns true if the CPU has AVX-512F features or false if not.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_HasAVX
+ * \sa SDL_HasAVX2
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_HasAVX512F(void);
+
+/**
+ * Determine whether the CPU has ARM SIMD (ARMv6) features.
+ *
+ * This is different from ARM NEON, which is a different instruction set.
+ *
+ * This always returns false on CPUs that aren't using ARM instruction sets.
+ *
+ * \returns true if the CPU has ARM SIMD features or false if not.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_HasNEON
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_HasARMSIMD(void);
+
+/**
+ * Determine whether the CPU has NEON (ARM SIMD) features.
+ *
+ * This always returns false on CPUs that aren't using ARM instruction sets.
+ *
+ * \returns true if the CPU has ARM NEON features or false if not.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_HasNEON(void);
+
+/**
+ * Determine whether the CPU has LSX (LOONGARCH SIMD) features.
+ *
+ * This always returns false on CPUs that aren't using LOONGARCH instruction
+ * sets.
+ *
+ * \returns true if the CPU has LOONGARCH LSX features or false if not.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_HasLSX(void);
+
+/**
+ * Determine whether the CPU has LASX (LOONGARCH SIMD) features.
+ *
+ * This always returns false on CPUs that aren't using LOONGARCH instruction
+ * sets.
+ *
+ * \returns true if the CPU has LOONGARCH LASX features or false if not.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ */
+extern SDL_DECLSPEC bool SDLCALL SDL_HasLASX(void);
+
+/**
+ * Get the amount of RAM configured in the system.
+ *
+ * \returns the amount of RAM configured in the system in MiB.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ */
+extern SDL_DECLSPEC int SDLCALL SDL_GetSystemRAM(void);
+
+/**
+ * Report the alignment this system needs for SIMD allocations.
+ *
+ * This will return the minimum number of bytes to which a pointer must be
+ * aligned to be compatible with SIMD instructions on the current machine. For
+ * example, if the machine supports SSE only, it will return 16, but if it
+ * supports AVX-512F, it'll return 64 (etc). This only reports values for
+ * instruction sets SDL knows about, so if your SDL build doesn't have
+ * SDL_HasAVX512F(), then it might return 16 for the SSE support it sees and
+ * not 64 for the AVX-512 instructions that exist but SDL doesn't know about.
+ * Plan accordingly.
+ *
+ * \returns the alignment in bytes needed for available, known SIMD
+ *          instructions.
+ *
+ * \threadsafety It is safe to call this function from any thread.
+ *
+ * \since This function is available since SDL 3.2.0.
+ *
+ * \sa SDL_aligned_alloc
+ * \sa SDL_aligned_free
+ */
+extern SDL_DECLSPEC size_t SDLCALL SDL_GetSIMDAlignment(void);
+
+/* Ends C function definitions when using C++ */
+#ifdef __cplusplus
+}
+#endif
+#include <SDL3/SDL_close_code.h>
+
+#endif /* SDL_cpuinfo_h_ */