opt: Implement build stripping and platform-specific size optimizations

Adds a 'STRIP_ALL' mode to minimize binary size for the final build, and refines size optimization flags for macOS.

- **STRIP_ALL Mode**: Added a 'DEMO_STRIP_ALL' CMake option that defines 'STRIP_ALL'. In this mode, command-line parsing is bypassed (forcing fullscreen), debug labels/error callbacks are removed from WebGPU, and non-essential code (like iostream) is stripped.

- **macOS Optimizations**: Updated CMake to use '-dead_strip' instead of GNU '--gc-sections' on Apple platforms to resolve linker errors and improve dead code elimination.

- **Documentation**: Updated HOWTO.md to document the new 'Final / Strip Build' process and FETCH_DEPS.md for optimized wgpu-native build guidance.

- **Task 7 & 8**: Marks these tasks as completed in PROJECT_CONTEXT.md.

1 files changed, 33 insertions, 1 deletions

diff --git a/src/gpu/gpu.cc b/src/gpu/gpu.cc
index a1da2db..9eb2437 100644
--- a/src/gpu/gpu.cc
+++ b/src/gpu/gpu.cc
@@ -8,9 +8,12 @@
 #include <algorithm>
 #include <cassert>
 #include <cstring>
-#include <iostream>
 #include <vector>
 
+#ifndef STRIP_ALL
+#include <iostream>
+#endif
+
 static WGPUInstance g_instance = nullptr;
 static WGPUAdapter g_adapter = nullptr;
 static WGPUDevice g_device = nullptr;
@@ -23,10 +26,16 @@ static WGPUBuffer g_uniform_buffer = nullptr;
 static WGPUBindGroup g_bind_group = nullptr;
 
 static WGPUStringView str_view(const char *str) {
+#ifndef STRIP_ALL
   if (!str) return {nullptr, 0};
   return {str, strlen(str)};
+#else
+  (void)str;
+  return {nullptr, 0};
+#endif
 }
 
+#ifndef STRIP_ALL
 static void handle_request_adapter(WGPURequestAdapterStatus status,
                                    WGPUAdapter adapter, WGPUStringView message,
                                    void *userdata1, void *userdata2) {
@@ -36,7 +45,17 @@ static void handle_request_adapter(WGPURequestAdapterStatus status,
     printf("Request adapter failed: %.*s\n", (int)message.length, message.data);
   }
 }
+#else
+static void handle_request_adapter(WGPURequestAdapterStatus status,
+                                   WGPUAdapter adapter, WGPUStringView message,
+                                   void *userdata1, void *userdata2) {
+  if (status == WGPURequestAdapterStatus_Success) {
+    *((WGPUAdapter *)userdata1) = adapter;
+  }
+}
+#endif
 
+#ifndef STRIP_ALL
 static void handle_request_device(WGPURequestDeviceStatus status,
                                   WGPUDevice device, WGPUStringView message,
                                   void *userdata1, void *userdata2) {
@@ -46,12 +65,23 @@ static void handle_request_device(WGPURequestDeviceStatus status,
     printf("Request device failed: %.*s\n", (int)message.length, message.data);
   }
 }
+#else
+static void handle_request_device(WGPURequestDeviceStatus status,
+                                  WGPUDevice device, WGPUStringView message,
+                                  void *userdata1, void *userdata2) {
+  if (status == WGPURequestDeviceStatus_Success) {
+    *((WGPUDevice *)userdata1) = device;
+  }
+}
+#endif
 
+#ifndef STRIP_ALL
 static void handle_device_error(WGPUDevice const *device, WGPUErrorType type,
                                 WGPUStringView message, void *userdata1,
                                 void *userdata2) {
   printf("WebGPU Error: %.*s\n", (int)message.length, message.data);
 }
+#endif
 
 const char *shader_wgsl_code = R"(
 struct Uniforms {
@@ -116,7 +146,9 @@ void gpu_init(GLFWwindow *window) {
 
   WGPUDeviceDescriptor device_desc = {};
   device_desc.label = str_view("Demo Device");
+#ifndef STRIP_ALL
   device_desc.uncapturedErrorCallbackInfo.callback = handle_device_error;
+#endif
 
   wgpuAdapterRequestDevice(
       g_adapter, &device_desc,