refactor: ShaderManager 引用计数句柄——shader 生命周期自动化

- ShaderObject 值类型 → ShaderRef（shared_ptr 语义，UE TShaderRef
  对标）：拷贝共享所有权，最后一个引用析构时自动销毁 VkShaderEXT，
  持有方零清理代码——"新 pass 忘销毁"整类 bug 被机制消灭
- 删除公开 Destroy 与全部持有方手工 destroyShaderEXT（8 处）；
  各类析构只清 sampler/layout/set 等其余原生句柄
- RecordFullscreen 参数收窄为 const ShaderRef&（裸 handle 无法
  再被误传）；bind 点统一 GetHandle()
- Outline/SelectionMask 撤掉上一轮补的 Device&（不再需要）
- 验证：5s 探针自动退出 → exit 0、无 VUID-05137 泄漏、无 SIGSEGV

Engine/src/Renderer/IblGenerator.cpp

@@ -78,19 +78,19 @@ namespace Engine
         m_ToCubeShader = shaders.CreateFromFile(
             dir + "ibl_equirect_to_cube.comp.spv",
             vk::ShaderStageFlagBits::eCompute,
-            {m_IblSetLayout}).Handle;
+            {m_IblSetLayout});
         m_IrrShader = shaders.CreateFromFile(
             dir + "ibl_irradiance.comp.spv",
             vk::ShaderStageFlagBits::eCompute,
-            {m_IblSetLayout}).Handle;
+            {m_IblSetLayout});
         m_PrefShader = shaders.CreateFromFile(
             dir + "ibl_prefilter.comp.spv",
             vk::ShaderStageFlagBits::eCompute,
-            {m_IblSetLayout}, sizeof(f32)).Handle;
+            {m_IblSetLayout}, sizeof(f32));
         m_LutShader = shaders.CreateFromFile(
             dir + "ibl_brdf_lut.comp.spv",
             vk::ShaderStageFlagBits::eCompute,
-            {m_LutSetLayout}).Handle;
+            {m_LutSetLayout});
 
         if (!m_ToCubeShader || !m_IrrShader || !m_PrefShader || !m_LutShader) {
             ENGINE_FATAL("IblGenerator: failed to create compute shaders");
@@ -125,10 +125,7 @@ namespace Engine
     {
         const auto& vkDevice = m_Device.GetDevice();
 
-        for (const vk::ShaderEXT s : {m_ToCubeShader, m_IrrShader,
-                                      m_PrefShader, m_LutShader}) {
-            if (s) vkDevice.destroyShaderEXT(s);
-        }
+        // shader 随 ShaderRef 引用计数自动销毁，此处只清其余原生句柄
         vkDevice.destroySampler(m_EquirectSampler);
         vkDevice.destroyPipelineLayout(m_IblLayout);
         vkDevice.destroyPipelineLayout(m_IblLayoutPush);
@@ -412,7 +409,7 @@ namespace Engine
         // 4c. ① equirect → env cube
         std::array<vk::ShaderStageFlagBits, 1> computeStage = {
             vk::ShaderStageFlagBits::eCompute};
-        std::array<vk::ShaderEXT, 1> toCubeShaders = {m_ToCubeShader};
+        std::array<vk::ShaderEXT, 1> toCubeShaders = {m_ToCubeShader.GetHandle()};
         m_Cmd.bindShadersEXT(computeStage, toCubeShaders);
         m_Cmd.bindDescriptorSets(vk::PipelineBindPoint::eCompute,
                                  m_IblLayout, 0, setToCube, {});
@@ -434,7 +431,7 @@ namespace Engine
             vk::DependencyInfo{}.setImageMemoryBarriers(envToRead));
 
         // 4d. ② irradiance（env → irr）
-        std::array<vk::ShaderEXT, 1> irrShaders = {m_IrrShader};
+        std::array<vk::ShaderEXT, 1> irrShaders = {m_IrrShader.GetHandle()};
         m_Cmd.bindShadersEXT(computeStage, irrShaders);
         m_Cmd.bindDescriptorSets(vk::PipelineBindPoint::eCompute,
                                  m_IblLayout, 0, setIrr, {});
@@ -443,7 +440,7 @@ namespace Engine
                        kFaceCount);
 
         // 4e. ③ prefilter（env → pref，每 mip 一次 dispatch + push roughness）
-        std::array<vk::ShaderEXT, 1> prefShaders = {m_PrefShader};
+        std::array<vk::ShaderEXT, 1> prefShaders = {m_PrefShader.GetHandle()};
         m_Cmd.bindShadersEXT(computeStage, prefShaders);
         for (u32 m = 0; m < kPrefilterMips; ++m) {
             const u32 dim = kPrefilterFaceSize >> m;
@@ -459,7 +456,7 @@ namespace Engine
         }
 
         // 4f. ④ BRDF LUT（独立纯数学）
-        std::array<vk::ShaderEXT, 1> lutShaders = {m_LutShader};
+        std::array<vk::ShaderEXT, 1> lutShaders = {m_LutShader.GetHandle()};
         m_Cmd.bindShadersEXT(computeStage, lutShaders);
         m_Cmd.bindDescriptorSets(vk::PipelineBindPoint::eCompute,
                                  m_LutLayout, 0, setLut, {});