fix: add negative cache for failed spell visual model loads

Spell visual M2 models that fail to load (missing file, empty model,
or GPU upload failure) were re-attempted on every subsequent spell cast,
causing repeated file I/O during combat. Now caches failed model IDs in
spellVisualFailedModels_ so they are skipped on subsequent attempts.

include/rendering/renderer.hpp

@@ -7,6 +7,7 @@
 #include <future>
 #include <cstddef>
 #include <unordered_map>
+#include <unordered_set>
 #include <glm/glm.hpp>
 #include <vulkan/vulkan.h>
 #include <vk_mem_alloc.h>
@@ -343,6 +344,7 @@ private:
     std::unordered_map<uint32_t, std::string> spellVisualCastPath_;   // visualId → cast M2 path
     std::unordered_map<uint32_t, std::string> spellVisualImpactPath_; // visualId → impact M2 path
     std::unordered_map<std::string, uint32_t> spellVisualModelIds_;   // M2 path → M2Renderer modelId
+    std::unordered_set<uint32_t> spellVisualFailedModels_;           // modelIds that failed to load (negative cache)
     uint32_t nextSpellVisualModelId_ = 999000; // Reserved range 999000-999799
     bool spellVisualDbcLoaded_ = false;
     void loadSpellVisualDbc();

src/rendering/renderer.cpp

@@ -2860,16 +2860,21 @@ void Renderer::playSpellVisual(uint32_t visualId, const glm::vec3& worldPosition
         spellVisualModelIds_[modelPath] = modelId;
     }
 
+    // Skip models that have previously failed to load (avoid repeated I/O)
+    if (spellVisualFailedModels_.count(modelId)) return;
+
     // Load the M2 model if not already loaded
     if (!m2Renderer->hasModel(modelId)) {
         auto m2Data = cachedAssetManager->readFile(modelPath);
         if (m2Data.empty()) {
             LOG_DEBUG("SpellVisual: could not read model: ", modelPath);
+            spellVisualFailedModels_.insert(modelId);
             return;
         }
         pipeline::M2Model model = pipeline::M2Loader::load(m2Data);
         if (model.vertices.empty() && model.particleEmitters.empty()) {
             LOG_DEBUG("SpellVisual: empty model: ", modelPath);
+            spellVisualFailedModels_.insert(modelId);
             return;
         }
         // Load skin file for WotLK-format M2s
@@ -2880,6 +2885,7 @@ void Renderer::playSpellVisual(uint32_t visualId, const glm::vec3& worldPosition
         }
         if (!m2Renderer->loadModel(model, modelId)) {
             LOG_WARNING("SpellVisual: failed to load model to GPU: ", modelPath);
+            spellVisualFailedModels_.insert(modelId);
             return;
         }
         LOG_DEBUG("SpellVisual: loaded model id=", modelId, " path=", modelPath);