Fix resource refcounting

src/backend_wgpu.c

@@ -259,8 +259,10 @@ void BindShaderWithSettings(Shader shader, PrimitiveType drawMode, RenderSetting
     impl->state.settings = settings;
     impl->state.colorAttachmentState = GetActiveRenderPass()->colorAttachmentState;
     WGPURenderPipeline activePipeline = PipelineHashMap_getOrCreate(&impl->pipelineCache, &impl->state, &impl->shaderModule, &impl->bglayout, &impl->layout);
-    wgpuRenderPassEncoderSetPipeline(g_renderstate.activeRenderpass->rpEncoder, activePipeline);
-    wgpuRenderPassEncoderSetBindGroup(g_renderstate.activeRenderpass->rpEncoder, 0, UpdateAndGetNativeBindGroup(&impl->bindGroup), 0, NULL);
+    if (activePipeline) {
+        wgpuRenderPassEncoderSetPipeline(g_renderstate.activeRenderpass->rpEncoder, activePipeline);
+        wgpuRenderPassEncoderSetBindGroup(g_renderstate.activeRenderpass->rpEncoder, 0, UpdateAndGetNativeBindGroup(&impl->bindGroup), 0, NULL);
+    }
 }
 
 void BindShader(Shader shader, PrimitiveType drawMode) {
@@ -654,6 +656,10 @@ void UpdateBindGroup(DescribedBindGroup *bg) {
     // std::cout << "Updating bindgroup with " << bg->desc.entryCount << " entries" << std::endl;
     // std::cout << "Updating bindgroup with " << bg->desc.entries[1].binding << " entries" << std::endl;
     if (bg->needsUpdate) {
+        if (bg->bindGroup) {
+            wgpuBindGroupRelease((WGPUBindGroup)bg->bindGroup);
+            bg->bindGroup = NULL;
+        }
         WGPUBindGroupDescriptor desc  = {0};
         WGPUBindGroupEntry* aswgpu = (WGPUBindGroupEntry*)RL_CALLOC(bg->entryCount, sizeof(WGPUBindGroupEntry));
         for (uint32_t i = 0; i < bg->entryCount; i++) {
@@ -665,10 +671,29 @@ void UpdateBindGroup(DescribedBindGroup *bg) {
             to->size = from->size;
             to->sampler = from->sampler;
             to->textureView = from->textureView;
+
+            if(bg->layout && bg->layout->entries){
+                if(to->sampler == NULL && bg->layout->entries[i].type == texture_sampler){
+                    to->sampler = (WGPUSampler)g_renderstate.defaultSampler.sampler;
+                }
+                if(to->textureView == NULL && bg->layout->entries[i].type == texture2d){
+                    to->textureView = (WGPUTextureView)g_renderstate.whitePixel.view;
+                }
+                if(to->buffer == NULL && (bg->layout->entries[i].type == uniform_buffer || bg->layout->entries[i].type == storage_buffer)){
+                    to->buffer = (WGPUBuffer)g_renderstate.identityMatrix->buffer;
+                    to->offset = 0;
+                    to->size = 64; // sizeof(Matrix)
+                }
+            }
         }
 
         desc.entries = aswgpu;
         desc.entryCount = bg->entryCount;
+        if (bg->layout == NULL) {
+            TRACELOG(LOG_ERROR, "UpdateBindGroup: bind group layout is NULL. Skipping bind group creation.");
+            RL_FREE(aswgpu);
+            return;
+        }
         desc.layout = bg->layout->layout;
         bg->bindGroup = wgpuDeviceCreateBindGroup(GetDevice(), &desc);
         bg->needsUpdate = false;
@@ -687,9 +712,9 @@ static inline uint64_t bgEntryHashRD(const ResourceDescriptor bge) {
 }
 
 void UpdateBindGroupEntry(DescribedBindGroup *bg, size_t index, ResourceDescriptor entry) {
-    if (index >= bg->entryCount) {
+    if (bg->entries == NULL || index >= bg->entryCount) {
         TRACELOG(LOG_WARNING, "Trying to set entry %d on a BindGroup with only %d entries", (int)index, (int)bg->entryCount);
-        // return;
+        return;
     }
     WGPUBuffer newpuffer = entry.buffer;
     WGPUTextureView newtexture = entry.textureView;
@@ -705,6 +730,9 @@ void UpdateBindGroupEntry(DescribedBindGroup *bg, size_t index, ResourceDescript
     if (entry.textureView) {
         wgpuTextureViewAddRef((WGPUTextureView)entry.textureView);
     }
+    if (entry.sampler) {
+        wgpuSamplerAddRef((WGPUSampler)entry.sampler);
+    }
 
     if (bg->entries[index].buffer) {
         wgpuBufferRelease((WGPUBuffer)bg->entries[index].buffer);
@@ -714,6 +742,10 @@ void UpdateBindGroupEntry(DescribedBindGroup *bg, size_t index, ResourceDescript
         wgpuTextureViewRelease((WGPUTextureView)bg->entries[index].textureView);
         bg->entries[index].textureView = 0;
     }
+    if (bg->entries[index].sampler) {
+        wgpuSamplerRelease((WGPUSampler)bg->entries[index].sampler);
+        bg->entries[index].sampler = 0;
+    }
 
     bg->entries[index] = entry;
     bg->descriptorHash ^= bgEntryHashRD(bg->entries[index]);
@@ -1595,19 +1627,18 @@ DescribedSampler LoadSamplerEx(TextureWrap amode, TextureFilter fmode, TextureFi
     return ret;
 }
 void SetBindgroupUniformBufferData(DescribedBindGroup *bg, uint32_t index, const void *data, size_t size) {
-
     const WGPUBufferDescriptor bufferDesc = {
         .size = size,
         .usage = WGPUBufferUsage_CopySrc | WGPUBufferUsage_CopyDst | WGPUBufferUsage_Uniform,
         .mappedAtCreation = false,
     };
     WGPUBuffer uniformBuffer = wgpuDeviceCreateBuffer((WGPUDevice)GetDevice(), &bufferDesc);
     wgpuQueueWriteBuffer((WGPUQueue)GetQueue(), uniformBuffer, 0, data, size);
+    // Use the existing binding index from the bind group layout if available
+    uint32_t bindingIdx = (bg->entries && index < bg->entryCount) ? bg->entries[index].binding : index;
+
     const ResourceDescriptor entry = {
-        .binding = index,
-        .buffer = uniformBuffer,
-        .size = size,
-    };
+        .binding = bindingIdx, .buffer = uniformBuffer, .size = size, .offset = 0, .sampler = 0, .textureView = 0};
 
     UpdateBindGroupEntry(bg, index, entry);
     wgpuBufferRelease(uniformBuffer);
@@ -1622,11 +1653,10 @@ void SetBindgroupStorageBufferData(DescribedBindGroup *bg, uint32_t index, const
     };
     WGPUBuffer storageBuffer = wgpuDeviceCreateBuffer((WGPUDevice)GetDevice(), &bufferDesc);
     wgpuQueueWriteBuffer((WGPUQueue)GetQueue(), storageBuffer, 0, data, size);
+    uint32_t bindingIdx = (bg->entries && index < bg->entryCount) ? bg->entries[index].binding : index;
+
     const ResourceDescriptor entry = {
-        .binding = index,
-        .buffer = storageBuffer,
-        .size = size,
-    };
+        .binding = bindingIdx, .buffer = storageBuffer, .size = size, .offset = 0, .sampler = 0, .textureView = 0};
 
     UpdateBindGroupEntry(bg, index, entry);
 
@@ -2841,8 +2871,26 @@ EntryPointSet getEntryPointsSPIRV(const uint32_t *shaderSourceSPIRV, uint32_t wo
 }
 
 void UnloadBindGroup(DescribedBindGroup *bg) {
-    free(bg->entries);
-    wgpuBindGroupRelease((WGPUBindGroup)bg->bindGroup);
+    if (bg->entries) {
+        for (uint32_t i = 0; i < bg->entryCount; i++) {
+            if (bg->entries[i].buffer) {
+                wgpuBufferRelease((WGPUBuffer)bg->entries[i].buffer);
+            }
+            if (bg->entries[i].textureView) {
+                wgpuTextureViewRelease((WGPUTextureView)bg->entries[i].textureView);
+            }
+            if (bg->entries[i].sampler) {
+                wgpuSamplerRelease((WGPUSampler)bg->entries[i].sampler);
+            }
+        }
+        free(bg->entries);
+    }
+    if (bg->bindGroup) {
+        wgpuBindGroupRelease((WGPUBindGroup)bg->bindGroup);
+    }
+    bg->bindGroup = NULL;
+    bg->entries = NULL;
+    bg->entryCount = 0;
 }
 void UnloadBindGroupLayout(DescribedBindGroupLayout *bglayout) {
     free(bglayout->entries);

src/internal_include/internals.h

@@ -866,6 +866,10 @@ static WGPURenderPipeline PipelineHashMap_getOrCreate(PipelineHashMap* cacheMap,
     }
     else{
         WGPURenderPipeline toEmplace = createSingleRenderPipe(mst, shaderModule, bglayout, pllayout);
+        if (!toEmplace) {
+            TRACELOG(LOG_ERROR, "Failed to create render pipeline, likely due to a shader compilation/translation error.");
+            return NULL;
+        }
         PipelineHashMap_put(cacheMap, *mst, toEmplace);
         return toEmplace;
     }

src/raygpu.c

@@ -2453,7 +2453,18 @@ DescribedBindGroup LoadBindGroup(const DescribedBindGroupLayout* bglayout, const
     DescribedBindGroup ret  = {0};
     if(entryCount > 0){
         ret.entries = (ResourceDescriptor*)RL_CALLOC(entryCount, sizeof(ResourceDescriptor));
-        memcpy(ret.entries, entries, entryCount * sizeof(ResourceDescriptor));
+        for (uint32_t i = 0; i < entryCount; i++) {
+            ret.entries[i] = entries[i];
+            if (entries[i].buffer) {
+                wgpuBufferAddRef((WGPUBuffer)entries[i].buffer);
+            }
+            if (entries[i].textureView) {
+                wgpuTextureViewAddRef((WGPUTextureView)entries[i].textureView);
+            }
+            if (entries[i].sampler) {
+                wgpuSamplerAddRef((WGPUSampler)entries[i].sampler);
+            }
+        }
     }
     ret.entryCount = entryCount;
     ret.layout = bglayout;
@@ -3660,18 +3671,28 @@ RGAPI ShaderImpl* GetShaderImplByID(uint32_t id){
 RGAPI ShaderImpl* GetShaderImpl(Shader shader){
     return GetShaderImplByID(shader.id);
 }
+
 RGAPI uint32_t getNextShaderID_shc(){
     if(nextShaderID_shc >= capacity_shc){
         uint32_t newCapacity = capacity_shc * 2 + (capacity_shc == 0) * 8;
         ShaderImpl* newAllocatedShaderIDs_shc = (ShaderImpl*)RL_CALLOC(newCapacity, sizeof(ShaderImpl));
         if(capacity_shc){
             memcpy(newAllocatedShaderIDs_shc, allocatedShaderIDs_shc, capacity_shc * sizeof(ShaderImpl));
+
+            // Fixup self-referential pointers after move
+            for (uint32_t i = 0; i < capacity_shc; i++) {
+                ShaderImpl* oldPtr = allocatedShaderIDs_shc + i;
+                ShaderImpl* newPtr = newAllocatedShaderIDs_shc + i;
+                if (newPtr->bindGroup.layout == &oldPtr->bglayout) {
+                    newPtr->bindGroup.layout = &newPtr->bglayout;
+                }
+            }
         }
         if(allocatedShaderIDs_shc){
             RL_FREE(allocatedShaderIDs_shc);
         }
         capacity_shc = newCapacity;
-        allocatedShaderIDs_shc = newAllocatedShaderIDs_shc; 
+        allocatedShaderIDs_shc = newAllocatedShaderIDs_shc;
     }
     return nextShaderID_shc++;
 }