WIP change BFS to DFS for rig building and writing

src/Libraries/MeshCompiler.cpp

@@ -17,7 +17,7 @@
 #include <fstream>
 #include <iostream>
 
-#include <queue>
+#include <stack>
 
 namespace SH_COMP
 {
@@ -138,7 +138,7 @@ namespace SH_COMP
     }
   }
 
-  void MeshCompiler::BuildHeaders(ModelAsset& asset) noexcept
+  void MeshCompiler::BuildHeaders(ModelRef asset) noexcept
   {
     // Mesh Headers
     asset.meshHeaders.resize(asset.meshes.size());
@@ -254,19 +254,19 @@ namespace SH_COMP
     return new RigNodeData(inNode->mName.C_Str(), aiTransformToMat4(inNode->mTransformation));
   }
 
-  void MeshCompiler::LoadFromFile(AssetPath path, ModelAsset& asset) noexcept
+  void MeshCompiler::LoadFromFile(AssetPath path, ModelRef asset) noexcept
   {
     const aiScene* scene = aiImporter.ReadFile(path.string().c_str(), 
-      aiProcess_Triangulate           // Make sure we get triangles rather than nvert polygons
+      aiProcess_Triangulate                 // Make sure we get triangles rather than nvert polygons
       | aiProcess_GenUVCoords               // Convert any type of mapping to uv mapping
       | aiProcess_TransformUVCoords         // preprocess UV transformations (scaling, translation ...)
       | aiProcess_FindInstances             // search for instanced meshes and remove them by references to one master
       | aiProcess_CalcTangentSpace          // calculate tangents and bitangents if possible
       | aiProcess_JoinIdenticalVertices     // join identical vertices/ optimize indexing
-      | aiProcess_FindInvalidData           // detect invalid model data, such as invalid normal vectors
+      | aiProcess_FindInvalidData           // detect invalid model data, such as invalid normal vector
       | aiProcess_FlipUVs                   // flip the V to match the Vulkans way of doing UVs
       | aiProcess_ValidateDataStructure     // checks all bones, animations and vertices are linked correctly
-      | aiProcess_LimitBoneWeights          // Limit number of bones effect vertices to 4
+      //| aiProcess_LimitBoneWeights          // Limit number of bones effect vertices to 4
     );
 
     if (!scene || !scene->HasMeshes())
@@ -286,25 +286,25 @@ namespace SH_COMP
   {
     // Build implementation copy of armature tree
     // node collection write done later when writing to file
-    std::queue<std::pair<RigNodeData*, AiNodeConstPtr>> nodeQueue;
+    std::stack<std::pair<RigNodeData*, AiNodeConstPtr>> nodeStack;
 
-    nodeQueue.emplace(PairHelper(baseNode));
+    nodeStack.emplace(PairHelper(baseNode));
-    rig.root = nodeQueue.front().first;
+    rig.root = nodeStack.top().first;
     rig.header.nodeCount++;
     rig.header.charCounts.push_back(rig.root->name.length());
 
-    while(!nodeQueue.empty())
+    while(!nodeStack.empty())
     {
-      auto currPair = nodeQueue.front();
+      auto currPair = nodeStack.top();
-      nodeQueue.pop();
+      nodeStack.pop();
       auto currNode = currPair.first;
       auto const& currAiNode = currPair.second;
-
+      int const iStart {static_cast<int>(currAiNode->mNumChildren - 1)};
-      for (auto i {0}; i < currAiNode->mNumChildren; ++i)
+      for (int i {iStart}; i >= 0 ; --i)
       {
         auto newPair = PairHelper(currAiNode->mChildren[i]);
         currNode->children.push_back(newPair.first);
-        nodeQueue.push(newPair);
+        nodeStack.push(newPair);
         
 		    rig.header.nodeCount++;
 		    rig.header.charCounts.push_back(newPair.first->name.length());

src/Libraries/MeshWriter.cpp

@@ -11,7 +11,7 @@
 #include "MeshWriter.h"
 #include <fstream>
 #include <iostream>
-#include <queue>
+#include <stack>
 
 namespace SH_COMP
 {
@@ -184,26 +184,27 @@ namespace SH_COMP
     std::vector<RigNodeDataWrite> dataToWrite;
     dataToWrite.reserve(rig.header.nodeCount);
 
-    std::queue<std::pair<RigWriteNode*, RigNodeData*>> nodeQueue;
+    std::stack<std::pair<RigWriteNode*, RigNodeData*>> nodeStack;
     treeRoot = new RigWriteNode;
     treeRoot->id = 0;
     treeRoot->children.clear();
-    nodeQueue.emplace(std::make_pair(treeRoot, rig.root));
+    nodeStack.emplace(std::make_pair(treeRoot, rig.root));
     dataToWrite.emplace_back(rig.root->name, rig.root->transform);
 
-    while(!nodeQueue.empty())
+    while(!nodeStack.empty())
     {
-	    auto currPair = nodeQueue.front();
+	    auto currPair = nodeStack.top();
-      nodeQueue.pop();
+      nodeStack.pop();
       auto currWriteNode = currPair.first;
       auto currDataNode = currPair.second;
 
-      for (auto child : currDataNode->children)
+      for (int i {static_cast<int>(currDataNode->children.size() - 1)}; i >= 0; --i)
       {
+        auto child = currDataNode->children[i];
 	      auto newPair = std::make_pair(new RigWriteNode(), child);
         newPair.first->id = dataToWrite.size();
         currWriteNode->children.push_back(newPair.first);
-        nodeQueue.push(newPair);
+        nodeStack.push(newPair);
 
         dataToWrite.emplace_back(child->name, child->transform);
       }
@@ -223,13 +224,13 @@ namespace SH_COMP
 
   void MeshWriter::WriteRigTree(FileReference file, RigWriteNode const* root)
   {
-    std::queue<RigWriteNode const*> nodeQueue;
+    std::stack<RigWriteNode const*> nodeStack;
-    nodeQueue.push(root);
+    nodeStack.push(root);
 
-    while(!nodeQueue.empty())
+    while(!nodeStack.empty())
     {
-      auto node = nodeQueue.front();
+      auto node = nodeStack.top();
-      nodeQueue.pop();
+      nodeStack.pop();
 
 	    file.write(
 				reinterpret_cast<char const*>(&node->id),
@@ -245,7 +246,7 @@ namespace SH_COMP
 
       for (auto child : node->children)
       {
-	      nodeQueue.push(child);
+	      nodeStack.push(child);
       }
     }
   }