build: add appropriate compile flags to cmake presets #34
Conversation
ViTeXFTW
changed the title
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
| alignas(16) glm::vec4 diffuseColor; // RGB + alpha | ||
| alignas(16) glm::vec4 emissiveColor; // RGB + intensity | ||
| alignas(16) glm::vec4 specularColor; // RGB + shininess | ||
| alignas(16) glm::vec3 hoverTint; // RGB tint for hover highlighting (1,1,1 = no tint) |
There was a problem hiding this comment.
Potential shader layout mismatch: vec3 in std140 layout takes 12 bytes but following fields may not align correctly. In GLSL push constants, after a vec3 at offset 48, the next field starts at offset 60. The C++ alignas(4) on flags may not match this layout, potentially causing the shader to read wrong values.
Consider using vec4 instead of vec3 for hoverTint, or verify the struct size matches expected shader layout (should be 76 bytes total).
Prompt To Fix With AI
This is a comment left during a code review.
Path: src/core/pipeline.hpp
Line: 71:71
Comment:
Potential shader layout mismatch: `vec3` in std140 layout takes 12 bytes but following fields may not align correctly. In GLSL push constants, after a `vec3` at offset 48, the next field starts at offset 60. The C++ `alignas(4)` on `flags` may not match this layout, potentially causing the shader to read wrong values.
Consider using `vec4` instead of `vec3` for `hoverTint`, or verify the struct size matches expected shader layout (should be 76 bytes total).
How can I resolve this? If you propose a fix, please make it concise.| // Store CPU copies for ray-triangle intersection | ||
| gpu.cpuVertices = subMesh.vertices; | ||
| gpu.cpuIndices = subMesh.indices; |
There was a problem hiding this comment.
Storing full CPU-side copies doubles mesh memory usage. Consider building a simplified collision mesh (e.g., just positions, no normals/UVs) if memory becomes a concern for large models.
Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!
Prompt To Fix With AI
This is a comment left during a code review.
Path: src/render/renderable_mesh.cpp
Line: 43:45
Comment:
Storing full CPU-side copies doubles mesh memory usage. Consider building a simplified collision mesh (e.g., just positions, no normals/UVs) if memory becomes a concern for large models.
<sub>Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!</sub>
How can I resolve this? If you propose a fix, please make it concise.
|
Closed in favor of #36 |
Greptile Overview
Greptile Summary
This PR fixes Windows shader compilation issues and adds a complete mouse hover detection system for meshes and skeleton bones.
Windows Shader Fix:
CMakeLists.txtandEmbedShaders.cmaketo avoid Windows command-line length limits[[maybe_unused]]attribute to suppress warnings when no shaders are presentHover Detection Feature:
raycast.cpp/hpp) with triangle, sphere, and line segment intersection algorithmsHoverDetectorclass to manage hover state and test against scene geometryraycast_test.cpp) with 100% coverage of intersection algorithmsBuild Improvements:
-O0 -gfor debug,-O3for release on GCC/Clang)/O2 /GLfor release with/LTCGlink-time optimization)Notes:
Confidence Score: 4/5
Important Files Changed