Function project_glb_size 

pub fn project_glb_size(content: &[u8], opts: &GltfOptions) -> GlbSizeProjection

Project the single-GLB size for content from pass 1 only — the world AABB + per-mesh byte sizes the bounded assembler already computes — WITHOUT meshing twice or allocating the output (#1516). Lets a caller pick single GLB vs multi-buffer glTF up front. Meshes the model once (the price of an exact size).