Evan Architecture

Evan is split into a small set of runtime layers so rendering code, platform code, and engine content stay separated.

Module Layout

graph TD
    Engine[Engine] --> Platform[Platform]
    Engine --> Device[DeviceContext]
    Engine --> Swapchain[SwapchainContext]
    Engine --> Renderer[Renderer]
 
    Renderer --> RenderObject[RenderObject]
    RenderObject --> GPUMesh[GPUMesh]
    Renderer --> MaterialManager[RessourceManager]
    MaterialManager --> GPUMaterial[GPUMaterial]
 
    Platform --> Backend[Backend: OpenXR / GLFW]

• Engine owns the frame loop and coordinates update, draw, and present.
• Platform abstracts the target backend (OpenXR or GLFW) and platform.
• Device & swapchain manage Vulkan resources and frame ownership.
• Renderer owns the registry of render objects and their GPU meshes, and draws them each frame; GPU materials are managed by the RessourceManager.

Frame Lifecycle

sequenceDiagram
    participant Engine
    participant Platform
    participant Device as DeviceContext
    participant Renderer
 
    loop each frame
        Engine->>Platform: poll events
        Engine->>Engine: update state
        Engine->>Renderer: drawFrame(render objects)
        Engine->>Device: record command buffers
        Engine->>Platform: present image
    end

Build Variants

• BUILD_FOR_GLFW enables desktop platform support.
• BUILD_FOR_OPENXR enables XR support.
• Android-specific paths are enabled when the selected platform is Android.

Design Goal

Keep the engine low-level enough for performance while still exposing a straightforward application lifecycle to the higher layers.