Changes in the shroud are now tracked. If a cell changes it will mark itself and its neighbors as dirty. During the render phase all dirty cells will have their vertices calculated and cached. If a cell is not dirty, the pre-calculated vertices are retrieved from cache. Then the sprite renderer is provided the sprite and the pre-calculated vertices to draw.
This prevents constant recalculation of vertices for the shroud in the render phase, requiring instead only dirty cells in the visible area. The update phase is reduced to a practical noop, instead incurring the cost only of changed cells each frame, rather than checking the visible area.
The HardwarePalette will now grow its palette buffer and texture in power-of-2 increments. This avoids it having to allocate memory for a full 256x256 texture up front. In practice the default mods use 22 or 23 palettes so a 32x256 texture is used. This means both the buffer and texture save neatly on memory. Additionally, HardwarePalette.ApplyModifiers sees a nice speedup as it has to transfer a much smaller amount of memory from the buffer to the texture.
To facilitate this change, the MaxPalettes constant is no more. Instead the PaletteReference deals with the calculation of the index and this is passed into the appropriate methods.
Textures, FrameBuffers and VertexBuffers allocated by the Sdl2 Renderer were only being released via finalizers. This could lead to OpenGL out of memory errors since resources may not be cleaned up in a timely manner. To avoid this, IDisposable has been implemented and transitively applied to classes that use these resources.
As a side-effect some static state is no longer static, particularly in Renderer, in order to facilitate this change and just for nicer design in general.
Also dispose some bitmaps.
- Cache the old resolution, scroll and zoom in BeginFrame, and don't bother updating the viewport parameters again until they change.
- Pass around scroll as an int2 to reduce the number of back-and-forth casts.
