Struct GreedyConfig 

pub struct GreedyConfig<D, FA, FL, FG, FO, FS, FP, FT> {
    pub data: D,
    pub draw_delta: Vec3<i32>,
    pub greedy_size: Vec3<usize>,
    pub greedy_size_cross: Vec3<usize>,
    pub get_ao: FA,
    pub get_light: FL,
    pub get_glow: FG,
    pub get_opacity: FO,
    pub should_draw: FS,
    pub push_quad: FP,
    pub make_face_texel: FT,
}

Fields

data: D

draw_delta: Vec3<i32>

The minimum position to mesh, in the coordinate system used for queries against the volume.

greedy_size: Vec3<usize>

For each dimension i, for faces drawn in planes parallel to i, represents the number of voxels considered along dimension i in those planes, starting from draw_delta.

greedy_size_cross: Vec3<usize>

For each dimension i, represents the number of planes considered orthogonal to dimension i, starting from draw_delta. This should usually be the same as greedy_size.

An important exception is during chunk rendering (where vertical faces at chunk boundaries would otherwise be rendered twice, and also force us to use more than 5 bits to represent x and y positions–though there may be a clever way around the latter). Thus, for chunk rendering we set the number of vertical planes to one less than the chunk size along the x and y dimensions, but keep the number of horizontal planes large enough to cover the whole chunk.

get_ao: FA

Given a position, return the AO information for the voxel at that position (0.0 - 1.0).

get_light: FL

Given a position, return the lighting information for the voxel at that position.

get_glow: FG

Given a position, return the glow information for the voxel at that position (i.e: additional non-sun light).

get_opacity: FO

Given a position, return the opacity information for the voxel at that position. Currently, we don’t support real translucent lighting, so the value should either be false (for opaque blocks) or true (otherwise).

should_draw: FS

Given a position and a normal, should we draw the face between the position and position - normal (i.e. the voxel “below” this vertex)? If so, provide its orientation, together with any other meta information required for the mesh that needs to split up faces. For example, terrain faces currently record a bit indicating whether they are exposed to water or not, so we should not merge faces where one is submerged in water and the other is not, even if they otherwise have the same orientation, dimensions, and are next to each other.

push_quad: FP

Create an opaque quad (used for only display rendering) from its top-left atlas position, the rectangle’s dimensions in (2D) atlas space, a world position, the u and v axes of the rectangle in (3D) world space, the normal facing out frmo the rectangle in world space, and meta information common to every voxel in this rectangle.

make_face_texel: FT

Given a position and the lighting information for a face at that position, return the texel for the face at that position.