Clarify that a Camera without RenderLayers only renders RenderLayers::layer(0)

[FrTerstappen]

How can Bevy's documentation be improved?

The documentation of RenderLayers in

bevy/crates/bevy_render/src/view/visibility/render_layers.rs

Lines 10 to 21 in 18e1bf1

	/// Describes which rendering layers an entity belongs to.
	///
	/// Cameras with this component will only render entities with intersecting
	/// layers.
	///
	/// Entities may belong to one or more layers, or no layer at all.
	///
	/// The [`Default`] instance of `RenderLayers` contains layer `0`, the first layer.
	///
	/// An entity with this component without any layers is invisible.
	///
	/// Entities without this component belong to layer `0`.

contains the following part:

/// Cameras with this component will only render entities with intersecting
/// layers.

This lead me to assume that a camera without this component renders entities without regard to the layer they belong to. This does not work as a camera without RenderLayers only renders entities belonging to RenderLayers::layer(0).

This can be seen in

bevy/crates/bevy_render/src/view/visibility/mod.rs

Line 533 in 18e1bf1

let view_mask = maybe_view_mask.unwrap_or_default();

where a missing RenderLayers is replaced with the default.

While it could be argued that the line

/// Entities without this component belong to layer `0`. 

also applies to the camera this was not clear to me.

I propose expanding the camera related section of the docs to something like this:

/// Cameras with this component will only render entities with intersecting
/// layers.
/// Cameras without this component will only render entities belonging to the default render layer (RenderLayers::layer(0)).

A further improvement would be to describe how to create a camera that renders all layers (if that is even possible).
I currently create a RenderLayers component containing all layers I use.
This feels like a workaround and is error prone because I have to remember to add new layers.