Simplify conditions

[Ixentus]

Objective

• Conditions don't have to be closures unless they have state or mutate.

Solution

• Simplify conditions when possible.

---

Changelog

The following run conditions are now regular systems:

• resource_exists
• resource_added
• resource_changed
• resource_exists_and_changed
• state_exists<S: States>
• state_changed<S: States>
• any_with_component<T: Component>

Migration Guide

• resource_exists() -> resource_exists
• resource_added() -> resource_added
• resource_changed() -> resource_changed
• resource_exists_and_changed() -> resource_exists_and_changed
• state_exists<S: States>() -> state_exists<S: States>
• state_changed<S: States>() -> state_changed<S: States>
• any_with_component<T: Component>() -> any_with_component<T: Component>

[alice-i-cecile]

@soqb @doonv reviews? Also let me know if either of you want to be added to the Bevy org so I can do things like "request reviewer" (and so you can label things).

[Ixentus]

Alright fixed all the docs tests. Messed up the pr commits though -_-

Edit: Fixed hopefully

[jdm]

Nice simplification!

[hymm]

They were originally made closures to keep the API consistent. i.e. you always call the built-in run conditions as functions when adding them in run_if.

[alice-i-cecile]

They were originally made closures to keep the API consistent. i.e. you always call the built-in run conditions as functions when adding them in run_if.

I think this is a mistake overall: we should model the correct, simple behavior for users. See the previous PR by this author, who thought that it was required to use closures. It's not the first time I've seen that, and I was confused about where the misconception was coming from.

[nvdaz]

Docs should be updated on changed conditions (eg. here) to reflect new behavior. It could be something like

    /// A [`Condition`](super::Condition)-satisfying system that returns `true`
    ...

[AxiomaticSemantics]

Everything here is pretty clear, looks correct and improves readability and flexibility.

[soqb]

LGTM, a clear ergonomic improvement. i think this also reduces confusion, since what is passed to run_if is more obviously a function, it's slightly easier for beginners to recognise that conditions aren't called once and then are called based on that stored flag.

btw @alice-i-cecile, i'd be happy to become an org member!

[Ixentus]

Docs should be updated on changed conditions

Done