Make default state change trigger self state transitions

[janhohenheim]

Objective

• enable same state transitions #19363 finally allows self-state transitions
• The naming is a bit meh (set_forced)
• Many people expect this to work out of the box
• That PR also missed a few spots where this distinction applies

Solution

• Expand on enable same state transitions #19363 by making it the default
• Rename set -> set_if_neq
• Rename set_forced -> set
• Add the two variants to
  • commands
  • reflection
  • computed states
• make the transition logs a bit less chatty in the common case
• add a test for the new behavior

Testing

• CI

[github-actions]

It looks like your PR is a breaking change, but you didn't provide a migration guide.

Please review the instructions for writing migration guides, then expand or revise the content in the migration guides directory to reflect your changes.

[alice-i-cecile]

So the 4 areas of access:
App.insert_state : Always transitions. (could add a insert_state_if_neq but meh)
Commands.set_state/set_state_if_neq: Always transitions, unless you dont set it.
NextState.set/set_if_neq: Always transitions, unless you dont set it.
ComputedStates.compute: Always transitions, unless the const flag makes you not set it.

In most/all these scenerios, the StateTransitionEvent is always doing same-state transitioning if current == next, but only differs in whether or not you actually provide a next == current.
So I think it should be possible to drop the field from the Event and just run the schedules by default, with the API of states supporting this set_if_neq ability.

• suggestion on Discord

[mockersf]

We shouldn't change the default on only one of the method, they should all behave the same 👍

[hukasu]

Nit: nowhere in std and bevy "not equal" is written as neq, it is always written as ne

[ickshonpe]

Confused here, if allow_same_state_transitions is true, then same state transitions are allowed, and the state transition systems should be run? Wouldn't state insertion be idempotent then when allow_same_state_transitions is false?

[janhohenheim]

No, since when deactivated, inserting the state without it pre-existing will trigger transitions, and then inserting it a second time won't.

[ickshonpe]

Ah right, too sleepy..

[ickshonpe]

Nit: nowhere in std and bevy "not equal" is written as neq, it is always written as ne

Component change detection has a set_is_neq function. It's not idiomatic rust, but I don't see a problem with it. Both APIs should match, so if we want to use ne instead of neq that should be a separate PR anyway I think.

[janhohenheim]

I believe the blocking feedback was based on a misconception. This PR changes the default on all methods, not just on one.

[ChristopherBiscardi]

it is unclear to me why this const exists.

[ChristopherBiscardi]

Did discovery. This exists because of this discord conversation: https://discord.com/channels/691052431525675048/749335865876021248/1437152386345996430

Which is intended to be a computed_states on/off flag for either always or never triggering. In this case the default is "always trigger".

It is used in state/state_set, which is updated in this PR.