Jonathan Cornaz

After some trial and errors I finally managed to make it work. Here is how to do: * `action-gh-release` must run on a different workflow reacting on tag push *...

I found a better way! If we have access to the tag of the existing release (release-please has a `tag_name` output), then we can simply give that as an input...

Yes. Conceptualy like this: ```rust pub struct PhysicsPlugin; // Same definition as before pub struct StagedPhysicsPlugin { /* fields */ } // New plugin impl Plugin for PhysicyPlugin { //...

mmh... sorry, I forgot about that one. But, I'm talking about both. Actually about all of them (incl. the `CorePlugin`). The goal is that no existing API is changed. The...

Thanks for the changes, it looks great. I'll have a deeper look soon (-ish). > Its not happening for me locally (granted i do have a weird setup with nixos)....

the CI passed now... I don't know, maybe there is a flakyness. But I doubt it was due to your code. When you think you don't have anything else to...

If I understand well, your need is to choose in what stage the actual physics step happen (the systems that are currently running on `CoreStage::PostUpdate` by default), and then ofc...

> For rollback you would need a constant physics tick rate and re-simulating multiple steps as it is now might get weird with the time delta stuff. It is possible...

Hey, you know what? While talking about our problem in the [bevy discrod](https://discord.com/channels/691052431525675048/810632276701347920/935096643051741194) I got an idea... What if, we only let the user choose in which stage to run...

> Hmm tbh im still a little confused on what each stage actually does. Do these comments describe it mostly correctly? https://github.com/JRMurr/heron/blob/2283b5974385765cb61b4c0b65dd611c32dd5aa0/rapier/src/lib.rs#L73 Here are what the stages do (in order):...