26. https://github.com/bevyengine/bevy/pull/18047
The trivial dependency bump that turned into a disaster due to questionable decisions from getrand on how to handle web support. Sorry! We should make an explicit call on what to do though (even if that's wait for now), so this can stay.
Objective New rand version, which means updating glam and encase to support the newer ecosystem update. Does mean that this changes how WASM builds need to be done in order to configure getrandom c...
27. https://github.com/bevyengine/bevy/pull/19232
Enable wayland by default. Trivial but controversial. Fine, this can stay, and we'll make a call quick. The release candidate process is good for testing risky user-facing changes like this.
28. https://github.com/bevyengine/bevy/pull/17889
Remote asset sources via https. I really want this so we can use better assets for our examples. This has been a huge pain. I would like to ship this, but Francois has had ongoing reasonable complaints. Let me ask him what we should do.
Objective This is a duplicate of #16366 because I made a mess of the commit history. See that pr for more context and discussion. Fixes Feature: Upstream bevy_web_asset, allowing assets loaded via...
29. https://github.com/bevyengine/bevy/pull/19340
Image node rotation. Neat! Definitely not milestone content: it's in draft due to a need for design with no movement for months. Cutting.
30. https://github.com/bevyengine/bevy/issues/18981
Disabled entities are not working properly with rendering, among other things. This is largely due to our decision to avoid firing change detection and lifecycle events when disabling and re-enabling entities.
Needs serious design; cutting :(
Bevy version 0.16.0 What you did Spawn an Entity with a mesh and the Disabled component. Remove the component later. What went wrong The mesh will not render. Additional information Minimal reprodu...
31. https://github.com/bevyengine/bevy/pull/19377
A random cleanup PR for some BRP method names. Just needs a second review. This can stay: I can review this even if needed. Better to get the breaking changes done sooner.
32. https://github.com/bevyengine/bevy/pull/19415
A hacky text measure functionality. Useful, but a reviewer suggested some serious improvements. Those haven't been implemented, so I'm bumping to the next milestone.
33. https://github.com/bevyengine/bevy/issues/19452
A simple request to experiment with type magic to avoid ergonomics regressions for state-scoped entities. Easy to test, leaving it in and I'll get to it this week.
I have reverted back to clear_events_on_exit_state, but unlike with 4 it requires specifying the event type each time. Quick question: if you use an impl Trait argument, rather than a generic, can ...
34. https://github.com/bevyengine/bevy/pull/19489
A proposed change (again...) to the behavior of `Single`. I'm on the fence about this: it's more consistent but less convenient. We should make the call now and stick with it; this stays in the milestone.
35. https://github.com/bevyengine/bevy/issues/19569
A "bug" about observers and command ordering. This is extremely complex and needs serious design. Cutting.
36. https://github.com/bevyengine/bevy/issues/19623
Observers panicking due to earlier observers despawning entities they were watching. Should be an easy fix; I'll take a crack at this. Panics are bad! Leaving it in.
Bevy version 0.16.1 What you did use bevy::prelude::*; fn main() -> AppExit { App::new() .add_plugins(MinimalPlugins) .add_systems(Startup, spawn) .add_systems(Update, despawn_transforms) .add_obse...
37. https://github.com/bevyengine/bevy/pull/19580
System combinators for resultful systems, treating Ok as true and Err as false. Fine, this can exist. Only one review needed, and I can tackle that. It gets to live for now, but I wouldn't leave it in if it was further off.
Objective This is a follow-up of #19553. I missed something. Right now the combinators (and, or, ...) only work with conditions returning bool. The same goes for condition_changed, condition_change...
38. https://github.com/bevyengine/bevy/issues/19660
Ah, one of several issues and PRs for "we should finalize the Event trait split before releasing 0.17". These sort of conceptual shifts are often worse for users than more serious breaks, so we need to get this right once and for all.
39. https://github.com/bevyengine/bevy/issues/19263
Similarly, we should figure out the final naming for various observer concepts and methods. It gets to stay: I'm going to need to wrangle Cart into a discussion on this I think.
40. https://github.com/bevyengine/bevy/issues/19715
Multiple copies of the `children!` macro override each other. This is basically "working as intended" and needs serious design to improve, Cutting.
41. https://github.com/bevyengine/bevy/pull/19755
Ah, the newtype PR for ComponentId inside of observers. This was from a new contributor, and the work was ultimately adopted. Closing.
Objective Add a newtype to store events, as mentioned in this HackMD. Solution From the Discord, we agreed on creating a EventKey struct that wraps around ComponentId. I created EventKey and refac...
42. https://github.com/bevyengine/bevy/issues/19844
A Cart issue! We silently lost staticness in our AssetPath types due to a well-intentioned cleanup PR, resulting in unnecessary clones. This needs to be fixed, even if that's just "revert and then add detailed comments explaining why the weird thing was done".
43. https://github.com/bevyengine/bevy/pull/19864
DLSS! This looks "basically done", but is blocked on the wgpu upgrade. We can leave this in: testing for complex features earlier is nice, and contributor happiness is important too :)
44. https://github.com/bevyengine/bevy/issues/19925
Ah cache invalidation. The request is very simple and reasonable, but we need an implementation PR here. Pinging some rendering folks for opinions.
45. https://github.com/bevyengine/bevy/issues/19644
Window component split. Sensible enough to include, but not critical. This appears to be blocked on tricky rendering work: I am absolutely going to punt rather than cram this in.
46. https://github.com/bevyengine/bevy/issues/19990
Another option for how to deal with the glTF coordinates problem. It gets to stay and we'll close out all but one of these PRs.
47. https://github.com/bevyengine/bevy/pull/20035
UI gradient shaders fixes. Definitely important to ship. Why is this blocked? Ah, linked PR to clean things up, which is now merged. Fixing the label and pinging the author.
48. https://github.com/bevyengine/bevy/issues/20040
Reflect derives for the new UI code. Very uncontroversial, nice to do. Why is *this* blocked? Oh, it's because bevy_ui *also* doesn't use feature flags properly for reflection. Boo. Well, adding that to the milestone: this is all very acheiveable.
49. https://github.com/bevyengine/bevy/pull/20069
A usability feature for rendering for working with bind groups. Sensible enough I think? Requesting another rendering review and leaving it.
@alice_i_cecile oh no, not again.
Single will become useless again with this change.
I would argue a change in the opposite direction would make a lot more sense.
Have Res systems params behave the same way as Single does now.
Unifying the behavior, and allow users to define to panic through configuration.
If this will be merged, using Query with single and an early out is a lot more ergonomic than `When<Single<&Player>>`.