bevy_ecs: add untyped methods for inserting components and bundles by jakobhellermann · Pull Request #5602 · bevyengine/bevy

jakobhellermann · 2022-08-07T10:41:14Z

Objective

[Merged by Bors] - untyped APIs for components and resources #4447 implemented untyped (using component ids instead of generics and TypeId) APIs for inserting/accessing resources and accessing components, but left inserting components for another PR (this one)

Solution

add EntityMut::{insert_by_id, insert_bundle_by_id}
split Bundle into DynamicBundle with get_components and Bundle: DynamicBundle. This allows the BundleInserter machinery to be reused for bundles that can only be written, not read, and have no statically available ComponentIds
add Bundles::bundle_ids_dynamic: HashMap<Vec<ComponentId>, BundleId> next to the HashMap<TypeId, BundleId>

Changelog

add methods for inserting bundles and components to: world.entity_mut(entity).insert_bundle_by_id/insert_by_id

Unresolved questions

can we get rid of some allocations?
- init_info_dynamic could remove a .clone() of a Vec<ComponentId> when HashMap::raw_entry stabilizes
- insert_by_id could not reuse insert_bundle_by_id to avoid the Vec overhead, or alternatively insert_bundle_by_id could maybe be changed to not require the Vec for the OwningPtrs
currently insert_bundle_by_ids requires takes a Vec<ComponentId> that must be sorted and a Vec<OwningPtr<'_> that must match the component ids. This is because internally we also need the Vec<ComponentId> so if the caller already has it we can just use that.
- should we just accept a &[(ComponentId, OwningPtr<'_>)] and internally collect that into a sorted vec of component ids?
- alternatively, should we sort the ComponentId vec in the function so that we don't have such a subtle safety requirement? sorting an already sorted vec is the best case scenario for insertion sort (which is used for std sort for <=20 elements) so it's just a small O(n) cost

crates/bevy_ecs/src/world/entity_ref.rs

NathanSWard · 2022-08-26T17:52:45Z

Pining here to see if there is any plans to pursue this PR?
I wanted to pick up #5074 however, it relies on a DynamicBundle like interface (which is what this PR provides).
I can always adopt this PR, at the discretion of @jakobhellermann of course.

NathanSWard · 2022-09-02T06:09:49Z

ping @jakobhellermann
Any plans to move forward w/ this PR?
I'm more than happy to take it over if that's easier.

jakobhellermann · 2022-09-02T07:43:51Z

ping @jakobhellermann Any plans to move forward w/ this PR? I'm more than happy to take it over if that's easier.

Sorry for not responding earlier. I'm a bit busy with Uni right now but feel free to adopt this PR.

jakobhellermann · 2022-09-20T18:54:30Z

I rebased the PR and adressed the review comments.

It's not quite ready for review yet because I made the DynamicBundle trait safe because I noticed that it had no actual safety requirements anymore and need to properly think about it to make sure that is correct. Also I don't know if Drop logic is correct right now, so I need to look into that as well. I will do both tomorrow probably.

NathanSWard · 2022-09-28T17:15:06Z

crates/bevy_ecs/src/bundle.rs

 pub struct Bundles {
    bundle_infos: Vec<BundleInfo>,
    bundle_ids: HashMap<TypeId, BundleId>,
+    bundle_ids_dynamic: HashMap<Vec<ComponentId>, BundleId>,


The Vec<ComponentId> could be reduced to a Box<[ComponentId]> since it's never appending elements, therefore the entire Vec is not needed.

Creating boxed slices will induce a lot of unnecessary reallocations which are far more expensive than the 8 bytes you save of not having a Vec.

ahh yes true, forgot the Vec would need to drop any excess storage and possibly reallocate. Not just a simple pointer transfer.

jakobhellermann · 2022-10-28T19:27:56Z

rebased the PR

alice-i-cecile

I've taken a look, and this seems solid from an API side. @DJMcNab, you made a very similar PR for me at one point, can I get your review?

DJMcNab

I'm not quite sure on the API design here; allocating vectors every time you need this is unfortunate.

I also think that the ordering things require more careful consideration.

DynamicBundle also needs to be unsafe I believe. Hmm, technically if all the APIs which use it require it to be implemented correctly as part of their contracts, maybe not?

OTOH, I don't mind making this slow that much; people should just write their game code in Rust.

DJMcNab · 2022-11-10T16:47:30Z

crates/bevy_ecs/src/world/entity_ref.rs

+        mut component_ids: Vec<ComponentId>,
+        components: I,
+    ) -> &mut Self {
+        component_ids.sort();


This is not obvious correct.

Do you not also need to simultaneously sort the components?

Suficio · 2022-11-11T00:52:08Z

This PR looks amazing and is super helpful!

I was looking into implementing something similar and would like to throw a suggestion up for discussion,

Would it be possible to move the ComponentId correctness check into init_info_dynamic which could be made safe and public?

The resulting BundleId could be stored by the user and used in fn insert_bundle_by_ids(&mut self, bundle_id, ...).

This would have the added benefit of being able to remove bundle_ids_dynamic, and init_info_dynamic would document that all dynamic queries will be assigned a unique BundleId (this is akin to what happens to "Components::init_component_with_descriptor").

Could rename init_info_dynamic to init_dynamic_bundle.

EDIT: Dumped the changes in jakobhellermann#69

alice-i-cecile · 2023-01-15T16:30:08Z

Controversial because we need to decide between this and #7204 :)

…7204) This MR is a rebased and alternative proposal to #5602 # Objective - #4447 implemented untyped (using component ids instead of generics and TypeId) APIs for inserting/accessing resources and accessing components, but left inserting components for another PR (this one) ## Solution - add `EntityMut::insert_by_id` - split `Bundle` into `DynamicBundle` with `get_components` and `Bundle: DynamicBundle`. This allows the `BundleInserter` machinery to be reused for bundles that can only be written, not read, and have no statically available `ComponentIds` - Compared to the original MR this approach exposes unsafe endpoints and requires the user to manage instantiated `BundleIds`. This is quite easy for the end user to do and does not incur the performance penalty of checking whether component input is correctly provided for the `BundleId`. - This MR does ensure that constructing `BundleId` itself is safe --- ## Changelog - add methods for inserting bundles and components to: `world.entity_mut(entity).insert_by_id`

jakobhellermann added C-Feature A new feature, making something new possible A-ECS Entities, components, systems, and events A-Modding labels Aug 7, 2022

BoxyUwU suggested changes Aug 7, 2022

View reviewed changes

crates/bevy_ecs/src/world/entity_ref.rs Outdated Show resolved Hide resolved

crates/bevy_ecs/src/world/entity_ref.rs Outdated Show resolved Hide resolved

BoxyUwU reviewed Aug 7, 2022

View reviewed changes

crates/bevy_ecs/src/world/entity_ref.rs Outdated Show resolved Hide resolved

jakobhellermann force-pushed the insert-components-by-id branch 3 times, most recently from 2e1fe5a to 199c646 Compare September 20, 2022 17:40

jakobhellermann marked this pull request as draft September 20, 2022 18:54

jakobhellermann force-pushed the insert-components-by-id branch from 199c646 to 34c9332 Compare September 21, 2022 10:33

jakobhellermann marked this pull request as ready for review September 21, 2022 10:55

NathanSWard suggested changes Sep 28, 2022

View reviewed changes

jakobhellermann added 3 commits October 28, 2022 21:26

bevy_ecs: add untyped methods for inserting components and bundles

301030a

sort component ids in insert_bundle_by_ids

44f23f0

get rid of allocation for OwningPtrs

25115e8

jakobhellermann force-pushed the insert-components-by-id branch from 34c9332 to 25115e8 Compare October 28, 2022 19:27

use spawn_empty() instead of spawn()

ed2169a

jakobhellermann requested a review from NathanSWard October 28, 2022 19:37

alice-i-cecile requested a review from BoxyUwU October 28, 2022 20:34

alice-i-cecile approved these changes Oct 28, 2022

View reviewed changes

Escapingbug mentioned this pull request Nov 10, 2022

What is still missing for bevy and/or this repo to declare new component entirely in dynamic fashion? jakobhellermann/bevy_mod_js_scripting#34

Open

DJMcNab reviewed Nov 10, 2022

View reviewed changes

Suficio mentioned this pull request Nov 11, 2022

Alt proposal jakobhellermann/bevy#69

Closed

Suficio mentioned this pull request Jan 15, 2023

bevy_ecs: add untyped methods for inserting components and bundles #7204

Merged

alice-i-cecile added the X-Needs-SME This type of work requires an SME to approve it. label Jan 15, 2023

jakobhellermann closed this Mar 21, 2023

Uh oh!

Conversation

jakobhellermann commented Aug 7, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Objective

Solution

Changelog

Unresolved questions

Uh oh!

Uh oh!

Uh oh!

Uh oh!

NathanSWard commented Aug 26, 2022

Uh oh!

NathanSWard commented Sep 2, 2022

Uh oh!

jakobhellermann commented Sep 2, 2022

Uh oh!

jakobhellermann commented Sep 20, 2022

Uh oh!

NathanSWard Sep 28, 2022

Choose a reason for hiding this comment

Uh oh!

orlp Sep 28, 2022

Choose a reason for hiding this comment

Uh oh!

NathanSWard Sep 28, 2022

Choose a reason for hiding this comment

Uh oh!

jakobhellermann commented Oct 28, 2022

Uh oh!

alice-i-cecile left a comment

Choose a reason for hiding this comment

Uh oh!

DJMcNab left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

DJMcNab Nov 10, 2022

Choose a reason for hiding this comment

Uh oh!

Suficio commented Nov 11, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

alice-i-cecile commented Jan 15, 2023

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

jakobhellermann commented Aug 7, 2022 •

edited

Loading

DJMcNab left a comment •

edited

Loading

Suficio commented Nov 11, 2022 •

edited

Loading