Update bevy, hecs, legion, specs and shipyard#30
Conversation
|
Who should I ask for a review before this can be merged? |
There was a problem hiding this comment.
i'm just a random bevy user coming through here, throwing out thoughts:
bevy_ecs0.8.1 is out, which didn't include any changes to 0.8.0 but may still make sense to bump just to be clear towards readers.bevy_ecs::component::Componentis part of thebevy_ecs::prelude::*so those lines could just be#[derive(Component)]instead of the fully qualified path.- bevy supports serde, though that should likely be a separate PR (i don't have experience with that part, but I'll throw a note on the bevy discord in case someone there wants to have a look)
These are just some small nitpicks, looks good to me otherwise.
| use cgmath::*; | ||
|
|
||
| #[derive(Copy, Clone)] | ||
| #[derive(Copy, Clone, bevy_ecs::component::Component)] |
There was a problem hiding this comment.
For an insertion-focused test, I would suggest using Bevy's sparse set storage rather than the default table storage. Ideally there'd be benches for both, but that may be hard to display.
There was a problem hiding this comment.
For using Sparse Set storage you would add a #[component(storage = "SparseSet")] below the Component derive, as it can be used per-component.
There was a problem hiding this comment.
Having bench only for SparseSet based storage wouldn't be fair.
It's a tradeoff with significant iteration speed disadvantage.
Having both would be perfect.
There was a problem hiding this comment.
Yep, I'm on board there. I'd love to see both storage types for iteration and insertion speed.
|
Is this going to get merged or what? |
6 participants
No description provided.