Rework wgpu-rs Context #6619
Rework wgpu-rs Context #6619
Conversation
wgpu/src/dispatch.rs
Outdated
| #[cfg(wgpu_core)] | ||
| #[inline] | ||
| #[allow(unused)] | ||
| pub fn as_core(&self) -> &$core { |
There was a problem hiding this comment.
Result would be nicer imho
depends a bit on the usecase tbf
There was a problem hiding this comment.
We'll get there later, no idea what the current behavior is if you mix context types, at least now it'll panic in a nice well defined place.
wgpu/src/dispatch.rs
Outdated
| pub enum $name { | ||
| #[cfg(wgpu_core)] | ||
| Core($core), | ||
| #[cfg(webgpu)] | ||
| WebGPU($webgpu), | ||
| } |
There was a problem hiding this comment.
So for #6330 we would need to add Custom(Arc<dyn $trait>) variant, but we would probably want to put it behind a cfg.
There was a problem hiding this comment.
Yeah, that's the idea.
cwfitzgerald
force-pushed
the
cw/ctx-rework-2
branch
2 times, most recently
from
5216d7d to
0e53174
Compare
| [[test]] | ||
| name = "wgpu-compile-test" | ||
| path = "compile_tests/root.rs" | ||
| harness = true |
There was a problem hiding this comment.
Lookie: Our test suite has grown a new feature. We can now ensure that a given bit of wgpu code does not compile. This is very useful to make sure that #6145 doesn't come back.
| crate::cmp::impl_eq_ord_hash_arc_address!(ContextWgpuCore => .0); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreAdapter => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreDevice => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreQueue => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreShaderModule => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreBindGroupLayout => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreBindGroup => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreTextureView => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreSampler => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreBuffer => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreTexture => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreBlas => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreTlas => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreQuerySet => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CorePipelineLayout => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreRenderPipeline => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreComputePipeline => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CorePipelineCache => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreCommandEncoder => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreComputePass => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreRenderPass => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreCommandBuffer => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreRenderBundleEncoder => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreRenderBundle => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreSurface => .id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreSurfaceOutputDetail => .surface_id); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreQueueWriteBuffer => .mapping.ptr); | ||
| crate::cmp::impl_eq_ord_hash_proxy!(CoreBufferMappedRange => .ptr); |
There was a problem hiding this comment.
Lookie: Now the backend types are responsible for implementing PartialEq etc how they see fit. Most of them use this proxy type, which just uses a single member for the partial eq behavior.
There was a problem hiding this comment.
Once we get rid of the registries in wgpu-core this will be easy to adapt, just use impl_eq_ord_hash_arc_address. Nice design!
There was a problem hiding this comment.
Lookie! This file is very important
cwfitzgerald
force-pushed
the
cw/ctx-rework-2
branch
from
c787967 to
55c839b
Compare
| impl Identifier { | ||
| pub fn create() -> Self { | ||
| let id = NEXT_ID.fetch_add(1, Ordering::Relaxed); | ||
| // Safety: Will take 7000+ years of constant incrementing to overflow. It's fine. |
|
There is something I was wondering about: Since the public pub struct X {
pub(crate) inner: dispatch::DispatchX,
}Wouldn't it be possible to simplify things further and use the |
Connections
Closes #6324
Closes #6145
Description
Here it is. This reworks how the
wgpucontext works so that there is:Dropcode into the backends as opposed to being in the context.Testing
Tested manually on WebGPU and native tests. Covered through CI. Added compile tests to make sure #6145 doesn't come back.
Checklist
cargo fmt.taplo format.cargo clippy. If applicable, add:--target wasm32-unknown-unknown--target wasm32-unknown-emscriptencargo xtask testto run tests.CHANGELOG.md. See simple instructions inside file.