Merge branch 'trunk' into glsl-constant-id-support

Author: 06wj

.github/workflows/ci.yml

@@ -691,7 +691,7 @@ jobs:
         run: taplo format --check --diff
 
       - name: Check for typos
-        uses: crate-ci/typos@v1.39.2
+        uses: crate-ci/typos@v1.40.0
 
   check-cts-runner:
     # runtime is normally 2 minutes

.github/workflows/cts.yml

@@ -99,7 +99,14 @@ jobs:
         if: matrix.os == 'ubuntu-24.04'
         uses: ./.github/actions/install-mesa
 
-      - name: run CTS
+      # Mixing --llvm-cov with deno (rusty_v8) has linking problems.
+      # Explicitly set the backend to avoid EGL messages in output,
+      # because these tests check stdout.
+      - name: Test cts_runner
+        shell: bash
+        run: DENO_WEBGPU_BACKEND=${{ matrix.backend }} cargo --locked test -p cts_runner
+
+      - name: Run CTS
         shell: bash
         run: cargo --locked xtask cts --llvm-cov --backend ${{ matrix.backend }}
 

CHANGELOG.md

@@ -106,7 +106,15 @@ One other breaking change worth noting is that in WGSL `@builtin(view_index)` no
 
 By @SupaMaggie70Incorporated in [#8206](https://github.com/gfx-rs/wgpu/pull/8206).
 
-#### Error Scopes are now thread-local
+#### Error scopes now use guards and are thread-local.
+
+```diff
+- device.push_error_scope(wgpu::ErrorFilter::Validation);
++ let scope = device.push_error_scope(wgpu::ErrorFilter::Validation);
+  // ... perform operations on the device ...
+- let error: Option<Error> = device.pop_error_scope().await;
++ let error: Option<Error> = scope.pop().await;
+```
 
 Device error scopes now operate on a per-thread basis. This allows them to be used easily within multithreaded contexts,
 without having the error scope capture errors from other threads.
@@ -148,7 +156,7 @@ By @cwfitzgerald in [#8579](https://github.com/gfx-rs/wgpu/pull/8579).
 - Validation errors from `CommandEncoder::finish()` will report the label of the invalid encoder. By @kpreid in [#8449](https://github.com/gfx-rs/wgpu/pull/8449).
 - Corrected documentation of the minimum alignment of the *end* of a mapped range of a buffer (it is 4, not 8). By @kpreid in [#8450](https://github.com/gfx-rs/wgpu/pull/8450).
 - `util::StagingBelt` now takes a `Device` when it is created instead of when it is used. By @kpreid in [#8462](https://github.com/gfx-rs/wgpu/pull/8462).
-- `wgpu_hal::vulkan::Device::texture_from_raw` now takes an `external_memory` argument. By @s-ol in [#8512](https://github.com/gfx-rs/wgpu/pull/8512)
+- `wgpu_hal::vulkan::Texture` API changes to handle externally-created textures and memory more flexibly. By @s-ol in [#8512](https://github.com/gfx-rs/wgpu/pull/8512), [#8521](https://github.com/gfx-rs/wgpu/pull/8521).
 
 #### Metal
 - Add support for mesh shaders. By @SupaMaggie70Incorporated in [#8139](https://github.com/gfx-rs/wgpu/pull/8139)

Cargo.lock

@@ -110,7 +110,6 @@ bytemuck = { version = "1.22", features = [
 cargo_metadata = "0.23"
 cfg_aliases = "0.2.1"
 cfg-if = "1"
-criterion = "0.7"
 codespan-reporting = { version = "0.12", default-features = false }
 ctor = "0.6"
 diff = "0.1"

Cargo.toml

@@ -27,7 +27,6 @@ unexpected_cfgs = { level = "warn", check-cfg = [
 anyhow.workspace = true
 bincode = { workspace = true, features = ["serde"] }
 bytemuck.workspace = true
-# criterion.workspace = true
 naga = { workspace = true, features = [
     "deserialize",
     "serialize",

benches/Cargo.toml

@@ -251,7 +251,7 @@ impl ComputepassState {
                 .create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
                     label: None,
                     bind_group_layouts: &[&bind_group_layout],
-                    push_constant_ranges: &[],
+                    immediates_ranges: &[],
                 });
 
         let pipeline =
@@ -346,7 +346,7 @@ impl ComputepassState {
                     .create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
                         label: None,
                         bind_group_layouts: &[&bindless_bind_group_layout],
-                        push_constant_ranges: &[],
+                        immediates_ranges: &[],
                     });
 
             let bindless_pipeline =

benches/benches/wgpu-benchmark/computepass.rs

@@ -151,7 +151,7 @@ impl RenderpassState {
                 .create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
                     label: None,
                     bind_group_layouts: &[&bind_group_layout],
-                    push_constant_ranges: &[],
+                    immediates_ranges: &[],
                 });
 
         let mut vertex_buffers = Vec::with_capacity(vertex_buffer_count as usize);
@@ -287,7 +287,7 @@ impl RenderpassState {
                     .create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
                         label: None,
                         bind_group_layouts: &[&bindless_bind_group_layout],
-                        push_constant_ranges: &[],
+                        immediates_ranges: &[],
                     });
 
             bindless_pipeline = Some(device_state.device.create_render_pipeline(

benches/benches/wgpu-benchmark/renderpass.rs

@@ -1 +1 @@
-3ba0344b8f33cc110df0d840e09d8c0fc7b29928
+4595051dc14c398aad7f5df2d19cc6861e7f59ba

cts_runner/revision.txt

@@ -372,7 +372,7 @@ impl GPUDevice {
     let wgpu_descriptor = wgpu_core::binding_model::PipelineLayoutDescriptor {
       label: crate::transform_label(descriptor.label.clone()),
       bind_group_layouts: Cow::Owned(bind_group_layouts),
-      push_constant_ranges: Default::default(),
+      immediates_ranges: Default::default(),
     };
 
     let (id, err) = self.instance.device_create_pipeline_layout(