refactor: s/once_cell::Lazy/std::sync::LazyLock

Weaken our dependence on the `once_cell` crate by using functionality
from `std` instead that was upstreamed from `once_cell`, this time with
what's available in Rust 1.80+.

It's not yet possible to eliminate this dependency entirely, but do what
we can for now.

.github/workflows/ci.yml

@@ -34,12 +34,12 @@ env:
   # We sometimes need nightly to use special things in CI.
   #
   # In order to prevent CI regressions, we pin the nightly version.
-  NIGHTLY_VERSION: "nightly-2023-12-17"
+  NIGHTLY_VERSION: "nightly-2024-08-11"
   # This is the MSRV used by `wgpu` itself and all surrounding infrastructure.
-  REPO_MSRV: "1.76"
+  REPO_MSRV: "1.80.1"
   # This is the MSRV used by the `wgpu-core`, `wgpu-hal`, and `wgpu-types` crates,
   # to ensure that they can be used with firefox.
-  CORE_MSRV: "1.76"
+  CORE_MSRV: "1.80.1"
 
   #
   # Environment variables

Cargo.toml

@@ -42,7 +42,7 @@ default-members = [
 
 [workspace.package]
 edition = "2021"
-rust-version = "1.76"
+rust-version = "1.80.1"
 keywords = ["graphics"]
 license = "MIT OR Apache-2.0"
 homepage = "https://wgpu.rs/"
@@ -77,7 +77,7 @@ bincode = "1"
 bit-vec = "0.8"
 bitflags = "2.6"
 bytemuck = { version = "1.19" }
-cfg_aliases = "0.1"
+cfg_aliases = "0.2.1"
 cfg-if = "1"
 criterion = "0.5"
 codespan-reporting = "0.11"

benches/Cargo.toml

@@ -21,6 +21,11 @@ path = "benches/root.rs"
 # tracy = ["dep:tracy-client", "profiling/profile-with-tracy"]
 # superluminal = ["profiling/profile-with-superluminal"]
 
+[lints.rust]
+unexpected_cfgs = { level = "warn", check-cfg = [
+    'cfg(feature, values("tracy"))',
+] }
+
 [dependencies]
 bincode.workspace = true
 bytemuck.workspace = true
@@ -38,7 +43,6 @@ naga = { workspace = true, features = [
     "wgsl-out",
 ] }
 nanorand.workspace = true
-once_cell.workspace = true
 pollster.workspace = true
 profiling.workspace = true
 rayon.workspace = true

benches/benches/computepass.rs

@@ -5,8 +5,8 @@ use std::{
 
 use criterion::{criterion_group, Criterion, Throughput};
 use nanorand::{Rng, WyRand};
-use once_cell::sync::Lazy;
 use rayon::iter::{IntoParallelIterator, ParallelIterator};
+use std::sync::LazyLock;
 
 use crate::DeviceState;
 
@@ -424,7 +424,7 @@ impl ComputepassState {
 }
 
 fn run_bench(ctx: &mut Criterion) {
-    let state = Lazy::new(ComputepassState::new);
+    let state = LazyLock::new(ComputepassState::new);
 
     let dispatch_count = dispatch_count();
     let dispatch_count_bindless = dispatch_count_bindless();
@@ -449,7 +449,7 @@ fn run_bench(ctx: &mut Criterion) {
             group.bench_function(
                 format!("{cpasses} computepasses x {dispatch_per_pass} dispatches ({label})"),
                 |b| {
-                    Lazy::force(&state);
+                    LazyLock::force(&state);
 
                     b.iter_custom(|iters| {
                         profiling::scope!("benchmark invocation");
@@ -498,7 +498,7 @@ fn run_bench(ctx: &mut Criterion) {
         group.bench_function(
             format!("{threads} threads x {dispatch_per_pass} dispatch"),
             |b| {
-                Lazy::force(&state);
+                LazyLock::force(&state);
 
                 b.iter_custom(|iters| {
                     profiling::scope!("benchmark invocation");
@@ -538,7 +538,7 @@ fn run_bench(ctx: &mut Criterion) {
     group.throughput(Throughput::Elements(dispatch_count_bindless as _));
 
     group.bench_function(format!("{dispatch_count_bindless} dispatch"), |b| {
-        Lazy::force(&state);
+        LazyLock::force(&state);
 
         b.iter_custom(|iters| {
             profiling::scope!("benchmark invocation");
@@ -579,7 +579,7 @@ fn run_bench(ctx: &mut Criterion) {
             texture_count + storage_texture_count + storage_buffer_count
         ),
         |b| {
-            Lazy::force(&state);
+            LazyLock::force(&state);
 
             b.iter(|| state.device_state.queue.submit([]));
         },

benches/benches/renderpass.rs

@@ -5,8 +5,8 @@ use std::{
 
 use criterion::{criterion_group, Criterion, Throughput};
 use nanorand::{Rng, WyRand};
-use once_cell::sync::Lazy;
 use rayon::iter::{IntoParallelIterator, ParallelIterator};
+use std::sync::LazyLock;
 
 use crate::DeviceState;
 
@@ -427,7 +427,7 @@ impl RenderpassState {
 }
 
 fn run_bench(ctx: &mut Criterion) {
-    let state = Lazy::new(RenderpassState::new);
+    let state = LazyLock::new(RenderpassState::new);
 
     let draw_count = draw_count();
     let vertex_buffer_count = draw_count * VERTEX_BUFFERS_PER_DRAW;
@@ -450,7 +450,7 @@ fn run_bench(ctx: &mut Criterion) {
             group.bench_function(
                 format!("{rpasses} renderpasses x {draws_per_pass} draws ({label})"),
                 |b| {
-                    Lazy::force(&state);
+                    LazyLock::force(&state);
 
                     b.iter_custom(|iters| {
                         profiling::scope!("benchmark invocation");
@@ -502,7 +502,7 @@ fn run_bench(ctx: &mut Criterion) {
     for threads in [2, 4, 8] {
         let draws_per_pass = draw_count / threads;
         group.bench_function(format!("{threads} threads x {draws_per_pass} draws"), |b| {
-            Lazy::force(&state);
+            LazyLock::force(&state);
 
             b.iter_custom(|iters| {
                 profiling::scope!("benchmark invocation");
@@ -541,7 +541,7 @@ fn run_bench(ctx: &mut Criterion) {
     group.throughput(Throughput::Elements(draw_count as _));
 
     group.bench_function(format!("{draw_count} draws"), |b| {
-        Lazy::force(&state);
+        LazyLock::force(&state);
 
         b.iter_custom(|iters| {
             profiling::scope!("benchmark invocation");
@@ -577,7 +577,7 @@ fn run_bench(ctx: &mut Criterion) {
             texture_count + vertex_buffer_count
         ),
         |b| {
-            Lazy::force(&state);
+            LazyLock::force(&state);
 
             b.iter(|| state.device_state.queue.submit([]));
         },

benches/benches/resource_creation.rs

@@ -1,13 +1,13 @@
 use std::time::{Duration, Instant};
 
 use criterion::{criterion_group, Criterion, Throughput};
-use once_cell::sync::Lazy;
 use rayon::iter::{IntoParallelIterator, ParallelIterator};
+use std::sync::LazyLock;
 
 use crate::DeviceState;
 
 fn run_bench(ctx: &mut Criterion) {
-    let state = Lazy::new(DeviceState::new);
+    let state = LazyLock::new(DeviceState::new);
 
     const RESOURCES_TO_CREATE: usize = 8;
 
@@ -19,7 +19,7 @@ fn run_bench(ctx: &mut Criterion) {
         group.bench_function(
             format!("{threads} threads x {resources_per_thread} resource"),
             |b| {
-                Lazy::force(&state);
+                LazyLock::force(&state);
 
                 b.iter_custom(|iters| {
                     profiling::scope!("benchmark invocation");

examples/src/boids/mod.rs

@@ -2,7 +2,6 @@
 // adapted from https://github.com/austinEng/webgpu-samples/blob/master/src/examples/computeBoids.ts
 
 use nanorand::{Rng, WyRand};
-use std::mem::size_of;
 use wgpu::util::DeviceExt;
 
 // number of boid particles to simulate

examples/src/bunnymark/mod.rs

@@ -1,6 +1,6 @@
 use bytemuck::{Pod, Zeroable};
 use nanorand::{Rng, WyRand};
-use std::{borrow::Cow, mem::size_of};
+use std::borrow::Cow;
 use wgpu::util::DeviceExt;
 use winit::{
     event::{ElementState, KeyEvent},

examples/src/cube/mod.rs

@@ -1,5 +1,5 @@
 use bytemuck::{Pod, Zeroable};
-use std::{f32::consts, mem::size_of};
+use std::f32::consts;
 use wgpu::util::DeviceExt;
 
 #[repr(C)]

examples/src/hello_compute/mod.rs

@@ -1,4 +1,4 @@
-use std::{mem::size_of_val, str::FromStr};
+use std::str::FromStr;
 use wgpu::util::DeviceExt;
 
 // Indicates a u32 overflow in an intermediate Collatz value

examples/src/hello_synchronization/mod.rs

@@ -1,5 +1,3 @@
-use std::mem::size_of_val;
-
 const ARR_SIZE: usize = 128;
 
 struct ExecuteResults {

examples/src/hello_workgroups/mod.rs

@@ -7,8 +7,6 @@
 //!
 //! Only parts specific to this example will be commented.
 
-use std::mem::size_of_val;
-
 use wgpu::util::DeviceExt;
 
 async fn run() {

examples/src/mipmap/mod.rs

@@ -1,5 +1,5 @@
 use bytemuck::{Pod, Zeroable};
-use std::{f32::consts, mem::size_of};
+use std::f32::consts;
 use wgpu::util::DeviceExt;
 
 const TEXTURE_FORMAT: wgpu::TextureFormat = wgpu::TextureFormat::Rgba8UnormSrgb;

examples/src/msaa_line/mod.rs

@@ -7,7 +7,7 @@
 //! *   Set the primitive_topology to PrimitiveTopology::LineList.
 //! *   Vertices and Indices describe the two points that make up a line.
 
-use std::{iter, mem::size_of};
+use std::iter;
 
 use bytemuck::{Pod, Zeroable};
 use wgpu::util::DeviceExt;

examples/src/repeated_compute/mod.rs

@@ -5,8 +5,6 @@
 //! hello-compute example does not such as mapping buffers
 //! and why use the async channels.
 
-use std::mem::size_of_val;
-
 const OVERFLOW: u32 = 0xffffffff;
 
 async fn run() {

examples/src/shadow/mod.rs

@@ -1,4 +1,4 @@
-use std::{f32::consts, iter, mem::size_of, ops::Range, sync::Arc};
+use std::{f32::consts, iter, ops::Range, sync::Arc};
 
 use bytemuck::{Pod, Zeroable};
 use wgpu::util::{align_to, DeviceExt};

examples/src/skybox/mod.rs

@@ -1,5 +1,5 @@
 use bytemuck::{Pod, Zeroable};
-use std::{f32::consts, mem::size_of};
+use std::f32::consts;
 use wgpu::{util::DeviceExt, AstcBlock, AstcChannel};
 
 const IMAGE_SIZE: u32 = 256;

examples/src/srgb_blend/mod.rs

@@ -1,5 +1,4 @@
 use bytemuck::{Pod, Zeroable};
-use std::mem::size_of;
 use wgpu::util::DeviceExt;
 
 #[repr(C)]

examples/src/stencil_triangles/mod.rs

@@ -1,5 +1,4 @@
 use bytemuck::{Pod, Zeroable};
-use std::mem::size_of;
 use wgpu::util::DeviceExt;
 
 #[repr(C)]

examples/src/storage_texture/mod.rs

@@ -14,8 +14,6 @@
 //! A lot of things aren't explained here via comments. See hello-compute and
 //! repeated-compute for code that is more thoroughly commented.
 
-use std::mem::size_of_val;
-
 #[cfg(not(target_arch = "wasm32"))]
 use crate::utils::output_image_native;
 #[cfg(target_arch = "wasm32")]

examples/src/texture_arrays/mod.rs

@@ -1,8 +1,5 @@
 use bytemuck::{Pod, Zeroable};
-use std::{
-    mem::size_of,
-    num::{NonZeroU32, NonZeroU64},
-};
+use std::num::{NonZeroU32, NonZeroU64};
 use wgpu::util::DeviceExt;
 
 #[repr(C)]