chore: bump to 0.5.4, fix clippy in benches/examples

- Version bump from 0.5.3 to 0.5.4
- Migrate remaining archmage 0.1 references in benches/examples to 0.4
  (Avx2Token → X64V3Token)
- Fix clippy lints: is_multiple_of, needless_borrow, allow(deprecated)

Author: lilith

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "mozjpeg-rs"
-version = "0.5.3"
+version = "0.5.4"
 edition = "2021"
 rust-version = "1.89.0"
 license = "BSD-3-Clause"

benches/dct.rs

@@ -88,11 +88,10 @@ fn bench_dct_combined(c: &mut Criterion) {
 
 #[cfg(target_arch = "x86_64")]
 fn bench_dct_archmage(c: &mut Criterion) {
-    use archmage::tokens::x86::Avx2Token;
-    use archmage::SimdToken;
+    use archmage::{SimdToken, X64V3Token};
     use mozjpeg_rs::dct::avx2_archmage::forward_dct_8x8_i32;
 
-    let Some(token) = Avx2Token::try_new() else {
+    let Some(token) = X64V3Token::try_new() else {
         eprintln!("AVX2 not available, skipping benchmark");
         return;
     };
@@ -113,11 +112,10 @@ fn bench_dct_archmage(c: &mut Criterion) {
 
 #[cfg(target_arch = "x86_64")]
 fn bench_dct_archmage_i16(c: &mut Criterion) {
-    use archmage::tokens::x86::Avx2Token;
-    use archmage::SimdToken;
+    use archmage::{SimdToken, X64V3Token};
     use mozjpeg_rs::dct::avx2_archmage::forward_dct_8x8_i16;
 
-    let Some(token) = Avx2Token::try_new() else {
+    let Some(token) = X64V3Token::try_new() else {
         eprintln!("AVX2 not available, skipping benchmark");
         return;
     };
@@ -215,11 +213,10 @@ fn bench_dct_batch_1000(c: &mut Criterion) {
 
     #[cfg(target_arch = "x86_64")]
     {
-        use archmage::tokens::x86::Avx2Token;
-        use archmage::SimdToken;
+        use archmage::{SimdToken, X64V3Token};
         use mozjpeg_rs::dct::avx2_archmage::{forward_dct_8x8_i16, forward_dct_8x8_i32};
 
-        if let Some(token) = Avx2Token::try_new() {
+        if let Some(token) = X64V3Token::try_new() {
             group.bench_function("archmage_i32", |b| {
                 b.iter(|| {
                     let mut coeffs = [0i16; 64];

examples/debug_dct.rs

@@ -1,9 +1,9 @@
 //! Debug DCT by tracing intermediate values
 
-#[cfg(all(target_arch = "x86_64", target_feature = "avx2"))]
-use archmage::tokens::x86::Avx2Token;
 #[cfg(all(target_arch = "x86_64", target_feature = "avx2"))]
 use archmage::SimdToken;
+#[cfg(all(target_arch = "x86_64", target_feature = "avx2"))]
+use archmage::X64V3Token;
 
 const DCTSIZE: usize = 8;
 const DCTSIZE2: usize = 64;
@@ -212,7 +212,7 @@ fn scalar_dct(samples: &[i16; DCTSIZE2], coeffs: &mut [i16; DCTSIZE2]) {
 #[target_feature(enable = "avx2")]
 unsafe fn avx2_i16_dct_debug(samples: &[i16; DCTSIZE2], coeffs: &mut [i16; DCTSIZE2]) {
     // SAFETY: We're inside a #[target_feature(enable = "avx2")] function
-    let token = Avx2Token::new_unchecked();
+    let token = X64V3Token::try_new().unwrap();
     mozjpeg_rs::dct::avx2::forward_dct_8x8_avx2_i16(token, samples, coeffs);
 }
 

examples/profile_dct.rs

@@ -80,10 +80,9 @@ fn main() {
     // Profile AVX2 archmage intrinsics
     #[cfg(all(target_arch = "x86_64", target_feature = "avx2"))]
     {
-        use archmage::tokens::x86::Avx2Token;
-        use archmage::SimdToken;
+        use archmage::{SimdToken, X64V3Token};
 
-        if let Some(token) = Avx2Token::try_new() {
+        if let Some(token) = X64V3Token::try_new() {
             let start = Instant::now();
             for _ in 0..ITERATIONS {
                 forward_dct_8x8_i32(token, black_box(&samples), black_box(&mut coeffs));

examples/synth_dct_overflow.rs

@@ -1,3 +1,4 @@
+#![allow(deprecated)] // Uses deprecated avx2_archmage module for overflow testing
 //! Detect i16 DCT overflow with synthetic worst-case patterns.
 //!
 //! Half-black/half-white blocks within a single 8x8 DCT block produce the largest
@@ -96,7 +97,7 @@ fn main() {
         }
 
         use archmage::SimdToken;
-        let token = archmage::Avx2Token::try_new().unwrap();
+        let token = archmage::X64V3Token::try_new().unwrap();
         let mut coeffs_i16 = [0i16; DCTSIZE2];
         let mut coeffs_i32 = [0i16; DCTSIZE2];
         #[allow(deprecated)]

examples/trellis_noise_bench.rs

@@ -324,7 +324,11 @@ fn make_high_freq_checker(w: u32, h: u32) -> Vec<u8> {
     let mut rgb = vec![0u8; (w * h * 3) as usize];
     for y in 0..h {
         for x in 0..w {
-            let val = if (x + y) % 2 == 0 { 200u8 } else { 55u8 };
+            let val = if (x + y).is_multiple_of(2) {
+                200u8
+            } else {
+                55u8
+            };
             let i = (y * w + x) as usize * 3;
             rgb[i] = val;
             rgb[i + 1] = val;
@@ -382,7 +386,7 @@ fn make_salt_pepper(w: u32, h: u32, density: f64, seed: u64) -> Vec<u8> {
     for pixel in rgb.chunks_mut(3) {
         let r = lcg(&mut s);
         if r < threshold {
-            let val = if r % 2 == 0 { 0u8 } else { 255u8 };
+            let val = if r.is_multiple_of(2) { 0u8 } else { 255u8 };
             pixel[0] = val;
             pixel[1] = val;
             pixel[2] = val;

examples/trellis_speed_bench.rs

@@ -42,7 +42,7 @@ fn main() {
 
     let png_files = png_files_in_dir(&cid22_dir);
     for path in &png_files {
-        if let Some((rgb, w, h)) = load_png(&path) {
+        if let Some((rgb, w, h)) = load_png(path) {
             let name = path
                 .file_name()
                 .unwrap_or_default()

examples/verify_dct.rs

@@ -20,10 +20,9 @@ fn main() {
 
     #[cfg(target_arch = "x86_64")]
     {
-        use archmage::tokens::x86::Avx2Token;
-        use archmage::SimdToken;
+        use archmage::{SimdToken, X64V3Token};
 
-        if let Some(token) = Avx2Token::try_new() {
+        if let Some(token) = X64V3Token::try_new() {
             mozjpeg_rs::dct::avx2_archmage::forward_dct_8x8_i32(token, &samples, &mut coeffs_new);
             println!(
                 "\nNew (archmage) DCT output (first 8): {:?}",