perf: add precomputed-hash for CachedSource (#128)

* perf: add precomputed-cache for `CachedSource`

* perf: case

* perf: case

Author: h-a-n-a

benches/bench.rs

@@ -1,14 +1,16 @@
 #![allow(missing_docs)]
 
+use std::collections::HashMap;
+
 #[cfg(not(codspeed))]
 pub use criterion::*;
 
 #[cfg(codspeed)]
 pub use codspeed_criterion_compat::*;
 
 use rspack_sources::{
-  CachedSource, ConcatSource, MapOptions, ReplaceSource, Source, SourceMap,
-  SourceMapSource, SourceMapSourceOptions,
+  BoxSource, CachedSource, ConcatSource, MapOptions, ReplaceSource, Source,
+  SourceExt, SourceMap, SourceMapSource, SourceMapSourceOptions,
 };
 
 const HELLOWORLD_JS: &str = include_str!(concat!(
@@ -195,6 +197,61 @@ fn benchmark_replace_large_minified_source(b: &mut Bencher) {
   });
 }
 
+fn benchmark_concat_generate_string_with_cache_as_key(b: &mut Bencher) {
+  let sms_minify = SourceMapSource::new(SourceMapSourceOptions {
+    value: HELLOWORLD_MIN_JS,
+    name: "helloworld.min.js",
+    source_map: SourceMap::from_json(HELLOWORLD_MIN_JS_MAP).unwrap(),
+    original_source: Some(HELLOWORLD_JS.to_string()),
+    inner_source_map: Some(SourceMap::from_json(HELLOWORLD_JS_MAP).unwrap()),
+    remove_original_source: false,
+  });
+  let sms_rollup = SourceMapSource::new(SourceMapSourceOptions {
+    value: BUNDLE_JS,
+    name: "bundle.js",
+    source_map: SourceMap::from_json(BUNDLE_JS_MAP).unwrap(),
+    original_source: None,
+    inner_source_map: None,
+    remove_original_source: false,
+  });
+  let concat = ConcatSource::new([sms_minify, sms_rollup]);
+  let cached = CachedSource::new(concat).boxed();
+
+  b.iter(|| {
+    let mut m = HashMap::<BoxSource, ()>::new();
+    m.insert(cached.clone(), ());
+    let _ = black_box(|| m.get(&cached));
+    let _ = black_box(|| m.get(&cached));
+  })
+}
+
+fn benchmark_concat_generate_string_as_key(b: &mut Bencher) {
+  let sms_minify = SourceMapSource::new(SourceMapSourceOptions {
+    value: HELLOWORLD_MIN_JS,
+    name: "helloworld.min.js",
+    source_map: SourceMap::from_json(HELLOWORLD_MIN_JS_MAP).unwrap(),
+    original_source: Some(HELLOWORLD_JS.to_string()),
+    inner_source_map: Some(SourceMap::from_json(HELLOWORLD_JS_MAP).unwrap()),
+    remove_original_source: false,
+  });
+  let sms_rollup = SourceMapSource::new(SourceMapSourceOptions {
+    value: BUNDLE_JS,
+    name: "bundle.js",
+    source_map: SourceMap::from_json(BUNDLE_JS_MAP).unwrap(),
+    original_source: None,
+    inner_source_map: None,
+    remove_original_source: false,
+  });
+  let concat = ConcatSource::new([sms_minify, sms_rollup]).boxed();
+
+  b.iter(|| {
+    let mut m = HashMap::<BoxSource, ()>::new();
+    m.insert(concat.clone(), ());
+    let _ = black_box(|| m.get(&concat));
+    let _ = black_box(|| m.get(&concat));
+  })
+}
+
 fn bench_rspack_sources(criterion: &mut Criterion) {
   let mut group = criterion.benchmark_group("rspack_sources");
   group.bench_function(
@@ -213,6 +270,14 @@ fn bench_rspack_sources(criterion: &mut Criterion) {
     "replace_large_minified_source",
     benchmark_replace_large_minified_source,
   );
+  group.bench_function(
+    "concat_generate_string_with_cache_as_key",
+    benchmark_concat_generate_string_with_cache_as_key,
+  );
+  group.bench_function(
+    "concat_generate_string_as_key",
+    benchmark_concat_generate_string_as_key,
+  );
   group.finish();
 }
 

src/cached_source.rs

@@ -1,6 +1,6 @@
 use std::{
   borrow::Cow,
-  hash::{BuildHasherDefault, Hash},
+  hash::{BuildHasherDefault, Hash, Hasher},
   sync::{Arc, OnceLock},
 };
 
@@ -52,6 +52,7 @@ pub struct CachedSource<T> {
   inner: Arc<T>,
   cached_buffer: Arc<OnceLock<Vec<u8>>>,
   cached_source: Arc<OnceLock<Arc<str>>>,
+  cached_hash: Arc<OnceLock<u64>>,
   cached_maps:
     Arc<DashMap<MapOptions, Option<SourceMap>, BuildHasherDefault<FxHasher>>>,
 }
@@ -63,6 +64,7 @@ impl<T> CachedSource<T> {
       inner: Arc::new(inner),
       cached_buffer: Default::default(),
       cached_source: Default::default(),
+      cached_hash: Default::default(),
       cached_maps: Default::default(),
     }
   }
@@ -89,11 +91,7 @@ impl<T: Source + Hash + PartialEq + Eq + 'static> Source for CachedSource<T> {
   }
 
   fn size(&self) -> usize {
-    let source = self.cached_source.get();
-    match source {
-      Some(source) => source.len(),
-      None => self.inner.size(),
-    }
+    self.source().len()
   }
 
   fn map(&self, options: &MapOptions) -> Option<SourceMap> {
@@ -160,20 +158,26 @@ impl<T: Source + Hash + PartialEq + Eq + 'static> StreamChunks<'_>
   }
 }
 
-impl<T: Source> Clone for CachedSource<T> {
+impl<T> Clone for CachedSource<T> {
   fn clone(&self) -> Self {
     Self {
       inner: self.inner.clone(),
       cached_buffer: self.cached_buffer.clone(),
       cached_source: self.cached_source.clone(),
+      cached_hash: self.cached_hash.clone(),
       cached_maps: self.cached_maps.clone(),
     }
   }
 }
 
-impl<T: Hash> Hash for CachedSource<T> {
+impl<T: Source + Hash + PartialEq + Eq + 'static> Hash for CachedSource<T> {
   fn hash<H: std::hash::Hasher>(&self, state: &mut H) {
-    self.inner.hash(state);
+    (self.cached_hash.get_or_init(|| {
+      let mut hasher = FxHasher::default();
+      hasher.write(self.source().as_bytes());
+      hasher.finish()
+    }))
+    .hash(state);
   }
 }
 

src/source.rs

@@ -522,7 +522,7 @@ mod tests {
     RawSource::from("g").boxed().hash(&mut state);
     (&RawSource::from("h") as &dyn Source).hash(&mut state);
     ReplaceSource::new(RawSource::from("i").boxed()).hash(&mut state);
-    assert_eq!(format!("{:x}", state.finish()), "8163b42b7cb1d8f0");
+    assert_eq!(format!("{:x}", state.finish()), "ef733b8b8ee61bf0");
   }
 
   #[test]