refactor: reduce memory consumption of CachedSource (#144)

* refactor: init

* perf: faster lines iterator for `Rope` (#145)

* refactor: init

* refactor: faster lines

* refactor: try

* chore: clippy

* test: more

* chore: more

* perf

* finish

* finish

Author: h-a-n-a

Cargo.lock

@@ -14,18 +14,18 @@ readme = "README.md"
 include = ["/src/**/*.rs", "/*.toml", "/LICENSE", "/README.md"]
 
 [lints.rust]
-unsafe_code  = "warn"
+unsafe_code = "warn"
 missing_docs = "warn"
 
 [lints.clippy]
-dbg_macro     = "warn"
-todo          = "warn"
+dbg_macro = "warn"
+todo = "warn"
 unimplemented = "warn"
-print_stdout  = "warn"
-print_stderr  = "warn"
+print_stdout = "warn"
+print_stderr = "warn"
 
 [[bench]]
-name    = "bench"
+name = "bench"
 path = "benches/bench.rs"
 harness = false
 
@@ -36,6 +36,8 @@ dyn-clone = "1"
 rustc-hash = "1"
 dashmap = "5"
 memchr = "2.6.4"
+itertools = "0.13"
+
 
 codspeed-criterion-compat = { version = "2.3.3", default-features = false, optional = true }
 static_assertions = "1.1.0"

Cargo.toml

@@ -12,6 +12,7 @@ use crate::{
     stream_and_get_source_and_map, stream_chunks_of_raw_source,
     stream_chunks_of_source_map, StreamChunks,
   },
+  rope::Rope,
   MapOptions, Source, SourceMap,
 };
 
@@ -50,8 +51,6 @@ use crate::{
 /// ```
 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>>>,
@@ -62,8 +61,6 @@ impl<T> CachedSource<T> {
   pub fn new(inner: T) -> Self {
     Self {
       inner: Arc::new(inner),
-      cached_buffer: Default::default(),
-      cached_source: Default::default(),
       cached_hash: Default::default(),
       cached_maps: Default::default(),
     }
@@ -77,17 +74,15 @@ impl<T> CachedSource<T> {
 
 impl<T: Source + Hash + PartialEq + Eq + 'static> Source for CachedSource<T> {
   fn source(&self) -> Cow<str> {
-    let cached = self
-      .cached_source
-      .get_or_init(|| self.inner.source().into());
-    Cow::Borrowed(cached)
+    self.inner.source()
+  }
+
+  fn rope(&self) -> Rope<'_> {
+    self.inner.rope()
   }
 
   fn buffer(&self) -> Cow<[u8]> {
-    let cached = self
-      .cached_buffer
-      .get_or_init(|| self.inner.buffer().to_vec());
-    Cow::Borrowed(cached)
+    self.inner.buffer()
   }
 
   fn size(&self) -> usize {
@@ -109,7 +104,7 @@ impl<T: Source + Hash + PartialEq + Eq + 'static> Source for CachedSource<T> {
   }
 }
 
-impl<T: Source + Hash + PartialEq + Eq + 'static> StreamChunks<'_>
+impl<T: Source + Hash + PartialEq + Eq + 'static> StreamChunks
   for CachedSource<T>
 {
   fn stream_chunks<'a>(
@@ -122,9 +117,7 @@ impl<T: Source + Hash + PartialEq + Eq + 'static> StreamChunks<'_>
     let cached_map = self.cached_maps.entry(options.clone());
     match cached_map {
       Entry::Occupied(entry) => {
-        let source = self
-          .cached_source
-          .get_or_init(|| self.inner.source().into());
+        let source = self.rope();
         if let Some(map) = entry.get() {
           #[allow(unsafe_code)]
           // SAFETY: We guarantee that once a `SourceMap` is stored in the cache, it will never be removed.
@@ -162,8 +155,6 @@ 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(),
     }
@@ -196,17 +187,14 @@ impl<T: std::fmt::Debug> std::fmt::Debug for CachedSource<T> {
   ) -> Result<(), std::fmt::Error> {
     f.debug_struct("CachedSource")
       .field("inner", self.inner.as_ref())
-      .field("cached_buffer", &self.cached_buffer.get().is_some())
-      .field("cached_source", &self.cached_source.get().is_some())
+      .field("cached_hash", self.cached_hash.as_ref())
       .field("cached_maps", &(!self.cached_maps.is_empty()))
       .finish()
   }
 }
 
 #[cfg(test)]
 mod tests {
-  use std::borrow::Borrow;
-
   use crate::{
     ConcatSource, OriginalSource, RawSource, ReplaceSource, SourceExt,
     SourceMapSource, WithoutOriginalOptions,
@@ -247,11 +235,6 @@ mod tests {
     source.size();
     source.map(&map_options);
 
-    assert_eq!(clone.cached_source.get().unwrap().borrow(), source.source());
-    assert_eq!(
-      *clone.cached_buffer.get().unwrap(),
-      source.buffer().to_vec()
-    );
     assert_eq!(
       *clone.cached_maps.get(&map_options).unwrap().value(),
       source.map(&map_options)

src/cached_source.rs

@@ -10,7 +10,7 @@ use crate::{
   helpers::{get_map, GeneratedInfo, OnChunk, OnName, OnSource, StreamChunks},
   linear_map::LinearMap,
   source::{Mapping, OriginalLocation},
-  BoxSource, MapOptions, Source, SourceExt, SourceMap,
+  BoxSource, MapOptions, Rope, Source, SourceExt, SourceMap,
 };
 
 /// Concatenate multiple [Source]s to a single [Source].
@@ -109,6 +109,20 @@ impl Source for ConcatSource {
     }
   }
 
+  fn rope(&self) -> Rope<'_> {
+    let children = self.children();
+    if children.len() == 1 {
+      children[0].rope()
+    } else {
+      let mut rope = Rope::new();
+      for child in children {
+        let child_rope = child.rope();
+        rope.append(child_rope);
+      }
+      rope
+    }
+  }
+
   fn buffer(&self) -> Cow<[u8]> {
     let children = self.children();
     if children.len() == 1 {
@@ -155,8 +169,8 @@ impl PartialEq for ConcatSource {
 }
 impl Eq for ConcatSource {}
 
-impl<'a> StreamChunks<'a> for ConcatSource {
-  fn stream_chunks(
+impl StreamChunks for ConcatSource {
+  fn stream_chunks<'a>(
     &'a self,
     options: &MapOptions,
     on_chunk: OnChunk<'_, 'a>,

src/concat_source.rs

@@ -8,12 +8,15 @@ pub type Result<T> = result::Result<T, Error>;
 pub enum Error {
   /// a JSON parsing related failure
   BadJson(simd_json::Error),
+  /// rope related failure
+  Rope(&'static str),
 }
 
 impl fmt::Display for Error {
   fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
     match self {
       Error::BadJson(err) => write!(f, "bad json: {err}"),
+      Error::Rope(err) => write!(f, "rope error: {err}"),
     }
   }
 }