Bump

CHANGELOG.md

@@ -1,5 +1,10 @@
 # Changelog
 
+## v2.6.0
+
+- Added source code locations to the C and JS APIs
+- Significant performance improvements and code size reductions
+
 ## v2.5.0
 
 - Source code locations for tags and other tokens.

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "lol_html"
-version = "2.5.0"
+version = "2.6.0"
 authors = ["Ivan Nikulin <[email protected], [email protected]>"]
 license = "BSD-3-Clause"
 description = "Streaming HTML rewriter/parser with CSS selector-based API"
@@ -32,21 +32,21 @@ bitflags = "2.9.1"
 cfg-if = "1.0.1"
 cssparser = "0.35"
 encoding_rs = "0.8.35"
-hashbrown = "0.15.3"
-memchr = "2.7.4"
-mime = "0.3.16"
-selectors = "0.27"
+hashbrown = "0.15.4"
+memchr = "2.7.5"
+mime = "0.3.17"
+selectors = "0.30"
 thiserror = "2.0"
 precomputed-hash = "0.1.1"
 
 [dev-dependencies]
-criterion = "0.5.1"
+criterion = "0.6.0"
 # Needed for criterion <= v0.5.1. See https://github.com/bheisler/criterion.rs/pull/703.
-clap = { version = "4.5.21", features = ["help"] }
-glob = "0.3.0"
-html5ever = "0.29"
-markup5ever_rcdom = "0.5.0-unofficial"
-hashbrown = { version = "0.15.0", features = ["serde"] }
+clap = { version = "4.5.41", features = ["help"] }
+glob = "0.3.2"
+html5ever = "0.35"
+markup5ever_rcdom = "0.35.0"
+hashbrown = { version = "0.15.4", features = ["serde"] }
 serde = "1.0.219"
 serde_derive = "1.0.219"
 serde_json = "1.0.140"

benches/bench.rs

@@ -47,7 +47,7 @@ macro_rules! create_runner {
         move |b, i: &Vec<Vec<u8>>| {
             b.iter(|| {
                 let mut rewriter = lol_html::HtmlRewriter::new($settings, |c: &[u8]| {
-                    black_box(c);
+                    std::hint::black_box(c);
                 });
 
                 for chunk in i {
@@ -63,7 +63,7 @@ macro_rules! create_runner {
 macro_rules! noop_handler {
     () => {
         |arg| {
-            black_box(arg);
+            std::hint::black_box(arg);
             Ok(())
         }
     };

src/base/bytes.rs

@@ -7,7 +7,7 @@ use std::str;
 
 /// An error used to indicate that an encoded string has replacements and can't be converted losslessly.
 #[derive(Debug, Default, PartialEq, Eq, PartialOrd, Ord)]
-#[allow(unnameable_types, reason = "accidentally exposed via `tag.set_name()`")]
+#[allow(unnameable_types)] // accidentally exposed via `tag.set_name()`
 pub struct HasReplacementsError;
 
 /// A thin wrapper around byte slice with handy APIs attached
@@ -17,7 +17,7 @@ pub(crate) struct Bytes<'b>(&'b [u8]);
 
 /// A thin wrapper around either byte slice or owned bytes with some handy APIs attached
 #[derive(Clone, PartialEq, Eq, Hash)]
-#[allow(unnameable_types, reason = "accidentally exposed via `tag.set_name()`")]
+#[allow(unnameable_types)] // accidentally exposed via `tag.set_name()`
 #[repr(transparent)]
 pub struct BytesCow<'b>(Cow<'b, [u8]>);
 

src/html/text_type.rs

@@ -67,8 +67,8 @@ cfg_if! {
             }
         }
 
+        #[allow(clippy::fallible_impl_from)]
         impl<'s> From<&'s str> for TextType {
-            #[allow(clippy::fallible_impl_from)]
             fn from(text_type: &'s str) -> Self {
                 match text_type {
                     "Data state" => Self::Data,

src/memory/limited_vec.rs

@@ -21,6 +21,7 @@ impl<T> LimitedVec<T> {
     }
 
     pub fn push(&mut self, element: T) -> Result<(), MemoryLimitExceededError> {
+        #[allow(clippy::branches_sharing_code)]
         if self.vec.capacity() - self.vec.len() >= 1 {
             // the two push calls are optimized into one, but need to be two so each gets its own capacity hint
             self.vec.push(element);

src/selectors_vm/ast.rs

@@ -141,10 +141,10 @@ impl From<&Component<SelectorImplDescriptor>> for Condition {
                 ),
             )),
             Component::Nth(data) if data.ty == NthType::Child => {
-                Self::OnTagName(OnTagNameExpr::NthChild(NthChild::new(data.a, data.b)))
+                Self::OnTagName(OnTagNameExpr::NthChild(NthChild::new(data.an_plus_b.0, data.an_plus_b.1)))
             }
             Component::Nth(data) if data.ty == NthType::OfType => {
-                Self::OnTagName(OnTagNameExpr::NthOfType(NthChild::new(data.a, data.b)))
+                Self::OnTagName(OnTagNameExpr::NthOfType(NthChild::new(data.an_plus_b.0, data.an_plus_b.1)))
             }
             // NOTE: the rest of the components are explicit namespace or
             // pseudo class-related. Ideally none of them should appear in

src/transform_stream/dispatcher.rs

@@ -270,7 +270,7 @@ where
                 } => {
                     get_flags_from_aux_info_res!(aux_info_req, &attributes, self_closing)
                 }
-                _ => Err(ActionError::internal(
+                TagTokenOutline::EndTag { .. } => Err(ActionError::internal(
                     "Tag should be a start tag at this point",
                 )),
             },

tests/harness/suites/html5lib_tests/feedback_tests/expected_tokens.rs

@@ -3,9 +3,9 @@ use hashbrown::HashMap;
 use html5ever::tendril::StrTendril;
 use html5ever::tokenizer::{
     BufferQueue, TagKind, Token, TokenSink, TokenSinkResult, Tokenizer, TokenizerOpts,
-    TokenizerResult,
 };
 use html5ever::tree_builder::{TreeBuilder, TreeBuilderOpts};
+use html5ever::TokenizerResult;
 use markup5ever_rcdom::RcDom;
 use std::cell::RefCell;
 use std::iter::FromIterator;