General Rust fixes and updates

Update to Rust edition 2024, update dependency versions, and run Clippy to apply some suggested improvements.

Author: domenic

Cargo.lock

@@ -2,7 +2,7 @@
 name = "html-build"
 version = "0.0.0"
 publish = false
-edition = "2021"
+edition = "2024"
 
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
@@ -12,7 +12,7 @@ html5ever = "0.26.0"
 markup5ever_rcdom = "0.2.0"
 regex = "1"
 delegate = "0.12.0"
-url = "2.2.2"
+url = "2.5.4"
 
 [dev-dependencies]
 tempfile = "3"

Cargo.toml

@@ -1,4 +1,4 @@
-FROM rust:1.73-slim as builder
+FROM rust:1.88-slim as builder
 WORKDIR /whatwg/html-build
 COPY Cargo.lock Cargo.toml ./
 COPY src ./src/

Dockerfile

@@ -1,6 +1,6 @@
 # This Dockerfile is just used to run on Travis CI in an environment that can easily and repeatedly
 # install our build dependencies.
-FROM rust:1.73-slim as builder
+FROM rust:1.88-slim as builder
 WORKDIR /whatwg/html-build
 COPY Cargo.lock Cargo.toml ./
 COPY src ./src/

ci-build/Dockerfile

@@ -5,7 +5,7 @@ use std::io;
 use std::rc::Rc;
 
 use html5ever::tendril::StrTendril;
-use html5ever::{local_name, namespace_url, ns, LocalName, QualName};
+use html5ever::{LocalName, QualName, local_name, namespace_url, ns};
 use markup5ever_rcdom::{Handle, NodeData};
 
 use crate::dom_utils::{self, NodeHandleExt};
@@ -142,7 +142,7 @@ impl Processor {
                 let mut variant_comment = None;
                 let mut variant_str = None;
                 for node in description.iter() {
-                    if let NodeData::Comment { ref contents } = node.data {
+                    if let NodeData::Comment { contents } = &node.data {
                         if contents.trim().starts_with("or:") {
                             variant_comment = Some(node);
                             variant_str = Some(StrTendril::from(contents.trim()[3..].trim_start()));
@@ -158,7 +158,7 @@ impl Processor {
                         .children
                         .borrow()
                         .iter()
-                        .filter(|c| variant_comment.map_or(true, |vc| !Rc::ptr_eq(c, vc)))
+                        .filter(|c| variant_comment.is_none_or(|vc| !Rc::ptr_eq(c, vc)))
                         .map(|c| c.deep_clone())
                         .collect(),
                     variant: variant_str,
@@ -167,7 +167,7 @@ impl Processor {
                 if existing.default.is_empty() {
                     existing.default = descriptions.default;
                 } else if !descriptions.default.is_empty() {
-                    if let NodeData::Text { ref contents } = existing.default.last().unwrap().data {
+                    if let NodeData::Text { contents } = &existing.default.last().unwrap().data {
                         let mut borrow = contents.borrow_mut();
                         if let Some(last_non_ws) = borrow.rfind(|c: char| !c.is_ascii_whitespace())
                         {
@@ -209,13 +209,13 @@ impl Processor {
                     let mut has_special_semantics = false;
                     let mut key = None;
                     dom_utils::scan_dom(dd, &mut |n| match &n.data {
-                        NodeData::Comment { ref contents } if contents.trim() == "no-annotate" => {
+                        NodeData::Comment { contents } if contents.trim() == "no-annotate" => {
                             can_annotate = false;
                         }
-                        NodeData::Comment { ref contents } if contents.trim() == "variant" => {
+                        NodeData::Comment { contents } if contents.trim() == "variant" => {
                             wants_variant_description = true;
                         }
-                        NodeData::Text { ref contents }
+                        NodeData::Text { contents }
                             if contents.borrow().contains("has special semantics") =>
                         {
                             has_special_semantics = true;
@@ -257,7 +257,7 @@ impl Processor {
             };
             let mut description: Vec<Handle> = match descriptions {
                 Descriptions {
-                    variant: Some(ref variant),
+                    variant: Some(variant),
                     ..
                 } if wants_variant_description => {
                     parser::parse_fragment_async(variant[..].as_bytes(), &dd).await?
@@ -268,22 +268,22 @@ impl Processor {
                         format!(
                             "Attribute {key} wants variant description, but no <!--or--> was found"
                         ),
-                    ))
-                }
-                Descriptions { ref default, .. } => {
-                    default.iter().map(|n| n.deep_clone()).collect()
+                    ));
                 }
+                Descriptions { default, .. } => default.iter().map(|n| n.deep_clone()).collect(),
             };
 
             let mut dd_children = dd.children.borrow_mut();
             if has_special_semantics {
                 // Replace the trailing period with a separating colon.
-                if let Some(NodeData::Text { contents }) = dd_children.last_mut().map(|n| &n.data) {
-                    let mut text = contents.borrow_mut();
-                    *text = StrTendril::from(
-                        text.trim_end_matches(|c: char| c.is_ascii_whitespace() || c == '.'),
-                    );
-                    text.push_slice(": ");
+                if let Some(last) = dd_children.last_mut() {
+                    if let NodeData::Text { contents } = &last.data {
+                        let mut text = contents.borrow_mut();
+                        *text = StrTendril::from(
+                            text.trim_end_matches(|c: char| c.is_ascii_whitespace() || c == '.'),
+                        );
+                        text.push_slice(": ");
+                    }
                 }
             } else {
                 // Insert an em dash.

src/annotate_attributes.rs

@@ -7,7 +7,7 @@ use std::io;
 use std::path::{Path, PathBuf};
 
 use html5ever::tendril::{self, SendTendril};
-use html5ever::{local_name, Attribute, LocalName, QualName};
+use html5ever::{Attribute, LocalName, QualName, local_name};
 use markup5ever_rcdom::{Handle, NodeData};
 use tokio::fs::File;
 use tokio::task::JoinHandle;
@@ -48,11 +48,11 @@ impl Processor {
     /// Identifies replacements which will be needed, and starts the necessary
     /// I/O.
     pub fn visit(&mut self, node: &Handle) {
-        match node.data {
+        match &node.data {
             // BOILERPLATE comments will need to be replaced with their
             // corresponding HTML, parsed. Open the file so that we can do so on
             // demand.
-            NodeData::Comment { ref contents } if contents.starts_with("BOILERPLATE ") => {
+            NodeData::Comment { contents } if contents.starts_with("BOILERPLATE ") => {
                 let path = Path::new(contents[12..].trim());
                 let file = if is_safe_path(path) {
                     tokio::spawn(File::open(self.path.join(path)))
@@ -67,12 +67,8 @@ impl Processor {
             // Pseudo-comments can also appear in element attributes. These are
             // not parsed as HTML, so we simply want to read them into memory so
             // they can be replaced.
-            NodeData::Element { ref attrs, .. } => {
-                for Attribute {
-                    ref name,
-                    ref value,
-                } in attrs.borrow().iter()
-                {
+            NodeData::Element { attrs, .. } => {
+                for Attribute { name, value } in attrs.borrow().iter() {
                     if value.starts_with("<!--BOILERPLATE ") && value.ends_with("-->") {
                         let path = Path::new(value[16..value.len() - 3].trim());
                         let file_contents = if is_safe_path(path) {
@@ -94,18 +90,18 @@ impl Processor {
             // <pre> and <pre><code> which contain EXAMPLE also need to be
             // replaced, but as plain text. These are loaded from the "examples"
             // directory instead.
-            NodeData::Text { ref contents } => {
+            NodeData::Text { contents } => {
                 let borrowed_contents = contents.borrow();
                 let text = borrowed_contents.trim();
                 if !text.starts_with("EXAMPLE ") {
                     return;
                 }
                 const PRE: LocalName = local_name!("pre");
                 const CODE: LocalName = local_name!("code");
-                let has_suitable_parent = node.parent_node().map_or(false, |p| {
+                let has_suitable_parent = node.parent_node().is_some_and(|p| {
                     p.is_html_element(&PRE)
                         || (p.is_html_element(&CODE)
-                            && p.parent_node().map_or(false, |p2| p2.is_html_element(&PRE)))
+                            && p.parent_node().is_some_and(|p2| p2.is_html_element(&PRE)))
                 });
                 if has_suitable_parent {
                     let path = Path::new(text[8..].trim());
@@ -179,7 +175,8 @@ mod tests {
         proc.apply().await?;
         assert_eq!(
             serialize_for_test(&[document]),
-            "<!DOCTYPE html><html><head></head><body><table><tbody><tr><td>en</td><td>English</td></tr></tbody></table></body></html>");
+            "<!DOCTYPE html><html><head></head><body><table><tbody><tr><td>en</td><td>English</td></tr></tbody></table></body></html>"
+        );
         Ok(())
     }
 
@@ -200,7 +197,8 @@ mod tests {
         proc.apply().await?;
         assert_eq!(
             serialize_for_test(&[document]),
-            "<!DOCTYPE html><html><head></head><body><a href=\"data:text/html,Hello, world!\">hello</a></body></html>");
+            "<!DOCTYPE html><html><head></head><body><a href=\"data:text/html,Hello, world!\">hello</a></body></html>"
+        );
         Ok(())
     }
 
@@ -218,7 +216,8 @@ mod tests {
         proc.apply().await?;
         assert_eq!(
             serialize_for_test(&[document]),
-            "<!DOCTYPE html><html><head></head><body><pre>first</pre><pre><code class=\"html\">second</code></pre><p>EXAMPLE ignored</p></body></html>" );
+            "<!DOCTYPE html><html><head></head><body><pre>first</pre><pre><code class=\"html\">second</code></pre><p>EXAMPLE ignored</p></body></html>"
+        );
         Ok(())
     }
 

src/boilerplate.rs

@@ -2,7 +2,7 @@ use std::cell::RefCell;
 use std::rc::Rc;
 
 use html5ever::tendril::StrTendril;
-use html5ever::{local_name, namespace_url, ns, Attribute, LocalName, QualName};
+use html5ever::{Attribute, LocalName, QualName, local_name, namespace_url, ns};
 use markup5ever_rcdom::{Handle, Node, NodeData};
 
 /// Extensions to the DOM interface to make manipulation more ergonimc.
@@ -231,7 +231,7 @@ impl NodeHandleExt for Handle {
                 name:
                     QualName {
                         ns: ns!(html),
-                        ref local,
+                        local,
                         ..
                     },
                 ..
@@ -247,7 +247,7 @@ impl NodeHandleExt for Handle {
             local: local_name!("class"),
         };
         self.get_attribute(&CLASS)
-            .map_or(false, |v| v.split_ascii_whitespace().any(|c| c == class))
+            .is_some_and(|v| v.split_ascii_whitespace().any(|c| c == class))
     }
 
     fn has_any_class(&self, classes: &[&str]) -> bool {
@@ -256,22 +256,21 @@ impl NodeHandleExt for Handle {
             ns: ns!(),
             local: local_name!("class"),
         };
-        self.get_attribute(&CLASS).map_or(false, |v| {
-            v.split_ascii_whitespace().any(|c| classes.contains(&c))
-        })
+        self.get_attribute(&CLASS)
+            .is_some_and(|v| v.split_ascii_whitespace().any(|c| classes.contains(&c)))
     }
 
     fn node_text(&self) -> Option<StrTendril> {
         match &self.data {
-            NodeData::Text { ref contents } => Some(contents.borrow().clone()),
+            NodeData::Text { contents } => Some(contents.borrow().clone()),
             _ => None,
         }
     }
 
     fn text_content(&self) -> StrTendril {
         let mut text = StrTendril::new();
         scan_dom(self, &mut |n| {
-            if let NodeData::Text { ref contents } = &n.data {
+            if let NodeData::Text { contents } = &n.data {
                 text.push_tendril(&contents.borrow());
             }
         });