qknight · February 19, 2025 12:37
diff --git a/gistfile1.txt b/gistfile1.txt
 # main.rs

 #[macro_use]
 extern crate html5ever;
 extern crate markup5ever_rcdom as rcdom;

 use html5ever::parse_document;
 use html5ever::tendril::TendrilSink;
 use rcdom::{Handle, NodeData, RcDom};
 use std::io::Cursor;
 // This is not proper HTML serialization, of course.

 fn walk(indent: usize, handle: &Handle) {
    let node = handle;
    for _ in 0..indent {
        print!(" ");
    }
    let close_tag = match node.data {
        NodeData::Document => {println!(""); None },

        NodeData::Doctype {
            ref name,
            ref public_id,
            ref system_id,
        } => {
            let mut doctype = format!("<!DOCTYPE {}", name);
            if !public_id.is_empty() {
                doctype += &format!(" \"{}\"", public_id);
            }
            if !system_id.is_empty() {
                doctype += &format!(" \"{}\"", system_id);
            }            
            doctype += ">";
            println!("{}", doctype);
            None
        },

        NodeData::Text { ref contents } => {
            println!("{}", contents.borrow().escape_default());
            None
        },

        NodeData::Comment { ref contents } => {
            println!("<!-- {} -->", contents.escape_default());
            None
        },

        NodeData::Element {
            ref name,
            ref attrs,
            ..
        } => {
            assert!(name.ns == ns!(html));
            print!("<{}", name.local);
            for attr in attrs.borrow().iter() {
                assert!(attr.name.ns == ns!());
                print!(" {}=\"{}\"", attr.name.local, attr.value);
            }
            println!(">");
            
            Some(format!("</{}>", name.local))
            
        },

        NodeData::ProcessingInstruction { .. } => {
            unreachable!();
            None
        },
    };

    for child in node.children.borrow().iter() {
        walk(indent + 2, child);
    }

    match close_tag {
        Some(tag) => {
            println!("{}{}", " ".repeat(indent), tag);
        },
        None => {}
    }
 }

 fn main() {
     let html = "<!DOCTYPE html><!-- asdf --><html><head><title>Test</title></head><body><h1>Hello, World!</h1></body></html>";
 let dom = parse_document(RcDom::default(), Default::default())
        .from_utf8()
        .read_from(&mut Cursor::new(html.as_bytes()))
        .unwrap();
    walk(0, &dom.document);

    if !dom.errors.borrow().is_empty() {
        println!("\nParse errors:");
        for err in dom.errors.borrow().iter() {
            println!("    {err}");
        }
    }
 }

 # Cargo.toml

 [package]
 name = "z"
 version = "0.1.0"
 edition = "2021"

 [dependencies]
 chrono = "0.4.39"
 html5ever = { version = "0.29.0", features = ["trace_tokenizer"] }
 lazy_static = "1.5.0"
 markup5ever = "0.14.0"
 markup5ever_rcdom = "0.5.0-unofficial"
 prettyish-html = "0.1.1"
 regex = "1.11.1"
 time-util = { version = "0.3.4", features = ["chrono"] }

    Finished `dev` profile [unoptimized + debuginfo] target(s) in 0.41s
     Running `target/debug/z`

 # output

  <!DOCTYPE html>
  <!--  asdf  -->
  <html>
    <head>
      <title>
        Test
      </title>
    </head>
    <body>
      <h1>
        Hello, World!
      </h1>
    </body>
  </html>
	# main.rs

	#[macro_use]
	extern crate html5ever;
	extern crate markup5ever_rcdom as rcdom;

	use html5ever::parse_document;
	use html5ever::tendril::TendrilSink;
	use rcdom::{Handle, NodeData, RcDom};
	use std::io::Cursor;
	// This is not proper HTML serialization, of course.

	fn walk(indent: usize, handle: &Handle) {
	let node = handle;
	for _ in 0..indent {
	print!(" ");
	}
	let close_tag = match node.data {
	NodeData::Document => {println!(""); None },

	NodeData::Doctype {
	ref name,
	ref public_id,
	ref system_id,
	} => {
	let mut doctype = format!("<!DOCTYPE {}", name);
	if !public_id.is_empty() {
	doctype += &format!(" \"{}\"", public_id);
	}
	if !system_id.is_empty() {
	doctype += &format!(" \"{}\"", system_id);
	}
	doctype += ">";
	println!("{}", doctype);
	None
	},

	NodeData::Text { ref contents } => {
	println!("{}", contents.borrow().escape_default());
	None
	},

	NodeData::Comment { ref contents } => {
	println!("<!-- {} -->", contents.escape_default());
	None
	},

	NodeData::Element {
	ref name,
	ref attrs,
	..
	} => {
	assert!(name.ns == ns!(html));
	print!("<{}", name.local);
	for attr in attrs.borrow().iter() {
	assert!(attr.name.ns == ns!());
	print!(" {}=\"{}\"", attr.name.local, attr.value);
	}
	println!(">");

	Some(format!("</{}>", name.local))

	},

	NodeData::ProcessingInstruction { .. } => {
	unreachable!();
	None
	},
	};

	for child in node.children.borrow().iter() {
	walk(indent + 2, child);
	}

	match close_tag {
	Some(tag) => {
	println!("{}{}", " ".repeat(indent), tag);
	},
	None => {}
	}
	}

	fn main() {
	let html = "<!DOCTYPE html><!-- asdf --><html><head><title>Test</title></head><body><h1>Hello, World!</h1></body></html>";
	let dom = parse_document(RcDom::default(), Default::default())
	.from_utf8()
	.read_from(&mut Cursor::new(html.as_bytes()))
	.unwrap();
	walk(0, &dom.document);

	if !dom.errors.borrow().is_empty() {
	println!("\nParse errors:");
	for err in dom.errors.borrow().iter() {
	println!(" {err}");
	}
	}
	}

	# Cargo.toml

	[package]
	name = "z"
	version = "0.1.0"
	edition = "2021"

	[dependencies]
	chrono = "0.4.39"
	html5ever = { version = "0.29.0", features = ["trace_tokenizer"] }
	lazy_static = "1.5.0"
	markup5ever = "0.14.0"
	markup5ever_rcdom = "0.5.0-unofficial"
	prettyish-html = "0.1.1"
	regex = "1.11.1"
	time-util = { version = "0.3.4", features = ["chrono"] }

	Finished `dev` profile [unoptimized + debuginfo] target(s) in 0.41s
	Running `target/debug/z`

	# output

	<!DOCTYPE html>
	<!-- asdf -->
	<html>
	<head>
	<title>
	Test
	</title>
	</head>
	<body>
	<h1>
	Hello, World!
	</h1>
	</body>
	</html>