Preserve whitespace when serializing without indentation, and do so by default

README.md

@@ -68,8 +68,5 @@ body = html.create_element("body")
 body.create_text("Hello Everyone ...")
 
 print(root.serialize())
-# <!DOCTYPE html>
-# <html lang="en">
-#   <body>Hello Everyone ...</body>
-# </html>
+# <!DOCTYPE html><html lang="en"><body>Hello Everyone ...</body></html>
 ```

docs/docs/parser.md

@@ -89,14 +89,9 @@ title.text()
 title.parent.name
 # QualName(local="head", ns="http://www.w3.org/1999/xhtml", prefix=None)
 
-ul = root.select_one("ul")
-ul.serialize()
-# <ul>
-#     <li><a href="https://www.example.com">Example Website</a></li>
-#     <li><a href="https://www.wikipedia.org">Wikipedia</a></li>
-#     <li><a href="https://www.bbc.com">BBC</a></li>
-#     <li><a href="https://www.microsoft.com">Microsoft</a></li>
-# </ul>
+li = root.select_one("li")
+li.serialize()
+# <li><a href="https://www.example.com">Example Website</a></li>
 ```
 
 !!! tip "Common task"
@@ -128,6 +123,28 @@ root.serialize()
 # </body></html>
 ```
 
+!!! warning "Pretty-printing"
+
+    Pass `indent` to pretty-print the output, indenting each level of elements by that many spaces:
+    ```python
+    root.serialize(indent=4)
+    # <!DOCTYPE html>
+    # <html>
+    #     <head>
+    #         <title>Incomplete Html</title>
+    #     </head>
+    #     <body>
+    #         <ul>
+    #             <li>
+    #                 <a href="https://www.example.com">Example Website</a>
+    #             </li>
+    # ...
+    ```
+
+    Pretty-printing adds and removes whitespace, which can change the meaning of the document: for example, it
+    changes the rendering of whitespace between inline elements, and the contents of `<pre>`, `<textarea>`, and
+    `<script>` elements. Use it only for display, and leave `indent` at its default of 0 to serialize the document exactly.
+
 ## Parsing XML
 Imagine this **`file.xml`** file:
 
@@ -187,13 +204,9 @@ for i in root.select("mag|*"):  # get all elements which has namespace 'mag'
 # Element(name=QualName(local="publisher", ns="http://www.example.com/magazines", prefix=Some("mag")), attrs=[], template=false, mathml_annotation_xml_integration_point=false)
 # Element(name=QualName(local="month", ns="http://www.example.com/magazines", prefix=Some("mag")), attrs=[], template=false, mathml_annotation_xml_integration_point=false)
 
-book = root.select_one("book")
-book.serialize()
-# <bk:book xmlns:bk="http://www.example.com/books">
-#   <bk:title>Programming for Beginners</bk:title>
-#   <bk:author>Jane Doe</bk:author>
-#   <bk:year>2021</bk:year>
-# </bk:book>
+title = root.select_one("book title")
+title.serialize()
+# <bk:title xmlns:bk="http://www.example.com/books">Programming for Beginners</bk:title>
 ```
 
 ## Using Parser

python/markupever/dom.py

@@ -41,7 +41,7 @@ def select_one(self, expr: str, offset: int = 0) -> typing.Optional["Element"]:
 
     def serialize_bytes(
         self,
-        indent: int = 4,
+        indent: int = 0,
         is_html: bool | None = None,
         include_self: bool = True,
     ) -> bytes:
@@ -52,7 +52,7 @@ def serialize_bytes(
 
     def serialize(
         self,
-        indent: int = 4,
+        indent: int = 0,
         is_html: bool | None = None,
         include_self: bool = True,
     ) -> str:
@@ -356,29 +356,35 @@ def text(self, separator: str = "", strip: bool = False) -> str:
 
     def serialize_bytes(
         self,
-        indent: int = 4,
+        indent: int = 0,
         is_html: bool | None = None,
         include_self: bool = True,
     ) -> bytes:
         """
         Serialize the current node and its subtree to bytes.
 
-        - indent (int, optional): Number of spaces for indentation. Defaults to 4.
+        - indent (int, optional): If nonzero, pretty-print the output, indenting each level of elements by this
+          many spaces. Pretty-printing adds and removes whitespace, which can change the meaning of the document:
+          for example, it changes the rendering of whitespace between inline elements, and the contents of
+          `<pre>`, `<textarea>`, and `<script>` elements. Defaults to 0, which serializes the document exactly.
         - is_html (bool, optional): Whether to serialize as HTML. Defaults to None.
         - include_self (bool, optional): Whether to include the current node in serialization. Defaults to True.
         """
         return _rustlib.serialize(self._raw, indent, include_self, is_html)
 
     def serialize(
         self,
-        indent: int = 4,
+        indent: int = 0,
         is_html: bool | None = None,
         include_self: bool = True,
     ) -> str:
         """
         Serialize the current node and its subtree to string.
 
-        - indent (int, optional): Number of spaces for indentation. Defaults to 4.
+        - indent (int, optional): If nonzero, pretty-print the output, indenting each level of elements by this
+          many spaces. Pretty-printing adds and removes whitespace, which can change the meaning of the document:
+          for example, it changes the rendering of whitespace between inline elements, and the contents of
+          `<pre>`, `<textarea>`, and `<script>` elements. Defaults to 0, which serializes the document exactly.
         - is_html (bool, optional): Whether to serialize as HTML. Defaults to None.
         - include_self (bool, optional): Whether to include the current node in serialization. Defaults to True.
         """

python/tests/test_dom.py

@@ -475,7 +475,7 @@ def test_serializer():
             '<!DOCTYPE html>\n<html lang="en">\n<head>\n  <meta charset="UTF-8">\n  <meta name="viewport" content="width=device-width, initial-scale=1.0">\n  <title>Complex Test</title>\n  <style>\n    body { font-family: Arial, sans-serif; }\n    .highlight { background-color: yellow; }\n  </style>\n</head>\n<body>\n  <header>\n    <h1>Welcome to the Test Page</h1>\n  </header>\n  <nav>\n    <ul>\n      <li><a href="#home">Home</a></li>\n      <li><a href="#about">About</a></li>\n      <li><a href="#contact">Contact</a></li>\n    </ul>\n  </nav>\n  <main>\n    <section id="home">\n      <p>Home content goes here.</p>\n    </section>\n    <section id="about">\n      <p>About content goes here.</p>\n    </section>\n    <section id="contact">\n      <p>Contact content goes here.</p>\n    </section>\n  </main>\n  <footer>\n    <p>&copy; 2025 Test Page. All rights reserved.</p>\n  </footer>\n</body>\n</html>\n',
             False,
             0,
-            '<!DOCTYPE html><html lang="en"><head><meta charset="UTF-8"><meta name="viewport" content="width=device-width, initial-scale=1.0"><title>Complex Test</title><style>\n    body { font-family: Arial, sans-serif; }\n    .highlight { background-color: yellow; }</style></head><body><header><h1>Welcome to the Test Page</h1></header><nav><ul><li><a href="#home">Home</a></li><li><a href="#about">About</a></li><li><a href="#contact">Contact</a></li></ul></nav><main><section id="home"><p>Home content goes here.</p></section><section id="about"><p>About content goes here.</p></section><section id="contact"><p>Contact content goes here.</p></section></main><footer><p>© 2025 Test Page. All rights reserved.</p></footer></body></html>',
+            '<!DOCTYPE html><html lang="en"><head>\n  <meta charset="UTF-8">\n  <meta name="viewport" content="width=device-width, initial-scale=1.0">\n  <title>Complex Test</title>\n  <style>\n    body { font-family: Arial, sans-serif; }\n    .highlight { background-color: yellow; }\n  </style>\n</head>\n<body>\n  <header>\n    <h1>Welcome to the Test Page</h1>\n  </header>\n  <nav>\n    <ul>\n      <li><a href="#home">Home</a></li>\n      <li><a href="#about">About</a></li>\n      <li><a href="#contact">Contact</a></li>\n    </ul>\n  </nav>\n  <main>\n    <section id="home">\n      <p>Home content goes here.</p>\n    </section>\n    <section id="about">\n      <p>About content goes here.</p>\n    </section>\n    <section id="contact">\n      <p>Contact content goes here.</p>\n    </section>\n  </main>\n  <footer>\n    <p>© 2025 Test Page. All rights reserved.</p>\n  </footer>\n\n\n</body></html>',
         ),
         _SerializerIndent(
             '<!DOCTYPE html>\n<html lang="en">\n<head>\n  <meta charset="UTF-8">\n  <meta name="viewport" content="width=device-width, initial-scale=1.0">\n  <title>Complex Test</title>\n  <style>\n    body { font-family: Arial, sans-serif; }\n    .highlight { background-color: yellow; }\n  </style>\n</head>\n<body>\n  <header>\n    <h1>Welcome to the Test Page</h1>\n  </header>\n  <nav>\n    <ul>\n      <li><a href="#home">Home</a></li>\n      <li><a href="#about">About</a></li>\n      <li><a href="#contact">Contact</a></li>\n    </ul>\n  </nav>\n  <main>\n    <section id="home">\n      <p>Home content goes here.</p>\n    </section>\n    <section id="about">\n      <p>About content goes here.</p>\n    </section>\n    <section id="contact">\n      <p>Contact content goes here.</p>\n    </section>\n  </main>\n  <footer>\n    <p>&copy; 2025 Test Page. All rights reserved.</p>\n  </footer>\n</body>\n</html>\n',
@@ -493,6 +493,29 @@ def test_serializer():
         assert dom.serialize(case.indent) == case.expected
 
 
+def test_serializer_preserves_whitespace():
+    for content, options in [
+        (
+            "<p><b>a</b> <i>b</i></p>\n<p>c </p>",
+            markupever.HtmlOptions(full_document=False),
+        ),
+        (
+            "<pre>  x\n\n</pre><textarea>  y\n</textarea>",
+            markupever.HtmlOptions(full_document=False),
+        ),
+        (
+            "<script> if (a < b) {}\n</script><style> p { }\n</style>",
+            markupever.HtmlOptions(full_document=False),
+        ),
+        ("<a> <b>x </b>\n  <c></c>\t</a>", markupever.XmlOptions()),
+    ]:
+        dom = markupever.parse(content, options)
+        assert dom.serialize(indent=0) == content
+        # indent=0 is the default.
+        assert dom.serialize() == content
+        assert dom.serialize_bytes() == content.encode()
+
+
 def test_add_itself():
     tag = markupever.parse("<p>").root().first_child
 

src/parser.rs

@@ -452,7 +452,7 @@ impl PyParser {
 }
 
 #[pyo3::pyfunction]
-#[pyo3(signature=(node, indent=4, include_self=true, is_html=None))]
+#[pyo3(signature=(node, indent=0, include_self=true, is_html=None))]
 pub fn serialize(
     node: crate::nodes::PyNodeRef,
     indent: usize,

treedom/src/dom.rs

@@ -230,7 +230,10 @@ impl<'a> Serializer<'a> {
                         serializer.write_processing_instruction(&pi.target, &pi.data)?
                     }
                     interface::Interface::Text(text) => {
-                        if !text.contents.trim_ascii().is_empty() {
+                        // Without indentation, whitespace is significant: write it unchanged.
+                        if self.indent == 0 {
+                            serializer.write_text(&text.contents)?;
+                        } else if !text.contents.trim_ascii().is_empty() {
                             serializer.write_text(text.contents.trim_ascii_end())?;
                         }
                     }

treedom/src/parser.rs

@@ -579,6 +579,27 @@ mod tests {
         assert_eq!(HTML, String::from_utf8_lossy(&buf));
     }
 
+    #[cfg(feature = "html5ever")]
+    #[test]
+    fn html_serializer_preserves_whitespace() {
+        use crate::Serializer;
+
+        const HTML: &str = "<!DOCTYPE html><html><head>\n</head>\n<body><p><b>a</b> <i>b</i></p>\n<pre>  x\n\n</pre>\n</body></html>";
+
+        let parser = ParserSink::parse_html(true, Default::default(), Default::default());
+        let dom = parser.one(HTML).into_dom();
+
+        let mut buf: Vec<u8> = Vec::new();
+        html5ever::serialize::serialize(
+            &mut buf,
+            &Serializer::new(&dom, dom.root().id(), 0),
+            Default::default(),
+        )
+        .unwrap();
+
+        assert_eq!(HTML, String::from_utf8_lossy(&buf));
+    }
+
     #[cfg(feature = "xml5ever")]
     #[test]
     fn xml_serializer() {