Normalize XML tag names

Fixes #1496

Author: jhy

CHANGES.md

@@ -5,7 +5,14 @@
 ### Changes
 
 ### Improvements
-* `Element.cssSelector()` will prefer to return shorter selectors by using ancestor IDs when available and unique. E.g. `#id > div > p` instead of  `html > body > div > div > p` [#2283](https://github.com/jhy/jsoup/pull/2283).
+
+* `Element.cssSelector()` will prefer to return shorter selectors by using ancestor IDs when available and unique. E.g.
+  `#id > div > p` instead of  `html > body > div > div > p` [#2283](https://github.com/jhy/jsoup/pull/2283).
+
+### Bug Fixes
+
+* When serializing a Document to XML, element names with characters that are invalid in XML are now
+  normalized. [#1496].(https://github.com/jhy/jsoup/issues/1496)
 
 ## 1.19.1 (2025-03-04)
 

src/main/java/org/jsoup/internal/Normalizer.java

@@ -1,5 +1,9 @@
 package org.jsoup.internal;
 
+import org.jsoup.nodes.Attribute;
+import org.jsoup.nodes.Document;
+import org.jspecify.annotations.Nullable;
+
 import java.util.Locale;
 
 /**
@@ -23,8 +27,7 @@ public static String normalize(final String input, boolean isStringLiteral) {
     }
 
     /** Minimal helper to get an otherwise OK HTML name like "foo<bar" to "foo_bar". */
-    public static String xmlSafeTagName(final String tagname) {
-        // todo - if required we could make a fuller version of this as in Attribute.getValidKey(syntax) in Element. for now, just minimal based on what HtmlTreeBuilder produces
-        return tagname.replace('<', '_');
+    @Nullable public static String xmlSafeTagName(final String tagname) {
+        return Attribute.getValidKey(tagname, Document.OutputSettings.Syntax.xml); // Reuses the Attribute key normal, which is same for xml tag names
     }
 }

src/main/java/org/jsoup/nodes/Element.java

@@ -1838,7 +1838,7 @@ void outerHtmlTail(Appendable accum, int depth, Document.OutputSettings out) thr
     }
 
     /* If XML syntax, normalizes < to _ in tag name. */
-    private String safeTagName(Document.OutputSettings.Syntax syntax) {
+    @Nullable private String safeTagName(Document.OutputSettings.Syntax syntax) {
         return syntax == xml ? Normalizer.xmlSafeTagName(tagName()) : tagName();
     }
 

src/test/java/org/jsoup/nodes/ElementTest.java

@@ -3035,4 +3035,16 @@ void prettySerializationRoundTrips(Document.OutputSettings settings) {
         Iterator<Element> iterator = div.iterator();
         assertIterates(iterator, expect);
     }
+
+    @Test void htmlToXmlNormalizes() {
+        // https://github.com/jhy/jsoup/issues/1496
+        String in = "<p\u226F\u0322>One</p\u226F\u0322>";
+        Document doc = Jsoup.parse(in);
+        doc.outputSettings().prettyPrint(false);
+        String html = doc.body().html();
+        doc.outputSettings().syntax(Document.OutputSettings.Syntax.xml);
+        String xml = doc.body().html();
+        assertEquals("<p≯̢>One</p≯̢>", html);
+        assertEquals("<p_>One</p_>", xml);
+    }
 }