Add support for LexicalHandler.startEntity and endEntity (#210)

pom.xml

@@ -199,6 +199,12 @@ SAX2 and Stax2 APIs
             <version>${version.junit}</version>
             <scope>test</scope>
         </dependency>
+        <dependency>
+            <groupId>org.mockito</groupId>
+            <artifactId>mockito-core</artifactId>
+            <version>4.11.0</version>
+            <scope>test</scope>
+        </dependency>
     </dependencies>
 
     <build>

release-notes/CREDITS

@@ -109,3 +109,8 @@ Stanimir Stamenkov (@stanio)
 
 * Reported, provided fix for #204: Non-conformant `XMLEventFactory.setLocation(null)`
  (7.0.0)
+
+Philipp Nanz (@philippn)
+
+* Contributed #209: SAXParser: Add support for `LexicalHandler.startEntity()` and `.endEntity()`
+ (7.1.0)

release-notes/VERSION

@@ -4,6 +4,11 @@ Project: woodstox
 === Releases ===
 ------------------------------------------------------------------------
 
+Not yet released:
+
+#209: SAXParser: Add support for `LexicalHandler.startEntity()` and `.endEntity()`
+ (contributed by Philipp N)
+
 7.0.0 (21-Jun-2024)
 
 #134: Increase JDK baseline to Java 8

src/main/java/com/ctc/wstx/sax/WstxSAXParser.java

@@ -774,7 +774,12 @@ private final void fireAuxEvent(int type, boolean inTree)
             /* Only occurs in non-entity-expanding mode; so effectively
              * we are skipping the entity?
              */
-            if (mContentHandler != null) {
+            if (mLexicalHandler != null) {
+                String text = mScanner.getText();
+                mLexicalHandler.startEntity(mScanner.getLocalName());
+                mContentHandler.characters(text.toCharArray(), 0, text.length());
+                mLexicalHandler.endEntity(mScanner.getLocalName());
+            } else if (mContentHandler != null) {
                 mContentHandler.skippedEntity(mScanner.getLocalName());
             }
             break;

src/test/java/wstxtest/sax/TestLexicalHandler.java

@@ -0,0 +1,126 @@
+package wstxtest.sax;
+
+import com.ctc.wstx.sax.WstxSAXParserFactory;
+import com.ctc.wstx.stax.WstxInputFactory;
+import org.mockito.InOrder;
+import org.mockito.Mockito;
+import org.xml.sax.Attributes;
+import org.xml.sax.InputSource;
+import org.xml.sax.SAXException;
+import org.xml.sax.ext.DefaultHandler2;
+import wstxtest.BaseWstxTest;
+
+import javax.xml.parsers.SAXParser;
+import java.net.URL;
+
+/**
+ * Unit tests that verify handling of entity references during parsing.
+ */
+public class TestLexicalHandler extends BaseWstxTest {
+
+    public void testReplaceEntityRefs() throws Exception {
+        WstxSAXParserFactory spf = new WstxSAXParserFactory();
+        SAXParser sp = spf.newSAXParser();
+        EventListener listener = Mockito.mock(EventListener.class);
+        sp.parse(getInputSource("eyephone.xml"), new EventListenerHandler(listener));
+
+        InOrder orderVerifier = Mockito.inOrder(listener);
+        orderVerifier.verify(listener).startElement("prodname");
+        orderVerifier.verify(listener).characters("eyePhone© 2.0");
+        orderVerifier.verify(listener).endElement("prodname");
+    }
+
+    public void testWithoutReplaceEntityRefs() throws Exception {
+        WstxInputFactory staxFactory = new WstxInputFactory();
+        staxFactory.getConfig().doReplaceEntityRefs(false);
+        WstxSAXParserFactory spf = new WstxSAXParserFactory(staxFactory);
+        SAXParser sp = spf.newSAXParser();
+        EventListener listener = Mockito.mock(EventListener.class);
+        sp.parse(getInputSource("eyephone.xml"), new EventListenerHandler(listener));
+
+        InOrder orderVerifier = Mockito.inOrder(listener);
+        orderVerifier.verify(listener).startElement("prodname");
+        orderVerifier.verify(listener).characters("eyePhone");
+        orderVerifier.verify(listener).skippedEntity("copyright");
+        orderVerifier.verify(listener).characters(" 2.0");
+        orderVerifier.verify(listener).endElement("prodname");
+    }
+
+    public void testWithoutReplaceEntityRefsAndWithLexicalHandler() throws Exception {
+        WstxInputFactory staxFactory = new WstxInputFactory();
+        staxFactory.getConfig().doReplaceEntityRefs(false);
+        WstxSAXParserFactory spf = new WstxSAXParserFactory(staxFactory);
+        SAXParser sp = spf.newSAXParser();
+        EventListener listener = Mockito.mock(EventListener.class);
+        sp.setProperty("http://xml.org/sax/properties/lexical-handler", new EventListenerHandler(listener));
+        sp.parse(getInputSource("eyephone.xml"), new EventListenerHandler(listener));
+
+        InOrder orderVerifier = Mockito.inOrder(listener);
+        orderVerifier.verify(listener).startElement("prodname");
+        orderVerifier.verify(listener).characters("eyePhone");
+        orderVerifier.verify(listener).startEntity("copyright");
+        orderVerifier.verify(listener).characters("©");
+        orderVerifier.verify(listener).endEntity("copyright");
+        orderVerifier.verify(listener).characters(" 2.0");
+        orderVerifier.verify(listener).endElement("prodname");
+    }
+
+    private InputSource getInputSource(String resource) {
+        URL url = TestLexicalHandler.class.getResource(resource);
+        return new InputSource(url.toString());
+    }
+
+    private static class EventListenerHandler extends DefaultHandler2 {
+
+        private final EventListener eventListener;
+
+        private EventListenerHandler(EventListener eventListener) {
+            this.eventListener = eventListener;
+        }
+
+        @Override
+        public void startEntity(String name) throws SAXException {
+            eventListener.startEntity(name);
+        }
+
+        @Override
+        public void endEntity(String name) throws SAXException {
+            eventListener.endEntity(name);
+        }
+
+        @Override
+        public void skippedEntity(String name) throws SAXException {
+            eventListener.skippedEntity(name);
+        }
+
+        @Override
+        public void characters(char[] ch, int start, int length) throws SAXException {
+            eventListener.characters(String.copyValueOf(ch, start, length));
+        }
+
+        @Override
+        public void startElement(String uri, String localName, String qName, Attributes attributes) throws SAXException {
+            eventListener.startElement(qName);
+        }
+
+        @Override
+        public void endElement(String uri, String localName, String qName) throws SAXException {
+            eventListener.endElement(qName);
+        }
+    }
+
+    private interface EventListener {
+
+        void startElement(String name);
+
+        void endElement(String name);
+
+        void startEntity(String name);
+
+        void endEntity(String name);
+
+        void skippedEntity(String name);
+
+        void characters(String content);
+    }
+}

src/test/resources/wstxtest/sax/eyephone.xml

@@ -0,0 +1,18 @@
+<!DOCTYPE product-info PUBLIC "-//Woodstox//DTD Sample V1.0//EN" "productinfo_v1.dtd">
+<product-info>
+    <meta>
+        <prodname>eyePhone&copyright; 2.0</prodname>
+        <company>MomCorp</company>
+        <prodtype>mobile phone</prodtype>
+        <source>http://theinfosphere.org/EyePhone</source>
+    </meta>
+    <!-- This is a comment -->
+    <description>
+        <para>The <b>eyePhone</b> is a new type of mobile phones released by MomCorp and one of the
+            products of Mom Store in 3010. It is called an eyePhone because it is located in the
+            eyes of its users and displays a screen in front of them. The eyePhone has a lot of
+            applications, such as Twit and can even make phone calls. Soon, everyone became addicted
+            to their new eyePhones, and Mom activated a Twit-worm so that she could have between 1
+            and 2 million zombies buying the machine's "upgraded" version, the eyePhone 2.0.</para>
+    </description>
+</product-info>

src/test/resources/wstxtest/sax/productinfo_v1.dtd

@@ -0,0 +1,23 @@
+<?xml encoding="UTF-8"?>
+
+<!-- The DTD for the sample documents -->
+
+<!ENTITY copyright "&#169;">
+
+<!ELEMENT product-info (meta,description)>
+
+<!ELEMENT meta (prodname,company,prodtype,source)>
+
+<!ELEMENT description (para)>
+
+<!ELEMENT prodname (#PCDATA)>
+
+<!ELEMENT company (#PCDATA)>
+
+<!ELEMENT prodtype (#PCDATA)>
+
+<!ELEMENT source (#PCDATA)>
+
+<!ELEMENT para (#PCDATA|b)*>
+
+<!ELEMENT b (#PCDATA)*>