Implement DOMNode.getTextContent() according to API

Fix #1695 Signed-off-by: Christoph Läubrich <[email protected]>
eclipse-lemminx · Nov 12, 2024 · 7ebf6a1 · 7ebf6a1
1 parent 0aadd16
commit 7ebf6a1
Show file tree

Hide file tree

Showing 2 changed files with 96 additions and 4 deletions.
diff --git a/org.eclipse.lemminx/src/main/java/org/eclipse/lemminx/dom/DOMNode.java b/org.eclipse.lemminx/src/main/java/org/eclipse/lemminx/dom/DOMNode.java
@@ -827,7 +827,39 @@ public DOMElement getOrphanEndElement(int offset, String tagName, boolean anyOrp
 	 */
 	@Override
 	public String getTextContent() throws DOMException {
-		return getNodeValue();
+
+		switch (getNodeType()) {
+		// Text like nodes simply return their node value
+		case Node.TEXT_NODE:
+		case Node.CDATA_SECTION_NODE:
+		case Node.COMMENT_NODE:
+		case Node.PROCESSING_INSTRUCTION_NODE:
+			return getNodeValue();
+		// These special types has to return null
+		case Node.DOCUMENT_NODE:
+		case Node.DOCUMENT_TYPE_NODE:
+		case Node.NOTATION_NODE:
+			return null;
+		// concatenation of the textContent attribute value of every child node
+		default:
+			if (this.children != null && children.size() > 0) {
+				final StringBuilder builder = new StringBuilder();
+				for (DOMNode child : children) {
+					short nodeType = child.getNodeType();
+					if (nodeType == Node.COMMENT_NODE || nodeType == Node.PROCESSING_INSTRUCTION_NODE) {
+						// excluding COMMENT_NODE and PROCESSING_INSTRUCTION_NODE nodes.
+						continue;
+					}
+					String text = child.getTextContent();
+					if (text != null && !text.isEmpty()) {
+						builder.append(text);
+					}
+				}
+				return builder.toString();
+			}
+			// empty string if the node has no children
+			return "";
+		}
 	}
 
 	@Override

diff --git a/org.eclipse.lemminx/src/test/java/org/eclipse/lemminx/dom/DOMParserTest.java b/org.eclipse.lemminx/src/test/java/org/eclipse/lemminx/dom/DOMParserTest.java
@@ -45,6 +45,65 @@ public void testNestedElement() {
 		assertDocument("<html><body></body></html>", html);
 	}
 
+	@Test
+	public void testGetTextContentWithSimpleContent() {
+		DOMDocument document = DOMParser.getInstance().parse("<a><b><c>Hello</c></b></a>", "uri", null);
+		String textContent = document.getDocumentElement().getTextContent();
+		assertNotNull(textContent);
+		assertEquals("Hello", textContent);
+	}
+
+	@Test
+	public void testGetTextContentWithMixedContent() {
+		DOMDocument document = DOMParser.getInstance().parse("<a>H<b>e<c>ll</c></b>o</a>", "uri", null);
+		String textContent = document.getDocumentElement().getTextContent();
+		assertNotNull(textContent);
+		assertEquals("Hello", textContent);
+	}
+
+	@Test
+	public void testGetTextContentWithComplexContent() {
+		DOMDocument document = DOMParser.getInstance().parse("<a><b>H</b><c>e</c><b>ll</b><x>o</x></b></a>", "uri",
+				null);
+		String textContent = document.getDocumentElement().getTextContent();
+		assertNotNull(textContent);
+		assertEquals("Hello", textContent);
+	}
+
+	@Test
+	public void testGetTextContentWithCharContent() {
+		DOMDocument document = DOMParser.getInstance().parse("<text>Hello</text>", "uri", null);
+		String textContent = document.getDocumentElement().getChild(0).getTextContent();
+		assertNotNull(textContent);
+		assertEquals("Hello", textContent);
+	}
+
+	@Test
+	public void testGetTextContentWithCDATAContent() {
+		DOMDocument document = DOMParser.getInstance().parse("<a><b><c><![CDATA[Hello]]></c></b></a>", "uri", null);
+		String textContent = document.getDocumentElement().getTextContent();
+		assertNotNull(textContent);
+		assertEquals("Hello", textContent);
+	}
+
+	@Test
+	public void testGetTextContentWithComment() {
+		DOMDocument document = DOMParser.getInstance()
+				.parse("<a><b><c>Hello</c><!-- comments must not be included --></b></a>", "uri", null);
+		String textContent = document.getDocumentElement().getTextContent();
+		assertNotNull(textContent);
+		assertEquals("Hello", textContent);
+	}
+
+	@Test
+	public void testGetTextContentWithPI() {
+		DOMDocument document = DOMParser.getInstance().parse("<a><b><c>Hello</c><?PI must not be included ?></b></a>",
+				"uri", null);
+		String textContent = document.getDocumentElement().getTextContent();
+		assertNotNull(textContent);
+		assertEquals("Hello", textContent);
+	}
+
 	@Test
 	public void testNestedElements() {
 		DOMNode head = createElement("head", 6, 12, 19, true);
@@ -56,6 +115,7 @@ public void testNestedElements() {
 		assertDocument("<html><head></head><body></body></html>", html);
 	}
 
+
 	@Test
 	public void testNestedNestedElements() {
 		DOMNode c = createElement("c", 6, 9, 13, true);
@@ -95,7 +155,7 @@ public void testEmptyTagT() {
 
 	@Test
 	public void singleEndTag() {
-		DOMElement meta = (DOMElement) createElement("meta", 0, 0, 7, false);
+		DOMElement meta = createElement("meta", 0, 0, 7, false);
 		assertDocument("</meta>", meta);
 		assertFalse(meta.hasStartTag());
 		assertTrue(meta.hasEndTag());
@@ -104,8 +164,8 @@ public void singleEndTag() {
 
 	@Test
 	public void insideEndTag() {
-		DOMElement meta = (DOMElement) createElement("meta", 6, 6, 13, false);
-		DOMElement html = (DOMElement) createElement("html", 0, 13, 20, true);
+		DOMElement meta = createElement("meta", 6, 6, 13, false);
+		DOMElement html = createElement("html", 0, 13, 20, true);
 		html.addChild(meta);
 
 		assertDocument("<html></meta></html>", html);