fix issue #115

taoqf · May 22, 2021 · d4d06ed · d4d06ed
1 parent 416fc9f
commit d4d06ed
Show file tree

Hide file tree

Showing 2 changed files with 44 additions and 2 deletions.
diff --git a/src/nodes/html.ts b/src/nodes/html.ts
@@ -133,6 +133,42 @@ export default class HTMLElement extends Node {
 
 		return JSON.stringify(attr.replace(/"/g, '&quot;'));
 	}
+
+  /**
+   * Trim all whitespace except single leading/trailing non-breaking space
+   * @param text string to trim
+   * @returns {string} trimmed value
+   * @private
+   */
+	private trimTextNodeWhitespace(text: string): string {
+    let i = 0;
+    let startPos;
+    let endPos;
+
+    while (i >= 0 && i < text.length) {
+      if (/\S/.test(text[i])) {
+        if (startPos === undefined) {
+          startPos = i;
+          i = text.length;
+        } else {
+          endPos = i;
+          i = void 0;
+        }
+      }
+
+      if (startPos === undefined) i++;
+      else i--;
+    }
+
+    if (startPos === undefined) startPos = 0;
+    if (endPos === undefined) endPos = text.length - 1;
+
+    const hasLeadingSpace = startPos > 0 && /[^\S\r\n]/.test(text[startPos-1]);
+    const hasTrailingSpace = endPos < (text.length - 1) && /[^\S\r\n]/.test(text[endPos+1]);
+
+    return (hasLeadingSpace ? ' ' : '') + text.slice(startPos, endPos + 1) + (hasTrailingSpace ? ' ' : '');
+  }
+
 	/**
 	 * Creates an instance of HTMLElement.
 	 * @param keyAttrs	id and class attribute
@@ -401,7 +437,7 @@ export default class HTMLElement extends Node {
 				if ((node as TextNode).isWhitespace) {
 					return;
 				}
-				node.rawText = node.rawText.trim();
+				node.rawText = this.trimTextNodeWhitespace(node.rawText);
 			} else if (node.nodeType === NodeType.ELEMENT_NODE) {
 				(node as HTMLElement).removeWhitespace();
 			}

diff --git a/test/html.js b/test/html.js
@@ -198,14 +198,20 @@ describe('HTML Parser', function () {
 
 		describe('#removeWhitespace()', function () {
 			it('should remove whitespaces while preserving nodes with content', function () {
-				const root = parseHTML('<p> \r \n  \t <h5> 123 </h5></p>');
+				const root = parseHTML('<p> \r \n  \t <h5>123</h5></p>');
 
 				const p = new HTMLElement('p', {}, '', root);
 				p.appendChild(new HTMLElement('h5', {}, ''))
 					.appendChild(new TextNode('123'));
 
 				root.firstChild.removeWhitespace().should.eql(p);
 			});
+
+			it('should preserve legitimate leading/trailing whitespace in TextNode', function () {
+			  parseHTML('<p>Hello  <em>World</em>!</p>').removeWhitespace().firstChild.text.should.eql('Hello World!');
+			  parseHTML('<p>\t\nHello\n\t<em>World</em>!</p>').removeWhitespace().firstChild.text.should.eql('HelloWorld!');
+			  parseHTML('<p>\t\n  Hello \n\t<em>World</em>!</p>').removeWhitespace().firstChild.text.should.eql(' Hello World!');
+      });
 		});
 
 		describe('#rawAttributes', function () {