/** Test {@link UniqueIdFilter}. */
@Test
public void duplicateIds() throws Exception {
String actual = "<p id=\"x\">1</p><p id=\"xy\">2</p><p id=\"x\">3</p>";
String expected = "<p id=\"x\">1</p><p id=\"xy\">2</p><p id=\"x0\">3</p>";
HTMLCleanerConfiguration config = this.mocker.getComponentUnderTest().getDefaultConfiguration();
List<HTMLFilter> filters = new ArrayList<HTMLFilter>(config.getFilters());
filters.add(this.mocker.<HTMLFilter>getInstance(HTMLFilter.class, "uniqueId"));
config.setFilters(filters);
Assert.assertEquals(
HEADER_FULL + expected + FOOTER,
HTMLUtils.toString(
this.mocker.getComponentUnderTest().clean(new StringReader(actual), config)));
}
@Override
public Document clean(Reader originalHtmlContent, HTMLCleanerConfiguration configuration) {
Document result;
// Note: Instantiation of an HtmlCleaner object is cheap so there's no need to cache an instance
// of it,
// especially since this makes it extra safe with regards to multithreading (even though HTML
// Cleaner is
// already supposed to be thread safe).
CleanerProperties cleanerProperties = getDefaultCleanerProperties(configuration);
HtmlCleaner cleaner = new HtmlCleaner(cleanerProperties);
cleaner.setTransformations(getDefaultCleanerTransformations(configuration));
TagNode cleanedNode;
try {
cleanedNode = cleaner.clean(originalHtmlContent);
} catch (Exception e) {
// This shouldn't happen since we're not doing any IO... I consider this a flaw in the design
// of HTML
// Cleaner.
throw new RuntimeException("Unhandled error when cleaning HTML", e);
}
// Serialize the cleanedNode TagNode into a w3c dom. Ideally following code should be enough.
// But SF's HTML Cleaner seems to omit the DocType declaration while serializing.
// See
// https://sourceforge.net/tracker/index.php?func=detail&aid=2062318&group_id=183053&atid=903696
// cleanedNode.setDocType(new DoctypeToken("html", "PUBLIC", "-//W3C//DTD XHTML 1.0
// Strict//EN",
// "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"));
// try {
// result = new DomSerializer(cleanerProperties, false).createDOM(cleanedNode);
// } catch(ParserConfigurationException ex) { }
// As a workaround, we must serialize the cleanedNode into a temporary w3c document, create a
// new w3c document
// with proper DocType declaration and move the root node from the temporary document to the new
// one.
try {
// Since there's a bug in SF's HTML Cleaner in that it doesn't recognize CDATA blocks we need
// to turn off
// character escaping (hence the false value passed) and do the escaping in
// XMLUtils.toString(). Note that
// this can cause problem for code not serializing the W3C DOM to a String since it won't have
// the
// characters escaped.
// See
// https://sourceforge.net/tracker/index.php?func=detail&aid=2691888&group_id=183053&atid=903696
Document tempDoc = new XWikiDOMSerializer(cleanerProperties, false).createDOM(cleanedNode);
DOMImplementation domImpl =
DocumentBuilderFactory.newInstance().newDocumentBuilder().getDOMImplementation();
DocumentType docType =
domImpl.createDocumentType(
QUALIFIED_NAME_HTML,
"-//W3C//DTD XHTML 1.0 Strict//EN",
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd");
result = domImpl.createDocument(null, QUALIFIED_NAME_HTML, docType);
result.replaceChild(
result.adoptNode(tempDoc.getDocumentElement()), result.getDocumentElement());
} catch (ParserConfigurationException ex) {
throw new RuntimeException("Error while serializing TagNode into w3c dom.", ex);
}
// Finally apply filters.
for (HTMLFilter filter : configuration.getFilters()) {
filter.filter(result, configuration.getParameters());
}
return result;
}