For #170 and #161 - Support for character entities.

Created two DTDs, one with just character entities for XHTML and one combined character entities for XHTML and MathML. Also: + Minor refactoring of entity resolver and catalog. + Documented the two new doctypes in author’s guide. + Made sure that other doctypes resolve to the empty string.
danfickle · Mar 12, 2018 · 19828ff · 19828ff
1 parent c584289
commit 19828ff
Show file tree

Hide file tree

Showing 7 changed files with 3,052 additions and 89 deletions.
diff --git a/openhtmltopdf-core/src/main/java/com/openhtmltopdf/resource/FSCatalog.java b/openhtmltopdf-core/src/main/java/com/openhtmltopdf/resource/FSCatalog.java
@@ -30,6 +30,7 @@
 import java.io.InputStream;
 import java.io.IOException;
 import java.net.URL;
+import java.util.Collections;
 import java.util.HashMap;
 import java.util.Map;
 import java.util.logging.Level;
@@ -56,9 +57,6 @@
  * @author Patrick Wright
  */
 public class FSCatalog {
-    /**
-     * Default constructor
-     */
     public FSCatalog() {
     }
 
@@ -68,17 +66,17 @@ public FSCatalog() {
      *
      * @param catalogURI A String URI to a catalog XML file on the classpath.
      */
-    public Map parseCatalog(String catalogURI) {
+    public Map<String, String> parseCatalog(String catalogURI) {
         URL url;
-        Map map = null;
+        Map<String, String> map = null;
         InputStream s = null;
         try {
             url = FSCatalog.class.getClassLoader().getResource(catalogURI);
             s = new BufferedInputStream(url.openStream());
             map = parseCatalog(new InputSource(s));
         } catch (Exception ex) {
             XRLog.xmlEntities(Level.WARNING, "Could not open XML catalog from URI '" + catalogURI + "'", ex);
-            map = new HashMap();
+            map = Collections.emptyMap();
         } finally {
             try {
                 if (s != null) {
@@ -97,7 +95,7 @@ public Map parseCatalog(String catalogURI) {
      *
      * @param inputSource A SAX InputSource to a catalog XML file on the classpath.
      */
-    public Map parseCatalog(InputSource inputSource) {
+    public Map<String, String> parseCatalog(InputSource inputSource) {
         XMLReader xmlReader = XMLResource.newXMLReader();
 
         CatalogContentHandler ch = new CatalogContentHandler();
@@ -151,16 +149,16 @@ public void warning(SAXParseException ex) {
      * parse, then call getEntityMap().
      */
     private static class CatalogContentHandler extends DefaultHandler {
-        private Map entityMap;
+        private final Map<String, String> entityMap;
 
         public CatalogContentHandler() {
-            this.entityMap = new HashMap();
+            this.entityMap = new HashMap<String, String>();
         }
 
         /**
          * Returns a Map of public Ids to local URIs
          */
-        public Map getEntityMap() {
+        public Map<String, String> getEntityMap() {
             return entityMap;
         }
 
@@ -197,5 +195,4 @@ private void setFeature(XMLReader xmlReader, String featureUri, boolean value) {
                     featureUri + ". Feature may be properly named, but not recognized by this parser.");
         }
     }
-}// end class
-
+}
diff --git a/openhtmltopdf-core/src/main/java/com/openhtmltopdf/resource/FSEntityResolver.java b/openhtmltopdf-core/src/main/java/com/openhtmltopdf/resource/FSEntityResolver.java
@@ -29,7 +29,9 @@
 
 import java.io.IOException;
 import java.io.InputStream;
+import java.io.StringReader;
 import java.net.URL;
+import java.util.Collections;
 import java.util.HashMap;
 import java.util.Map;
 import java.util.logging.Level;
@@ -57,49 +59,26 @@ public class FSEntityResolver implements EntityResolver {
     /**
      * Singleton instance, use {@link #instance()} to retrieve.
      */
-    private static FSEntityResolver instance;
+    private static final FSEntityResolver instance = new FSEntityResolver();
 
-    private final Map entities = new HashMap();
-
-    // fill the list of URLs
+    private final Map<String, String> entities = new HashMap<String, String>();
+
     /**
-     * Constructor for the FSEntityResolver object
+     * Constructor for the FSEntityResolver object, fill the map of public ids to local urls.
      */
     private FSEntityResolver() {
         FSCatalog catalog = new FSCatalog();
 
-        // The HTML 4.01 DTDs; includes entities. Load from catalog file.
-        entities.putAll(catalog.parseCatalog("resources/schema/html-4.01/catalog-html-4.01.xml"));
-
-        // XHTML common (shared declarations)
-        entities.putAll(catalog.parseCatalog("resources/schema/xhtml/catalog-xhtml-common.xml"));
-
-        // The XHTML 1.0 DTDs
-        entities.putAll(catalog.parseCatalog("resources/schema/xhtml/catalog-xhtml-1.0.xml"));
-
-        // The XHMTL 1.1 DTD
-        entities.putAll(catalog.parseCatalog("resources/schema/xhtml/catalog-xhtml-1.1.xml"));
-
-        // DocBook DTDs
-        entities.putAll(catalog.parseCatalog("resources/schema/docbook/catalog-docbook.xml"));
-
-        // The XHTML 1.1 element sets
+        entities.putAll(catalog.parseCatalog("resources/schema/openhtmltopdf/catalog-special.xml"));
     }
 
-    /**
-     * Description of the Method
-     *
-     * @param publicID PARAM
-     * @param systemID PARAM
-     * @return Returns
-     * @throws SAXException Throws
-     */
+    @Override
     public InputSource resolveEntity(String publicID,
                                      String systemID)
             throws SAXException {
-
         InputSource local = null;
-        String url = (String) getEntities().get(publicID);
+        String url = getEntities().get(publicID);
+
         if (url != null) {
             URL realUrl = GeneralUtil.getURLFromClasspath(this, url);
             InputStream is = null;
@@ -123,63 +102,26 @@ public InputSource resolveEntity(String publicID,
             XRLog.xmlEntities(Level.FINE, "Entity public: " + publicID + " -> " + url +
                     (local == null ? ", NOT FOUND" : " (local)"));
         } else {
-            XRLog.xmlEntities("Entity public: " + publicID + ", no local mapping. Parser will probably pull from network.");
+            XRLog.xmlEntities("Entity public: " + publicID + ", no local mapping. Returning empty entity to avoid pulling from network.");
+            local = new InputSource(new StringReader(""));
         }
         return local;
     }
 
     /**
      * Gets an instance of this class.
      *
-     * @return An instance of .
+     * @return An instance of FSEntityResolver.
      */
-    public static synchronized FSEntityResolver instance() {
-        if (instance == null) {
-            instance = new FSEntityResolver();
-        }
+    public static FSEntityResolver instance() {
         return instance;
     }
 
     /**
-     * Returns a map of entities parsed by this resolver.
-     * @return a map of entities parsed by this resolver. 
+     * Returns an unmodifiable map of entities parsed by this resolver.
+     * @return an unmodifiable map of entities parsed by this resolver. 
      */
-    public Map getEntities() {
-        return new HashMap(entities);
+    public Map<String, String> getEntities() {
+        return Collections.unmodifiableMap(entities);
     }
 }
-
-/*
- * $Id$
- *
- * $Log$
- * Revision 1.8  2008/12/01 20:37:24  pdoubleya
- * Expose copy of parsed entities from catalog.
- *
- * Revision 1.7  2007/05/21 22:13:02  peterbrant
- * Code cleanup (patch from Sean Bright)
- *
- * Revision 1.6  2007/05/20 23:25:34  peterbrant
- * Various code cleanups (e.g. remove unused imports)
- *
- * Patch from Sean Bright
- *
- * Revision 1.5  2005/06/13 06:50:15  tobega
- * Fixed a bug in table content resolution.
- * Various "tweaks" in other stuff.
- *
- * Revision 1.4  2005/03/28 14:24:48  pdoubleya
- * Changed to resolve all entities using simple catalog files.
- *
- * Revision 1.3  2005/03/27 18:36:26  pdoubleya
- * Added separate logging for entity resolution.
- *
- * Revision 1.2  2005/03/21 09:13:50  pdoubleya
- * Added XHTML 1.1 references (Kevin).
- *
- * Revision 1.1  2005/02/03 20:39:34  pdoubleya
- * Added to CVS.
- *
- *
- */
-
diff --git a/openhtmltopdf-core/src/main/java/com/openhtmltopdf/resource/XMLResource.java b/openhtmltopdf-core/src/main/java/com/openhtmltopdf/resource/XMLResource.java
@@ -173,7 +173,7 @@ private void setXmlReaderSecurityFeatures(XMLReader xmlReader) {
           	 xmlReader.setFeature("http://apache.org/xml/features/disallow-doctype-decl", false);
            	 xmlReader.setFeature("http://xml.org/sax/features/external-general-entities", false);
            	 xmlReader.setFeature("http://xml.org/sax/features/external-parameter-entities", false);
-           	 xmlReader.setFeature("http://apache.org/xml/features/nonvalidating/load-external-dtd", false);
+           	 xmlReader.setFeature("http://apache.org/xml/features/nonvalidating/load-external-dtd", true);
            	 xmlReader.setFeature(XMLConstants.FEATURE_SECURE_PROCESSING, true);
            } catch (SAXNotSupportedException e) {
            	 XRLog.load(Level.SEVERE, "Unable to disable XML External Entities, which might put you at risk to XXE attacks", e);

diff --git a/openhtmltopdf-core/src/main/resources/resources/schema/openhtmltopdf/catalog-special.xml b/openhtmltopdf-core/src/main/resources/resources/schema/openhtmltopdf/catalog-special.xml
@@ -0,0 +1,13 @@
+<?xml version='1.0'?>
+<catalog xmlns="urn:oasis:names:tc:entity:xmlns:xml:catalog" prefer="public">
+
+<!-- ...................................................................... -->
+<!-- XML Catalog data for special DTD created for OpenHTMLToPDF ................................ -->
+<!-- ...................................................................... -->
+
+  <public publicId="-//OPENHTMLTOPDF//DOC XHTML Character Entities Only 1.0//EN" uri="resources/schema/openhtmltopdf/char-entities-xhtml-only.ent" />
+  <public publicId="-//OPENHTMLTOPDF//MATH XHTML Character Entities With MathML 1.0//EN" uri="resources/schema/openhtmltopdf/char-entities-xhtml-mathml.ent" />
+
+<!-- ...................................................................... -->
+
+</catalog>