Incorrect handling of & in CDATA sections in XmlFixerReader fixed (fixes

#134)

src/main/java/com/sun/syndication/io/impl/XmlFixerReader.java

@@ -38,6 +38,7 @@ public class XmlFixerReader extends Reader {
     }
 
     private boolean trimmed;
+    private boolean cdata = false;
     private final StringBuffer buffer;
     private int bufferPos;
     private int state = 0;
@@ -197,7 +198,18 @@ public class XmlFixerReader extends Reader {
                             buffer.setLength(0);
                             bufferPos = 0;
                             buffer.append((char) c);
-                            state = 1;
+                            loop = true;
+                        } else if (c == '<') {
+                            state = 4;
+                            buffer.setLength(0);
+                            bufferPos = 0;
+                            buffer.append((char) c);
+                            loop = true;
+                        } else if (c == ']' && cdata) {
+                            state = 5;
+                            buffer.setLength(0);
+                            bufferPos = 0;
+                            buffer.append((char) c);
                             loop = true;
                         } else {
                             loop = false;
@@ -219,7 +231,9 @@ public class XmlFixerReader extends Reader {
                         } else {
                             // no ';' to match the '&' lets just make the '&'
                             // a legal xml character entity '&amp;'
-                            buffer.insert(1, "amp;");
+                            if (!cdata) {
+                                buffer.insert(1, "amp;");
+                            }
                             buffer.append((char) c);
                             state = 3;
                             loop = true;
@@ -227,7 +241,9 @@ public class XmlFixerReader extends Reader {
                     } else {
                         // no ';' to match the '&' lets just make the '&'
                         // a legal xml character entity '&amp;'
-                        buffer.insert(1, "amp;");
+                        if (!cdata) {
+                            buffer.insert(1, "amp;");
+                        }
                         state = 3;
                         loop = true;
                     }
@@ -253,6 +269,58 @@ public class XmlFixerReader extends Reader {
                         loop = true;
                     }
                     break;
+                case 4: // checking for CDATA
+                    c = in.read();
+                    loop = true;
+                    state = 3;
+                    switch (c) {
+                        case -1:
+                            // end of stream
+                            break;
+                        case ' ':
+                        case '>':
+                        case '/':
+                            // tag end or something like this
+                            buffer.append((char) c);
+                            break;
+                        case '[':
+                            buffer.append((char) c);
+                            final String actBufferContent = buffer.toString();
+                            if ("<![CDATA[".equals(actBufferContent)) {
+                                cdata = true;
+                            } else {
+                                state = 4;
+                            }
+                            break;
+                        default:
+                            state = 4;
+                            buffer.append((char) c);
+                    }
+                    break;
+                case 5: // checking end of CDATA
+                    c = in.read();
+                    loop = true;
+                    state = 3;
+                    switch (c) {
+                        case -1:
+                            // end of stream
+                            break;
+                        case ']':
+                            buffer.append((char) c);
+                            state = 5;
+                            break;
+                        case '>':
+                            buffer.append((char) c);
+                            final String actBufferContent = buffer.toString();
+                            if ("]]>".equals(actBufferContent)) {
+                                cdata = false;
+                            }
+                            break;
+                        default:
+                            buffer.append((char) c);
+                            break;
+                    }
+                    break;
                 default:
                     throw new IOException("It shouldn't happen");
             }

src/test/java/com/sun/syndication/unittest/issues/Issue134Test.java

@@ -0,0 +1,22 @@
+package com.sun.syndication.unittest.issues;
+
+import com.sun.syndication.feed.synd.SyndFeed;
+import com.sun.syndication.unittest.FeedTest;
+
+/**
+ * Test for #134: Incorrect handling of CDATA sections.
+ * @author Martin Kurz
+ *
+ */
+public class Issue134Test extends FeedTest {
+
+    public Issue134Test() {
+        super("CDATATestFeed.xml");
+    }
+
+    public void testCDataLinks() throws Exception {
+        final SyndFeed feed = this.getCachedSyndFeed();
+        assertEquals("links differ", feed.getEntries().get(0).getLink(), feed.getEntries().get(1).getLink());
+    }
+
+}

src/test/resources/CDATATestFeed.xml

@@ -0,0 +1,32 @@
+<?xml version="1.0" encoding="utf-8"?>
+<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" version="2.0">
+	<channel>
+		<title>test</title>
+		<link>test</link>
+		<description>test</description>
+		<language>de</language>
+		<copyright>no</copyright>
+		<pubDate>Tue, 04 Aug 2009 11:25:23 +0200</pubDate>
+		<ttl>15</ttl>	
+		<image>
+			<url>https://rome.dev.java.net/branding/images/header_jnet_new.jpg</url>
+			<title>java.net</title>
+			<link>https://rome.dev.java.net/</link>
+			<description>no thanks</description>
+		</image>
+		<item>
+			<title>item1</title>
+			<description>Some Description</description>
+			<link>http://rtl-now.rtl.de/awz.php?container_id=26908&amp;paytype=ppv&amp;productdetail=1&amp;na=1</link>
+			<guid><![CDATA[http://rtl-now.rtl.de/awz.php?container_id=26908&paytype=ppv&productdetail=1&na=1]]></guid>
+			<pubDate>Wed, 05 Aug 2009 07:30:00 +0200</pubDate>
+		</item>
+		<item>
+			<title>item2</title>
+			<description>Some Description</description>
+			<link><![CDATA[http://rtl-now.rtl.de/awz.php?container_id=26908&paytype=ppv&productdetail=1&na=1]]></link>
+			<guid><![CDATA[http://rtl-now.rtl.de/awz.php?container_id=26908&paytype=ppv&productdetail=1&na=1]]></guid>
+			<pubDate>Wed, 05 Aug 2009 07:30:00 +0200</pubDate>
+		</item>
+	</channel>
+</rss>