fix more encooding problems in yacysearch.rss.

- URL encoding for search terms where required
- removed "ugly" CDATA escaping
- UTF-8 encoding for the XML
- no HTML style escaping for XML/RSS element values
Note: some unicode characters might still be encooded in a wrong way.

git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@4140 6c8d7289-2bf4-0310-a012-ef5d649a1542

htroot/yacysearch.java

@@ -51,6 +51,7 @@ import java.net.MalformedURLException;
 import java.util.HashMap;
 import java.util.TreeSet;
 
+import de.anomic.data.htmlTools;
 import de.anomic.http.httpHeader;
 import de.anomic.index.indexURLEntry;
 import de.anomic.kelondro.kelondroBitfield;
@@ -381,8 +382,8 @@ public class yacysearch {
             // adding some additional properties needed for the rss feed
             String hostName = (String) header.get("Host", "localhost");
             if (hostName.indexOf(":") == -1) hostName += ":" + serverCore.getPortNr(env.getConfig("port", "8080"));
-            prop.put("searchBaseURL", "http://" + hostName + "/yacysearch.html");
-            prop.put("rssYacyImageURL", "http://" + hostName + "/env/grafics/yacy.gif");
+            prop.putASIS("searchBaseURL", "http://" + hostName + "/yacysearch.html");
+            prop.putASIS("rssYacyImageURL", "http://" + hostName + "/env/grafics/yacy.gif");
         }
         
         prop.put("searchagain", (global) ? 1 : 0);
@@ -408,6 +409,11 @@ public class yacysearch {
         prop.put("input_contentdomCheckImage", (contentdomCode == plasmaSearchQuery.CONTENTDOM_IMAGE) ? 1 : 0);
         prop.put("input_contentdomCheckApp", (contentdomCode == plasmaSearchQuery.CONTENTDOM_APP) ? 1 : 0);
         
+        // for RSS: don't HTML encode some elements
+        String q = htmlTools.encodeUnicode2xml(post.get("search", ""));
+        prop.put("rss_query", q);
+        prop.put("rss_queryenc", yacyURL.escape(q));
+        
         // return rewrite properties
         return prop;
     }

htroot/yacysearch.rss

@@ -1,4 +1,4 @@
-<?xml version="1.0"?>
+<?xml version="1.0" encoding="UTF-8"?>
 <?xml-stylesheet type='text/xsl' href='/yacysearch.xsl' version='1.0'?>
 <rss version="2.0"
 	xmlns:yacyTopwords="http://www.yacy.net/yacy/topwords"
@@ -6,19 +6,19 @@
 	xmlns:atom="http://www.w3.org/2005/Atom">
 	<!-- YaCy P2P Web Search - Results; http://yacy.net -->
 	<channel>
-		<title>YaCy P2P-Search for #[former]#</title>
-		<description>Search for #[former]#</description>
-		<link>#[searchBaseURL]#?search=#[former]#&amp;resource=#[input_resource]#&amp;contentdom=#[input_contentdom]#</link>
+		<title>YaCy P2P-Search for #[rss_query]#</title>
+		<description>Search for #[rss_query]#</description>
+		<link>#[searchBaseURL]#?search=#[rss_queryenc]#&amp;resource=#[input_resource]#&amp;contentdom=#[input_contentdom]#</link>
 		<image>
 			<url>#[rssYacyImageURL]#</url>
-			<title>Search for #[former]#</title>
-			<link>#[searchBaseURL]#?search=#[former]#&amp;resource=#[input_resource]#&amp;contentdom=#[input_contentdom]#</link>
+			<title>Search for #[rss_query]#</title>
+			<link>#[searchBaseURL]#?search=#[rss_queryenc]#&amp;resource=#[input_resource]#&amp;contentdom=#[input_contentdom]#</link>
 		</image>
 		<opensearch:totalResults>#[num-results_totalcount]#</opensearch:totalResults>
 		<opensearch:startIndex>#[num-results_offset]#</opensearch:startIndex>
 		<opensearch:itemsPerPage>#[num-results_itemsPerPage]#</opensearch:itemsPerPage>
 		<atom:link rel="related" href="opensearchdescription.xml" type="application/opensearchdescription+xml"/>
-		<opensearch:Query role="request" searchTerms="#[former]#" />
+		<opensearch:Query role="request" searchTerms="#[rss_queryenc]#" />
 		
 		#{results}#
 <!--#include virtual="yacysearchitem.html?rss=true&item=#[item]#&eventID=#[eventID]#" -->

htroot/yacysearchitem.html

@@ -41,9 +41,9 @@
 #(/content)#
 #(rss)#::
 <item>
-<title><![CDATA[#[title]#]]></title>
+<title>#[title]#</title>
 <link>#[link]#</link>
-<description><![CDATA[#[description]#]]></description>
+<description>#[description]#></description>
 <pubDate>#[date]#</pubDate>
 <guid isPermaLink="false">#[urlhash]#</guid>
 </item>

htroot/yacysearchitem.java

@@ -32,6 +32,7 @@ import java.util.Iterator;
 import java.util.Set;
 import java.util.TreeSet;
 
+import de.anomic.data.htmlTools;
 import de.anomic.http.httpHeader;
 import de.anomic.kelondro.kelondroMSetTools;
 import de.anomic.kelondro.kelondroNaturalOrder;
@@ -174,8 +175,8 @@ public class yacysearchitem {
         if (rss) {
             // text search for rss output
             prop.put("rss", 1); // switch on specific content
-            prop.put("rss_title", result.title());
-            prop.put("rss_description", result.textSnippet().getLineRaw());
+            prop.putASIS("rss_title", htmlTools.encodeUnicode2xml(result.title()));
+            prop.putASIS("rss_description", htmlTools.encodeUnicode2xml(result.textSnippet().getLineRaw()));
             prop.put("rss_link", result.urlstring());
             prop.put("rss_urlhash", result.hash());
             prop.put("rss_date", plasmaSwitchboard.dateString822(result.modified()));

source/de/anomic/data/htmlTools.java

@@ -2,30 +2,62 @@ package de.anomic.data;
 
 public class htmlTools {
 
-    /** Replaces characters in a string with other characters defined in an array.
+    /** Replaces characters in a string with other entities according to HTML standards.
       * @param text a string that possibly contains special characters
-      * @param entities array that contains characters to be replaced and characters it will be replaced by
+      * @param includingAmpersand if <code>false</code> ampersands are not encoded
       * @return the string with all characters replaced by the corresponding character from array
       */
       //[FB], changes by [MN], re-implemented by [MC]
     public static String encodeUnicode2html(String text, boolean includingAmpersand) {
         if (text == null) return null;
-        int pos = 0;
+        int spos = (includingAmpersand ? 0 : 2);
+        int epos = mapping.length;
+
+        return encode(text, mapping, spos, epos);
+    }
+
+    /**
+     * Replaces special entities ampersand, quotation marks, and less than/graiter than
+     * by the escaping entities allowed in XML documents.
+     * 
+     * @param text the original String
+     * @return the encoded String
+     */
+    public static String encodeUnicode2xml(String text) {
+        if (text == null) return null;
+        int spos = 0;
+        int epos = 8;
+
+        return encode(text, mapping, spos, epos);
+    }
+
+    /**
+     * Generic method that replaces occurences of special character entities defined in map 
+     * array with their corresponding mapping.
+     * @param text The String too process.
+     * @param map  An array defining the entity mapping.
+     * @param spos It is possible to use a subset of the map only. This parameter defines the
+     *             starting point in the map array. 
+     * @param epos The ending point, see above.
+     * @return A copy of the original String with all entities defined in map replaced.
+     */
+    public static String encode(String text, final String[] map, int spos, int epos) {
         StringBuffer sb = new StringBuffer(text.length());
-        search: while (pos < text.length()) {
+        search: while (spos < text.length()) {
             // find a (forward) mapping
-            loop: for (int i = (includingAmpersand) ? 0 : 2; i < mapping.length; i += 2) {
-                if (text.charAt(pos) != mapping[i].charAt(0)) continue loop;
+            loop: for (int i = spos; i < epos; i += 2) {
+                if (text.charAt(spos) != map[i].charAt(0)) continue loop;
                 // found match
-                sb.append(mapping[i + 1]);
-                pos++;
+                sb.append(map[i + 1]);
+                spos++;
                 continue search;
             }
             // not found match
-            sb.append(text.charAt(pos));
-            pos++;
+            sb.append(text.charAt(spos));
+            spos++;
         }
-        return new String(sb);
+
+        return sb.toString();
     }
     
     public static String decodeHtml2Unicode(String text) {