Fixed stored URL in web cache when redirection(s) occurs.

Associate cached content to the last redirection location, instead of
the first URL of a redirection(s) chain :
 - for proper base URL processing in parsers (fixes mantis 636 -
http://mantis.tokeek.de/view.php?id=636)
 - to prevent duplicated content in Solr index when recrawling a
redirected URL

source/net/yacy/crawler/retrieval/Response.java

@@ -257,6 +257,11 @@ public class Response {
         return this.request.name();
     }
 
+	/**
+	 * @return the requested URL that produced this response. When redirection(s)
+	 *         occurred, this is not the initial URL, but the last redirection
+	 *         target.
+	 */
     public DigestURL url() {
         return this.request.url();
     }

source/net/yacy/repository/LoaderDispatcher.java

@@ -267,7 +267,11 @@ public final class LoaderDispatcher {
         final String storeError = response.shallStoreCacheForCrawler();
         if (storeError == null) {
             try {
-                Cache.store(url, response.getResponseHeader(), response.getContent());
+            	/* Important : we associate here the loaded content with the URL response.url(). 
+            	 * On eventual redirection(s), response.url() provides the last redirection location. 
+            	 * If instead we associated content with the initial url (beginning of the redirection(s) chain),
+            	 * the parsers would then have a wrong base URL when following links with relative URLs. */
+                Cache.store(response.url(), response.getResponseHeader(), response.getContent());
             } catch (final IOException e) {
                 LoaderDispatcher.log.warn("cannot write " + response.url() + " to Cache (3): " + e.getMessage(), e);
             }