From 22649408ad9ead0ab020e01e7baee5d5ab3d02bc Mon Sep 17 00:00:00 2001
From: theli <theli@6c8d7289-2bf4-0310-a012-ef5d649a1542>
Date: Tue, 10 Oct 2006 10:14:03 +0000
Subject: [PATCH] *) Better errorhandling for charset encoding problem during
 content parsing    See: http://www.yacy-forum.de/viewtopic.php?t=2952

git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@2737 6c8d7289-2bf4-0310-a012-ef5d649a1542
---
 source/de/anomic/plasma/plasmaCrawlEURL.java | 1 +
 source/de/anomic/plasma/plasmaParser.java    | 5 +++++
 2 files changed, 6 insertions(+)

diff --git a/source/de/anomic/plasma/plasmaCrawlEURL.java b/source/de/anomic/plasma/plasmaCrawlEURL.java
index 8d9918354..f53dd0f55 100644
--- a/source/de/anomic/plasma/plasmaCrawlEURL.java
+++ b/source/de/anomic/plasma/plasmaCrawlEURL.java
@@ -86,6 +86,7 @@ public class plasmaCrawlEURL extends indexURL {
     
     // wrong content
     public static final String DENIED_WRONG_MIMETYPE_OR_EXT = "denied_(wrong_mimetype_or_extension)";
+    public static final String DENIED_UNSUPPORTED_CHARSET = "denied_(unsupported_charset)";
     public static final String DENIED_REDIRECTION_HEADER_EMPTY = "denied_(redirection_header_empty)";
     public static final String DENIED_REDIRECTION_COUNTER_EXCEEDED = "denied_(redirection_counter_exceeded)";
     public static final String DENIED_WRONG_HTTP_STATUSCODE = "denied_(wrong_http_status_code_";
diff --git a/source/de/anomic/plasma/plasmaParser.java b/source/de/anomic/plasma/plasmaParser.java
index 81404dc06..c11984b29 100644
--- a/source/de/anomic/plasma/plasmaParser.java
+++ b/source/de/anomic/plasma/plasmaParser.java
@@ -52,6 +52,7 @@ import java.io.FileInputStream;
 import java.io.FilenameFilter;
 import java.io.IOException;
 import java.io.InputStream;
+import java.io.UnsupportedEncodingException;
 import java.net.MalformedURLException;
 import java.net.URI;
 import java.util.Arrays;
@@ -600,6 +601,10 @@ public final class plasmaParser {
             }
             return doc;
             
+        } catch (UnsupportedEncodingException e) {
+            String errorMsg = "Unsupported charset encoding: " + e.getMessage();
+            this.theLogger.logSevere("Unable to parse '" + location + "'. " + errorMsg, e);
+            throw new ParserException(errorMsg,location,plasmaCrawlEURL.DENIED_UNSUPPORTED_CHARSET);                	
         } catch (Exception e) {
             // Interrupted- and Parser-Exceptions should pass through
             if (e instanceof InterruptedException) throw (InterruptedException) e;