From 8eb4181737dbb0fc8fe93debfc2ab96f4a93469e Mon Sep 17 00:00:00 2001
From: borg-0300 <borg-0300@6c8d7289-2bf4-0310-a012-ef5d649a1542>
Date: Wed, 28 Dec 2005 10:32:21 +0000
Subject: [PATCH] BUGFIX for regular expression

git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@1261 6c8d7289-2bf4-0310-a012-ef5d649a1542
---
 source/de/anomic/htmlFilter/htmlFilterContentScraper.java | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/source/de/anomic/htmlFilter/htmlFilterContentScraper.java b/source/de/anomic/htmlFilter/htmlFilterContentScraper.java
index ffd66871e..0e9e4953e 100644
--- a/source/de/anomic/htmlFilter/htmlFilterContentScraper.java
+++ b/source/de/anomic/htmlFilter/htmlFilterContentScraper.java
@@ -164,7 +164,7 @@ public class htmlFilterContentScraper extends htmlFilterAbstractScraper implemen
 
         // (this is different from previous normal forms where a '/' must not appear in root paths; here it must appear. Makes everything easier.)
         if (path.length() == 0 || path.charAt(0) != '/') path = "/" + path;
-
+        
         Pattern pathPattern = Pattern.compile("(/[^/\\.]+/)[.]{2}(?=/)|/\\.(?=/)|/(?=/)");
         Matcher matcher = pathPattern.matcher(path);
         while (matcher.find()) {
@@ -172,6 +172,10 @@ public class htmlFilterContentScraper extends htmlFilterAbstractScraper implemen
             matcher.reset(path);
         }
 
+        while (path.startsWith("/../")) {
+            path = path.substring(3);
+        }
+
         if (defaultPort) return url.getProtocol() + "://" + url.getHost() + path;
         return url.getProtocol() + "://" + url.getHost() + ":" + url.getPort() + path;
     }