From c40fca08a2432f9114f80b5c0935244543124eaf Mon Sep 17 00:00:00 2001
From: orbiter <orbiter@6c8d7289-2bf4-0310-a012-ef5d649a1542>
Date: Mon, 2 Oct 2006 10:21:14 +0000
Subject: [PATCH] fixed bad handling of string separation you can now use a new
 encoding attribute to create strings from byte arrays

git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@2693 6c8d7289-2bf4-0310-a012-ef5d649a1542
---
 source/de/anomic/tools/nxTools.java | 33 ++++++++++++++++++++++++++---
 1 file changed, 30 insertions(+), 3 deletions(-)

diff --git a/source/de/anomic/tools/nxTools.java b/source/de/anomic/tools/nxTools.java
index 2eaf2664b..086080927 100644
--- a/source/de/anomic/tools/nxTools.java
+++ b/source/de/anomic/tools/nxTools.java
@@ -43,6 +43,7 @@
 
 package de.anomic.tools;
 
+import java.io.UnsupportedEncodingException;
 import java.util.ArrayList;
 import java.util.Enumeration;
 import java.util.HashMap;
@@ -149,13 +150,39 @@ public class nxTools {
     }
 
     public static ArrayList strings(byte[] a) {
+        return strings(a, null);
+    }
+    
+    public static ArrayList strings(byte[] a, String encoding) {
         int s = 0;
         int e;
         ArrayList v = new ArrayList();
+        byte b;
         while (s < a.length) {
-            e = s; while (e < a.length) if (a[e++] < 32) {e--; break;}
-            v.add(new String(a, s, e - s));
-            s = e; while (s < a.length) if (a[s++] >= 32) {s--; break;}
+            // find eol
+            e = s;
+            while (e < a.length) {
+                b = a[e];
+                if ((b == 10) || (b == 13)) break;
+                e++;
+            }
+            
+            // read line
+            if (encoding == null) {
+                v.add(new String(a, s, e - s));
+            } else try {
+                v.add(new String(a, s, e - s, encoding));
+            } catch (UnsupportedEncodingException xcptn) {
+                return v;
+            }
+            
+            // eat up additional eol bytes
+            s = e + 1;
+            while (s < a.length) {
+                b = a[s];
+                if ((b != 10) && (b != 13)) break;
+                s++;
+            }
         }
         return v;
     }