From 1bb0b135ac5dab0adab423d89612f7b1e13f2e61 Mon Sep 17 00:00:00 2001
From: luccioman <luccioman@users.noreply.github.com>
Date: Tue, 27 Sep 2016 07:53:08 +0200
Subject: [PATCH] Avoid duplication of various MS Windows file URLs flavors

Fix for mantis 692 (http://mantis.tokeek.de/view.php?id=692)
---
 source/net/yacy/cora/document/id/MultiProtocolURL.java |  7 ++++---
 test/java/net/yacy/cora/document/id/DigestURLTest.java | 10 ++++++++--
 2 files changed, 12 insertions(+), 5 deletions(-)

diff --git a/source/net/yacy/cora/document/id/MultiProtocolURL.java b/source/net/yacy/cora/document/id/MultiProtocolURL.java
index f4136350c..856a857c8 100644
--- a/source/net/yacy/cora/document/id/MultiProtocolURL.java
+++ b/source/net/yacy/cora/document/id/MultiProtocolURL.java
@@ -284,12 +284,13 @@ public class MultiProtocolURL implements Serializable, Comparable<MultiProtocolU
                     this.path = h.substring(2); // "/path"  or "/c:/path"
                 } else if (h.startsWith("//")) { // "//host/path" or "//host/c:/path"
                     if (h.length() > 4 && h.charAt(3) == ':' && h.charAt(4) != '/' && h.charAt(4) != '\\') {
-                        // wrong windows path, after the doublepoint there should be a backslash
-                        h = h.substring(0, 4) + '\\' + h.substring(4);
+                        // wrong windows path, after the doublepoint there should be a backslash. Let's add a slash, as it will be slash in the normal form
+                        h = h.substring(0, 4) + '/' + h.substring(4);
                     }
                     int q = h.indexOf('/', 2);
                     if (q < 0 || h.length() > 3 && h.charAt(3) == ':') {
-                        this.path = h.substring(2); // "path"  or "c:/path"
+                    	// Missing root slash such as "path" or "c:/path" accepted, but the path attribute must by after all start with it
+                        this.path = "/" + h.substring(2); 
                     } else {
                         this.host = h.substring(2, q ); // TODO: handle "c:"  ?
                         if (this.host.equalsIgnoreCase(Domains.LOCALHOST)) this.host = null;
diff --git a/test/java/net/yacy/cora/document/id/DigestURLTest.java b/test/java/net/yacy/cora/document/id/DigestURLTest.java
index ffdca39c4..7756376f9 100644
--- a/test/java/net/yacy/cora/document/id/DigestURLTest.java
+++ b/test/java/net/yacy/cora/document/id/DigestURLTest.java
@@ -4,7 +4,6 @@ import java.net.MalformedURLException;
 import java.util.HashSet;
 import java.util.Set;
 import junit.framework.TestCase;
-import static junit.framework.TestCase.assertEquals;
 import net.yacy.cora.document.encoding.ASCII;
 import org.junit.Test;
 
@@ -43,11 +42,18 @@ public class DigestURLTest extends TestCase {
         String javaUrlStr = "file:///C:/tmp/test.html"; // allowed Java notation for Windows file system
 
         // allowed Windows notation
-        Set<String> testUrls = new HashSet();
+        Set<String> testUrls = new HashSet<String>();
+        /* URLs mixing slashes and backslashes */
         testUrls.add("file:///C:\\tmp\\test.html");
         testUrls.add("file:///C:/tmp\\test.html");
         testUrls.add("file:///C:\\tmp/test.html");
         testUrls.add("file:///C:/tmp/test.html");
+        
+        /* Wrong URLs missing slashes, however accepted by DigestURL and MultiProtocolURL constructors */
+        testUrls.add("file://C:/tmp/test.html");
+        testUrls.add("file://C:\\tmp\\test.html");
+        testUrls.add("file://C:tmp/test.html");
+        testUrls.add("file://C:tmp\\test.html");
 
         DigestURL javaUrl = new DigestURL(javaUrlStr);
         String javaHashResult = ASCII.String(javaUrl.hash());