fixes to SMB crawler

git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@6900 6c8d7289-2bf4-0310-a012-ef5d649a1542

source/de/anomic/crawler/retrieval/SMBLoader.java

@@ -29,10 +29,16 @@ package de.anomic.crawler.retrieval;
 
 import java.io.IOException;
 import java.io.InputStream;
+import java.net.MalformedURLException;
+import java.net.UnknownHostException;
 import java.util.ArrayList;
 import java.util.Date;
 import java.util.List;
 
+import jcifs.smb.SmbException;
+import jcifs.smb.SmbFile;
+import jcifs.smb.SmbFileInputStream;
+
 import de.anomic.http.server.HeaderFramework;
 import de.anomic.http.server.RequestHeader;
 import de.anomic.http.server.ResponseHeader;
@@ -72,8 +78,6 @@ public class SMBLoader {
         
         // process directories: transform them to html with meta robots=noindex (using the ftpc lib)
         if (url.isDirectory()) {
-            List<String> list = new ArrayList<String>();
-            String u = url.toNormalform(true, true);
             String[] l = url.list();
             if (l == null) {
                 // this can only happen if there is no connection or the directory does not exist
@@ -81,7 +85,16 @@ public class SMBLoader {
                 sb.crawlQueues.errorURL.push(request, this.sb.peers.mySeed().hash.getBytes(), new Date(), 1, "directory listing not available. URL = " + request.url().toString());
                 throw new IOException("directory listing not available. URL = " + request.url().toString());
             }
-            for (String s: l) list.add(u + s);
+            String u = url.toNormalform(true, true);
+            List<String> list = new ArrayList<String>();
+            for (String s: l) {
+                if (!s.endsWith("/") && !s.endsWith("\\")) {
+                    // check if this is a directory
+                    SmbFile sf = new SmbFile(u + s);
+                    if (sf.isDirectory()) s = s + "/";
+                }
+                list.add(u + s);
+            }
          
             StringBuilder content = ftpc.dirhtml(u, null, null, null, list, true);
             
@@ -148,4 +161,31 @@ public class SMBLoader {
         return response;
     }
  
+    public static void main(String[] args) {
+        //jcifs.Config.setProperty( "jcifs.netbios.wins", "192.168.1.220" );
+        //NtlmPasswordAuthentication auth = new NtlmPasswordAuthentication("domain", "username", "password");
+        SmbFileInputStream in;
+        try {
+            SmbFile sf = new SmbFile(args[0]);
+            if (sf.isDirectory()) {
+                String[] s = sf.list();
+                for (String t: s) System.out.println(t);
+            } else {
+                in = new SmbFileInputStream(sf);
+                byte[] b = new byte[8192];
+                int n;
+                while(( n = in.read( b )) > 0 ) {
+                    System.out.write( b, 0, n );
+                }
+            }
+        } catch (SmbException e) {
+            e.printStackTrace();
+        } catch (MalformedURLException e) {
+            e.printStackTrace();
+        } catch (UnknownHostException e) {
+            e.printStackTrace();
+        } catch (IOException e) {
+            e.printStackTrace();
+        }
+    }
 }

source/de/anomic/net/ftpc.java

@@ -2665,7 +2665,7 @@ public class ftpc {
         entryInfo info;
         for (final String line : list) {
             info = parseListData(line);
-            if(info != null) {
+            if (info != null) {
                 // with link
                 nameStart = line.indexOf(info.name);
                 page.append(line.substring(0, nameStart));
@@ -2674,9 +2674,11 @@ public class ftpc {
                 if (line.length() > nameEnd) {
                     page.append(line.substring(nameEnd));
                 }
+            } else if (line.startsWith("http://") || line.startsWith("ftp://") || line.startsWith("smb://")) {
+                page.append("<a href=\"" + line + "\">" + line + "</a>");
             } else {
-                // raw
-                page.append(line);
+               // raw
+               page.append(line);
             }
             page.append('\n');
         }

source/de/anomic/search/Switchboard.java

@@ -1749,7 +1749,7 @@ public final class Switchboard extends serverSwitch {
                 // process the next hyperlink
                 nextUrl = nextEntry.getKey();
                 String u = nextUrl.toNormalform(true, true, true);
-                if (!(u.startsWith("http") || u.startsWith("ftp"))) continue;
+                if (!(u.startsWith("http") || u.startsWith("ftp") || u.startsWith("smb"))) continue;
                 // enqueue the hyperlink into the pre-notice-url db
                 try {
                     crawlStacker.enqueueEntry(new Request(
@@ -1829,13 +1829,13 @@ public final class Switchboard extends serverSwitch {
 
         if (condenser == null || document.indexingDenied()) {
             if (this.log.isInfo()) log.logInfo("Not Indexed Resource '" + queueEntry.url().toNormalform(false, true) + "': denied by rule in document, process case=" + processCase);
-            addURLtoErrorDB(queueEntry.url(), (referrerURL == null) ? null : referrerURL.hash(), queueEntry.initiator(), dc_title, "unknown indexing process case"  + processCase);
+            addURLtoErrorDB(queueEntry.url(), (referrerURL == null) ? null : referrerURL.hash(), queueEntry.initiator(), dc_title, "denied by rule in document");
             return;
         }
         
         if (!queueEntry.profile().indexText() && !queueEntry.profile().indexMedia()) {
             if (this.log.isInfo()) log.logInfo("Not Indexed Resource '" + queueEntry.url().toNormalform(false, true) + "': denied by profile rule, process case=" + processCase);
-            addURLtoErrorDB(queueEntry.url(), (referrerURL == null) ? null : referrerURL.hash(), queueEntry.initiator(), dc_title, "unknown indexing process case"  + processCase);
+            addURLtoErrorDB(queueEntry.url(), (referrerURL == null) ? null : referrerURL.hash(), queueEntry.initiator(), dc_title, "denied by profile rule");
             return;
         }
         

source/net/yacy/kelondro/data/meta/DigestURI.java

@@ -1037,7 +1037,8 @@ public class DigestURI implements Serializable {
      */
     public SmbFile getSmbFile() throws MalformedURLException {
         if (!isSMB()) throw new UnsupportedOperationException();
-        return new SmbFile(this.toNormalform(false, true));
+        String url = this.toNormalform(false, true);
+        return new SmbFile(url);
     }
     
     // some methods that let the DigestURI look like a java.io.File object
@@ -1162,10 +1163,13 @@ public class DigestURI implements Serializable {
     public String[] list() {
         if (isFile()) return getFSFile().list();
         if (isSMB()) try {
-            return getSmbFile().list();
-        } catch (SmbException e) {
-            Log.logWarning("DigestURI", "SMB.list SmbException for " + this.toString() + ": " + e.getMessage());
-            return null;
+            SmbFile sf = getSmbFile();
+            try {
+                return sf.list();
+            } catch (SmbException e) {
+                Log.logWarning("DigestURI", "SMB.list SmbException for " + sf.toString() + ": " + e.getMessage());
+                return null;
+            }
         } catch (MalformedURLException e) {
             Log.logWarning("DigestURI", "SMB.list MalformedURLException for " + this.toString() + ": " + e.getMessage());
             return null;