Commit emails of the iDok project

["AFS account Stadler Hans Christian" <stadler_h AT savannah.psi.ch>]

Author: stadler_h
Date: Mon Oct 13 14:24:19 2008
New Revision: 275

Log:
Fixed bug causing timeout while extracting text from MS Office Documents

Modified:
   trunk/java/ch/idok/dmsd/impl/extractor/microsoft/MSExtractorProcess.java
   
trunk/java/ch/idok/dmsd/impl/extractor/microsoft/MSOfficeExtractorFactory.java

Modified: 
trunk/java/ch/idok/dmsd/impl/extractor/microsoft/MSExtractorProcess.java
==============================================================================
--- trunk/java/ch/idok/dmsd/impl/extractor/microsoft/MSExtractorProcess.java  
  (original)
+++ trunk/java/ch/idok/dmsd/impl/extractor/microsoft/MSExtractorProcess.java  
  Mon Oct 13 14:24:19 2008
@@ -218,6 +218,7 @@
      *        - args[2]: Log level
      */
     public static void main(String[] args) {
+        int retval = 1;
         try {
             Level level = Level.parse(args[2]);
             logger = Logger.getLogger("dmsd.impl.extractor.microsoft");
@@ -233,13 +234,13 @@
             logger.finest("Retrieved desktop object from Open Office");
             msep.printText(desktop, args[0], type);
             logger.finest("Printed document text to stdout, exiting with 
status 0");
-            System.exit(0);
+            retval = 0;
         } catch (DmsException ex) {
             logger.warning(Util.mergeMessages(ex.getUserMessage(), 
ex.getDetailedMessage()+"\n"+ex));
         } catch (Throwable th) {
             logger.warning("Failed to extract and print document text, 
exiting with status 1:\n"+th);
         }
-        System.exit(1);
+        System.exit(retval);
     }
 
     /**

Modified: 
trunk/java/ch/idok/dmsd/impl/extractor/microsoft/MSOfficeExtractorFactory.java
==============================================================================
--- 
trunk/java/ch/idok/dmsd/impl/extractor/microsoft/MSOfficeExtractorFactory.java
      (original)
+++ 
trunk/java/ch/idok/dmsd/impl/extractor/microsoft/MSOfficeExtractorFactory.java
      Mon Oct 13 14:24:19 2008
@@ -245,22 +245,25 @@
                 // Read the stdout output of the extraction process
                 is = proc.getInputStream();
                 int length = readLength(is)+1; // allow for 1 character of 
slack
+                int offset = 0;
                 logger.finest("length is "+length);
                 byte[] buf = new byte[0];
                 if (length > 0) {
                     buf = new byte[length];
-                    int offset = 0;
                     do {
                         int len = is.read(buf, offset, length);
                         if (len < 0) {
-                            logger.finest("Finished reading the input 
stream, nchars="+offset);
                             break;
+                        } else if (len == 0) {
+                            Thread.sleep(1000);
+                        } else {
+                            length -= len;
+                            offset += len;
                         }
-                        length -= len;
-                        offset += len;
-                    } while (true);
+                    } while (length > 0);
                 }
                 is.close(); is = null;
+                logger.finest("Finished reading the input stream, 
nchars="+offset);
                 // Log the stderr output, if any, of the extraction process
                 errThread.join(interruptDelay);
                 String errStr = errOut.getErrorOutput();