Hi,
I had the problems from version 1.0 to the last from CVS. When I tried read XSL
from URL, POI throws an exception:
java.io.IOException: Unable to read entire block; 352 bytes read; expected
512 bytes
at
org.apache.poi.poifs.storage.RawDataBlock.<init>(RawDataBlock.java:98)
at
org.apache.poi.poifs.storage.RawDataBlockList.<init>(RawDataBlockList.ja
va:88)
at
org.apache.poi.poifs.filesystem.POIFSFileSystem.<init>(POIFSFileSystem.ja
va:123)
at
org.apache.poi.hssf.usermodel.HSSFWorkbook.<init>(HSSFWorkbook.java:
175)
The problem is, that the method stream.read reads less bytes from
InputStream, but there wasn't still eof. BufferedInputStream didn't solve the
problem so I tried to rewrite sourcecode of
org.apache.poi.poifs.storage.RawDataBlock.java. diff -u output is here
(RawDataBlock2.java is my rewriten version):
--- RawDataBlock.java Tue Mar 26 00:23:00 2002
+++ RawDataBlock2.java Tue Mar 26 00:21:00 2002
@@ -1,4 +1,3 @@
-
/*
==============================================================
======
* The Apache Software License, Version 1.1
*
@@ -84,22 +83,17 @@
throws IOException
{
_data = new byte[ POIFSConstants.BIG_BLOCK_SIZE ];
- int count = stream.read(_data);
-
- if (count == -1)
- {
+ int count;
+ int countfill=0;
+ do { /* if use remote InputStream */
+ /* stream.read reads less bytes, try and wait to
get another */
+ count=
stream.read(_data,countfill,POIFSConstants.BIG_BLOCK_SIZE-
countfill);
+ if (count == -1) {
_eof = true;
- }
- else if (count != POIFSConstants.BIG_BLOCK_SIZE)
- {
- String type = " byte" + ((count == 1) ? ("")
- : ("s"));
-
- throw new IOException("Unable to read entire
block; " + count
- + type + " read; expected "
- +
POIFSConstants.BIG_BLOCK_SIZE + " bytes");
- }
- else
+ break;
+ } else countfill+=count;
+ } while (countfill!=POIFSConstants.BIG_BLOCK_SIZE);
+ if (count!=-1)
{
_eof = false;
}
The problem is now solved, my program runs perfectly, but now I'm not sure
if I haven't damaged the error recognition.
Tomas Kulhanek
