Merged revisions 709570,709598,710114,710134,710136,711505,711513-711515,711694,711739,711741,711746,711749 via svnmerge from

https://svn.apache.org/repos/asf/poi/trunk

........
  r709570 | josh | 2008-10-31 14:17:08 -0700 (Fri, 31 Oct 2008) | 1 line
  
  made DrawingSelectionRecord into a plain BIFF record (not an escher holder aggregate). Added some interpretation of fields
........
  r709598 | josh | 2008-10-31 16:24:41 -0700 (Fri, 31 Oct 2008) | 1 line
  
  Simplified multiple record expansion logic
........
  r710114 | yegor | 2008-11-03 09:54:01 -0800 (Mon, 03 Nov 2008) | 1 line
  
  fixed #46122: Picture#getEscherBSERecord threw NullPointerException if EscherContainerRecord.BSTORE_CONTAINER was not found
........
  r710134 | yegor | 2008-11-03 11:19:39 -0800 (Mon, 03 Nov 2008) | 1 line
  
  fixed bug #46033: table cells had incorrect text type resulting in corrupted style info
........
  r710136 | yegor | 2008-11-03 11:23:52 -0800 (Mon, 03 Nov 2008) | 1 line
  
  updated status of the fixed bug #46033
........
  r711505 | josh | 2008-11-04 19:50:31 -0800 (Tue, 04 Nov 2008) | 1 line
  
  Refactored test case
........
  r711513 | josh | 2008-11-04 21:45:17 -0800 (Tue, 04 Nov 2008) | 1 line
  
  Converted ConstantValueParser to use plain Strings instead of UnicodeStrings
........
  r711514 | josh | 2008-11-04 21:52:35 -0800 (Tue, 04 Nov 2008) | 1 line
  
  Converted SupBookRecord to use plain Strings instead of UnicodeStrings
........
  r711515 | josh | 2008-11-04 22:15:59 -0800 (Tue, 04 Nov 2008) | 1 line
  
  Refactored test case
........
  r711694 | josh | 2008-11-05 12:46:00 -0800 (Wed, 05 Nov 2008) | 1 line
  
  Fixed bug in conversion to/from text cells
........
  r711739 | josh | 2008-11-05 15:28:55 -0800 (Wed, 05 Nov 2008) | 1 line
  
  Refactoring test case
........
  r711741 | josh | 2008-11-05 15:35:02 -0800 (Wed, 05 Nov 2008) | 1 line
  
  Refactoring test case
........
  r711746 | josh | 2008-11-05 15:45:42 -0800 (Wed, 05 Nov 2008) | 1 line
  
  Fixed mistake in test case.  Constant value was 4 bytes too large (should be max *data* size not max *record* size).
........
  r711749 | josh | 2008-11-05 17:12:41 -0800 (Wed, 05 Nov 2008) | 1 line
  
  Introduced ContinuableRecord to help fix serialization of StringRecords with large data.  Fixed TextObjectRecord to only write 16bit unicode when needed.  Simplification in UnicodeString.
........


git-svn-id: https://svn.apache.org/repos/asf/poi/branches/ooxml@711755 13f79535-47bb-0310-9956-ffa450edef68

src/documentation/content/xdocs/changes.xml

@@ -37,6 +37,8 @@
 
 		<!-- Don't forget to update status.xml too! -->
         <release version="3.5-beta4" date="2008-??-??">
+           <action dev="POI-DEVELOPERS" type="fix">46033 - fixed TableCell to correctly set text type</action>
+           <action dev="POI-DEVELOPERS" type="fix">46122 - fixed Picture.draw to skip rendering if picture data was not found</action>
            <action dev="POI-DEVELOPERS" type="fix">15716 - memory usage optimisation - converted Ptg arrays into Formula objects</action>
            <action dev="POI-DEVELOPERS" type="add">46065 - added implementation for VALUE function</action>
            <action dev="POI-DEVELOPERS" type="add">45966 - added implementation for FIND function</action>

src/documentation/content/xdocs/status.xml

@@ -34,6 +34,8 @@
 	<!-- Don't forget to update changes.xml too! -->
     <changes>
         <release version="3.5-beta4" date="2008-??-??">
+           <action dev="POI-DEVELOPERS" type="fix">46033 - fixed TableCell to correctly set text type</action>
+           <action dev="POI-DEVELOPERS" type="fix">46122 - fixed Picture.draw to skip rendering if picture data was not found</action>
            <action dev="POI-DEVELOPERS" type="fix">15716 - memory usage optimisation - converted Ptg arrays into Formula objects</action>
            <action dev="POI-DEVELOPERS" type="add">46065 - added implementation for VALUE function</action>
            <action dev="POI-DEVELOPERS" type="add">45966 - added implementation for FIND function</action>

src/java/org/apache/poi/hssf/model/LinkTable.java

@@ -29,8 +29,9 @@ import org.apache.poi.hssf.record.ExternalNameRecord;
 import org.apache.poi.hssf.record.NameRecord;
 import org.apache.poi.hssf.record.Record;
 import org.apache.poi.hssf.record.SupBookRecord;
-import org.apache.poi.hssf.record.UnicodeString;
+import org.apache.poi.hssf.record.formula.Area3DPtg;
 import org.apache.poi.hssf.record.formula.NameXPtg;
+import org.apache.poi.hssf.record.formula.Ref3DPtg;
 
 /**
  * Link Table (OOO pdf reference: 4.10.3 ) <p/>
@@ -311,10 +312,10 @@ final class LinkTable {
 			return null;
 		}
 		int shIx = _externSheetRecord.getFirstSheetIndexFromRefIndex(extRefIndex);
-		UnicodeString usSheetName = ebr.getSheetNames()[shIx];
+		String usSheetName = ebr.getSheetNames()[shIx];
 		return new String[] {
 				ebr.getURL(),
-				usSheetName.getString(),
+				usSheetName,
 		};
 	}
 
@@ -345,9 +346,9 @@ final class LinkTable {
 		return result;
 	}
 
-	private static int getSheetIndex(UnicodeString[] sheetNames, String sheetName) {
+	private static int getSheetIndex(String[] sheetNames, String sheetName) {
 		for (int i = 0; i < sheetNames.length; i++) {
-			if (sheetNames[i].getString().equals(sheetName)) {
+			if (sheetNames[i].equals(sheetName)) {
 				return i;
 			}
 			

src/java/org/apache/poi/hssf/record/DrawingSelectionRecord.java

@@ -17,25 +17,129 @@
 
 package org.apache.poi.hssf.record;
 
-public final class DrawingSelectionRecord extends AbstractEscherHolderRecord {
+import org.apache.poi.util.HexDump;
-    public static final short sid = 0xED;
+import org.apache.poi.util.LittleEndianByteArrayOutputStream;
+import org.apache.poi.util.LittleEndianInput;
+import org.apache.poi.util.LittleEndianOutput;
 
-    public DrawingSelectionRecord()
+/**
-    {
+ * MsoDrawingSelection (0x00ED)<p/>
+ * Reference:
+ * [MS-OGRAPH].pdf sec 2.4.69
+ * 
+ * @author Josh Micich
+ */
+public final class DrawingSelectionRecord extends Record {
+	public static final short sid = 0x00ED;
+
+	/**
+	 * From [MS-ODRAW].pdf sec 2.2.1<br/>
+	 * TODO - make EscherRecordHeader {@link LittleEndianInput} aware and refactor with this
+	 */
+	private static final class OfficeArtRecordHeader {
+		public static final int ENCODED_SIZE = 8;
+		/**
+		 * lower 4 bits is 'version' usually 0x01 or 0x0F (for containers)<br/>
+		 * upper 12 bits is 'instance'
+		 */
+		private final int _verAndInstance;
+		/** value should be between 0xF000 and 0xFFFF */
+		private final int _type;
+		private final int _length;
+
+		public OfficeArtRecordHeader(LittleEndianInput in) {
+			_verAndInstance = in.readUShort();
+			_type = in.readUShort();
+			_length = in.readInt();
 		}
 
-    public DrawingSelectionRecord( RecordInputStream in )
+		public void serialize(LittleEndianOutput out) {
-    {
+			out.writeShort(_verAndInstance);
-        super( in );
+			out.writeShort(_type);
+			out.writeInt(_length);
 		}
 
-    protected String getRecordName()
+		public String debugFormatAsString() {
-    {
+			StringBuffer sb = new StringBuffer(32);
-        return "MSODRAWINGSELECTION";
+			sb.append("ver+inst=").append(HexDump.shortToHex(_verAndInstance));
+			sb.append(" type=").append(HexDump.shortToHex(_type));
+			sb.append(" len=").append(HexDump.intToHex(_length));
+			return sb.toString();
+		}
 	}
 
-    public short getSid()
+	// [MS-OGRAPH].pdf says that the data of this record is an OfficeArtFDGSL structure
-    {
+	// as described in[MS-ODRAW].pdf sec 2.2.33
+	private OfficeArtRecordHeader _header;
+	private int _cpsp;
+	/** a MSODGSLK enum value for the current selection mode */
+	private int _dgslk;
+	private int _spidFocus;
+	/** selected shape IDs (e.g. from EscherSpRecord.ShapeId) */
+	private int[] _shapeIds;
+
+	public DrawingSelectionRecord(RecordInputStream in) {
+		_header = new OfficeArtRecordHeader(in);
+		_cpsp = in.readInt();
+		_dgslk = in.readInt();
+		_spidFocus = in.readInt();
+		int nShapes = in.available() / 4;
+		int[] shapeIds = new int[nShapes];
+		for (int i = 0; i < nShapes; i++) {
+			shapeIds[i] = in.readInt();
+		}
+		_shapeIds = shapeIds;
+	}
+
+	public short getSid() {
 		return sid;
 	}
+
+	protected int getDataSize() {
+		return OfficeArtRecordHeader.ENCODED_SIZE 
+			+ 12 // 3 int fields
+			+ _shapeIds.length * 4;
+	}
+
+	public int serialize(int offset, byte[] data) {
+		int dataSize = getDataSize();
+		int recSize = 4 + dataSize;
+		LittleEndianOutput out = new LittleEndianByteArrayOutputStream(data, offset, recSize);
+		out.writeShort(sid);
+		out.writeShort(dataSize);
+		_header.serialize(out);
+		out.writeInt(_cpsp);
+		out.writeInt(_dgslk);
+		out.writeInt(_spidFocus);
+		for (int i = 0; i < _shapeIds.length; i++) {
+			out.writeInt(_shapeIds[i]);
+		}
+		return recSize;
+	}
+
+	public Object clone() {
+		// currently immutable
+		return this;
+	}
+
+	public String toString() {
+		StringBuffer sb = new StringBuffer();
+
+		sb.append("[MSODRAWINGSELECTION]\n");
+		sb.append("    .rh       =(").append(_header.debugFormatAsString()).append(")\n");
+		sb.append("    .cpsp     =").append(HexDump.intToHex(_cpsp)).append('\n');
+		sb.append("    .dgslk    =").append(HexDump.intToHex(_dgslk)).append('\n');
+		sb.append("    .spidFocus=").append(HexDump.intToHex(_spidFocus)).append('\n');
+		sb.append("    .shapeIds =(");
+		for (int i = 0; i < _shapeIds.length; i++) {
+			if (i > 0) {
+				sb.append(", ");
+			}
+			sb.append(HexDump.intToHex(_shapeIds[i]));
+		}
+		sb.append(")\n");
+
+		sb.append("[/MSODRAWINGSELECTION]\n");
+		return sb.toString();
+	}
 }

src/java/org/apache/poi/hssf/record/RecordFactory.java

@@ -49,7 +49,7 @@ public final class RecordFactory {
 	 * contains the classes for all the records we want to parse.<br/>
 	 * Note - this most but not *every* subclass of Record.
 	 */
-	private static final Class[] records = {
+	private static final Class[] recordClasses = {
 		ArrayRecord.class,
 		BackupRecord.class,
 		BlankRecord.class,
@@ -163,7 +163,7 @@ public final class RecordFactory {
 	/**
 	 * cache of the recordsToMap();
 	 */
-	private static Map recordsMap  = recordsToMap(records);
+	private static Map recordsMap  = recordsToMap(recordClasses);
 
 	private static short[] _allKnownRecordSIDs;
 	
@@ -172,16 +172,33 @@ public final class RecordFactory {
 	 * are returned digested into the non-mul form.
 	 */
 	public static Record [] createRecord(RecordInputStream in) {
+		
+		Record record = createSingleRecord(in);
+		if (record instanceof DBCellRecord) {
+			// Not needed by POI.  Regenerated from scratch by POI when spreadsheet is written
+			return new Record[] { null, };
+		}
+		if (record instanceof RKRecord) {
+			return new Record[] { convertToNumberRecord((RKRecord) record), };
+		}
+		if (record instanceof MulRKRecord) {
+			return convertRKRecords((MulRKRecord)record);
+		}
+		if (record instanceof MulBlankRecord) {
+			return convertMulBlankRecords((MulBlankRecord)record);
+		}
+		return new Record[] { record, };
+	}
+	
+	private static Record createSingleRecord(RecordInputStream in) {
 		Constructor constructor = (Constructor) recordsMap.get(new Short(in.getSid()));
 
 		if (constructor == null) {
-			return new Record[] { new UnknownRecord(in), };
+			return new UnknownRecord(in);
 		}
 		
-		Record retval;
-
 		try {
-			retval = ( Record ) constructor.newInstance(new Object[] { in });
+			return (Record) constructor.newInstance(new Object[] { in });
 		} catch (InvocationTargetException e) {
 			throw new RecordFormatException("Unable to construct record instance" , e.getTargetException());
 		} catch (IllegalArgumentException e) {
@@ -191,28 +208,28 @@ public final class RecordFactory {
 		} catch (IllegalAccessException e) {
 			throw new RuntimeException(e);
 		}
+	}
 
-		if (retval instanceof RKRecord) {
+	/**
-			// RK record is a slightly smaller alternative to NumberRecord
+	 * RK record is a slightly smaller alternative to NumberRecord
-			// POI likes NumberRecord better
+	 * POI likes NumberRecord better
-			RKRecord	 rk  = ( RKRecord ) retval;
+	 */
+	private static NumberRecord convertToNumberRecord(RKRecord rk) {
 		NumberRecord num = new NumberRecord();
 
 		num.setColumn(rk.getColumn());
 		num.setRow(rk.getRow());
 		num.setXFIndex(rk.getXFIndex());
 		num.setValue(rk.getRKNumber());
-			return new Record[] { num, };
+		return num;
 	}
-		if (retval instanceof DBCellRecord) {
-			// Not needed by POI.  Regenerated from scratch by POI when spreadsheet is written
-			return new Record[] { null, };
-		}
-		// expand multiple records where necessary
-		if (retval instanceof MulRKRecord) {
-			MulRKRecord mrk = ( MulRKRecord ) retval;
 
-			Record[] mulRecs = new Record[ mrk.getNumColumns() ];
+	/**
+	 * Converts a {@link MulRKRecord} into an equivalent array of {@link NumberRecord}s
+	 */
+	private static NumberRecord[] convertRKRecords(MulRKRecord mrk) {
+
+		NumberRecord[] mulRecs = new NumberRecord[mrk.getNumColumns()];
 		for (int k = 0; k < mrk.getNumColumns(); k++) {
 			NumberRecord nr = new NumberRecord();
 
@@ -224,10 +241,13 @@ public final class RecordFactory {
 		}
 		return mulRecs;
 	}
-		if (retval instanceof MulBlankRecord) {
-			MulBlankRecord mb = ( MulBlankRecord ) retval;
 
-			Record[] mulRecs = new Record[ mb.getNumColumns() ];
+	/**
+	 * Converts a {@link MulBlankRecord} into an equivalent array of {@link BlankRecord}s
+	 */
+	private static BlankRecord[] convertMulBlankRecords(MulBlankRecord mb) {
+
+		BlankRecord[] mulRecs = new BlankRecord[mb.getNumColumns()];
 		for (int k = 0; k < mb.getNumColumns(); k++) {
 			BlankRecord br = new BlankRecord();
 
@@ -238,8 +258,6 @@ public final class RecordFactory {
 		}
 		return mulRecs;
 	}
-		return new Record[] { retval, };
-	}
 
 	/**
 	 * @return an array of all the SIDS for all known records
@@ -325,19 +343,26 @@ public final class RecordFactory {
 				// After EOF, Excel seems to pad block with zeros
 				continue;
 			}
-			Record[] recs = createRecord(recStream);   // handle MulRK records
+			Record record = createSingleRecord(recStream);
 
-			if (recs.length > 1) {
+			if (record instanceof DBCellRecord) {
-				for (int k = 0; k < recs.length; k++) {
+				// Not needed by POI.  Regenerated from scratch by POI when spreadsheet is written
-					records.add(recs[ k ]);			   // these will be number records
-				}
 				continue;
 			}
-			Record record = recs[ 0 ];
 
-			if (record == null) {
+			if (record instanceof RKRecord) {
+				records.add(convertToNumberRecord((RKRecord) record));
 				continue;
 			}
+			if (record instanceof MulRKRecord) {
+				addAll(records, convertRKRecords((MulRKRecord)record));
+				continue;
+			}
+			if (record instanceof MulBlankRecord) {
+				addAll(records, convertMulBlankRecords((MulBlankRecord)record));
+				continue;
+			}
+
 			if (record.getSid() == DrawingGroupRecord.sid
 				   && lastRecord instanceof DrawingGroupRecord) {
 				DrawingGroupRecord lastDGRecord = (DrawingGroupRecord) lastRecord;
@@ -354,8 +379,6 @@ public final class RecordFactory {
 					records.add(record);
 				} else if (lastRecord instanceof DrawingGroupRecord) {
 					((DrawingGroupRecord)lastRecord).processContinueRecord(contRec.getData());
-				} else if (lastRecord instanceof StringRecord) {
-					((StringRecord)lastRecord).processContinueRecord(contRec.getData());
 				} else if (lastRecord instanceof UnknownRecord) {
 					//Gracefully handle records that we don't know about,
 					//that happen to be continued
@@ -373,4 +396,10 @@ public final class RecordFactory {
 		}
 		return records;
 	}
+
+	private static void addAll(List destList, Record[] srcRecs) {
+		for (int i = 0; i < srcRecs.length; i++) {
+			destList.add(srcRecs[i]);
+		}
+	}
 }

src/java/org/apache/poi/hssf/record/RecordInputStream.java

@@ -320,16 +320,6 @@ public final class RecordInputStream extends InputStream implements LittleEndian
 		}
 	}
 
-  /** Returns an excel style unicode string from the bytes reminaing in the record.
-   * <i>Note:</i> Unicode strings differ from <b>normal</b> strings due to the addition of
-   * formatting information.
-   *
-   * @return The unicode string representation of the remaining bytes.
-   */
-  public UnicodeString readUnicodeString() {
-    return new UnicodeString(this);
-  }
-
 	/** Returns the remaining bytes for the current record.
 	 *
 	  * @return The remaining bytes of the current record.

src/java/org/apache/poi/hssf/record/SSTRecord.java

@@ -17,14 +17,16 @@
 
 package org.apache.poi.hssf.record;
 
+import java.util.Iterator;
+
+import org.apache.poi.hssf.record.cont.ContinuableRecord;
+import org.apache.poi.hssf.record.cont.ContinuableRecordOutput;
 import org.apache.poi.util.IntMapper;
 import org.apache.poi.util.LittleEndianConsts;
 
-import java.util.Iterator;
-
 /**
- * Title:        Static String Table Record
+ * Title:        Static String Table Record (0x00FC)<p/>
- * <P>
+ *
  * Description:  This holds all the strings for LabelSSTRecords.
  * <P>
  * REFERENCE:    PG 389 Microsoft Excel 97 Developer's Kit (ISBN:
@@ -37,27 +39,20 @@ import java.util.Iterator;
  * @see org.apache.poi.hssf.record.LabelSSTRecord
  * @see org.apache.poi.hssf.record.ContinueRecord
  */
-public final class SSTRecord extends Record {
+public final class SSTRecord extends ContinuableRecord {
     public static final short sid = 0x00FC;
 
-      private static UnicodeString EMPTY_STRING = new UnicodeString("");
+    private static final UnicodeString EMPTY_STRING = new UnicodeString("");
-
-    /** how big can an SST record be? As big as any record can be: 8228 bytes */
-    static final int MAX_RECORD_SIZE = 8228;
 
+    // TODO - move these constants to test class (the only consumer)
     /** standard record overhead: two shorts (record id plus data space size)*/
-    static final int STD_RECORD_OVERHEAD =
+    static final int STD_RECORD_OVERHEAD = 2 * LittleEndianConsts.SHORT_SIZE;
-            2 * LittleEndianConsts.SHORT_SIZE;
 
     /** SST overhead: the standard record overhead, plus the number of strings and the number of unique strings -- two ints */
-    static final int SST_RECORD_OVERHEAD =
+    static final int SST_RECORD_OVERHEAD = STD_RECORD_OVERHEAD + 2 * LittleEndianConsts.INT_SIZE;
-            ( STD_RECORD_OVERHEAD + ( 2 * LittleEndianConsts.INT_SIZE ) );
 
     /** how much data can we stuff into an SST record? That would be _max minus the standard SST record overhead */
-    static final int MAX_DATA_SPACE = MAX_RECORD_SIZE - SST_RECORD_OVERHEAD;
+    static final int MAX_DATA_SPACE = RecordInputStream.MAX_RECORD_DATA_SIZE - 8;
-
-    /** overhead for each string includes the string's character count (a short) and the flag describing its characteristics (a byte) */
-    static final int STRING_MINIMAL_OVERHEAD = LittleEndianConsts.SHORT_SIZE + LittleEndianConsts.BYTE_SIZE;
 
     /** union of strings in the SST and EXTSST */
     private int field_1_num_strings;
@@ -133,37 +128,6 @@ public final class SSTRecord extends Record {
         return field_2_num_unique_strings;
     }
 
-    /**
-     * USE THIS METHOD AT YOUR OWN PERIL: THE <code>addString</code>
-     * METHODS MANIPULATE THE NUMBER OF STRINGS AS A SIDE EFFECT; YOUR
-     * ATTEMPTS AT MANIPULATING THE STRING COUNT IS LIKELY TO BE VERY
-     * WRONG AND WILL RESULT IN BAD BEHAVIOR WHEN THIS RECORD IS
-     * WRITTEN OUT AND ANOTHER PROCESS ATTEMPTS TO READ THE RECORD
-     *
-     * @param count  number of strings
-     *
-     */
-
-    public void setNumStrings( final int count )
-    {
-        field_1_num_strings = count;
-    }
-
-    /**
-     * USE THIS METHOD AT YOUR OWN PERIL: THE <code>addString</code>
-     * METHODS MANIPULATE THE NUMBER OF UNIQUE STRINGS AS A SIDE
-     * EFFECT; YOUR ATTEMPTS AT MANIPULATING THE UNIQUE STRING COUNT
-     * IS LIKELY TO BE VERY WRONG AND WILL RESULT IN BAD BEHAVIOR WHEN
-     * THIS RECORD IS WRITTEN OUT AND ANOTHER PROCESS ATTEMPTS TO READ
-     * THE RECORD
-     *
-     * @param count  number of strings
-     */
-
-    public void setNumUniqueStrings( final int count )
-    {
-        field_2_num_unique_strings = count;
-    }
 
     /**
      * Get a particular string by its index
@@ -178,11 +142,6 @@ public final class SSTRecord extends Record {
         return (UnicodeString) field_3_strings.get( id );
     }
 
-    public boolean isString16bit( final int id )
-    {
-        UnicodeString unicodeString = ( (UnicodeString) field_3_strings.get( id  ) );
-        return ( ( unicodeString.getOptionFlags() & 0x01 ) == 1 );
-    }
 
     /**
      * Return a debugging string representation
@@ -350,29 +309,11 @@ public final class SSTRecord extends Record {
         return field_3_strings.size();
     }
 
-    /**
+    protected void serialize(ContinuableRecordOutput out) {
-     * called by the class that is responsible for writing this sucker.
+        SSTSerializer serializer = new SSTSerializer(field_3_strings, getNumStrings(), getNumUniqueStrings() );
-     * Subclasses should implement this so that their data is passed back in a
+        serializer.serialize(out);
-     * byte array.
-     *
-     * @return size
-     */
-
-    public int serialize( int offset, byte[] data )
-    {
-        SSTSerializer serializer = new SSTSerializer(
-               field_3_strings, getNumStrings(), getNumUniqueStrings() );
-        int bytes = serializer.serialize( offset, data );
         bucketAbsoluteOffsets = serializer.getBucketAbsoluteOffsets();
         bucketRelativeOffsets = serializer.getBucketRelativeOffsets();
-        return bytes;
-    }
-
-
-    protected int getDataSize() {
-        SSTRecordSizeCalculator calculator = new SSTRecordSizeCalculator(field_3_strings);
-        int recordSize = calculator.getRecordSize();
-        return recordSize-4;
     }
 
     SSTDeserializer getDeserializer()

src/java/org/apache/poi/hssf/record/SSTRecordHeader.java

@@ -1,76 +0,0 @@
-
-/* ====================================================================
-   Licensed to the Apache Software Foundation (ASF) under one or more
-   contributor license agreements.  See the NOTICE file distributed with
-   this work for additional information regarding copyright ownership.
-   The ASF licenses this file to You under the Apache License, Version 2.0
-   (the "License"); you may not use this file except in compliance with
-   the License.  You may obtain a copy of the License at
-
-       http://www.apache.org/licenses/LICENSE-2.0
-
-   Unless required by applicable law or agreed to in writing, software
-   distributed under the License is distributed on an "AS IS" BASIS,
-   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-   See the License for the specific language governing permissions and
-   limitations under the License.
-==================================================================== */
-        
-
-package org.apache.poi.hssf.record;
-
-import org.apache.poi.util.LittleEndian;
-import org.apache.poi.util.LittleEndianConsts;
-
-/**
- * Write out an SST header record.
- *
- * @author Glen Stampoultzis (glens at apache.org)
- */
-class SSTRecordHeader
-{
-    int numStrings;
-    int numUniqueStrings;
-
-    public SSTRecordHeader( int numStrings, int numUniqueStrings )
-    {
-        this.numStrings = numStrings;
-        this.numUniqueStrings = numUniqueStrings;
-    }
-
-    /**
-     * Writes out the SST record.  This consists of the sid, the record size, the number of
-     * strings and the number of unique strings.
-     *
-     * @param data          The data buffer to write the header to.
-     * @param bufferIndex   The index into the data buffer where the header should be written.
-     * @param recSize       The number of records written.
-     *
-     * @return The bufer of bytes modified.
-     */
-    public int writeSSTHeader( UnicodeString.UnicodeRecordStats stats, byte[] data, int bufferIndex, int recSize )
-    {
-        int offset = bufferIndex;
-
-        LittleEndian.putShort( data, offset, SSTRecord.sid );
-        offset += LittleEndianConsts.SHORT_SIZE;
-        stats.recordSize += LittleEndianConsts.SHORT_SIZE;
-        stats.remainingSize -= LittleEndianConsts.SHORT_SIZE;
-        //Delay writing the length
-        stats.lastLengthPos = offset;
-        offset += LittleEndianConsts.SHORT_SIZE;
-        stats.recordSize += LittleEndianConsts.SHORT_SIZE;
-        stats.remainingSize -= LittleEndianConsts.SHORT_SIZE;
-        LittleEndian.putInt( data, offset, numStrings );
-        offset += LittleEndianConsts.INT_SIZE;
-        stats.recordSize += LittleEndianConsts.INT_SIZE;
-        stats.remainingSize -= LittleEndianConsts.INT_SIZE;
-        LittleEndian.putInt( data, offset, numUniqueStrings );
-        offset += LittleEndianConsts.INT_SIZE;
-        stats.recordSize += LittleEndianConsts.INT_SIZE;
-        stats.remainingSize -= LittleEndianConsts.INT_SIZE;
-
-        return offset - bufferIndex;
-    }
-
-}

src/java/org/apache/poi/hssf/record/SSTRecordSizeCalculator.java

@@ -1,51 +0,0 @@
-/* ====================================================================
-   Licensed to the Apache Software Foundation (ASF) under one or more
-   contributor license agreements.  See the NOTICE file distributed with
-   this work for additional information regarding copyright ownership.
-   The ASF licenses this file to You under the Apache License, Version 2.0
-   (the "License"); you may not use this file except in compliance with
-   the License.  You may obtain a copy of the License at
-
-       http://www.apache.org/licenses/LICENSE-2.0
-
-   Unless required by applicable law or agreed to in writing, software
-   distributed under the License is distributed on an "AS IS" BASIS,
-   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-   See the License for the specific language governing permissions and
-   limitations under the License.
-==================================================================== */
-        
-
-package org.apache.poi.hssf.record;
-
-import org.apache.poi.util.IntMapper;
-
-/**
- * Used to calculate the record sizes for a particular record.  This kind of
- * sucks because it's similar to the SST serialization code.  In general
- * the SST serialization code needs to be rewritten.
- *
- * @author Glen Stampoultzis (glens at apache.org)
- * @author Jason Height (jheight at apache.org)
- */
-class SSTRecordSizeCalculator
-{
-    private IntMapper strings;
-
-    public SSTRecordSizeCalculator(IntMapper strings)
-    {
-        this.strings = strings;
-    }
-
-    public int getRecordSize() {
-        UnicodeString.UnicodeRecordStats rs = new UnicodeString.UnicodeRecordStats();
-        rs.remainingSize -= SSTRecord.SST_RECORD_OVERHEAD;
-        rs.recordSize += SSTRecord.SST_RECORD_OVERHEAD;
-        for (int i=0; i < strings.size(); i++ )
-        {
-          UnicodeString unistr = ( (UnicodeString) strings.get(i));
-          unistr.getRecordSize(rs);
-    }
-        return rs.recordSize;
-    }
-}

src/java/org/apache/poi/hssf/record/SSTSerializer.java

@@ -1,4 +1,3 @@
-
 /* ====================================================================
    Licensed to the Apache Software Foundation (ASF) under one or more
    contributor license agreements.  See the NOTICE file distributed with
@@ -16,11 +15,10 @@
    limitations under the License.
 ==================================================================== */
 
-
 package org.apache.poi.hssf.record;
 
+import org.apache.poi.hssf.record.cont.ContinuableRecordOutput;
 import org.apache.poi.util.IntMapper;
-import org.apache.poi.util.LittleEndian;
 
 /**
  * This class handles serialization of SST records.  It utilizes the record processor
@@ -28,70 +26,49 @@ import org.apache.poi.util.LittleEndian;
  *
  * @author Glen Stampoultzis (glens at apache.org)
  */
-class SSTSerializer
+final class SSTSerializer {
-{
 
-    // todo: make private again
+	private final int _numStrings;
-    private IntMapper strings;
+	private final int _numUniqueStrings;
 
-    private SSTRecordHeader sstRecordHeader;
+    private final IntMapper strings;
 
     /** Offsets from the beginning of the SST record (even across continuations) */
-    int[] bucketAbsoluteOffsets;
+    private final int[] bucketAbsoluteOffsets;
     /** Offsets relative the start of the current SST or continue record */
-    int[] bucketRelativeOffsets;
+    private final int[] bucketRelativeOffsets;
     int startOfSST, startOfRecord;
 
     public SSTSerializer( IntMapper strings, int numStrings, int numUniqueStrings )
     {
         this.strings = strings;
-        this.sstRecordHeader = new SSTRecordHeader( numStrings, numUniqueStrings );
+		_numStrings = numStrings;
+		_numUniqueStrings = numUniqueStrings;
 
         int infoRecs = ExtSSTRecord.getNumberOfInfoRecsForStrings(strings.size());
         this.bucketAbsoluteOffsets = new int[infoRecs];
         this.bucketRelativeOffsets = new int[infoRecs];
     }
 
-    /**
+    public void serialize(ContinuableRecordOutput out) {
-     * Create a byte array consisting of an SST record and any
+        out.writeInt(_numStrings);
-     * required Continue records, ready to be written out.
+        out.writeInt(_numUniqueStrings);
-     * <p>
-     * If an SST record and any subsequent Continue records are read
-     * in to create this instance, this method should produce a byte
-     * array that is identical to the byte array produced by
-     * concatenating the input records' data.
-     *
-     * @return the byte array
-     */
-    public int serialize(int offset, byte[] data )
-    {
-      UnicodeString.UnicodeRecordStats stats = new UnicodeString.UnicodeRecordStats();
-      sstRecordHeader.writeSSTHeader( stats, data, 0 + offset, 0 );
-      int pos = offset + SSTRecord.SST_RECORD_OVERHEAD;
 
         for ( int k = 0; k < strings.size(); k++ )
         {
             if (k % ExtSSTRecord.DEFAULT_BUCKET_SIZE == 0)
             {
+                int rOff = out.getTotalSize();
               int index = k/ExtSSTRecord.DEFAULT_BUCKET_SIZE;
               if (index < ExtSSTRecord.MAX_BUCKETS) {
                 //Excel only indexes the first 128 buckets.
-              bucketAbsoluteOffsets[index] = pos-offset;
+                    bucketAbsoluteOffsets[index] = rOff;
-              bucketRelativeOffsets[index] = pos-offset;
+                    bucketRelativeOffsets[index] = rOff;
                 }
             }
           UnicodeString s = getUnicodeString(k);
-          pos += s.serialize(stats, pos, data);
+          s.serialize(out);
         }
-      //Check to see if there is a hanging continue record length
-      if (stats.lastLengthPos != -1) {
-        short lastRecordLength = (short)(pos - stats.lastLengthPos-2);
-        if (lastRecordLength > 8224)
-          throw new InternalError();
-
-        LittleEndian.putShort(data, stats.lastLengthPos, lastRecordLength);
-                  }
-      return pos - offset;
     }
 
 

src/java/org/apache/poi/hssf/record/StringRecord.java

@@ -17,19 +17,23 @@
 
 package org.apache.poi.hssf.record;
 
-import org.apache.poi.util.LittleEndian;
+import org.apache.poi.hssf.record.cont.ContinuableRecord;
+import org.apache.poi.hssf.record.cont.ContinuableRecordOutput;
 import org.apache.poi.util.StringUtil;
 
 /**
- * Supports the STRING record structure. (0x0207)
+ * STRING (0x0207)<p/>
+ * 
+ * Stores the cached result of a text formula
  *
  * @author Glen Stampoultzis (glens at apache.org)
  */
-public class StringRecord extends Record {
+public final class StringRecord extends ContinuableRecord {
+
 	public final static short sid = 0x0207;
-    private int                 field_1_string_length;
+
-    private byte                field_2_unicode_flag;
+	private boolean _is16bitUnicode;
-    private String              field_3_string;
+	private String _text;
 
 
     public StringRecord()
@@ -39,77 +43,24 @@ public class StringRecord extends Record {
     /**
      * @param in the RecordInputstream to read the record from
      */
-    public StringRecord( RecordInputStream in)
+    public StringRecord( RecordInputStream in) {
-    {
+        int field_1_string_length           = in.readUShort();
-        field_1_string_length           = in.readShort();
+        _is16bitUnicode            = in.readByte() != 0x00;
-        field_2_unicode_flag            = in.readByte();
-        byte[] data = in.readRemainder();
-        //Why isn't this using the in.readString methods???
-        if (isUnCompressedUnicode())
-        {
-            field_3_string = StringUtil.getFromUnicodeLE(data, 0, field_1_string_length );
-        }
-        else
-        {
-            field_3_string = StringUtil.getFromCompressedUnicode(data, 0, field_1_string_length);
-        }
-    }
         
-    public void processContinueRecord(byte[] data) {
+        if (_is16bitUnicode){
-    	if(isUnCompressedUnicode()) {
+            _text = in.readUnicodeLEString(field_1_string_length);
-    		field_3_string += StringUtil.getFromUnicodeLE(data, 0, field_1_string_length - field_3_string.length());
         } else {
-    		field_3_string += StringUtil.getFromCompressedUnicode(data, 0, field_1_string_length - field_3_string.length());
+            _text = in.readCompressedUnicode(field_1_string_length);
         }
     }
 
-    private int getStringByteLength()
+
-    {
+    protected void serialize(ContinuableRecordOutput out) {
-        return isUnCompressedUnicode() ? field_1_string_length * 2 : field_1_string_length;
+        out.writeShort(_text.length());
+        out.writeStringData(_text);
     }
 
-    protected int getDataSize() {
-        return 2 + 1 + getStringByteLength();
-    }
 
-    /**
-     * is this uncompressed unicode (16bit)?  Or just 8-bit compressed?
-     * @return isUnicode - True for 16bit- false for 8bit
-     */
-    public boolean isUnCompressedUnicode()
-    {
-        return (field_2_unicode_flag == 1);
-    }
-
-    /**
-     * called by the class that is responsible for writing this sucker.
-     * Subclasses should implement this so that their data is passed back in a
-     * byte array.
-     *
-     * @param offset to begin writing at
-     * @param data byte array containing instance data
-     * @return number of bytes written
-     */
-    public int serialize( int offset, byte[] data )
-    {
-        LittleEndian.putUShort(data, 0 + offset, sid);
-        LittleEndian.putUShort(data, 2 + offset, 3 + getStringByteLength());
-        LittleEndian.putUShort(data, 4 + offset, field_1_string_length);
-        data[6 + offset] = field_2_unicode_flag;
-        if (isUnCompressedUnicode())
-        {
-            StringUtil.putUnicodeLE(field_3_string, data, 7 + offset);
-        }
-        else
-        {
-            StringUtil.putCompressedUnicode(field_3_string, data, 7 + offset);
-        }
-        return getRecordSize();
-    }
-
-    /**
-     * return the non static version of the id for this record.
-     */
     public short getSid()
     {
         return sid;
@@ -120,26 +71,16 @@ public class StringRecord extends Record {
      */
     public String getString()
     {
-        return field_3_string;
+        return _text;
     }
 
-    /**
-     * Sets whether the string is compressed or not
-     * @param unicode_flag   1 = uncompressed, 0 = compressed
-     */
-    public void setCompressedFlag( byte unicode_flag )
-    {
-        this.field_2_unicode_flag = unicode_flag;
-    }
 
     /**
      * Sets the string represented by this record.
      */
-    public void setString( String string )
+    public void setString(String string) {
-    {
+        _text = string;
-        this.field_1_string_length = string.length();
+        _is16bitUnicode = StringUtil.hasMultibyte(string);        
-        this.field_3_string = string;
-        setCompressedFlag(StringUtil.hasMultibyte(string) ?  (byte)1 : (byte)0);        
     }
 
     public String toString()
@@ -148,16 +89,15 @@ public class StringRecord extends Record {
 
         buffer.append("[STRING]\n");
         buffer.append("    .string            = ")
-            .append(field_3_string).append("\n");
+            .append(_text).append("\n");
         buffer.append("[/STRING]\n");
         return buffer.toString();
     }
     
     public Object clone() {
         StringRecord rec = new StringRecord();
-        rec.field_1_string_length = this.field_1_string_length;
+        rec._is16bitUnicode= _is16bitUnicode;
-        rec.field_2_unicode_flag= this.field_2_unicode_flag;
+        rec._text = _text;
-        rec.field_3_string = this.field_3_string;
         return rec;
     }
 }

src/java/org/apache/poi/hssf/record/SupBookRecord.java

@@ -17,11 +17,12 @@
 
 package org.apache.poi.hssf.record;
 
-import org.apache.poi.hssf.record.UnicodeString.UnicodeRecordStats;
+import org.apache.poi.util.LittleEndianByteArrayOutputStream;
-import org.apache.poi.util.LittleEndian;
+import org.apache.poi.util.LittleEndianOutput;
+import org.apache.poi.util.StringUtil;
 
 /**
- * Title:        Sup Book (EXTERNALBOOK) <P>
+ * Title:        Sup Book - EXTERNALBOOK (0x01AE) <p/>
  * Description:  A External Workbook Description (Supplemental Book)
  *               Its only a dummy record for making new ExternSheet Record <P>
  * REFERENCE:  5.38<P>
@@ -31,15 +32,15 @@ import org.apache.poi.util.LittleEndian;
  */
 public final class SupBookRecord extends Record {
 
-    public final static short sid = 0x1AE;
+    public final static short sid = 0x01AE;
 
     private static final short SMALL_RECORD_SIZE = 4;
     private static final short TAG_INTERNAL_REFERENCES = 0x0401;
     private static final short TAG_ADD_IN_FUNCTIONS = 0x3A01;
 
     private short field_1_number_of_sheets;
-    private UnicodeString     field_2_encoded_url;
+    private String field_2_encoded_url;
-    private UnicodeString[]   field_3_sheet_names;
+    private String[] field_3_sheet_names;
     private boolean _isAddInFunctions;
 
 
@@ -49,7 +50,7 @@ public final class SupBookRecord extends Record {
     public static SupBookRecord createAddInFunctions() {
         return new SupBookRecord(true, (short)0);
     }
-    public static SupBookRecord createExternalReferences(UnicodeString url, UnicodeString[] sheetNames) {
+    public static SupBookRecord createExternalReferences(String url, String[] sheetNames) {
         return new SupBookRecord(url, sheetNames);
     }
     private SupBookRecord(boolean isAddInFuncs, short numberOfSheets) {
@@ -59,7 +60,7 @@ public final class SupBookRecord extends Record {
         field_3_sheet_names = null;
         _isAddInFunctions = isAddInFuncs;
     }
-    public SupBookRecord(UnicodeString url, UnicodeString[] sheetNames) {
+    public SupBookRecord(String url, String[] sheetNames) {
         field_1_number_of_sheets = (short) sheetNames.length;
         field_2_encoded_url = url;
         field_3_sheet_names = sheetNames;
@@ -92,10 +93,10 @@ public final class SupBookRecord extends Record {
             // 5.38.1 External References
             _isAddInFunctions = false;
 
-            field_2_encoded_url = in.readUnicodeString();
+            field_2_encoded_url = in.readString();
-            UnicodeString[] sheetNames = new UnicodeString[field_1_number_of_sheets];
+            String[] sheetNames = new String[field_1_number_of_sheets];
             for (int i = 0; i < sheetNames.length; i++) {
-                sheetNames[i] = in.readUnicodeString();
+                sheetNames[i] = in.readString();
             }
             field_3_sheet_names = sheetNames;
             return;
@@ -143,18 +144,14 @@ public final class SupBookRecord extends Record {
             return SMALL_RECORD_SIZE;
         }
         int sum = 2; // u16 number of sheets
-        UnicodeRecordStats urs = new UnicodeRecordStats();
+
-        field_2_encoded_url.getRecordSize(urs);
+        sum += StringUtil.getEncodedSize(field_2_encoded_url);
-        sum += urs.recordSize;
 
         for(int i=0; i<field_3_sheet_names.length; i++) {
-            urs = new UnicodeRecordStats();
+            sum += StringUtil.getEncodedSize(field_3_sheet_names[i]);
-            field_3_sheet_names[i].getRecordSize(urs);
-            sum += urs.recordSize;
         }
         return sum;
     }
-
     /**
      * called by the class that is responsible for writing this sucker.
      * Subclasses should implement this so that their data is passed back in a
@@ -165,29 +162,26 @@ public final class SupBookRecord extends Record {
      * @return number of bytes written
      */
     public int serialize(int offset, byte [] data) {
-        LittleEndian.putShort(data, 0 + offset, sid);
         int dataSize = getDataSize();
-        LittleEndian.putShort(data, 2 + offset, (short) dataSize);
+        int recordSize = 4 + dataSize;
-        LittleEndian.putShort(data, 4 + offset, field_1_number_of_sheets);
+        LittleEndianOutput out = new LittleEndianByteArrayOutputStream(data, offset, recordSize);
+
+        out.writeShort(sid);
+        out.writeShort(dataSize);
+        out.writeShort(field_1_number_of_sheets);
 
         if(isExternalReferences()) {
-            
+            StringUtil.writeUnicodeString(out, field_2_encoded_url);
-            int currentOffset = 6 + offset;
-            UnicodeRecordStats urs = new UnicodeRecordStats();
-            field_2_encoded_url.serialize(urs, currentOffset, data);
-            currentOffset += urs.recordSize;
 
             for(int i=0; i<field_3_sheet_names.length; i++) {
-                urs = new UnicodeRecordStats();
+                StringUtil.writeUnicodeString(out, field_3_sheet_names[i]);
-                field_3_sheet_names[i].serialize(urs, currentOffset, data);
-                currentOffset += urs.recordSize;
             }
         } else {
-            short field2val = _isAddInFunctions ? TAG_ADD_IN_FUNCTIONS : TAG_INTERNAL_REFERENCES;
+            int field2val = _isAddInFunctions ? TAG_ADD_IN_FUNCTIONS : TAG_INTERNAL_REFERENCES;
 
-            LittleEndian.putShort(data, 6 + offset, field2val);
+            out.writeShort(field2val);
         }
-        return dataSize + 4;
+        return recordSize;
     }
 
     public void setNumberOfSheets(short number){
@@ -203,7 +197,7 @@ public final class SupBookRecord extends Record {
         return sid;
     }
     public String getURL() {
-        String encodedUrl = field_2_encoded_url.getString();
+        String encodedUrl = field_2_encoded_url;
         switch(encodedUrl.charAt(0)) {
             case 0: // Reference to an empty workbook name
                 return encodedUrl.substring(1); // will this just be empty string?
@@ -230,7 +224,7 @@ public final class SupBookRecord extends Record {
 
         */
     }
-    public UnicodeString[] getSheetNames() {
+    public String[] getSheetNames() {
-        return (UnicodeString[]) field_3_sheet_names.clone();
+        return (String[]) field_3_sheet_names.clone();
     }
 }

src/java/org/apache/poi/hssf/record/TextObjectRecord.java

@@ -17,16 +17,13 @@
 
 package org.apache.poi.hssf.record;
 
-import java.io.UnsupportedEncodingException;
+import org.apache.poi.hssf.record.cont.ContinuableRecord;
-
+import org.apache.poi.hssf.record.cont.ContinuableRecordOutput;
 import org.apache.poi.hssf.record.formula.Ptg;
 import org.apache.poi.hssf.usermodel.HSSFRichTextString;
 import org.apache.poi.util.BitField;
 import org.apache.poi.util.BitFieldFactory;
 import org.apache.poi.util.HexDump;
-import org.apache.poi.util.LittleEndian;
-import org.apache.poi.util.LittleEndianByteArrayOutputStream;
-import org.apache.poi.util.LittleEndianOutput;
 
 /**
  * The TXO record (0x01B6) is used to define the properties of a text box. It is
@@ -36,7 +33,7 @@ import org.apache.poi.util.LittleEndianOutput;
  * 
  * @author Glen Stampoultzis (glens at apache.org)
  */
-public final class TextObjectRecord extends Record {
+public final class TextObjectRecord extends ContinuableRecord {
 	public final static short sid = 0x01B6;
 
 	private static final int FORMAT_RUN_ENCODED_SIZE = 8; // 2 shorts and 4 bytes reserved
@@ -163,30 +160,7 @@ public final class TextObjectRecord extends Record {
 		return sid;
 	}
 
-	/**
+	private void serializeTXORecord(ContinuableRecordOutput out) {
-	 * Only for the current record. does not include any subsequent Continue
-	 * records
-	 */
-	private int getCurrentRecordDataSize() {
-		int result = 2 + 2 + 2 + 2 + 2 + 2 + 2 + 4;
-		if (_linkRefPtg != null) {
-			result += 2 // formula size
-				+ 4  // unknownInt
-				+_linkRefPtg.getSize();
-			if (_unknownPostFormulaByte != null) {
-				result += 1;
-			}
-		}
-		return result;
-	}
-
-	private int serializeTXORecord(int offset, byte[] data) {
-		int dataSize = getCurrentRecordDataSize();
-		int recSize = dataSize+4;
-		LittleEndianOutput out = new LittleEndianByteArrayOutputStream(data, offset, recSize);
-		
-		out.writeShort(TextObjectRecord.sid);
-		out.writeShort(dataSize);
 		
 		out.writeShort(field_1_options);
 		out.writeShort(field_2_textOrientation);
@@ -206,79 +180,23 @@ public final class TextObjectRecord extends Record {
 				out.writeByte(_unknownPostFormulaByte.byteValue());
 			}
 		}
-		return recSize;
 	}
 
-	private int serializeTrailingRecords(int offset, byte[] data) {
+	private void serializeTrailingRecords(ContinuableRecordOutput out) {
-		byte[] textBytes;
+		out.writeContinue();
-		try {
+		out.writeStringData(_text.getString());
-			textBytes = _text.getString().getBytes("UTF-16LE");
+		out.writeContinue();
-		} catch (UnsupportedEncodingException e) {
+		writeFormatData(out, _text);
-			throw new RuntimeException(e.getMessage(), e);
-		}
-		int remainingLength = textBytes.length;
-
-		int countTextBytesWritten = 0;
-		int pos = offset;
-		// (regardless what was read, we always serialize double-byte
-		// unicode characters (UTF-16LE).
-		Byte unicodeFlag = new Byte((byte)1);
-		while (remainingLength > 0) {
-			int chunkSize = Math.min(RecordInputStream.MAX_RECORD_DATA_SIZE - 2, remainingLength);
-			remainingLength -= chunkSize;
-			pos += ContinueRecord.write(data, pos, unicodeFlag, textBytes, countTextBytesWritten, chunkSize);
-			countTextBytesWritten += chunkSize;
 	}
 
-		byte[] formatData = createFormatData(_text);
+	protected void serialize(ContinuableRecordOutput out) {
-		pos += ContinueRecord.write(data, pos, null, formatData);
-		return pos - offset;
-	}
-
-	private int getTrailingRecordsSize() {
-		if (_text.length() < 1) {
-			return 0;
-		}
-		int encodedTextSize = 0;
-		int textBytesLength = _text.length() * LittleEndian.SHORT_SIZE;
-		while (textBytesLength > 0) {
-			int chunkSize = Math.min(RecordInputStream.MAX_RECORD_DATA_SIZE - 2, textBytesLength);
-			textBytesLength -= chunkSize;
-
-			encodedTextSize += 4;           // +4 for ContinueRecord sid+size
-			encodedTextSize += 1+chunkSize; // +1 for compressed unicode flag, 
-		}
-
-		int encodedFormatSize = (_text.numFormattingRuns() + 1) * FORMAT_RUN_ENCODED_SIZE
-			+ 4;  // +4 for ContinueRecord sid+size
-		return encodedTextSize + encodedFormatSize;
-	}
-
-
-	public int serialize(int offset, byte[] data) {
-
-		int expectedTotalSize = getRecordSize();
-		int totalSize = serializeTXORecord(offset, data);
 
+		serializeTXORecord(out);
 		if (_text.getString().length() > 0) {
-			totalSize += serializeTrailingRecords(offset+totalSize, data);
+			serializeTrailingRecords(out);
 		} 
-		
-		if (totalSize != expectedTotalSize)
-			throw new RecordFormatException(totalSize
-					+ " bytes written but getRecordSize() reports " + expectedTotalSize);
-		return totalSize;
 	}
 
-	/**
-	 * Note - this total size includes all potential {@link ContinueRecord}s written
-	 * but it is not the "ushort size" value to be written at the start of the first BIFF record 
-	 */
-	protected int getDataSize() {
-		return getCurrentRecordDataSize() + getTrailingRecordsSize();
-	}
-
-	
 	private int getFormattingDataLength() {
 		if (_text.length() < 1) {
 			// important - no formatting data if text is empty 
@@ -287,25 +205,17 @@ public final class TextObjectRecord extends Record {
 		return (_text.numFormattingRuns() + 1) * FORMAT_RUN_ENCODED_SIZE;
 	}
 
-	private static byte[] createFormatData(HSSFRichTextString str) {
+	private static void writeFormatData(ContinuableRecordOutput out , HSSFRichTextString str) {
 		int nRuns = str.numFormattingRuns();
-		byte[] result = new byte[(nRuns + 1) * FORMAT_RUN_ENCODED_SIZE];
-		int pos = 0;
 		for (int i = 0; i < nRuns; i++) {
-			LittleEndian.putUShort(result, pos, str.getIndexOfFormattingRun(i));
+			out.writeShort(str.getIndexOfFormattingRun(i));
-			pos += 2;
 			int fontIndex = str.getFontOfFormattingRun(i);
-			LittleEndian.putUShort(result, pos, fontIndex == str.NO_FONT ? 0 : fontIndex);
+			out.writeShort(fontIndex == str.NO_FONT ? 0 : fontIndex);
-			pos += 2;
+			out.writeInt(0); // skip reserved
-			pos += 4; // skip reserved
 		}
-		LittleEndian.putUShort(result, pos, str.length());
+		out.writeShort(str.length());
-		pos += 2;
+		out.writeShort(0);
-		LittleEndian.putUShort(result, pos, 0);
+		out.writeInt(0); // skip reserved
-		pos += 2;
-		pos += 4; // skip reserved
-
-		return result;
 	}
 
 	/**

src/java/org/apache/poi/hssf/record/UnicodeString.java

@@ -17,30 +17,31 @@
 
 package org.apache.poi.hssf.record;
 
-import org.apache.poi.util.BitField;
-import org.apache.poi.util.BitFieldFactory;
-import org.apache.poi.util.LittleEndian;
-import org.apache.poi.util.HexDump;
-
-import java.util.Iterator;
-import java.util.List;
 import java.util.ArrayList;
 import java.util.Collections;
+import java.util.Iterator;
+import java.util.List;
+
+import org.apache.poi.hssf.record.cont.ContinuableRecordOutput;
+import org.apache.poi.util.BitField;
+import org.apache.poi.util.BitFieldFactory;
+import org.apache.poi.util.HexDump;
+import org.apache.poi.util.LittleEndianInput;
+import org.apache.poi.util.LittleEndianOutput;
 
 /**
- * Title: Unicode String<P>
+ * Title: Unicode String<p/>
- * Description:  Unicode String record.  We implement these as a record, although
+ * Description:  Unicode String - just standard fields that are in several records.
- *               they are really just standard fields that are in several records.
+ *               It is considered more desirable then repeating it in all of them.<p/>
- *               It is considered more desirable then repeating it in all of them.<P>
+ * REFERENCE:  PG 264 Microsoft Excel 97 Developer's Kit (ISBN: 1-57231-498-2)<p/>
- * REFERENCE:  PG 264 Microsoft Excel 97 Developer's Kit (ISBN: 1-57231-498-2)<P>
  * @author  Andrew C. Oliver
  * @author Marc Johnson (mjohnson at apache dot org)
  * @author Glen Stampoultzis (glens at apache.org)
  */
 public final class UnicodeString implements Comparable {
-    private short             field_1_charCount;     // = 0;
+    private short             field_1_charCount;
-    private byte              field_2_optionflags;   // = 0;
+    private byte              field_2_optionflags;
-    private String            field_3_string;        // = null;
+    private String            field_3_string;
     private List field_4_format_runs;
     private byte[] field_5_ext_rst;
     private static final BitField   highByte  = BitFieldFactory.getInstance(0x1);
@@ -56,6 +57,10 @@ public final class UnicodeString implements Comparable {
             this.fontIndex = fontIndex;
         }
 
+        public FormatRun(LittleEndianInput in) {
+            this(in.readShort(), in.readShort());
+        }
+
         public short getCharacterPos() {
             return character;
         }
@@ -65,13 +70,12 @@ public final class UnicodeString implements Comparable {
         }
 
         public boolean equals(Object o) {
-        if ((o == null) || (o.getClass() != this.getClass()))
+            if (!(o instanceof FormatRun)) {
-    {
                 return false;
             }
             FormatRun other = ( FormatRun ) o;
 
-        return ((character == other.character) && (fontIndex == other.fontIndex));
+            return character == other.character && fontIndex == other.fontIndex;
         }
 
         public int compareTo(Object obj) {
@@ -86,6 +90,11 @@ public final class UnicodeString implements Comparable {
         public String toString() {
             return "character="+character+",fontIndex="+fontIndex;
         }
+
+        public void serialize(LittleEndianOutput out) {
+            out.writeShort(character);
+            out.writeShort(fontIndex);
+        }
     }
 
     private UnicodeString() {
@@ -116,13 +125,12 @@ public final class UnicodeString implements Comparable {
      */
     public boolean equals(Object o)
     {
-        if ((o == null) || (o.getClass() != this.getClass()))
+        if (!(o instanceof UnicodeString)) {
-        {
             return false;
         }
         UnicodeString other = (UnicodeString) o;
 
-        //Ok lets do this in stages to return a quickly, first check the actual string
+        //OK lets do this in stages to return a quickly, first check the actual string
         boolean eq = ((field_1_charCount == other.field_1_charCount)
                 && (field_2_optionflags == other.field_2_optionflags)
                 && field_3_string.equals(other.field_3_string));
@@ -203,9 +211,7 @@ public final class UnicodeString implements Comparable {
         if (isRichText() && (runCount > 0)) {
           field_4_format_runs = new ArrayList(runCount);
           for (int i=0;i<runCount;i++) {
-            field_4_format_runs.add(new FormatRun(in.readShort(), in.readShort()));
+            field_4_format_runs.add(new FormatRun(in));
-            //read reserved
-            //in.readInt();
           }
         }
 
@@ -372,11 +378,8 @@ public final class UnicodeString implements Comparable {
       field_2_optionflags = richText.clearByte(field_2_optionflags);
     }
 
-    public byte[] getExtendedRst() {
-       return this.field_5_ext_rst;
-    }
 
-    public void setExtendedRst(byte[] ext_rst) {
+    void setExtendedRst(byte[] ext_rst) {
       if (ext_rst != null)
         field_2_optionflags = extBit.setByte(field_2_optionflags);
       else field_2_optionflags = extBit.clearByte(field_2_optionflags);
@@ -442,351 +445,43 @@ public final class UnicodeString implements Comparable {
         return buffer.toString();
     }
 
-    private int writeContinueIfRequired(UnicodeRecordStats stats, final int requiredSize, int offset, byte[] data) {
+    public void serialize(ContinuableRecordOutput out) {
-      //Basic string overhead
+        int numberOfRichTextRuns = 0;
-      if (stats.remainingSize < requiredSize) {
+        int extendedDataSize = 0;
-        //Check if be are already in a continue record, if so make sure that
+        if (isRichText() && field_4_format_runs != null) {
-        //we go back and write out our length
+            numberOfRichTextRuns = field_4_format_runs.size();
-        if (stats.lastLengthPos != -1) {
+        }
-          short lastRecordLength = (short)(offset - stats.lastLengthPos - 2);
+        if (isExtendedText() && field_5_ext_rst != null) {
-          if (lastRecordLength > 8224)
+            extendedDataSize = field_5_ext_rst.length;
-            throw new InternalError();
-          LittleEndian.putShort(data, stats.lastLengthPos, lastRecordLength);
         }
     
-        LittleEndian.putShort(data, offset, ContinueRecord.sid);
+        out.writeString(field_3_string, numberOfRichTextRuns, extendedDataSize);
-        offset+=2;
-        //Record the location of the last continue length position, but don't write
-        //anything there yet (since we don't know what it will be!)
-        stats.lastLengthPos = offset;
-        offset += 2;
 
-        stats.recordSize += 4;
+        if (numberOfRichTextRuns > 0) {
-        stats.remainingSize = SSTRecord.MAX_RECORD_SIZE-4;
-      }
-      return offset;
-    }
-
-    public int serialize(UnicodeRecordStats stats, final int offset, byte [] data)
-    {
-      int pos = offset;
-
-      //Basic string overhead
-      pos = writeContinueIfRequired(stats, 3, pos, data);
-      LittleEndian.putShort(data, pos, getCharCount());
-      pos += 2;
-      data[ pos ] = getOptionFlags();
-      pos += 1;
-      stats.recordSize += 3;
-      stats.remainingSize-= 3;
-
-      if (isRichText()) {
-        if (field_4_format_runs != null) {
-          pos = writeContinueIfRequired(stats, 2, pos, data);
-
-          LittleEndian.putShort(data, pos, (short) field_4_format_runs.size());
-          pos += 2;
-          stats.recordSize += 2;
-          stats.remainingSize -= 2;
-        }
-      }
-      if ( isExtendedText() )
-      {
-        if (this.field_5_ext_rst != null) {
-          pos = writeContinueIfRequired(stats, 4, pos, data);
-
-          LittleEndian.putInt(data, pos, field_5_ext_rst.length);
-          pos += 4;
-          stats.recordSize += 4;
-          stats.remainingSize -= 4;
-        }
-      }
-
-      int charsize = isUncompressedUnicode() ? 2 : 1;
-      int strSize = (getString().length() * charsize);
-
-      byte[] strBytes = null;
-        try {
-            String unicodeString = getString();
-              if (!isUncompressedUnicode())
-            {
-                strBytes = unicodeString.getBytes("ISO-8859-1");
-            }
-            else
-            {
-                  strBytes = unicodeString.getBytes("UTF-16LE");
-            }
-        }
-        catch (Exception e) {
-              throw new InternalError();
-        }
-          if (strSize != strBytes.length)
-            throw new InternalError("That shouldnt have happened!");
-
-      //Check to see if the offset occurs mid string, if so then we need to add
-      //the byte to start with that represents the first byte of the continue record.
-      if (strSize > stats.remainingSize) {
-        //OK the offset occurs half way through the string, that means that
-        //we need an extra byte after the continue record ie we didnt finish
-        //writing out the string the 1st time through
-
-        //But hang on, how many continue records did we span? What if this is
-        //a REALLY long string. We need to work this all out.
-        int amountThatCantFit = strSize;
-        int strPos = 0;
-        while (amountThatCantFit > 0) {
-          int amountWritten = Math.min(stats.remainingSize, amountThatCantFit);
-          //Make sure that the amount that can't fit takes into account
-          //whether we are writing double byte unicode
-          if (isUncompressedUnicode()) {
-            //We have the '-1' here because whether this is the first record or
-            //subsequent continue records, there is always the case that the
-            //number of bytes in a string on double byte boundaries is actually odd.
-            if ( ( (amountWritten ) % 2) == 1)
-              amountWritten--;
-          }
-          System.arraycopy(strBytes, strPos, data, pos, amountWritten);
-          pos += amountWritten;
-          strPos += amountWritten;
-          stats.recordSize += amountWritten;
-          stats.remainingSize -= amountWritten;
-
-          //Ok lets subtract what we can write
-          amountThatCantFit -= amountWritten;
-
-          //Each iteration of this while loop is another continue record, unless
-          //everything  now fits.
-          if (amountThatCantFit > 0) {
-            //We know that a continue WILL be requied, but use this common method
-            pos = writeContinueIfRequired(stats, amountThatCantFit, pos, data);
-
-            //The first byte after a continue mid string is the extra byte to
-            //indicate if this run is compressed or not.
-            data[pos] = (byte) (isUncompressedUnicode() ? 0x1 : 0x0);
-            pos++;
-            stats.recordSize++;
-            stats.remainingSize --;
-          }
-        }
-      } else {
-        if (strSize > (data.length-pos))
-          System.out.println("Hmm shouldnt happen");
-        //Ok the string fits nicely in the remaining size
-        System.arraycopy(strBytes, 0, data, pos, strSize);
-        pos += strSize;
-        stats.recordSize += strSize;
-        stats.remainingSize -= strSize;
-      }
-
-
-      if (isRichText() && (field_4_format_runs != null)) {
-        int count = field_4_format_runs.size();
 
           //This will ensure that a run does not split a continue
-        for (int i=0;i<count;i++) {
+          for (int i=0;i<numberOfRichTextRuns;i++) {
-          pos = writeContinueIfRequired(stats, 4, pos, data);
+              if (out.getAvailableSpace() < 4) {
+                  out.writeContinue();
+              }
                 FormatRun r = (FormatRun)field_4_format_runs.get(i);
-          LittleEndian.putShort(data, pos, r.character);
+                r.serialize(out);
-          pos += 2;
-          LittleEndian.putShort(data, pos, r.fontIndex);
-          pos += 2;
-
-          //Each run count is four bytes
-          stats.recordSize += 4;
-          stats.remainingSize -=4;
           }
         }
 
-      if (isExtendedText() && (field_5_ext_rst != null)) {
+        if (extendedDataSize > 0) {
-        //Ok ExtRst is actually not documented, so i am going to hope
+            // OK ExtRst is actually not documented, so i am going to hope
             // that we can actually continue on byte boundaries
-        int ammountThatCantFit = field_5_ext_rst.length - stats.remainingSize;
+
             int extPos = 0;
-        if (ammountThatCantFit > 0) {
+            while (true) {
-          while (ammountThatCantFit > 0) {
+                int nBytesToWrite = Math.min(extendedDataSize - extPos, out.getAvailableSpace());
-            //So for this record we have already written
+                out.write(field_5_ext_rst, extPos, nBytesToWrite);
-            int ammountWritten = Math.min(stats.remainingSize, ammountThatCantFit);
+                extPos += nBytesToWrite;
-            System.arraycopy(field_5_ext_rst, extPos, data, pos, ammountWritten);
+                if (extPos >= extendedDataSize) {
-            pos += ammountWritten;
+                    break;
-            extPos += ammountWritten;
-            stats.recordSize += ammountWritten;
-            stats.remainingSize -= ammountWritten;
-
-            //Ok lets subtract what we can write
-            ammountThatCantFit -= ammountWritten;
-            if (ammountThatCantFit > 0) {
-              pos = writeContinueIfRequired(stats, 1, pos, data);
                 }
-          }
+                out.writeContinue();
-        } else {
-          //We can fit wholey in what remains.
-          System.arraycopy(field_5_ext_rst, 0, data, pos, field_5_ext_rst.length);
-          pos +=  field_5_ext_rst.length;
-          stats.remainingSize -= field_5_ext_rst.length;
-          stats.recordSize += field_5_ext_rst.length;
-        }
-      }
-
-        return pos - offset;
-    }
-
-
-    public void setCompressedUnicode() {
-      field_2_optionflags = highByte.setByte(field_2_optionflags);
-    }
-
-    public void setUncompressedUnicode() {
-      field_2_optionflags = highByte.clearByte(field_2_optionflags);
-    }
-
-    private boolean isUncompressedUnicode()
-    {
-        return highByte.isSet(getOptionFlags());
-    }
-
-    /** Returns the size of this record, given the amount of record space
-     * remaining, it will also include the size of writing a continue record.
-     */
-
-    public static class UnicodeRecordStats {
-      public int recordSize;
-      public int remainingSize = SSTRecord.MAX_RECORD_SIZE;
-      public int lastLengthPos = -1;
-    }
-    public void getRecordSize(UnicodeRecordStats stats) {
-      //Basic string overhead
-      if (stats.remainingSize < 3) {
-        //Needs a continue
-         stats.recordSize += 4;
-         stats.remainingSize = SSTRecord.MAX_RECORD_SIZE-4;
-      }
-      stats.recordSize += 3;
-      stats.remainingSize-= 3;
-
-      //Read the number of rich runs if rich text.
-      if ( isRichText() )
-    {
-          //Run count
-          if (stats.remainingSize < 2) {
-            //Needs a continue
-            //Reset the available space.
-            stats.remainingSize = SSTRecord.MAX_RECORD_SIZE-4;
-            //continue record overhead
-            stats.recordSize+=4;
-    }
-
-          stats.recordSize += 2;
-          stats.remainingSize -=2;
-      }
-      //Read the size of extended data if present.
-      if ( isExtendedText() )
-    {
-        //Needs a continue
-          //extension length
-          if (stats.remainingSize < 4) {
-            //Reset the available space.
-            stats.remainingSize = SSTRecord.MAX_RECORD_SIZE-4;
-            //continue record overhead
-            stats.recordSize+=4;
-          }
-
-          stats.recordSize += 4;
-          stats.remainingSize -=4;
-      }
-
-      int charsize = isUncompressedUnicode() ? 2 : 1;
-      int strSize = (getString().length() * charsize);
-      //Check to see if the offset occurs mid string, if so then we need to add
-      //the byte to start with that represents the first byte of the continue record.
-      if (strSize > stats.remainingSize) {
-        //Ok the offset occurs half way through the string, that means that
-        //we need an extra byte after the continue record ie we didnt finish
-        //writing out the string the 1st time through
-
-        //But hang on, how many continue records did we span? What if this is
-        //a REALLY long string. We need to work this all out.
-        int ammountThatCantFit = strSize;
-        while (ammountThatCantFit > 0) {
-          int ammountWritten = Math.min(stats.remainingSize, ammountThatCantFit);
-          //Make sure that the ammount that cant fit takes into account
-          //whether we are writing double byte unicode
-          if (isUncompressedUnicode()) {
-            //We have the '-1' here because whether this is the first record or
-            //subsequent continue records, there is always the case that the
-            //number of bytes in a string on doube byte boundaries is actually odd.
-            if ( ( (ammountWritten) % 2) == 1)
-              ammountWritten--;
-          }
-          stats.recordSize += ammountWritten;
-          stats.remainingSize -= ammountWritten;
-
-          //Ok lets subtract what we can write
-          ammountThatCantFit -= ammountWritten;
-
-          //Each iteration of this while loop is another continue record, unless
-          //everything  now fits.
-          if (ammountThatCantFit > 0) {
-            //Reset the available space.
-            stats.remainingSize = SSTRecord.MAX_RECORD_SIZE-4;
-            //continue record overhead
-            stats.recordSize+=4;
-
-            //The first byte after a continue mid string is the extra byte to
-            //indicate if this run is compressed or not.
-            stats.recordSize++;
-            stats.remainingSize --;
-          }
-        }
-      } else {
-        //Ok the string fits nicely in the remaining size
-        stats.recordSize += strSize;
-        stats.remainingSize -= strSize;
-      }
-
-      if (isRichText() && (field_4_format_runs != null)) {
-        int count = field_4_format_runs.size();
-
-        //This will ensure that a run does not split a continue
-        for (int i=0;i<count;i++) {
-          if (stats.remainingSize < 4) {
-            //Reset the available space.
-            stats.remainingSize = SSTRecord.MAX_RECORD_SIZE-4;
-            //continue record overhead
-            stats.recordSize+=4;
-          }
-
-          //Each run count is four bytes
-          stats.recordSize += 4;
-          stats.remainingSize -=4;
-        }
-      }
-
-      if (isExtendedText() && (field_5_ext_rst != null)) {
-        //Ok ExtRst is actually not documented, so i am going to hope
-        //that we can actually continue on byte boundaries
-        int ammountThatCantFit = field_5_ext_rst.length - stats.remainingSize;
-        if (ammountThatCantFit > 0) {
-          while (ammountThatCantFit > 0) {
-            //So for this record we have already written
-            int ammountWritten = Math.min(stats.remainingSize, ammountThatCantFit);
-            stats.recordSize += ammountWritten;
-            stats.remainingSize -= ammountWritten;
-
-            //Ok lets subtract what we can write
-            ammountThatCantFit -= ammountWritten;
-            if (ammountThatCantFit > 0) {
-              //Each iteration of this while loop is another continue record.
-
-              //Reset the available space.
-              stats.remainingSize = SSTRecord.MAX_RECORD_SIZE-4;
-              //continue record overhead
-              stats.recordSize += 4;
-            }
-          }
-        } else {
-          //We can fit wholey in what remains.
-          stats.remainingSize -= field_5_ext_rst.length;
-          stats.recordSize += field_5_ext_rst.length;
             }
         }
     }
@@ -801,9 +496,9 @@ public final class UnicodeString implements Comparable {
         if (result != 0)
           return result;
 
-        //Ok string appears to be equal but now lets compare formatting runs
+        //OK string appears to be equal but now lets compare formatting runs
         if ((field_4_format_runs == null) && (str.field_4_format_runs == null))
-          //Strings are equal, and there are no formtting runs.
+          //Strings are equal, and there are no formatting runs.
           return 0;
 
         if ((field_4_format_runs == null) && (str.field_4_format_runs != null))
@@ -850,12 +545,12 @@ public final class UnicodeString implements Comparable {
         return 0;
     }
 
-    public boolean isRichText()
+    private boolean isRichText()
     {
       return richText.isSet(getOptionFlags());
     }
 
-    public boolean isExtendedText()
+    private boolean isExtendedText()
         {
         return extBit.isSet(getOptionFlags());
     }
@@ -881,6 +576,4 @@ public final class UnicodeString implements Comparable {
 
         return str;
     }
-
-
 }

src/java/org/apache/poi/hssf/record/constant/ConstantValueParser.java

@@ -17,8 +17,6 @@
 
 package org.apache.poi.hssf.record.constant;
 
-import org.apache.poi.hssf.record.UnicodeString;
-import org.apache.poi.hssf.record.UnicodeString.UnicodeRecordStats;
 import org.apache.poi.util.LittleEndianInput;
 import org.apache.poi.util.LittleEndianOutput;
 import org.apache.poi.util.StringUtil;
@@ -65,7 +63,7 @@ public final class ConstantValueParser {
 			case TYPE_NUMBER:
 				return new Double(in.readDouble());
 			case TYPE_STRING:
-				return new UnicodeString(StringUtil.readUnicodeString(in));
+				return StringUtil.readUnicodeString(in);
 			case TYPE_BOOLEAN:
 				return readBoolean(in);
 			case TYPE_ERROR_CODE:
@@ -111,10 +109,8 @@ public final class ConstantValueParser {
 		if(cls == Boolean.class || cls == Double.class || cls == ErrorConstant.class) {
 			return 8;
 		}
-		UnicodeString strVal = (UnicodeString)object;
+		String strVal = (String)object;
-		UnicodeRecordStats urs = new UnicodeRecordStats();
+		return StringUtil.getEncodedSize(strVal);
-		strVal.getRecordSize(urs);
-		return urs.recordSize;
 	}
 
 	public static void encode(LittleEndianOutput out, Object[] values) {
@@ -142,10 +138,10 @@ public final class ConstantValueParser {
 			out.writeDouble(dVal.doubleValue());
 			return;
 		}
-		if (value instanceof UnicodeString) {
+		if (value instanceof String) {
-			UnicodeString usVal = (UnicodeString) value;
+			String val = (String) value;
 			out.writeByte(TYPE_STRING);
-			StringUtil.writeUnicodeString(out, usVal.getString());
+			StringUtil.writeUnicodeString(out, val);
 			return;
 		}
 		if (value instanceof ErrorConstant) {

src/java/org/apache/poi/hssf/record/cont/ContinuableRecord.java

@@ -0,0 +1,69 @@
+/* ====================================================================
+   Licensed to the Apache Software Foundation (ASF) under one or more
+   contributor license agreements.  See the NOTICE file distributed with
+   this work for additional information regarding copyright ownership.
+   The ASF licenses this file to You under the Apache License, Version 2.0
+   (the "License"); you may not use this file except in compliance with
+   the License.  You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.
+==================================================================== */
+
+package org.apache.poi.hssf.record.cont;
+
+import org.apache.poi.hssf.record.ContinueRecord;
+import org.apache.poi.hssf.record.Record;
+import org.apache.poi.util.LittleEndianByteArrayOutputStream;
+import org.apache.poi.util.LittleEndianOutput;
+
+/**
+ * Common superclass of all records that can produce {@link ContinueRecord}s while being serialized.
+ * 
+ * @author Josh Micich
+ */
+public abstract class ContinuableRecord extends Record {
+
+	protected ContinuableRecord() {
+		// no fields to initialise 
+	}
+	/**
+	 * Serializes this record's content to the supplied data output.<br/>
+	 * The standard BIFF header (ushort sid, ushort size) has been handled by the superclass, so 
+	 * only BIFF data should be written by this method.  Simple data types can be written with the
+	 * standard {@link LittleEndianOutput} methods.  Methods from {@link ContinuableRecordOutput} 
+	 * can be used to serialize strings (with {@link ContinueRecord}s being written as required).
+	 * If necessary, implementors can explicitly start {@link ContinueRecord}s (regardless of the
+	 * amount of remaining space).
+	 * 
+	 * @param out a data output stream
+	 */
+	protected abstract void serialize(ContinuableRecordOutput out);
+
+
+	/**
+	 * @return four less than the total length of the encoded record(s) 
+	 * (in the case when no {@link ContinueRecord} is needed, this is the 
+	 *  same ushort value that gets encoded after the record sid
+	 */
+	protected final int getDataSize() {
+		ContinuableRecordOutput out = ContinuableRecordOutput.createForCountingOnly();
+		serialize(out);
+		out.terminate();
+		return out.getTotalSize() - 4;
+	}
+
+	public final int serialize(int offset, byte[] data) {
+
+		LittleEndianOutput leo = new LittleEndianByteArrayOutputStream(data, offset);
+		ContinuableRecordOutput out = new ContinuableRecordOutput(leo, getSid());
+		serialize(out);
+		out.terminate();
+		return out.getTotalSize();
+	}
+}

src/java/org/apache/poi/hssf/record/cont/ContinuableRecordOutput.java

@@ -0,0 +1,257 @@
+/* ====================================================================
+   Licensed to the Apache Software Foundation (ASF) under one or more
+   contributor license agreements.  See the NOTICE file distributed with
+   this work for additional information regarding copyright ownership.
+   The ASF licenses this file to You under the Apache License, Version 2.0
+   (the "License"); you may not use this file except in compliance with
+   the License.  You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.
+==================================================================== */
+
+package org.apache.poi.hssf.record.cont;
+
+import org.apache.poi.hssf.record.ContinueRecord;
+import org.apache.poi.util.DelayableLittleEndianOutput;
+import org.apache.poi.util.LittleEndianOutput;
+import org.apache.poi.util.StringUtil;
+
+/**
+ * An augmented {@link LittleEndianOutput} used for serialization of {@link ContinuableRecord}s.
+ * This class keeps track of how much remaining space is available in the current BIFF record and
+ * can start new {@link ContinueRecord}s as required. 
+ * 
+ * @author Josh Micich
+ */
+public final class ContinuableRecordOutput implements LittleEndianOutput {
+	
+	private final LittleEndianOutput _out;
+	private UnknownLengthRecordOutput _ulrOutput;
+	private int _totalPreviousRecordsSize;
+
+	ContinuableRecordOutput(LittleEndianOutput out, int sid) {
+		_ulrOutput = new UnknownLengthRecordOutput(out, sid);
+		_out = out;
+		_totalPreviousRecordsSize = 0;
+	}
+	
+	public static ContinuableRecordOutput createForCountingOnly() {
+		return new ContinuableRecordOutput(NOPOutput, -777); // fake sid
+	}
+
+	/**
+	 * @return total number of bytes written so far (including all BIFF headers) 
+	 */
+	public int getTotalSize() {
+		return _totalPreviousRecordsSize + _ulrOutput.getTotalSize();
+	}
+	/**
+	 * Terminates the last record (also updates its 'ushort size' field)
+	 */
+	void terminate() {
+		_ulrOutput.terminate();
+	}
+	/**
+	 * @return number of remaining bytes of space in current record
+	 */
+	public int getAvailableSpace() {
+		return _ulrOutput.getAvailableSpace();
+	}
+	
+	/**
+	 * Terminates the current record and starts a new {@link ContinueRecord} (regardless
+	 * of how much space is still available in the current record).
+	 */
+	public void writeContinue() {
+		_ulrOutput.terminate();
+		_totalPreviousRecordsSize += _ulrOutput.getTotalSize();
+		_ulrOutput = new UnknownLengthRecordOutput(_out, ContinueRecord.sid);
+	}
+	public void writeContinueIfRequired(int requiredContinuousSize) {
+		if (_ulrOutput.getAvailableSpace() < requiredContinuousSize) {
+			writeContinue();
+		}
+	}
+
+	/**
+	 * Writes the 'optionFlags' byte and encoded character data of a unicode string.  This includes:
+	 * <ul>
+	 * <li>byte optionFlags</li>
+	 * <li>encoded character data (in "ISO-8859-1" or "UTF-16LE" encoding)</li>
+	 * </ul>
+	 * 
+	 * Notes:
+	 * <ul>
+	 * <li>The value of the 'is16bitEncoded' flag is determined by the actual character data 
+	 * of <tt>text</tt></li>
+	 * <li>The string options flag is never separated (by a {@link ContinueRecord}) from the
+	 * first chunk of character data it refers to.</li>
+	 * <li>The 'ushort length' field is assumed to have been explicitly written earlier.  Hence, 
+	 * there may be an intervening {@link ContinueRecord}</li>
+	 * </ul>
+	 */
+	public void writeStringData(String text) {
+		boolean is16bitEncoded = StringUtil.hasMultibyte(text);
+		// calculate total size of the header and first encoded char
+		int keepTogetherSize = 1 + 1; // ushort len, at least one character byte
+		int optionFlags = 0x00;
+		if (is16bitEncoded) {
+			optionFlags |= 0x01;
+			keepTogetherSize += 1; // one extra byte for first char
+		}
+		writeContinueIfRequired(keepTogetherSize);
+		writeByte(optionFlags);
+		writeCharacterData(text, is16bitEncoded);
+	}
+	/**
+	 * Writes a unicode string complete with header and character data.  This includes:
+	 * <ul>
+	 * <li>ushort length</li>
+	 * <li>byte optionFlags</li>
+	 * <li>ushort numberOfRichTextRuns (optional)</li>
+	 * <li>ushort extendedDataSize (optional)</li>
+	 * <li>encoded character data (in "ISO-8859-1" or "UTF-16LE" encoding)</li>
+	 * </ul>
+	 * 
+	 * The following bits of the 'optionFlags' byte will be set as appropriate:
+	 * <table border='1'>
+	 * <tr><th>Mask</th><th>Description</th></tr>
+	 * <tr><td>0x01</td><td>is16bitEncoded</td></tr>
+	 * <tr><td>0x04</td><td>hasExtendedData</td></tr>
+	 * <tr><td>0x08</td><td>isRichText</td></tr>
+	 * </table>
+	 * Notes:
+	 * <ul> 
+	 * <li>The value of the 'is16bitEncoded' flag is determined by the actual character data 
+	 * of <tt>text</tt></li>
+	 * <li>The string header fields are never separated (by a {@link ContinueRecord}) from the
+	 * first chunk of character data (i.e. the first character is always encoded in the same
+	 * record as the string header).</li>
+	 * </ul>
+	 */
+	public void writeString(String text, int numberOfRichTextRuns, int extendedDataSize) {
+		boolean is16bitEncoded = StringUtil.hasMultibyte(text);
+		// calculate total size of the header and first encoded char
+		int keepTogetherSize = 2 + 1 + 1; // ushort len, byte optionFlags, at least one character byte
+		int optionFlags = 0x00;
+		if (is16bitEncoded) {
+			optionFlags |= 0x01;
+			keepTogetherSize += 1; // one extra byte for first char
+		}
+		if (numberOfRichTextRuns > 0) {
+			optionFlags |= 0x08;
+			keepTogetherSize += 2;
+		}
+		if (extendedDataSize > 0) {
+			optionFlags |= 0x04;
+			keepTogetherSize += 4;
+		}
+		writeContinueIfRequired(keepTogetherSize);
+		writeShort(text.length());
+		writeByte(optionFlags);
+		if (numberOfRichTextRuns > 0) {
+			writeShort(numberOfRichTextRuns);
+		}
+		if (extendedDataSize > 0) {
+			writeInt(extendedDataSize);
+		}
+		writeCharacterData(text, is16bitEncoded);
+	}
+
+
+	private void writeCharacterData(String text, boolean is16bitEncoded) {
+		int nChars = text.length();
+		int i=0;
+		if (is16bitEncoded) {
+			while(true) {
+				int nWritableChars = Math.min(nChars-i, _ulrOutput.getAvailableSpace() / 2);
+				for ( ; nWritableChars > 0; nWritableChars--) {
+					_ulrOutput.writeShort(text.charAt(i++));
+				}
+				if (i >= nChars) {
+					break;
+				}
+				writeContinue();
+				writeByte(0x01); 
+			}
+		} else {
+			while(true) {
+				int nWritableChars = Math.min(nChars-i, _ulrOutput.getAvailableSpace() / 1);
+				for ( ; nWritableChars > 0; nWritableChars--) {
+					_ulrOutput.writeByte(text.charAt(i++));
+				}
+				if (i >= nChars) {
+					break;
+				}
+				writeContinue();
+				writeByte(0x00); 
+			}
+		}
+	}
+
+	public void write(byte[] b) {
+		writeContinueIfRequired(b.length);
+		_ulrOutput.write(b);
+	}
+	public void write(byte[] b, int offset, int len) {
+		writeContinueIfRequired(len);
+		_ulrOutput.write(b, offset, len);
+	}
+	public void writeByte(int v) {
+		writeContinueIfRequired(1);
+		_ulrOutput.writeByte(v);
+	}
+	public void writeDouble(double v) {
+		writeContinueIfRequired(8);
+		_ulrOutput.writeDouble(v);
+	}
+	public void writeInt(int v) {
+		writeContinueIfRequired(4);
+		_ulrOutput.writeInt(v);
+	}
+	public void writeLong(long v) {
+		writeContinueIfRequired(8);
+		_ulrOutput.writeLong(v);
+	}
+	public void writeShort(int v) {
+		writeContinueIfRequired(2);
+		_ulrOutput.writeShort(v);
+	}
+	
+	/**
+	 * Allows optimised usage of {@link ContinuableRecordOutput} for sizing purposes only.
+	 */
+    private static final LittleEndianOutput NOPOutput = new DelayableLittleEndianOutput() {
+
+		public LittleEndianOutput createDelayedOutput(int size) {
+			return this;
+		}
+		public void write(byte[] b) {
+			// does nothing
+		}
+		public void write(byte[] b, int offset, int len) {
+			// does nothing
+		}
+		public void writeByte(int v) {
+			// does nothing
+		}
+		public void writeDouble(double v) {
+			// does nothing
+		}
+		public void writeInt(int v) {
+			// does nothing
+		}
+		public void writeLong(long v) {
+			// does nothing
+		}
+		public void writeShort(int v) {
+			// does nothing
+		}
+    };
+}

src/java/org/apache/poi/hssf/record/cont/UnknownLengthRecordOutput.java

@@ -0,0 +1,114 @@
+/* ====================================================================
+   Licensed to the Apache Software Foundation (ASF) under one or more
+   contributor license agreements.  See the NOTICE file distributed with
+   this work for additional information regarding copyright ownership.
+   The ASF licenses this file to You under the Apache License, Version 2.0
+   (the "License"); you may not use this file except in compliance with
+   the License.  You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.
+==================================================================== */
+
+package org.apache.poi.hssf.record.cont;
+
+import org.apache.poi.hssf.record.RecordInputStream;
+import org.apache.poi.util.DelayableLittleEndianOutput;
+import org.apache.poi.util.LittleEndianByteArrayOutputStream;
+import org.apache.poi.util.LittleEndianOutput;
+/**
+ * Allows the writing of BIFF records when the 'ushort size' header field is not known in advance.
+ * When the client is finished writing data, it calls {@link #terminate()}, at which point this 
+ * class updates the 'ushort size' with its final value. 
+ * 
+ * @author Josh Micich
+ */
+final class UnknownLengthRecordOutput implements LittleEndianOutput {
+	private static final int MAX_DATA_SIZE = RecordInputStream.MAX_RECORD_DATA_SIZE;
+
+	private final LittleEndianOutput _originalOut;
+	/** for writing the 'ushort size'  field once its value is known */
+	private final LittleEndianOutput _dataSizeOutput;
+	private final byte[] _byteBuffer;
+	private LittleEndianOutput _out;
+	private int _size;
+	
+	public UnknownLengthRecordOutput(LittleEndianOutput out, int sid) {
+		_originalOut = out;
+		out.writeShort(sid);
+		if (out instanceof DelayableLittleEndianOutput) {
+			// optimisation
+			DelayableLittleEndianOutput dleo = (DelayableLittleEndianOutput) out;
+			_dataSizeOutput = dleo.createDelayedOutput(2);
+			_byteBuffer = null;
+			_out = out;
+		} else {
+			// otherwise temporarily write all subsequent data to a buffer
+			_dataSizeOutput = out;
+			_byteBuffer = new byte[RecordInputStream.MAX_RECORD_DATA_SIZE];
+			_out = new LittleEndianByteArrayOutputStream(_byteBuffer, 0);
+		}
+	}
+	/**
+	 * includes 4 byte header
+	 */
+	public int getTotalSize() {
+		return 4 + _size;
+	}
+	public int getAvailableSpace() {
+		if (_out == null) {
+			throw new IllegalStateException("Record already terminated");
+		}
+		return MAX_DATA_SIZE - _size;
+	}
+	/**
+	 * Finishes writing the current record and updates 'ushort size' field.<br/>
+	 * After this method is called, only {@link #getTotalSize()} may be called.
+	 */
+	public void terminate() {
+		if (_out == null) {
+			throw new IllegalStateException("Record already terminated");
+		}
+		_dataSizeOutput.writeShort(_size);
+		if (_byteBuffer != null) {
+			_originalOut.write(_byteBuffer, 0, _size);
+			_out = null;
+			return;
+		}
+		_out = null;
+	}
+	
+	public void write(byte[] b) {
+		_out.write(b);
+		_size += b.length;
+	}
+	public void write(byte[] b, int offset, int len) {
+		_out.write(b, offset, len);
+		_size += len;
+	}
+	public void writeByte(int v) {
+		_out.writeByte(v);
+		_size += 1;
+	}
+	public void writeDouble(double v) {
+		_out.writeDouble(v);
+		_size += 8;
+	}
+	public void writeInt(int v) {
+		_out.writeInt(v);
+		_size += 4;
+	}
+	public void writeLong(long v) {
+		_out.writeLong(v);
+		_size += 8;
+	}
+	public void writeShort(int v) {
+		_out.writeShort(v);
+		_size += 2;
+	}
+}

src/java/org/apache/poi/hssf/record/formula/ArrayPtg.java

@@ -206,8 +206,8 @@ public final class ArrayPtg extends Ptg {
 		if (o == null) {
 			throw new RuntimeException("Array item cannot be null");
 		}
-		if (o instanceof UnicodeString) {
+		if (o instanceof String) {
-			return "\"" + ((UnicodeString)o).getString() + "\"";
+			return "\"" + (String)o + "\"";
 		}
 		if (o instanceof Double) {
 			return ((Double)o).toString();

src/java/org/apache/poi/hssf/usermodel/HSSFCell.java

@@ -43,7 +43,6 @@ import org.apache.poi.hssf.record.NumberRecord;
 import org.apache.poi.hssf.record.ObjRecord;
 import org.apache.poi.hssf.record.Record;
 import org.apache.poi.hssf.record.RecordBase;
-import org.apache.poi.hssf.record.StringRecord;
 import org.apache.poi.hssf.record.SubRecord;
 import org.apache.poi.hssf.record.TextObjectRecord;
 import org.apache.poi.hssf.record.UnicodeString;
@@ -336,38 +335,23 @@ public class HSSFCell implements Cell {
                 break;
 
             case CELL_TYPE_STRING :
-                LabelSSTRecord lrec = null;
+                LabelSSTRecord lrec;
 
-                if (cellType != this.cellType)
+                if (cellType == this.cellType) {
-                {
-                    lrec = new LabelSSTRecord();
-                }
-                else
-                {
                     lrec = (LabelSSTRecord) record;
-                }
+                } else {
+                    lrec = new LabelSSTRecord();
                     lrec.setColumn(col);
                     lrec.setRow(row);
                     lrec.setXFIndex(styleIndex);
-                if (setValue)
-                {
-                    if ((getStringCellValue() != null)
-                            && (!getStringCellValue().equals("")))
-                    {
-                        int sst = 0;
-
-                        UnicodeString str = getRichStringCellValue().getUnicodeString();
-//jmh                        if (encoding == ENCODING_COMPRESSED_UNICODE)
-//jmh                        {
-//                      jmh                            str.setCompressedUnicode();
-//                      jmh                        } else if (encoding == ENCODING_UTF_16)
-//                      jmh                        {
-//                      jmh                            str.setUncompressedUnicode();
-//                      jmh                        }
-                        sst = book.getWorkbook().addSSTString(str);
-                        lrec.setSSTIndex(sst);
-                        getRichStringCellValue().setUnicodeString(book.getWorkbook().getSSTString(sst));
                 }
+                if (setValue) {
+                    String str = convertCellValueToString();
+                    int sstIndex = book.getWorkbook().addSSTString(new UnicodeString(str));
+                    lrec.setSSTIndex(sstIndex);
+                    UnicodeString us = book.getWorkbook().getSSTString(sstIndex);
+                    stringValue = new HSSFRichTextString();
+                    stringValue.setUnicodeString(us);
                 }
                 record = lrec;
                 break;
@@ -782,7 +766,9 @@ public class HSSFCell implements Cell {
             case CELL_TYPE_BOOLEAN:
                 return (( BoolErrRecord ) record).getBooleanValue();
             case CELL_TYPE_STRING:
-                return Boolean.valueOf(((StringRecord)record).getString()).booleanValue();
+                int sstIndex = ((LabelSSTRecord)record).getSSTIndex();
+                String text = book.getWorkbook().getSSTString(sstIndex).getString();
+                return Boolean.valueOf(text).booleanValue();
             case CELL_TYPE_NUMERIC:
                 return ((NumberRecord)record).getValue() != 0;
 
@@ -796,6 +782,26 @@ public class HSSFCell implements Cell {
         }
         throw new RuntimeException("Unexpected cell type (" + cellType + ")");
     }
+    private String convertCellValueToString() {
+
+        switch (cellType) {
+            case CELL_TYPE_BLANK:
+                return "";
+            case CELL_TYPE_BOOLEAN:
+                return ((BoolErrRecord) record).getBooleanValue() ? "TRUE" : "FALSE";
+            case CELL_TYPE_STRING:
+                int sstIndex = ((LabelSSTRecord)record).getSSTIndex();
+                return book.getWorkbook().getSSTString(sstIndex).getString();
+            case CELL_TYPE_NUMERIC:
+                return String.valueOf(((NumberRecord)record).getValue());
+            case CELL_TYPE_ERROR:
+                   return HSSFErrorConstants.getText(((BoolErrRecord) record).getErrorValue());
+            case CELL_TYPE_FORMULA:
+                // should really evaluate, but HSSFCell can't call HSSFFormulaEvaluator
+                return "";
+        }
+        throw new RuntimeException("Unexpected cell type (" + cellType + ")");
+    }
 
     /**
      * get the value of the cell as a boolean.  For strings, numbers, and errors, we throw an exception.

src/java/org/apache/poi/util/DelayableLittleEndianOutput.java

@@ -0,0 +1,34 @@
+/* ====================================================================
+   Licensed to the Apache Software Foundation (ASF) under one or more
+   contributor license agreements.  See the NOTICE file distributed with
+   this work for additional information regarding copyright ownership.
+   The ASF licenses this file to You under the Apache License, Version 2.0
+   (the "License"); you may not use this file except in compliance with
+   the License.  You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.
+==================================================================== */
+
+package org.apache.poi.util;
+/**
+ * Implementors of this interface allow client code to 'delay' writing to a certain section of a 
+ * data output stream.<br/>
+ * A typical application is for writing BIFF records when the size is not known until well after
+ * the header has been written.  The client code can call {@link #createDelayedOutput(int)}
+ * to reserve two bytes of the output for the 'ushort size' header field.  The delayed output can
+ * be written at any stage. 
+ * 
+ * @author Josh Micich
+ */
+public interface DelayableLittleEndianOutput extends LittleEndianOutput {
+	/**
+	 * Creates an output stream intended for outputting a sequence of <tt>size</tt> bytes.
+	 */
+	LittleEndianOutput createDelayedOutput(int size);
+}

src/java/org/apache/poi/util/LittleEndianByteArrayOutputStream.java

@@ -24,7 +24,7 @@ package org.apache.poi.util;
  * 
  * @author Josh Micich
  */
-public final class LittleEndianByteArrayOutputStream implements LittleEndianOutput {
+public final class LittleEndianByteArrayOutputStream implements LittleEndianOutput, DelayableLittleEndianOutput {
 	private final byte[] _buf;
 	private final int _endIndex;
 	private int _writeIndex;
@@ -89,4 +89,10 @@ public final class LittleEndianByteArrayOutputStream implements LittleEndianOutp
 	public int getWriteIndex() {
 		return _writeIndex;
 	}
+	public LittleEndianOutput createDelayedOutput(int size) {
+		checkPosition(size);
+		LittleEndianOutput result = new LittleEndianByteArrayOutputStream(_buf, _writeIndex, _writeIndex+size);
+		_writeIndex += size;
+		return result;
+	}
 }

src/java/org/apache/poi/util/StringUtil.java

@@ -162,6 +162,15 @@ public class StringUtil {
 		}
 	}
 
+	/**
+	 * @return the number of bytes that would be written by {@link #writeUnicodeString(LittleEndianOutput, String)}
+	 */
+	public static int getEncodedSize(String value) {
+		int result = 2 + 1;
+		result += value.length() * (StringUtil.hasMultibyte(value) ? 2 : 1);
+		return result;
+	}
+
 	/**
 	 * Takes a unicode (java) string, and returns it as 8 bit data (in ISO-8859-1
 	 * codepage).

src/scratchpad/src/org/apache/poi/hslf/model/Picture.java

@@ -196,10 +196,14 @@ public class Picture extends SimpleShape {
         Document doc = ppt.getDocumentRecord();
         EscherContainerRecord dggContainer = doc.getPPDrawingGroup().getDggContainer();
         EscherContainerRecord bstore = (EscherContainerRecord)Shape.getEscherChild(dggContainer, EscherContainerRecord.BSTORE_CONTAINER);
-
+        if(bstore == null) {
+            logger.log(POILogger.DEBUG, "EscherContainerRecord.BSTORE_CONTAINER was not found ");
+            return null;
+        }
         List lst = bstore.getChildRecords();
         int idx = getPictureIndex();
         if (idx == 0){
+            logger.log(POILogger.DEBUG, "picture index was not found, returning ");
             return null;
         } else {
             return (EscherBSERecord)lst.get(idx-1);
@@ -263,7 +267,7 @@ public class Picture extends SimpleShape {
         ShapePainter.paint(this, graphics);
 
         PictureData data = getPictureData();
-        data.draw(graphics, this);
+        if(data != null) data.draw(graphics, this);
 
         graphics.setTransform(at);
     }

src/scratchpad/src/org/apache/poi/hslf/model/TableCell.java

@@ -56,8 +56,8 @@ public class TableCell extends TextBox {
         super(parent);
 
         setShapeType(ShapeTypes.Rectangle);
-        _txtrun.setRunType(TextHeaderAtom.HALF_BODY_TYPE);
+        //_txtrun.setRunType(TextHeaderAtom.HALF_BODY_TYPE);
-        _txtrun.getRichTextRuns()[0].setFlag(false, 0, false);
+        //_txtrun.getRichTextRuns()[0].setFlag(false, 0, false);
     }
 
     protected EscherContainerRecord createSpContainer(boolean isChild){

src/scratchpad/testcases/org/apache/poi/hslf/model/TestPicture.java

@@ -20,9 +20,12 @@ import junit.framework.*;
 
 import java.io.FileOutputStream;
 import java.io.File;
+import java.io.IOException;
 import java.awt.*;
+import java.awt.image.BufferedImage;
 
 import org.apache.poi.hslf.usermodel.SlideShow;
+import org.apache.poi.hslf.usermodel.PictureData;
 import org.apache.poi.hslf.HSLFSlideShow;
 import org.apache.poi.ddf.EscherBSERecord;
 
@@ -70,4 +73,24 @@ public class TestPicture extends TestCase {
 
     }
 
+    /**
+     * Picture#getEscherBSERecord threw NullPointerException if EscherContainerRecord.BSTORE_CONTAINER
+     * was not found. The correct behaviour is to return null.
+     */
+    public void test46122() throws IOException {
+        SlideShow ppt = new SlideShow();
+        Slide slide = ppt.createSlide();
+
+        Picture pict = new Picture(-1); //index to non-existing picture data
+        pict.setSheet(slide);
+        PictureData data = pict.getPictureData();
+        assertNull(data);
+
+        BufferedImage img = new BufferedImage(100, 100, BufferedImage.TYPE_INT_RGB);
+        Graphics2D graphics = img.createGraphics();
+        pict.draw(graphics);
+
+        assertTrue("no errors rendering Picture with null data", true);
+    }
+
 }

src/scratchpad/testcases/org/apache/poi/hslf/model/TestTable.java

@@ -24,6 +24,7 @@ import java.awt.geom.Rectangle2D;
 
 import org.apache.poi.hslf.usermodel.SlideShow;
 import org.apache.poi.hslf.HSLFSlideShow;
+import org.apache.poi.hslf.record.TextHeaderAtom;
 
 /**
  * Test <code>Table</code> object.
@@ -43,6 +44,10 @@ public class TestTable extends TestCase {
         Table tbl = new Table(2, 5);
         slide.addShape(tbl);
 
+        TableCell cell = tbl.getCell(0, 0);
+        //table cells have type=TextHeaderAtom.OTHER_TYPE, see bug #46033
+        assertEquals(TextHeaderAtom.OTHER_TYPE, cell.getTextRun().getRunType());
+
         assertTrue(slide.getShapes()[0] instanceof Table);
         Table tbl2 = (Table)slide.getShapes()[0];
         assertEquals(tbl.getNumberOfColumns(), tbl2.getNumberOfColumns());

src/testcases/org/apache/poi/hssf/record/TestRecordFactory.java

@@ -48,7 +48,6 @@ public final class TestRecordFactory extends TestCase {
         byte[]   data    = {
             0, 6, 5, 0, -2, 28, -51, 7, -55, 64, 0, 0, 6, 1, 0, 0
         };
-        short    size    = 16;
         Record[] record  = RecordFactory.createRecord(TestcaseRecordInputStream.create(recType, data));
 
         assertEquals(BOFRecord.class.getName(),
@@ -64,7 +63,6 @@ public final class TestRecordFactory extends TestCase {
         assertEquals(5, bofRecord.getType());
         assertEquals(1536, bofRecord.getVersion());
         recType = MMSRecord.sid;
-        size    = 2;
         data    = new byte[]
         {
             0, 0
@@ -93,7 +91,6 @@ public final class TestRecordFactory extends TestCase {
         byte[]   data    = {
             0, 0, 0, 0, 21, 0, 0, 0, 0, 0
         };
-        short    size    = 10;
         Record[] record  = RecordFactory.createRecord(TestcaseRecordInputStream.create(recType, data));
 
         assertEquals(NumberRecord.class.getName(),
@@ -154,34 +151,34 @@ public final class TestRecordFactory extends TestCase {
      */
     public void testMixedContinue() throws Exception {
         /**
-         *  Taken from a real test sample file 39512.xls. See Bug 39512 for details.
+         *  Adapted from a real test sample file 39512.xls (Offset 0x4854). 
+         *  See Bug 39512 for details.
          */
         String dump =
                 //OBJ
-                "5D, 00, 48, 00, 15, 00, 12, 00, 0C, 00, 3C, 00, 11, 00, A0, 2E, 03, 01, CC, 42, " +
+                "5D 00 48 00 15 00 12 00 0C 00 3C 00 11 00 A0 2E 03 01 CC 42 " +
-                "CF, 00, 00, 00, 00, 00, 0A, 00, 0C, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, " +
+                "CF 00 00 00 00 00 0A 00 0C 00 00 00 00 00 00 00 00 00 00 00 " +
-                "03, 00, 0B, 00, 06, 00, 28, 01, 03, 01, 00, 00, 12, 00, 08, 00, 00, 00, 00, 00, " +
+                "03 00 0B 00 06 00 28 01 03 01 00 00 12 00 08 00 00 00 00 00 " +
-                "00, 00, 03, 00, 11, 00, 04, 00, 3D, 00, 00, 00, 00, 00, 00, 00, " +
+                "00 00 03 00 11 00 04 00 3D 00 00 00 00 00 00 00 " +
                  //MSODRAWING
-                "EC, 00, 08, 00, 00, 00, 0D, F0, 00, 00, 00, 00, " +
+                "EC 00 08 00 00 00 0D F0 00 00 00 00 " +
-                //TXO
+                //TXO (and 2 trailing CONTINUE records)
-                "B6, 01, 12, 00, 22, 02, 00, 00, 00, 00, 00, 00, 00, 00, 10, 00, 10, 00, 00, 00, " +
+                "B6 01 12 00 22 02 00 00 00 00 00 00 00 00 10 00 10 00 00 00 00 00 " +
-                "00, 00, 3C, 00, 21, 00, 01, 4F, 00, 70, 00, 74, 00, 69, 00, 6F, 00, 6E, 00, 20, " +
+                "3C 00 11 00 00 4F 70 74 69 6F 6E 20 42 75 74 74 6F 6E 20 33 39 " +
-                "00, 42, 00, 75, 00, 74, 00, 74, 00, 6F, 00, 6E, 00, 20, 00, 33, 00, 39, 00, 3C, " +
+                "3C 00 10 00 00 00 05 00 00 00 00 00 10 00 00 00 00 00 00 00 " +
-                "00, 10, 00, 00, 00, 05, 00, 00, 00, 00, 00, 10, 00, 00, 00, 00, 00, 00, 00, " +
+                // another CONTINUE
-                //CONTINUE
+                "3C 00 7E 00 0F 00 04 F0 7E 00 00 00 92 0C 0A F0 08 00 00 00 " +
-                "3C, 00, 7E, 00, 0F, 00, 04, F0, 7E, 00, 00, 00, 92, 0C, 0A, F0, 08, 00, 00, 00, " +
+                "3D 04 00 00 00 0A 00 00 A3 00 0B F0 3C 00 00 00 7F 00 00 01 " +
-                "3D, 04, 00, 00, 00, 0A, 00, 00, A3, 00, 0B, F0, 3C, 00, 00, 00, 7F, 00, 00, 01, " +
+                "00 01 80 00 8C 01 03 01 85 00 01 00 00 00 8B 00 02 00 00 00 " +
-                "00, 01, 80, 00, 8C, 01, 03, 01, 85, 00, 01, 00, 00, 00, 8B, 00, 02, 00, 00, 00, " +
+                "BF 00 08 00 1A 00 7F 01 29 00 29 00 81 01 41 00 00 08 BF 01 " +
-                "BF, 00, 08, 00, 1A, 00, 7F, 01, 29, 00, 29, 00, 81, 01, 41, 00, 00, 08, BF, 01, " +
+                "00 00 10 00 C0 01 40 00 00 08 FF 01 00 00 08 00 00 00 10 F0 " +
-                "00, 00, 10, 00, C0, 01, 40, 00, 00, 08, FF, 01, 00, 00, 08, 00, 00, 00, 10, F0, " +
+                "12 00 00 00 02 00 02 00 A0 03 18 00 B5 00 04 00 30 02 1A 00 " +
-                "12, 00, 00, 00, 02, 00, 02, 00, A0, 03, 18, 00, B5, 00, 04, 00, 30, 02, 1A, 00, " +
+                "00 00 00 00 11 F0 00 00 00 00 " +
-                "00, 00, 00, 00, 11, F0, 00, 00, 00, 00, " +
                 //OBJ
-                "5D, 00, 48, 00, 15, 00, 12, 00, 0C, 00, 3D, 00, 11, 00, 8C, 01, 03, 01, C8, 59, CF, 00, 00, " +
+                "5D 00 48 00 15 00 12 00 0C 00 3D 00 11 00 8C 01 03 01 C8 59 CF 00 00 " +
-                "00, 00, 00, 0A, 00, 0C, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 03, 00, 0B, 00, 06, 00, " +
+                "00 00 00 0A 00 0C 00 00 00 00 00 00 00 00 00 00 00 03 00 0B 00 06 00 " +
-                "7C, 16, 03, 01, 00, 00, 12, 00, 08, 00, 00, 00, 00, 00, 00, 00, 03, 00, 11, 00, 04, 00, 01, " +
+                "7C 16 03 01 00 00 12 00 08 00 00 00 00 00 00 00 03 00 11 00 04 00 01 " +
-                "00, 00, 00, 00, 00, 00, 00";
+                "00 00 00 00 00 00 00";
         byte[] data = HexRead.readFromString(dump);
 
         List records = RecordFactory.createRecords(new ByteArrayInputStream(data));

src/testcases/org/apache/poi/hssf/record/TestSSTRecordSizeCalculator.java

@@ -1,4 +1,3 @@
-
 /* ====================================================================
    Licensed to the Apache Software Foundation (ASF) under one or more
    contributor license agreements.  See the NOTICE file distributed with
@@ -16,11 +15,11 @@
    limitations under the License.
 ==================================================================== */
 
-
 package org.apache.poi.hssf.record;
 
 import junit.framework.TestCase;
 
+import org.apache.poi.hssf.record.cont.ContinuableRecordOutput;
 import org.apache.poi.util.IntMapper;
 
 /**
@@ -28,120 +27,93 @@ import org.apache.poi.util.IntMapper;
  * 
  * @author Glen Stampoultzis (glens at apache.org)
  */
-public class TestSSTRecordSizeCalculator
+public final class TestSSTRecordSizeCalculator extends TestCase {
-        extends TestCase
-{
 	private static final String SMALL_STRING = "Small string";
 	private static final int COMPRESSED_PLAIN_STRING_OVERHEAD = 3;
-//    private List recordLengths;
-    private IntMapper strings;
 	private static final int OPTION_FIELD_SIZE = 1;
 	
-    public TestSSTRecordSizeCalculator( String s )
+	private final IntMapper strings = new IntMapper();
-    {
+
-        super( s );
+	
+	private void confirmSize(int expectedSize) {
+		ContinuableRecordOutput cro = ContinuableRecordOutput.createForCountingOnly();
+		SSTSerializer ss = new SSTSerializer(strings, 0, 0);
+		ss.serialize(cro);
+		assertEquals(expectedSize, cro.getTotalSize());
 	}
 
-    public void testBasic()
+	public void testBasic() {
-            throws Exception
-    {
 		strings.add(makeUnicodeString(SMALL_STRING));
-        SSTRecordSizeCalculator calculator = new SSTRecordSizeCalculator(strings);
+		confirmSize(SSTRecord.SST_RECORD_OVERHEAD
-        assertEquals(SSTRecord.SST_RECORD_OVERHEAD + COMPRESSED_PLAIN_STRING_OVERHEAD + SMALL_STRING.length(),
+				+ COMPRESSED_PLAIN_STRING_OVERHEAD
-                calculator.getRecordSize());
+				+ SMALL_STRING.length());
 	}
 
-    public void testBigStringAcrossUnicode()
+	public void testBigStringAcrossUnicode() {
-            throws Exception
+		int bigString = SSTRecord.MAX_DATA_SPACE + 100;
-    {
-        String bigString = new String(new char[SSTRecord.MAX_DATA_SPACE + 100]);
 		strings.add(makeUnicodeString(bigString));
-        SSTRecordSizeCalculator calculator = new SSTRecordSizeCalculator(strings);
+		confirmSize(SSTRecord.SST_RECORD_OVERHEAD
-        assertEquals(SSTRecord.SST_RECORD_OVERHEAD
 				+ COMPRESSED_PLAIN_STRING_OVERHEAD
 				+ SSTRecord.MAX_DATA_SPACE
 				+ SSTRecord.STD_RECORD_OVERHEAD
 				+ OPTION_FIELD_SIZE
-                + 100,
+				+ 100);
-                calculator.getRecordSize());
 	}
 
-    public void testPerfectFit()
+	public void testPerfectFit() {
-            throws Exception
+		int perfectFit = SSTRecord.MAX_DATA_SPACE - COMPRESSED_PLAIN_STRING_OVERHEAD;
-    {
-        String perfectFit = new String(new char[SSTRecord.MAX_DATA_SPACE - COMPRESSED_PLAIN_STRING_OVERHEAD]);
 		strings.add(makeUnicodeString(perfectFit));
-        SSTRecordSizeCalculator calculator = new SSTRecordSizeCalculator(strings);
+		confirmSize(SSTRecord.SST_RECORD_OVERHEAD
-        assertEquals(SSTRecord.SST_RECORD_OVERHEAD
 				+ COMPRESSED_PLAIN_STRING_OVERHEAD
-                + perfectFit.length(),
+				+ perfectFit);
-                calculator.getRecordSize());
 	}
 
-    public void testJustOversized()
+	public void testJustOversized() {
-            throws Exception
+		int tooBig = SSTRecord.MAX_DATA_SPACE - COMPRESSED_PLAIN_STRING_OVERHEAD + 1;
-    {
-        String tooBig = new String(new char[SSTRecord.MAX_DATA_SPACE - COMPRESSED_PLAIN_STRING_OVERHEAD + 1]);
 		strings.add(makeUnicodeString(tooBig));
-        SSTRecordSizeCalculator calculator = new SSTRecordSizeCalculator(strings);
+		confirmSize(SSTRecord.SST_RECORD_OVERHEAD
-        assertEquals(SSTRecord.SST_RECORD_OVERHEAD
 				+ COMPRESSED_PLAIN_STRING_OVERHEAD
-                + tooBig.length() - 1
+				+ tooBig - 1
 				// continue record
 				+ SSTRecord.STD_RECORD_OVERHEAD
-                + OPTION_FIELD_SIZE
+				+ OPTION_FIELD_SIZE + 1);
-                + 1,
-                calculator.getRecordSize());
 
 	}
 
-    public void testSecondStringStartsOnNewContinuation()
+	public void testSecondStringStartsOnNewContinuation() {
-            throws Exception
+		int perfectFit = SSTRecord.MAX_DATA_SPACE - COMPRESSED_PLAIN_STRING_OVERHEAD;
-    {
-        String perfectFit = new String(new char[SSTRecord.MAX_DATA_SPACE - COMPRESSED_PLAIN_STRING_OVERHEAD]);
 		strings.add(makeUnicodeString(perfectFit));
 		strings.add(makeUnicodeString(SMALL_STRING));
-        SSTRecordSizeCalculator calculator = new SSTRecordSizeCalculator(strings);
+		confirmSize(SSTRecord.SST_RECORD_OVERHEAD
-        assertEquals(SSTRecord.SST_RECORD_OVERHEAD
 				+ SSTRecord.MAX_DATA_SPACE
 				// second string
 				+ SSTRecord.STD_RECORD_OVERHEAD
 				+ COMPRESSED_PLAIN_STRING_OVERHEAD
-                + SMALL_STRING.length(),
+				+ SMALL_STRING.length());
-                calculator.getRecordSize());
 	}
 
-    public void testHeaderCrossesNormalContinuePoint()
+	public void testHeaderCrossesNormalContinuePoint() {
-            throws Exception
+		int almostPerfectFit = SSTRecord.MAX_DATA_SPACE - COMPRESSED_PLAIN_STRING_OVERHEAD - 2;
-    {
-        String almostPerfectFit = new String(new char[SSTRecord.MAX_DATA_SPACE - COMPRESSED_PLAIN_STRING_OVERHEAD - 2]);
 		strings.add(makeUnicodeString(almostPerfectFit));
 		String oneCharString = new String(new char[1]);
 		strings.add(makeUnicodeString(oneCharString));
-        SSTRecordSizeCalculator calculator = new SSTRecordSizeCalculator(strings);
+		confirmSize(SSTRecord.SST_RECORD_OVERHEAD
-        assertEquals(SSTRecord.SST_RECORD_OVERHEAD
 				+ COMPRESSED_PLAIN_STRING_OVERHEAD
-                + almostPerfectFit.length()
+				+ almostPerfectFit
 				// second string
 				+ SSTRecord.STD_RECORD_OVERHEAD
 				+ COMPRESSED_PLAIN_STRING_OVERHEAD
-                + oneCharString.length(),
+				+ oneCharString.length());
-                calculator.getRecordSize());
 
 	}
-
+	private static UnicodeString makeUnicodeString(int size) {
-
+		String s = new String(new char[size]);
-    public void setUp()
+		return makeUnicodeString(s);
-    {
-        strings = new IntMapper();
 	}
 
-
+	private static UnicodeString makeUnicodeString(String s) {
-    private UnicodeString makeUnicodeString( String s )
-    {
 		UnicodeString st = new UnicodeString(s);
 		st.setOptionFlags((byte) 0);
 		return st;
 	}
-
 }

src/testcases/org/apache/poi/hssf/record/TestStringRecord.java

@@ -18,6 +18,12 @@
 package org.apache.poi.hssf.record;
 
 
+import org.apache.poi.util.HexRead;
+import org.apache.poi.util.LittleEndian;
+import org.apache.poi.util.LittleEndianByteArrayInputStream;
+import org.apache.poi.util.LittleEndianInput;
+
+import junit.framework.AssertionFailedError;
 import junit.framework.TestCase;
 
 /**
@@ -28,12 +34,12 @@ import junit.framework.TestCase;
  * @author Glen Stampoultzis (glens at apache.org)
  */
 public final class TestStringRecord extends TestCase {
-    byte[] data = new byte[] {
+	private static final byte[] data = HexRead.readFromString(
-        (byte)0x0B,(byte)0x00,   // length
+			"0B 00 " + // length
-        (byte)0x00,              // option
+			"00 " +    // option
 			// string
-        (byte)0x46,(byte)0x61,(byte)0x68,(byte)0x72,(byte)0x7A,(byte)0x65,(byte)0x75,(byte)0x67,(byte)0x74,(byte)0x79,(byte)0x70
+			"46 61 68 72 7A 65 75 67 74 79 70"
-    };
+	);
 
 	public void testLoad() {
 
@@ -43,8 +49,7 @@ public final class TestStringRecord extends TestCase {
 		assertEquals( 18, record.getRecordSize() );
 	}
 
-    public void testStore()
+	public void testStore() {
-    {
 		StringRecord record = new StringRecord();
 		record.setString("Fahrzeugtyp");
 
@@ -53,4 +58,42 @@ public final class TestStringRecord extends TestCase {
 		for (int i = 0; i < data.length; i++)
 			assertEquals("At offset " + i, data[i], recordBytes[i+4]);
 	}
+    
+	public void testContinue() {
+		int MAX_BIFF_DATA = RecordInputStream.MAX_RECORD_DATA_SIZE;
+		int TEXT_LEN = MAX_BIFF_DATA + 1000; // deliberately over-size
+		String textChunk = "ABCDEGGHIJKLMNOP"; // 16 chars
+		StringBuffer sb = new StringBuffer(16384);
+		while (sb.length() < TEXT_LEN) {
+			sb.append(textChunk);
+		}
+		sb.setLength(TEXT_LEN);
+
+		StringRecord sr = new StringRecord();
+		sr.setString(sb.toString());
+		byte[] ser = sr.serialize();
+		assertEquals(StringRecord.sid, LittleEndian.getUShort(ser, 0));
+		if (LittleEndian.getUShort(ser, 2) > MAX_BIFF_DATA) {
+			throw new AssertionFailedError(
+					"StringRecord should have been split with a continue record");
+		}
+		// Confirm expected size of first record, and ushort strLen.
+		assertEquals(MAX_BIFF_DATA, LittleEndian.getUShort(ser, 2));
+		assertEquals(TEXT_LEN, LittleEndian.getUShort(ser, 4));
+
+		// Confirm first few bytes of ContinueRecord
+		LittleEndianInput crIn = new LittleEndianByteArrayInputStream(ser, (MAX_BIFF_DATA + 4));
+		int nCharsInFirstRec = MAX_BIFF_DATA - (2 + 1); // strLen, optionFlags
+		int nCharsInSecondRec = TEXT_LEN - nCharsInFirstRec;
+		assertEquals(ContinueRecord.sid, crIn.readUShort());
+		assertEquals(1 + nCharsInSecondRec, crIn.readUShort());
+		assertEquals(0, crIn.readUByte());
+		assertEquals('N', crIn.readUByte());
+		assertEquals('O', crIn.readUByte());
+
+		// re-read and make sure string value is the same
+		RecordInputStream in = TestcaseRecordInputStream.create(ser);
+		StringRecord sr2 = new StringRecord(in);
+		assertEquals(sb.toString(), sr2.getString());
+	}
 }

src/testcases/org/apache/poi/hssf/record/TestSupBookRecord.java

@@ -69,10 +69,10 @@ public final class TestSupBookRecord extends TestCase {
         assertEquals( 34, record.getRecordSize() );  //sid+size+data
         
         assertEquals("testURL", record.getURL());
-        UnicodeString[] sheetNames = record.getSheetNames();
+        String[] sheetNames = record.getSheetNames();
         assertEquals(2, sheetNames.length);
-        assertEquals("Sheet1", sheetNames[0].getString());
+        assertEquals("Sheet1", sheetNames[0]);
-        assertEquals("Sheet2", sheetNames[1].getString());
+        assertEquals("Sheet2", sheetNames[1]);
     }
     
     /**
@@ -97,11 +97,8 @@ public final class TestSupBookRecord extends TestCase {
     }   
     
     public void testStoreER() {
-        UnicodeString url = new UnicodeString("testURL");
+        String url = "testURL";
-        UnicodeString[] sheetNames = {
+        String[] sheetNames = { "Sheet1", "Sheet2", };
-                new UnicodeString("Sheet1"),
-                new UnicodeString("Sheet2"),
-        };
         SupBookRecord record = SupBookRecord.createExternalReferences(url, sheetNames);
 
         TestcaseRecordInputStream.confirmRecordEncoding(0x01AE, dataER, record.serialize());

src/testcases/org/apache/poi/hssf/record/TestTextObjectBaseRecord.java

@@ -44,9 +44,9 @@ public final class TestTextObjectBaseRecord extends TestCase {
         "00 00" +
         "00 00 " +
         "3C 00 " + // ContinueRecord.sid
-        "05 00 " + // size 5
+        "03 00 " + // size 3
-        "01 " + // unicode uncompressed
+        "00 " + // unicode compressed
-        "41 00 42 00 " + // 'AB'
+        "41 42 " + // 'AB'
         "3C 00 " + // ContinueRecord.sid
         "10 00 " + // size 16 
         "00 00 18 00 00 00 00 00 " +
@@ -63,7 +63,7 @@ public final class TestTextObjectBaseRecord extends TestCase {
         assertEquals(true, record.isTextLocked());
         assertEquals(TextObjectRecord.TEXT_ORIENTATION_ROT_RIGHT, record.getTextOrientation());
 
-        assertEquals(51, record.getRecordSize() );
+        assertEquals(49, record.getRecordSize() );
     }
 
     public void testStore()

src/testcases/org/apache/poi/hssf/record/TestTextObjectRecord.java

@@ -41,10 +41,8 @@ public final class TestTextObjectRecord extends TestCase {
         "12 02 00 00 00 00 00 00" +
         "00 00 0D 00 08 00    00 00" +
         "00 00 " +
-    	"3C 00 1B 00 " +
+        "3C 00 0E 00 " +
-    	"01 48 00 65 00 6C 00 6C 00 6F 00 " +
+        "00 48 65 6C 6C 6F 2C 20 57 6F 72 6C 64 21 " +
-    	"2C 00 20 00 57 00 6F 00 72 00 6C " +
-    	"00 64 00 21 00 " + 
         "3C 00 08 " +
         "00 0D 00 00 00 00 00 00 00"
     );
@@ -169,9 +167,9 @@ public final class TestTextObjectRecord extends TestCase {
             "13 " +             // ??
 
             "3C 00 " + // ContinueRecord.sid
-    	    "05 00 " + // size 5
+            "03 00 " + // size 3
-    	    "01 " + // unicode uncompressed
+            "00 " + // unicode compressed
-    	    "41 00 42 00 " + // 'AB'
+            "41 42 " + // 'AB'
             "3C 00 " + // ContinueRecord.sid
             "10 00 " + // size 16
             "00 00 18 00 00 00 00 00 " +

src/testcases/org/apache/poi/hssf/record/TestUnicodeString.java

@@ -15,115 +15,123 @@
    limitations under the License.
 ==================================================================== */
 
-
 package org.apache.poi.hssf.record;
 
-import org.apache.poi.util.HexRead;
-
 import junit.framework.TestCase;
 
+import org.apache.poi.hssf.record.cont.ContinuableRecordOutput;
+
 /**
- * Tests that records size calculates correctly.
+ * Tests that {@link UnicodeString} record size calculates correctly.  The record size
+ * is used when serializing {@link SSTRecord}s.
  *
  * @author Jason Height (jheight at apache.org)
  */
 public final class TestUnicodeString extends TestCase {
+    private static final int MAX_DATA_SIZE = RecordInputStream.MAX_RECORD_DATA_SIZE;
 
+    /** a 4 character string requiring 16 bit encoding */
+    private static final String STR_16_BIT = "A\u591A\u8A00\u8A9E";
+
+    private static void confirmSize(int expectedSize, UnicodeString s) {
+        confirmSize(expectedSize, s, 0);
+    }
+    /**
+     * Note - a value of zero for <tt>amountUsedInCurrentRecord</tt> would only ever occur just
+     * after a {@link ContinueRecord} had been started.  In the initial {@link SSTRecord} this 
+     * value starts at 8 (for the first {@link UnicodeString} written).  In general, it can be
+     * any value between 0 and {@link #MAX_DATA_SIZE}
+     */
+    private static void confirmSize(int expectedSize, UnicodeString s, int amountUsedInCurrentRecord) {
+        ContinuableRecordOutput out = ContinuableRecordOutput.createForCountingOnly();
+        out.writeContinue();
+        for(int i=amountUsedInCurrentRecord; i>0; i--) {
+            out.writeByte(0);
+        }
+        int size0 = out.getTotalSize();
+        s.serialize(out);
+        int size1 = out.getTotalSize();
+        int actualSize = size1-size0;
+        assertEquals(expectedSize, actualSize);
+    }
 
     public void testSmallStringSize() {
         //Test a basic string
         UnicodeString s = makeUnicodeString("Test");
-        UnicodeString.UnicodeRecordStats stats = new UnicodeString.UnicodeRecordStats();
+        confirmSize(7, s);
-        s.getRecordSize(stats);
-        assertEquals(7, stats.recordSize);
 
         //Test a small string that is uncompressed
+        s = makeUnicodeString(STR_16_BIT);
         s.setOptionFlags((byte)0x01);
-        stats = new UnicodeString.UnicodeRecordStats();
+        confirmSize(11, s);
-        s.getRecordSize(stats);
-        assertEquals(11, stats.recordSize);
 
         //Test a compressed small string that has rich text formatting
+        s.setString("Test");
         s.setOptionFlags((byte)0x8);
         UnicodeString.FormatRun r = new UnicodeString.FormatRun((short)0,(short)1);
         s.addFormatRun(r);
         UnicodeString.FormatRun r2 = new UnicodeString.FormatRun((short)2,(short)2);
         s.addFormatRun(r2);
-        stats = new UnicodeString.UnicodeRecordStats();
+        confirmSize(17, s);
-        s.getRecordSize(stats);
-        assertEquals(17, stats.recordSize);
 
         //Test a uncompressed small string that has rich text formatting
+        s.setString(STR_16_BIT);
         s.setOptionFlags((byte)0x9);
-        stats = new UnicodeString.UnicodeRecordStats();
+        confirmSize(21, s);
-        s.getRecordSize(stats);
-        assertEquals(21, stats.recordSize);
 
         //Test a compressed small string that has rich text and extended text
+        s.setString("Test");
         s.setOptionFlags((byte)0xC);
         s.setExtendedRst(new byte[]{(byte)0x1,(byte)0x2,(byte)0x3,(byte)0x4,(byte)0x5});
-        stats = new UnicodeString.UnicodeRecordStats();
+        confirmSize(26, s);
-        s.getRecordSize(stats);
-        assertEquals(26, stats.recordSize);
 
         //Test a uncompressed small string that has rich text and extended text
+        s.setString(STR_16_BIT);
         s.setOptionFlags((byte)0xD);
-        stats = new UnicodeString.UnicodeRecordStats();
+        confirmSize(30, s);
-        s.getRecordSize(stats);
-        assertEquals(30, stats.recordSize);
     }
 
     public void testPerfectStringSize() {
       //Test a basic string
-      UnicodeString s = makeUnicodeString(SSTRecord.MAX_RECORD_SIZE-2-1);
+      UnicodeString s = makeUnicodeString(MAX_DATA_SIZE-2-1);
-      UnicodeString.UnicodeRecordStats stats = new UnicodeString.UnicodeRecordStats();
+      confirmSize(MAX_DATA_SIZE, s);
-      s.getRecordSize(stats);
-      assertEquals(SSTRecord.MAX_RECORD_SIZE, stats.recordSize);
 
       //Test an uncompressed string
       //Note that we can only ever get to a maximim size of 8227 since an uncompressed
       //string is writing double bytes.
-      s = makeUnicodeString((SSTRecord.MAX_RECORD_SIZE-2-1)/2);
+      s = makeUnicodeString((MAX_DATA_SIZE-2-1)/2, true);
       s.setOptionFlags((byte)0x1);
-      stats = new UnicodeString.UnicodeRecordStats();
+      confirmSize(MAX_DATA_SIZE-1, s);
-      s.getRecordSize(stats);
-      assertEquals(SSTRecord.MAX_RECORD_SIZE-1, stats.recordSize);
     }
 
     public void testPerfectRichStringSize() {
       //Test a rich text string
-      UnicodeString s = makeUnicodeString(SSTRecord.MAX_RECORD_SIZE-2-1-8-2);
+      UnicodeString s = makeUnicodeString(MAX_DATA_SIZE-2-1-8-2);
       s.addFormatRun(new UnicodeString.FormatRun((short)1,(short)0));
       s.addFormatRun(new UnicodeString.FormatRun((short)2,(short)1));
-      UnicodeString.UnicodeRecordStats stats = new UnicodeString.UnicodeRecordStats();
       s.setOptionFlags((byte)0x8);
-      s.getRecordSize(stats);
+      confirmSize(MAX_DATA_SIZE, s);
-      assertEquals(SSTRecord.MAX_RECORD_SIZE, stats.recordSize);
 
       //Test an uncompressed rich text string
-      //Note that we can only ever get to a maximim size of 8227 since an uncompressed
+      //Note that we can only ever get to a maximum size of 8227 since an uncompressed
       //string is writing double bytes.
-      s = makeUnicodeString((SSTRecord.MAX_RECORD_SIZE-2-1-8-2)/2);
+      s = makeUnicodeString((MAX_DATA_SIZE-2-1-8-2)/2, true);
       s.addFormatRun(new UnicodeString.FormatRun((short)1,(short)0));
       s.addFormatRun(new UnicodeString.FormatRun((short)2,(short)1));
       s.setOptionFlags((byte)0x9);
-      stats = new UnicodeString.UnicodeRecordStats();
+      confirmSize(MAX_DATA_SIZE-1, s);
-      s.getRecordSize(stats);
-      assertEquals(SSTRecord.MAX_RECORD_SIZE-1, stats.recordSize);
     }
 
     public void testContinuedStringSize() {
       //Test a basic string
-      UnicodeString s = makeUnicodeString(SSTRecord.MAX_RECORD_SIZE-2-1+20);
+      UnicodeString s = makeUnicodeString(MAX_DATA_SIZE-2-1+20);
-      UnicodeString.UnicodeRecordStats stats = new UnicodeString.UnicodeRecordStats();
+      confirmSize(MAX_DATA_SIZE+4+1+20, s);
-      s.getRecordSize(stats);
-      assertEquals(SSTRecord.MAX_RECORD_SIZE+4+1+20, stats.recordSize);
     }
 
     /** Tests that a string size calculation that fits neatly in two records, the second being a continue*/
     public void testPerfectContinuedStringSize() {
       //Test a basic string
-      int strSize = SSTRecord.MAX_RECORD_SIZE*2;
+      int strSize = MAX_DATA_SIZE*2;
       //String overhead
       strSize -= 3;
       //Continue Record overhead
@@ -131,25 +139,29 @@ public final class TestUnicodeString extends TestCase {
       //Continue Record additional byte overhead
       strSize -= 1;
       UnicodeString s = makeUnicodeString(strSize);
-      UnicodeString.UnicodeRecordStats stats = new UnicodeString.UnicodeRecordStats();
+      confirmSize(MAX_DATA_SIZE*2, s);
-      s.getRecordSize(stats);
-      assertEquals(SSTRecord.MAX_RECORD_SIZE*2, stats.recordSize);
     }
 
 
-
+    private static UnicodeString makeUnicodeString(String s) {
-
-    private static UnicodeString makeUnicodeString( String s )
-    {
       UnicodeString st = new UnicodeString(s);
       st.setOptionFlags((byte)0);
       return st;
     }
 
     private static UnicodeString makeUnicodeString(int numChars) {
+        return makeUnicodeString(numChars, false);
+    }
+    /**
+     * @param is16Bit if <code>true</code> the created string will have characters > 0x00FF
+     * @return a string of the specified number of characters
+     */
+    private static UnicodeString makeUnicodeString(int numChars, boolean is16Bit) {
       StringBuffer b = new StringBuffer(numChars);
+      int charBase = is16Bit ? 0x8A00 : 'A';
       for (int i=0;i<numChars;i++) {
-        b.append(i%10);
+        char ch = (char) ((i%16)+charBase);
+        b.append(ch);
       }
       return makeUnicodeString(b.toString());
     }

src/testcases/org/apache/poi/hssf/record/constant/TestConstantValueParser.java

@@ -22,7 +22,6 @@ import java.util.Arrays;
 import junit.framework.TestCase;
 
 import org.apache.poi.hssf.record.TestcaseRecordInputStream;
-import org.apache.poi.hssf.record.UnicodeString;
 import org.apache.poi.hssf.usermodel.HSSFErrorConstants;
 import org.apache.poi.util.HexRead;
 import org.apache.poi.util.LittleEndianByteArrayOutputStream;
@@ -36,7 +35,7 @@ public final class TestConstantValueParser extends TestCase {
 			Boolean.TRUE,
 			null,
 			new Double(1.1),
-			new UnicodeString("Sample text"),
+			"Sample text",
 			ErrorConstant.valueOf(HSSFErrorConstants.ERROR_DIV_0),
 		};
 	private static final byte[] SAMPLE_ENCODING = HexRead.readFromString(

src/testcases/org/apache/poi/hssf/record/formula/TestArrayPtg.java

@@ -66,10 +66,10 @@ public final class TestArrayPtg extends TestCase {
 		
 		
 		assertEquals(Boolean.TRUE, values[0][0]);
-		assertEquals(new UnicodeString("ABCD"), values[0][1]);
+		assertEquals("ABCD", values[0][1]);
 		assertEquals(new Double(0), values[1][0]);
 		assertEquals(Boolean.FALSE, values[1][1]);
-		assertEquals(new UnicodeString("FG"), values[1][2]);
+		assertEquals("FG", values[1][2]);
 		
 		byte[] outBuf = new byte[ENCODED_CONSTANT_DATA.length];
 		ptg.writeTokenValueBytes(new LittleEndianByteArrayOutputStream(outBuf, 0));

src/testcases/org/apache/poi/hssf/usermodel/TestHSSFCell.java

@@ -28,8 +28,8 @@ import org.apache.poi.hssf.model.Sheet;
 import org.apache.poi.hssf.util.HSSFColor;
 
 /**
- * Tests various functionity having to do with HSSFCell.  For instance support for
+ * Tests various functionality having to do with {@link HSSFCell}.  For instance support for
- * paticular datatypes, etc.
+ * particular datatypes, etc.
  * @author Andrew C. Oliver (andy at superlinksoftware dot com)
  * @author  Dan Sherman (dsherman at isisph.com)
  * @author Alex Jacoby (ajacoby at gmail.com)
@@ -349,7 +349,7 @@ public final class TestHSSFCell extends TestCase {
 	 * Test to ensure we can only assign cell styles that belong
 	 *  to our workbook, and not those from other workbooks.
 	 */
-    public void testCellStyleWorkbookMatch() throws Exception {
+	public void testCellStyleWorkbookMatch() {
 		HSSFWorkbook wbA = new HSSFWorkbook();
 		HSSFWorkbook wbB = new HSSFWorkbook();
 
@@ -381,5 +381,46 @@ public final class TestHSSFCell extends TestCase {
 			fail();
 		} catch (IllegalArgumentException e) {}
 	}
+
+	public void testChangeTypeStringToBool() {
+		HSSFCell cell = new HSSFWorkbook().createSheet("Sheet1").createRow(0).createCell(0);
+
+		cell.setCellValue(new HSSFRichTextString("TRUE"));
+		assertEquals(HSSFCell.CELL_TYPE_STRING, cell.getCellType());
+		try {
+			cell.setCellType(HSSFCell.CELL_TYPE_BOOLEAN);
+		} catch (ClassCastException e) {
+			throw new AssertionFailedError(
+					"Identified bug in conversion of cell from text to boolean");
+		}
+
+		assertEquals(HSSFCell.CELL_TYPE_BOOLEAN, cell.getCellType());
+		assertEquals(true, cell.getBooleanCellValue());
+		cell.setCellType(HSSFCell.CELL_TYPE_STRING);
+		assertEquals("TRUE", cell.getRichStringCellValue().getString());
+
+		// 'false' text to bool and back
+		cell.setCellValue(new HSSFRichTextString("FALSE"));
+		cell.setCellType(HSSFCell.CELL_TYPE_BOOLEAN);
+		assertEquals(HSSFCell.CELL_TYPE_BOOLEAN, cell.getCellType());
+		assertEquals(false, cell.getBooleanCellValue());
+		cell.setCellType(HSSFCell.CELL_TYPE_STRING);
+		assertEquals("FALSE", cell.getRichStringCellValue().getString());
+	}
+
+	public void testChangeTypeBoolToString() {
+		HSSFCell cell = new HSSFWorkbook().createSheet("Sheet1").createRow(0).createCell(0);
+		cell.setCellValue(true);
+		try {
+			cell.setCellType(HSSFCell.CELL_TYPE_STRING);
+		} catch (IllegalStateException e) {
+			if (e.getMessage().equals("Cannot get a text value from a boolean cell")) {
+				throw new AssertionFailedError(
+						"Identified bug in conversion of cell from boolean to text");
+			}
+			throw e;
+		}
+		assertEquals("TRUE", cell.getRichStringCellValue().getString());
+	}
 }