44929 - Improved error handling in HSSFWorkbook when attempting to read a BIFF5 file

git-svn-id: https://svn.apache.org/repos/asf/poi/trunk@653117 13f79535-47bb-0310-9956-ffa450edef68

src/documentation/content/xdocs/changes.xml

@@ -37,6 +37,7 @@
 
 		<!-- Don't forget to update status.xml too! -->
         <release version="3.1-beta2" date="2008-05-??">
+           <action dev="POI-DEVELOPERS" type="fix">44929 - Improved error handling in HSSFWorkbook when attempting to read a BIFF5 file</action>
            <action dev="POI-DEVELOPERS" type="fix">44675 - Parameter operand classes (function metadata) required to encode SUM() etc properly. Added parse validation for number of parameters</action>
            <action dev="POI-DEVELOPERS" type="fix">44921 - allow Ptg.writeBytes() to be called on relative ref Ptgs (RefN* and AreaN*)</action>
            <action dev="POI-DEVELOPERS" type="fix">44914 - Fix/suppress warning message "WARN. Unread n bytes of record 0xNN"</action>

src/documentation/content/xdocs/status.xml

@@ -34,6 +34,7 @@
 	<!-- Don't forget to update changes.xml too! -->
     <changes>
         <release version="3.1-beta2" date="2008-05-??">
+           <action dev="POI-DEVELOPERS" type="fix">44929 - Improved error handling in HSSFWorkbook when attempting to read a BIFF5 file</action>
            <action dev="POI-DEVELOPERS" type="fix">44675 - Parameter operand classes (function metadata) required to encode SUM() etc properly. Added parse validation for number of parameters</action>
            <action dev="POI-DEVELOPERS" type="fix">44921 - allow Ptg.writeBytes() to be called on relative ref Ptgs (RefN* and AreaN*)</action>
            <action dev="POI-DEVELOPERS" type="fix">44914 - Fix/suppress warning message "WARN. Unread n bytes of record 0xNN"</action>

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

@@ -165,7 +165,43 @@ public class HSSFWorkbook extends POIDocument
     public HSSFWorkbook(POIFSFileSystem fs, boolean preserveNodes)
             throws IOException
     {
-    	this(fs.getRoot(), fs, preserveNodes);
+        this(fs.getRoot(), fs, preserveNodes);
+    }
+
+    /**
+     * Normally, the Workbook will be in a POIFS Stream
+     * called "Workbook". However, some weird XLS generators use "WORKBOOK"
+     */
+    private static final String[] WORKBOOK_DIR_ENTRY_NAMES = {
+        "Workbook", // as per BIFF8 spec
+        "WORKBOOK",
+    };
+
+
+    private static String getWorkbookDirEntryName(DirectoryNode directory) {
+
+        String[] potentialNames = WORKBOOK_DIR_ENTRY_NAMES;
+        for (int i = 0; i < potentialNames.length; i++) {
+            String wbName = potentialNames[i];
+            try {
+                directory.getEntry(wbName);
+                return wbName;
+            } catch (FileNotFoundException e) {
+                // continue - to try other options
+            }
+        }
+
+        // check for previous version of file format
+        try {
+            directory.getEntry("Book");
+            throw new IllegalArgumentException("The supplied spreadsheet seems to be Excel 5.0/7.0 (BIFF5) format. "
+                    + "POI only supports BIFF8 format (from Excel versions 97/2000/XP/2003)");
+        } catch (FileNotFoundException e) {
+            // fall through
+        }
+
+        throw new IllegalArgumentException("The supplied POIFSFileSystem does not contain a BIFF8 'Workbook' entry. "
+            + "Is it really an excel file?");
     }
 
     /**
@@ -185,7 +221,9 @@ public class HSSFWorkbook extends POIDocument
     public HSSFWorkbook(DirectoryNode directory, POIFSFileSystem fs, boolean preserveNodes)
             throws IOException
     {
-    	super(directory, fs);
+        super(directory, fs);
+        String workbookName = getWorkbookDirEntryName(directory);
+
         this.preserveNodes = preserveNodes;
 
         // If we're not preserving nodes, don't track the
@@ -198,27 +236,8 @@ public class HSSFWorkbook extends POIDocument
         sheets = new ArrayList(INITIAL_CAPACITY);
         names  = new ArrayList(INITIAL_CAPACITY);
 
-        // Normally, the Workbook will be in a POIFS Stream
-        //  called "Workbook". However, some wierd XLS generators
-        //  put theirs in one called "WORKBOOK"
-        String workbookName = "Workbook";
-        try {
-        	directory.getEntry(workbookName);
-        	// Is the default name
-        } catch(FileNotFoundException fe) {
-        	// Try the upper case form
-        	try {
-        		workbookName = "WORKBOOK";
-        		directory.getEntry(workbookName);
-        	} catch(FileNotFoundException wfe) {
-        		// Doesn't contain it in either form
-        		throw new IllegalArgumentException("The supplied POIFSFileSystem contained neither a 'Workbook' entry, nor a 'WORKBOOK' entry. Is it really an excel file?");
-        	}
-        }
-
-        
         // Grab the data from the workbook stream, however
-        //  it happens to be spelt.
+        //  it happens to be spelled.
         InputStream stream = directory.createDocumentInputStream(workbookName);
 
         EventRecordFactory factory = new EventRecordFactory();
@@ -516,12 +535,12 @@ public class HSSFWorkbook extends POIDocument
      */
     public int getSheetIndex(HSSFSheet sheet)
     {
-    	for(int i=0; i<sheets.size(); i++) {
+        for(int i=0; i<sheets.size(); i++) {
-    		if(sheets.get(i) == sheet) {
+            if(sheets.get(i) == sheet) {
-    			return i;
+                return i;
-    		}
+            }
-    	}
+        }
-    	return -1;
+        return -1;
     }
 
     /**
@@ -532,7 +551,7 @@ public class HSSFWorkbook extends POIDocument
      *  named range things.
      */
     public short getExternalSheetIndex(int internalSheetIndex) {
-    	return workbook.checkExternSheet(internalSheetIndex);
+        return workbook.checkExternSheet(internalSheetIndex);
     }
 
     /**
@@ -576,18 +595,18 @@ public class HSSFWorkbook extends POIDocument
         sheets.add(clonedSheet);
         int i=1;
         while (true) {
-        	//Try and find the next sheet name that is unique
+            //Try and find the next sheet name that is unique
-        	String name = srcName;
+            String name = srcName;
-        	String index = Integer.toString(i++);
+            String index = Integer.toString(i++);
-        	if (name.length()+index.length()+2<31)
+            if (name.length()+index.length()+2<31)
-        	  name = name + "("+index+")";
+              name = name + "("+index+")";
-        	else name = name.substring(0, 31-index.length()-2)+"("+index+")";
+            else name = name.substring(0, 31-index.length()-2)+"("+index+")";
 
-        	//If the sheet name is unique, then set it otherwise move on to the next number.
+            //If the sheet name is unique, then set it otherwise move on to the next number.
-        	if (workbook.getSheetIndex(name) == -1) {
+            if (workbook.getSheetIndex(name) == -1) {
               workbook.setSheetName(sheets.size()-1, name);
               break;
-        	}
+            }
         }
         return clonedSheet;
       }
@@ -663,7 +682,7 @@ public class HSSFWorkbook extends POIDocument
     }
 
     public SheetReferences getSheetReferences() {
-    	return workbook.getSheetReferences();
+        return workbook.getSheetReferences();
     }
 
     /**
@@ -999,13 +1018,13 @@ public class HSSFWorkbook extends POIDocument
         writeProperties(fs, excepts);
 
         if (preserveNodes) {
-			// Don't write out the old Workbook, we'll be doing our new one
+            // Don't write out the old Workbook, we'll be doing our new one
             excepts.add("Workbook");
-			// If the file had WORKBOOK instead of Workbook, we'll write it
+            // If the file had WORKBOOK instead of Workbook, we'll write it
-			//  out correctly shortly, so don't include the old one
+            //  out correctly shortly, so don't include the old one
             excepts.add("WORKBOOK");
 
-			// Copy over all the other nodes to our new poifs
+            // Copy over all the other nodes to our new poifs
             copyNodes(this.filesystem,fs,excepts);
         }
         fs.writeFilesystem(stream);
@@ -1127,76 +1146,76 @@ public class HSSFWorkbook extends POIDocument
      * @return the string representation of the defined or external name
      */
     public String resolveNameXText(int refIndex, int definedNameIndex) {
-    	return workbook.resolveNameXText(refIndex, definedNameIndex);
+        return workbook.resolveNameXText(refIndex, definedNameIndex);
     }
 
 
-	/**
+    /**
-	 * Sets the printarea for the sheet provided
+     * Sets the printarea for the sheet provided
-	 * <p>
+     * <p>
-	 * i.e. Reference = $A$1:$B$2
+     * i.e. Reference = $A$1:$B$2
-	 * @param sheetIndex Zero-based sheet index (0 Represents the first sheet to keep consistent with java)
+     * @param sheetIndex Zero-based sheet index (0 Represents the first sheet to keep consistent with java)
-	 * @param reference Valid name Reference for the Print Area
+     * @param reference Valid name Reference for the Print Area
-	 */
+     */
-	public void setPrintArea(int sheetIndex, String reference)
+    public void setPrintArea(int sheetIndex, String reference)
-	{
+    {
-		NameRecord name = workbook.getSpecificBuiltinRecord(NameRecord.BUILTIN_PRINT_AREA, sheetIndex+1);
+        NameRecord name = workbook.getSpecificBuiltinRecord(NameRecord.BUILTIN_PRINT_AREA, sheetIndex+1);
 
 
-		if (name == null)
+        if (name == null)
-			name = workbook.createBuiltInName(NameRecord.BUILTIN_PRINT_AREA, sheetIndex+1);
+            name = workbook.createBuiltInName(NameRecord.BUILTIN_PRINT_AREA, sheetIndex+1);
        //adding one here because 0 indicates a global named region; doesnt make sense for print areas
 
-	    short externSheetIndex = getWorkbook().checkExternSheet(sheetIndex);
+        short externSheetIndex = getWorkbook().checkExternSheet(sheetIndex);
-		name.setExternSheetNumber(externSheetIndex);
+        name.setExternSheetNumber(externSheetIndex);
-		name.setAreaReference(reference);
+        name.setAreaReference(reference);
 
 
-	}
+    }
 
-	/**
+    /**
-	 * For the Convenience of Java Programmers maintaining pointers.
+     * For the Convenience of Java Programmers maintaining pointers.
-	 * @see #setPrintArea(int, String)
+     * @see #setPrintArea(int, String)
-	 * @param sheetIndex Zero-based sheet index (0 = First Sheet)
+     * @param sheetIndex Zero-based sheet index (0 = First Sheet)
-	 * @param startColumn Column to begin printarea
+     * @param startColumn Column to begin printarea
-	 * @param endColumn Column to end the printarea
+     * @param endColumn Column to end the printarea
-	 * @param startRow Row to begin the printarea
+     * @param startRow Row to begin the printarea
-	 * @param endRow Row to end the printarea
+     * @param endRow Row to end the printarea
-	 */
+     */
-	public void setPrintArea(int sheetIndex, int startColumn, int endColumn,
+    public void setPrintArea(int sheetIndex, int startColumn, int endColumn,
-							  int startRow, int endRow) {
+                              int startRow, int endRow) {
 
-		//using absolute references because they don't get copied and pasted anyway
+        //using absolute references because they don't get copied and pasted anyway
-		CellReference cell = new CellReference(startRow, startColumn, true, true);
+        CellReference cell = new CellReference(startRow, startColumn, true, true);
-		String reference = cell.formatAsString();
+        String reference = cell.formatAsString();
 
-		cell = new CellReference(endRow, endColumn, true, true);
+        cell = new CellReference(endRow, endColumn, true, true);
-		reference = reference+":"+cell.formatAsString();
+        reference = reference+":"+cell.formatAsString();
 
-		setPrintArea(sheetIndex, reference);
+        setPrintArea(sheetIndex, reference);
-	}
+    }
 
 
-	/**
+    /**
-	 * Retrieves the reference for the printarea of the specified sheet, the sheet name is appended to the reference even if it was not specified.
+     * Retrieves the reference for the printarea of the specified sheet, the sheet name is appended to the reference even if it was not specified.
-	 * @param sheetIndex Zero-based sheet index (0 Represents the first sheet to keep consistent with java)
+     * @param sheetIndex Zero-based sheet index (0 Represents the first sheet to keep consistent with java)
-	 * @return String Null if no print area has been defined
+     * @return String Null if no print area has been defined
-	 */
+     */
-	public String getPrintArea(int sheetIndex)
+    public String getPrintArea(int sheetIndex)
-	{
+    {
-		NameRecord name = workbook.getSpecificBuiltinRecord(NameRecord.BUILTIN_PRINT_AREA, sheetIndex+1);
+        NameRecord name = workbook.getSpecificBuiltinRecord(NameRecord.BUILTIN_PRINT_AREA, sheetIndex+1);
-		if (name == null) return null;
+        if (name == null) return null;
-		//adding one here because 0 indicates a global named region; doesnt make sense for print areas
+        //adding one here because 0 indicates a global named region; doesnt make sense for print areas
 
-		return name.getAreaReference(this);
+        return name.getAreaReference(this);
-	}
+    }
 
     /**
      * Delete the printarea for the sheet specified
      * @param sheetIndex Zero-based sheet index (0 = First Sheet)
      */
     public void removePrintArea(int sheetIndex) {
-    	getWorkbook().removeBuiltinRecord(NameRecord.BUILTIN_PRINT_AREA, sheetIndex+1);
+        getWorkbook().removeBuiltinRecord(NameRecord.BUILTIN_PRINT_AREA, sheetIndex+1);
     }
 
     /** creates a new named range and add it to the model
@@ -1252,9 +1271,9 @@ public class HSSFWorkbook extends POIDocument
      * @see org.apache.poi.hssf.record.Record
      */
     public HSSFDataFormat createDataFormat() {
-	if (formatter == null)
+    if (formatter == null)
-	    formatter = new HSSFDataFormat(workbook);
+        formatter = new HSSFDataFormat(workbook);
-	return formatter;
+    return formatter;
     }
 
     /** remove the named range by his name
@@ -1433,7 +1452,7 @@ public class HSSFWorkbook extends POIDocument
      * Is the workbook protected with a password (not encrypted)?
      */
     public boolean isWriteProtected() {
-    	return this.workbook.isWriteProtected();
+        return this.workbook.isWriteProtected();
     }
 
     /**