2008-01-29 Ito Kazumitsu <kaz@maczuka.gcd.org>

	Fixes bug #22941
	* java/io/DataInputStream.java(readUTFLong): New method.
	(readUTF): New private method.
	* java/io/DataOutputStream.java(getUTFlength): Made package-private,
	Return type changed to long.
	(writeUTF): Use the new method writeUTFShort.
	(writeUTFShort): New package-private method.
	(writeUTFLong): New package-private method.
	(writeUTFBytes): New private method.
	* java/io/ObjectInputStream.java(parseContent): Separate the
	handling of TC_LONGSTRING from TC_STRING.
	* java/io/ObjectOutputStream.java(writeObject): When to write a
	String, use writeUTFShort or writeUTFLong depending on the byte length.

5 files changed, 158 insertions, 27 deletions

diff --git a/ChangeLog b/ChangeLog
index da1c640ab..f625f84b9 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,19 @@
+2008-01-29  Ito Kazumitsu  <kaz@maczuka.gcd.org>
+
+	Fixes bug #22941
+	* java/io/DataInputStream.java(readUTFLong): New method.
+	(readUTF): New private method.
+	* java/io/DataOutputStream.java(getUTFlength): Made package-private,
+	Return type changed to long.
+	(writeUTF): Use the new method writeUTFShort.
+	(writeUTFShort): New package-private method.
+	(writeUTFLong): New package-private method.
+	(writeUTFBytes): New private method.
+	* java/io/ObjectInputStream.java(parseContent): Separate the
+	handling of TC_LONGSTRING from TC_STRING.
+	* java/io/ObjectOutputStream.java(writeObject): When to write a
+	String, use writeUTFShort or writeUTFLong depending on the byte length.
+
 2008-01-27  Bernhard Fischer  <rep.dot.nop@gmail.com>
 
 	* java/util/SimpleTimeZone.java (SimpleTimeZone): Fix typo in comment.
diff --git a/java/io/DataInputStream.java b/java/io/DataInputStream.java
index d2604b51f..8b223ea84 100644
--- a/java/io/DataInputStream.java
+++ b/java/io/DataInputStream.java
@@ -1,5 +1,6 @@
 /* DataInputStream.java -- FilteredInputStream that implements DataInput
-   Copyright (C) 1998, 1999, 2000, 2001, 2003, 2005  Free Software Foundation
+   Copyright (C) 1998, 1999, 2000, 2001, 2003, 2005, 2008
+   Free Software Foundation
 
 This file is part of GNU Classpath.
 
@@ -590,13 +591,56 @@ public class DataInputStream extends FilterInputStream implements DataInput
   public static final String readUTF(DataInput in) throws IOException
   {
     final int UTFlen = in.readUnsignedShort ();
-    byte[] buf = new byte [UTFlen];
+    
+    return readUTF(in, UTFlen);
+  }
+
+  /**
+   * This method is similar to <code>readUTF</code>, but the
+   * UTF-8 byte length is in 64 bits.
+   * This method is not public. It is used by <code>ObjectInputStream</code>.
+   * 
+   * @return The <code>String</code> read
+   *
+   * @exception EOFException If end of file is reached before reading
+   * the String
+   * @exception UTFDataFormatException If the data is not in UTF-8 format
+   * @exception IOException If any other error occurs
+   *
+   * @see DataOutput#writeUTFLong
+   */
+  final String readUTFLong () throws IOException
+  {
+    long l = readLong ();
+    if (l > Integer.MAX_VALUE)
+      throw new IOException("The string length > Integer.MAX_VALUE");
+    final int UTFlen = (int)l;
+    return readUTF (this, UTFlen);
+  }
+
+  /**
+   * This method performs the main task of <code>readUTF</code> and
+   * <code>readUTFLong</code>.
+   *
+   * @param in The <code>DataInput</code> source to read from
+   *
+   * @param len The UTF-8 byte length of the String to be read
+   *
+   * @return The String read from the source
+   *
+   * @exception IOException If an error occurs
+   *
+   * @see DataInput#readUTF
+   */
+  private static final String readUTF(DataInput in, int len) throws IOException
+  {
+    byte[] buf = new byte [len];
 
     // This blocks until the entire string is available rather than
     // doing partial processing on the bytes that are available and then
     // blocking.  An advantage of the latter is that Exceptions
     // could be thrown earlier.  The former is a bit cleaner.
-    in.readFully (buf, 0, UTFlen);
+    in.readFully (buf, 0, len);
 
     return convertFromUTF (buf);
   }
diff --git a/java/io/DataOutputStream.java b/java/io/DataOutputStream.java
index 6670c2dba..435ff76d1 100644
--- a/java/io/DataOutputStream.java
+++ b/java/io/DataOutputStream.java
@@ -1,5 +1,5 @@
 /* DataOutputStream.java -- Writes primitive Java datatypes to streams
-   Copyright (C) 1998, 2001, 2003, 2005  Free Software Foundation, Inc.
+   Copyright (C) 1998, 2001, 2003, 2005, 2008  Free Software Foundation, Inc.
 
 This file is part of GNU Classpath.
 
@@ -379,19 +379,20 @@ public class DataOutputStream extends FilterOutputStream implements DataOutput
 
   /**
    *  Calculate the length, in bytes, of a <code>String</code> in Utf8 format.
+   *  This method is package-private so that <code>ObjectOutputStream</code>
+   *  may use it.  The return type is long so that a long string whose
+   *  Utf8 byte count is 64 bit long may be handled.
    *
    *  @param value The <code>String</code> to measure
    *  @param start String index at which to begin count
    *  @param sum Starting Utf8 byte count
    *
-   *  @throws UTFDataFormatException if result would exceed 65535
    */
-  private int getUTFlength(String value, int start, int sum)
-    throws IOException
+  long getUTFlength(String value, int start, long sum)
   {
     int len = value.length();
 
-    for (int i = start; i < len && sum <= 65535; ++i)
+    for (int i = start; i < len; ++i)
       {
 	char c = value.charAt(i);
 	if (c >= '\u0001' && c <= '\u007f')
@@ -402,9 +403,6 @@ public class DataOutputStream extends FilterOutputStream implements DataOutput
 	  sum += 3;
       }
 
-    if (sum > 65535)
-      throw new UTFDataFormatException ();
-
     return sum;
   }
   
@@ -442,10 +440,70 @@ public class DataOutputStream extends FilterOutputStream implements DataOutput
    */
   public final synchronized void writeUTF(String value) throws IOException
   {
+    long l = getUTFlength(value, 0, 0);
+    if (l > 65535)
+      throw new UTFDataFormatException ();
+    writeUTFShort(value, (int)l);
+  }
+
+  /**
+   * This method performs the main task of <code>writeUTF</code>.
+   * This method is package-private because ObjectOutputStream uses it.
+   *
+   * @param value The <code>String</code> to write to the output in UTF format
+   *
+   * @param bytelen The UTF-8 byte length of the <code>String</code>. When
+   * this method is called, the expected byte length must have been calculated
+   * by <code>getUTFlength</code>.
+   *
+   * @exception IOException If an error occurs
+   *
+   * @see DataInput#readUTF
+   */
+  final synchronized void writeUTFShort(String value, int bytelen)
+    throws IOException
+  {
+    writeShort(bytelen);
+    writeUTFBytes(value);
+  }
+
+  /**
+   * This method is similar to <code>writeUTF</code>, but it writes the
+   * UTF-8 byte length in 64 bits.
+   * This method is not public but <code>ObjectOutputStream</code> uses it.
+   *
+   * @param value The <code>String</code> to write to the output in UTF format
+   *
+   * @param bytelen The UTF-8 byte length of the <code>String</code>. When
+   * this method is called, the expected byte length must have been calculated
+   * by <code>getUTFlength</code>.
+   *
+   * @exception IOException If an error occurs
+   *
+   */
+  final synchronized void writeUTFLong(String value, long bytelen)
+    throws IOException
+  {
+    writeLong(bytelen);
+    writeUTFBytes(value);
+  }
+
+  /**
+   * This method performes the main task of <code>writeUTF</code> and
+   * <code>WriteUTFLong</code>, which is to write the UTF-8 byte
+   * sequence to the output.
+   *
+   * @param value The <code>String</code> to write to the output in UTF format
+   *
+   * @exception IOException If an error occurs
+   *
+   */
+  private final synchronized void writeUTFBytes(String value)
+    throws IOException
+  {
     int len = value.length();
     int i = 0;
     int pos = 0;
-    boolean lengthWritten = false;
 
     if (buf == null)
       buf = new byte[512];
@@ -472,14 +530,6 @@ public class DataOutputStream extends FilterOutputStream implements DataOutput
 		buf[pos++] = (byte) (0x80 | (0x3f & c));
 	      }
 	  }
-	if (! lengthWritten)
-	  {
-	    if (i == len)
-	      writeShort(pos);
-	    else
-	      writeShort(getUTFlength(value, i, pos));
-	    lengthWritten = true;
-	  }
 	write(buf, 0, pos);
 	pos = 0;
      }
diff --git a/java/io/ObjectInputStream.java b/java/io/ObjectInputStream.java
index ea3ffde61..36e70d6a4 100644
--- a/java/io/ObjectInputStream.java
+++ b/java/io/ObjectInputStream.java
@@ -1,5 +1,5 @@
 /* ObjectInputStream.java -- Class used to read serialized objects
-   Copyright (C) 1998, 1999, 2000, 2001, 2002, 2003, 2005, 2006
+   Copyright (C) 1998, 1999, 2000, 2001, 2002, 2003, 2005, 2006, 2008
    Free Software Foundation, Inc.
 
 This file is part of GNU Classpath.
@@ -356,7 +356,6 @@ public class ObjectInputStream extends InputStream
  	}
  	
        case TC_STRING:
-       case TC_LONGSTRING:
  	{
  	  if(dump) dumpElement("STRING=");
  	  String s = this.realInputStream.readUTF();
@@ -366,6 +365,16 @@ public class ObjectInputStream extends InputStream
  	  break;
  	}
  
+       case TC_LONGSTRING:
+ 	{
+ 	  if(dump) dumpElement("STRING=");
+ 	  String s = this.realInputStream.readUTFLong();
+ 	  if(dump) dumpElementln(s);
+ 	  ret_val = processResolution(null, s, assignNewHandle(s,shared),
+				      shared);
+ 	  break;
+ 	}
+
        case TC_ARRAY:
  	{
  	  if(dump) dumpElementln("ARRAY");
diff --git a/java/io/ObjectOutputStream.java b/java/io/ObjectOutputStream.java
index 5cb2a73f9..c09487536 100644
--- a/java/io/ObjectOutputStream.java
+++ b/java/io/ObjectOutputStream.java
@@ -1,5 +1,5 @@
 /* ObjectOutputStream.java -- Class used to write serialized objects
-   Copyright (C) 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006
+   Copyright (C) 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2008
    Free Software Foundation, Inc.
 
 This file is part of GNU Classpath.
@@ -363,10 +363,22 @@ public class ObjectOutputStream extends OutputStream
 
 	    if (obj instanceof String)
 	      {
-		realOutput.writeByte(TC_STRING);
-		if (shared)
-		  assignNewHandle(obj);
-		realOutput.writeUTF((String)obj);
+                String s = (String)obj;
+                long l = realOutput.getUTFlength(s, 0, 0);
+                if (l <= 65535)
+                  {
+		    realOutput.writeByte(TC_STRING);
+		    if (shared)
+		      assignNewHandle(obj);
+		    realOutput.writeUTFShort(s, (int)l);
+                  }
+                else
+                  {
+		    realOutput.writeByte(TC_LONGSTRING);
+		    if (shared)
+		      assignNewHandle(obj);
+		    realOutput.writeUTFLong(s, l);
+                  }
 		break;
 	      }