From 520042837423-0001@t-online.de Wed Nov 01 16:29:41 2000
Received: from mailout03.sul.t-online.com ([194.25.134.81])
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13r5S0-00077H-00
	for <classpath@gnu.org>; Wed, 01 Nov 2000 16:29:41 -0500
Received: from fwd04.sul.t-online.com 
	by mailout03.sul.t-online.com with smtp 
	id 13r5Rz-0006Bw-06; Wed, 01 Nov 2000 22:29:39 +0100
Received: from turbosport.com (520042837423-0001@[217.2.235.236]) by fwd04.sul.t-online.com
	with smtp id 13r5Rq-0x0MfAC; Wed, 1 Nov 2000 22:29:30 +0100
From: 520042837423-0001@t-online.de (ECS Startup Info)
To: "&lt;classpath@gnu.org&gt;" <classpath@gnu.org>
Subject: Personal StartUP-Information
Mime-Version: 1.0
Content-Type: text/html; charset="ISO-8859-1"
Date: Wed, 1 Nov 2000 22:28:57 +0100
Content-Transfer-Encoding: 8bit
Message-ID: <13r5Rq-0x0MfAC@fwd04.sul.t-online.com>
X-Sender: 520042837423-0001@t-dialin.net
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 FINAL//EN">

<HTML>
<HEAD><META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=UTF-8">

<TITLE>Home</TITLE>
<META NAME="Generator" CONTENT="Don't know">

</HEAD>

<BODY  BGCOLOR="#FFFFFF" LINK="#0000FF" VLINK="#990000" TEXT="#000000" TOPMARGIN=16 LEFTMARGIN=10 MARGINWIDTH=10 MARGINHEIGHT=16>

  <TABLE BORDER=0 CELLSPACING=0 CELLPADDING=0 WIDTH=541>
   <TR VALIGN="top" ALIGN="left">
	<TD WIDTH=541 COLSPAN=2>
<P ALIGN="CENTER"><A HREF="http://www.geocities.com/bigearn909/easycashsystem.html"><FONT SIZE="+3"><U>Easy-Cash-System</U></FONT></A></U></FONT>
<TABLE WIDTH=539 BORDER=0 CELLSPACING=0 CELLPADDING=0 ALIGN=LEFT><TR>
<TD>
<P>I apologize if this e-mail came to you by mistake. If you don't want to receive future e-mail messages about Easy-Cash-System, please do NOT reply to this e-mail message. <BR>Instead, use this link to beremoved from this mailing list.<FONT COLOR="#669900"> </FONT><A HREF="mailto:b_earn@hotmail.com?subject=REMOVE"><FONT COLOR="#669900"></FONT><FONT COLOR="#999900">REMOVE NOW!</FONT></A><FONT COLOR="#999900"><BR></FONT>Otherwise, if you do nothing you should even be removed from this list. <BR>&nbsp;</TD></TR></TABLE>
<FONT SIZE="+3"><U><A HREF="http://www.geocities.com/bigearn999/easycashsystem.html"><FONT SIZE="+3"><U></U></FONT></A><FONT SIZE="+3"><U></U></FONT></TD>
   </TR>
   <TR VALIGN="top" ALIGN="left">
	<TD WIDTH=539 HEIGHT =33><IMG SRC="./clearpixel.gif" WIDTH =539 HEIGHT=1 BORDER=0></TD>
	<TD WIDTH=2><IMG SRC="./clearpixel.gif" WIDTH =2 HEIGHT=1 BORDER=0></TD>
   </TR>
   <TR VALIGN="top" ALIGN="left">
	<TD WIDTH=539>
<P ALIGN="CENTER"><B><FONT SIZE="+1">Still reading?</P>
<P ALIGN="CENTER"></FONT></B><FONT SIZE="+1">So read on, and it could bring you another opportunity to earn <BR></FONT><FONT COLOR="#FF0033" SIZE="+2"><B>cash</B></FONT><FONT SIZE="+1"> while you are </FONT><FONT COLOR="#FF0033" SIZE="+2"><B>surfing</B></FONT><FONT COLOR="#FF0033" SIZE="+1"><B>.</B></FONT></TD>
	<TD></TD>
   </TR>
   <TR VALIGN="top" ALIGN="left">
	<TD COLSPAN=2 HEIGHT =60></TD>
   </TR>
   <TR VALIGN="top" ALIGN="left">
	<TD WIDTH=539>
<P ALIGN="CENTER">The following is a list of mostly recommended payment systems I&#194;&#180;ve found.</P>
<P ALIGN="CENTER">If you like you can use the following links to start making moneyright away!!!! <BR><BR>It&#194;&#180;s also possible to visit my page when you <A HREF="http://www.geocities.com/bigearn999/easycashsystem.html">click here</A>. There&#194;&#180;s the possibility to sign up for the newsletter which includes only the best paying programms.</TD>
	<TD></TD>
   </TR>
   <TR VALIGN="top" ALIGN="left">
	<TD COLSPAN=2 HEIGHT =28></TD>
   </TR>
   <TR VALIGN="top" ALIGN="left">
	<TD WIDTH=539>
<P ALIGN="CENTER">When you decide to use this pages, please,<B> use these links!<BR></B>So, it&#194;&#180;s possible to send you more of this programs directly to your home!</P>
<P ALIGN="CENTER"><FONT SIZE="+2">My favourite pages with </FONT><FONT COLOR="#FF0033" SIZE="+2"><B>high payouts</B></FONT><FONT SIZE="+2"> are:</P>
<P ALIGN="CENTER"></FONT><A HREF="http://www.bepaid.com/users.rhtml?REFID=11041966"><FONT SIZE="+2">BePaid</FONT></A><FONT SIZE="+2"><BR></FONT><A HREF="http://www.cashfiesta.com/php/join.php?ref=bigearn999"><FONT SIZE="+2">CashFiesta</FONT></A><FONT SIZE="+2"><BR></FONT><A HREF="http://net-4-biz.affiliatetracking.net/al/a?96232"><FONT SIZE="+2">NetBiz</FONT></A><FONT SIZE="+2"><BR></FONT><A HREF="http://www.jackpot.com/win.asp?ref=BigEarns"><FONT SIZE="+2">JACKPOT!</FONT></A><FONT SIZE="+2"><BR>&nbsp;</P>
<P ALIGN="CENTER"></FONT>So, please<B> use the above links</B>.<BR>Thank you for spending your time reading this.<BR><A HREF="http://www.geocities.com/bigearn909/easycashsystem.html">Easy-Cash-System<BR></A>&nbsp;</TD>
	<TD></TD>
   </TR>
  </TABLE></BODY>
</HTML>


From bryce@albatross.co.nz Thu Nov 02 05:03:08 2000
Received: from waitaki.otago.ac.nz ([139.80.75.140])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13rHD8-0006Ea-00
	for <classpath@gnu.org>; Thu, 02 Nov 2000 05:03:06 -0500
Received: from albatross.co.nz (IDENT:bryce@reason.otago.ac.nz [139.80.75.188])
	by waitaki.otago.ac.nz (8.9.3/8.9.3) with ESMTP id XAA15798;
	Thu, 2 Nov 2000 23:02:56 +1300
Message-ID: <3A013C0E.5CFC7F02@albatross.co.nz>
Date: Thu, 02 Nov 2000 23:03:59 +1300
From: Bryce McKinlay <bryce@albatross.co.nz>
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.4.0-test9 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: java-patches@sources.redhat.com, classpath@gnu.org
Subject: PATCH: More collections List updates
Content-Type: multipart/mixed;
 boundary="------------9DE53B4993AF87E7337C78A0"
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

This is a multi-part message in MIME format.
--------------9DE53B4993AF87E7337C78A0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Hi,

I have rewritten most of LinkedList. The new implementation is simpler,
faster and conforms to the spec more closely. While working on this I
realised the flaw in my previous change (removal of) the SubList
iterator in AbstractList: Although correct, it resulted in poor
performance for iterators on sublists of linked lists, as the iterator
would ultimately traverse the linked list via its get() method,
requiring quadratic-ish time. The fix was to restore the existing
iterator wrapper approach. A few other minor bugs are also fixed by this
patch.

Unfortunately, gcj can not build AbstractList any more due to a static
inner class bug (specifically, it cant handle the references to modCount
from the Sublist iterator anonymous class). So, what I'm going to do is
commit this patch to both classpath and libgcj, and then comment out the
offending lines in libgcj as a workaround until the compiler is fixed.

regards

  [ bryce ]


--------------9DE53B4993AF87E7337C78A0
Content-Type: text/plain; charset=us-ascii;
 name="libgcj-collections-2.patch"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="libgcj-collections-2.patch"

2000-11-02  Bryce McKinlay  <bryce@albatross.co.nz>

	* java/util/AbstractList.java: Throw messages with 
	IndexOutOfBoundsExceptions.
	 (listIterator()): Call listIterator(0).
	(size): New field. Initialize to size().
	(hasNext): Test position against size, not size().
	(remove): Increment knownMod by one instead of resetting it from 
	modCount.
	(add): Ditto.
	(SubList.upMod): Removed.
	(SubList.set): Don't call upMod() or update knownMod.
	(SubList.add(int,Object)): Increment modCount instead of caling upMod().
	(SubList.remove): Ditto.
	(SubList.addAll): Don't call backingList.size(). Increment size from 
	c.size().
	(SubList.iterator): New method. Call listIterator(0).
	(SubList.listIterator): New method. Restore code to return an anonymous
	listIterator implementation (with some changes).
	* java/util/AbstractSequentialList.java: Throw messages with 
	IndexOutOfBoundsExceptions.
	(addAll): Add a specnote.
	* java/util/ArrayList.java (removeRange): Get the math right.
	(addAll): Increment modCount _before_ creating iterator.
	* java/util/LinkedList.java: Rewritten, mostly.

Index: AbstractList.java
===================================================================
RCS file: /cvs/java/libgcj/libjava/java/util/AbstractList.java,v
retrieving revision 1.2
diff -u -r1.2 AbstractList.java
--- AbstractList.java	2000/10/29 05:06:10	1.2
+++ AbstractList.java	2000/11/02 09:48:01
@@ -145,15 +145,22 @@
     return -1;
   }
 
+  /**
+   * Return an Iterator over this List. This implementation calls
+   * listIterator(0).
+   *
+   * @return an Iterator over this List
+   */
   public ListIterator listIterator()
   {
-    return new AbstractListItr(0);
+    return listIterator(0);
   }
 
   public ListIterator listIterator(int index)
   {
     if (index < 0 || index > size())
-      throw new IndexOutOfBoundsException();
+      throw new IndexOutOfBoundsException("Index: " + index + ", Size:" + 
+                                          size());
 
     return new AbstractListItr(index);
   }
@@ -193,10 +200,10 @@
     throw new UnsupportedOperationException();
   }
 
-  public List subList(final int fromIndex, final int toIndex)
+  public List subList(int fromIndex, int toIndex)
   {
     if (fromIndex > toIndex)
-      throw new IllegalArgumentException();
+      throw new IllegalArgumentException(fromIndex + " > " + toIndex);
     if (fromIndex < 0 || toIndex > size())
       throw new IndexOutOfBoundsException();
 
@@ -208,6 +215,7 @@
     private int knownMod = modCount;
     private int position;
     private int lastReturned = -1;
+    private int size = size();
 
     AbstractListItr(int start_pos)
     {
@@ -223,7 +231,7 @@
     public boolean hasNext()
     {
       checkMod();
-      return position < size();
+      return position < size;
     }
 
     public boolean hasPrevious()
@@ -235,7 +243,7 @@
     public Object next()
     {
       checkMod();
-      if (position < size())
+      if (position < size)
 	{
 	  lastReturned = position++;
 	  return get(lastReturned);
@@ -280,7 +288,7 @@
 	  throw new IllegalStateException();
 	}
       AbstractList.this.remove(lastReturned);
-      knownMod = modCount;
+      knownMod++;
       position = lastReturned;
       lastReturned = -1;
     }
@@ -298,7 +306,7 @@
       checkMod();
       AbstractList.this.add(position++, o);
       lastReturned = -1;
-      knownMod = modCount;
+      knownMod++;
     }
   }				// AbstractList.Iterator
 
@@ -311,7 +319,9 @@
     public SubList(AbstractList backing, int fromIndex, int toIndex)
     {
       backingList = backing;
-      upMod();
+      // FIXME: The `this' prefixes in this class are a workaround for a
+      // gcj bug. They should be removed later.
+      this.modCount = backingList.modCount;
       offset = fromIndex;
       size = toIndex - fromIndex;
     }
@@ -332,17 +342,6 @@
     }
 
     /**
-     * This method is called after every method that causes a structural
-     * modification to the backing list. It updates the local modCount field
-     * to match that of the backing list.
-     * Note that since this method is private, it will be inlined.
-     */
-    private void upMod()
-    {
-      this.modCount = backingList.modCount;
-    }
-
-    /**
      * This method checks that a value is between 0 and size (inclusive). If
      * it is not, an exception is thrown.
      * Note that since this method is private, it will be inlined.
@@ -352,7 +351,8 @@
     private void checkBoundsInclusive(int index)
     {
       if (index < 0 || index > size)
-	throw new IndexOutOfBoundsException();
+	throw new IndexOutOfBoundsException("Index: " + index + ", Size:" + 
+                                            size);
     }
 
     /**
@@ -365,7 +365,8 @@
     private void checkBoundsExclusive(int index)
     {
       if (index < 0 || index >= size)
-	throw new IndexOutOfBoundsException();
+	throw new IndexOutOfBoundsException("Index: " + index + ", Size:" + 
+                                            size);
     }
 
     public int size()
@@ -379,7 +380,6 @@
       checkMod();
       checkBoundsExclusive(index);
       o = backingList.set(index + offset, o);
-      upMod();
       return o;
     }
 
@@ -395,7 +395,7 @@
       checkMod();
       checkBoundsInclusive(index);
       backingList.add(index + offset, o);
-      upMod();
+      this.modCount++;
       size++;
     }
 
@@ -404,7 +404,7 @@
       checkMod();
       checkBoundsExclusive(index);
       Object o = backingList.remove(index + offset);
-      upMod();
+      this.modCount++;
       size--;
       return o;
     }
@@ -417,19 +417,122 @@
 
       // this call will catch the toIndex < fromIndex condition
       backingList.removeRange(offset + fromIndex, offset + toIndex);
-      upMod();
+      this.modCount = backingList.modCount;
       size -= toIndex - fromIndex;
     }
-
+    
     public boolean addAll(int index, Collection c)
     {
       checkMod();
       checkBoundsInclusive(index);
-      int s = backingList.size();
+      int csize = c.size();
       boolean result = backingList.addAll(offset + index, c);
-      upMod();
-      size += backingList.size() - s;
+      this.modCount = backingList.modCount;
+      size += csize;
       return result;
+    }
+    
+    public Iterator iterator()
+    {
+      return listIterator(0);
+    }
+
+    public ListIterator listIterator(final int index)
+    {      
+      checkMod();
+      checkBoundsInclusive(index);
+      
+      return new ListIterator() 
+      {
+        ListIterator i = backingList.listIterator(index + offset);
+        int position = index;
+        
+        public boolean hasNext()
+	{
+          checkMod();
+          return position < size;
+        }
+        
+        public boolean hasPrevious()
+	{
+          checkMod();
+          return position > 0;
+        }
+        
+        public Object next()
+	{
+          if (position < size)
+	    {
+              Object o = i.next();
+              position++;
+              return o;
+            }
+	  else
+            throw new NoSuchElementException();
+	}
+        
+        public Object previous()
+	{
+          if (position > 0)
+	    {
+              Object o = i.previous();
+              position--;
+              return o;
+            }
+	  else
+            throw new NoSuchElementException();
+        }
+        
+        public int nextIndex()
+	{
+          return offset + i.nextIndex();
+        }
+        
+        public int previousIndex()
+	{
+          return offset + i.previousIndex();
+        }
+
+        public void remove()
+	{
+          i.remove();
+	  SubList.this.modCount++;
+          size--;
+          position = nextIndex();
+        }
+        
+        public void set(Object o)
+	{
+          i.set(o);
+        }
+        
+        public void add(Object o)
+	{
+          i.add(o);
+	  SubList.this.modCount++;
+          size++;
+          position++;
+        }
+
+        // Here is the reason why the various modCount fields are mostly
+        // ignored in this wrapper listIterator.
+        // IF the backing listIterator is failfast, then the following holds:
+        //   Using any other method on this list will call a corresponding
+        //   method on the backing list *after* the backing listIterator
+        //   is created, which will in turn cause a ConcurrentModException
+        //   when this listIterator comes to use the backing one. So it is
+        //   implicitly failfast.
+        // If the backing listIterator is NOT failfast, then the whole of
+        //   this list isn't failfast, because the modCount field of the
+        //   backing list is not valid. It would still be *possible* to
+        //   make the iterator failfast wrt modifications of the sublist
+        //   only, but somewhat pointless when the list can be changed under
+        //   us.
+        // Either way, no explicit handling of modCount is needed.
+        // However modCount++ must be executed in add and remove, and size
+        // must also be updated in these two methods, since they do not go
+        // through the corresponding methods of the subList.
+      };
     }
-  }				// AbstractList.SubList
+  }  // AbstractList.SubList
 }
Index: AbstractSequentialList.java
===================================================================
RCS file: /cvs/java/libgcj/libjava/java/util/AbstractSequentialList.java,v
retrieving revision 1.2
diff -u -r1.2 AbstractSequentialList.java
--- AbstractSequentialList.java	2000/10/29 05:06:10	1.2
+++ AbstractSequentialList.java	2000/11/02 09:48:01
@@ -64,6 +64,12 @@
     i.add(o);
   }
 
+  /**
+   * @specnote The spec in the JDK1.3 online docs is wrong. The implementation
+   *           should not call next() to skip over new elements as they are
+   *           added, because iterator.add() should add new elements BEFORE
+   *           the cursor.
+   */
   public boolean addAll(int index, Collection c)
   {
     boolean modified = false;
@@ -81,7 +87,8 @@
   {
     ListIterator i = listIterator(index);
     if (index < 0 || index > size())
-      throw new IndexOutOfBoundsException();
+      throw new IndexOutOfBoundsException("Index: " + index + ", Size:" + 
+                                          size());
     return i.next();
   }
 
@@ -100,7 +107,8 @@
   {
     ListIterator i = listIterator(index);
     if (index < 0 || index > size())
-      throw new IndexOutOfBoundsException();
+      throw new IndexOutOfBoundsException("Index: " + index + ", Size:" + 
+                                          size());
     Object removed = i.next();
     i.remove();
     return removed;
@@ -110,7 +118,8 @@
   {
     ListIterator i = listIterator(index);
     if (index < 0 || index > size())
-      throw new IndexOutOfBoundsException();
+      throw new IndexOutOfBoundsException("Index: " + index + ", Size:" + 
+                                          size());
     Object old = i.next();
     i.set(o);
     return old;
Index: ArrayList.java
===================================================================
RCS file: /cvs/java/libgcj/libjava/java/util/ArrayList.java,v
retrieving revision 1.2
diff -u -r1.2 ArrayList.java
--- ArrayList.java	2000/10/29 05:06:10	1.2
+++ ArrayList.java	2000/11/02 09:48:01
@@ -187,7 +187,7 @@
     if (fromIndex != toIndex)
       {
 	System.arraycopy(data, toIndex, data, fromIndex, size - toIndex);
-	size -= (fromIndex - toIndex);
+	size -= (toIndex - fromIndex);
       }
   }
 
@@ -219,9 +219,9 @@
    */
   public boolean addAll(Collection c)
   {
+    modCount++;
     Iterator itr = c.iterator();
     int csize = c.size();
-    modCount++;
     ensureCapacity(size + csize);
     for (int pos = 0; pos < csize; pos++)
       {
@@ -240,13 +240,13 @@
    */
   public boolean addAll(int index, Collection c)
   {
-    Iterator itr = c.iterator();
-    int csize = c.size();
-
-    modCount++;
     if (index < 0 || index > size)
       throw new IndexOutOfBoundsException("Index: " + index + ", Size:" + 
                                           size);
+    modCount++;
+    Iterator itr = c.iterator();
+    int csize = c.size();
+
     ensureCapacity(size + csize);
     int end = index + csize;
     if (size > 0 && index != size)
Index: LinkedList.java
===================================================================
RCS file: /cvs/java/libgcj/libjava/java/util/LinkedList.java,v
retrieving revision 1.1
diff -u -r1.1 LinkedList.java
--- LinkedList.java	2000/08/27 22:06:44	1.1
+++ LinkedList.java	2000/11/02 09:48:01
@@ -7,7 +7,7 @@
 it under the terms of the GNU General Public License as published by
 the Free Software Foundation; either version 2, or (at your option)
 any later version.
- 
+
 GNU Classpath is distributed in the hope that it will be useful, but
 WITHOUT ANY WARRANTY; without even the implied warranty of
 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
@@ -30,16 +30,17 @@
 import java.io.ObjectOutputStream;
 import java.io.ObjectInputStream;
 import java.io.IOException;
+import java.lang.reflect.Array;
 
 // TO DO:
 // ~ Doc comment for the class.
 // ~ Doc comments for the non-list methods.
-// ~ Some commenting on the Backing API and other general implementation notes.
+// ~ other general implementation notes.
 
 /**
  * Linked list implementation of the List interface.
  */
-public class LinkedList extends AbstractSequentialList 
+public class LinkedList extends AbstractSequentialList
   implements Serializable, Cloneable
 {
   static final long serialVersionUID = 876323262645176354L;
@@ -49,7 +50,8 @@
    * previous field) of the list. The data field is null. If the list is empty,
    * both the head and the tail point to ends itself.
    */
-  transient Entry ends = new Entry();
+  transient Entry first;
+  transient Entry last;
 
   /**
    * The current length of the list.
@@ -59,260 +61,86 @@
   /**
    * Class to represent an entry in the list. Holds a single element.
    */
-  private static class Entry {
-
-    /**
-     * The list element.
-     */
-    Object data = null;
-
-    /**
-     * The next entry in the list. If this is the last entry in the list, the
-     * ends field of the list is held here.
-     */
+  private static class Entry
+  {
+    Object data;
     Entry next;
-
-    /**
-     * The previous entry in the list. If this is the first entry in the list,
-     * the ends field of the list is held here.
-     */
     Entry previous;
-
-    /**
-     * Create an entry with given data and linkage.
-     */
-    Entry(Object d, Entry n, Entry p) {
-      data = d;
-      next = n;
-      previous = p;
+    
+    Entry(Object data)
+    {
+      this.data = data;
     }
-
-    /**
-     * Create an entry with no data and linking to itself, for use as the ends
-     * field of the list.
-     */
-    Entry() {
-      next = previous = this;
-    }
-
-    /**
-     * Remove this entry.
-     */
-    Object remove() {
-      previous.next = next;
-      next.previous = previous;
-      return data;
-    }
-  }
-
-  private static interface Backing {
-    void checkMod(int known);
-    void upMod();
-    void incSize(int by);
-    void decSize(int by);
   }
-
-  private final Backing back = new Backing() {
-    public void checkMod(int known) {
-      if (known != modCount) {
-	throw new ConcurrentModificationException();
-      }
-    }
-    public void upMod() {
-      modCount++;
-    }
-    public void incSize(int by) {
-      size += by;
-    }
-    public void decSize(int by) {
-      size -= by;
-    }
-  };
-
-  /** A ListIterator over the list. This class keeps track of its
-   * position in the list, the size of the list, and the two list
-   * entries it is between.  This enables it to be used identically
-   * for both the list itself and a sublist of the list.
-   */
-  private static class Iter implements ListIterator {
-
-    /**
-     * The index of the element that will be returned by next().
-     */
-    int pos;
-
-    /**
-     * The size of the backing list.
-     */
-    int size;
-
-    /**
-     * The entry containing the element that will be returned by next().
-     */
-    Entry next;
-
-    /**
-     * The entry containing the element that will be returned by previous().
-     */
-    Entry previous;
-
-    /**
-     * The entry that will be affected by remove() or set().
-     */
-    Entry recent;
-
-    /**
-     * The known value of the modCount of the backing list.
-     */
-    int knownMod;
-
-    private final Backing b;
-
-    /**
-     * Create a new Iter starting at a given Entry within the list, at a given
-     * position, in a list of given size.
-     *
-     * @param index the index to begin iteration.
-     * @exception IndexOutOfBoundsException if index < 0 || index > size.
-     */
-    Iter(Backing backing, Entry n, int index, int s, int modCount) {
-      b = backing;
-      pos = index;
-      size = s;
-      next = n;
-      previous = n.previous;
-      knownMod = modCount;
-    }
-
-    public int nextIndex() {
-      b.checkMod(knownMod);
-      return pos;
-    }
-
-    public int previousIndex() {
-      b.checkMod(knownMod);
-      return pos - 1;
-    }
-
-    public boolean hasNext() {
-      b.checkMod(knownMod);
-      return pos < size;
-    }
-
-    public boolean hasPrevious() {
-      b.checkMod(knownMod);
-      return pos > 0;
-    }
-
-    public Object next() {
-      b.checkMod(knownMod);
-      if (pos >= size) {
-	throw new NoSuchElementException();
-      } else {
-	pos++;
-	recent = previous = next;
-	next = recent.next;
-	return recent.data;
-      }
-    }
-
-    public Object previous() {
-      b.checkMod(knownMod);
-      if (pos <= 0) {
-	throw new NoSuchElementException();
-      } else {
-	pos--;
-	recent = next = previous;
-	previous = recent.previous;
-	return recent.data;
-      }
-    }
-
-    public void remove() {
-      b.checkMod(knownMod);
-      if (recent == null) {
-	throw new IllegalStateException();
-      }
-
-      // Adjust the position to before the removed element
-      if (recent == previous) pos--;
-
-      // Could use recent.remove() but this way is quicker, and also correctly
-      // fixes next and previous.
-      next = recent.previous.next = recent.next;
-      previous = recent.next.previous = recent.previous;
-      size--;
-      b.decSize(1);
-      knownMod++;
-      b.upMod();
-      recent = null;
-    }
-
-    public void add(Object o) {
-      b.checkMod(knownMod);
-      previous.next = next.previous = new Entry(o, next, previous);
-
-      // New for 1.2RC1 - the semantics changed so that the iterator is
-      // positioned *after* the new element.
-      previous = previous.next;
-      pos++;
-
-      size++;
-      b.incSize(1);
-      knownMod++;
-      b.upMod();
-      recent = null;
-    }
-
-    public void set(Object o) {
-      b.checkMod(knownMod);
-      if (recent == null) {
-	throw new IllegalStateException();
-      }
-      recent.data = o;
-    }
-  }
-
+  
   /**
    * Obtain the Entry at a given position in a list. This method of course
    * takes linear time, but it is intelligent enough to take the shorter of the
    * paths to get to the Entry required. This implies that the first or last
    * entry in the list is obtained in constant time, which is a very desirable
    * property.
-   * For speed and flexibility in which ranges are valid, range checking is not
-   * done in this method, and if n is outside the range -1 <= n <= size, the
-   * result will be wrong (but no exception will be thrown).
-   * Note that you *can* obtain entries at position -1 and size, which are
-   * equal to prehead and posttail respectively.
-   * This method is static so that it can also be used in subList.
+   * For speed and flexibility, range checking is not done in this method:
+   * Incorrect values will be returned if (n < 0) or (n >= size).
    *
    * @param n the number of the entry to get.
-   * @param size the size of the list to get the entry in.
-   * @param head the entry before the first element of the list (usually ends).
-   * @param tail the entry after the last element of the list (usually ends).
    */
-  static Entry getEntry(int n, int size, Entry head, Entry tail) {
-
-    // n less than size/2, iterate from start
-    if (n < size >> 1) {
-      while (n-- >= 0) {
-	head = head.next;
+  private Entry getEntry(int n)
+  {
+    Entry e;
+    if (n < size / 2)
+      {
+        e = first;
+	// n less than size/2, iterate from start
+	while (n-- > 0)
+	  {
+	    e = e.next;
+	  }
       }
-      return head;
-
-    // n greater than size/2, iterate from end
-    } else {
-      while (++n <= size) {
-	tail = tail.previous;
+    else
+      {
+        e = last;      
+	// n greater than size/2, iterate from end
+	while (++n < size)
+	  {
+	    e = e.previous;
+	  }
+      }
+    return e;
+  }
+  
+  /** Remove an entry from the list. This will adjust size and deal with
+   *  `first' and  `last' appropriatly. It does not effect modCount, that is 
+   *  the responsibility of the caller. */
+  private void removeEntry(Entry e)
+  {
+    if (size == 1)
+      first = last = null;
+    else
+      {
+	if (e == first)
+	  {
+	    first = e.next;
+	    e.next.previous = null;
+	  }
+	else if (e == last)
+	  {
+	    last = e.previous;
+	    e.previous.next = null;
+	  }
+	else
+	  {
+	    e.next.previous = e.previous;	
+	    e.previous.next = e.next;
+	  }
       }
-      return tail;
-    }
+    size--;
   }
 
   /**
    * Create an empty linked list.
    */
-  public LinkedList() {
+  public LinkedList()
+  {
     super();
   }
 
@@ -322,7 +150,8 @@
    *
    * @param c the collection to populate this list from.
    */
-  public LinkedList(Collection c) {
+  public LinkedList(Collection c)
+  {
     super();
     // Note: addAll could be made slightly faster, but not enough so to justify
     // re-implementing it from scratch. It is just a matter of a relatively
@@ -330,84 +159,279 @@
     addAll(c);
   }
 
-  public Object getFirst() {
-    if (size == 0) {
+  public Object getFirst()
+  {
+    if (size == 0)
       throw new NoSuchElementException();
-    }
-    return ends.next.data;
+    return first.data;
   }
 
-  public Object getLast() {
-    if (size == 0) {
+  public Object getLast()
+  {
+    if (size == 0)
       throw new NoSuchElementException();
-    }
-    return ends.previous.data;
+    return last.data;
   }
 
-  public Object removeFirst() {
-    if (size == 0) {
+  public Object removeFirst()
+  {
+    if (size == 0)
       throw new NoSuchElementException();
-    }
     size--;
     modCount++;
-    return ends.next.remove();
+    Object r = first.data;
+    
+    if (first.next != null)
+      first.next.previous = null;
+    return r;
   }
 
-  public Object removeLast() {
-    if (size == 0) {
+  public Object removeLast()
+  {
+    if (size == 0)
       throw new NoSuchElementException();
-    }
     size--;
     modCount++;
-    return ends.previous.remove();
+    Object r = last.data;
+    
+    if (last.previous != null)
+      last.previous.next = null;    
+    return r;
   }
 
-  public void addFirst(Object o) {
-    ends.next.previous = ends.next = new Entry(o, ends.next, ends);
-    size++;
+  public void addFirst(Object o)
+  {
     modCount++;
+    Entry e = new Entry(o);
+    
+    if (size == 0)
+      first = last = e;
+    else
+      {
+	e.next = first;
+        first.previous = e;
+	first = e;
+      }    
+    size++;
   }
 
-  public void addLast(Object o) {
-    ends.previous.next = ends.previous = new Entry(o, ends, ends.previous);
-    size++;
+  public void addLast(Object o)
+  {
     modCount++;
+    addLastEntry(new Entry(o));
+  }
+  
+  private void addLastEntry(Entry e)
+  {
+    if (size == 0)
+      first = last = e;
+    else
+      {
+	e.previous = last;
+        last.next = e;
+	last = e;
+      }
+    size++;
   }
 
-  /**
-   * Obtain the number of elements currently in this list.
-   *
-   * @returns the number of elements currently in this list.
-   */
-  public int size() {
+  public boolean contains(Object o)
+  {
+    Entry e = first;
+    while (e != null)
+      {
+        if (e.data == null ? o == null : o.equals(e.data))
+	  return true;
+        e = e.next;
+      }
+    return false;
+  }
+
+  public int size()
+  {
     return size;
   }
+  
+  public boolean add(Object o)
+  {
+    modCount++;
+    addLastEntry(new Entry(o));
+    return true;
+  }
+  
+  public boolean remove(Object o)
+  {
+    modCount++;
+    Entry e = first;
+    while (e != null)
+      {
+        if (e.data == null ? o == null : o.equals(e.data))
+	  {
+	    removeEntry(e);
+	    return true;
+	  }
+        e = e.next;
+      }
+    return false;
+  }
 
-  /**
-   * Remove a range of elements from this list.
-   *
-   * @param fromIndex the index, inclusive, to remove from.
-   * @param toIndex the index, exclusive, to remove to.
-   * @exception IndexOutOfBoundsException if fromIndex > toIndex || fromIndex <
-   *   0 || toIndex > size().
-   */
-  // Note: normally removeRange is provided to allow efficient ways to
-  // implement clear() on subLists. However, in this case clear on subLists
-  // works anyway, so this implementation is included just for completeness
-  // and because subclasses might try to use it.
-  protected void removeRange(int fromIndex, int toIndex) {
-    subList(fromIndex, toIndex).clear();
+  public boolean addAll(Collection c)
+  {
+    return addAll(size, c);
   }
+  
+  public boolean addAll(int index, Collection c)
+  {
+    if (index < 0 || index > size)
+      throw new IndexOutOfBoundsException("Index: " + index + ", Size:" + 
+                                          size);
+    modCount++;
+    int csize = c.size();
 
-  /**
-   * Clear the list.
-   */
-  public void clear() {
-    ends.next = ends.previous = ends;
+    if (csize == 0)
+      return false;
+
+    Iterator itr = c.iterator();
+    
+    // Get the entries just before and after index. If index is at the start
+    // of the list, BEFORE is null. If index is at the end of thelist, AFTER is
+    // null. If the list is empty, both are null.
+    Entry after = null;
+    Entry before = null;    
+    if (index != size)
+      {
+	after = getEntry(index);
+	before = after.previous;
+      }
+    else
+      before = last;
+      
+    // Create the first new entry. We do not yet set the link from `before'
+    // to the first entry, in order to deal with the case where (c == this). 
+    // [Actually, we don't have to handle this case to fufill the 
+    // contract for addAll(), but Sun's implementation appears to.]
+    Entry e = new Entry(itr.next());
+    e.previous = before;
+    Entry prev = e;
+    Entry firstNew = e;
+    
+    // Create and link all the remaining entries.
+    for (int pos = 1; pos < csize; pos++)
+      {
+        e = new Entry(itr.next());
+	e.previous = prev;	
+	prev.next = e;
+	prev = e;
+      }
+    // Fix up the links between the last new entry and the following entry.
+    prev.next = after;
+    if (after != null)
+      after.previous = e;
+    else
+      last = e;
+    
+    if (before != null)
+      before.next = firstNew;
+    else
+      first = firstNew;
+    
+    size += csize;
+    return true;
+  }
+
+  public void clear()
+  {
     modCount++;
+    first = null;
+    last = null;
     size = 0;
   }
 
+  public Object get(int index)
+  {
+    if (index < 0 || index >= size)
+      throw new IndexOutOfBoundsException("Index: " + index + ", Size:" + 
+                                          size);
+    Entry e = getEntry(index);
+    return e.data;
+  }
+  
+  public Object set(int index, Object o)
+  {
+    if (index < 0 || index >= size)
+      throw new IndexOutOfBoundsException("Index: " + index + ", Size:" + 
+                                          size);
+    Entry e = getEntry(index);
+    Object old = e.data;
+    e.data = o;
+    return old;
+  }
+
+  public void add(int index, Object o)
+  {
+    if (index < 0 || index > size)
+      throw new IndexOutOfBoundsException("Index: " + index + ", Size:" + 
+                                          size);
+    modCount++;
+    addEntry(index, new Entry(o));    
+  }
+  
+  private void addEntry(int index, Entry e)
+  {
+    if (index < size)
+      {
+	Entry after = getEntry(index);
+	e.next = after;
+	e.previous = after.previous;
+	if (after.previous == null)
+	  first = e;
+	else
+	  after.previous.next = e;
+	after.previous = e;
+	size++;        
+      }
+    else
+      addLastEntry(e);
+  }
+  
+  public Object remove(int index)
+  {
+    if (index < 0 || index >= size)
+      throw new IndexOutOfBoundsException("Index: " + index + ", Size:" + 
+                                          size);
+    modCount++;
+    Entry e = getEntry(index);
+    removeEntry(e);
+    return e.data;
+  }
+  
+  public int indexOf(Object o)
+  {
+    int index = 0;
+    Entry e = first;
+    while (e != null)
+      {
+        if (e.data == null ? o == null : o.equals(e.data))
+	  return index;
+	++index;
+        e = e.next;
+      }
+    return -1;
+  }
+  
+  public int lastIndexOf(Object o)
+  {
+    int index = size - 1;
+    Entry e = last;
+    while (e != null)
+      {
+        if (e.data == null ? o == null : o.equals(e.data))
+	  return index;
+	--index;
+        e = e.previous;
+      }
+    return -1;  
+  }
+
   /**
    * Obtain a ListIterator over this list, starting at a given index. The
    * ListIterator returned by this method supports the add, remove and set
@@ -416,121 +440,13 @@
    * @param index the index of the element to be returned by the first call to
    *   next(), or size() to be initially positioned at the end of the list.
    * @exception IndexOutOfBoundsException if index < 0 || index > size().
-   */
-  public ListIterator listIterator(int index) {
-
-    // Check bounds
-    if (index < 0 || index > size) {
-      throw new IndexOutOfBoundsException();
-    }
-
-    return new Iter(back, getEntry(index, size, ends, ends), 
-		    index, size, modCount);
-  }
-
-  /**
-   * Obtain a List view of a subsection of this list, from fromIndex
-   * (inclusive) to toIndex (exclusive). The returned list is modifiable in
-   * every respect. Changes to the returned list are reflected in this list. If
-   * this list is structurally modified is any way other than through the
-   * returned list, any subsequent operations on the returned list will result
-   * in a ConcurrentModificationException (that is, the returned list is
-   * fail-fast).
-   *
-   * @param fromIndex the index that the returned list should start from
-   *    (inclusive).
-   * @param toIndex the index that the returned list should go to (exclusive).
-   * @returns a List backed by a subsection of this list.
-   * @exception IndexOutOfBoundsException if fromIndex < 0 || toIndex > size()
-   *   || fromIndex > toIndex.
    */
-  public List subList(int fromIndex, int toIndex) {
-
-    // Check bounds
-    if (fromIndex > toIndex || fromIndex < 0 || toIndex > size) {
-      throw new IndexOutOfBoundsException();
-    }
-
-    return new SubLinkedList(back, modCount,
-			     getEntry(fromIndex - 1, size, ends, ends),
-			     getEntry(toIndex, size, ends, ends),
-			     toIndex - fromIndex);
-  }
-
-  private static class SubLinkedList extends AbstractSequentialList {
-
-    Entry head; // entry before the beginning
-    Entry tail; // entry after the end
-    int size;
-    private final Backing b;
-
-    private final Backing back = new Backing() {
-      public void checkMod(int known) {
-	if (known != modCount) {
-	  throw new ConcurrentModificationException();
-	}
-      }
-      public void upMod() {
-	modCount++;
-      }
-      public void incSize(int by) {
-	size += by;
-      }
-      public void decSize(int by) {
-	size -= by;
-      }
-    };
-
-    SubLinkedList(Backing backing, int knownMod, Entry h, Entry t, int s) {
-      this.modCount = knownMod;
-      b = backing;
-      head = h;
-      tail = t;
-      size = s;
-    }
-
-    public int size() {
-      b.checkMod(this.modCount);
-      return size;
-    }
-
-    public ListIterator listIterator(int index) {
-      b.checkMod(this.modCount);
-
-      // Check bounds
-      if (index < 0 || index > size) {
-	throw new IndexOutOfBoundsException();
-      }
-
-      return new Iter(back, getEntry(index, size, head, tail), 
-		      index, size, modCount);
-    }
-
-    public void clear() {
-      b.checkMod(this.modCount);
-      head.next = tail;
-      tail.previous = head;
-      size = 0;
-      b.decSize(size);
-      modCount++;
-      b.upMod();
-    }
-
-    // No removeRange because this class cannot be publically subclassed.
-
-    public List subList(int fromIndex, int toIndex) {
-      b.checkMod(this.modCount);
-
-      // Check bounds
-      if (fromIndex > toIndex || fromIndex < 0 || toIndex > size) {
-	throw new IndexOutOfBoundsException();
-      }
-
-      return new SubLinkedList(back, this.modCount,
-			       getEntry(fromIndex - 1, size, head, tail),
-			       getEntry(toIndex, size, head, tail),
-			       toIndex - fromIndex);
-    }
+  public ListIterator listIterator(int index)
+  {
+    if (index < 0 || index > size)
+      throw new IndexOutOfBoundsException("Index: " + index + ", Size:" + 
+                                          size);
+    return new LinkedListItr(index);
   }
 
   /**
@@ -538,34 +454,60 @@
    * @return an object of the same class as this object, containing the
    * same elements in the same order.
    */
-  public Object clone() 
+  public Object clone()
   {
-    LinkedList copy;
+    LinkedList copy = null;
     try
       {
 	copy = (LinkedList) super.clone();
       }
     catch (CloneNotSupportedException ex)
       {
-	throw new InternalError(ex.getMessage());
       }
     copy.size = 0;
-    copy.ends = new Entry();
     copy.addAll(this);
     return copy;
   }
+  
+  public Object[] toArray()
+  {
+    Object[] array = new Object[size];
+    Entry e = first;
+    for (int i = 0; i < size; i++)
+      {
+        array[i] = e.data;
+        e = e.next;
+      }
+    return array;
+  }
+  
+  public Object[] toArray(Object[] array)
+  {
+    if (array.length < size)
+      array = (Object[]) Array.newInstance(array.getClass().getComponentType(), 
+        				   size);
+    else if (array.length > size)
+      array[size] = null;
+    Entry e = first;
+    for (int i = 0; i < size; i++)
+      {
+        array[i] = e.data;
+        e = e.next;
+      }
+    return array;  
+  }
 
   /**
    * Serialize an object to a stream.
    * @serialdata the size of the list (int), followed by all the elements
    * (Object) in proper order.
    */
-  private void writeObject(ObjectOutputStream s)
-    throws IOException
+  private void writeObject(ObjectOutputStream s) throws IOException
   {
     s.writeInt(size);
-    for (Iterator i = iterator(); i.hasNext(); )
-      s.writeObject(i.next());
+    Iterator itr = iterator();
+    for (int i = 0; i < size; i++)
+      s.writeObject(itr.next());
   }
 
   /**
@@ -577,8 +519,133 @@
     throws IOException, ClassNotFoundException
   {
     int serialSize = s.readInt();
-    ends = new Entry();
-    for (int i=0; i< serialSize; i++)
-      addLast(s.readObject());
+    for (int i = 0; i < serialSize; i++)
+      addLastEntry(new Entry(s.readObject()));
   }
+  
+  /** A ListIterator over the list. This class keeps track of its
+   * position in the list, the size of the list, and the two list
+   * entries it is between.  This enables it to be used identically
+   * for both the list itself and a sublist of the list.
+   */
+  class LinkedListItr implements ListIterator
+  {
+    int knownMod;
+    Entry next;         // entry that will be returned by next().
+    Entry previous;     // entry that will be returned by previous().
+    Entry lastReturned; // entry that will be affected by remove() or set().
+    int position;       // index of `next'.
+
+    /**
+     * Create a new Iter starting at a given Entry within the list, at a given
+     * position, in a list of given size.
+     */
+    LinkedListItr(int index)
+    {      
+      if (index == size)
+        {
+          next = null;
+	  previous = last;
+	}
+      else
+        {
+          next = getEntry(index);
+	  previous = next.previous;
+	}
+      position = index;
+      knownMod = modCount;
+    }
+
+    private void checkMod()
+    {
+      if (knownMod != modCount)
+	throw new ConcurrentModificationException();
+    }
+
+    public int nextIndex()
+    {
+      checkMod();
+      return position;
+    }
+
+    public int previousIndex()
+    {
+      checkMod();
+      return position - 1;
+    }
+
+    public boolean hasNext()
+    {
+      checkMod();
+      return (next != null);
+    }
+
+    public boolean hasPrevious()
+    {
+      checkMod();
+      return (previous != null);
+    }
+
+    public Object next()
+    {
+      checkMod();
+      if (next == null)
+	throw new NoSuchElementException();
+      position++;
+      lastReturned = previous = next;
+      next = lastReturned.next;
+      return lastReturned.data;
+    }
+
+    public Object previous()
+    {
+      checkMod();
+      if (previous == null)
+	throw new NoSuchElementException();
+      position--;
+      lastReturned = next = previous;
+      previous = lastReturned.previous;
+      return lastReturned.data;
+    }
+
+    public void remove()
+    {
+      checkMod();
+      if (lastReturned == null)
+	throw new IllegalStateException();
+
+      // Adjust the position to before the removed element, if the element
+      // being removed is behind the cursor.
+      if (lastReturned == previous)
+	position--;
+
+      next = lastReturned.next;
+      previous = lastReturned.previous;
+      // Because the list is being manipulated directly, there's no need to 
+      // touch either modCount or knownMod here.
+      removeEntry(lastReturned);
+      
+      lastReturned = null;
+    }
+
+    public void add(Object o)
+    {
+      checkMod();
+      // Because the list is being manipulated directly, there's no need to 
+      // touch either modCount or knownMod here.
+      Entry e = new Entry(o);
+      addEntry(position, e);
+      previous = e;
+      position++;
+      lastReturned = null;
+    }
+
+    public void set(Object o)
+    {
+      checkMod();
+      if (lastReturned == null)
+	throw new IllegalStateException();
+      lastReturned.data = o;
+    }
+  }  // class LinkedListItr  
 }

--------------9DE53B4993AF87E7337C78A0--



From abies@pg.gda.pl Thu Nov 02 05:18:16 2000
Received: from serwer.osi.gda.pl ([213.25.180.210] helo=lite.osi.gda.pl)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13rHRk-0006hp-00
	for <classpath@gnu.org>; Thu, 02 Nov 2000 05:18:16 -0500
Received: from pg.gda.pl (ip100.osi [192.168.1.100])
	by lite.osi.gda.pl (Postfix) with ESMTP id AE2D99201
	for <classpath@gnu.org>; Fri,  1 Jan 1999 20:06:19 +0100 (CET)
Message-ID: <3A013D3D.27437D72@pg.gda.pl>
Date: Thu, 02 Nov 2000 11:09:01 +0100
From: Artur Biesiadowski <abies@pg.gda.pl>
X-Mailer: Mozilla 4.75 [en] (Windows NT 5.0; U)
X-Accept-Language: pl,en
MIME-Version: 1.0
To: ClassPath List <classpath@gnu.org>
Subject: Re: PATCH: More collections List updates
References: <3A013C0E.5CFC7F02@albatross.co.nz>
Content-Type: text/plain; charset=iso-8859-2
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

I'm not sure if such constructs are best thing as far as readibility is
concerned.

lastReturned = previous = next;

Maybe splitting them up would be better. Or is it only me that find
myself puzzling half a minute which way it works ? :)

Artur


From bryce@albatross.co.nz Thu Nov 02 05:25:15 2000
Received: from waitaki.otago.ac.nz ([139.80.75.140])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13rHYX-0006nk-00
	for <classpath@gnu.org>; Thu, 02 Nov 2000 05:25:14 -0500
Received: from albatross.co.nz (IDENT:bryce@reason.otago.ac.nz [139.80.75.188])
	by waitaki.otago.ac.nz (8.9.3/8.9.3) with ESMTP id XAA15858;
	Thu, 2 Nov 2000 23:24:56 +1300
Message-ID: <3A014137.F490883@albatross.co.nz>
Date: Thu, 02 Nov 2000 23:25:59 +1300
From: Bryce McKinlay <bryce@albatross.co.nz>
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.4.0-test9 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Artur Biesiadowski <abies@pg.gda.pl>
CC: ClassPath List <classpath@gnu.org>
Subject: Re: PATCH: More collections List updates
References: <3A013C0E.5CFC7F02@albatross.co.nz> <3A013D3D.27437D72@pg.gda.pl>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

Artur Biesiadowski wrote:

> I'm not sure if such constructs are best thing as far as readibility is
> concerned.
>
> lastReturned = previous = next;
>
> Maybe splitting them up would be better. Or is it only me that find
> myself puzzling half a minute which way it works ? :)

I'm inclined to agree with you. (Actually, I think that particular line
came from the existing implementation). I'll change it.

regards

  [ bryce ]




From sballard@netreach.net Thu Nov 02 09:47:50 2000
Received: from fatuka.netreach.net ([207.29.195.121] helo=smtp.netreach.net)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13rLeg-00086O-00
	for <classpath@gnu.org>; Thu, 02 Nov 2000 09:47:50 -0500
Received: (qmail 9905 invoked by uid 102); 2 Nov 2000 14:48:15 -0000
Received: from rainbow.netreach.net (HELO netreach.net) (209.152.221.64)
  by smtp.netreach.net with SMTP; 2 Nov 2000 14:48:15 -0000
Message-ID: <3A017E8D.23468FAD@netreach.net>
Date: Thu, 02 Nov 2000 09:47:41 -0500
From: Stuart Ballard <sballard@netreach.net>
X-Mailer: Mozilla 4.72 [en] (X11; I; Linux 2.2.17 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Bryce McKinlay <bryce@albatross.co.nz>
CC: java-patches@sources.redhat.com, classpath@gnu.org
Subject: Re: PATCH: More collections List updates
References: <3A013C0E.5CFC7F02@albatross.co.nz>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

Bryce McKinlay wrote:
> 
> I have rewritten most of LinkedList. The new implementation is simpler,
> faster and conforms to the spec more closely.

As the original author of LinkedList - thanks! I think I was trying to
be too clever with my original code, resulting in an obfuscated mess.
Thanks for cleaning up after me :)

However, there is one feature of my implementation which I would like to
see preserved: In a naive implementation (including yours, and also
Sun's), every call to listIterator(i) on the result of
subList(fromIndex, toIndex) is O(fromIndex + i), because it goes through
AbstractList.SubList.listIterator(fromIndex). My implementation included
an override of LinkedList.subList() to store a pointer to the Entry
objects of the first-1 and last+1 nodes. Thus listIterator(i) on the
subList is only O(i). 

You can tell that Sun does not make this optimization because the
JavaDocs for LinkedList (as of 1.3) show the subList() method as being
inherited from AbstractList.

One way to implement this would be to make LinkedList.subList() itself
return a LinkedList (and set its first and last pointers accordingly).
In order to get the correct ConcurrentModification behavior, you may
need to make this a *subclass* of LinkedList, including the appropriate
modifications to modCount.

Now, obviously since Sun has this flaw in their implementation, I'm not
sure that we *need* to do this, but it does seem to me that we should
take every opportunity to be better than Sun's implementation where we
can.

> While working on this I
> realised the flaw in my previous change (removal of) the SubList
> iterator in AbstractList: Although correct, it resulted in poor
> performance for iterators on sublists of linked lists, as the iterator
> would ultimately traverse the linked list via its get() method,
> requiring quadratic-ish time. The fix was to restore the existing
> iterator wrapper approach. A few other minor bugs are also fixed by this
> patch.

Yes, this is the reason behind the original implementation (and it
applies even if you make the change above, because there could be other
"sequential" lists than LinkedList).

FWIW, I noticed recently that you changed AbstractList.iterator() to
simply return this.listIterator(). I agree with this change, but it's
worth mentioning that (at least as of the late betas of JDK1.2) Sun did
not agree with it, because I sent that very question to Sun before I
made the original implementation. With 2 years of hindsight, I don't
think I agree with their answer (which amounted to "users might assume
it's a ListIterator and cast to it, which would sort of be a security
hole because then they could do things they can't do with a regular
iterator" - I'll see whether I can dig up the message from the Sun guy)
but it is at least worth noting.

Stuart.


From bryce@albatross.co.nz Thu Nov 02 22:56:32 2000
Received: from waitaki.otago.ac.nz ([139.80.75.140])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13rXxv-0000TE-00
	for <classpath@gnu.org>; Thu, 02 Nov 2000 22:56:31 -0500
Received: from albatross.co.nz (IDENT:bryce@reason.otago.ac.nz [139.80.75.188])
	by waitaki.otago.ac.nz (8.9.3/8.9.3) with ESMTP id QAA18563;
	Fri, 3 Nov 2000 16:56:16 +1300
Message-ID: <3A0237A3.ED44678F@albatross.co.nz>
Date: Fri, 03 Nov 2000 16:57:23 +1300
From: Bryce McKinlay <bryce@albatross.co.nz>
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.4.0-test9 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: java-patches@sources.redhat.com, classpath@gnu.org
Subject: PATCH: More list updates
Content-Type: multipart/mixed;
 boundary="------------B2313F15BCFF523C2A1D2017"
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

This is a multi-part message in MIME format.
--------------B2313F15BCFF523C2A1D2017
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

This patch makes AbstractList.SubList a top-level private class in 0rder
to work around gcj's static inner class problems (arguably, this also
makes the code cleaner). It also fixes a bug in LinkedList: modification
counts must always be updated in the iterator add() and remove() methods
so that other iterators on the same list will detect concurrent
modification correctly against modifications made through the iterator
view.

This patch is against libgcj, but I'm checking the same into classpath.

regards

  [ bryce ]


--------------B2313F15BCFF523C2A1D2017
Content-Type: text/plain; charset=us-ascii;
 name="libgcj-collections-3.patch"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="libgcj-collections-3.patch"

2000-11-03  Bryce McKinlay  <bryce@albatross.co.nz>

	* java/util/AbstractList.java (SubList): Make it a top-level private
	class.
	* java/util/LinkedList.java (remove): Do update modCount and knownMod.
	(add): Ditto.
	* Makefile.am (ordinary_java_source_files): Add LinkedList.java.
	* Makefile.in: Rebuilt.

Index: java/util/AbstractList.java
===================================================================
RCS file: /cvs/java/libgcj/libjava/java/util/AbstractList.java,v
retrieving revision 1.4
diff -u -b -r1.4 AbstractList.java
--- AbstractList.java	2000/11/02 10:17:15	1.4
+++ AbstractList.java	2000/11/03 03:46:22
@@ -309,9 +309,12 @@
       knownMod++;
     }
   }				// AbstractList.Iterator
+}
 
-  static class SubList extends AbstractList
-  {
+
+
+class SubList extends AbstractList
+{
     private AbstractList backingList;
     private int offset;
     private int size;
@@ -319,9 +322,7 @@
     public SubList(AbstractList backing, int fromIndex, int toIndex)
     {
       backingList = backing;
-      // FIXME: The `this' prefixes in this class are a workaround for a
-      // gcj bug. They should be removed later.
-      this.modCount = backingList.modCount;
+    modCount = backingList.modCount;
       offset = fromIndex;
       size = toIndex - fromIndex;
     }
@@ -337,7 +338,7 @@
      */
     private void checkMod()
     {
-      if (this.modCount != backingList.modCount)
+    if (modCount != backingList.modCount)
 	throw new ConcurrentModificationException();
     }
 
@@ -496,8 +497,7 @@
         public void remove()
 	{
           i.remove();
-	  // FIXME: Uncomment the following line once the compiler is fixed.
-	  //SubList.this.modCount++;
+	modCount++;
           size--;
           position = nextIndex();
         }
@@ -510,8 +510,7 @@
         public void add(Object o)
 	{
           i.add(o);
-	  // FIXME: Uncomment the following line once the compiler is fixed.
-	  //SubList.this.modCount++;
+	modCount++;
           size++;
           position++;
         }
@@ -536,5 +535,4 @@
         // through the corresponding methods of the subList.
       };
     }
-  }  // AbstractList.SubList
-}
+}  // SubList
Index: java/util/LinkedList.java
===================================================================
RCS file: /cvs/java/libgcj/libjava/java/util/LinkedList.java,v
retrieving revision 1.2
diff -u -b -r1.2 LinkedList.java
--- LinkedList.java	2000/11/02 10:08:03	1.2
+++ LinkedList.java	2000/11/03 03:46:22
@@ -322,7 +322,7 @@
 	prev.next = e;
 	prev = e;
       }
-    // Fix up the links between the last new entry and the following entry.
+    // Link the new chain of entries into the list.
     prev.next = after;
     if (after != null)
       after.previous = e;
@@ -621,8 +621,8 @@
 
       next = lastReturned.next;
       previous = lastReturned.previous;
-      // Because the list is being manipulated directly, there's no need to 
-      // touch either modCount or knownMod here.
+      modCount++;
+      knownMod++;
       removeEntry(lastReturned);
       
       lastReturned = null;
@@ -631,11 +631,27 @@
     public void add(Object o)
     {
       checkMod();
-      // Because the list is being manipulated directly, there's no need to 
-      // touch either modCount or knownMod here.
+      modCount++;
+      knownMod++;
       Entry e = new Entry(o);
-      addEntry(position, e);
+      e.previous = previous;
+      e.next = next;
+
+      if (previous != null)
+	previous.next = e;
+      else
+	first = e;
+
+      if (next != null)
+        {
+	  next.previous = e;
+	  next = next.next;
+	}
+      else
+	last = e;
+
       previous = e;
+      size++;
       position++;
       lastReturned = null;
     }
Index: Makefile.am
===================================================================
RCS file: /cvs/java/libgcj/libjava/Makefile.am,v
retrieving revision 1.100
diff -u -b -r1.100 Makefile.am
--- Makefile.am	2000/11/02 20:33:05	1.100
+++ Makefile.am	2000/11/03 03:46:23
@@ -935,6 +935,7 @@
 java/util/HashMap.java \
 java/util/Hashtable.java \
 java/util/Iterator.java	\
+java/util/LinkedList.java \
 java/util/List.java \
 java/util/ListIterator.java \
 java/util/ListResourceBundle.java \

--------------B2313F15BCFF523C2A1D2017--



From bryce@albatross.co.nz Thu Nov 02 23:29:15 2000
Received: from waitaki.otago.ac.nz ([139.80.75.140])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13rYTa-00013Y-00
	for <classpath@gnu.org>; Thu, 02 Nov 2000 23:29:14 -0500
Received: from albatross.co.nz (IDENT:bryce@reason.otago.ac.nz [139.80.75.188])
	by waitaki.otago.ac.nz (8.9.3/8.9.3) with ESMTP id RAA18716;
	Fri, 3 Nov 2000 17:28:20 +1300
Message-ID: <3A023F28.CA48BD56@albatross.co.nz>
Date: Fri, 03 Nov 2000 17:29:29 +1300
From: Bryce McKinlay <bryce@albatross.co.nz>
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.4.0-test9 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Stuart Ballard <sballard@netreach.net>
CC: java-patches@sources.redhat.com, classpath@gnu.org
Subject: Re: PATCH: More collections List updates
References: <3A013C0E.5CFC7F02@albatross.co.nz> <3A017E8D.23468FAD@netreach.net>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

Stuart Ballard wrote:

> My implementation included
> an override of LinkedList.subList() to store a pointer to the Entry
> objects of the first-1 and last+1 nodes. Thus listIterator(i) on the
> subList is only O(i).

I'm not sure I agree with you about overriding subList() in LinkedList. Sure,
it would make the listIterator() call faster on a LinkedList-backed SubList,
but is it worth all the extra complexity and code duplication that is
necessary to do it? Note that your existing implementation of this was not
correct, because it did not add concurrent modification checks in all the
necessary places.

It seems like might be reasonable, however, to extend AbstractList.SubList
from LinkedList and only override its listIterator() method?  Unfortunatly I
seem to recall thinking there was some reason why this wouldn't work. But I
will look into it...

> FWIW, I noticed recently that you changed AbstractList.iterator() to
> simply return this.listIterator(). I agree with this change, but it's
> worth mentioning that (at least as of the late betas of JDK1.2) Sun did
> not agree with it, because I sent that very question to Sun before I
> made the original implementation.

Right. I don't really see a problem with it either, especially since all the
various other list implementations (SubList, LinkedList, etc) do exactly
that. Worth thinking about though, if we want to get every little detail
compatible with Sun's implementation. The original implementation returned
anonymous classes for the iterators and duplicated code between them, which I
didn't like, but now that I think about it we should be able to simply have
the ListIterator implementation extend the Iterator implementation, without
much extra code.

regards

  [ bryce ]




From warrenl@cygnus.com Fri Nov 03 03:12:49 2000
Received: from runyon.cygnus.com ([205.180.230.5] helo=cygnus.com)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13rbxx-0005cj-00
	for <classpath@gnu.org>; Fri, 03 Nov 2000 03:12:49 -0500
Received: from fencer.cygnus.com (fencer.cygnus.com [205.180.230.60])
	by runyon.cygnus.com (8.8.7-cygnus/8.8.7) with ESMTP id AAA27989
	for <classpath@gnu.org>; Fri, 3 Nov 2000 00:12:47 -0800 (PST)
Received: from localhost (warrenl@localhost) by fencer.cygnus.com (8.9.3/8.6.4) with ESMTP id AAA22918 for <classpath@gnu.org>; Fri, 3 Nov 2000 00:12:47 -0800
X-Authentication-Warning: fencer.cygnus.com: warrenl owned process doing -bs
Date: Fri, 3 Nov 2000 00:12:47 -0800 (PST)
From: Warren Levy <warrenl@cygnus.com>
To: classpath@gnu.org
Subject: Patch: Code to dump serialized objects
Message-ID: <Pine.LNX.4.10.10011030009010.4788-100000@fencer.cygnus.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

Folks,
I've instrumented java.io.ObjectInputStream to work as a serialized object
dumper by reworking the commented out debug statements in the module.

Now, if you call the static method
java.io.ObjectInputStream.setDump(true); you will see the contents
of the serialized object printed out for you as it is deserialized with
readObject().  This can be useful if you are trying to implement
serialization for a special case class and you're trying to understand
the (sometimes limited) Serialized Form doc better.

Enjoy!
--warrenl


2000-11-02  Warren Levy  <warrenl@cygnus.com>

	* java/io/ObjectInputStream.java (readObject): Added code to
	conditionally dump out the serialized data.
	Handle ENDBLOCKDATA case a bit more gracefully since the current
	behavior doesn't seem to work as expected.
	(readStreamHeader): Added code for serialized data dumper.
	(readNextBlock): Ditto.
	(readFields): Ditto.
	(dump): New private static field for turning on/off dumper.
	(setDump): New method.
	(dumpElement): New method.
	(dumpElementln): New method.


Index: java/io/ObjectInputStream.java
===================================================================
RCS file: /cvs/classpath/java/io/ObjectInputStream.java,v
retrieving revision 1.15
diff -u -p -r1.15 ObjectInputStream.java
--- java/io/ObjectInputStream.java	2000/10/06 00:03:07	1.15
+++ java/io/ObjectInputStream.java	2000/11/03 08:11:06
@@ -106,24 +106,30 @@ public class ObjectInputStream extends I
 
     this.isDeserializing = true;
 
-//    DEBUG ("MARKER ");
     byte marker = this.realInputStream.readByte ();
+    dumpElement ("MARKER: 0x" + Integer.toHexString(marker) + " ");
 
     switch (marker)
     {
       case TC_BLOCKDATA:
       case TC_BLOCKDATALONG:
+	if (marker == TC_BLOCKDATALONG) 
+	  dumpElementln ("BLOCKDATALONG");
+	else
+	  dumpElementln ("BLOCKDATA");
 	readNextBlock (marker);
 	throw new BlockDataException (this.blockDataBytes);
 
       case TC_NULL:
+	dumpElementln ("NULL");
 	ret_val = null;
 	break;
 
       case TC_REFERENCE:
       {
-//	DEBUG ("REFERENCE ");
+	dumpElement ("REFERENCE ");
 	Integer oid = new Integer (this.realInputStream.readInt ());
+	dumpElementln (Integer.toHexString(oid.intValue()));
 	ret_val = ((ObjectIdentityWrapper)
 		   this.objectLookupTable.get (oid)).object;
 	break;
@@ -131,6 +137,7 @@ public class ObjectInputStream extends I
 
       case TC_CLASS:
       {
+	dumpElementln ("CLASS");
 	ObjectStreamClass osc = (ObjectStreamClass)readObject ();
 	Class clazz = osc.forClass ();
 	assignNewHandle (clazz);
@@ -140,14 +147,15 @@ public class ObjectInputStream extends I
 
       case TC_CLASSDESC:
       {
-//	DEBUG ("CLASSDESC NAME ");
+	dumpElement ("CLASSDESC NAME=");
 	String name = this.realInputStream.readUTF ();
-//	DEBUG ("UID ");
+	dumpElement (name + "; UID=");
 	long uid = this.realInputStream.readLong ();
-//	DEBUG ("FLAGS ");
+	dumpElement (Long.toHexString(uid) + "; FLAGS=");
 	byte flags = this.realInputStream.readByte ();
-//	DEBUG ("FIELD COUNT ");
+	dumpElement (Integer.toHexString(flags) + "; FIELD COUNT=");
 	short field_count = this.realInputStream.readShort ();
+	dumpElementln (Short.toString(field_count));
 	ObjectStreamField[] fields = new ObjectStreamField[field_count];
 
 	ObjectStreamClass osc = new ObjectStreamClass (name, uid,
@@ -156,10 +164,11 @@ public class ObjectInputStream extends I
 
 	for (int i=0; i < field_count; i++)
 	{
-//	  DEBUG ("TYPE CODE ");
+	  dumpElement ("  TYPE CODE=");
 	  char type_code = (char)this.realInputStream.readByte ();
-//	  DEBUG ("FIELD NAME ");
+	  dumpElement (type_code + "; FIELD NAME=");
 	  String field_name = this.realInputStream.readUTF ();
+	  dumpElementln (field_name);
 	  String class_name;
 
 	  if (type_code == 'L' || type_code == '[')
@@ -177,9 +186,9 @@ public class ObjectInputStream extends I
 	osc.setClass (resolveClass (osc));
 	setBlockDataMode (false);
 
-//	DEBUG ("ENDBLOCKDATA ");
 	if (this.realInputStream.readByte () != TC_ENDBLOCKDATA)
 	  throw new IOException ("Data annotated to class was not consumed.");
+	dumpElementln ("ENDBLOCKDATA ");
 
 	osc.setSuperclass ((ObjectStreamClass)readObject ());
 	ret_val = osc;
@@ -188,27 +197,33 @@ public class ObjectInputStream extends I
 
       case TC_STRING:
       {
-//	DEBUG ("STRING ");
+	dumpElement ("STRING=");
 	String s = this.realInputStream.readUTF ();
+	dumpElementln (s);
 	ret_val = processResolution (s, assignNewHandle (s));
 	break;
       }
 
       case TC_ARRAY:
       {
+	dumpElementln ("ARRAY");
 	ObjectStreamClass osc = (ObjectStreamClass)readObject ();
 	Class componenetType = osc.forClass ().getComponentType ();
-//	DEBUG ("ARRAY LENGTH ");
+	dumpElement ("ARRAY LENGTH=");
 	int length = this.realInputStream.readInt ();
+	dumpElementln (length + "; COMPONENT TYPE=" + componenetType);
 	Object array = Array.newInstance (componenetType, length);
 	int handle = assignNewHandle (array);
 	readArrayElements (array, componenetType);
+	for (int i=0, len=Array.getLength(array); i < len; i++)
+	  dumpElementln ("  ELEMENT[" + i + "]=" + Array.get(array, i).toString());
 	ret_val = processResolution (array, handle);
 	break;
       }
 
       case TC_OBJECT:
       {
+	dumpElementln ("OBJECT");
 	ObjectStreamClass osc = (ObjectStreamClass)readObject ();
 	Class clazz = osc.forClass ();
 
@@ -284,8 +299,8 @@ public class ObjectInputStream extends I
 	{
 	  this.currentObjectStreamClass = hierarchy[i];
 
-//	  DEBUGln ("Reading fields of "
-//		   + this.currentObjectStreamClass.getName ());
+	  dumpElementln ("Reading fields of "
+		   + this.currentObjectStreamClass.getName ());
 
 	  has_read = true;
 
@@ -308,9 +323,23 @@ public class ObjectInputStream extends I
 
 	  if (has_read)
 	  {
-//	    DEBUG ("ENDBLOCKDATA? ");
-	    if (this.realInputStream.readByte () != TC_ENDBLOCKDATA)
-	      throw new IOException ("No end of block data seen for class with readObject (ObjectInputStream) method.");
+	    dumpElement ("ENDBLOCKDATA? ");
+	    try
+	      {
+		// FIXME: XXX: This try block is to catch EOF which is
+		// thrown for some objects.  That indicates a bug in the logic.
+	        if (this.realInputStream.readByte () != TC_ENDBLOCKDATA)
+		  throw new IOException ("No end of block data seen for class with readObject (ObjectInputStream) method.");
+	        dumpElementln ("yes");
+	      }
+	    catch (EOFException e)
+	      {
+	        dumpElementln ("no, got EOFException");
+	      }
+	    catch (IOException e)
+	      {
+	        dumpElementln ("no, got IOException");
+	      }
 	  }
 	}
 
@@ -321,13 +350,16 @@ public class ObjectInputStream extends I
       }
 
       case TC_RESET:
+	dumpElementln ("RESET");
 	clearHandles ();
 	ret_val = readObject ();
 	break;
 
       case TC_EXCEPTION:
       {
+	dumpElement ("EXCEPTION=");
 	Exception e = (Exception)readObject ();
+	dumpElementln (e.toString());
 	clearHandles ();
 	throw new WriteAbortedException ("Exception thrown during writing of stream", e);
       }
@@ -512,11 +544,11 @@ public class ObjectInputStream extends I
   protected void readStreamHeader ()
     throws IOException, StreamCorruptedException
   {
-//    DEBUG ("STREAM MAGIC ");
+    dumpElement ("STREAM MAGIC ");
     if (this.realInputStream.readShort () != STREAM_MAGIC)
       throw new StreamCorruptedException ("Invalid stream magic number");
 
-//    DEBUG ("STREAM VERSION ");
+    dumpElementln ("STREAM VERSION ");
     if (this.realInputStream.readShort () != STREAM_VERSION)
       throw new StreamCorruptedException ("Invalid stream version number");
   }
@@ -982,7 +1014,7 @@ public class ObjectInputStream extends I
 
   private void readNextBlock () throws IOException
   {
-//    DEBUG ("MARKER ");
+//  DEBUGln ("In readNextBlock ");
     readNextBlock (this.realInputStream.readByte ());
   }
 
@@ -991,13 +1023,15 @@ public class ObjectInputStream extends I
   {
     if (marker == TC_BLOCKDATA)
     {
-//      DEBUG ("BLOCK DATA SIZE ");
+      dumpElement ("BLOCK DATA SIZE=");
       this.blockDataBytes = this.realInputStream.readUnsignedByte ();
+      dumpElementln (Integer.toString(this.blockDataBytes));
     }
     else if (marker == TC_BLOCKDATALONG)
     {
-//      DEBUG ("BLOCK DATA LONG SIZE ");
+      dumpElement ("BLOCK DATA LONG SIZE=");
       this.blockDataBytes = this.realInputStream.readInt ();
+      dumpElementln (Integer.toString(this.blockDataBytes));
     }
     else
     {
@@ -1088,8 +1122,10 @@ public class ObjectInputStream extends I
 			   ObjectStreamClass stream_osc)
     throws ClassNotFoundException, IOException
   {
+//  DEBUGln ("In readFields");
     if (call_read_method)
     {
+//    DEBUGln ("  call_read_method is true");
       fieldsAlreadyRead = false;
       setBlockDataMode (true);
       callReadMethod (obj, stream_osc.forClass ());
@@ -1157,6 +1193,8 @@ public class ObjectInputStream extends I
       {
 	boolean value =
 	  default_initialize ? false : this.realInputStream.readBoolean ();
+	if (!default_initialize && set_value)
+	  dumpElementln ("  " + field_name + ": " + value);
 	if (set_value)
 	  setBooleanField (obj, field_name, value);
       }
@@ -1164,6 +1202,8 @@ public class ObjectInputStream extends I
       {
 	byte value =
 	  default_initialize ? 0 : this.realInputStream.readByte ();
+	if (!default_initialize && set_value)
+	  dumpElementln ("  " + field_name + ": " + value);
 	if (set_value)
 	  setByteField (obj, field_name, value);
       }
@@ -1171,6 +1211,8 @@ public class ObjectInputStream extends I
       {
 	char value =
 	  default_initialize ? (char)0 : this.realInputStream.readChar ();
+	if (!default_initialize && set_value)
+	  dumpElementln ("  " + field_name + ": " + value);
 	if (set_value)
 	  setCharField (obj, field_name, value);
       }
@@ -1178,6 +1220,8 @@ public class ObjectInputStream extends I
       {
 	double value =
 	  default_initialize ? 0 : this.realInputStream.readDouble ();
+	if (!default_initialize && set_value)
+	  dumpElementln ("  " + field_name + ": " + value);
 	if (set_value)
 	  setDoubleField (obj, field_name, value);
       }
@@ -1185,6 +1229,8 @@ public class ObjectInputStream extends I
       {
 	float value =
 	  default_initialize ? 0 : this.realInputStream.readFloat ();
+	if (!default_initialize && set_value)
+	  dumpElementln ("  " + field_name + ": " + value);
 	if (set_value)
 	  setFloatField (obj, field_name, value);
       }
@@ -1192,6 +1238,8 @@ public class ObjectInputStream extends I
       {
 	int value =
 	  default_initialize ? 0 : this.realInputStream.readInt ();
+	if (!default_initialize && set_value)
+	  dumpElementln ("  " + field_name + ": " + value);
 	if (set_value)
 	  setIntField (obj, field_name, value);
       }
@@ -1199,6 +1247,8 @@ public class ObjectInputStream extends I
       {
 	long value =
 	  default_initialize ? 0 : this.realInputStream.readLong ();
+	if (!default_initialize && set_value)
+	  dumpElementln ("  " + field_name + ": " + value);
 	if (set_value)
 	  setLongField (obj, field_name, value);
       }
@@ -1206,6 +1256,8 @@ public class ObjectInputStream extends I
       {
 	short value =
 	  default_initialize ? (short)0 : this.realInputStream.readShort ();
+	if (!default_initialize && set_value)
+	  dumpElementln ("  " + field_name + ": " + value);
 	if (set_value)
 	  setShortField (obj, field_name, value);
       }
@@ -1323,6 +1375,25 @@ public class ObjectInputStream extends I
   private boolean fieldsAlreadyRead;
   private Vector validators;
 
+  private static boolean dump;
+
+  public static void setDump (boolean dump)
+  {
+    this.dump = dump;
+  }
+
+  private void dumpElement (String msg);
+  {
+    if (dump)
+      Systm.out.print (msg);
+  }
+
+  private void dumpElementln (String msg);
+  {
+    if (dump)
+      Systm.out.println (msg);
+  }
+
   static
   {
     System.loadLibrary ("javaio");
@@ -1339,7 +1410,6 @@ public class ObjectInputStream extends I
   {
     System.out.println (msg);
   }
-
 }
 
 



From bryce@albatross.co.nz Fri Nov 03 03:26:04 2000
Received: from 210-54-227-153.adsl.xtra.co.nz ([210.54.227.153] helo=hokkaido)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13rcAl-0005rf-00
	for <classpath@gnu.org>; Fri, 03 Nov 2000 03:26:03 -0500
Received: from albatross.co.nz (IDENT:bryce@localhost [127.0.0.1])
	by hokkaido (8.11.0/8.11.0) with ESMTP id eA38R9V10015;
	Fri, 3 Nov 2000 21:27:09 +1300
Message-ID: <3A0276DC.E2F55E71@albatross.co.nz>
Date: Fri, 03 Nov 2000 21:27:08 +1300
From: Bryce McKinlay <bryce@albatross.co.nz>
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.2.15 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Warren Levy <warrenl@cygnus.com>
CC: Java Patch List <java-patches@sourceware.cygnus.com>, classpath@gnu.org
Subject: Re: Patch: Code to dump serialized objects
References: <Pine.LNX.4.10.10011022343150.4788-100000@fencer.cygnus.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

Warren Levy wrote:

> Folks,
> I've instrumented java.io.ObjectInputStream to work as a serialized object
> dumper by reworking the commented out debug statements in the module.
>
> Now, if you've configured with --enable-libgcj-debug and call the static
> method java.io.ObjectInputStream.setDump(true); you will see the contents
> of the serialized object printed out for you as it is deserialized with
> readObject().  This can be useful if you are trying to implement
> serialization for a special case class and you're trying to understand
> the (sometimes limited) Serialized Form doc better.
>
> I've made the code conditionally compiled this way to optimize things a
> little bit and to avoid this accidently being turned on in production
> code.  Enjoy!

Hi Warren,

This sounds cool! However, since we're trying to create a _compatible_ Java implementation, adding new public methods to the
standard classes is a bad idea!

Instead, I would suggest that this stuff gets activated via a system property (eg "gcj.dumpobjects") that can be set either by
calling "System.setProperty()" or by using the -D option at compile-time.

Even better, why not make a small utility (ala 'jcf-dump') that hooks into ObjectInputStream and gets installed with libgcj, ie
'ser-dump' or something?

regards

  [ bryce ]




From warrenl@cygnus.com Fri Nov 03 04:25:55 2000
Received: from runyon.cygnus.com ([205.180.230.5] helo=cygnus.com)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13rd6h-0007mv-00
	for <classpath@gnu.org>; Fri, 03 Nov 2000 04:25:55 -0500
Received: from fencer.cygnus.com (fencer.cygnus.com [205.180.230.60])
	by runyon.cygnus.com (8.8.7-cygnus/8.8.7) with ESMTP id BAA00960;
	Fri, 3 Nov 2000 01:25:53 -0800 (PST)
Received: from localhost (warrenl@localhost) by fencer.cygnus.com (8.9.3/8.6.4) with ESMTP id BAA22975; Fri, 3 Nov 2000 01:25:53 -0800
X-Authentication-Warning: fencer.cygnus.com: warrenl owned process doing -bs
Date: Fri, 3 Nov 2000 01:25:52 -0800 (PST)
From: Warren Levy <warrenl@cygnus.com>
To: Bryce McKinlay <bryce@albatross.co.nz>
cc: Java Patch List <java-patches@sourceware.cygnus.com>, classpath@gnu.org
Subject: Re: Patch: Code to dump serialized objects
In-Reply-To: <3A0276DC.E2F55E71@albatross.co.nz>
Message-ID: <Pine.LNX.4.10.10011030109570.4788-100000@fencer.cygnus.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

On Fri, 3 Nov 2000, Bryce McKinlay wrote:

> Warren Levy wrote:
> 
> > I've instrumented java.io.ObjectInputStream to work as a serialized object
> > dumper by reworking the commented out debug statements in the module.
>
> This sounds cool! However, since we're trying to create a _compatible_ Java implementation, adding new public methods to the
> standard classes is a bad idea!
>
> Instead, I would suggest that this stuff gets activated via a system property (eg "gcj.dumpobjects") that can be set either by
> calling "System.setProperty()" or by using the -D option at compile-time.

Yeah, I thought about the issue with adding a new public method but
hadn't considered about using system properties as a way to get around it.
Thanks!  I'll try to rework that next week, but I've got to take care
of some other responsibilities first (so I may not get to it for a little
bit).
 
> Even better, why not make a small utility (ala 'jcf-dump') that hooks into ObjectInputStream and gets installed with libgcj, ie
> 'ser-dump' or something?

I thought about this, but decided I preferred instrumenting the code this
way to get the benefit of being able to trace/debug as the object is being
deserialized.  I figured that there would be times when this is valuable
and just having a small utility that understood the format might not
always cut it (e.g. esp. in cases where the data is being massaged by the
object or with special read*() methods called directly from the
serializable class's readObject() ).

Also, I think I've run into a couple things in the
ObjectInputStream.readObject code that I don't think quite work right, so
I figured instrumenting it this way would be helpful in fixing those (as
well as helping out if/when there are additions to the Object
Serialization spec and new code has to be added/debugged/maintained).

I suppose it wouldn't hurt to have both.
--warrenl



From green@redhat.com Fri Nov 03 08:54:34 2000
Received: from runyon.cygnus.com ([205.180.230.5] helo=cygnus.com)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13rhIf-0001VK-00
	for <classpath@gnu.org>; Fri, 03 Nov 2000 08:54:33 -0500
Received: from locutus (locutus.cygnus.com [205.180.230.91])
	by runyon.cygnus.com (8.8.7-cygnus/8.8.7) with SMTP id FAA09491;
	Fri, 3 Nov 2000 05:54:25 -0800 (PST)
Received: by localhost with Microsoft MAPI; Fri, 3 Nov 2000 05:57:30 -0800
Message-ID: <01C0455A.F3F58160.green@redhat.com>
From: Anthony Green <green@redhat.com>
Reply-To: "green@cygnus.com" <green@cygnus.com>
To: "'Bryce McKinlay'" <bryce@albatross.co.nz>,
        Warren Levy
	 <warrenl@cygnus.com>
Cc: Java Patch List <java-patches@sourceware.cygnus.com>,
        "classpath@gnu.org" <classpath@gnu.org>
Subject: RE: Patch: Code to dump serialized objects
Date: Fri, 3 Nov 2000 05:56:22 -0800
Organization: Red Hat, Inc.
X-Mailer: Microsoft Internet E-mail/MAPI - 8.0.0.4211
Encoding: 19 TEXT
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

Bryce wrote:
> Instead, I would suggest that this stuff gets activated via
> a system property (eg "gcj.dumpobjects") that can be set either by
> calling "System.setProperty()" or by using the -D option at compile-time.
>
> Even better, why not make a small utility (ala 'jcf-dump') that hooks
> into ObjectInputStream and gets installed with libgcj, ie
> 'ser-dump' or something?

Or we could build two versions of ObjectInputStream, with and without 
instrumentation.  The normal one goes in libgcj.so and the instrumented one 
goes in ser-dump.so, so under Linux at least, you just run MyProgram like 
so....

$ LD_PRELOAD=ser-dump.so MyProgram

...to get the new behaviour with zero overhead otherwise.

AG



From jewel@pixie.co.za Wed Nov 15 16:18:31 2000
Received: from jhb-proxy.mweb.co.za ([196.2.48.243])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13w9ws-00039R-00
	for <classpath@gnu.org>; Wed, 15 Nov 2000 16:18:31 -0500
Received: from pta-dial-196-31-184-65.mweb.co.za ([196.31.184.65])
 by jhb-proxy.mweb.co.za
 (Sun Internet Mail Server sims.3.5.2000.03.23.18.03.p10)
 with ESMTP id <0G43008EI4EC1G@jhb-proxy.mweb.co.za> for classpath@gnu.org;
 Wed, 15 Nov 2000 23:16:34 +0200 (GMT-2)
Date: Wed, 15 Nov 2000 21:14:51 +0000 (UTC)
From: John Leuner <jewel@pixie.co.za>
Subject: java.util.zip
In-reply-to: <20000904132636.B25288@urbanophile.com>
X-Sender: jewel@skami.mweb.co.za
To: classpath@gnu.org
Message-id: <Pine.LNX.4.21.0011152111100.5735-100000@skami.mweb.co.za>
MIME-version: 1.0
Content-type: TEXT/PLAIN; charset=US-ASCII
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

I've started a port of the zlib code to java.

This is a non-trivial exercise and I don't want to promise that I'll
finish it.

So far I've written a lot of the inflation and huffman tree building code,
but none of it works yet.

If anyone wants to help out with the zlib engine stuff (or with the Java
wrapper classes, java.util.jar in particular), feel free to contact me.

I assume that if I do finish this I'll have to sign some paperwork, could
someone give me advice on this?

John Leuner



From arenn@shell-1.enteract.com Wed Nov 15 16:29:14 2000
Received: from mail.enteract.com ([207.229.143.33])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13wA7G-0003eJ-00
	for <classpath@gnu.org>; Wed, 15 Nov 2000 16:29:14 -0500
Received: from shell-1.enteract.com (arenn@shell-1.enteract.com [207.229.143.40])
	by mail.enteract.com (8.9.3/8.9.3) with ESMTP id PAA28875;
	Wed, 15 Nov 2000 15:29:11 -0600 (CST)
	(envelope-from arenn@shell-1.enteract.com)
Received: (from arenn@localhost)
	by shell-1.enteract.com (8.9.3/8.9.3) id PAA25323;
	Wed, 15 Nov 2000 15:29:09 -0600 (CST)
	(envelope-from arenn)
Date: Wed, 15 Nov 2000 15:29:09 -0600
From: "Aaron M. Renn" <arenn@urbanophile.com>
To: John Leuner <jewel@pixie.co.za>
Cc: classpath@gnu.org
Subject: Re: java.util.zip
Message-ID: <20001115152909.B24908@urbanophile.com>
Mail-Followup-To: John Leuner <jewel@pixie.co.za>, classpath@gnu.org
References: <20000904132636.B25288@urbanophile.com> <Pine.LNX.4.21.0011152111100.5735-100000@skami.mweb.co.za>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Mailer: Mutt 1.0pre3i
In-Reply-To: <Pine.LNX.4.21.0011152111100.5735-100000@skami.mweb.co.za>
Organization: GNU's Not Unix!
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

John Leuner (jewel@pixie.co.za) wrote:
> I've started a port of the zlib code to java.
> 
> This is a non-trivial exercise and I don't want to promise that I'll
> finish it.
> 
> So far I've written a lot of the inflation and huffman tree building code,
> but none of it works yet.
> 
> If anyone wants to help out with the zlib engine stuff (or with the Java
> wrapper classes, java.util.jar in particular), feel free to contact me.
> 
> I assume that if I do finish this I'll have to sign some paperwork, could
> someone give me advice on this?

Hi.  We've merged with libgcj over at Cygnus and they have already
written this package.  What we need to do is port it over to Classpath.
The problem is that we still have some questions outstanding on how to
handle native code, so this package is in a bit of limbo, but you
should not write a new implementation from scratch.

-- 
Aaron M. Renn (arenn@urbanophile.com) http://www.urbanophile.com/arenn/


From tromey@creche.cygnus.com Wed Nov 15 17:00:33 2000
Received: from gasherbrumii.peakpeak.com ([207.174.178.17] helo=gash2.peakpeak.com)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13wAbY-0004Dj-00
	for <classpath@gnu.org>; Wed, 15 Nov 2000 17:00:33 -0500
Received: from creche.cygnus.com (06-088.026.popsite.net [64.24.106.88])
	by gash2.peakpeak.com (8.9.3/8.9.3) with ESMTP id PAA11468;
	Wed, 15 Nov 2000 15:00:03 -0700
Received: (from tromey@localhost)
	by creche.cygnus.com (8.9.3/8.9.3) id PAA19749;
	Wed, 15 Nov 2000 15:06:11 -0700
To: "Aaron M. Renn" <arenn@urbanophile.com>
Cc: John Leuner <jewel@pixie.co.za>, classpath@gnu.org
Subject: Re: java.util.zip
References: <20000904132636.B25288@urbanophile.com> <Pine.LNX.4.21.0011152111100.5735-100000@skami.mweb.co.za> <20001115152909.B24908@urbanophile.com>
X-Zippy:  I'm not available for comment..
X-Attribution:  Tom
Reply-To: tromey@cygnus.com
From: Tom Tromey <tromey@cygnus.com>
Date: 15 Nov 2000 15:06:10 -0700
In-Reply-To: "Aaron M. Renn"'s message of "Wed, 15 Nov 2000 15:29:09 -0600"
Message-ID: <87d7fw99gt.fsf@creche.cygnus.com>
Lines: 13
X-Mailer: Gnus v5.7/Emacs 20.5
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

Aaron> What we need to do is port it over to Classpath.  The problem
Aaron> is that we still have some questions outstanding on how to
Aaron> handle native code, so this package is in a bit of limbo, but
Aaron> you should not write a new implementation from scratch.

I thought someone had ported the java.util.zip native code from CNI to
JNI already.  We could just check that in as a temporary measure until
the CNI/JNI integration problem is solved.

I've been looking at the integration problem lately.  I don't have
anything ready to show yet.

Tom


From information@technojobs.co.uk Wed Nov 15 17:05:30 2000
Received: from node-1.minx.net.uk ([212.85.249.131] helo=clustermail.minx.net.uk)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13wAgM-0004U3-00
	for <classpath@gnu.org>; Wed, 15 Nov 2000 17:05:30 -0500
Received: from [212.85.230.157] (helo=freecall775252)
	by clustermail.minx.net.uk with smtp (Exim 3.15 #2)
	id 13w9wa-0006le-00
	for classpath@gnu.org; Wed, 15 Nov 2000 21:18:12 +0000
From: Technojobs.co.uk<information@technojobs.co.uk>
To: classpath@gnu.org
Subject: Search the recruitment agencies for technical jobs on-line.
X-Mailer: Technojobs
Reply-To: information@technojobs.co.uk
Date: Wed, 15 Nov 2000 21:17:31 -0000
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Message-Id: <E13w9wa-0006le-00@clustermail.minx.net.uk>
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

Search the recruitment agencies for technical jobs on-line.

at:       http://www.technojobs.co.uk

Hundreds of new technical jobs posted every day...............

Technojobs New Advanced Site Now On-Line

Dear Job Seeker

Searching the recruitment agencies for technical jobs on-line has never been easier; our new site interface is now on line. Your new look new feel technojobs has been
updated with some top job-hunting tools. www.technojobs.co.uk

Search The Agencies
Our new advanced intelligent search tool is fast as it is accurate. Once you have used our CV centre to register your CV, simply click the link in a job add, and technojobs will automatically send your CV with a covering letter directly to that agency in reply to their job add.

CV Centre
Upload your CV via our secure site, logon at any time with your personal ID to update your CV, Activate it de-activate it, or if you're in work and looking to further your career use the Head Hunter Mode.

Jobs By Email
Our jobs by email keyword filter system, now employees the same technology that drives our main search.

Agencies
Technojobs hosts jobs for over 400 agencies, including some of the world top names in IT recruitment. We receive around 300 new technology jobs a day. If you have registered your CV with our CV centre then our agencies can search for you, we also distribute your CV when activated to agencies automatically, what's more your CV is scanned and sent to consultants that are specifically looking for candidates with your skill set.

Technojobs.co.uk Limited

Remove yourself from this mail list
Send a blank email to mailto:remove_097463@technojobs.co.uk , and place the word remove in the subject field.


From mark@klomp.org Wed Nov 15 18:29:28 2000
Received: from node1c5f.a2000.nl ([62.108.28.95] helo=nescio.klomp.org)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13wBzc-00077H-00
	for <classpath@gnu.org>; Wed, 15 Nov 2000 18:29:28 -0500
Received: from mark by nescio.klomp.org with local (Exim 3.16 #1 (Debian))
	id 13wBzO-0004ka-00; Thu, 16 Nov 2000 00:29:14 +0100
Date: Thu, 16 Nov 2000 00:29:14 +0100
From: Mark Wielaard <mark@klomp.org>
To: Tom Tromey <tromey@cygnus.com>
Cc: "Aaron M. Renn" <arenn@urbanophile.com>,
	John Leuner <jewel@pixie.co.za>, classpath@gnu.org
Subject: Re: java.util.zip
Message-ID: <20001116002914.A18179@klomp.org>
References: <20000904132636.B25288@urbanophile.com> <Pine.LNX.4.21.0011152111100.5735-100000@skami.mweb.co.za> <20001115152909.B24908@urbanophile.com> <87d7fw99gt.fsf@creche.cygnus.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5i
In-Reply-To: <87d7fw99gt.fsf@creche.cygnus.com>; from tromey@cygnus.com on Wed, Nov 15, 2000 at 03:06:10PM -0700
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

Hi,

On Wed, Nov 15, 2000 at 03:06:10PM -0700, Tom Tromey wrote:
> Aaron> What we need to do is port it over to Classpath.  The problem
> Aaron> is that we still have some questions outstanding on how to
> Aaron> handle native code, so this package is in a bit of limbo, but
> Aaron> you should not write a new implementation from scratch.
> 
> I thought someone had ported the java.util.zip native code from CNI to
> JNI already.  We could just check that in as a temporary measure until
> the CNI/JNI integration problem is solved.

Jochen Hoenicke has made a CNI to JNI conversion that can be found at
<http://www.informatik.uni-oldenburg.de/~delwi/classpath/JCL/>
We could check that in to Classpath as a temporary solution.
I have been adding Javadoc to the java.util.zip java files from
libgcj but have not yet checked that in the libgcj tree.
(It is not completely finished.) There is now also a java.util.jar
implementation in Classpath. But it does not contain a jar file/Manifest
verifier yet.

If you could look at the current libgcj implementation, Jopchen his native
code conversion and the jar implementation of Classpath and compare that
to what you have that would be greatly appreciated.

Cheers,

Mark


From tromey@creche.cygnus.com Wed Nov 15 18:44:17 2000
Received: from gasherbrumii.peakpeak.com ([207.174.178.17] helo=gash2.peakpeak.com)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13wCDw-0007RO-00
	for <classpath@gnu.org>; Wed, 15 Nov 2000 18:44:16 -0500
Received: from creche.cygnus.com (06-088.026.popsite.net [64.24.106.88])
	by gash2.peakpeak.com (8.9.3/8.9.3) with ESMTP id QAA28385;
	Wed, 15 Nov 2000 16:43:57 -0700
Received: (from tromey@localhost)
	by creche.cygnus.com (8.9.3/8.9.3) id QAA20241;
	Wed, 15 Nov 2000 16:50:04 -0700
To: Mark Wielaard <mark@klomp.org>
Cc: "Aaron M. Renn" <arenn@urbanophile.com>, John Leuner <jewel@pixie.co.za>,
        classpath@gnu.org
Subject: Re: java.util.zip
References: <20000904132636.B25288@urbanophile.com> <Pine.LNX.4.21.0011152111100.5735-100000@skami.mweb.co.za> <20001115152909.B24908@urbanophile.com> <87d7fw99gt.fsf@creche.cygnus.com> <20001116002914.A18179@klomp.org>
X-Zippy:  It's a lot of fun being alive...  I wonder if my bed is made?!?
X-Attribution:  Tom
Reply-To: tromey@cygnus.com
From: Tom Tromey <tromey@cygnus.com>
Date: 15 Nov 2000 16:50:04 -0700
In-Reply-To: Mark Wielaard's message of "Thu, 16 Nov 2000 00:29:14 +0100"
Message-ID: <87r94c7q37.fsf@creche.cygnus.com>
Lines: 20
X-Mailer: Gnus v5.7/Emacs 20.5
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

Mark> I have been adding Javadoc to the java.util.zip java files from
Mark> libgcj but have not yet checked that in the libgcj tree.
Mark> (It is not completely finished.)

Don't feel like you have to have it completely finished to do a
commit.  Any additions to the javadoc comments are worth committing.

Mark> If you could look at the current libgcj implementation, Jopchen
Mark> his native code conversion and the jar implementation of
Mark> Classpath and compare that to what you have that would be
Mark> greatly appreciated.

I definitely don't have time to do that :-(

However all changes to java.util.zip should be documented in the
libgcj ChangeLog.  I imagine the .java files can simply be copied from
the libgcj repository, which means you only have to check for changes
to the .cc files.

Tom


From mark@klomp.org Wed Nov 15 19:18:35 2000
Received: from node1c5f.a2000.nl ([62.108.28.95] helo=nescio.klomp.org)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13wCl8-0008JF-00
	for <classpath@gnu.org>; Wed, 15 Nov 2000 19:18:34 -0500
Received: from mark by nescio.klomp.org with local (Exim 3.16 #1 (Debian))
	id 13wCky-0004u4-00; Thu, 16 Nov 2000 01:18:24 +0100
Date: Thu, 16 Nov 2000 01:18:24 +0100
From: Mark Wielaard <mark@klomp.org>
To: Tom Tromey <tromey@cygnus.com>
Cc: John Leuner <jewel@pixie.co.za>, classpath@gnu.org
Subject: Re: java.util.zip
Message-ID: <20001116011824.A18800@klomp.org>
References: <20000904132636.B25288@urbanophile.com> <Pine.LNX.4.21.0011152111100.5735-100000@skami.mweb.co.za> <20001115152909.B24908@urbanophile.com> <87d7fw99gt.fsf@creche.cygnus.com> <20001116002914.A18179@klomp.org> <87r94c7q37.fsf@creche.cygnus.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5i
In-Reply-To: <87r94c7q37.fsf@creche.cygnus.com>; from tromey@cygnus.com on Wed, Nov 15, 2000 at 04:50:04PM -0700
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

Hi,

On Wed, Nov 15, 2000 at 04:50:04PM -0700, Tom Tromey wrote:
> Mark> I have been adding Javadoc to the java.util.zip java files from
> Mark> libgcj but have not yet checked that in the libgcj tree.
> Mark> (It is not completely finished.)
> 
> Don't feel like you have to have it completely finished to do a
> commit.  Any additions to the javadoc comments are worth committing.
Yes you are right. Will do.

> Mark> If you could look at the current libgcj implementation, Jopchen
> Mark> his native code conversion and the jar implementation of
> Mark> Classpath and compare that to what you have that would be
> Mark> greatly appreciated.
> 
> I definitely don't have time to do that :-(
Sorry, wrong To: field. I was trying to encourage the original poster,
John Leuner, to please help use with reviewing the current code even
if that means that we won't use his implementation right away.
Now that he has looked into zip, jar and zlib for his own implementation
plans he would be great to have as a reviewer of the current code.

Cheers,

Mark


From tom_evo2000@yahoo.com Wed Nov 15 21:08:12 2000
Received: from ip196.reno.quik.com ([209.213.147.196] helo=ip186.reno.quik.com)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13wETD-0002mp-00; Wed, 15 Nov 2000 21:08:12 -0500
From: tom_evo2000@yahoo.com
To: golfer<>
Subject: Free 10.5 DEGREE EVO golf driver
Date: Wed, 15 Nov 2000 18:06:30 -0600
X-Sender: tom_evo2000@yahoo.com
X-Mailer: QUALCOMM Windows Eudora Pro Version 4.1
Content-Type: text/plain; charset="us-ascii"
X-Priority: 3
X-MSMail-Priority: Normal
Message-Id: <E13wETD-0002mp-00@fencepost.gnu.org>
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

The United States Golf Institute is conducting a series of Golf clinics in Arizona this week. 

Everyone in attendance will receive the Evolution Golf 10.5 degree driver as a Free Gift for attending. 

To get all the details on this exciting Golf Clinic and to insure you get your Free Driver click here:

                                       http:/www.geocities.com/george_evg/index.html

                                              or call toll free          800 319 7266













We adhere to responsible mailing rules and ethics 
to be removed from this list please hit reply and type remove in the subject box.





From jewel@pixie.co.za Thu Nov 16 10:59:10 2000
Received: from jhb-proxy.mweb.co.za ([196.2.48.243])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13wRRO-0001jK-00
	for <classpath@gnu.org>; Thu, 16 Nov 2000 10:59:10 -0500
Received: from pta-dial-196-31-187-124.mweb.co.za ([196.31.187.124])
 by jhb-proxy.mweb.co.za
 (Sun Internet Mail Server sims.3.5.2000.03.23.18.03.p10)
 with ESMTP id <0G44006PXK6UIK@jhb-proxy.mweb.co.za> for classpath@gnu.org;
 Thu, 16 Nov 2000 17:54:32 +0200 (GMT-2)
Date: Thu, 16 Nov 2000 15:54:01 +0000 (UTC)
From: John Leuner <jewel@pixie.co.za>
Subject: Re: java.util.zip
In-reply-to: <20001116002914.A18179@klomp.org>
X-Sender: jewel@skami.mweb.co.za
To: classpath@gnu.org
Message-id: <Pine.LNX.4.21.0011161550320.5735-100000@skami.mweb.co.za>
MIME-version: 1.0
Content-type: TEXT/PLAIN; charset=US-ASCII
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

> 
> On Wed, Nov 15, 2000 at 03:06:10PM -0700, Tom Tromey wrote:
> > Aaron> What we need to do is port it over to Classpath.  The problem
> > Aaron> is that we still have some questions outstanding on how to
> > Aaron> handle native code, so this package is in a bit of limbo, but
> > Aaron> you should not write a new implementation from scratch.
> > 
> > I thought someone had ported the java.util.zip native code from CNI to
> > JNI already.  We could just check that in as a temporary measure until
> > the CNI/JNI integration problem is solved.
> 
> Jochen Hoenicke has made a CNI to JNI conversion that can be found at
> <http://www.informatik.uni-oldenburg.de/~delwi/classpath/JCL/>
> We could check that in to Classpath as a temporary solution.
> I have been adding Javadoc to the java.util.zip java files from
> libgcj but have not yet checked that in the libgcj tree.
> (It is not completely finished.) There is now also a java.util.jar
> implementation in Classpath. But it does not contain a jar file/Manifest
> verifier yet.
> 
> If you could look at the current libgcj implementation, Jopchen his native
> code conversion and the jar implementation of Classpath and compare that
> to what you have that would be greatly appreciated.

I should have made it more clear that this is a port to pure java code. My
hidden agenda here is to use the code in a future java OS (see
http://jos.org), which is why I didn't want to use zlib.

I will have a look at the libgcj code and Jochen's code, to see what I can
borrow / contribute in terms of code and javadoc.

John Leuner



From cam_evo@yahoo.com Thu Nov 16 15:00:14 2000
Received: from dialup-216-222-34-33.reno.rmci.net ([216.222.34.33])
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13wVCf-0008HT-00; Thu, 16 Nov 2000 15:00:14 -0500
From: cam_evo@yahoo.com
To: owner<>
Subject: FREE EVO 10.5 Degree Golf Driver
Date: Thu, 16 Nov 2000 11:58:43 -0600
X-Sender: cam_evo@yahoo.com
X-Mailer: QUALCOMM Windows Eudora Pro Version 4.1
Content-Type: text/plain; charset="us-ascii"
X-Priority: 3
X-MSMail-Priority: Normal
Message-Id: <E13wVCf-0008HT-00@fencepost.gnu.org>
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

The United States Golf Institute is conducting a series of Golf clinics in Arizona this week. 

Everyone in attendance will receive the Evolution Golf 10.5 degree driver as a Free Gift for attending. 

To get all the details on this exciting Golf Clinic and to insure you get your Free Driver click here:

                                       http:/www.geocities.com/george_evg/index.html

                                              or call toll free          800 319 7266













We adhere to responsible mailing rules and ethics 
to be removed from this list please hit reply and type remove in the subject box.





From jim_evo2000@yahoo.com Thu Nov 16 18:22:09 2000
Received: from ip177.reno.quik.com ([209.213.147.177])
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13wYM4-0005IF-00; Thu, 16 Nov 2000 18:22:08 -0500
From: jim_evo2000@yahoo.com
To: golfer<>
Subject: FREE EVO 10.5 DEGREE Golf Driver
Date: Thu, 16 Nov 2000 15:20:37 -0600
X-Sender: jim_evo2000@yahoo.com
X-Mailer: QUALCOMM Windows Eudora Pro Version 4.1
Content-Type: text/plain; charset="us-ascii"
X-Priority: 3
X-MSMail-Priority: Normal
Message-Id: <E13wYM4-0005IF-00@fencepost.gnu.org>
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

The United States Golf Institute is conducting a series of Golf clinics in Arizona this week. 

Everyone in attendance will receive the Evolution Golf 10.5 degree driver as a Free Gift for attending. 

To get all the details on this exciting Golf Clinic and to insure you get your Free Driver click here:

                                       http:/www.geocities.com/george_evg/index.html

                                              or call toll free          800 319 7266













We adhere to responsible mailing rules and ethics 
to be removed from this list please hit reply and type remove in the subject box.





From tromey@creche.cygnus.com Thu Nov 16 19:40:24 2000
Received: from gasherbrumii.peakpeak.com ([207.174.178.17] helo=gash2.peakpeak.com)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13wZZo-0007dU-00
	for <classpath@gnu.org>; Thu, 16 Nov 2000 19:40:24 -0500
Received: from creche.cygnus.com (06-088.026.popsite.net [64.24.106.88])
	by gash2.peakpeak.com (8.9.3/8.9.3) with ESMTP id RAA14813;
	Thu, 16 Nov 2000 17:40:19 -0700
Received: (from tromey@localhost)
	by creche.cygnus.com (8.9.3/8.9.3) id RAA11569;
	Thu, 16 Nov 2000 17:46:44 -0700
To: Classpath Project <classpath@gnu.org>
Subject: ByteArrayOutputStream.close
X-Zippy:  ..I'll make you an ASHTRAY!!
X-Attribution:  Tom
Reply-To: tromey@cygnus.com
From: Tom Tromey <tromey@cygnus.com>
Date: 16 Nov 2000 17:46:43 -0700
Message-ID: <871ywb2znw.fsf@creche.cygnus.com>
Lines: 128
X-Mailer: Gnus v5.7/Emacs 20.5
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

Consider this program:

    import java.io.*;
    public class foo
    {
      public static void main (String[] args)
      {
	try
	  {
	    ByteArrayOutputStream baos = new ByteArrayOutputStream ();
	    baos.write (23);
	    baos.close ();
	    baos.write (27);
	    System.out.println (baos.size ());
	    System.out.println (baos.toByteArray ().length);
	  }
	catch (IOException _)
	  {
	    System.out.println (_);
	  }
      }
    }

What should it print?

With libgcj (and thus the shared-with-Classpath ByteArrayOutputStream)
it prints:

    2
    2

Sun's JDK 1.2.2 also prints this.

However the online JDK docs imply that the second write() should cause
an exception to be thrown.  The docs for close() say:

    Closes this output stream and releases any system resources
    associated with this stream. A closed stream cannot perform output
    operations and cannot be reopened.

Any thoughts on this?

The appended patch changes this to work the way I think it ought to
work.  However, I'm reluctant to check this in.  Maybe it is just a
bug in the docs.  Or maybe I'm reading too much into the close()
documentation.

Tom

Index: java/io/ByteArrayOutputStream.java
===================================================================
RCS file: /cvs/java/libgcj/libjava/java/io/ByteArrayOutputStream.java,v
retrieving revision 1.5
diff -u -r1.5 ByteArrayOutputStream.java
--- ByteArrayOutputStream.java	2000/08/19 19:54:36	1.5
+++ ByteArrayOutputStream.java	2000/11/17 00:39:11
@@ -57,6 +57,7 @@
   public ByteArrayOutputStream ()
   {
     this (initial_buffer_size);
+    closed = false;
   }
 
   /**
@@ -69,6 +70,7 @@
   {
     buf = new byte[size];
     count = 0;
+    closed = false;
   }
 
   /**
@@ -83,6 +85,16 @@
   }
 
   /**
+   * Close the output stream.
+   * Future attempts to write to this stream will cause exceptions.
+   * A closed stream cannot be reopened.
+   */
+  public void close () throws IOException
+  {
+    closed = true;
+  }
+
+  /**
    * This method returns the number of bytes that have been written to
    * the buffer so far.  This is the same as the value of the protected
    * <code>count</code> variable.  If the <code>reset</code> method is
@@ -162,7 +174,7 @@
    * @return A <code>String</code> containing the data written to this
    * stream so far
    *
-   * @deprecrated
+   * @deprecated
    */
   public String toString (int hibyte)
   {
@@ -191,6 +203,8 @@
    */
   public synchronized void write (int oneByte)
   {
+    if (closed)
+      throw new IOException ("stream closed");
     resize (1);
     buf[count++] = (byte) oneByte;
   }
@@ -206,6 +220,8 @@
    */
   public synchronized void write (byte[] buffer, int offset, int add)
   {
+    if (closed)
+      throw new IOException ("stream closed");
     // If ADD < 0 then arraycopy will throw the appropriate error for
     // us.
     if (add >= 0)
@@ -236,6 +252,11 @@
    * The number of bytes that have been written to the buffer
    */
   protected int count;
+
+  /**
+   * Whether the stream has been closed.
+   */
+  private boolean closed;
 
   /**
    * The default initial buffer size.  Specified by the JCL.


From cbj@gnu.org Thu Nov 16 22:11:13 2000
Received: from rdu25-12-037.nc.rr.com ([24.25.12.37])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13wbvk-0002oJ-00
	for <classpath@gnu.org>; Thu, 16 Nov 2000 22:11:12 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eAH3B6G18351;
	Thu, 16 Nov 2000 22:11:06 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: tromey@cygnus.com
Cc: Classpath Project <classpath@gnu.org>
Subject: Re: ByteArrayOutputStream.close
References: <871ywb2znw.fsf@creche.cygnus.com>
From: Brian Jones <cbj@gnu.org>
Date: 16 Nov 2000 22:11:06 -0500
In-Reply-To: Tom Tromey's message of "16 Nov 2000 17:46:43 -0700"
Message-ID: <m3hf5747jp.fsf@rdu25-12-037.nc.rr.com>
Lines: 29
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

Tom Tromey <tromey@cygnus.com> writes:

> What should it print?
> 
> With libgcj (and thus the shared-with-Classpath ByteArrayOutputStream)
> it prints:
> 
>     2
>     2
> 
> Sun's JDK 1.2.2 also prints this.
> 
> The appended patch changes this to work the way I think it ought to
> work.  However, I'm reluctant to check this in.  Maybe it is just a
> bug in the docs.  Or maybe I'm reading too much into the close()
> documentation.
> 
> Tom

According to the Fundamental Classes Reference, none of the methods of
ByteArrayOutputStream throw an IOException because the data is written
to an array instead of an actual I/O device.  It's possibly more
interesting to check if there is a difference between the two
(sun/gcj) if you check the size of the array returned from toCharArray
() before and after the close.

Brian
-- 
Brian Jones <cbj@gnu.org>


From cbj@gnu.org Thu Nov 16 22:17:16 2000
Received: from rdu25-12-037.nc.rr.com ([24.25.12.37])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13wc1b-0002tv-00
	for <classpath@gnu.org>; Thu, 16 Nov 2000 22:17:15 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eAH3H1118367;
	Thu, 16 Nov 2000 22:17:01 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: John Leuner <jewel@pixie.co.za>
Cc: classpath@gnu.org
Subject: Re: java.util.zip
References: <Pine.LNX.4.21.0011161550320.5735-100000@skami.mweb.co.za>
From: Brian Jones <cbj@gnu.org>
Date: 16 Nov 2000 22:17:01 -0500
In-Reply-To: John Leuner's message of "Thu, 16 Nov 2000 15:54:01 +0000 (UTC)"
Message-ID: <m3d7fv479u.fsf@rdu25-12-037.nc.rr.com>
Lines: 21
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

John Leuner <jewel@pixie.co.za> writes:

> I should have made it more clear that this is a port to pure java code. My
> hidden agenda here is to use the code in a future java OS (see
> http://jos.org), which is why I didn't want to use zlib.
> 
> I will have a look at the libgcj code and Jochen's code, to see what I can
> borrow / contribute in terms of code and javadoc.

That's great.  It is what I thought you were doing.  It's probably a
good idea to borrow the javadoc and get the similar zlib type code
working in Java.  I definitely prefer a pure java implementation to
the native zlib backed implementation.  I've cc'd Paul F. on this so
he can communicate with you about paper work.

Any idea how long you think it will take to implement... possibly how
the problem can be broken down into bits different people can work on?

Brian
-- 
Brian Jones <cbj@gnu.org>


From cbj@gnu.org Thu Nov 16 22:46:33 2000
Received: from rdu25-12-037.nc.rr.com ([24.25.12.37])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13wcTu-0003sZ-00; Thu, 16 Nov 2000 22:46:30 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eAH3kSc18469;
	Thu, 16 Nov 2000 22:46:28 -0500
Date: Thu, 16 Nov 2000 22:46:28 -0500
From: cbj@gnu.org
Message-Id: <200011170346.eAH3kSc18469@rdu25-12-037.nc.rr.com>
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: classpath@gnu.org
Cc: rao@gnu.org, arenn@gnu.org, gcb@gnu.org, shalom@gnu.org
Subject: (no subject)
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

corvus@gnu.org, jochen@gnu.org
Subject: commit mail
From: Brian Jones <cbj@gnu.org>
Date: 16 Nov 2000 22:46:28 -0500
Message-ID: <m3zoiz2rcb.fsf@rdu25-12-037.nc.rr.com>
Lines: 13
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii

I've redirected the classpath-commits email to commit-classpath.  You
may subscribe to commit-classpath by sending an email to
commit-classpath-request@gnu.org with the subject 'subscribe'.

I've especially cc'd those with gnu accounts who are specified to
receive the classpath-commits email via the sendmail alias.  Send me
email after you've subscribed to the new list and I'll remove you from
the alias to avoid duplicate commit messages going your way.  I'm too
unfamiliar with the mailman software to manually add you.

Brian
-- 
Brian Jones <cbj@gnu.org>


From cbj@gnu.org Thu Nov 16 23:19:02 2000
Received: from rdu25-12-037.nc.rr.com ([24.25.12.37])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13wczN-0004ia-00
	for <classpath@gnu.org>; Thu, 16 Nov 2000 23:19:02 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eAH4J0018556;
	Thu, 16 Nov 2000 23:19:00 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: classpath@gnu.org
Subject: java/util/Locale.java
From: Brian Jones <cbj@gnu.org>
Date: 16 Nov 2000 23:19:00 -0500
Message-ID: <m3k8a32pu3.fsf@rdu25-12-037.nc.rr.com>
Lines: 7
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

Could someone who last updated Locale.java please remove the cvs diff
marks "<<<<<<" and ">>>>>>>" and any other improper code between?  I'm
not sure the real intent or I'd just do it.

Brian
-- 
Brian Jones <cbj@gnu.org>


From bryce@albatross.co.nz Thu Nov 16 23:40:46 2000
Received: from waitaki.otago.ac.nz ([139.80.75.140])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13wdKO-00053a-00; Thu, 16 Nov 2000 23:40:45 -0500
Received: from albatross.co.nz (IDENT:bryce@reason.otago.ac.nz [139.80.75.188])
	by waitaki.otago.ac.nz (8.9.3/8.9.3) with ESMTP id RAA02949;
	Fri, 17 Nov 2000 17:40:41 +1300
Message-ID: <3A14B771.69FAD216@albatross.co.nz>
Date: Fri, 17 Nov 2000 17:43:29 +1300
From: Bryce McKinlay <bryce@albatross.co.nz>
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.4.0-test9 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Brian Jones <cbj@gnu.org>
CC: classpath@gnu.org
Subject: Re: java/util/Locale.java
References: <m3k8a32pu3.fsf@rdu25-12-037.nc.rr.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

Brian Jones wrote:

> Could someone who last updated Locale.java please remove the cvs diff
> marks "<<<<<<" and ">>>>>>>" and any other improper code between?  I'm
> not sure the real intent or I'd just do it.

My bad. Fixed.

regards

  [ bryce ]




From Jochen.Hoenicke@informatik.uni-oldenburg.de Fri Nov 17 09:36:19 2000
Received: from arbi.informatik.uni-oldenburg.de ([134.106.1.7])
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13wmcl-0008R9-00
	for <classpath@gnu.org>; Fri, 17 Nov 2000 09:36:19 -0500
Received: from brecht.Informatik.Uni-Oldenburg.DE ([134.106.15.4])
	by arbi.Informatik.Uni-Oldenburg.DE (Exim 3.16)
	id 13wmca-0005bh-00; Fri, 17 Nov 2000 15:36:08 +0100
Received: from huxley.Informatik.Uni-Oldenburg.DE (huxley.Informatik.Uni-Oldenburg.DE [134.106.19.9])
	by brecht.Informatik.Uni-Oldenburg.DE (8.8.7/8.8.7) with ESMTP id PAA02404;
	Fri, 17 Nov 2000 15:36:01 +0100 (MET)
Received: (from hoenicke@localhost)
	by huxley.Informatik.Uni-Oldenburg.DE (8.8.7/8.8.7) id PAA20479;
	Fri, 17 Nov 2000 15:36:01 +0100 (MET)
From: "Jochen Hoenicke" <Jochen.Hoenicke@Informatik.Uni-Oldenburg.DE>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <14869.16976.847819.337124@huxley.Informatik.Uni-Oldenburg.DE>
Date: Fri, 17 Nov 2000 15:36:00 +0100 (MET)
To: John Leuner <jewel@pixie.co.za>
Cc: classpath@gnu.org
Subject: Re: java.util.zip
In-Reply-To: <Pine.LNX.4.21.0011161550320.5735-100000@skami.mweb.co.za>
References: <20001116002914.A18179@klomp.org>
	<Pine.LNX.4.21.0011161550320.5735-100000@skami.mweb.co.za>
X-Mailer: VM 6.75 under 20.0 XEmacs Lucid (beta28)
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

On Nov 16, John Leuner wrote:
> 
> I should have made it more clear that this is a port to pure java code. My
> hidden agenda here is to use the code in a future java OS (see
> http://jos.org), which is why I didn't want to use zlib.

IIRC, everything except Deflater and Inflater are already pure java in
libgcj.  Deflater and Inflater are just interfaces to zlib, so it
makes sense for you to rewrite them from scratch.  You should borrow
some code from zlib, of course.  And make sure you have RFC 1950-1952
handy.

> I will have a look at the libgcj code and Jochen's code, to see what I can
> borrow / contribute in terms of code and javadoc.

You can probably take all other classes in java.zip from libgcj
without changes.  Regarding javadoc I think there is very little in
libgcj.

  Jochen


From mark@klomp.org Fri Nov 17 18:44:05 2000
Received: from node1c5f.a2000.nl ([62.108.28.95] helo=nescio.klomp.org)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13wvAq-0006JF-00
	for <classpath@gnu.org>; Fri, 17 Nov 2000 18:44:04 -0500
Received: from mark by nescio.klomp.org with local (Exim 3.16 #1 (Debian))
	id 13wvAJ-0002rq-00; Sat, 18 Nov 2000 00:43:31 +0100
Date: Sat, 18 Nov 2000 00:43:31 +0100
From: Mark Wielaard <mark@klomp.org>
To: Jochen Hoenicke <Jochen.Hoenicke@Informatik.Uni-Oldenburg.DE>
Cc: John Leuner <jewel@pixie.co.za>, classpath@gnu.org
Subject: Re: java.util.zip
Message-ID: <20001118004331.A10129@klomp.org>
References: <20001116002914.A18179@klomp.org> <Pine.LNX.4.21.0011161550320.5735-100000@skami.mweb.co.za> <14869.16976.847819.337124@huxley.Informatik.Uni-Oldenburg.DE>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5i
In-Reply-To: <14869.16976.847819.337124@huxley.Informatik.Uni-Oldenburg.DE>; from Jochen.Hoenicke@Informatik.Uni-Oldenburg.DE on Fri, Nov 17, 2000 at 03:36:00PM +0100
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

Hi,

On Fri, Nov 17, 2000 at 03:36:00PM +0100, Jochen Hoenicke wrote:
> On Nov 16, John Leuner wrote:
> > 
> > I should have made it more clear that this is a port to pure java code. My
> > hidden agenda here is to use the code in a future java OS (see
> > http://jos.org), which is why I didn't want to use zlib.
Nice. But doesn't JOS support JNI? How are other non-pure-java libraries
suported? You could even consider implementing CNI support for JOS if
you are concerned about the speed barrier of JNI.

> IIRC, everything except Deflater and Inflater are already pure java in
> libgcj.  Deflater and Inflater are just interfaces to zlib, so it
> makes sense for you to rewrite them from scratch.  You should borrow
> some code from zlib, of course.  And make sure you have RFC 1950-1952
> handy.

I think having a pure java.util.zip is very sexy but keep in mind that
the zlib library has been very well tested and by porting it to java
you could introduce subtle bugs. So maybe only writing the JNI/CNI
wrapper isn't such a bad choice.

> > I will have a look at the libgcj code and Jochen's code, to see what I can
> > borrow / contribute in terms of code and javadoc.
> 
> You can probably take all other classes in java.zip from libgcj
> without changes.  Regarding javadoc I think there is very little in
> libgcj.

I just comitted my Javadoc additions to the libgcj zip code and
Tom Tromey just comitted a couple of small code aditions so make
sure you get the very latest version from CVS.

Cheers,

Mark


From mark@klomp.org Fri Nov 17 20:08:50 2000
Received: from node1c5f.a2000.nl ([62.108.28.95] helo=nescio.klomp.org)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13wwUs-0000J2-00
	for <classpath@gnu.org>; Fri, 17 Nov 2000 20:08:50 -0500
Received: from mark by nescio.klomp.org with local (Exim 3.16 #1 (Debian))
	id 13wwUq-0003iH-00; Sat, 18 Nov 2000 02:08:48 +0100
Date: Sat, 18 Nov 2000 02:08:48 +0100
From: Mark Wielaard <mark@klomp.org>
To: classpath@gnu.org
Subject: (Re)merge java.util.jar code with libgcj
Message-ID: <20001118020848.A14264@klomp.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5i
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

Hi,

I forgot to merge some small changes in the java.util.jar code from
libgcj that were made by me and Anthony Green a while back.
I have just comitted them to Classpath.

        Merge with libgcj (fixes by Anthony Green and Mark Wielaard)
        * java/util/jar/Attributes.java: Correct comment spelling.
        * java/util/jar/JarFile.java: Constructor that takes mode now
        calls super. Don't call java.util.zip.ZipFile.getEntry twice.

Cheers,

Mark


From tromey@creche.cygnus.com Sat Nov 18 12:11:11 2000
Received: from gasherbrumii.peakpeak.com ([207.174.178.17] helo=gash2.peakpeak.com)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13xBWB-00028o-00; Sat, 18 Nov 2000 12:11:11 -0500
Received: from creche.cygnus.com (06-088.026.popsite.net [64.24.106.88])
	by gash2.peakpeak.com (8.9.3/8.9.3) with ESMTP id KAA32550;
	Sat, 18 Nov 2000 10:11:05 -0700
Received: (from tromey@localhost)
	by creche.cygnus.com (8.9.3/8.9.3) id KAA25104;
	Sat, 18 Nov 2000 10:17:58 -0700
To: Brian Jones <cbj@gnu.org>
Cc: Classpath Project <classpath@gnu.org>
Subject: Re: ByteArrayOutputStream.close
References: <871ywb2znw.fsf@creche.cygnus.com> <m3hf5747jp.fsf@rdu25-12-037.nc.rr.com>
X-Zippy:  ..  One FISHWICH coming up!!
X-Attribution:  Tom
Reply-To: tromey@cygnus.com
From: Tom Tromey <tromey@cygnus.com>
Date: 18 Nov 2000 10:17:57 -0700
In-Reply-To: Brian Jones's message of "16 Nov 2000 22:11:06 -0500"
Message-ID: <87snopurlm.fsf@creche.cygnus.com>
Lines: 14
X-Mailer: Gnus v5.7/Emacs 20.5
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0beta6
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>, <mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: http://mail.gnu.org/pipermail/classpath/

>>>>> "Brian" == Brian Jones <cbj@gnu.org> writes:

Brian> According to the Fundamental Classes Reference, none of the
Brian> methods of ByteArrayOutputStream throw an IOException because
Brian> the data is written to an array instead of an actual I/O
Brian> device.  It's possibly more interesting to check if there is a
Brian> difference between the two (sun/gcj) if you check the size of
Brian> the array returned from toCharArray () before and after the
Brian> close.

It prints what you'd expect if close() did nothing.  Apparently that
is how it works.  I'll drop the patch.

Tom


From joinda1@hotmail.com Mon Nov 20 11:29:07 2000
Received: from [211.104.46.232] (helo=hsh)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13xtmV-0000io-00
	for <classpath@gnu.org>; Mon, 20 Nov 2000 11:27:14 -0500
From: "이영섭." <joinda1@hotmail.com>
To: classpath@gnu.org
Subject: 안녕하세요 classpath@gnu.org님
Mime-Version: 1.0
Content-Type: text/html; charset="iso-2022-kr"
Date: Tue, 21 Nov 2000 01:18:29
Message-Id: <E13xtmV-0000io-00@fencepost.gnu.org>
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

<html>

<head>
<meta http-equiv="Content-Language" content="ko">
<meta http-equiv="Content-Type" content="text/html; charset=ks_c_5601-1987">
<meta name="GENERATOR" content="Microsoft FrontPage 4.0">
<meta name="ProgId" content="FrontPage.Editor.Document">
<title>안녕하세요</title>
</head>

<body>

<p><b><font size="4">제게 메일을 보내실때는 꼭 <A href="mailto:stopbus@hitel.net">stopbus@hitel.net</A>으로    
보내주세요</font></b></p>   
<hr width="70%" align="left" >
   
<p><font size="4" face="굴림"><strong><span style="LETTER-SPACING: 0px"> 안녕하세요   
classpath@gnu.org님...</span></strong></font></p>    
<p><font size="4" face="굴림"><strong><span style="LETTER-SPACING: 0px">우연히 http://miso.sarang.net/softwar에 방문했다가 제      
나름대로 이 정보가 서로에게</span></strong></font></p>     
<p><font size="4" face="굴림"><strong><span style="LETTER-SPACING: 0px">도움이 될수있다고      
생각되기에 이렇게 메일을 드려봅니다.</span></strong></font></p>     
<p align="left"><font size="4" face="굴림"><strong><span style="LETTER-SPACING: 0px">기분 상하셨다면      
죄송하구요. 넓은 마음으로 이해 바랍니다.</span></strong></font></p>    
<p align="left"><font size="4" face="굴림"><strong><span style="LETTER-SPACING: 0px">그리고,     
요즘 날씨도 추운데 감기 조심하시구요</span></strong></font></p>    
<p align="left"><span style="LETTER-SPACING: 0px"><strong><font size="4" face="굴림">건강     
하시고 하시는 일 잘 되시길</font></strong> <strong><font size="4" face="굴림">바랍니다.</font></strong></span></p>    
<p align="left"><strong><font size="4" face="굴림"><span style="LETTER-SPACING: 0px">하지만 아래 사항은      
분명한 사실임을 약속드립니다.</span></font></strong></p>     
<p><font size="4" face="굴림"><strong><span style="LETTER-SPACING: 0px">기타    
문의사항은 </span></strong></font><font size="4" face="굴림"><strong><span style="LETTER-SPACING: 0px"><A href="mailto:stopbus@hitel.net">stopbus@hitel.net</A>    
으로 메일을 주시면 언제든</span></strong></font></p>   
<p><font size="4" face="굴림"><strong><span style="LETTER-SPACING: 0px">자세히     
알려드리겠습니다.</span></strong></font></p>    
<hr width="70%" align="left"><b><span style="LETTER-SPACING: 0px"><font face="견고딕" color="#0000ff" size="4">
<P><STRONG><B><SPAN style="LETTER-SPACING: 0px"><FONT color=#0000ff face=견고딕 
size=4>*전기요금 인상 걱정 마세요</FONT><FONT color=#ff0000 face=견고딕 size=4>-</FONT><FONT 
color=#ff0000 face=견고딕 size=4>전기를 10~30% 절약할 수 
있습니다.</FONT></SPAN></B></STRONG></P>
<P><STRONG><B><SPAN style="LETTER-SPACING: 0px"><FONT color=#0000ff face=견고딕 
size=4>*유해 전자파 걱정하지 마세요</FONT><FONT color=#ff0000 face=견고딕 size=4>-전자파를 차단시킬 수 
있습니다.</FONT></SPAN></B></STRONG></P>
<P>      
 </font></span></b><B><SPAN style="LETTER-SPACING: 0px"><FONT color=#0000ff 
face=견고딕 size=4>*저희 제품에 관심이 있으신분은 누구라도 환영합니다 [개인딜러 환영]</FONT></SPAN></B></P>   
<p><font color="#ff0000" size="6" face="견고딕"><b><span style="LETTER-SPACING: 0px">100% 환불 및 교환을      
해 드립니다.</span></b></font></p>    
<p><font face="견고딕" color="#0000ff" size="4"><b><span style="LETTER-SPACING: 0px"><span style="LETTER-SPACING: 0px">전국 어디나 직접      
배달 설치를 해드립니다.(단 지방은 5개이상)</span></span></b></font></p><b><span style="LETTER-SPACING: 0px"><font face="신그래픽" size="5" color="#ff0066">
<P><FONT color=#0000ff size=4>간단히 전기소켓에 꽂아만 주시면 됩니다</FONT></P>     
      </font></span></b>
<P><B><SPAN style="LETTER-SPACING: 0px"><FONT color=#ff0066 face=신그래픽 size=5>*가격 
: 가정용 55000원 , 업소용 : 99000원</FONT><font face="신그래픽" size="4">　</font></SPAN></B></P>    
<p><strong><b><span style="LETTER-SPACING: 0px"><font face="견고딕" size="4" color="#0000ff">      
 </font></span></b></strong>&nbsp;</p>
<P><FONT face=굴림 size=4><STRONG><SPAN style="LETTER-SPACING: 0px">그 외에  
<FONT color=#ff0000>다양한 기능이 </FONT>있지만 이정도만 
하겠습니다.</SPAN></STRONG></FONT></P>     
<p><font size="4" face="굴림"><strong><span style="LETTER-SPACING: 0px">자세한 사항은&nbsp; 
아래 사이트에서 직접 눈으로 확인해 보세요</span></strong></font></p>     
<p><font size="4" face="굴림"><span style="LETTER-SPACING: 0px"><strong>또한 <font color="#0000ff">검색어</font>를 "<font color="#ff0000"></strong><b>슈퍼에너지</b><strong>"</strong></font><strong>라고      
치시면 많은 정보를 찾으실 수 있으며</strong></SPAN></FONT></p>     
<p><font size="4" face="굴림"><strong><span style="LETTER-SPACING: 0px">각종 <font color="#ff0000">일간지      
및 KBS 9시 뉴스</font>에 방영 되었습니다.</span></strong></font></p>     
<p></p>    
<p><a href="http://203.248.251.56/asx.cgi?title=KBS뉴스&amp;url=2000$05$24$230"><span style="LETTER-SPACING: 0px"><b><font size="4">동영상보기</font></b></span></a>     
 <font size="4"><b><span style="LETTER-SPACING: 0px"><font face="견고딕" color="#0000ff">[KBS 9시 뉴스]</font>&nbsp; 관련     
사이트 : <a href="http://www.spenergy.com/">http://www.spenergy.com/</a> 외     
다수</span></b></font></p>    
<p><font face="견고딕" size="3"><strong><span style="LETTER-SPACING: 0px">       
</span></strong></font>&nbsp;</p>    
<p><font face="견고딕" size="3"><strong><span style="LETTER-SPACING: 0px">상담및 
문의처 :&nbsp;슈퍼에너지 인터넷 사업부&nbsp;&nbsp;&nbsp;&nbsp;</p></SPAN></STRONG></FONT>     
<p><strong><font face="굴림"><span style="LETTER-SPACING: 0px">  02-744-0209 <FONT color=#0000ff  
size=4>협력업체 모집중(개인도 가능)</FONT>  </span></font></strong></p>
<P><FONT face=굴림><STRONG>현재 전화량이 많아 통화가 않될때는 음성메세지로 연락처를 
남겨주시거나</STRONG></FONT></P>
<P><FONT face=굴림><STRONG>메일로 연락처를 알려주시면 바로 연락을 드리겠습니다.</STRONG></FONT></P>     
<p><strong><font face="굴림"><span style="LETTER-SPACING: 0px"><FONT color=#0000ff size=4>&nbsp;&nbsp; </FONT> <A href="mailto:stopbus@hitel.net">stopbus@hitel.net</A></span></font></strong></p>     
<p>&nbsp;</p>    
    
</body>    
    
</html>    


From jewel@pixie.co.za Mon Nov 20 12:36:01 2000
Received: from earthquake.mweb.co.za ([196.2.53.139])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13xurF-0002gL-00
	for <classpath@gnu.org>; Mon, 20 Nov 2000 12:36:01 -0500
Received: from pta-dial-196-31-185-3.mweb.co.za ([196.31.185.3])
 by earthquake.mweb.co.za
 (Sun Internet Mail Server sims.3.5.2000.03.23.18.03.p10)
 with ESMTP id <0G4C00JKJ3HLSQ@earthquake.mweb.co.za> for classpath@gnu.org;
 Mon, 20 Nov 2000 19:34:35 +0200 (SAT)
Date: Mon, 20 Nov 2000 17:39:51 +0000 (UTC)
From: John Leuner <jewel@pixie.co.za>
Subject: Re: java.util.zip
In-reply-to: <20001118004331.A10129@klomp.org>
X-Sender: jewel@skami.mweb.co.za
To: classpath@gnu.org
Message-id: <Pine.LNX.4.21.0011201734340.1073-100000@skami.mweb.co.za>
MIME-version: 1.0
Content-type: TEXT/PLAIN; charset=US-ASCII
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

> > > I should have made it more clear that this is a port to pure java code. My
> > > hidden agenda here is to use the code in a future java OS (see
> > > http://jos.org), which is why I didn't want to use zlib.
> Nice. But doesn't JOS support JNI? How are other non-pure-java libraries
> suported? You could even consider implementing CNI support for JOS if
> you are concerned about the speed barrier of JNI.

We have JNI support in one of the current alpha implementations, but again
it's a political / philosophical issue, not one of practicality.

I assume CNI requires C++ support?
 
> > IIRC, everything except Deflater and Inflater are already pure java in
> > libgcj.  Deflater and Inflater are just interfaces to zlib, so it
> > makes sense for you to rewrite them from scratch.  You should borrow
> > some code from zlib, of course.  And make sure you have RFC 1950-1952
> > handy.
> 
> I think having a pure java.util.zip is very sexy but keep in mind that
> the zlib library has been very well tested and by porting it to java
> you could introduce subtle bugs. So maybe only writing the JNI/CNI
> wrapper isn't such a bad choice.

There is always that risk. I can test against the JNI implementation and
other established utilities anyway, so there isn't such a great risk. 

> > > I will have a look at the libgcj code and Jochen's code, to see what I can
> > > borrow / contribute in terms of code and javadoc.
> > 
> > You can probably take all other classes in java.zip from libgcj
> > without changes.  Regarding javadoc I think there is very little in
> > libgcj.
> 
> I just comitted my Javadoc additions to the libgcj zip code and
> Tom Tromey just comitted a couple of small code aditions so make
> sure you get the very latest version from CVS.

I'll do that. 
Thanks.

John Leuner



From jewel@pixie.co.za Mon Nov 20 12:46:52 2000
Received: from jhb-proxy.mweb.co.za ([196.2.48.243])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13xv1n-00031f-00
	for <classpath@gnu.org>; Mon, 20 Nov 2000 12:46:52 -0500
Received: from pta-dial-196-31-185-3.mweb.co.za ([196.31.185.3])
 by jhb-proxy.mweb.co.za
 (Sun Internet Mail Server sims.3.5.2000.03.23.18.03.p10)
 with ESMTP id <0G4C00DHX3TXGM@jhb-proxy.mweb.co.za> for classpath@gnu.org;
 Mon, 20 Nov 2000 19:41:59 +0200 (GMT-2)
Date: Mon, 20 Nov 2000 17:47:17 +0000 (UTC)
From: John Leuner <jewel@pixie.co.za>
Subject: Re: java.util.zip
In-reply-to: <14869.16976.847819.337124@huxley.Informatik.Uni-Oldenburg.DE>
X-Sender: jewel@skami.mweb.co.za
To: classpath@gnu.org
Message-id: <Pine.LNX.4.21.0011201740200.1073-100000@skami.mweb.co.za>
MIME-version: 1.0
Content-type: TEXT/PLAIN; charset=US-ASCII
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

> > I should have made it more clear that this is a port to pure java code. My
> > hidden agenda here is to use the code in a future java OS (see
> > http://jos.org), which is why I didn't want to use zlib.
> 
> IIRC, everything except Deflater and Inflater are already pure java in
> libgcj.  Deflater and Inflater are just interfaces to zlib, so it
> makes sense for you to rewrite them from scratch.  You should borrow
> some code from zlib, of course.  And make sure you have RFC 1950-1952
> handy.

I'm working from the zlib C source and the rfc documents.

> > I will have a look at the libgcj code and Jochen's code, to see what I can
> > borrow / contribute in terms of code and javadoc.
> 
> You can probably take all other classes in java.zip from libgcj
> without changes.  Regarding javadoc I think there is very little in
> libgcj.

Well the latest CVS seems to be pretty good, I'll try merge what docs I
have and suggest some patches.
 
John Leuner



From cbj@gnu.org Mon Nov 20 13:17:16 2000
Received: from rdu25-12-037.nc.rr.com ([24.25.12.37])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13xvVE-0003gn-00
	for <classpath@gnu.org>; Mon, 20 Nov 2000 13:17:16 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eAKIHAE09395;
	Mon, 20 Nov 2000 13:17:10 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: GNU Classpath <classpath@gnu.org>
Cc: japhar@hungry.com
Subject: classpath-0.01 released
From: Brian Jones <cbj@gnu.org>
Date: 20 Nov 2000 13:17:10 -0500
Message-ID: <m3y9yecxuh.fsf@rdu25-12-037.nc.rr.com>
Lines: 36
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

I've packaged and released classpath-0.01.  This release features
numerous improvements and additions.  Places to download the source
code and precompiled binaries are listed below.

Compiling from source:

        To compile from source code you must have JDK 1.1.x (1.1.8
        preferred) and Jikes in addition to japhar-0.09 and nspr.
        Links for downloading source code of the latter are available
        from lyta.haphazard.org below.  Use configure flags
        --with-japhar=<prefix> and --with-javah=... and
        --with-classlib=... the latter two should point to the JDK
        relevant files.

Binaries:

        To use the provided rpms, install nspr-pthreads, japhar, and
        classpath rpms.

        CLASSPATH=/usr/japhar/share/classpath
        PATH=/usr/japhar/bin
        LD_LIBRARY_PATH=/usr/japhar/lib/classpath:/usr/japhar/lib

Download:

        ftp://alpha.gnu.org/gnu/classpath/classpath-0.01.tar.gz
        http://lyta.haphazard.org/~cbj/classpath.html

Have a happy Thanksgiving for those of you in the U.S.  I'll be in PA
for the rest of the week until late Saturday.  I hope releases will
not be spread year(s) apart again.  This release met my "Hello World"
executed correctly confidence level.

Brian
-- 
Brian Jones <cbj@gnu.org>


From tromey@creche.cygnus.com Mon Nov 20 13:22:02 2000
Received: from gasherbrumii.peakpeak.com ([207.174.178.17] helo=gash2.peakpeak.com)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13xvZp-0003x1-00
	for <classpath@gnu.org>; Mon, 20 Nov 2000 13:22:01 -0500
Received: from creche.cygnus.com (01-055.026.popsite.net [64.24.105.55])
	by gash2.peakpeak.com (8.9.3/8.9.3) with ESMTP id LAA20676;
	Mon, 20 Nov 2000 11:21:55 -0700
Received: (from tromey@localhost)
	by creche.cygnus.com (8.9.3/8.9.3) id LAA11485;
	Mon, 20 Nov 2000 11:29:18 -0700
To: Classpath Project <classpath@gnu.org>
Subject: cvs2cl
X-Zippy:  I always liked FLAG DAY!!
X-Attribution:  Tom
Reply-To: tromey@cygnus.com
From: Tom Tromey <tromey@cygnus.com>
Date: 20 Nov 2000 11:29:18 -0700
Message-ID: <878zqeo5tt.fsf@creche.cygnus.com>
Lines: 29
X-Mailer: Gnus v5.7/Emacs 20.5
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

I noticed that Brian regenerated ChangeLog using cvs2cl yesterday.
This made it much less readable, imho.  For instance the correctly
("correct" being determined by the GNU coding standards) formatted
entry:

2000-11-17  Mark Wielaard  <mark@klomp.org>

        Merge with libgcj (fixes by Anthony Green and Mark Wielaard)
        * java/util/jar/Attributes.java: Correct comment spelling.
        * java/util/jar/JarFile.java: Constructor that takes mode now
        calls super. Don't call java.util.zip.ZipFile.getEntry twice.

changed to this:

2000-11-18 01:06  Mark Wielaard <mark@klomp.org>

	* java/util/jar/: Attributes.java, JarFile.java:	Merge with
	libgcj (fixes by Anthony Green and Mark Wielaard)
		* java/util/jar/Attributes.java: Correct comment spelling.
		* java/util/jar/JarFile.java: Constructor that takes mode
	now
		calls super. Don't call java.util.zip.ZipFile.getEntry
	twice.


This change also results in gratuitous changes to the file that I had
to merge with my local copy.  Do we have to use it?

Tom


From jkeiser@iname.com Mon Nov 20 13:26:49 2000
Received: from [207.225.27.130] (helo=colorado.ieffects.com)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13xveT-0004Ei-00
	for <classpath@gnu.org>; Mon, 20 Nov 2000 13:26:49 -0500
Received: (qmail 25107 invoked by uid 501); 20 Nov 2000 19:27:36 -0000
Received: from delaware.ieffects.com (HELO delaware) (10.1.1.3)
  by colorado.ieffects.com with SMTP; 20 Nov 2000 19:27:36 -0000
From: "John Keiser" <jkeiser@iname.com>
To: "Brian Jones" <cbj@gnu.org>,
	"GNU Classpath" <classpath@gnu.org>
Cc: <japhar@hungry.com>
Subject: RE: classpath-0.01 released
Date: Mon, 20 Nov 2000 11:26:07 -0700
Message-ID: <NEBBJENFPMAHNAJLBGMKCEOHCFAA.jkeiser@iname.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
Importance: Normal
In-Reply-To: <m3y9yecxuh.fsf@rdu25-12-037.nc.rr.com>
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Yaaaaay!  Thanks so much, Brian!

So the rumors that it doesn't work with current Japhar are exaggerated?

--John

> -----Original Message-----
> From: cbj@rdu25-12-037.nc.rr.com [mailto:cbj@rdu25-12-037.nc.rr.com]On
> Behalf Of Brian Jones
> Sent: Monday, November 20, 2000 11:17 AM
> To: GNU Classpath
> Cc: japhar@hungry.com
> Subject: classpath-0.01 released
> 
> 
> I've packaged and released classpath-0.01.  This release features
> numerous improvements and additions.  Places to download the source
> code and precompiled binaries are listed below.
> 
> Compiling from source:
> 
>         To compile from source code you must have JDK 1.1.x (1.1.8
>         preferred) and Jikes in addition to japhar-0.09 and nspr.
>         Links for downloading source code of the latter are available
>         from lyta.haphazard.org below.  Use configure flags
>         --with-japhar=<prefix> and --with-javah=... and
>         --with-classlib=... the latter two should point to the JDK
>         relevant files.
> 
> Binaries:
> 
>         To use the provided rpms, install nspr-pthreads, japhar, and
>         classpath rpms.
> 
>         CLASSPATH=/usr/japhar/share/classpath
>         PATH=/usr/japhar/bin
>         LD_LIBRARY_PATH=/usr/japhar/lib/classpath:/usr/japhar/lib
> 
> Download:
> 
>         ftp://alpha.gnu.org/gnu/classpath/classpath-0.01.tar.gz
>         http://lyta.haphazard.org/~cbj/classpath.html
> 
> Have a happy Thanksgiving for those of you in the U.S.  I'll be in PA
> for the rest of the week until late Saturday.  I hope releases will
> not be spread year(s) apart again.  This release met my "Hello World"
> executed correctly confidence level.
> 
> Brian
> -- 
> Brian Jones <cbj@gnu.org>
> 


From cbj@gnu.org Mon Nov 20 14:33:02 2000
Received: from rdu25-12-037.nc.rr.com ([24.25.12.37])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13xwgX-0005rN-00
	for <classpath@gnu.org>; Mon, 20 Nov 2000 14:33:01 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eAKJWWF09592;
	Mon, 20 Nov 2000 14:32:32 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: "John Keiser" <jkeiser@iname.com>
Cc: "GNU Classpath" <classpath@gnu.org>, <japhar@hungry.com>
Subject: Re: classpath-0.01 released
References: <NEBBJENFPMAHNAJLBGMKCEOHCFAA.jkeiser@iname.com>
From: Brian Jones <cbj@gnu.org>
Date: 20 Nov 2000 14:32:32 -0500
In-Reply-To: "John Keiser"'s message of "Mon, 20 Nov 2000 11:26:07 -0700"
Message-ID: <m366limobz.fsf@rdu25-12-037.nc.rr.com>
Lines: 15
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

"John Keiser" <jkeiser@iname.com> writes:

> Yaaaaay!  Thanks so much, Brian!
> 
> So the rumors that it doesn't work with current Japhar are exaggerated?
> 
> --John

I haven't heard those, but it works with a simple Hello World
example.  The first thing I want to fix is to make exceptions actually
print a valid stack trace which is not the case today.

Brian
-- 
Brian Jones <cbj@gnu.org>


From jkeiser@iname.com Mon Nov 20 16:15:58 2000
Received: from [207.225.27.130] (helo=colorado.ieffects.com)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13xyI9-0000Jr-00
	for <classpath@gnu.org>; Mon, 20 Nov 2000 16:15:58 -0500
Received: (qmail 26624 invoked by uid 501); 20 Nov 2000 22:17:24 -0000
Received: from delaware.ieffects.com (HELO delaware) (10.1.1.3)
  by colorado.ieffects.com with SMTP; 20 Nov 2000 22:17:24 -0000
From: "John Keiser" <jkeiser@iname.com>
To: "Brian Jones" <cbj@gnu.org>
Cc: "Classpath" <classpath@gnu.org>
Subject: RE: classpath-0.01 released
Date: Mon, 20 Nov 2000 14:15:55 -0700
Message-ID: <NEBBJENFPMAHNAJLBGMKMEOJCFAA.jkeiser@iname.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
Importance: Normal
In-Reply-To: <m366limobz.fsf@rdu25-12-037.nc.rr.com>
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Yeah, I was able to run it before too, around when we had the reports.

Ooh, exceptions, that's a fun one :)  Shouldn't be awfully bad unless Japhar
doesn't support it already.

--John Keiser



> -----Original Message-----
> From: cbj@rdu25-12-037.nc.rr.com [mailto:cbj@rdu25-12-037.nc.rr.com]On
> Behalf Of Brian Jones
> Sent: Monday, November 20, 2000 12:33 PM
> To: John Keiser
> Cc: GNU Classpath; japhar@hungry.com
> Subject: Re: classpath-0.01 released
>
>
> "John Keiser" <jkeiser@iname.com> writes:
>
> > Yaaaaay!  Thanks so much, Brian!
> >
> > So the rumors that it doesn't work with current Japhar are exaggerated?
> >
> > --John
>
> I haven't heard those, but it works with a simple Hello World
> example.  The first thing I want to fix is to make exceptions actually
> print a valid stack trace which is not the case today.
>
> Brian
> --
> Brian Jones <cbj@gnu.org>
>



From cbj@gnu.org Tue Nov 21 00:38:56 2000
Received: from rdu25-12-037.nc.rr.com ([24.25.12.37])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13y68t-0005ME-00
	for <classpath@gnu.org>; Tue, 21 Nov 2000 00:38:55 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eAL5clF11039;
	Tue, 21 Nov 2000 00:38:47 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: GNU Classpath <classpath@gnu.org>
Subject: JOS
From: Brian Jones <cbj@gnu.org>
Date: 21 Nov 2000 00:38:47 -0500
Message-ID: <m38zqdsx3s.fsf@rdu25-12-037.nc.rr.com>
Lines: 8
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

The following article was interesting to me.  I don't know what the
state of the JOS project is or who our liason is.  Could someone from
that project bring us up to date with respect to this email referenced
below?

http://www.jos.org/pipermail/arch/1999-December/000358.html
-- 
Brian Jones <cbj@gnu.org>


From bryce@albatross.co.nz Wed Nov 22 06:49:44 2000
Received: from waitaki.otago.ac.nz ([139.80.75.140])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13yYPH-0007qE-00
	for <classpath@gnu.org>; Wed, 22 Nov 2000 06:49:43 -0500
Received: from albatross.co.nz (IDENT:bryce@reason.otago.ac.nz [139.80.75.188])
	by waitaki.otago.ac.nz (8.9.3/8.9.3) with ESMTP id AAA21088
	for <classpath@gnu.org>; Thu, 23 Nov 2000 00:49:27 +1300
Message-ID: <3A1BB394.95600DAB@albatross.co.nz>
Date: Thu, 23 Nov 2000 00:52:52 +1300
From: Bryce McKinlay <bryce@albatross.co.nz>
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.4.0-test9 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: classpath@gnu.org
Subject: PATCH: java.util.Vector
Content-Type: multipart/mixed;
 boundary="------------919B584153876FE9BCA7B267"
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

This is a multi-part message in MIME format.
--------------919B584153876FE9BCA7B267
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Classpath's Vector implementation was not synchronized. It should be. It
also had a bug where indexOf() and lastIndexOf() would not handle null
entries correctly.

This patch implements all the methods specified in the JDK 1.3 docs.
Many of these are simply synchronized wrappers around the default
implementations in AbstractList/AbstractCollection. However, where a
collections method is also implemented in ArrayList, I've implemented it
likewise in Vector.

There are also a couple of minor formatting fixes here.

regards

  [ bryce ]

P.S. Whats with the mangled classpath ChangeLog? Do we really want to be
generating this from the cvs logs? I find manual changelogs to be more
useful, and anyone can generate a CVS log if they want one.



--------------919B584153876FE9BCA7B267
Content-Type: text/plain; charset=us-ascii;
 name="classpath-collections-1123.patch"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="classpath-collections-1123.patch"

2000-11-23  Bryce McKinlay  <bryce@albatross.co.nz>

	* java/util/ArrayList.java (addAll(Collection)): Call 
	addAll(int,Collection) instead of duplicating code.
	(indexOf): Clean up int initialization.
	(clear): Set cleared array entries to null, to allow garbage 
	collection.
	* java/util/List.java: Minor formatting fixes.
	* java/util/Random.java: ditto.
	* java/util/SimpleTimeZone.java: ditto.
	* java/util/Vector.java: Make many methods synchronized. Increment 
	modCount unconditionally in line with the JCL spec.
	(Vector): Simplify constructors. Use for loop in Collection copy
	constructor for efficiency.
	(indexOf): Handle `null' elements.
	(lastIndexOf): Ditto.
	(toArray): Replace with simpler implementation from ArrayList.
	(add): Implemented.
	(remove(int)): Add bounds check.
	(containsAll): Implemented.
	(addAll): Implemented.
	(removeAll): Implemented.
	(retainAll): Implemented.
	(equals): Implemented.
	(hashCode): Implemented.
	(toString): Use string concatenation, not StringBuffer.
	(subList): Implemented.
	(removeRange): Implemented.

Index: ArrayList.java
===================================================================
RCS file: /cvs/classpath/java/util/ArrayList.java,v
retrieving revision 1.9
diff -u -r1.9 ArrayList.java
--- ArrayList.java	2000/11/02 10:12:57	1.9
+++ ArrayList.java	2000/11/22 11:29:07
@@ -43,7 +43,7 @@
  * to or removing from the end of a list, checking the size, &c.
  *
  * @author        Jon A. Zeppieri
- * @version       $Id: ArrayList.java,v 1.9 2000/11/02 10:12:57 bryce Exp $
+ * @version       $Id: ArrayList.java,v 1.3 2000/11/02 10:08:03 bryce Exp $
  * @see           java.util.AbstractList
  * @see           java.util.List
  */
@@ -219,15 +219,7 @@
    */
   public boolean addAll(Collection c)
   {
-    modCount++;
-    Iterator itr = c.iterator();
-    int csize = c.size();
-    ensureCapacity(size + csize);
-    for (int pos = 0; pos < csize; pos++)
-      {
-	data[size++] = itr.next();
-      }
-    return (csize > 0);
+    return addAll(size, c);
   }
 
   /** 
@@ -295,9 +287,7 @@
    */
   public int indexOf(Object e)
   {
-    int i;
-
-    for (i = 0; i < size; i++)
+    for (int i = 0; i < size; i++)
       {
 	if (e == null ? data[i] == null : e.equals(data[i]))
 	  return i;
@@ -330,6 +320,10 @@
   public void clear()
   {
     modCount++;
+    for (int i = 0; i < size; i++)
+      {
+	data[i] = null;
+      }    
     size = 0;
   }
 
@@ -364,8 +358,8 @@
   }
 
   /**
-   * Returns an Array whse component type is the runtime component type of
-   * the passes-in Array.  The returned Array is populated with all of the
+   * Returns an Array whose component type is the runtime component type of
+   * the passed-in Array.  The returned Array is populated with all of the
    * elements in this ArrayList.  If the passed-in Array is not large enough
    * to store all of the elements in this List, a new Array will be created 
    * and returned; if the passed-in Array is <i>larger</i> than the size
Index: List.java
===================================================================
RCS file: /cvs/classpath/java/util/List.java,v
retrieving revision 1.5
diff -u -r1.5 List.java
--- List.java	2000/10/26 10:19:01	1.5
+++ List.java	2000/11/22 11:29:07
@@ -311,7 +311,7 @@
    * @returns an array of type Object[] and length equal to the length of this
    *   list, containing the elements currently in this list, in order.
    */
-  Object[]toArray();
+  Object[] toArray();
 
   /**
    * Copy the current contents of this list into an array. If the array passed
@@ -329,5 +329,5 @@
    * @exception ArrayStoreException if the type of any element of the
    *   collection is not a subtype of the element type of a.
    */
-  Object[]toArray(Object[]a);
+  Object[] toArray(Object[] a);
 }
Index: Random.java
===================================================================
RCS file: /cvs/classpath/java/util/Random.java,v
retrieving revision 1.6
diff -u -r1.6 Random.java
--- Random.java	2000/10/26 10:19:01	1.6
+++ Random.java	2000/11/22 11:29:07
@@ -166,7 +166,7 @@
    * @param bytes The byte array that should be filled.
    * @since JDK1.1
    */
-  public void nextBytes(byte[]bytes)
+  public void nextBytes(byte[] bytes)
     /*{ require { bytes != null :: "bytes is null"; } } */
   {
     int random;
Index: SimpleTimeZone.java
===================================================================
RCS file: /cvs/classpath/java/util/SimpleTimeZone.java,v
retrieving revision 1.10
diff -u -r1.10 SimpleTimeZone.java
--- SimpleTimeZone.java	2000/10/27 09:53:52	1.10
+++ SimpleTimeZone.java	2000/11/22 11:29:08
@@ -763,7 +763,7 @@
     else
       {
 	int length = input.readInt();
-	byte[]byteArray = new byte[length];
+	byte[] byteArray = new byte[length];
 	input.read(byteArray, 0, length);
 	if (length >= 4)
 	  {
Index: Vector.java
===================================================================
RCS file: /cvs/classpath/java/util/Vector.java,v
retrieving revision 1.10
diff -u -r1.10 Vector.java
--- Vector.java	2000/10/26 10:19:01	1.10
+++ Vector.java	2000/11/22 11:29:08
@@ -27,6 +27,7 @@
 
 package java.util;
 import java.lang.reflect.Array;
+import java.io.Serializable;
 
 /**
  * the <b>Vector</b> classes implements growable arrays of Objects.
@@ -41,11 +42,16 @@
  *
  * Vector implements the JDK 1.2 List interface, and is therefor a fully
  * compliant Collection object.
+ *
+ * @specnote The JCL claims that various methods in this class throw
+ * IndexOutOfBoundsException, which would be consistent with other collections
+ * classes. ArrayIndexOutOfBoundsException is actually thrown, per the online 
+ * docs, even for List method implementations.
  * 
  * @author Scott G. Miller
  */
-public class Vector extends AbstractList implements List,
-  Cloneable, java.io.Serializable
+public class Vector extends AbstractList 
+  implements List, Cloneable, Serializable
 {
   /**
    * The amount the Vector's internal array should be increased in size when
@@ -53,14 +59,14 @@
    * or when {@link #ensureCapacity} is called.
    * @serial
    */
-  protected int capacityIncrement;
+  protected int capacityIncrement = 0;
 
   /**
    * The number of elements currently in the vector, also returned by
    * {@link #size}.
    * @serial
    */
-  protected int elementCount;
+  protected int elementCount = 0;
 
   /**
    * The internal array used to hold members of a Vector
@@ -76,7 +82,7 @@
    */
   public Vector()
   {
-    this(10, 0);
+    this(10);
   }
 
   /**
@@ -88,13 +94,13 @@
    */
   public Vector(Collection c)
   {
-    this(c.size(), 0);
-    elementCount = elementData.length;
-    Iterator collectionIterator = c.iterator();
-    int counter = 0;
-    while (collectionIterator.hasNext())
+    int csize = c.size();
+    elementData = new Object[csize];
+    elementCount = csize;
+    Iterator itr = c.iterator();
+    for (int i = 0; i < csize; i++)
       {
-	elementData[counter++] = collectionIterator.next();
+	elementData[i] = itr.next();
       }
   }
 
@@ -109,10 +115,8 @@
    */
   public Vector(int initialCapacity, int capacityIncrement)
   {
-    modCount = 0;
     elementData = new Object[initialCapacity];
     this.capacityIncrement = capacityIncrement;
-    elementCount = 0;
   }
 
   /**
@@ -122,21 +126,21 @@
    */
   public Vector(int initialCapacity)
   {
-    this(initialCapacity, 0);
+    elementData = new Object[initialCapacity];
   }
 
   /**
    * Copies the contents of a provided array into the Vector.  If the 
-   * array is too large to fit in the Vector, an IndexOutOfBoundsException
+   * array is too large to fit in the Vector, an ArrayIndexOutOfBoundsException
    * is thrown.  Old elements in the Vector are overwritten by the new
    * elements
    *
    * @param anArray An array from which elements will be copied into the Vector
    * 
-   * @throws IndexOutOfBoundsException the array being copied
+   * @throws ArrayIndexOutOfBoundsException the array being copied
    * is larger than the Vectors internal data array
    */
-  public void copyInto(Object[]anArray)
+  public synchronized void copyInto(Object[] anArray)
   {
     System.arraycopy(elementData, 0, anArray, 0, elementCount);
   }
@@ -146,7 +150,7 @@
    * than the number of Objects its holding, a new array is constructed
    * that precisely holds the elements.  
    */
-  public void trimToSize()
+  public synchronized void trimToSize()
   {
     // Check if the Vector is already trimmed, to save execution time
     if (elementCount == elementData.length)
@@ -169,16 +173,19 @@
    * @param minCapacity The minimum capacity the internal array should be
    * able to handle after executing this method
    */
-  public void ensureCapacity(int minCapacity)
+  public synchronized void ensureCapacity(int minCapacity)
   {
+    modCount++;
     if (elementData.length >= minCapacity)
       return;
 
-    int newCapacity = elementData.length + capacityIncrement;
+    int newCapacity; 
     if (capacityIncrement <= 0)
       newCapacity = elementData.length * 2;
-
-    Object[]newArray = new Object[Math.max(newCapacity, minCapacity)];
+    else
+      newCapacity = elementData.length + capacityIncrement;
+      
+    Object[] newArray = new Object[Math.max(newCapacity, minCapacity)];
 
     System.arraycopy(elementData, 0, newArray, 0, elementData.length);
     elementData = newArray;
@@ -187,20 +194,18 @@
   /**
    * Explicitly sets the size of the internal data array, copying the 
    * old values to the new internal array.  If the new array is smaller
-   * than the old one, old values that don't fit are lost.
+   * than the old one, old values that don't fit are lost. If the new size
+   * is larger than the old one, the vector is padded with null entries.
    *
    * @param newSize The new size of the internal array
    */
-  public void setSize(int newSize)
+  public synchronized void setSize(int newSize)
   {
-    if (newSize < elementCount)
-      {
-	modCount++;
-	elementCount = newSize;
-      }
+    modCount++;
     Object[] newArray = new Object[newSize];
-    System.arraycopy(elementData, 0, newArray, 0,
-		     Math.min(newSize, elementData.length));
+    System.arraycopy(elementData, 0, newArray, 0, 
+                     Math.min(newSize, elementCount));
+    elementCount = newSize;
     elementData = newArray;
   }
 
@@ -240,16 +245,16 @@
    * and returns the index of the first occurence of this Object.  If
    * the object is not found, -1 is returned
    *
-   * @param elem The Object to search for
+   * @param e The Object to search for
    * @param index Start searching at this index
    * @returns The index of the first occurence of <b>elem</b>, or -1
    * if it is not found
    */
-  public int indexOf(Object elem, int index)
+  public synchronized int indexOf(Object e, int index)
   {
     for (int i = index; i < elementCount; i++)
       {
-	if (elementData[i].equals(elem))
+	if (e == null ? elementData[i] == null : e.equals(elementData[i]))
 	  return i;
       }
     return -1;
@@ -284,17 +289,18 @@
    * backwards from <b>index</b>.  If the object does not occur in this Vector,
    * -1 is returned.
    *
-   * @param elem The object to search for
+   * @param eThe object to search for
    * @param index The index to start searching in reverse from
    * @returns The index of the Object if found, -1 otherwise
    */
-  public int lastIndexOf(Object elem, int index)
+  public synchronized int lastIndexOf(Object e, int index)
   {
-    if (index > elementCount - 1)
+    if (index >= elementCount)
       throw new ArrayIndexOutOfBoundsException(index);
+
     for (int i = index; i >= 0; i--)
       {
-	if ((elementData[i] == elem) || elementData[i].equals(elem))
+	if (e == null ? elementData[i] == null : e.equals(elementData[i]))
 	  return i;
       }
     return -1;
@@ -321,7 +327,7 @@
    * @throws ArrayIndexOutOfBoundsException <b>index</b> is
    * larger than the Vector
    */
-  public Object elementAt(int index)
+  public synchronized Object elementAt(int index)
   {
     //Within the bounds of this Vector does not necessarily mean within 
     //the bounds of the internal array
@@ -338,9 +344,9 @@
    * @returns The first Object in the Vector
    * @throws NoSuchElementException the Vector is empty
    */
-  public Object firstElement()
+  public synchronized Object firstElement()
   {
-    if (isEmpty())
+    if (elementCount == 0)
       throw new NoSuchElementException();
 
     return elementAt(0);
@@ -353,9 +359,9 @@
    * @returns The last Object in the Vector
    * @throws NoSuchElementException the Vector is empty
    */
-  public Object lastElement()
+  public synchronized Object lastElement()
   {
-    if (isEmpty())
+    if (elementCount == 0)
       throw new NoSuchElementException();
 
     return elementAt(elementCount - 1);
@@ -370,7 +376,7 @@
    * @param index The position in the Vector to store the object
    * @throws ArrayIndexOutOfBoundsException the index is out of range
    */
-  public void setElementAt(Object obj, int index)
+  public synchronized void setElementAt(Object obj, int index)
   {
     if ((index < 0) || (index >= elementCount))
       throw new ArrayIndexOutOfBoundsException(index);
@@ -386,8 +392,9 @@
    * @param element The Object to store in the Vector
    * @returns The previous object at the specified index
    * @throws ArrayIndexOutOfBoundsException the index is out of range
+   *
    */
-  public Object set(int index, Object element)
+  public synchronized Object set(int index, Object element)
   {
     if (index >= elementCount)
       throw new ArrayIndexOutOfBoundsException(index);
@@ -403,7 +410,7 @@
    *
    * @param index The index of the element to remove
    */
-  public void removeElementAt(int index)
+  public synchronized void removeElementAt(int index)
   {
     if (index >= elementCount)
       throw new ArrayIndexOutOfBoundsException(index);
@@ -445,7 +452,7 @@
    *
    * @param obj The object to add to the Vector
    */
-  public void addElement(Object obj)
+  public synchronized void addElement(Object obj)
   {
     ensureCapacity(elementCount + 1);
     modCount++;
@@ -460,32 +467,28 @@
    * @param obj The object to remove from the Vector
    * @returns true if the Object was in the Vector, false otherwise
    */
-  public boolean removeElement(Object obj)
+  public synchronized boolean removeElement(Object obj)
   {
-    int ix = indexOf(obj);
-    if (ix != -1)
+    int idx = indexOf(obj);
+    if (idx != -1)
       {
-	removeElementAt(ix);
+	removeElementAt(idx);
 	return true;
       }
-    else
-      {
-	return false;
-      }
+    return false;
   }
 
   /**
    * Removes all elements from the Vector.  Note that this does not
    * resize the internal data array.
    */
-  public void removeAllElements()
+  public synchronized void removeAllElements()
   {
-    if (size() != 0)
-      modCount++;
-    else
+    modCount++;
+    if (elementCount == 0)
       return;
 
-    for (int i = 0; i < elementData.length; i++)
+    for (int i = 0; i < elementCount; i++)
       {
 	elementData[i] = null;
       }
@@ -495,12 +498,12 @@
   /**
    * Creates a new Vector with the same contents as this one.
    */
-  public Object clone()
+  public synchronized Object clone()
   {
     try
       {
 	Vector clone = (Vector) super.clone();
-	clone.elementData = (Object[])elementData.clone();
+	clone.elementData = (Object[]) elementData.clone();
 	return clone;
       }
     catch (CloneNotSupportedException ex)
@@ -519,7 +522,7 @@
    * @returns An Object[] containing the contents of this Vector in order
    *
    */
-  public Object[] toArray()
+  public synchronized Object[] toArray()
   {
     Object[] newArray = new Object[elementCount];
     copyInto(newArray);
@@ -540,32 +543,20 @@
    * array.
    *
    *
-   * @param a An array to copy the Vector into if large enough
+   * @param array An array to copy the Vector into if large enough
    * @returns An array with the contents of this Vector in order
    * @throws ArrayStoreException the runtime type of the provided array
    * cannot hold the elements of the Vector
    */
-  public Object[] toArray(Object[]a)
+  public synchronized Object[] toArray(Object[] array)
   {
-    if (a.length >= elementCount)
-      {
-	copyInto(a);
-	if (a.length > elementCount)
-	  {
-	    a[elementCount] = null;
-	  }
-	return a;
-      }
+    if (array.length < elementCount)
+      array = (Object[]) Array.newInstance(array.getClass().getComponentType(), 
+        				   elementCount);
     else
-      {
-	//This seems like a kludge.. Is there a better way to find the
-	//Runtime type of an array?
-	Object[]newArray =
-	  (Object[])Array.newInstance(a.getClass().getComponentType(),
-				      elementCount);
-	copyInto(newArray);
-	return newArray;
-      }
+      array[elementCount] = null;
+    System.arraycopy(elementData, 0, array, 0, elementCount);
+    return array;
   }
 
   /**
@@ -575,7 +566,7 @@
    * @throws ArrayIndexOutOfBoundsException the index is not within the 
    * range of the Vector
    */
-  public Object get(int index)
+  public synchronized Object get(int index)
   {
     return elementAt(index);
   }
@@ -593,6 +584,17 @@
   }
 
   /**
+   * Adds an object to the Vector.
+   *
+   * @param o The element to add to the Vector
+   */
+  public synchronized boolean add(Object o)
+  {
+    addElement(o);
+    return true;
+  }
+
+  /**
    * Adds an object at the specified index.  Elements at or above
    * index are shifted up one position.
    *
@@ -612,9 +614,11 @@
    * @throws ArrayIndexOutOfBoundsException the index was out of the range
    * of the Vector
    */
-  public Object remove(int index)
+  public synchronized Object remove(int index)
   {
-    modCount++;
+    if (index >= elementCount)
+      throw new ArrayIndexOutOfBoundsException(index);
+  
     Object temp = elementData[index];
     removeElementAt(index);
     return temp;
@@ -628,22 +632,80 @@
     removeAllElements();
   }
 
+  public synchronized boolean containsAll(Collection c)
+  {
+    Iterator itr = c.iterator();
+    int size = c.size();
+    for (int pos = 0; pos < size; pos++)
+      {
+	if (!contains(itr.next()))
+	  return false;
+      }
+    return true;
+  }
+
+  public synchronized boolean addAll(Collection c)
+  {
+    return addAll(elementCount, c);
+  }
+  
+  public synchronized boolean removeAll(Collection c)
+  {
+    return super.removeAll(c);
+  }
+  
+  public synchronized boolean retainAll(Collection c)
+  {
+    return super.retainAll(c);
+  }
+
+  public synchronized boolean addAll(int index, Collection c)
+  {
+    if (index < 0 || index > elementCount)
+      throw new ArrayIndexOutOfBoundsException(index);
+    modCount++;
+    Iterator itr = c.iterator();
+    int csize = c.size();
+
+    ensureCapacity(elementCount + csize);
+    int end = index + csize;
+    if (elementCount > 0 && index != elementCount)
+      System.arraycopy(elementData, index, elementData, end, csize);
+    elementCount += csize;
+    for (; index < end; index++)
+      {
+        elementData[index] = itr.next();
+      }
+    return (csize > 0);  
+  }
+
+  public synchronized boolean equals(Object c)
+  {
+    return super.equals(c);
+  }
+
+  public synchronized int hashCode()
+  {
+    return super.hashCode();
+  }
+
   /**
    * Returns a string representation of this Vector in the form 
    * [element0, element1, ... elementN]
    *
    * @returns the String representation of this Vector
    */
-  public String toString()
+  public synchronized String toString()
   {
-    StringBuffer toReturn = new StringBuffer("[");
-    String comma = "";
+    String r = "[";
     for (int i = 0; i < elementCount; i++)
       {
-	toReturn.append(comma).append(elementData[i].toString());
-	comma = ", ";
+	r += elementData[i];
+	if (i < elementCount - 1)
+	  r += ", ";
       }
-    return toReturn.append("]").toString();
+    r += "]";
+    return r;
   }
 
   /**
@@ -654,21 +716,42 @@
    *
    * @returns an Enumeration
    */
-  public Enumeration elements()
+  public synchronized Enumeration elements()
   {
     return new Enumeration()
     {
       int i = 0;
-      public final boolean hasMoreElements()
+      public boolean hasMoreElements()
       {
-	return (i < size());
+	return (i < elementCount);
       }
-      public final Object nextElement()
+      public Object nextElement()
       {
-	if (i >= size())
+	if (i >= elementCount)
 	  throw new NoSuchElementException();
 	return (elementAt(i++));
       }
     };
+  }
+  
+  public List subList(int fromIndex, int toIndex)
+  {
+    List sub = super.subList(fromIndex, toIndex);
+    return Collections.synchronizedList(sub);
+  }
+  
+  /** @specnote This is not specified as synchronized in the JCL, but it seems
+    * to me that is should be. If it isn't, a clear() operation on a sublist
+    * will not be synchronized w.r.t. the Vector object.
+    */
+  protected synchronized void removeRange(int fromIndex, int toIndex)
+  {
+    modCount++;
+    if (fromIndex != toIndex)
+      {
+	System.arraycopy(elementData, toIndex, elementData, fromIndex, 
+	                 elementCount - toIndex);
+	elementCount -= (toIndex - fromIndex);
+      }
   }
-}				// Vector
+}

--------------919B584153876FE9BCA7B267--



From tmiller@haverford.edu Wed Nov 22 13:17:52 2000
Received: from 165-82-108-36.students.haverford.edu ([165.82.108.36] helo=azaphrael.haverford.edu)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13yeSu-0003zW-00; Wed, 22 Nov 2000 13:17:52 -0500
Received: from localhost (localhost [127.0.0.1])
	by azaphrael.haverford.edu (8.9.3/8.9.3) with ESMTP id NAA01004;
	Wed, 22 Nov 2000 13:21:08 -0500
Date: Wed, 22 Nov 2000 13:21:08 -0500 (EST)
From: "Todd L. Miller" <tmiller@haverford.edu>
To: Brian Jones <cbj@gnu.org>
cc: GNU Classpath <classpath@gnu.org>
Subject: Re: JOS
In-Reply-To: <m38zqdsx3s.fsf@rdu25-12-037.nc.rr.com>
Message-ID: <Pine.LNX.4.20.0011221318460.984-100000@azaphrael.haverford.edu>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

	Unfortunately, right after that email, I had to put a stall on my
JVM project, so I don't actually have any patches to for classpath.  John
Leuner may -- he's apparently gotten classpath working as the library for
linux kernel module JVM -- but, AFAIK, we don't have an official contact
yet.  I'll forward this to right place for JOS, and report back.

-_Quinn



From sballard@netreach.net Wed Nov 22 14:06:24 2000
Received: from fatuka.netreach.net ([207.29.195.121] helo=smtp.netreach.net)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13yfDs-0005x7-00
	for <classpath@gnu.org>; Wed, 22 Nov 2000 14:06:24 -0500
Received: (qmail 18652 invoked by uid 102); 22 Nov 2000 19:06:49 -0000
Received: from rainbow.netreach.net (HELO netreach.net) (@209.152.221.64)
  by smtp.netreach.net with SMTP; 22 Nov 2000 19:06:49 -0000
Message-ID: <3A1C1922.1F508DFF@netreach.net>
Date: Wed, 22 Nov 2000 14:06:10 -0500
From: Stuart Ballard <sballard@netreach.net>
X-Mailer: Mozilla 4.72 [en] (X11; I; Linux 2.2.17 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: GNU Classpath <classpath@gnu.org>
Subject: Re: classpath-0.01 released
References: <m3y9yecxuh.fsf@rdu25-12-037.nc.rr.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Brian Jones wrote:
> 
> I've packaged and released classpath-0.01.  This release features
> numerous improvements and additions.  Places to download the source
> code and precompiled binaries are listed below.

Awesome news, and great work!

Questions:
- Did any of the available interim versions of java.util.zip make it
into this release? This has been a showstopper for some of the things I
want to do with classpath.
- Have announcements been made to freshmeat and similar?

And one question that just occurred to me but doesn't have anything to
do with the release, really...

- Are there any plans to integrate any of the third-party VMs supported
(such as IBM's OpenRuntime, or any number of smaller projects whose
names I forget) into the main source tree? Ideally, I'd like to see
support for Kaffe as well, but this seems to get too much political
opposition (or at least non-enthusiasm) from both sides. At least it
would be nice to say "these are the VMs we support" rather than "this is
the VM we support".

Again, thanks (to everyone here) for all the effort put into this
release :)

Stuart.


From nferrier@tapsellferrier.co.uk Wed Nov 22 14:41:53 2000
Received: from tf2.tapsellferrier.co.uk ([195.7.242.227] helo=tapsellferrier.co.uk)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13yfmC-0007Nt-00
	for <classpath@gnu.org>; Wed, 22 Nov 2000 14:41:52 -0500
Received: from tf-Message_Server by tapsellferrier.co.uk
	with Novell_GroupWise; Wed, 22 Nov 2000 19:50:15 +0000
Message-Id: <sa1c2377.028@tapsellferrier.co.uk>
X-Mailer: Novell GroupWise 5.2
Date: Wed, 22 Nov 2000 19:42:37 +0000
From: "Nic Ferrier" <nferrier@tapsellferrier.co.uk>
To: classpath@gnu.org
Subject: Re: classpath-0.01 released
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Disposition: inline
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

>>> Stuart Ballard <sballard@netreach.net> 22-Nov-00 7:06:10 PM >>>

>Ideally, I'd like to see support for Kaffe as well, but 
>this seems to get too much political opposition 
>(or at least non-enthusiasm) from both sides. 

I have just started looking at Kaffe integration for Classpath.

I don't think it's that hard and it seems to me that should the work
get done then people's enthusiasm levels might increase.

So I'm going to start doing it and see what I can get done. I'll
report back here.


Nic Ferrier


From nferrier@tapsellferrier.co.uk Wed Nov 22 15:48:58 2000
Received: from tf2.tapsellferrier.co.uk ([195.7.242.227] helo=tapsellferrier.co.uk)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13ygp7-0002de-00
	for <classpath@gnu.org>; Wed, 22 Nov 2000 15:48:57 -0500
Received: from tf-Message_Server by tapsellferrier.co.uk
	with Novell_GroupWise; Wed, 22 Nov 2000 20:57:35 +0000
Message-Id: <sa1c333f.029@tapsellferrier.co.uk>
X-Mailer: Novell GroupWise 5.2
Date: Wed, 22 Nov 2000 20:48:38 +0000
From: "Nic Ferrier" <nferrier@tapsellferrier.co.uk>
To: classpath@gnu.org
Subject: Kaffe and Classpath (again)
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Disposition: inline
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

I've decided I'll take on the integration project in my spare time
(which is minimal) but I'm a little confused as to what the current
status is.

As far as I can see from the website and HACKING what has to be done
is:

- provide new Classpath versions of Kaffe's classes:
  o java.lang.Runtime
  o java.lang.Thread
  o java.lang.Class
  o java.lang.reflect.Constructor
  o java.lang.reflect.Method
  o java.lang.reflect.Field
These will have to do all the things that Kaffe's need to do in terms
of VM integration but must also talk to Classpath (ie: the example on
the website, Thread needs to call Classpath's ThreadGroup.add()
method).

- provide implementations of:
   o java.lang.VMObject
   o java.lang.VMClassLoader
   o java.lang.VMSystem
   o java.lang.VMSecurityManager
   o java.lang.StackFrame
   o java.lang.StackTrace
using Kaffe native methods, eg: VMClassLoader will use Kaffe's
defineClass0 native method to turn bytes into Class objects.

As has been stated *a lot* by people on this list (and on the CP
website) I don't think this is tremendously taxing.


Here are some questions that would make my job a whole lot easier:

- are the StackFrame and StackTrace classes supposed to go into
java.lang or gnu.vm.stack?

- how is the config system supposed to work? ie: how do we select
what the VM target is?
There is talk in the HACKING file of a kaffe-config file but I can't
find that anywhere (even on the mailing list where it is said to be).



Nic


From sballard@netreach.net Wed Nov 22 16:08:31 2000
Received: from fatuka.netreach.net ([207.29.195.121] helo=smtp.netreach.net)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13yh83-0003N3-00
	for <classpath@gnu.org>; Wed, 22 Nov 2000 16:08:31 -0500
Received: (qmail 813 invoked by uid 102); 22 Nov 2000 21:09:07 -0000
Received: from rainbow.netreach.net (HELO netreach.net) (@209.152.221.64)
  by smtp.netreach.net with SMTP; 22 Nov 2000 21:09:07 -0000
Message-ID: <3A1C35CE.BF2CC365@netreach.net>
Date: Wed, 22 Nov 2000 16:08:30 -0500
From: Stuart Ballard <sballard@netreach.net>
X-Mailer: Mozilla 4.72 [en] (X11; I; Linux 2.2.17 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Nic Ferrier <nferrier@tapsellferrier.co.uk>
CC: classpath@gnu.org
Subject: Re: Kaffe and Classpath (again)
References: <sa1c333f.029@tapsellferrier.co.uk>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Nic Ferrier wrote:
> 
> I've decided I'll take on the integration project in my spare time
> (which is minimal) but I'm a little confused as to what the current
> status is.

Yay!

> As far as I can see from the website and HACKING what has to be done
> is:
> 
> - provide new Classpath versions of Kaffe's classes:
>   o java.lang.Runtime
>   o java.lang.Thread
>   o java.lang.Class
>   o java.lang.reflect.Constructor
>   o java.lang.reflect.Method
>   o java.lang.reflect.Field
> These will have to do all the things that Kaffe's need to do in terms
> of VM integration but must also talk to Classpath (ie: the example on
> the website, Thread needs to call Classpath's ThreadGroup.add()
> method).

As I understand it (and I wasn't directly involved, but I was on the
mailing list at the time of the original discussions) this isn't
necessary. Classpath provides implementations of these classes, which
*use*...

>    o java.lang.VMObject
>    o java.lang.VMClassLoader
>    o java.lang.VMSystem
>    o java.lang.VMSecurityManager
>    o java.lang.StackFrame
>    o java.lang.StackTrace
> using Kaffe native methods, eg: VMClassLoader will use Kaffe's
> defineClass0 native method to turn bytes into Class objects.

... these classes to do the stuff that is *really* VM-specific. In other
words, it is *these* classes that should do the things that Kaffe's
existing versions do, and (the majority of) classpath integration is
handled by the publically visible classes. I'm not sure what the
discussion of Thread and ThreadGroup.add is about, because I would have
assumed that Classpath's Thread class could do that, but I'm sure
someone (John?) knows how this works and can clarify for us.

> - are the StackFrame and StackTrace classes supposed to go into
> java.lang or gnu.vm.stack?
> - how is the config system supposed to work? ie: how do we select
> what the VM target is?

Don't know about this stuff, sorry. My only guess would be wherever the
Japhar ones (in the vm/reference directory) are today... who was it that
wrote the build system?

Stuart.


From jewel@pixie.co.za Wed Nov 22 16:20:22 2000
Received: from jhb-imta.mweb.co.za ([196.2.48.244])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13yhJV-0003w9-00
	for <classpath@gnu.org>; Wed, 22 Nov 2000 16:20:22 -0500
Received: from pta-dial-196-31-187-19.mweb.co.za ([196.31.187.19])
 by jhb-imta.mweb.co.za
 (Sun Internet Mail Server sims.3.5.2000.03.23.18.03.p10) with ESMTP id
 <0G4G0066F34Y43@jhb-imta.mweb.co.za> for classpath@gnu.org; Wed,
 22 Nov 2000 23:17:24 +0200 (GMT-2)
Date: Wed, 22 Nov 2000 21:24:12 +0000 (UTC)
From: John Leuner <jewel@pixie.co.za>
Subject: Re: Kaffe and Classpath (again)
In-reply-to: <3A1C35CE.BF2CC365@netreach.net>
X-Sender: jewel@skami.mweb.co.za
To: Stuart Ballard <sballard@netreach.net>
Cc: Nic Ferrier <nferrier@tapsellferrier.co.uk>, classpath@gnu.org
Message-id: <Pine.LNX.4.21.0011222121250.1073-100000@skami.mweb.co.za>
MIME-version: 1.0
Content-type: TEXT/PLAIN; charset=US-ASCII
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

> > These will have to do all the things that Kaffe's need to do in terms
> > of VM integration but must also talk to Classpath (ie: the example on
> > the website, Thread needs to call Classpath's ThreadGroup.add()
> > method).
> 
> As I understand it (and I wasn't directly involved, but I was on the
> mailing list at the time of the original discussions) this isn't
> necessary. Classpath provides implementations of these classes, which
> *use*...
> 
> >    o java.lang.VMObject
> >    o java.lang.VMClassLoader
> >    o java.lang.VMSystem
> >    o java.lang.VMSecurityManager
> >    o java.lang.StackFrame
> >    o java.lang.StackTrace
> > using Kaffe native methods, eg: VMClassLoader will use Kaffe's
> > defineClass0 native method to turn bytes into Class objects.
> 
> ... these classes to do the stuff that is *really* VM-specific. 

The problem is that some VMs (my example being my JVM, kissme) have VM
specific stuff in those top classes, ie java/lang/Thread and
java/lang/Class etc.

That stuff really should be moved into the VM* classes, but since it is
tied in with a whole lot of native code, it's not going to be a pleasant
experience.

At the moment I have just added the Classpath code to my own, merging
whatever methods are common by hand. 

John Leuner



From tromey@creche.cygnus.com Wed Nov 22 16:32:10 2000
Received: from gasherbrumii.peakpeak.com ([207.174.178.17] helo=gash2.peakpeak.com)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13yhUv-0004Tb-00
	for <classpath@gnu.org>; Wed, 22 Nov 2000 16:32:09 -0500
Received: from creche.cygnus.com (04-185.026.popsite.net [64.24.105.185])
	by gash2.peakpeak.com (8.9.3/8.9.3) with ESMTP id OAA01846;
	Wed, 22 Nov 2000 14:32:05 -0700
Received: (from tromey@localhost)
	by creche.cygnus.com (8.9.3/8.9.3) id OAA06103;
	Wed, 22 Nov 2000 14:40:06 -0700
To: John Leuner <jewel@pixie.co.za>
Cc: classpath@gnu.org
Subject: Re: java.util.zip
References: <Pine.LNX.4.21.0011201734340.1073-100000@skami.mweb.co.za>
X-Zippy:  Somewhere in suburban Honolulu, an unemployed bellhop is whipping up
 a batch of illegal psilocybin chop suey!!
X-Attribution:  Tom
Reply-To: tromey@cygnus.com
From: Tom Tromey <tromey@cygnus.com>
Date: 22 Nov 2000 14:40:05 -0700
In-Reply-To: John Leuner's message of "Mon, 20 Nov 2000 17:39:51 +0000 (UTC)"
Message-ID: <87vgtfadoq.fsf@creche.cygnus.com>
Lines: 7
X-Mailer: Gnus v5.7/Emacs 20.5
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

>>>>> "John" == John Leuner <jewel@pixie.co.za> writes:

John> I assume CNI requires C++ support?

Yes.

Tom


From ckrintz@cs.ucsd.edu Wed Nov 22 17:10:19 2000
Received: from pacers.cs.utk.edu ([128.169.93.179] helo=redwood.cs.utk.edu)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13yi5r-000699-00
	for <classpath@gnu.org>; Wed, 22 Nov 2000 17:10:19 -0500
Received: by cs.ucsd.edu
	via sendmail from stdin
	id <m13ygjy-000UwZC@redwood.cs.utk.edu> (Debian Smail3.2.0.102)
	for classpath@gnu.org; Wed, 22 Nov 2000 15:43:38 -0500 (EST) 
Message-Id: <m13ygjy-000UwZC@redwood.cs.utk.edu>
From: ckrintz@cs.ucsd.edu (Chandra Krintz)
Subject: zip file of class files
To: classpath@gnu.org
Date: Wed, 22 Nov 2000 15:43:38 -0500 (EST)
X-Mailer: ELM [version 2.4ME+ PL65 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Hi.

Would it be possible for someone to compile
the .java files that are included in the new release
to .class files and make some sort of archive of
them available?

I would like to use them directly without the
rest of the infrastructure.  Others have shown
interest (from the ORP mailing list) in 
this also.  In an early release a .zip file enabled
this type of access.

Thanks much!
Chandra Krintz


From nferrier@tapsellferrier.co.uk Wed Nov 22 19:16:24 2000
Received: from tf2.tapsellferrier.co.uk ([195.7.242.227] helo=tapsellferrier.co.uk)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13yk3r-0002Hu-00
	for <classpath@gnu.org>; Wed, 22 Nov 2000 19:16:23 -0500
Received: from tf-Message_Server by tapsellferrier.co.uk
	with Novell_GroupWise; Thu, 23 Nov 2000 00:25:00 +0000
Message-Id: <sa1c63dc.030@tapsellferrier.co.uk>
X-Mailer: Novell GroupWise 5.2
Date: Thu, 23 Nov 2000 00:16:45 +0000
From: "Nic Ferrier" <nferrier@tapsellferrier.co.uk>
To: classpath@gnu.org
Subject: Re: Kaffe and Classpath (again)
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Disposition: inline
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

>>> John Leuner <jewel@pixie.co.za> 22-Nov-00 9:24:12 PM >>>

>The problem is that some VMs (my example being my 
>JVM, kissme) have VM specific stuff in those top classes,
>ie java/lang/Thread and java/lang/Class etc.
>That stuff really should be moved into the VM* classes, but 
>since it is tied in with a whole lot of native code, it's not going

>to be a pleasant experience.

As far as I can see the website advice for VM integrators is pretty
accurate.

The situation is that Classpath has an almost complete java.*
implementation which uses gnu.java.* classes to do the implementation
hidden stuff.

For those java.lang classes I mentioned there needs to be one
implementation for each VM that is integrated. So, using Class as an
example, with current aims there will be:

   japhar/java.lang.Class
   kaffe/java.lang.Class
   eletrical-fire/java.lang.Class
   etc...

Each one of these VM specific versions needs to do certain things to
make Classpath work. Those things seem to be well documented in the
"hooks" section on the VM-integrators page.

It's not too onerous a requirement IMHO to have one set of these
classes per VM.


>At the moment I have just added the Classpath code to my 
>own, merging whatever methods are common by hand. 

That's one approach. I'm going to start by taking the other
approach... changing the Kaffe classes so that they use the Classpath
hooks.

If I can achieve that Kaffe *could* be altered to use adapted
versions of those classes on some configuration request (or even
dynamically but I'm not going to aim for that) and Classpath *could*
include the altered Kaffe classes in the existing:

   vm/kaffe

directory.


What I really need to know about is the history and workings of the
config system.

I presume it's just a system for selecting which version of the
classes under vm/ is compiled into the classes.jar.


The other thing I would quite like to know is more about StackFrame.
As I understand it I have to implement gnu.vm.stack.StackFrame  but
there is a gnu.java.StackFrame as well. 

I would be very gratefull if some classpath old-hand would explain
the intention with all of that. 


Nic


From bryce@albatross.co.nz Wed Nov 22 19:21:51 2000
Received: from waitaki.otago.ac.nz ([139.80.75.140])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13yk98-00037f-00
	for <classpath@gnu.org>; Wed, 22 Nov 2000 19:21:50 -0500
Received: from albatross.co.nz (IDENT:bryce@reason.otago.ac.nz [139.80.75.188])
	by waitaki.otago.ac.nz (8.9.3/8.9.3) with ESMTP id NAA23239;
	Thu, 23 Nov 2000 13:21:39 +1300
Message-ID: <3A1C63E4.F3BA772E@albatross.co.nz>
Date: Thu, 23 Nov 2000 13:25:09 +1300
From: Bryce McKinlay <bryce@albatross.co.nz>
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.4.0-test9 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: tromey@cygnus.com
CC: Classpath Project <classpath@gnu.org>
Subject: Re: cvs2cl
References: <878zqeo5tt.fsf@creche.cygnus.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Tom Tromey wrote:

> I noticed that Brian regenerated ChangeLog using cvs2cl yesterday.
> This made it much less readable, imho.

I agree. Generated cvs changelogs suck. The main problem is that its
difficult to specify different text for different groups of changed files
without doing multiple check-ins. People also tend to write more interesting
and useful entries when the ChangeLog is maintained by hand.

1 vote for a hand maintained ChangeLog.

  [ bryce ]




From bryce@albatross.co.nz Thu Nov 23 00:02:58 2000
Received: from waitaki.otago.ac.nz ([139.80.75.140])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13yoXA-0004N3-00
	for <classpath@gnu.org>; Thu, 23 Nov 2000 00:02:56 -0500
Received: from albatross.co.nz (IDENT:bryce@reason.otago.ac.nz [139.80.75.188])
	by waitaki.otago.ac.nz (8.9.3/8.9.3) with ESMTP id SAA24081;
	Thu, 23 Nov 2000 18:02:41 +1300
Message-ID: <3A1CA5C4.254DC238@albatross.co.nz>
Date: Thu, 23 Nov 2000 18:06:12 +1300
From: Bryce McKinlay <bryce@albatross.co.nz>
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.4.0-test9 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: java-discuss@sources.redhat.com, classpath@gnu.org
Subject: A java DEBUG flag
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

I think it would be useful to have a java file somewhere that defines
constants which can be used to enable/disable bits of java code at
compile time. The serialized object dumping facility in
ObjectInputStream, for example, currently adds a lot of overhead even
when its turned off, because the java code makes calls into native code
regardless of whether debugging is enabled or not. By setting a constant
somewhere, the compiler can optimize it out when not doing a debug
build.

I propose an interface (this ensures that only constants can be defined)
called gnu.classpath.Flags (or CompileSettings? DebugFlags?) which would
be generated at configure time from a Flags.java.in, and would change
based on whether "--enable-libgcj-debug" (or the classpath equivilant)
was specified at configure time.

Initially it will look something like this:

package gnu.classpath;

public interface Flags
{
  boolean SERIALIZATION_DEBUGGING = true;
}

Comments?

regards

  [ bryce ]




From tromey@creche.cygnus.com Thu Nov 23 01:30:13 2000
Received: from gasherbrumii.peakpeak.com ([207.174.178.17] helo=gash2.peakpeak.com)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13yptd-0007Iv-00
	for <classpath@gnu.org>; Thu, 23 Nov 2000 01:30:13 -0500
Received: from creche.cygnus.com (04-161.026.popsite.net [64.24.105.161])
	by gash2.peakpeak.com (8.9.3/8.9.3) with ESMTP id XAA23136;
	Wed, 22 Nov 2000 23:30:06 -0700
Received: (from tromey@localhost)
	by creche.cygnus.com (8.9.3/8.9.3) id XAA10209;
	Wed, 22 Nov 2000 23:38:12 -0700
To: Bryce McKinlay <bryce@albatross.co.nz>
Cc: java-discuss@sources.redhat.com, classpath@gnu.org
Subject: Re: A java DEBUG flag
References: <3A1CA5C4.254DC238@albatross.co.nz>
X-Zippy:  It's OKAY --- I'm an INTELLECTUAL, too.
X-Attribution:  Tom
Reply-To: tromey@cygnus.com
From: Tom Tromey <tromey@cygnus.com>
Date: 22 Nov 2000 23:38:11 -0700
In-Reply-To: Bryce McKinlay's message of "Thu, 23 Nov 2000 18:06:12 +1300"
Message-ID: <874s0z6vn0.fsf@creche.cygnus.com>
Lines: 15
X-Mailer: Gnus v5.7/Emacs 20.5
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

>>>>> "Bryce" == Bryce McKinlay <bryce@albatross.co.nz> writes:

Bryce> I propose an interface (this ensures that only constants can be
Bryce> defined) called gnu.classpath.Flags (or CompileSettings?
Bryce> DebugFlags?)

Maybe `Configuration'?

    if (Configuration.DEBUG) { ... }

Bryce> Comments?

Sounds like a good idea to me.

Tom


From mark@klomp.org Thu Nov 23 03:36:07 2000
Received: from node1c5f.a2000.nl ([62.108.28.95] helo=nescio.klomp.org)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13yrrT-0002ob-00
	for <classpath@gnu.org>; Thu, 23 Nov 2000 03:36:07 -0500
Received: from mark by nescio.klomp.org with local (Exim 3.16 #1 (Debian))
	id 13yrrK-0008ND-00; Thu, 23 Nov 2000 09:35:58 +0100
Date: Thu, 23 Nov 2000 09:35:58 +0100
From: Mark Wielaard <mark@klomp.org>
To: Tom Tromey <tromey@cygnus.com>
Cc: Classpath Project <classpath@gnu.org>
Subject: Re: cvs2cl
Message-ID: <20001123093558.A32159@klomp.org>
References: <878zqeo5tt.fsf@creche.cygnus.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5i
In-Reply-To: <878zqeo5tt.fsf@creche.cygnus.com>; from tromey@cygnus.com on Mon, Nov 20, 2000 at 11:29:18AM -0700
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Hi,

On Mon, Nov 20, 2000 at 11:29:18AM -0700, Tom Tromey wrote:
> I noticed that Brian regenerated ChangeLog using cvs2cl yesterday.
> This made it much less readable, imho.  For instance the correctly
> ("correct" being determined by the GNU coding standards) formatted
> entry:
> 
> 2000-11-17  Mark Wielaard  <mark@klomp.org>
> 
>         Merge with libgcj (fixes by Anthony Green and Mark Wielaard)
>         * java/util/jar/Attributes.java: Correct comment spelling.
>         * java/util/jar/JarFile.java: Constructor that takes mode now
>         calls super. Don't call java.util.zip.ZipFile.getEntry twice.
> 
> changed to this:
> 
> 2000-11-18 01:06  Mark Wielaard <mark@klomp.org>
> 
> 	* java/util/jar/: Attributes.java, JarFile.java:	Merge with
> 	libgcj (fixes by Anthony Green and Mark Wielaard)
> 		* java/util/jar/Attributes.java: Correct comment spelling.
> 		* java/util/jar/JarFile.java: Constructor that takes mode
> 	now
> 		calls super. Don't call java.util.zip.ZipFile.getEntry
> 	twice.

Although I would like to have a hand maintained ChangeLog file please
note that the cvs2cl utility has a --fsf flag that generates the somewhat
more readable:

2000-11-18 02:06  mark

        * ChangeLog, java/util/jar/Attributes.java,
        java/util/jar/JarFile.java: Merge with libgcj (fixes by Anthony
        Green and Mark Wielaard) 
        * java/util/jar/Attributes.java: Correct comment spelling.  
        * java/util/jar/JarFile.java: Constructor that takes mode now calls
        super. Don't call java.util.zip.ZipFile.getEntry twice.  

Cheers,

Mark


From mark@klomp.org Thu Nov 23 03:53:03 2000
Received: from node1c5f.a2000.nl ([62.108.28.95] helo=nescio.klomp.org)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13ys7r-0003zM-00
	for <classpath@gnu.org>; Thu, 23 Nov 2000 03:53:03 -0500
Received: from mark by nescio.klomp.org with local (Exim 3.16 #1 (Debian))
	id 13ys7U-0008Od-00; Thu, 23 Nov 2000 09:52:40 +0100
Date: Thu, 23 Nov 2000 09:52:40 +0100
From: Mark Wielaard <mark@klomp.org>
To: Bryce McKinlay <bryce@albatross.co.nz>
Cc: java-discuss@sources.redhat.com, classpath@gnu.org
Subject: Re: A java DEBUG flag
Message-ID: <20001123095240.B32159@klomp.org>
References: <3A1CA5C4.254DC238@albatross.co.nz>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5i
In-Reply-To: <3A1CA5C4.254DC238@albatross.co.nz>; from bryce@albatross.co.nz on Thu, Nov 23, 2000 at 06:06:12PM +1300
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Hi,

On Thu, Nov 23, 2000 at 06:06:12PM +1300, Bryce McKinlay wrote:
> I think it would be useful to have a java file somewhere that defines
> constants which can be used to enable/disable bits of java code at
> compile time. The serialized object dumping facility in
> ObjectInputStream, for example, currently adds a lot of overhead even
> when its turned off, because the java code makes calls into native code
> regardless of whether debugging is enabled or not. By setting a constant
> somewhere, the compiler can optimize it out when not doing a debug
> build.
Does it has to be a compile time constansts?
I would like to be able to have a class library where I could selectively
turn on/off some debugging code at runtime. If we used something like
System.getProperty("gnu.java.io.serialisation.debugging") to set these
constants (you can even do that in an interface) then at runtime the
user could do a -Dgnu.java.io.serialisation.debugging=yes to turn on
the debugging. (It should at least be an option to have these 'constants'
compiled in so they can be set dynamicly at runtime IMHO.)

Cheers,

Mark


From nferrier@tapsellferrier.co.uk Thu Nov 23 09:30:41 2000
Received: from tf2.tapsellferrier.co.uk ([195.7.242.227] helo=tapsellferrier.co.uk)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13yxOb-00031D-00
	for <classpath@gnu.org>; Thu, 23 Nov 2000 09:30:41 -0500
Received: from tf-Message_Server by tapsellferrier.co.uk
	with Novell_GroupWise; Thu, 23 Nov 2000 14:39:21 +0000
Message-Id: <sa1d2c19.036@tapsellferrier.co.uk>
X-Mailer: Novell GroupWise 5.2
Date: Thu, 23 Nov 2000 14:36:45 +0000
From: "Nic Ferrier" <nferrier@tapsellferrier.co.uk>
To: classpath@gnu.org
Subject: Kaffe integration
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Disposition: inline
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

java.lang.Throwable depends on:

   gnu.vm.stack.StackTrace 
   gnu.vm.stack.StackFrame

so I think that one of 2 things needs to be done to allow VM
integration:

- a requirement to implement gnu.vm.stack classes
- a requirement to implement java.lang.Throwable


It seems to me that the former is the best plan. This just means that
the vm-integration page on the website needs to be updated to state
clearly that the package:

   gnu.vm.stack 

needs to be implemented (and probably "there is an example
implementation (for japhar) in vm/reference").


Nic


From bryce@albatross.co.nz Thu Nov 23 18:16:01 2000
Received: from waitaki.otago.ac.nz ([139.80.75.140])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13z5ay-0004lC-00
	for <classpath@gnu.org>; Thu, 23 Nov 2000 18:16:00 -0500
Received: from albatross.co.nz (IDENT:bryce@reason.otago.ac.nz [139.80.75.188])
	by waitaki.otago.ac.nz (8.9.3/8.9.3) with ESMTP id MAA26674;
	Fri, 24 Nov 2000 12:15:32 +1300
Message-ID: <3A1DA5EB.E0698D4C@albatross.co.nz>
Date: Fri, 24 Nov 2000 12:19:07 +1300
From: Bryce McKinlay <bryce@albatross.co.nz>
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.4.0-test9 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Mark Wielaard <mark@klomp.org>
CC: java-discuss@sources.redhat.com, classpath@gnu.org
Subject: Re: A java DEBUG flag
References: <3A1CA5C4.254DC238@albatross.co.nz> <20001123095240.B32159@klomp.org>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Mark Wielaard wrote:

> Does it has to be a compile time constansts?

Yes.

> If we used something like
> System.getProperty("gnu.java.io.serialisation.debugging") to set these
> constants (you can even do that in an interface)

I don't think so ;-)

> then at runtime the
> user could do a -Dgnu.java.io.serialisation.debugging=yes to turn on
> the debugging. (It should at least be an option to have these 'constants'
> compiled in so they can be set dynamicly at runtime IMHO.)

We'll do that as well. To get the debug output you'll have to configure with
"--enable-libgcj-debug" _and_ set the runtime property. But we do want to
have the compile time constants in order to ensure there is _no_ overhead
from the debugging code for an optimized build.

regards

  [ bryce ]




From mark@klomp.org Fri Nov 24 03:33:19 2000
Received: from node1c5f.a2000.nl ([62.108.28.95] helo=nescio.klomp.org)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13zEIJ-0005R3-00
	for <classpath@gnu.org>; Fri, 24 Nov 2000 03:33:19 -0500
Received: from mark by nescio.klomp.org with local (Exim 3.16 #1 (Debian))
	id 13zEI2-0001Vi-00; Fri, 24 Nov 2000 09:33:02 +0100
Date: Fri, 24 Nov 2000 09:33:02 +0100
From: Mark Wielaard <mark@klomp.org>
To: Bryce McKinlay <bryce@albatross.co.nz>
Cc: java-discuss@sources.redhat.com, classpath@gnu.org
Subject: Re: A java DEBUG flag
Message-ID: <20001124093302.A5495@klomp.org>
References: <3A1CA5C4.254DC238@albatross.co.nz> <20001123095240.B32159@klomp.org> <3A1DA5EB.E0698D4C@albatross.co.nz>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5i
In-Reply-To: <3A1DA5EB.E0698D4C@albatross.co.nz>; from bryce@albatross.co.nz on Fri, Nov 24, 2000 at 12:19:07PM +1300
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Hi,

On Fri, Nov 24, 2000 at 12:19:07PM +1300, Bryce McKinlay wrote:
> Mark Wielaard wrote:
> 
> > If we used something like
> > System.getProperty("gnu.java.io.serialisation.debugging") to set these
> > constants (you can even do that in an interface)
> 
> I don't think so ;-)

What don't you think so?
What I meant was something like the following:

interface I {
    boolean SERIALISATION_DEBUG =
        Boolean.getBoolean("gnu.java.io.serialisation.debugging");
}
 
> > then at runtime the
> > user could do a -Dgnu.java.io.serialisation.debugging=yes to turn on
> > the debugging. (It should at least be an option to have these 'constants'
> > compiled in so they can be set dynamicly at runtime IMHO.)
> 
> We'll do that as well. To get the debug output you'll have to configure with
> "--enable-libgcj-debug" _and_ set the runtime property. But we do want to
> have the compile time constants in order to ensure there is _no_ overhead
> from the debugging code for an optimized build.

OK that is fair. But I would really like to see a solution that generated
a interface like I did above so that you could dynamicly turn it on and
a compile time configuration option to turn the above interface into

interface I {
    boolean SERIALISATION_DEBUG = false;
}

But I think the default should be that you can dynamicly turn the debugging
code on.

Cheers,

Mark


From bryce@albatross.co.nz Fri Nov 24 05:20:17 2000
Received: from 210-55-152-59.adsl.xtra.co.nz ([210.55.152.59] helo=localhost.localdomain)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13zFxo-0000O7-00
	for <classpath@gnu.org>; Fri, 24 Nov 2000 05:20:17 -0500
Received: from albatross.co.nz (IDENT:bryce@hokkaido [127.0.0.1])
	by localhost.localdomain (8.11.0/8.11.0) with ESMTP id eAOALDo01043;
	Fri, 24 Nov 2000 23:21:13 +1300
Message-ID: <3A1E4118.FB916F24@albatross.co.nz>
Date: Fri, 24 Nov 2000 23:21:12 +1300
From: Bryce McKinlay <bryce@albatross.co.nz>
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.4.0-test10 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Mark Wielaard <mark@klomp.org>
CC: java-discuss@sources.redhat.com, classpath@gnu.org
Subject: Re: A java DEBUG flag
References: <3A1CA5C4.254DC238@albatross.co.nz> <20001123095240.B32159@klomp.org> <3A1DA5EB.E0698D4C@albatross.co.nz> <20001124093302.A5495@klomp.org>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Mark Wielaard wrote:

> > > If we used something like
> > > System.getProperty("gnu.java.io.serialisation.debugging") to set these
> > > constants (you can even do that in an interface)
> >
> > I don't think so ;-)
>
> What don't you think so?
> What I meant was something like the following:
>
> interface I {
>     boolean SERIALISATION_DEBUG =
>         Boolean.getBoolean("gnu.java.io.serialisation.debugging");
> }

But this is not a constant! It can only be evaluated at runtime. A bytecode
compiler or native compiler like gcj cannot determine the value of the flag, and
must emit a branch. Granted, a smart JIT will be able to take advantage of it
being a final field and eliminate the debug code from what it generates. However,
if it is used extensively in the runtime then it is probable that the flag would
be evaluated before the system property is set, making it useless.

regards

  [ bryce ]




From mark@klomp.org Fri Nov 24 06:28:06 2000
Received: from node1c5f.a2000.nl ([62.108.28.95] helo=nescio.klomp.org)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13zH1S-00075T-00
	for <classpath@gnu.org>; Fri, 24 Nov 2000 06:28:06 -0500
Received: from mark by nescio.klomp.org with local (Exim 3.16 #1 (Debian))
	id 13zH1D-0001nL-00; Fri, 24 Nov 2000 12:27:51 +0100
Date: Fri, 24 Nov 2000 12:27:51 +0100
From: Mark Wielaard <mark@klomp.org>
To: Bryce McKinlay <bryce@albatross.co.nz>
Cc: java-discuss@sources.redhat.com, classpath@gnu.org
Subject: Re: A java DEBUG flag
Message-ID: <20001124122751.A6660@klomp.org>
References: <3A1CA5C4.254DC238@albatross.co.nz> <20001123095240.B32159@klomp.org> <3A1DA5EB.E0698D4C@albatross.co.nz> <20001124093302.A5495@klomp.org> <3A1E4118.FB916F24@albatross.co.nz>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5i
In-Reply-To: <3A1E4118.FB916F24@albatross.co.nz>; from bryce@albatross.co.nz on Fri, Nov 24, 2000 at 11:21:12PM +1300
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Hi,

On Fri, Nov 24, 2000 at 11:21:12PM +1300, Bryce McKinlay wrote:
> Mark Wielaard wrote:
> 
> > What I meant was something like the following:
> >
> > interface I {
> >     boolean SERIALISATION_DEBUG =
> >         Boolean.getBoolean("gnu.java.io.serialisation.debugging");
> > }
> 
> But this is not a constant! It can only be evaluated at runtime. A bytecode
> compiler or native compiler like gcj cannot determine the value of the flag,
> and must emit a branch.
OK. We had different goals. You wanted to disable the code during compile
time. I wanted to be able to enable the code during runtime.

I should have read java-patches were you write:
> To get the debugging output, you should do BOTH of the following:
>   
> 1. configure libgcj with "--enable-libgcj-debug"
> 2. set "gcj.dumpobjects" to anything other than null or the empty
> string. You can do this by calling System.setProperty("gcj.dumpobjects",
> "true") in program code or by using "-Dgcj.dumpobjects=true" on the gcj
> command line.
So you already solved both our problems. Thanks.

> Granted, a smart JIT will be able to take advantage
> of it being a final field and eliminate the debug code from what it
> generates. However, if it is used extensively in the runtime then it is
> probable that the flag would be evaluated before the system property is set,
> making it useless.
So you have to reevaluate the System property every time (as you seem to
do in your ObjectInputStream patch). I was hoping that you could set system
properties on the command line with gcj compiled binaries then you could
be sure that such a property was always set (or not set). This is what
gij seems to do.

Cheers,

Mark


From brandon@zenemusic.hu Fri Nov 24 08:28:49 2000
Received: from zenemusic2.city.tvnet.hu ([195.38.101.219] helo=shelter.zenemusic.hu)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13zIu7-0002B1-00
	for <classpath@gnu.org>; Fri, 24 Nov 2000 08:28:39 -0500
Received: from maestro.zenemusic.hu (maestro2.zenemusic.hu [10.1.2.9])
	by shelter.zenemusic.hu (8.9.3/8.9.3) with SMTP id OAA13721
	for <classpath@gnu.org>; Fri, 24 Nov 2000 14:27:27 +0100 (CET)
	(envelope-from brandon@zenemusic.hu)
Message-Id: <200011241327.OAA13721@shelter.zenemusic.hu>
Received: from 10.1.1.5 ([10.1.1.5]) by maestro.zenemusic.hu (WinRoute Pro 4.1) with SMTP; Fri, 24 Nov 2000 14:04:22 +0100
Date: Fri, 24 Nov 2000 14:01:48 +0100
From: "Brandon Krueger" <brandon@zenemusic.hu>
To: classpath@gnu.org
Subject: Resource: Music of Eastern Europe via MP3
MIME-Version: 1.0
Content-type: text/plain
X-Mailer: FletMail 1.103
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Hello,

Note: this is a one-time mailing.

Finally a site has launched which solves two answers which music consumers and artists have faced recently: 1)  Selling tracks via MP3 while compensating artists and 2) Selling tracks from Central and Eastern Europe to which an eager North American audience hasn't had access. Try out XenoMusic.com today and download one track for FREE. We are a new website and are eager for feedback from fellow artists, music lovers, industry writers, and academics who have interest in both our legal arrangements with artists, as well as exotic music from places such as Russia, Poland, Bulgaria, Hungary and more.


CHECK US OUT TODAY AT:      www.xenomusic.com


Or contact me personally at brandon@zenemusic.hu




From bryce@albatross.co.nz Fri Nov 24 17:33:13 2000
Received: from 210-54-232-25.adsl.xtra.co.nz ([210.54.232.25] helo=localhost.localdomain)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13zRP6-0003HN-00
	for <classpath@gnu.org>; Fri, 24 Nov 2000 17:33:13 -0500
Received: from albatross.co.nz (IDENT:bryce@hokkaido [127.0.0.1])
	by localhost.localdomain (8.11.0/8.11.0) with ESMTP id eAOMXN801016;
	Sat, 25 Nov 2000 11:33:23 +1300
Message-ID: <3A1EECB2.66342D7F@albatross.co.nz>
Date: Sat, 25 Nov 2000 11:33:22 +1300
From: Bryce McKinlay <bryce@albatross.co.nz>
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.4.0-test10 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Mark Wielaard <mark@klomp.org>
CC: java-discuss@sources.redhat.com, classpath@gnu.org
Subject: Re: A java DEBUG flag
References: <3A1CA5C4.254DC238@albatross.co.nz> <20001123095240.B32159@klomp.org> <3A1DA5EB.E0698D4C@albatross.co.nz> <20001124093302.A5495@klomp.org> <3A1E4118.FB916F24@albatross.co.nz> <20001124122751.A6660@klomp.org>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Mark Wielaard wrote:

> > Granted, a smart JIT will be able to take advantage
> > of it being a final field and eliminate the debug code from what it
> > generates. However, if it is used extensively in the runtime then it is
> > probable that the flag would be evaluated before the system property is set,
> > making it useless.
> So you have to reevaluate the System property every time (as you seem to
> do in your ObjectInputStream patch). I was hoping that you could set system
> properties on the command line with gcj compiled binaries then you could
> be sure that such a property was always set (or not set). This is what
> gij seems to do.

Setting a property on the command line doesn't stop the application changing that
property later. I figured it would be nice to be able to disable debugging at
runtime as well, hence the re-checking the property in the ObjectInputStream
constructor.

regards

  [ bryce ]




From burton@relativity.yi.org Fri Nov 24 20:16:01 2000
Received: from kestrel.prod.itd.earthlink.net ([207.217.121.155])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13zTwe-0007mQ-00
	for <classpath@gnu.org>; Fri, 24 Nov 2000 20:16:01 -0500
Received: from relativity.yi.org (user-vcaumir.dsl.mindspring.com [216.175.90.91])
	by kestrel.prod.itd.earthlink.net (8.9.3-EL_1_3/8.9.3) with ESMTP id RAA28110;
	Fri, 24 Nov 2000 17:15:53 -0800 (PST)
Received: from localhost.localdomain (IDENT:root@user-vcaumed.dsl.mindspring.com [216.175.89.205])
	by relativity.yi.org (8.9.3/8.8.7) with ESMTP id FAA21222;
	Sat, 25 Nov 2000 05:15:51 -0800
Received: (from burton@localhost)
	by localhost.localdomain (8.11.0/8.11.0) id eAP1CQf24073;
	Fri, 24 Nov 2000 17:12:26 -0800
Resent-Message-Id: <200011250112.eAP1CQf24073@localhost.localdomain>
X-Authentication-Warning: localhost.localdomain: burton set sender to burton@relativity.yi.org using -f
Resent-To: java-discuss@sources.redhat.com, classpath@gnu.org
Resent-From: burton@relativity.yi.org (Kevin A. Burton)
Resent-Date: 24 Nov 2000 17:07:05 -0800
X-From-Line: nobody Thu Nov 23 19:44:33 2000
To: Mark Wielaard <mark@klomp.org>
Cc: Bryce McKinlay <bryce@albatross.co.nz>, java-discuss@sources.redhat.com,
        classpath@gnu.org
Subject: Re: A java DEBUG flag
References: <3A1CA5C4.254DC238@albatross.co.nz>
	<20001123095240.B32159@klomp.org>
From: burton@relativity.yi.org (Kevin A. Burton)
Date: 23 Nov 2000 19:44:32 -0800
In-Reply-To: Mark Wielaard's message of "Thu, 23 Nov 2000 09:52:40 +0100"
Message-ID: <m3lmuaf2zj.fsf@relativity.yi.org>
User-Agent: Gnus/5.0807 (Gnus v5.8.7) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Lines: 49
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Mark Wielaard <mark@klomp.org> writes:

<snip>

> Does it has to be a compile time constansts?
> I would like to be able to have a class library where I could selectively
> turn on/off some debugging code at runtime. If we used something like
> System.getProperty("gnu.java.io.serialisation.debugging") to set these
> constants (you can even do that in an interface) then at runtime the
> user could do a -Dgnu.java.io.serialisation.debugging=yes to turn on
> the debugging. (It should at least be an option to have these 'constants'
> compiled in so they can be set dynamicly at runtime IMHO.)

SUN has a JSR for JDK 1.4 (Merlin) for assertions.  Basically you can:

assert i > 4;

And then prior at runtime you can turn this off... I would say any
implementation of this should use the JSR for a spec.  I read it and although at
first I didn't like it... I came around and now think it is the best.

As an asside... I have always wanted a standard place to read java properties
files instead of having to pass -D to the JVM.  Maybe we need to have an
/etc/java.properties file that GJC has to read.  Since -D won't work for
statically compiled apps in GJC maybe this is the better way.

Kevin

- -- 
Kevin A. Burton ( burton@apache.org, burton@openprivacy.org, burtonator@acm.org )
             Cell: 408-910-6145 URL: http://relativity.yi.org

proprietary == evil
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.4 (GNU/Linux)
Comment: Get my public key at: http://relativity.yi.org/pgpkey.txt

iD8DBQE6HeQgAwM6xb2dfE0RAnuzAJ4lWvcpBx2FNPg7CYRHd3/YNX1YEACeMTfM
HHXImBMHweABt3FrXcb/Rvc=
=WYmK
-----END PGP SIGNATURE-----



cracking Clinton Peking BATF $400 million in gold bullion arrangements PLO
strategic Cocaine assassination Mossad Marxist South Africa Legion of Doom DES


From arenn@urbanophile.com Sat Nov 25 07:34:47 2000
Received: from mail.enteract.com ([207.229.143.33])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13zeXX-0004Nh-00; Sat, 25 Nov 2000 07:34:47 -0500
Received: from urbanophile.com (IDENT:root@216-80-74-221.dsl.enteract.com [216.80.74.221])
	by mail.enteract.com (8.9.3/8.9.3) with ESMTP id GAA71720;
	Sat, 25 Nov 2000 06:34:46 -0600 (CST)
	(envelope-from arenn@urbanophile.com)
From: arenn@urbanophile.com
Received: (from arenn@localhost)
	by urbanophile.com (8.9.3/8.8.7) id HAA02306;
	Sat, 25 Nov 2000 07:38:55 -0600
Date: Sat, 25 Nov 2000 07:38:50 -0600
To: Brian Jones <cbj@gnu.org>
Cc: GNU Classpath <classpath@gnu.org>
Subject: Re: JOS
Message-ID: <20001125073850.P2100@urbanophile.com>
Mail-Followup-To: Brian Jones <cbj@gnu.org>,
	GNU Classpath <classpath@gnu.org>
References: <m38zqdsx3s.fsf@rdu25-12-037.nc.rr.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Mailer: Mutt 1.0.1i
In-Reply-To: <m38zqdsx3s.fsf@rdu25-12-037.nc.rr.com>; from cbj@gnu.org on Tue, Nov 21, 2000 at 12:38:47AM -0500
Organization: GNU's Not Unix!
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Brian Jones (cbj@gnu.org) wrote:
> The following article was interesting to me.  I don't know what the
> state of the JOS project is or who our liason is.  Could someone from
> that project bring us up to date with respect to this email referenced
> below?
> 
> http://www.jos.org/pipermail/arch/1999-December/000358.html

It is interesting.  If I might extract their points:

1.  They think that replacement classes should be written in an alternate
namespace, then ported over when completed.  This certainly solves some
chicken and egg problems, but it creates new ones of its own.  I don't
object to that approach, but what we currently have seems to be working
fine.

2.  Classpath/libgcj merger.  No impact to JOS except that perhaps
JOS could be compiled at some point.

3.  VM specific aspects of Classpath.  Yes, we have them.  I don't see
any way around them.  And as noted, we do keep them separate from the
main tree.

This is what I gather after a very brief review.

-- 
Aaron M. Renn (arenn@urbanophile.com) http://www.urbanophile.com/arenn/


From arenn@urbanophile.com Sat Nov 25 07:47:13 2000
Received: from mail.enteract.com ([207.229.143.33])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13zejY-0004m8-00
	for <classpath@gnu.org>; Sat, 25 Nov 2000 07:47:12 -0500
Received: from urbanophile.com (IDENT:root@216-80-74-221.dsl.enteract.com [216.80.74.221])
	by mail.enteract.com (8.9.3/8.9.3) with ESMTP id GAA75802;
	Sat, 25 Nov 2000 06:47:12 -0600 (CST)
	(envelope-from arenn@urbanophile.com)
From: arenn@urbanophile.com
Received: (from arenn@localhost)
	by urbanophile.com (8.9.3/8.8.7) id HAA02341;
	Sat, 25 Nov 2000 07:51:20 -0600
Date: Sat, 25 Nov 2000 07:51:10 -0600
To: Chandra Krintz <ckrintz@cs.ucsd.edu>
Cc: classpath@gnu.org
Subject: Re: zip file of class files
Message-ID: <20001125075110.S2100@urbanophile.com>
Mail-Followup-To: Chandra Krintz <ckrintz@cs.ucsd.edu>, classpath@gnu.org
References: <m13ygjy-000UwZC@redwood.cs.utk.edu>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Mailer: Mutt 1.0.1i
In-Reply-To: <m13ygjy-000UwZC@redwood.cs.utk.edu>; from ckrintz@cs.ucsd.edu on Wed, Nov 22, 2000 at 03:43:38PM -0500
Organization: GNU's Not Unix!
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Chandra Krintz (ckrintz@cs.ucsd.edu) wrote:
> Would it be possible for someone to compile
> the .java files that are included in the new release
> to .class files and make some sort of archive of
> them available?
> 
> I would like to use them directly without the
> rest of the infrastructure.  Others have shown
> interest (from the ORP mailing list) in 
> this also.  In an early release a .zip file enabled
> this type of access.

It is possible and we should do it.  However, right now Classpath is
still lacking zip support, so it can't even load classes out of its
own archive file if you try to use them zipped.

-- 
Aaron M. Renn (arenn@urbanophile.com) http://www.urbanophile.com/arenn/


From bryce@albatross.co.nz Sat Nov 25 21:00:11 2000
Received: from waitaki.otago.ac.nz ([139.80.75.140])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13zr6w-0000VJ-00
	for <classpath@gnu.org>; Sat, 25 Nov 2000 21:00:11 -0500
Received: from albatross.co.nz (IDENT:bryce@reason.otago.ac.nz [139.80.75.188])
	by waitaki.otago.ac.nz (8.9.3/8.9.3) with ESMTP id OAA01793;
	Sun, 26 Nov 2000 14:59:29 +1300
Message-ID: <3A206F5F.ECC2B1F5@albatross.co.nz>
Date: Sun, 26 Nov 2000 15:03:11 +1300
From: Bryce McKinlay <bryce@albatross.co.nz>
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.4.0-test9 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: "Kevin A. Burton" <burton@relativity.yi.org>
CC: Mark Wielaard <mark@klomp.org>, java-discuss@sources.redhat.com,
        classpath@gnu.org
Subject: Re: A java DEBUG flag
References: <3A1CA5C4.254DC238@albatross.co.nz>
		<20001123095240.B32159@klomp.org> <m3lmuaf2zj.fsf@relativity.yi.org>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

"Kevin A. Burton" wrote:

> As an asside... I have always wanted a standard place to read java properties
> files instead of having to pass -D to the JVM.

The JDK has a few runtime config files such as font.properties, so we may eventually
want a mechanism like this. Perhaps we should consider generalizing it to extend to
system properties for all applications?

> Maybe we need to have an
> /etc/java.properties file that GJC has to read.  Since -D won't work for
> statically compiled apps in GJC maybe this is the better way.

gcj can use -D. You can either compile properties into your executable on the gcj
command line, or use "gij" as a front end to your application and pass a -D argument
there.

regards

  [ bryce ]




From cbj@gnu.org Sat Nov 25 21:51:17 2000
Received: from rdu25-12-037.nc.rr.com ([24.25.12.37])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13zruP-00023y-00
	for <classpath@gnu.org>; Sat, 25 Nov 2000 21:51:17 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eAQ2mn522566;
	Sat, 25 Nov 2000 21:48:49 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: Mark Wielaard <mark@klomp.org>
Cc: Tom Tromey <tromey@cygnus.com>, Classpath Project <classpath@gnu.org>
Subject: Re: cvs2cl
References: <878zqeo5tt.fsf@creche.cygnus.com> <20001123093558.A32159@klomp.org>
From: Brian Jones <cbj@gnu.org>
Date: 25 Nov 2000 21:48:44 -0500
In-Reply-To: Mark Wielaard's message of "Thu, 23 Nov 2000 09:35:58 +0100"
Message-ID: <m3itpbqwhf.fsf@rdu25-12-037.nc.rr.com>
Lines: 31
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Mark Wielaard <mark@klomp.org> writes:

> Although I would like to have a hand maintained ChangeLog file please
> note that the cvs2cl utility has a --fsf flag that generates the somewhat
> more readable:
> 
> 2000-11-18 02:06  mark
> 
>         * ChangeLog, java/util/jar/Attributes.java,
>         java/util/jar/JarFile.java: Merge with libgcj (fixes by Anthony
>         Green and Mark Wielaard) 
>         * java/util/jar/Attributes.java: Correct comment spelling.  
>         * java/util/jar/JarFile.java: Constructor that takes mode now calls
>         super. Don't call java.util.zip.ZipFile.getEntry twice.  
> 
> Cheers,
> 
> Mark

Thanks for pointing that out Mark.  I'll gladly make that change and
regenerate the file.  I'm of the opinion that we need to generate a
ChangeLog which catches us up to the current date now matter what.
All other ChangeLogs are rather incomplete.  

I would prefer to generate the log file from cvs comments... as the
two are supposed to be the same thing... the separate maintenance only
buys hand-formatted text as opposed to machine I suppose.

Brian
-- 
Brian Jones <cbj@gnu.org>


From cbj@gnu.org Sat Nov 25 22:01:47 2000
Received: from rdu25-12-037.nc.rr.com ([24.25.12.37])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13zs4U-0002BC-00
	for <classpath@gnu.org>; Sat, 25 Nov 2000 22:01:47 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eAQ303I22580;
	Sat, 25 Nov 2000 22:00:03 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: GNU Classpath <classpath@gnu.org>
Subject: Re: JOS
References: <m38zqdsx3s.fsf@rdu25-12-037.nc.rr.com> <20001125073850.P2100@urbanophile.com>
From: Brian Jones <cbj@gnu.org>
Date: 25 Nov 2000 21:59:58 -0500
In-Reply-To: arenn@urbanophile.com's message of "Sat, 25 Nov 2000 07:38:50 -0600"
Message-ID: <m3elzzqvyp.fsf@rdu25-12-037.nc.rr.com>
Lines: 32
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

arenn@urbanophile.com writes:

> Brian Jones (cbj@gnu.org) wrote:
> > The following article was interesting to me.  I don't know what the
> > state of the JOS project is or who our liason is.  Could someone from
> > that project bring us up to date with respect to this email referenced
> > below?
> > 
> > http://www.jos.org/pipermail/arch/1999-December/000358.html
> 
> It is interesting.  If I might extract their points:
> 
> 1.  They think that replacement classes should be written in an alternate
> namespace, then ported over when completed.  This certainly solves some
> chicken and egg problems, but it creates new ones of its own.  I don't
> object to that approach, but what we currently have seems to be working
> fine.

Point (1) was interesting in that it might be easier to test code
against a more reliable VM, making it much easier to just assume the
problem is in gcp (GNU Classpath) until proven otherwise.  I'm
somewhat eager to take the VM out of the picture when trying to debug
the 98% (swag) of gcp which doesn't have a VM specific hook.

Unlike the article however, I'd prefer to keep CVS code in the current
package name space and look at using a java parser to change the
package names of user defined parts of the tree for testing.  Has
something like this already been written elsewhere?

Brian
-- 
Brian Jones <cbj@gnu.org>


From idstewart@softhome.net Sat Nov 25 22:05:43 2000
Received: from tango.softhome.net ([204.144.231.49])
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 13zs8M-0002QY-00
	for <classpath@gnu.org>; Sat, 25 Nov 2000 22:05:43 -0500
Received: (qmail 13099 invoked by uid 417); 26 Nov 2000 03:06:43 -0000
Received: from dyn1-tnt1-14.columbus.oh.ameritech.net (HELO localhost.localdomain) (209.18.16.14)
  by smtpa.softhome.net with SMTP; 26 Nov 2000 03:06:43 -0000
From: Ian D. Stewart <idstewart@softhome.net>
To: gnue-forms@lists.gnue.org
Subject: libart tarball
Date: Sat, 25 Nov 2000 22:18:49 -0500
X-Mailer: KMail [version 1.0.28]
Content-Type: text/plain
Cc: classpath@gnu.org
MIME-Version: 1.0
Message-Id: <00112522241902.07289@localhost.localdomain>
Content-Transfer-Encoding: 8bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

For those who may want to build dia from source, you will need the libart 2D
graphics library installed.  Conventional wisdom holds that the only way to get
this library is to check it out from GNOME CVS.  However, after some research,
I've tracked down a tarball at
http://www.artofcode.com/libart/libart-2.3.3.tar.gz

Brian, I'm copying this e-mail to the ClassPath list as the GTK-based AWT
implementation also requires libart.


Regards,
Ian D. Stewart

--
"Rumors of my demise have been greatly exagerated."
Samuel Clemins,
AKA Mark Twain


From cbj@gnu.org Sat Nov 25 22:23:35 2000
Received: from rdu25-12-037.nc.rr.com ([24.25.12.37])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13zsPe-0002yf-00
	for <classpath@gnu.org>; Sat, 25 Nov 2000 22:23:34 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eAQ3Kwp22662;
	Sat, 25 Nov 2000 22:20:58 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: "Nic Ferrier" <nferrier@tapsellferrier.co.uk>
Cc: classpath@gnu.org
Subject: Re: Kaffe and Classpath (again)
References: <sa1c63dc.030@tapsellferrier.co.uk>
From: Brian Jones <cbj@gnu.org>
Date: 25 Nov 2000 22:20:53 -0500
In-Reply-To: "Nic Ferrier"'s message of "Thu, 23 Nov 2000 00:16:45 +0000"
Message-ID: <m3aeanquzu.fsf@rdu25-12-037.nc.rr.com>
Lines: 59
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

"Nic Ferrier" <nferrier@tapsellferrier.co.uk> writes:

>    japhar/java.lang.Class
>    kaffe/java.lang.Class
>    eletrical-fire/java.lang.Class
>    etc...

That's right, vm/reference is essentially the same as saying
vm/japhar.  The rest of the directories are named for the vm
involved.  So you'd use vm/kaffe/java/lang/*, etc.

> If I can achieve that Kaffe *could* be altered to use adapted
> versions of those classes on some configuration request (or even
> dynamically but I'm not going to aim for that) and Classpath *could*
> include the altered Kaffe classes in the existing:
> 
>    vm/kaffe
> 
> directory.
> 

If you get Kaffe to work in some fashion I can add whatever to the
configure system to make the details easier for the user.  I think I
have it setup to simply bomb out of configure if you specify
--with-kaffe, but the majority of the work is already there
configure-wise.

> What I really need to know about is the history and workings of the
> config system.

I believe I intended to use a kaffe-config shell script generated by
autoconf when configuring kaffe to determine how it was compiled and
linked... tends to matter for Japhar for instance due to
nspr... similar cases probably exist for kaffe.  This script would
have to be accepted by the kaffe team and included in their
distribution.  Otherwise, we can to make patches for each and every new
version of Kaffe.

> I presume it's just a system for selecting which version of the
> classes under vm/ is compiled into the classes.jar.

The --with-kaffe flag selects vm/kaffe and creates the symbolic link
/vm/current to point to the right directory.  On systems without
symbolic links I think the idea is to copy the entire directory to
vm/current producing the same result with more fuss.

> The other thing I would quite like to know is more about StackFrame.
> As I understand it I have to implement gnu.vm.stack.StackFrame  but
> there is a gnu.java.StackFrame as well. 

I do not believe that the stack stuff works at all in Japhar.  Last I
looked it had problems just with a visual inspection.  That's
something I want to get into and fix so it works with Japhar.  

John Keiser did the Japhar work as well as the vm integration guide.  

Brian
-- 
Brian Jones <cbj@gnu.org>


From cbj@gnu.org Sat Nov 25 22:26:19 2000
Received: from rdu25-12-037.nc.rr.com ([24.25.12.37])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13zsSI-0003Gj-00
	for <classpath@gnu.org>; Sat, 25 Nov 2000 22:26:18 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eAQ3Nxc22671;
	Sat, 25 Nov 2000 22:23:59 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: arenn@urbanophile.com
Cc: Chandra Krintz <ckrintz@cs.ucsd.edu>, classpath@gnu.org
Subject: Re: zip file of class files
References: <m13ygjy-000UwZC@redwood.cs.utk.edu> <20001125075110.S2100@urbanophile.com>
From: Brian Jones <cbj@gnu.org>
Date: 25 Nov 2000 22:23:53 -0500
In-Reply-To: arenn@urbanophile.com's message of "Sat, 25 Nov 2000 07:51:10 -0600"
Message-ID: <m366lbquuu.fsf@rdu25-12-037.nc.rr.com>
Lines: 27
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

arenn@urbanophile.com writes:

> Chandra Krintz (ckrintz@cs.ucsd.edu) wrote:
> > Would it be possible for someone to compile
> > the .java files that are included in the new release
> > to .class files and make some sort of archive of
> > them available?
> > 
> > I would like to use them directly without the
> > rest of the infrastructure.  Others have shown
> > interest (from the ORP mailing list) in 
> > this also.  In an early release a .zip file enabled
> > this type of access.
> 
> It is possible and we should do it.  However, right now Classpath is
> still lacking zip support, so it can't even load classes out of its
> own archive file if you try to use them zipped.
> 

Even so, there is a glibj.zip included with the .tar.gz distribution.
You're not actually required to compile the java source in the
distribution unless you want to... the classes were compiled with
jikes-1.12-1 from Red Hat.

Brian
-- 
Brian Jones <cbj@gnu.org>


From cbj@gnu.org Sat Nov 25 22:39:42 2000
Received: from rdu26-46-004.nc.rr.com ([66.26.46.4] helo=rdu25-12-037.nc.rr.com)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13zsfG-0003PA-00
	for <classpath@gnu.org>; Sat, 25 Nov 2000 22:39:42 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eAQ3bRJ22968;
	Sat, 25 Nov 2000 22:37:27 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: "Ian D. Stewart" <idstewart@softhome.net>
Cc: gnue-forms@lists.gnue.org, classpath@gnu.org
Subject: Re: libart tarball
References: <00112522241902.07289@localhost.localdomain>
From: Brian Jones <cbj@gnu.org>
Date: 25 Nov 2000 22:37:22 -0500
In-Reply-To: "Ian D. Stewart"'s message of "Sat, 25 Nov 2000 22:18:49 -0500"
Message-ID: <m3itpb9zf1.fsf@rdu25-12-037.nc.rr.com>
Lines: 20
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

"Ian D. Stewart" <idstewart@softhome.net> writes:

> For those who may want to build dia from source, you will need the libart 2D
> graphics library installed.  Conventional wisdom holds that the only way to get
> this library is to check it out from GNOME CVS.  However, after some research,
> I've tracked down a tarball at
> http://www.artofcode.com/libart/libart-2.3.3.tar.gz
> 
> Brian, I'm copying this e-mail to the ClassPath list as the GTK-based AWT
> implementation also requires libart.

Sorry, there are a number of other requirements the hacking file
mentions which I probably didn't account for.  Guess I'll look at
putting this information together on a web page soon.

libart is provided in Red Hat gnome-libs-1.2.4-11 as well.

Brian
-- 
Brian Jones <cbj@gnu.org>


From cbj@gnu.org Sat Nov 25 22:44:58 2000
Received: from rdu26-46-004.nc.rr.com ([66.26.46.4] helo=rdu25-12-037.nc.rr.com)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 13zskL-0003i0-00
	for <classpath@gnu.org>; Sat, 25 Nov 2000 22:44:57 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eAQ3ghP22979;
	Sat, 25 Nov 2000 22:42:43 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: arenn@urbanophile.com
Cc: Chandra Krintz <ckrintz@cs.ucsd.edu>, classpath@gnu.org
Subject: Re: zip file of class files
References: <m13ygjy-000UwZC@redwood.cs.utk.edu> <20001125075110.S2100@urbanophile.com> <m366lbquuu.fsf@rdu25-12-037.nc.rr.com>
From: Brian Jones <cbj@gnu.org>
Date: 25 Nov 2000 22:42:38 -0500
In-Reply-To: Brian Jones's message of "25 Nov 2000 22:23:53 -0500"
Message-ID: <m3elzz9z69.fsf@rdu25-12-037.nc.rr.com>
Lines: 8
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Brian Jones <cbj@gnu.org> writes:

> jikes-1.12-1 from Red Hat.

That should have read "built by Red Hat."  Obviously Mo Dejong and the
rest of the jikes team created it.
-- 
Brian Jones <cbj@gnu.org>


From idstewart@softhome.net Sun Nov 26 10:10:45 2000
Received: from tango.softhome.net ([204.144.231.49])
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 1403S1-0002cV-00
	for <classpath@gnu.org>; Sun, 26 Nov 2000 10:10:45 -0500
Received: (qmail 9626 invoked by uid 417); 26 Nov 2000 15:11:49 -0000
Received: from dyn1-tnt1-126.columbus.oh.ameritech.net (HELO localhost.localdomain) (209.18.16.126)
  by smtpb.softhome.net with SMTP; 26 Nov 2000 15:11:49 -0000
From: Ian D. Stewart <idstewart@softhome.net>
To: Brian Jones <cbj@gnu.org>,
 "Nic Ferrier" <nferrier@tapsellferrier.co.uk>
Subject: Re: Kaffe and Classpath (again)
Date: Sun, 26 Nov 2000 10:27:18 -0500
X-Mailer: KMail [version 1.0.28]
Content-Type: text/plain
Cc: classpath@gnu.org
References: <sa1c63dc.030@tapsellferrier.co.uk> <m3aeanquzu.fsf@rdu25-12-037.nc.rr.com>
In-Reply-To: <m3aeanquzu.fsf@rdu25-12-037.nc.rr.com>
MIME-Version: 1.0
Message-Id: <00112610293500.20871@localhost.localdomain>
Content-Transfer-Encoding: 8bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

On Sat, 25 Nov 2000, Brian Jones wrote:
> 
> The --with-kaffe flag selects vm/kaffe and creates the symbolic link
> /vm/current to point to the right directory.  On systems without
> symbolic links I think the idea is to copy the entire directory to
> vm/current producing the same result with more fuss.
> 

Is it the intent of the Classpath team then to support Classpath on Wintel (and
other non-UNIX) platforms?


Regards,
Ian


From cbj@gnu.org Sun Nov 26 10:15:39 2000
Received: from rdu26-46-004.nc.rr.com ([66.26.46.4] helo=rdu25-12-037.nc.rr.com)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 1403Wk-0002hK-00
	for <classpath@gnu.org>; Sun, 26 Nov 2000 10:15:38 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eAQFD1q00785;
	Sun, 26 Nov 2000 10:13:01 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: tromey@cygnus.com
Cc: Bryce McKinlay <bryce@albatross.co.nz>, java-discuss@sources.redhat.com,
   classpath@gnu.org
Subject: Re: A java DEBUG flag
References: <3A1CA5C4.254DC238@albatross.co.nz> <874s0z6vn0.fsf@creche.cygnus.com>
From: Brian Jones <cbj@gnu.org>
Date: 26 Nov 2000 10:12:56 -0500
In-Reply-To: Tom Tromey's message of "22 Nov 2000 23:38:11 -0700"
Message-ID: <m3zoim3gxz.fsf@rdu25-12-037.nc.rr.com>
Lines: 23
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Tom Tromey <tromey@cygnus.com> writes:

> >>>>> "Bryce" == Bryce McKinlay <bryce@albatross.co.nz> writes:
> 
> Bryce> I propose an interface (this ensures that only constants can be
> Bryce> defined) called gnu.classpath.Flags (or CompileSettings?
> Bryce> DebugFlags?)
> 
> Maybe `Configuration'?
> 
>     if (Configuration.DEBUG) { ... }
> 
> Bryce> Comments?
> 
> Sounds like a good idea to me.

Seems like a good idea to me.  Will you be modifying the configure
system so that the debugging can be toggled (--enable-debugging) or
should I?  

Brian
-- 
Brian Jones <cbj@gnu.org>


From cbj@gnu.org Sun Nov 26 11:49:21 2000
Received: from rdu26-46-004.nc.rr.com ([66.26.46.4] helo=rdu25-12-037.nc.rr.com)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 1404zR-00059L-00
	for <classpath@gnu.org>; Sun, 26 Nov 2000 11:49:21 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eAQGkrI04115;
	Sun, 26 Nov 2000 11:46:53 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: "Ian D. Stewart" <idstewart@softhome.net>
Cc: "Nic Ferrier" <nferrier@tapsellferrier.co.uk>, classpath@gnu.org
Subject: Re: Kaffe and Classpath (again)
References: <sa1c63dc.030@tapsellferrier.co.uk> <m3aeanquzu.fsf@rdu25-12-037.nc.rr.com> <00112610293500.20871@localhost.localdomain>
From: Brian Jones <cbj@gnu.org>
Date: 26 Nov 2000 11:46:48 -0500
In-Reply-To: "Ian D. Stewart"'s message of "Sun, 26 Nov 2000 10:27:18 -0500"
Message-ID: <m3puji3clj.fsf@rdu25-12-037.nc.rr.com>
Lines: 21
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

"Ian D. Stewart" <idstewart@softhome.net> writes:

> On Sat, 25 Nov 2000, Brian Jones wrote:
> > 
> > The --with-kaffe flag selects vm/kaffe and creates the symbolic link
> > /vm/current to point to the right directory.  On systems without
> > symbolic links I think the idea is to copy the entire directory to
> > vm/current producing the same result with more fuss.
> > 
> 
> Is it the intent of the Classpath team then to support Classpath on
> Wintel (and other non-UNIX) platforms?
> 

Not a primary goal.  This is clearly articulated on the website I
believe.  We're more likely to support cygwin way before natively
supporting wintel... mingw may get you there though.

Brian
-- 
Brian Jones <cbj@gnu.org>


From bryce@albatross.co.nz Sun Nov 26 17:39:39 2000
Received: from waitaki.otago.ac.nz ([139.80.75.140])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 140ASQ-0006YO-00; Sun, 26 Nov 2000 17:39:38 -0500
Received: from albatross.co.nz (IDENT:bryce@reason.otago.ac.nz [139.80.75.188])
	by waitaki.otago.ac.nz (8.9.3/8.9.3) with ESMTP id LAA04370;
	Mon, 27 Nov 2000 11:39:35 +1300
Message-ID: <3A219214.3CB70382@albatross.co.nz>
Date: Mon, 27 Nov 2000 11:43:32 +1300
From: Bryce McKinlay <bryce@albatross.co.nz>
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.4.0-test9 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Brian Jones <cbj@gnu.org>
CC: tromey@cygnus.com, java-discuss@sources.redhat.com, classpath@gnu.org
Subject: Re: A java DEBUG flag
References: <3A1CA5C4.254DC238@albatross.co.nz> <874s0z6vn0.fsf@creche.cygnus.com> <m3zoim3gxz.fsf@rdu25-12-037.nc.rr.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Brian Jones wrote:

> > Bryce> I propose an interface (this ensures that only constants can be
> > Bryce> defined) called gnu.classpath.Flags (or CompileSettings?
> > Bryce> DebugFlags?)
>
> Seems like a good idea to me.  Will you be modifying the configure
> system so that the debugging can be toggled (--enable-debugging) or
> should I?

I was planning to look at classpath's configure and mirror the changes
there, but if you want to do it that would be even better ;-)

For reference, the libgcj patch is here:

http://sources.redhat.com/ml/java-patches/2000-q4/msg00128.html

regards

  [ bryce ]




From bryce@albatross.co.nz Sun Nov 26 17:48:07 2000
Received: from waitaki.otago.ac.nz ([139.80.75.140])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 140Aab-0006hm-00; Sun, 26 Nov 2000 17:48:06 -0500
Received: from albatross.co.nz (IDENT:bryce@reason.otago.ac.nz [139.80.75.188])
	by waitaki.otago.ac.nz (8.9.3/8.9.3) with ESMTP id LAA04392;
	Mon, 27 Nov 2000 11:48:02 +1300
Message-ID: <3A219410.1A9E3108@albatross.co.nz>
Date: Mon, 27 Nov 2000 11:52:00 +1300
From: Bryce McKinlay <bryce@albatross.co.nz>
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.4.0-test9 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Brian Jones <cbj@gnu.org>
CC: Mark Wielaard <mark@klomp.org>, Tom Tromey <tromey@cygnus.com>,
        Classpath Project <classpath@gnu.org>
Subject: Re: cvs2cl
References: <878zqeo5tt.fsf@creche.cygnus.com> <20001123093558.A32159@klomp.org> <m3itpbqwhf.fsf@rdu25-12-037.nc.rr.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Brian Jones wrote:

> Thanks for pointing that out Mark.  I'll gladly make that change and
> regenerate the file.  I'm of the opinion that we need to generate a
> ChangeLog which catches us up to the current date now matter what.
> All other ChangeLogs are rather incomplete.

How about we generate a cvs2cl changelog to get historical changes, but hand
maintain it from now on? Or is that what's happening already?

> I would prefer to generate the log file from cvs comments... as the
> two are supposed to be the same thing... the separate maintenance only
> buys hand-formatted text as opposed to machine I suppose.

It would be great if the cvs comments were sufficient, but in reality they don't
work as well as handcrafted entries. My main problem with them is that you can't
easily write a separate entry for each file changed without doing multiple
commits.

regards

  [ bryce ]




From jkeiser@iname.com Mon Nov 27 15:36:25 2000
Received: from [207.225.27.130] (helo=colorado.ieffects.com)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 140V0i-0007ON-00
	for <classpath@gnu.org>; Mon, 27 Nov 2000 15:36:25 -0500
Received: (qmail 19354 invoked by uid 501); 27 Nov 2000 21:37:04 -0000
Received: from delaware.ieffects.com (HELO delaware) (10.1.1.3)
  by colorado.ieffects.com with SMTP; 27 Nov 2000 21:37:04 -0000
From: "John Keiser" <jkeiser@iname.com>
To: "Nic Ferrier" <nferrier@tapsellferrier.co.uk>,
	<classpath@gnu.org>
Subject: RE: Kaffe integration
Date: Mon, 27 Nov 2000 13:35:50 -0700
Message-ID: <NEBBJENFPMAHNAJLBGMKOEAKCGAA.jkeiser@iname.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
In-Reply-To: <sa1d2c19.036@tapsellferrier.co.uk>
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400
Importance: Normal
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Nic,

Sorry it took so long to respond.  I've been having a nice, long,
non-email-reading weekend.

Thread and Class: You are correct. Thread and Class are supposed to be
implemented per VM.  I initially made VMThread and VMClass, but these
classes are almost entirely native and can usually be done more easily and
efficiently if implemented per VM.  This way you can also customize private
data members.  Code reuse and module separation sometimes goes overboard, so
I relaxed and let each VM do it.

StackTrace and StackFrame: this part of the integration is troublesome for
two reasons.  First, it has never been tested (doesn't work w/Japhar).
Second, I think in hindsight that it was a bad idea.  There are probably
lots of cool optimizations a VM could do if it owned Throwable.  Also, just
the overhead of creating a Java object for StackTrace and StackFrame is
probably pretty bad.  Really, what should happen is Throwable should become
VM-specific.  This will likely solve your problem and Brian's in the most
efficient possible manner.  I do not have the time to do this, but it would
not be hard even to switch with Japhar.

Thanks for getting this going, more VMs is always better IMO :)

Incidentally, we do support other VMs apparently, IBM is using Classpath in
theirs.

--John Keiser

> -----Original Message-----
> From: classpath-admin@gnu.org [mailto:classpath-admin@gnu.org]On Behalf
> Of Nic Ferrier
> Sent: Thursday, November 23, 2000 7:37 AM
> To: classpath@gnu.org
> Subject: Kaffe integration
>
>
> java.lang.Throwable depends on:
>
>    gnu.vm.stack.StackTrace
>    gnu.vm.stack.StackFrame
>
> so I think that one of 2 things needs to be done to allow VM
> integration:
>
> - a requirement to implement gnu.vm.stack classes
> - a requirement to implement java.lang.Throwable
>
>
> It seems to me that the former is the best plan. This just means that
> the vm-integration page on the website needs to be updated to state
> clearly that the package:
>
>    gnu.vm.stack
>
> needs to be implemented (and probably "there is an example
> implementation (for japhar) in vm/reference").
>
>
> Nic
>
> _______________________________________________
> Classpath mailing list
> Classpath@gnu.org
> http://mail.gnu.org/mailman/listinfo/classpath
>



From cbj@gnu.org Mon Nov 27 19:10:20 2000
Received: from rdu26-46-004.nc.rr.com ([66.26.46.4] helo=rdu25-12-037.nc.rr.com)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 140YLk-0005w3-00
	for <classpath@gnu.org>; Mon, 27 Nov 2000 19:10:20 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eAS08bj16482;
	Mon, 27 Nov 2000 19:08:37 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: java-discuss@sources.redhat.com, classpath@gnu.org
Subject: Re: A java DEBUG flag
References: <3A1CA5C4.254DC238@albatross.co.nz> <874s0z6vn0.fsf@creche.cygnus.com> <m3zoim3gxz.fsf@rdu25-12-037.nc.rr.com> <3A219214.3CB70382@albatross.co.nz>
From: Brian Jones <cbj@gnu.org>
Date: 27 Nov 2000 19:08:37 -0500
In-Reply-To: Bryce McKinlay's message of "Mon, 27 Nov 2000 11:43:32 +1300"
Message-ID: <m3itp92c1m.fsf@rdu25-12-037.nc.rr.com>
Lines: 31
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="=-=-="
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

--=-=-=

Bryce McKinlay <bryce@albatross.co.nz> writes:

> Brian Jones wrote:
> 
> > > Bryce> I propose an interface (this ensures that only constants can be
> > > Bryce> defined) called gnu.classpath.Flags (or CompileSettings?
> > > Bryce> DebugFlags?)
> >
> > Seems like a good idea to me.  Will you be modifying the configure
> > system so that the debugging can be toggled (--enable-debugging) or
> > should I?
> 
> I was planning to look at classpath's configure and mirror the changes
> there, but if you want to do it that would be even better ;-)
> 
> For reference, the libgcj patch is here:
> 
> http://sources.redhat.com/ml/java-patches/2000-q4/msg00128.html

Thanks, the attached is the patch to classpath for the same.  I've
left the arguments to configure the same... seems like
--enable-java-debug should add '-g' to put debug symbols in class
files... and I couldn't really change the define in the java file
anyway without breaking gcj...

Brian
-- 
Brian Jones <cbj@gnu.org>


--=-=-=
Content-Disposition: attachment; filename=debug.diff

? ChangeLog.one
? gnu/classpath/Configuration.java
Index: acconfig.h
===================================================================
RCS file: /cvs/classpath/acconfig.h,v
retrieving revision 1.1
diff -r1.1 acconfig.h
20a21,23
> /* Define as 1 if building with debugging code included, this is not -g */
> #define DEBUG 0
> 
Index: configure.in
===================================================================
RCS file: /cvs/classpath/configure.in,v
retrieving revision 1.63
diff -r1.63 configure.in
197a198,217
> 
> dnl -----------------------------------------------------------
> dnl Enable debugging statements at compile time.  By default
> dnl these statements should be optimized out of the bytecode
> dnl produced by an optimizing Java compiler and not hinder 
> dnl performance.
> dnl -----------------------------------------------------------
> LIBGCJDEBUG="false"
> AC_ARG_ENABLE(libgcj-debug,
> [  --enable-libgcj-debug   enable runtime debugging code],
>   if test "$enable_libgcj_debug" = yes; then
>     AC_DEFINE(DEBUG)
>     LIBGCJDEBUG="true"
>   fi)
> ]
> AC_SUBST(LIBGCJDEBUG)
> 
> dnl -----------------------------------------------------------
> 
> 
267a288,289
> gnu/classpath/Makefile
> gnu/classpath/Configuration.java
Index: gnu/Makefile.am
===================================================================
RCS file: /cvs/classpath/gnu/Makefile.am,v
retrieving revision 1.5
diff -r1.5 Makefile.am
4c4
< SUBDIRS = java tools 
---
> SUBDIRS = classpath java tools 
Index: gnu/classpath/.cvsignore
===================================================================
RCS file: .cvsignore
diff -N .cvsignore
0a1,2
> Makefile
> Makefile.in
Index: gnu/classpath/Configuration.java.in
===================================================================
RCS file: Configuration.java.in
diff -N Configuration.java.in
0a1,41
> /* gnu.classpath.Configuration
>    Copyright (C) 1998 Free Software Foundation, Inc.
> 
> This file is part of GNU Classpath.
> 
> GNU Classpath is free software; you can redistribute it and/or modify
> it under the terms of the GNU General Public License as published by
> the Free Software Foundation; either version 2, or (at your option)
> any later version.
>  
> GNU Classpath is distributed in the hope that it will be useful, but
> WITHOUT ANY WARRANTY; without even the implied warranty of
> MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> General Public License for more details.
> 
> You should have received a copy of the GNU General Public License
> along with GNU Classpath; see the file COPYING.  If not, write to the
> Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
> 02111-1307 USA.
> 
> As a special exception, if you link this library with other files to
> produce an executable, this library does not by itself cause the
> resulting executable to be covered by the GNU General Public License.
> This exception does not however invalidate any other reasons why the
> executable file might be covered by the GNU General Public License. */
> 
> package gnu.classpath;
> 
> /**
>  * This file defines compile-time constants that can be accessed by
>  * java code. It is pre-processed by configure.  
>  */
> public interface Configuration
> {
>   // The value of DEBUG is substituted according to whether the
>   // "--enable-libgcj-debug" argument was passed to configure. Code
>   // which is made conditional based on the value of this flag will
>   // be removed by the optimizer in a non-debug build.
>     
>   boolean DEBUG = @LIBGCJDEBUG@;
> }
Index: gnu/classpath/Makefile.am
===================================================================
RCS file: Makefile.am
diff -N Makefile.am
0a1,5
> ## Input file for automake to generate the Makefile.in used by configure
> 
> EXTRA_DIST = \
> Configuration.java.in
> 

--=-=-=--


From michal.cierniak@intel.com Mon Nov 27 19:24:25 2000
Received: from thalia.fm.intel.com ([132.233.247.11])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 140YZN-0006OQ-00
	for <classpath@gnu.org>; Mon, 27 Nov 2000 19:24:25 -0500
Received: from SMTP (fmsmsxvs02-1.fm.intel.com [132.233.42.202])
	by thalia.fm.intel.com (8.9.1a+p1/8.9.1/d: relay.m4,v 1.33 2000/11/21 19:27:27 smothers Exp $) with SMTP id AAA18938;
	Tue, 28 Nov 2000 00:25:14 GMT
Received: from fmsmsx18.intel.com ([132.233.48.18]) by 132.233.48.202
  (Norton AntiVirus for Internet Email Gateways 1.0) ;
  Tue, 28 Nov 2000 00:23:54 0000 (GMT)
Received: by fmsmsx18.fm.intel.com with Internet Mail Service (5.5.2650.21)
	id <XN84HR1M>; Mon, 27 Nov 2000 16:23:49 -0800
Message-ID: <794826DE8867D411BAB8009027AE9EB901EF6466@FMSMSX38>
From: "Cierniak, Michal" <michal.cierniak@intel.com>
To: "'Stuart Ballard'" <sballard@netreach.net>,
        GNU Classpath
	 <classpath@gnu.org>
Subject: RE: classpath-0.01 released
Date: Mon, 27 Nov 2000 16:23:40 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2650.21)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Stuart Ballard wrote:

> Brian Jones wrote:
> > 
> > I've packaged and released classpath-0.01.  This release features
> > numerous improvements and additions.  Places to download the source
> > code and precompiled binaries are listed below.
> 
> Awesome news, and great work!

Yes, great job!

I have however one request.  We have submitted a few bug fixes that did not
end up in that snapshot.  One of those bugs, in java.io.DataInputStream,
prevents a SPEC JVM98 benchmark from running.  Could whoever maintains
DataInputStream add it in?  You can find the fix in the Classpath mail
archive.  It's the #1 in
http://www.mail-archive.com/classpath%40gnu.org/msg02599.html  The problem
is that the current version of DataInputStream.readLine in Classpath does
not return null at the eof.  Thanks!


> - Are there any plans to integrate any of the third-party VMs 
> supported
> (such as IBM's OpenRuntime, 
Stuart, do you mean the Open Runtime Platform
(http://intel.com/research/mrl/orp) from Intel?  Or does IBM have something
with a similar name?

As to your question about integrating support for ORP into Classpath, all of
the native methods implemented by us for Classpath are included in the
distribution of ORP and we require only one modification to the Java sources
of Classpath (adding one extra field to java.lang.Thread).  We simply
describe that modification in our readme.txt file.  Also, one of ORP users
has made a precompiled version of Classpath for ORP available publicly.  You
can find information about that on our mailing list
(http://www.egroups.com/group/orp).  

I guess one could add vm/orp in parallel to vm/reference with this one extra
line in Thread.java being the only difference.

Michal



From bryce@albatross.co.nz Mon Nov 27 23:58:36 2000
Received: from waitaki.otago.ac.nz ([139.80.75.140])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 140cqh-0007DU-00
	for <classpath@gnu.org>; Mon, 27 Nov 2000 23:58:35 -0500
Received: from albatross.co.nz (IDENT:bryce@reason.otago.ac.nz [139.80.75.188])
	by waitaki.otago.ac.nz (8.9.3/8.9.3) with ESMTP id RAA09138;
	Tue, 28 Nov 2000 17:58:10 +1300
Message-ID: <3A233C58.F470F094@albatross.co.nz>
Date: Tue, 28 Nov 2000 18:02:16 +1300
From: Bryce McKinlay <bryce@albatross.co.nz>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.4.0-test9 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: "Cierniak, Michal" <michal.cierniak@intel.com>
CC: "'Stuart Ballard'" <sballard@netreach.net>,
        GNU Classpath <classpath@gnu.org>
Subject: Re: classpath-0.01 released
References: <794826DE8867D411BAB8009027AE9EB901EF6466@FMSMSX38>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

"Cierniak, Michal" wrote:

> I have however one request.  We have submitted a few bug fixes that did not
> end up in that snapshot.  One of those bugs, in java.io.DataInputStream,
> prevents a SPEC JVM98 benchmark from running.  Could whoever maintains
> DataInputStream add it in?  You can find the fix in the Classpath mail
> archive.  It's the #1 in
> http://www.mail-archive.com/classpath%40gnu.org/msg02599.html  The problem
> is that the current version of DataInputStream.readLine in Classpath does
> not return null at the eof.  Thanks!

I have just merged the libgcj implementations of DataInputStream and
BufferedReader into classpath. I believe that these issues should now be fixed.
If there is still a problem, please get back to me.

thanks

  [ bryce ]




From tromey@creche.cygnus.com Tue Nov 28 00:32:22 2000
Received: from gasherbrumii.peakpeak.com ([207.174.178.17] helo=gash2.peakpeak.com)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 140dNO-0000FC-00
	for <classpath@gnu.org>; Tue, 28 Nov 2000 00:32:22 -0500
Received: from creche.cygnus.com (07-121.026.popsite.net [64.24.106.121])
	by gash2.peakpeak.com (8.9.3/8.9.3) with ESMTP id WAA29308;
	Mon, 27 Nov 2000 22:31:54 -0700
Received: (from tromey@localhost)
	by creche.cygnus.com (8.9.3/8.9.3) id WAA14105;
	Mon, 27 Nov 2000 22:41:20 -0700
To: Bryce McKinlay <bryce@albatross.co.nz>
Cc: "Cierniak, Michal" <michal.cierniak@intel.com>,
        "'Stuart Ballard'" <sballard@netreach.net>,
        GNU Classpath <classpath@gnu.org>
Subject: Re: classpath-0.01 released
References: <794826DE8867D411BAB8009027AE9EB901EF6466@FMSMSX38> <3A233C58.F470F094@albatross.co.nz>
X-Zippy:  Where's my SOCIAL WORKER?
X-Attribution:  Tom
Reply-To: tromey@cygnus.com
From: Tom Tromey <tromey@cygnus.com>
Date: 27 Nov 2000 22:41:18 -0700
In-Reply-To: Bryce McKinlay's message of "Tue, 28 Nov 2000 18:02:16 +1300"
Message-ID: <87g0kcd56p.fsf@creche.cygnus.com>
Lines: 21
X-Mailer: Gnus v5.7/Emacs 20.5
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

>>>>> "Bryce" == Bryce McKinlay <bryce@albatross.co.nz> writes:

Bryce> I have just merged the libgcj implementations of
Bryce> DataInputStream and BufferedReader into classpath. I believe
Bryce> that these issues should now be fixed.  If there is still a
Bryce> problem, please get back to me.

I think your change broke Classpath builds.

Classpath's java.io.RandomAccessFile depended on the
DataInputStream.convertTo* methods, which you removed.

I looked at doing this merge a while back but stopped because of this
problem.

FWIW the Classpath approach might be better here.  libgcj assumes that
all the file-like things can be handled similarly, but this isn't true
on all platforms.  Offhand I don't know whether RandomAccessFile is
one of the places we'll need to change though.

Tom


From bryce@albatross.co.nz Tue Nov 28 01:15:26 2000
Received: from waitaki.otago.ac.nz ([139.80.75.140])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 140e33-0001b8-00
	for <classpath@gnu.org>; Tue, 28 Nov 2000 01:15:25 -0500
Received: from albatross.co.nz (IDENT:bryce@reason.otago.ac.nz [139.80.75.188])
	by waitaki.otago.ac.nz (8.9.3/8.9.3) with ESMTP id TAA09333;
	Tue, 28 Nov 2000 19:14:23 +1300
Message-ID: <3A234E35.41FEA98@albatross.co.nz>
Date: Tue, 28 Nov 2000 19:18:29 +1300
From: Bryce McKinlay <bryce@albatross.co.nz>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.4.0-test9 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: tromey@cygnus.com
CC: "Cierniak, Michal" <michal.cierniak@intel.com>,
        "'Stuart Ballard'" <sballard@netreach.net>,
        GNU Classpath <classpath@gnu.org>
Subject: Re: classpath-0.01 released
References: <794826DE8867D411BAB8009027AE9EB901EF6466@FMSMSX38> <3A233C58.F470F094@albatross.co.nz> <87g0kcd56p.fsf@creche.cygnus.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Tom Tromey wrote:

> I think your change broke Classpath builds.
>
> Classpath's java.io.RandomAccessFile depended on the
> DataInputStream.convertTo* methods, which you removed.

You're right. Oops.

> FWIW the Classpath approach might be better here.  libgcj assumes that
> all the file-like things can be handled similarly, but this isn't true
> on all platforms.  Offhand I don't know whether RandomAccessFile is
> one of the places we'll need to change though.

I'm not sure exactly what you mean here, but it sounds like a lower level
problem to me. There is a portability issue where we assume that the
native file descriptors are ints, but afaik this is confined to
FileDescriptor.java in libgcj. Classpath's RandomAccessFile does have
this problem though:

/**
  * The native file descriptor for this file
  */
private int native_fd;

It also uses a native method to implement the 1.2 method setLength().
Probibly this would be better done as a new package-private native method
in FileDescriptor.

Will investigate and fix soon.

regards

  [ bryce ]




From burton@relativity.yi.org Tue Nov 28 07:13:32 2000
Received: from merlin.prod.itd.earthlink.net ([207.217.120.156])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 140jdb-0007jA-00; Tue, 28 Nov 2000 07:13:31 -0500
Received: from relativity.yi.org (user-vcauqk9.dsl.mindspring.com [216.175.106.137])
	by merlin.prod.itd.earthlink.net (8.9.3-EL_1_3/8.9.3) with ESMTP id EAA04460;
	Tue, 28 Nov 2000 04:13:29 -0800 (PST)
Received: from localhost.localdomain (IDENT:root@user-vcaum19.dsl.mindspring.com [216.175.88.41])
	by relativity.yi.org (8.9.3/8.8.7) with ESMTP id EAA02205;
	Tue, 28 Nov 2000 04:13:27 -0800
Received: (from burton@localhost)
	by localhost.localdomain (8.11.0/8.11.0) id eASC9tY10140;
	Tue, 28 Nov 2000 04:09:55 -0800
X-Authentication-Warning: localhost.localdomain: burton set sender to burton@relativity.yi.org using -f
To: Brian Jones <cbj@gnu.org>
Cc: java-discuss@sources.redhat.com, classpath@gnu.org
Subject: Re: A java DEBUG flag
References: <3A1CA5C4.254DC238@albatross.co.nz>
	<874s0z6vn0.fsf@creche.cygnus.com> <m3zoim3gxz.fsf@rdu25-12-037.nc.rr.com>
	<3A219214.3CB70382@albatross.co.nz>
	<m3itp92c1m.fsf@rdu25-12-037.nc.rr.com>
From: burton@relativity.yi.org (Kevin A. Burton)
Date: 28 Nov 2000 03:47:31 -0800
In-Reply-To: <m3itp92c1m.fsf@rdu25-12-037.nc.rr.com>
Message-ID: <m3bsv0jp2k.fsf@relativity.yi.org>
User-Agent: Gnus/5.0808 (Gnus v5.8.8) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Lines: 27
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Brian Jones <cbj@gnu.org> writes:
<snip>

> 
> Thanks, the attached is the patch to classpath for the same.  I've
> left the arguments to configure the same... seems like
> --enable-java-debug should add '-g' to put debug symbols in class
> files... and I couldn't really change the define in the java file
> anyway without breaking gcj...

This essentially works out to being a precompiler... precompilers were
explicitly left out of Java during its early design just like multiple
inheritance and pointers... thus I don't think it is a good idea to include it
in GCJ..

Kevin

-- 
Kevin A. Burton ( burton@apache.org, burton@openprivacy.org, burtonator@acm.org )
        Cell: 408-910-6145 URL: http://relativity.yi.org ICQ: 73488596 

The more you tighten your grip, the more systems will slip through your fingers.



Legion of Doom security NSA Serbian Honduras quiche Nazi strategic Treasury
Clinton Albanian SDI supercomputer spy $400 million in gold bullion


From abb999@pchome.com.tw Tue Nov 28 07:26:40 2000
Received: from [207.219.173.8] (helo=mercury.apfnet.org)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 140jqJ-00088d-00; Tue, 28 Nov 2000 07:26:39 -0500
Received: from pchome.com.tw (210.65.221.27) by mercury.apfnet.org (NPlex 5.1.039)
        id 3A215D6B00001A94; Tue, 28 Nov 2000 01:43:50 -0800
Message-ID: <3A215D6B00001A94@mercury.apfnet.org> (added by postmaster@mercury.apfnet.org)
From: "monkey" <sa8825002@scc3.scc.edu.tw>
To: "abb999@pchome.com.tw"
Subject: =?big5?B?p0Gquqbntdu48rXbrsmlTqZixdw=?=
Date: Tue, 28 Nov 2000 15:41:10 +0800
MIME-Version: 1.0
Content-Type: multipart/related;
	boundary="----=_NextPart_000_0009_01C05951.A7F44340";
	type="multipart/alternative"
X-Priority: 3
X-MSMail-Priority: Normal
X-Unsent: 1
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6600
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

This is a multi-part message in MIME format.

------=_NextPart_000_0009_01C05951.A7F44340
Content-Type: multipart/alternative;
	boundary="----=_NextPart_001_000A_01C05951.A7F44340"


------=_NextPart_001_000A_01C05951.A7F44340
Content-Type: text/plain;
	charset="big5"
Content-Transfer-Encoding: quoted-printable

=B7s=BA=F4=AD=B61
----- Original Message -----=20
Sent: Wednesday, October 04, 2000 8:28 PM
Subject: =A7A=AA=BA=A6=E7=B5=DB=B8=F2=B5=DB=AE=C9=A5N=A6b=C5=DC



=A1@=20
=A1@=20



=20










------=_NextPart_001_000A_01C05951.A7F44340
Content-Type: text/html;
	charset="big5"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD><TITLE>=B7s=BA=F4=AD=B61</TITLE><!DOCTYPE HTML PUBLIC =
"-//W3C//DTD HTML 4.0 Transitional//EN">
<META content=3D"text/html; charset=3Dbig5" http-equiv=3DContent-Type>
<META content=3D"MSHTML 5.00.2919.6307" name=3DGENERATOR>
<META content=3DFrontPage.Editor.Document name=3DProgId>
<STYLE type=3Dtext/css>.textanimlink {
	TEXT-DECORATION: none
}
A {
	TEXT-DECORATION: none
}
P.main {
	FONT-FAMILY: =BC=D0=B7=A2=C5=E9; FONT-SIZE: 28pt
}
</STYLE>
</HEAD>
<BODY bgColor=3D#000000 onload=3DScroll()>
<DIV>&nbsp;</DIV>
<DIV style=3D"FONT: 10pt =B7s=B2=D3=A9=FA=C5=E9">----- Original Message =
-----=20
<DIV><B>Sent:</B> Wednesday, October 04, 2000 8:28 PM</DIV>
<DIV><B>Subject:</B> =
=A7A=AA=BA=A6=E7=B5=DB=B8=F2=B5=DB=AE=C9=A5N=A6b=C5=DC</DIV></DIV>
<DIV><BR></DIV>
<DIV style=3D"FONT: 10pt =B7s=B2=D3=A9=FA=C5=E9">
<DIV>&nbsp;</DIV></DIV>
<DIV>
<DIV id=3Dtextanim onclick=3Dbreakf()=20
style=3D"HEIGHT: 1px; LEFT: 0px; POSITION: absolute; TOP: 100px; WIDTH: =
80px">=A1@=20
<P>=A1@ </P></DIV><BR></DIV>
<DIV style=3D"FONT: 10pt =B7s=B2=D3=A9=FA=C5=E9"></DIV>
<DIV><BR>&nbsp;</DIV>
<DIV><BR></DIV>
<SCRIPT language=3DJavascript>=20
<!-- Hiding=20
bname=3Dnavigator.appName;=20
bversion=3DparseInt(navigator.appVersion)=20
if ((bname=3D=3D"Netscape" && bversion>=3D4) || (bname=3D=3D"Microsoft =
Internet Explorer" && bversion>=3D4))=20
window.onload=3Dstart=20
else=20
stop();=20
window.onunload=3Dstop=20
if (bname=3D=3D"Netscape"){=20
brows=3Dtrue=20
dt=3D2=20
}=20
else{=20
brows=3Dfalse=20
dt=3D5=20
}=20
var z=3D0;=20
var msg=3D0;=20
var rgb=3D0;=20
var link=3Dtrue;=20
var status=3Dtrue;=20
var updwn=3Dfalse;=20
var message=3D new Array();=20
var value=3D0;=20
var h=3Dwindow.innerHeight;=20
var w=3Dwindow.innerWidth;=20
var timer1;=20
var timer2;=20
var timer3;=20
var convert =3D new Array()=20
var hexbase=3D new Array("0", "1", "2", "3", "4", "5", "6", "7", "8", =
"9", "A", "B", "C", "D", "E", "F");=20
var bgcolor=3D"#ac2d17"; =20
var color=3D"#00F000";  =20
=20
message[0]=3D'=A7A=AA=BA=A6=E7=B5=DB=B8=F2=B5=DB=AE=C9=A5N=A6b=C5=DC...'
message[1]=3D'=AC=B0=A6=F3=A7A=AA=BA=C6[=A9=C0=A4=A3=AF=E0=A7=EF=C5=DC???=
'
message[2]=3D'=A4=B0=BB=F2=A4~=ACO=ADP=B4I=AA=BA=C6[=A9=C0=A1=E3'
message[3]=3D'=ACO=A1u=A7A=A6s=A6h=A4=D6=BF=FA=A1H=A1v'
message[4]=3D'=A1u=BF=EF=A4=B0=BB=F2=BC=CB=AA=BA=A7=EB=B8=EA...=A1v'
message[5]=3D'=C1=D9=ACO=A1u=A5[=AA=F8=A7A=AA=BA=A4u=A7@=AE=C9=B6=A1?=A1v=
'
message[6]=3D'___________________________'
message[7]=3D'=A4=BD =A4=B8 2000 =A6~ =AD=B7 =BC=C9'
message[8]=3D'=A5u =ADn =A4@ =A6~'
message[9]=3D'-=C5=FD =A7A =C5=DC =A4=D1-'
message[10]=3D'=A5b=A6~=C5=FD=A7A=C1`=A6=AC=A4J=A2=B5=A2=B3=B8U'
message[11]=3D'=A4@=AD=D3=A4=EB=C5=FD=A7A=C1=C8=A2=B1=A2=B6=B8U'
message[12]=3D'=A1y=A7=EB=B8=EA=B3=F8=B9S=B2v=A1z=B6W=B9L 600%'
message[13]=3D'=B3o=B4N=ACO'
message[14]=3D'|'
message[15]=3D'=A1i =A4@ =A6~ =ADP =B4I =A4=E8 =AE=D7 =A1j'
message[16]=3D'.....'
message[17]=3D'=A7A=ACO=A7=DA=AD=CC=ADn=A7=E4=AA=BA=A4H=B6=DC???'
message[18]=3D'=A7=DA=AD=CC=B1N=ACD=BF=EF=A4=D6=BC=C6=AA=BA=B8g=C0=E7=AA=CC=
'
message[19]=3D'=A7A=B7=C7=B3=C6=A6n=A4F=B6=DC?'


for (x=3D0; x<16; x++){=20
for (y=3D0; y<16; y++){=20
convert[value]=3D hexbase[x] + hexbase[y];=20
value++;=20
}=20
}=20
=20
redx=3Dcolor.substring(1,3);=20
greenx=3Dcolor.substring(3,5);=20
bluex=3Dcolor.substring(5,7);=20
hred=3Deval(parseInt(redx,16));=20
hgreen=3Deval(parseInt(greenx,16));=20
hblue=3Deval(parseInt(bluex,16));=20
eredx=3Dbgcolor.substring(1,3);=20
egreenx=3Dbgcolor.substring(3,5);=20
ebluex=3Dbgcolor.substring(5,7);=20
ered=3Deval(parseInt(eredx,16));=20
egreen=3Deval(parseInt(egreenx,16));=20
eblue=3Deval(parseInt(ebluex,16));=20
red=3Dered;=20
green=3Degreen;=20
blue=3Deblue;=20
=20
function start(){=20
if ((bname=3D=3D"Netscape" && bversion>=3D4) || (bname=3D=3D"Microsoft =
Internet Explorer" && bversion>=3D4)){=20
link=3Dfalse;=20
updwn=3Dtrue;=20
if (brows)=20
res=3Ddocument.layers['textanim'].top=20
else{=20
textanim.style.width=3Ddocument.body.offsetWidth-20;=20
textanim.innerHTML=3D'<Pre><P Class=3D"main" =
Align=3D"Center">'+message[msg]+'</P></Pre>'=20
res=3Dtextanim.style.top=20
for (x=3D0; x<document.all.length; x++)=20
if(document.all[x].id=3D=3D"textanimlink")=20
link=3Dtrue;=20
}=20
up()=20
}=20
}=20
=20
function stop(){=20
clearTimeout(timer1);=20
clearTimeout(timer2);=20
clearTimeout(timer3);=20
}=20
=20
function resz(){=20
h=3Dwindow.innerHeight;=20
w=3Dwindow.innerWidth;=20
if (updwn)=20
timer1=3DsetTimeout('up()',1000)=20
else=20
timer2=3DsetTimeout('down()',1000)=20
}=20
=20
function breakf(){=20
if (status){=20
clearTimeout(timer1);=20
clearTimeout(timer2);=20
status=3Dfalse=20
return;=20
}=20
else{=20
status=3Dtrue;=20
if (updwn)=20
timer1=3DsetTimeout('up()',dt)=20
else=20
timer2=3DsetTimeout('down()',dt)=20
}=20
}=20
=20
=20
function up(){=20
if (red<hred){=20
if ((red+5)<hred)=20
red+=3D5;=20
else=20
red=3Dhred=20
redx =3D convert[red]=20
}=20
else{=20
if ((red-5)>hred)=20
red-=3D5;=20
else=20
red=3Dhred=20
redx =3D convert[red]=20
}=20
=20
if (green<hgreen){=20
if ((green+5)<hgreen)=20
green+=3D5;=20
else=20
green=3Dhgreen=20
greenx =3D convert[green]=20
}=20
else{=20
if ((green-5)>hgreen)=20
green-=3D5;=20
else=20
green=3Dhgreen=20
greenx =3D convert[green]=20
}=20
=20
if (blue<hblue){=20
if ((blue+5)<hblue)=20
blue+=3D5;=20
else=20
blue=3Dhblue=20
bluex =3D convert[blue]=20
}=20
else{=20
if ((blue-5)>hblue)=20
blue-=3D5;=20
else=20
blue=3Dhblue=20
bluex =3D convert[blue]=20
}=20
=20
rgb =3D "#"+redx+greenx+bluex;=20
if (brows){=20
document.layers['textanim'].document.linkColor=3Drgb;=20
document.layers['textanim'].document.vlinkColor=3Drgb;=20
if (window.innerHeight!=3Dh || window.innerWidth!=3Dw){=20
clearTimeout(timer1);=20
resz()=20
return;=20
}=20
else{=20
document.layers['textanim'].document.write('<Pre><P Class=3D"main" =
Align=3D"Center"><font =
color=3D"'+rgb+'">'+message[msg]+'</font></P></Pre>')=20
document.layers['textanim'].document.close();=20
}=20
}=20
else{=20
textanim.style.color=3Drgb;=20
if(link)=20
textanimlink.style.color=3Drgb;=20
}=20
if (z<38){=20
if (brows)=20
document.layers['textanim'].top--=20
else=20
textanim.style.posTop--=20
z++=20
timer1=3DsetTimeout('up()',dt)=20
}=20
else=20
{=20
updwn=3Dfalse;=20
down()=20
}=20
}=20
=20
=20
function down(){=20
if (red<ered){=20
if ((red+5)<ered)=20
red+=3D5;=20
else=20
red=3Dered=20
redx =3D convert[red]=20
}=20
else{=20
if ((red-5)>ered)=20
red-=3D5;=20
else=20
red=3Dered=20
redx =3D convert[red]=20
}=20
=20
if (green<egreen){=20
if ((green+5)<egreen)=20
green+=3D5;=20
else=20
green=3Degreen=20
greenx =3D convert[green]=20
}=20
else{=20
if ((green-5)>egreen)=20
green-=3D5;=20
else=20
green=3Degreen=20
greenx =3D convert[green]=20
}=20
=20
if (blue<eblue){=20
if ((blue+5)<eblue)=20
blue+=3D5;=20
else=20
blue=3Deblue=20
bluex =3D convert[blue]=20
}=20
else{=20
if ((blue-5)>eblue)=20
blue-=3D5;=20
else=20
blue=3Deblue=20
bluex =3D convert[blue]=20
}=20
=20
rgb =3D "#"+redx+greenx+bluex;=20
if (brows){=20
document.layers['textanim'].document.linkColor=3Drgb;=20
document.layers['textanim'].document.vlinkColor=3Drgb;=20
if (window.innerHeight!=3Dh || window.innerWidth!=3Dw){=20
clearTimeout(timer2);=20
resz()=20
return;=20
}=20
else{=20
document.layers['textanim'].document.write('<Pre><P Class=3D"main" =
Align=3D"Center"><font =
color=3D"'+rgb+'">'+message[msg]+'</font></P></Pre>')=20
document.layers['textanim'].document.close();=20
}=20
}=20
else{=20
textanim.style.color=3Drgb;=20
if(link)=20
textanimlink.style.color=3Drgb;=20
}=20
=20
if (z<76){=20
if (brows)=20
document.layers['textanim'].top--=20
else=20
textanim.style.posTop--=20
z++=20
timer2=3DsetTimeout('down()',dt)=20
}=20
else=20
{=20
if (brows){=20
document.layers['textanim'].document.write('')=20
document.layers['textanim'].document.close();=20
}=20
else=20
textanim.innerHTML=3D'';=20
window.clearInterval(timer2);=20
if(msg<message.length-1){=20
msg++;=20
z=3D0;=20
if (brows){=20
document.layers['textanim'].top=3Dres;=20
}=20
else=20
textanim.style.top=3Dres;=20
timer3=3DsetTimeout('start()',100);=20
}=20
else=20
{=20
msg=3D0;=20
z=3D0;=20
if (brows)=20
document.layers['textanim'].top=3Dres;=20
else=20
textanim.style.top=3Dres;=20
timer3=3DsetTimeout('start()',1000);=20
}=20
}=20
}=20
=20
=20
// done hiding -->=20
</SCRIPT>

<DIV>&nbsp;</DIV>
<DIV>&nbsp;</DIV>
<DIV>&nbsp;</DIV>
<DIV align=3Dcenter>&nbsp;</DIV>
<DIV align=3Dcenter>&nbsp;</DIV>
<DIV align=3Dcenter>&nbsp;</DIV>
<DIV align=3Dcenter><A =
href=3D"http://www.hot-life.com.tw/a105289/st.htm"><IMG=20
align=3Dbaseline alt=3D"" border=3D0 hspace=3D0=20
src=3D"cid:000401c0590e$934d3a20$1bdd41d2@INNET"></A></DIV></BODY></HTML>=


------=_NextPart_001_000A_01C05951.A7F44340--

------=_NextPart_000_0009_01C05951.A7F44340
Content-Type: image/gif;
	name="wellcome.gif"
Content-Transfer-Encoding: base64
Content-ID: <000401c0590e$934d3a20$1bdd41d2@INNET>
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=

------=_NextPart_000_0009_01C05951.A7F44340--


From bryce@albatross.co.nz Tue Nov 28 07:30:19 2000
Received: from 210-55-119-205.adsl.xtra.co.nz ([210.55.119.205] helo=localhost.localdomain)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 140jtq-0008E1-00
	for <classpath@gnu.org>; Tue, 28 Nov 2000 07:30:18 -0500
Received: from albatross.co.nz (IDENT:bryce@hokkaido [127.0.0.1])
	by localhost.localdomain (8.11.0/8.11.0) with ESMTP id eASCV3903632;
	Wed, 29 Nov 2000 01:31:03 +1300
Message-ID: <3A23A587.EAC25311@albatross.co.nz>
Date: Wed, 29 Nov 2000 01:31:03 +1300
From: Bryce McKinlay <bryce@albatross.co.nz>
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.4.0-test11-ac4 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: "Kevin A. Burton" <burton@relativity.yi.org>
CC: java-discuss@sources.redhat.com, classpath@gnu.org
Subject: Re: A java DEBUG flag
References: <3A1CA5C4.254DC238@albatross.co.nz>
		<874s0z6vn0.fsf@creche.cygnus.com> <m3zoim3gxz.fsf@rdu25-12-037.nc.rr.com>
		<3A219214.3CB70382@albatross.co.nz>
		<m3itp92c1m.fsf@rdu25-12-037.nc.rr.com> <m3bsv0jp2k.fsf@relativity.yi.org>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

"Kevin A. Burton" wrote:

> > Thanks, the attached is the patch to classpath for the same.  I've
> > left the arguments to configure the same... seems like
> > --enable-java-debug should add '-g' to put debug symbols in class
> > files... and I couldn't really change the define in the java file
> > anyway without breaking gcj...
>
> This essentially works out to being a precompiler... precompilers were
> explicitly left out of Java during its early design just like multiple
> inheritance and pointers... thus I don't think it is a good idea to include it
> in GCJ..

Can you suggest a better solution?

regards

  [ bryce ]




From cbj@gnu.org Tue Nov 28 07:45:27 2000
Received: from rdu26-46-178.nc.rr.com ([66.26.46.178] helo=rdu25-12-037.nc.rr.com)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 140k8V-0000Vy-00
	for <classpath@gnu.org>; Tue, 28 Nov 2000 07:45:27 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eASChcb18730;
	Tue, 28 Nov 2000 07:43:38 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: burton@relativity.yi.org (Kevin A. Burton)
Cc: java-discuss@sources.redhat.com, classpath@gnu.org
Subject: Re: A java DEBUG flag
References: <3A1CA5C4.254DC238@albatross.co.nz> <874s0z6vn0.fsf@creche.cygnus.com> <m3zoim3gxz.fsf@rdu25-12-037.nc.rr.com> <3A219214.3CB70382@albatross.co.nz> <m3itp92c1m.fsf@rdu25-12-037.nc.rr.com> <m3bsv0jp2k.fsf@relativity.yi.org>
From: Brian Jones <cbj@gnu.org>
Date: 28 Nov 2000 07:43:38 -0500
In-Reply-To: burton@relativity.yi.org's message of "28 Nov 2000 03:47:31 -0800"
Message-ID: <m3itp8z2px.fsf@rdu25-12-037.nc.rr.com>
Lines: 37
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

burton@relativity.yi.org (Kevin A. Burton) writes:

> Brian Jones <cbj@gnu.org> writes:
> <snip>
> 
> > 
> > Thanks, the attached is the patch to classpath for the same.  I've
> > left the arguments to configure the same... seems like
> > --enable-java-debug should add '-g' to put debug symbols in class
> > files... and I couldn't really change the define in the java file
> > anyway without breaking gcj...
> 
> This essentially works out to being a precompiler... precompilers
> were explicitly left out of Java during its early design just like
> multiple inheritance and pointers... thus I don't think it is a good
> idea to include it in GCJ..

What do you mean by not a good idea?  The gcj guys did this to get
more information when debugging classes in a simple way which can be
compiled out easily when you don't want to debug.  So my post was just
to say I've added the same configure stuff to support this to
classpath.  We share java source code between projects so it makes
sense for us to do this.

Relating to your objection, we're not adding a precompiler capability
to classpath or gcj with this which you can use generically wherever
you want.  This is specifically for us, the people who develop
classpath, to use when we want to debug.  

I have thought of one problem however.  It might be more useful to
break down debugging into at least package specific flags as opposed
to turning on debugging for the entire world with this one flag... and
I'd like a better name than GCJDEBUG... ideas?

Brian
-- 
Brian Jones <cbj@gnu.org>


From sballard@netreach.net Tue Nov 28 10:17:38 2000
Received: from fatuka.netreach.net ([207.29.195.121] helo=smtp.netreach.net)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 140mVm-0005EJ-00
	for <classpath@gnu.org>; Tue, 28 Nov 2000 10:17:38 -0500
Received: (qmail 22595 invoked by uid 102); 28 Nov 2000 15:18:14 -0000
Received: from rainbow.netreach.net (HELO netreach.net) (@209.152.221.64)
  by smtp.netreach.net with SMTP; 28 Nov 2000 15:18:14 -0000
Message-ID: <3A23CC91.FA6F25A0@netreach.net>
Date: Tue, 28 Nov 2000 10:17:37 -0500
From: Stuart Ballard <sballard@netreach.net>
X-Mailer: Mozilla 4.72 [en] (X11; I; Linux 2.2.17 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: "Cierniak, Michal" <michal.cierniak@intel.com>
CC: GNU Classpath <classpath@gnu.org>
Subject: Re: classpath-0.01 released
References: <794826DE8867D411BAB8009027AE9EB901EF6466@FMSMSX38>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

"Cierniak, Michal" wrote:
> 
> Stuart Ballard wrote:
> 
> > - Are there any plans to integrate any of the third-party VMs
> > supported
> > (such as IBM's OpenRuntime,
> Stuart, do you mean the Open Runtime Platform
> (http://intel.com/research/mrl/orp) from Intel?  Or does IBM have something
> with a similar name?

D'oh! Yes, that was what I meant. Too many big companies starting with I
that do a lot of java stuff...

(open mouth, insert foot)

> As to your question about integrating support for ORP into Classpath, all of
> the native methods implemented by us for Classpath are included in the
> distribution of ORP and we require only one modification to the Java sources
> of Classpath (adding one extra field to java.lang.Thread).
[snip]
> I guess one could add vm/orp in parallel to vm/reference with this one extra
> line in Thread.java being the only difference.

Hmm. That does seem a little wasteful... I wonder if there's any build
system magic that we could do?

Is the "one line" something that wouldn't hurt other VMs? If so, we
could add it to vm/reference and make ORP work the same way Japhar does
(--with-orp could just be a synonym for --with-japhar which would both
really just mean "with vm/reference")

Thoughts, anyone?

Stuart.


From sballard@netreach.net Tue Nov 28 10:27:51 2000
Received: from fatuka.netreach.net ([207.29.195.121] helo=smtp.netreach.net)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 140mff-0005OL-00
	for <classpath@gnu.org>; Tue, 28 Nov 2000 10:27:51 -0500
Received: (qmail 23586 invoked by uid 102); 28 Nov 2000 15:28:28 -0000
Received: from rainbow.netreach.net (HELO netreach.net) (@209.152.221.64)
  by smtp.netreach.net with SMTP; 28 Nov 2000 15:28:28 -0000
Message-ID: <3A23CEF6.2FC70376@netreach.net>
Date: Tue, 28 Nov 2000 10:27:50 -0500
From: Stuart Ballard <sballard@netreach.net>
X-Mailer: Mozilla 4.72 [en] (X11; I; Linux 2.2.17 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: classpath@gnu.org
Subject: Re: A java DEBUG flag
References: <3A1CA5C4.254DC238@albatross.co.nz> <874s0z6vn0.fsf@creche.cygnus.com> <m3zoim3gxz.fsf@rdu25-12-037.nc.rr.com> <3A219214.3CB70382@albatross.co.nz> <m3itp92c1m.fsf@rdu25-12-037.nc.rr.com> <m3bsv0jp2k.fsf@relativity.yi.org> <m3itp8z2px.fsf@rdu25-12-037.nc.rr.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Brian Jones wrote:
> 
> I have thought of one problem however.  It might be more useful to
> break down debugging into at least package specific flags as opposed
> to turning on debugging for the entire world with this one flag... and
> I'd like a better name than GCJDEBUG... ideas?

Suggestion:

Create a bunch of classes in the gnu.java.debug package called DebugFoo
(where Foo is the particular thing to be debugged). Each of these
classes has a single public static final boolean ON.

In the configure process, provide a way to specify which values of "Foo"
should be debugged (eg --with-debug=Foo,Bar) and set ON to true or false
based on whether Foo is in the list passed to with-debug.

Each file can then import gnu.java.debug.* and then use "if
(DebugFoo.ON)" to perform debugging.

Adding a new Foo is as easy as adding a new DebugFoo.java.in file in
gnu/java/debug (because the configure system would be designed to work
with all files in that directory, rather than having a hardcoded list of
available debug flags).

This could be made as granular or non-granular as we want - for example,
we might have DebugSerialization separate from other java.io debugging,
but then have all java.awt debugging done by the same flag (despite
there being lots of packages). Thus, the debug flag is "per thing that
has useful debugging messages" rather than per-package or
per-anything-else.

Thoughts?

Stuart.


From michal.cierniak@intel.com Tue Nov 28 12:36:58 2000
Received: from thalia.fm.intel.com ([132.233.247.11])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 140ogZ-0001We-00
	for <classpath@gnu.org>; Tue, 28 Nov 2000 12:36:58 -0500
Received: from SMTP (fmsmsxvs04-1.fm.intel.com [132.233.42.204])
	by thalia.fm.intel.com (8.9.1a+p1/8.9.1/d: relay.m4,v 1.33 2000/11/21 19:27:27 smothers Exp $) with SMTP id RAA03214;
	Tue, 28 Nov 2000 17:37:56 GMT
Received: from fmsmsx26.fm.intel.com ([132.233.48.26]) by 132.233.48.204
  (Norton AntiVirus for Internet Email Gateways 1.0) ;
  Tue, 28 Nov 2000 17:36:37 0000 (GMT)
Received: by fmsmsx26.fm.intel.com with Internet Mail Service (5.5.2650.21)
	id <XN88Y4RW>; Tue, 28 Nov 2000 09:36:36 -0800
Message-ID: <794826DE8867D411BAB8009027AE9EB901EF646A@FMSMSX38>
From: "Cierniak, Michal" <michal.cierniak@intel.com>
To: "'Stuart Ballard'" <sballard@netreach.net>
Cc: GNU Classpath <classpath@gnu.org>
Subject: RE: classpath-0.01 released
Date: Tue, 28 Nov 2000 09:36:23 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2650.21)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

> > I guess one could add vm/orp in parallel to vm/reference 
> > with this one extra
> > line in Thread.java being the only difference.
> 
> Hmm. That does seem a little wasteful...
[...]
> Is the "one line" something that wouldn't hurt other VMs?
It's an extra field (of type java.lang.Object) we use to quickly map a
java.lang.Thread instance to our internal data structures.

> If so, we
> could add it to vm/reference and make ORP work the same way 
> Japhar does
> (--with-orp could just be a synonym for --with-japhar which would both
> really just mean "with vm/reference")

That would work for us and would make Classpath work straight out of the box
with one extra VM.  It would be great for ORP users who could download a new
version of Classpath without any extra work.  I would also imagine that it
would be good for Classpath users because they would have one other VM to
work with.

I guess the question is if it's acceptable for the Japhar folks to have one
extra field for every java.lang.Thread instance.  I would think that it
would be acceptable, but then I don't use Japhar.

Comments?

Michal



From tromey@creche.cygnus.com Tue Nov 28 14:02:44 2000
Received: from gasherbrumii.peakpeak.com ([207.174.178.17] helo=gash2.peakpeak.com)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 140q1b-0004Uz-00
	for <classpath@gnu.org>; Tue, 28 Nov 2000 14:02:43 -0500
Received: from creche.cygnus.com (07-121.026.popsite.net [64.24.106.121])
	by gash2.peakpeak.com (8.9.3/8.9.3) with ESMTP id MAA20883;
	Tue, 28 Nov 2000 12:02:32 -0700
Received: (from tromey@localhost)
	by creche.cygnus.com (8.9.3/8.9.3) id MAA11458;
	Tue, 28 Nov 2000 12:12:07 -0700
To: Bryce McKinlay <bryce@albatross.co.nz>
Cc: "Cierniak, Michal" <michal.cierniak@intel.com>,
        "'Stuart Ballard'" <sballard@netreach.net>,
        GNU Classpath <classpath@gnu.org>
Subject: Re: classpath-0.01 released
References: <794826DE8867D411BAB8009027AE9EB901EF6466@FMSMSX38> <3A233C58.F470F094@albatross.co.nz> <87g0kcd56p.fsf@creche.cygnus.com> <3A234E35.41FEA98@albatross.co.nz>
X-Zippy:  I had a lease on an OEDIPUS COMPLEX back in '81...
X-Attribution:  Tom
Reply-To: tromey@cygnus.com
From: Tom Tromey <tromey@cygnus.com>
Date: 28 Nov 2000 12:12:07 -0700
In-Reply-To: Bryce McKinlay's message of "Tue, 28 Nov 2000 19:18:29 +1300"
Message-ID: <87k89nap2w.fsf@creche.cygnus.com>
Lines: 13
X-Mailer: Gnus v5.7/Emacs 20.5
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

>>>>> "Bryce" == Bryce McKinlay <bryce@albatross.co.nz> writes:

Bryce> I'm not sure exactly what you mean here, but it sounds like a
Bryce> lower level problem to me. There is a portability issue where
Bryce> we assume that the native file descriptors are ints, but afaik
Bryce> this is confined to FileDescriptor.java in libgcj.

My recollection (based on a conversation we had a while back on
java-discuss) was that the `int' problem was part of it, but also we
needed to handle file descriptors for sockets differently from file
descriptors for files.

Tom


From olk@olk.co.kr Tue Nov 28 14:26:13 2000
Received: from [211.49.129.168] (helo=smtp7.thrunet.com)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 140qOL-0005QI-00
	for <classpath@gnu.org>; Tue, 28 Nov 2000 14:26:13 -0500
Received: from thrunet ([210.205.197.216]) by smtp7.thrunet.com  with Microsoft SMTPSVC(5.5.1877.197.19);
	 Wed, 29 Nov 2000 03:38:01 +0900
Date: Wed, 29 Nov 2000 03:40:03 +0900
From: "Online Korea"<olk@olk.co.kr>
To: classpath@gnu.org
Subject: classpath님 안녕하세요? 
X-AD2000-Serial: 1033
X-AD2000-Register: "박재현" <olk@olk.co.kr>
MIME-Version: 1.0
Content-Type: text/html
Content-Transfer-Encoding: base64
Message-ID: <01a420138181cb0SMTP7@smtp7.thrunet.com>
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

PEhUTUw+DQo8SEVBRD4NCjxNRVRBIE5BTUU9IkdFTkVSQVRPUiIgQ29udGVudD0iTWljcm9z
b2Z0IERIVE1MIEVkaXRpbmcgQ29udHJvbCI+DQo8VElUTEU+PC9USVRMRT4NCjwvSEVBRD4N
CjxCT0RZPg0KPFA+PEZPTlQgZmFjZT2xvLiyw7wgc2l6ZT0yPqLEIL/AtMPAxyDAr7jTIMfR
uLa18CCixTwvRk9OVD48L1A+DQo8UD48Rk9OVCBmYWNlPbG8uLLDvCBzaXplPTI+Jm5ic3A7
ICJEYWQsIEkgZG9uJ3Qgd2FudCB0byBnbyB0byBzY2hvb2wgdG9kYXkuLCIgc2FpZCANCnRo
ZSBib3kuPEJSPiZuYnNwOyAiV2h5IG5vdCwgc29uPyI8QlI+Jm5ic3A7ICJXZWxsLCBvbmUg
b2YgdGhlIGNoaWNrZW5zIG9uIHRoZSANCnNjaG9vbCBmYXJtIGRpZWQgbGFzdCB3ZWVrIGFu
ZCA8QlI+Jm5ic3A7Jm5ic3A7IHdlIGhhZCBjaGlja2VuIHNvdXAgZm9yIGx1bmNoIA0KdGhl
IG5leHQgZGF5LiBBbmQgdGhyZWUgZGF5cyBhZ28gPEJSPiZuYnNwOyZuYnNwOyBvbmUgb2Yg
dGhlIHBpZ3MgZGllZCBhbmQgd2UgDQpoYWQgcm9hc3QgcG9yayB0aGUgbmV4dCBkYXksLCwu
IjxCUj4mbmJzcDsgIkJ1dCB3aHkgZG9uJ3QgeW91IHdhbnQgdG8gZ28gDQp0b2RheT8iPEJS
PiZuYnNwOyAiQmVjYXVzZSB0aGUgRW5nbGlzaCB0ZWFjaGVyIGRpZWQgeWVzdGVyZGF5ISI8
QlI+Jm5ic3A7IA0KIi4uLi4uIjwvRk9OVD48L1A+DQo8UD48Rk9OVCBmYWNlPbG8uLLDvCBz
aXplPTI+Jm5ic3A7ICK+xrr8LCCzqiC/wLTDIMfQsbMgsKGx4iC9yL7uLiIgx8+w7SC80rPi
wMwguLvH37TZLjxCUj4mbmJzcDsgIr/WILChseIgDQq9yL7uPyI8QlI+Jm5ic3A7ICLB9rOt
IMHWv6Egx9CxsyCz88Dlv6G8rSC03yDH0SC4triusKEgwde++rTCtaUgtNnAvbOvIMGhvcnA
uLfOIDxCUj4mbmJzcDsmbmJzcDsgtN8gvbrHwbimILjUvvq+7i4gDQqx17iusO0gM8DPIMD8
v6G0wiC1xcH2IMfRILi2uK6woSDB1776tMK1pSA8QlI+Jm5ic3A7Jm5ic3A7ILHXILTZwL0g
s6+/obTCILXFwfYgutKw7bHiuKYguNS++rDtLi4uLiI8QlI+Jm5ic3A7IA0KIrHXt7G1pSC/
wLTDwLogv9Ygx9Cxs7+hILChseIgvci+7j8iPEJSPiZuYnNwOyAivu7BpiC/tb7uILyxu/20
1MDMILW5vsawobzMtNwguLu+3yEiPEJSPiZuYnNwOyANCiIuLi4uLiI8L0ZPTlQ+PC9QPg0K
PFA+PEZPTlQgZmFjZT2xvLiyw7wgc2l6ZT0yPqLEIL/AtMPAxyC/tb7uIMfRuLa18CCixSZu
YnNwOyZuYnNwOyZuYnNwOyA8L0ZPTlQ+PC9QPg0KPFA+PEZPTlQgZmFjZT2xvLiyw7wgc2l6
ZT0yPqLCIEkgZmVsdCBvbiB0b3Agb2YgdGhlIA0Kd29ybGQmbmJzcDsmbmJzcDsmbmJzcDsm
bmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsm
bmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsgDQo8L0ZPTlQ+PC9QPg0KPFA+PEZPTlQgZmFjZT2x
vLiyw7wgc2l6ZT0yPrvntvfAuiC0qbG4s6ogvLqw+MfPuOkgseK60MDMIMHBwLogsM3AuiC0
57+sx8/B9r/kLiA8QlI+ubC30Cwgvu62siC757b3tenAuiC8urD4x9/AuyC2p7+hILT1IA0K
sOK81cfYwfbB9ri4ILTrus660MDHILvntve16cC6IDxCUj5bs6q4piC6wSEgs7uwoSDD1rDt
vt9dtvOw7SC7/bCix9W0z7TZLiA8QlI+wMy/zSCwsMC6ILHiutDAuyCzqsW4s7u0wiDHpcf2
wMwgJ29uIHRvcCBvZiANCnRoZSB3b3JsZCe2+LTPtNkuPC9GT05UPjwvUD4NCjxQPjxGT05U
IGZhY2U9sby4ssO8IHNpemU9Mj5BOiBJIGhlYXJkIHlvdSB3ZXJlIGVuZ2FnZWQgdG8gTWFy
aWFuLjwvRk9OVD48L1A+DQo8UD48Rk9OVCBmYWNlPbG8uLLDvCBzaXplPTI+QjogQWN0dWFs
bHksIEkgd2FzLjwvRk9OVD48L1A+DQo8UD48Rk9OVCBmYWNlPbG8uLLDvCBzaXplPTI+QTog
T2gsIHJlYWxseT8gU28sIGhvdyBkaWQgeW91IGZlZWwgb24gdGhlIG5pZ2h0IG9mIA0KeW91
ciBlbmdhZ2VtZW50PzwvRk9OVD48L1A+DQo8UD48Rk9OVCBmYWNlPbG8uLLDvCBzaXplPTI+
QjogSSBmZWx0IGxpa2UgSSB3YXMgb24gdG9wIG9mIHRoZSB3b3JsZC48L0ZPTlQ+PC9QPg0K
PFA+PEZPTlQgZmFjZT2xvLiyw7wgc2l6ZT0yPqHhJ29uIHRvcCBvZiB0aGUgd29ybGQnuKYg
wfe/qsfPuOkgW7y8u/MgssC067Hiv6FdtvO0wiDAx7nMwMy45yANCjxCUj4mbmJzcDsgJ0kg
ZmVsdCBvbiB0b3Agb2YgdGhlIHdvcmxkLie0wiAivLy78yCywLTrseK/oSDA1rTCILTAs6bA
zL76tNkuIjxCUj4mbmJzcDsgwe8sICixssDlyPcgDQqx4rrQwMwgwcG+0rTZKbimILbmx9W0
z7TZLjxCUj4mbmJzcDsgyOfI9yBbseK60MDMIMHBvtK+7l2287TCIMDHucy3ziAnSSBmZWx0
IGdvb2QuJ7i4wLsgu/2wosfPseIgvay/7LOqIA0KPEJSPiZuYnNwOyDAzCDHpcf2tbUgvsuw
7SDA1sC4uOkgsKjBpMDHIMelx/bAuyC09SDA3yDH0iC89iDA1sC7ILDFv6G/5C48L0ZPTlQ+
PC9QPg0KPFA+PEZPTlQgZmFjZT2xvLiyw7wgc2l6ZT0yPqLBICdvbiB0b3Agb2YgdGhlIHdv
cmxkJ7TCILy8u/MgssC067Hiv6EgwNa0wiC0wLOmIS4gsde3uLTZuOkgDQo8QlI+Jm5ic3A7
Jm5ic3A7ICdvdXQgb2YgdGhpcyB3b3JsZCe0wj8gwMwgvLy788DHIMDPILCwwfYgvsqw1CCx
2bvnx9Fbd29uZGVyZnVsXSwgDQo8QlI+Jm5ic3A7Jm5ic3A7IMikwLogyK+788D7wM5bZmFu
dGFzdGljXcDMtvUgtubAzL+hv+QuPC9GT05UPjwvUD4NCjxQPjxGT05UIGZhY2U9sby4ssO8
IHNpemU9Mj5BOiC43riuvsiw+iC+4Milx9G02bDtILXpvvq0wrWlLjxCUj5COiC7573HIMDM
ucwgx9++7i48QlI+QTogvu4sIA0KwaS4u8DMvt8/Jm5ic3A7ILHXt6EsvuDIpb3Ex9Egsdez
ryC547+hILHiutDAzCC+7rauvu4/PEJSPkI6ILy8u/PAuyC+8sC6ILHiutDAzL76wfYuPC9G
T05UPjwvUD4NCjxQPjxGT05UIGZhY2U9sby4ssO8IHNpemU9Mj6ixCC/wLTDwMcgwM++7iZu
YnNwOyDH0bi2tfAgosU8L0ZPTlQ+PC9QPg0KPFA+PEZPTlQgZmFjZT2xvLiyw7wgc2l6ZT0y
PiZuYnNwOyZuYnNwOyCh2MDPvu61tSC/tb7uv80gsLDAuiDH/MXCt84gsbi8uiC1x77uIA0K
wNa9wLTPtNkuPEJSPi0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t
LS0tLS0tLS0tLS0tLS0tLS0tLTxCUj5jbGFzc3BhdGi01CANCr7Is+fHz7y8v+Q/IDwvRk9O
VD48L1A+DQo8UD48Rk9OVCBmYWNlPbG8uLLDvCBzaXplPTI+wPrI8bTCIMD8yK24piDAzL/r
x9i8rSCwrbvnv80gMToxt84gv9yxub7uIMfQvcDAuyDH0iC89iDA1rTCIDxCUj6ixCBPbmxp
bmUgS29yZWEgDQpbJm5ic3A7IDxBIGhyZWY9Imh0dHA6Ly93d3cub2xrLmNvLmtyIj5odHRw
Oi8vd3d3Lm9say5jby5rcjwvQT4mbmJzcDsgXSCixSC287DtIA0Kx9W0z7TZLjwvRk9OVD48
L1A+DQo8UD48Rk9OVCBmYWNlPbG8uLLDvCBzaXplPTI+uNXA+iwgx+O29L74wMwgxu3B9iC1
5bfBIMHLvNvH1bTPtNkuILrOtfAgs8qx17evv+4gv+u8rbimLi4uLi4uPC9GT05UPjwvUD4N
CjxQPjxGT05UIGZhY2U9sby4ssO8IHNpemU9Mj7A+sjxIMi4u+e/obyttMIgv7W+7iDAr7jT
v80gu/3IsCDIuMitv6EgsPy9ySDA1rTCILrQtemysiC4xcDPKL/5LbHdKSwgPEJSPr+1vu4g
wK+407/NIL3HIA0Ku/3IsL+hIMfKv+TH0SC7/ciwyLjIrSDH0SC5rsDlvr/AuyC5q7fht84g
urizuyC15biusO0gwNa+7r/kLjwvRk9OVD48L1A+DQo8UD48Rk9OVCBmYWNlPbG8uLLDvCBz
aXplPTI+wKe/zbCwwLogs7u/68DHILytuvG9urimILnevsa6uLHiIL/4x8+9w7jpIKHsIDxB
IA0KaHJlZj0ibWFpbHRvOm9sa0BvbGsuY28ua3IiPm9sa0BvbGsuY28ua3I8L0E+IKHtt84g
PEJSPiJ5ZXMitvO0wiCzu7/rwMcgtOTA5cC7IMHWvcOx4iANCrnZtvi0z7TZLjwvRk9OVD48
L1A+DQo8UD48Rk9OVCBmYWNlPbG8uLLDvCBzaXplPTI+sde4rrDtLCDA+sjxIMD8yK0gv9yx
ub7uILCtwMewoSCxw7Hdx8+9xSC60LXpwLsgwKfH2CZuYnNwOyA8QSANCmhyZWY9Imh0dHA6
Ly93d3cub2xrLmNvLmtyL2xlY3R1cmVfZnJhbWUuaHRtIj6iwiC9w7n8sK3AxyCiwiA8L0E+
tbUgPEJSPjHIuL+hIMfRx9i8rSC5q7fht84gDQq9x73Dx8+w7SDA1sC4tM8gx9G5+CC16b7u
uriw7SC9zcC4vcO46SDB9rHdIL3Fw7vH2CDB1ry8v+QuPC9GT05UPjwvUD4NCjxQPjxGT05U
IGZhY2U9sby4ssO8IHNpemU9Mj65q7fhIL3DufywrcDHIL3Fw7sguea5/cC6IDxBIA0KaHJl
Zj0iaHR0cDovL3d3dy5vbGsuY28ua3IvbGVjdHVyZV9mcmFtZS5odG0iPqK6IL+pseIgorgg
PC9BPrimIMWsuK/Hz73FIMjEILCtwMcgvcXDu7b1v6EgDQrA1rTCPEJSPr3DufywrcDHuKYg
vcXDu8fPvcO46SDA+sjxsKEgwPzIrbimILXlt8EgvcO5/LCtwMe4piDH2CC15biwtOS0z7TZ
LjwvRk9OVD48L1A+DQo8UD48Rk9OVCBmYWNlPbG8uLLDvCBzaXplPTI+ubC30CwgwPzIrSAw
Mi01ODgtMDUxMCDAuLfOtbUgvcXDu8fPvccgvPYgwNaxuL/kLjwvRk9OVD48L1A+DQo8UD48
Rk9OVCBmYWNlPbG8uLLDvCBzaXplPTI+vsa5q8LJt88gwMwgx9C9wLn9wMwgY2xhc3NwYXRo
tNTAxyDIuMitIL3Ht8Igx+K787+hIMDbwLogurjFxsDMIA0Ktce++sC4uOk8QlI+x9W0z7TZ
LjwvRk9OVD48L1A+DQo8UD48Rk9OVCBmYWNlPbG8uLLDvCBzaXplPTI+sKi758fVtM+02S48
L0ZPTlQ+PC9QPg0KPFA+PEZPTlQgZmFjZT2xvLiyw7wgc2l6ZT0yPmNsYXNzcGF0aLTUsrK0
wiBodHRwOi8vZnJlZWRvbS5zYXJhbmcubmV0L3NvZnR3YXJlL2NsYXNzcGF0aC9jbGFzc3Bh
dGguaHRtbL+hIMDWtMIgwda80rimILq4sO0guN7AzyC15biusNQgDQq1x776vcC0z7TZLjwv
Rk9OVD48L1A+DQo8UD48Rk9OVCBmYWNlPbG8uLLDvCBzaXplPTI+utLHyr/kx9EgwaS6uL+0
tNm46SDBpLi7IMHLvNvH1bTPtNkuPEJSPjwvRk9OVD48L1A+DQo8L0JPRFk+DQo8L0hUTUw+
DQo=




From jeff.sturm@appnet.com Tue Nov 28 14:39:37 2000
Received: from toronto.sigma6.com ([204.176.48.147])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 140qbI-0005sk-00
	for <classpath@gnu.org>; Tue, 28 Nov 2000 14:39:37 -0500
Received: from appnet.com (IDENT:jsturm@localhost.localdomain [127.0.0.1])
	by toronto.sigma6.com (8.9.3/8.9.3) with ESMTP id PAA00938;
	Tue, 28 Nov 2000 15:50:54 -0500
Message-ID: <3A241AAE.F810AC8F@appnet.com>
Date: Tue, 28 Nov 2000 15:50:54 -0500
From: Jeff Sturm <jeff.sturm@appnet.com>
Reply-To: jeff.sturm@commerceone.com
Organization: Commerce One
X-Mailer: Mozilla 4.72 [en] (X11; U; Linux 2.2.16 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: tromey@cygnus.com
CC: Bryce McKinlay <bryce@albatross.co.nz>, GNU Classpath <classpath@gnu.org>
Subject: Re: classpath-0.01 released
References: <794826DE8867D411BAB8009027AE9EB901EF6466@FMSMSX38> <3A233C58.F470F094@albatross.co.nz> <87g0kcd56p.fsf@creche.cygnus.com> <3A234E35.41FEA98@albatross.co.nz> <87k89nap2w.fsf@creche.cygnus.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Tom Tromey wrote:
> Bryce> I'm not sure exactly what you mean here, but it sounds like a
> Bryce> lower level problem to me. There is a portability issue where
> Bryce> we assume that the native file descriptors are ints, but afaik
> Bryce> this is confined to FileDescriptor.java in libgcj.
> 
> My recollection (based on a conversation we had a while back on
> java-discuss) was that the `int' problem was part of it, but also we
> needed to handle file descriptors for sockets differently from file
> descriptors for files.

Right.  Using a socket as a native file descriptor (i.e. expecting
read/write/close system calls to work) turns out to be non-portable.  Libgcj
relies on FileDescriptor in java.net for convenience, not because it is required
by the API.  I don't know what Classpath does.

--
Jeff Sturm
jeff.sturm@commerceone.com


From burton@relativity.yi.org Tue Nov 28 23:38:48 2000
Received: from eagle.prod.itd.earthlink.net ([207.217.120.24])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 140z16-00066h-00; Tue, 28 Nov 2000 23:38:48 -0500
Received: from relativity.yi.org (user-vcauqk9.dsl.mindspring.com [216.175.106.137])
	by eagle.prod.itd.earthlink.net (8.9.3-EL_1_3/8.9.3) with ESMTP id UAA26012;
	Tue, 28 Nov 2000 20:38:49 -0800 (PST)
Received: from localhost.localdomain (IDENT:root@user-vcaum6q.dsl.mindspring.com [216.175.88.218])
	by relativity.yi.org (8.9.3/8.8.7) with ESMTP id UAA11100;
	Tue, 28 Nov 2000 20:38:44 -0800
Received: (from burton@localhost)
	by localhost.localdomain (8.11.0/8.11.0) id eAT4ZCr17729;
	Tue, 28 Nov 2000 20:35:12 -0800
X-Authentication-Warning: localhost.localdomain: burton set sender to burton@relativity.yi.org using -f
To: Brian Jones <cbj@gnu.org>
Cc: java-discuss@sources.redhat.com, classpath@gnu.org
Subject: Re: A java DEBUG flag
References: <3A1CA5C4.254DC238@albatross.co.nz>
	<874s0z6vn0.fsf@creche.cygnus.com> <m3zoim3gxz.fsf@rdu25-12-037.nc.rr.com>
	<3A219214.3CB70382@albatross.co.nz>
	<m3itp92c1m.fsf@rdu25-12-037.nc.rr.com> <m3bsv0jp2k.fsf@relativity.yi.org>
	<m3itp8z2px.fsf@rdu25-12-037.nc.rr.com>
From: burton@relativity.yi.org (Kevin A. Burton)
Date: 28 Nov 2000 20:31:39 -0800
In-Reply-To: <m3itp8z2px.fsf@rdu25-12-037.nc.rr.com>
Message-ID: <m3elzviel0.fsf@relativity.yi.org>
User-Agent: Gnus/5.0808 (Gnus v5.8.8) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Lines: 61
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Brian Jones <cbj@gnu.org> writes:
<snip>
> What do you mean by not a good idea?  The gcj guys did this to get
> more information when debugging classes in a simple way which can be
> compiled out easily when you don't want to debug.  So my post was just
> to say I've added the same configure stuff to support this to
> classpath.  We share java source code between projects so it makes
> sense for us to do this.

Well... I am kind of at a disadvantate here because the decision has already
been made.  It would be hard to migrate away from this even if I was right.

The theory at least is that you never use a precompiler.  Just just if () {} out
your code because it REALLY pays off in code maintenance in the long run.  The
IBM Log4J stuff is a good example.  You leave the code the same but when you
deploy you just turn off the debug and only get about a 0.5% lag.  Worth it if
you consider the time it will take maintaining this code in the long run.

.... there is a small section about this in the original JLS I believe....

There is also a SUN JSR for this stuff about assertions.  The assertion stuff is
a language extension so that you can exclude it on compiles.  

<snip>

> I have thought of one problem however.  It might be more useful to
> break down debugging into at least package specific flags as opposed
> to turning on debugging for the entire world with this one flag... and
> I'd like a better name than GCJDEBUG... ideas?

This is in the SUN Assertion JSR.
http://java.sun.com/aboutJava/communityprocess/jsr/asrt_prop.html

... The documentation is public but we certainly can not participate in its
development because of SUNs *stupid* IP restrictions in the JCP.  There are good
ideas here though and since this stuff will eventually make it into JDK 1.4 I
would say it should eventually belong here.

Kevin
- -- 
Kevin A. Burton ( burton@apache.org, burton@openprivacy.org, burtonator@acm.org )
        Cell: 408-910-6145 URL: http://relativity.yi.org ICQ: 73488596 

Intellectual property does not exist!  Get over it!
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.4 (GNU/Linux)
Comment: Get my public key at: http://relativity.yi.org/pgpkey.txt

iD8DBQE6JIarAwM6xb2dfE0RArIWAJ9zs3r5mRIZpQcdG+NH9DOMRAB//gCg08J7
cD6r7/rMZyHX+HWJRUnnAVA=
=FvuO
-----END PGP SIGNATURE-----



Ortega Kennedy DES explosion assassination Serbian [Hello to all my fans in
domestic surveillance] Rule Psix Marxist supercomputer Khaddafi quiche World
Trade Center $400 million in gold bullion North Korea


From cbj@gnu.org Tue Nov 28 23:47:49 2000
Received: from rdu26-46-178.nc.rr.com ([66.26.46.178] helo=rdu25-12-037.nc.rr.com)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 140z9o-0006Ey-00
	for <classpath@gnu.org>; Tue, 28 Nov 2000 23:47:49 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eAT4jlZ20480;
	Tue, 28 Nov 2000 23:45:47 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: "Online Korea"<olk@olk.co.kr>
Cc: classpath@gnu.org
Subject: Re: classpath
References: <01a420138181cb0SMTP7@smtp7.thrunet.com>
From: Brian Jones <cbj@gnu.org>
Date: 28 Nov 2000 23:45:47 -0500
In-Reply-To: "Online Korea"'s message of "Wed, 29 Nov 2000 03:40:03 +0900"
Message-ID: <m3pujfwflw.fsf@rdu25-12-037.nc.rr.com>
Lines: 12
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

"Online Korea"<olk@olk.co.kr> writes:

> <mime encoded web page with jokes and classpath link deleted>

So does anyone know why we're getting sent a web page from Online
Korea which contains a linke to
http://freedom.sarang.net/software/classpath/classpath.html which
appears to just be a FSF mirror?  

Brian
-- 
Brian Jones <cbj@gnu.org>


From cbj@gnu.org Wed Nov 29 00:23:17 2000
Received: from rdu26-46-178.nc.rr.com ([66.26.46.178] helo=rdu25-12-037.nc.rr.com)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 140zi8-0007fU-00
	for <classpath@gnu.org>; Wed, 29 Nov 2000 00:23:17 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eAT5LJh20577;
	Wed, 29 Nov 2000 00:21:19 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: Stuart Ballard <sballard@netreach.net>
Cc: classpath@gnu.org
Subject: Re: A java DEBUG flag
References: <3A1CA5C4.254DC238@albatross.co.nz> <874s0z6vn0.fsf@creche.cygnus.com> <m3zoim3gxz.fsf@rdu25-12-037.nc.rr.com> <3A219214.3CB70382@albatross.co.nz> <m3itp92c1m.fsf@rdu25-12-037.nc.rr.com> <m3bsv0jp2k.fsf@relativity.yi.org> <m3itp8z2px.fsf@rdu25-12-037.nc.rr.com> <3A23CEF6.2FC70376@netreach.net>
From: Brian Jones <cbj@gnu.org>
Date: 29 Nov 2000 00:21:19 -0500
In-Reply-To: Stuart Ballard's message of "Tue, 28 Nov 2000 10:27:50 -0500"
Message-ID: <m3lmu3wdyo.fsf@rdu25-12-037.nc.rr.com>
Lines: 48
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Stuart Ballard <sballard@netreach.net> writes:

> Brian Jones wrote:
> > 
> > I have thought of one problem however.  It might be more useful to
> > break down debugging into at least package specific flags as opposed
> > to turning on debugging for the entire world with this one flag... and
> > I'd like a better name than GCJDEBUG... ideas?
> 
> Suggestion:
> 
> Create a bunch of classes in the gnu.java.debug package called DebugFoo
> (where Foo is the particular thing to be debugged). Each of these
> classes has a single public static final boolean ON.
> 
> In the configure process, provide a way to specify which values of "Foo"
> should be debugged (eg --with-debug=Foo,Bar) and set ON to true or false
> based on whether Foo is in the list passed to with-debug.
> 
> Each file can then import gnu.java.debug.* and then use "if
> (DebugFoo.ON)" to perform debugging.

I think I'd rather keep it to one java class and just use multiple
public final static variables.  

Configure works like a shell script.  We have to define variables by
name and set them equal to true or false.  Then we have to use
AC_SUBST(VARIABLE_NAME).  I can't think of a way to do this
generically at the moment.  I don't think I've ever seen an example of
such either.

So can you do the following in configure?

VAR="true"
AC_SUBST(VAR)

where VAR is defined as a variable... ie $foo.

$foo="true"
AC_SUBST($foo)

Unlike Perl... I don't think shells support this.  I could be wrong
though... it's happened before.  Anyway if this doesn't work then
again I'd just like to isolate debugging either on a package by
package basis and/or concept basis (like serialization).

-- 
Brian Jones <cbj@gnu.org>


From hans_boehm@hp.com Wed Nov 29 13:20:15 2000
Received: from deimos.hpl.hp.com ([192.6.19.190])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 141Bq2-000199-00; Wed, 29 Nov 2000 13:20:14 -0500
Received: from hplms2.hpl.hp.com (hplms2.hpl.hp.com [15.0.152.33])
	by deimos.hpl.hp.com (8.9.3 (PHNE_18546)/HPL-PA Relay) with ESMTP id KAA13262;
	Wed, 29 Nov 2000 10:20:06 -0800 (PST)
Received: from hplex1.hpl.hp.com (hplex1.hpl.hp.com [15.0.152.182])
	by hplms2.hpl.hp.com (8.10.2/8.10.2 HPL-PA Hub) with SMTP id eATIK6g07669;
	Wed, 29 Nov 2000 10:20:07 -0800 (PST)
Received: from 15.0.152.182 by hplex1.hpl.hp.com (InterScan E-Mail VirusWall NT); Wed, 29 Nov 2000 10:20:06 -0800 (Pacific Standard Time)
Received: by hplex1.hpl.hp.com with Internet Mail Service (5.5.2650.21)
	id <XT8HNDQ7>; Wed, 29 Nov 2000 10:20:05 -0800
Message-ID: <140D21516EC2D3119EE700902787664401E3A753@hplex1.hpl.hp.com>
From: "Boehm, Hans" <hans_boehm@hp.com>
To: "'burton@relativity.yi.org'" <burton@relativity.yi.org>,
        Brian Jones
	 <cbj@gnu.org>
Cc: java-discuss@sources.redhat.com, classpath@gnu.org
Subject: RE: A java DEBUG flag
Date: Wed, 29 Nov 2000 10:19:57 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2650.21)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

> This is in the SUN Assertion JSR.
> http://java.sun.com/aboutJava/communityprocess/jsr/asrt_prop.html
> 
> ... The documentation is public but we certainly can not 
> participate in its
> development because of SUNs *stupid* IP restrictions in the 
> JCP.  There are good
> ideas here though and since this stuff will eventually make 
> it into JDK 1.4 I
> would say it should eventually belong here.
> 
True.  But isn't this an awfully weak proposal?  It doesn't seem to allow me
to cleanly compute additional state that's used only in assertion checking.
I find that I need to do that all the time in order to be able to write
reasonable assertions.  For example, I might explicitly keep track of who
holds a particular lock, so that I can assert that I hold it at some point.
Or I might need to count opened files so that I can assert that I closed
everything I opened in a certain section of code.  Or ...

Hans


From cedric@wireless-networks.com Wed Nov 29 13:30:36 2000
Received: from ns.aenet.net ([157.22.214.1] helo=crewsoft.com)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 141C03-0001om-00; Wed, 29 Nov 2000 13:30:36 -0500
Received: from [63.197.8.222] (HELO wireless-networks.com)
  by crewsoft.com (CommuniGate Pro SMTP 3.3b5)
  with ESMTP id 363250; Wed, 29 Nov 2000 10:32:48 -0800
Message-ID: <3A4CD86A.DC603BD2@wireless-networks.com>
Date: Fri, 29 Dec 2000 10:31:06 -0800
From: Cedric Berger <cedric@wireless-networks.com>
X-Mailer: Mozilla 4.73 [en] (Windows NT 5.0; U)
X-Accept-Language: en
MIME-Version: 1.0
To: "Boehm, Hans" <hans_boehm@hp.com>
CC: "'burton@relativity.yi.org'" <burton@relativity.yi.org>,
 	Brian Jones <cbj@gnu.org>, java-discuss@sources.redhat.com,
 	classpath@gnu.org
Subject: Re: A java DEBUG flag
References: <140D21516EC2D3119EE700902787664401E3A753@hplex1.hpl.hp.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

"Boehm, Hans" wrote:
> 
> > This is in the SUN Assertion JSR.
> > http://java.sun.com/aboutJava/communityprocess/jsr/asrt_prop.html
> >
> > ... The documentation is public but we certainly can not
> > participate in its
> > development because of SUNs *stupid* IP restrictions in the
> > JCP.  There are good
> > ideas here though and since this stuff will eventually make
> > it into JDK 1.4 I
> > would say it should eventually belong here.
> >
> True.  But isn't this an awfully weak proposal?  It doesn't seem to allow me
> to cleanly compute additional state that's used only in assertion checking.
> I find that I need to do that all the time in order to be able to write
> reasonable assertions.  For example, I might explicitly keep track of who
> holds a particular lock, so that I can assert that I hold it at some point.
> Or I might need to count opened files so that I can assert that I closed
> everything I opened in a certain section of code.  Or ...
> 
> Hans

Can't you put all this stuff a special debug class to keep things clean?
This class will only gets loaded if assertions are enabled!

class Debugging {
    static int openFilesCount;
    static boolean openFile() {
        return openFilesCount < LIMIT;
    }
    static boolean closeFile() {
        return openFilesCount > 0;
    }
}
class Main {
    void doSomething() {
        assert(Debugging.openFile());       
        // do something
        assert(Debugging.closeFile());
    }
}

Cedric


From hans_boehm@hp.com Wed Nov 29 14:05:36 2000
Received: from deimos.hpl.hp.com ([192.6.19.190])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 141CXw-0002Yu-00; Wed, 29 Nov 2000 14:05:36 -0500
Received: from hplms2.hpl.hp.com (hplms2.hpl.hp.com [15.0.152.33])
	by deimos.hpl.hp.com (8.9.3 (PHNE_18546)/HPL-PA Relay) with ESMTP id LAA16491;
	Wed, 29 Nov 2000 11:05:23 -0800 (PST)
Received: from hplex1.hpl.hp.com (hplex1.hpl.hp.com [15.0.152.182])
	by hplms2.hpl.hp.com (8.10.2/8.10.2 HPL-PA Hub) with SMTP id eATJ5Ng09603;
	Wed, 29 Nov 2000 11:05:24 -0800 (PST)
Received: from 15.0.152.182 by hplex1.hpl.hp.com (InterScan E-Mail VirusWall NT); Wed, 29 Nov 2000 11:05:23 -0800 (Pacific Standard Time)
Received: by hplex1.hpl.hp.com with Internet Mail Service (5.5.2650.21)
	id <XT8HN1CS>; Wed, 29 Nov 2000 11:05:22 -0800
Message-ID: <140D21516EC2D3119EE700902787664401E3A756@hplex1.hpl.hp.com>
From: "Boehm, Hans" <hans_boehm@hp.com>
To: "'Cedric Berger'" <cedric@wireless-networks.com>,
        "Boehm, Hans"
	 <hans_boehm@hp.com>
Cc: "'burton@relativity.yi.org'" <burton@relativity.yi.org>,
        Brian Jones
	 <cbj@gnu.org>, java-discuss@sources.redhat.com,
        classpath@gnu.org
Subject: RE: A java DEBUG flag
Date: Wed, 29 Nov 2000 11:05:15 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2650.21)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

I'm still not sure how to do this cleanly.  The problem is not factoring out
the assertion checking code.  It's that you want to maintain e.g. a counter
only if debugging is enabled.  Thus you want to conditionally execute
something with a side-effect, where the condition is evaluated at compile
time.  Expressions in assertions aren't supposed to have side-effects
(though I guess that isn't enforced).  There doesn't seem to be another way
to condition something on whether assertions are enabled.

You could maintain the added state unconditionally, and hope that the
compiler recognizes it as dead code if debugging is disabled.  But that's
often not very realistic.  For example, if I look at some of the code in the
GC that's covered by #ifdef GC_ASSERTIONS, much of it is already hard to
prove dead with assertions disabled, and it would be still harder if you had
to preserve null pointer exceptions. 

Hans

> -----Original Message-----
> From: Cedric Berger [mailto:cedric@wireless-networks.com]
> Sent: Friday, December 29, 2000 10:31 AM
> To: Boehm, Hans
> Cc: 'burton@relativity.yi.org'; Brian Jones;
> java-discuss@sources.redhat.com; classpath@gnu.org
> Subject: Re: A java DEBUG flag
> 
> 
> 
> "Boehm, Hans" wrote:
> > 
> > > This is in the SUN Assertion JSR.
> > > http://java.sun.com/aboutJava/communityprocess/jsr/asrt_prop.html
> > >
> > > ... The documentation is public but we certainly can not
> > > participate in its
> > > development because of SUNs *stupid* IP restrictions in the
> > > JCP.  There are good
> > > ideas here though and since this stuff will eventually make
> > > it into JDK 1.4 I
> > > would say it should eventually belong here.
> > >
> > True.  But isn't this an awfully weak proposal?  It doesn't 
> seem to allow me
> > to cleanly compute additional state that's used only in 
> assertion checking.
> > I find that I need to do that all the time in order to be 
> able to write
> > reasonable assertions.  For example, I might explicitly 
> keep track of who
> > holds a particular lock, so that I can assert that I hold 
> it at some point.
> > Or I might need to count opened files so that I can assert 
> that I closed
> > everything I opened in a certain section of code.  Or ...
> > 
> > Hans
> 
> Can't you put all this stuff a special debug class to keep 
> things clean?
> This class will only gets loaded if assertions are enabled!
> 
> class Debugging {
>     static int openFilesCount;
>     static boolean openFile() {
>         return openFilesCount < LIMIT;
>     }
>     static boolean closeFile() {
>         return openFilesCount > 0;
>     }
> }
> class Main {
>     void doSomething() {
>         assert(Debugging.openFile());       
>         // do something
>         assert(Debugging.closeFile());
>     }
> }
> 
> Cedric
> 


From cbj@gnu.org Wed Nov 29 18:53:08 2000
Received: from rdu26-46-178.nc.rr.com ([66.26.46.178] helo=rdu25-12-037.nc.rr.com)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 141H2B-0001HZ-00
	for <classpath@gnu.org>; Wed, 29 Nov 2000 18:53:08 -0500
Received: (from cbj@localhost)
	by rdu25-12-037.nc.rr.com (8.11.0/8.11.0) id eATNp2e26266;
	Wed, 29 Nov 2000 18:51:02 -0500
X-Authentication-Warning: rdu25-12-037.nc.rr.com: cbj set sender to cbj@gnu.org using -f
To: GNU Classpath <classpath@gnu.org>
Cc: tullmann@cs.utah.edu
Subject: Re: A java DEBUG flag
From: Brian Jones <cbj@gnu.org>
Date: 29 Nov 2000 18:51:02 -0500
Message-ID: <m366l6wd5l.fsf@rdu25-12-037.nc.rr.com>
Lines: 37
User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Pat Tullmann wrote:

> I do a very similar thing (preprocesssing Java code to help
> debugging).  I've got a framework that allows me to turn debugging
> on/off with the preprocessor, and then I can select subsets of the
> printfs() at run-time via a system property.  Its easier to look at
> than explain, so if you're interested fetch:
> 	http://www.cs.utah.edu/flux/janos/downloads/janos-20000905.tar.gz
>
> Untar it an look at:
> 	janos-20000905/java/Defines.cpp (these are the CPP macros I use)
> 	janos-20000905/java/nodeos/Trace.Java (this is what the macros 
> 			call)
> Then grep around janos-20000905/java/nodeos/ for calls to the macros
> (DBG*) to see how its used.
>
> I find it exceedingly handy to be able to turn debugging printfs (and
> other code) on/off without recompiling.

Pat sent me this and after taking a look I think he's possibly
provided an easy solution to two problems.

a) You can define package names and have the preprocessor name the
classes.  I'd vote to not try to rename some things in java.lang as
they are just plainly VM specific and you'd never want to test them
against anything other than that VM anyway.

b) You can #ifdef debugging code and this doesn't rely on a java
compiler to optimize the debugging code out of the class.

Making these sorts of changes would be time consuming and indeed the
gcj team might have problems with it as it would affect their build as
well as ours.  Can I get buy in for either (a) or (b)?

Brian
-- 
Brian Jones <cbj@gnu.org>


From nferrier@tapsellferrier.co.uk Wed Nov 29 19:47:41 2000
Received: from tf2.tapsellferrier.co.uk ([195.7.242.227] helo=tapsellferrier.co.uk)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 141Hsy-0003xm-00; Wed, 29 Nov 2000 19:47:41 -0500
Received: from tf-Message_Server by tapsellferrier.co.uk
	with Novell_GroupWise; Thu, 30 Nov 2000 00:56:34 +0000
Message-Id: <sa25a5c2.001@tapsellferrier.co.uk>
X-Mailer: Novell GroupWise 5.2
Date: Thu, 30 Nov 2000 00:56:00 +0000
From: "Nic Ferrier" <nferrier@tapsellferrier.co.uk>
To: cbj@gnu.org, classpath@gnu.org
Cc: tullmann@cs.utah.edu
Subject: Re: A java DEBUG flag
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Disposition: inline
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

>>> Brian Jones <cbj@gnu.org> 29-Nov-00 11:51:02 PM >>>

>Pat sent me this and after taking a look I think he's 
>possibly provided an easy solution to two problems.

Pat's solution is quite good but why don't you just use gcc to
pre-process the .java files?

I'm sure I got that to work for a project I worked on in the early
java days.


Nic Ferrier


From tullmann@moab.cs.utah.edu Wed Nov 29 19:52:17 2000
Received: from moab.cs.utah.edu ([155.99.212.88])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 141HxQ-0004Eb-00; Wed, 29 Nov 2000 19:52:16 -0500
Received: (from tullmann@localhost)
	by moab.cs.utah.edu (8.9.1/8.9.1) id RAA12924;
	Wed, 29 Nov 2000 17:51:25 -0700 (MST)
Date: Wed, 29 Nov 2000 17:51:25 -0700
From: Patrick Tullmann <tullmann@cs.utah.edu>
To: Nic Ferrier <nferrier@tapsellferrier.co.uk>
Cc: cbj@gnu.org, classpath@gnu.org
Subject: Re: A java DEBUG flag
Message-ID: <20001129175125.I6408@cs.utah.edu>
References: <sa25a5c2.000@tapsellferrier.co.uk>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Mailer: Mutt 1.0pre3i
In-Reply-To: <sa25a5c2.000@tapsellferrier.co.uk>
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

Nic Ferrier wrote:
> Pat's solution is quite good but why don't you just use gcc to
> pre-process the .java files?

FYI, I use: 'gcc -E -x c -P -Wcomment -nostdinc -undef -C'.  

Or are you asking about the ".Java" extension I use?  ... or are you
asking about the custom pre-processor that was brought up earlier?

The nice thing about using a pre-process in gcj is that you guys won't 
have to hunt around to find out if the GNU cpp is installed.  :)

-Pat

----- ----- ---- ---  ---  --   -    -      -         -               -
Pat Tullmann                                       tullmann@cs.utah.edu
    That which does not kill you merely postpones the inevitable.


From tromey@creche.cygnus.com Wed Nov 29 21:42:16 2000
Received: from gasherbrumii.peakpeak.com ([207.174.178.17] helo=gash2.peakpeak.com)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 141Jfs-0007Ck-00; Wed, 29 Nov 2000 21:42:16 -0500
Received: from creche.cygnus.com (06-079.026.popsite.net [64.24.106.79])
	by gash2.peakpeak.com (8.9.3/8.9.3) with ESMTP id TAA18409;
	Wed, 29 Nov 2000 19:42:10 -0700
Received: (from tromey@localhost)
	by creche.cygnus.com (8.9.3/8.9.3) id TAA11295;
	Wed, 29 Nov 2000 19:52:07 -0700
To: Brian Jones <cbj@gnu.org>
Cc: Stuart Ballard <sballard@netreach.net>, classpath@gnu.org
Subject: Re: A java DEBUG flag
References: <3A1CA5C4.254DC238@albatross.co.nz> <874s0z6vn0.fsf@creche.cygnus.com> <m3zoim3gxz.fsf@rdu25-12-037.nc.rr.com> <3A219214.3CB70382@albatross.co.nz> <m3itp92c1m.fsf@rdu25-12-037.nc.rr.com> <m3bsv0jp2k.fsf@relativity.yi.org> <m3itp8z2px.fsf@rdu25-12-037.nc.rr.com> <3A23CEF6.2FC70376@netreach.net> <m3lmu3wdyo.fsf@rdu25-12-037.nc.rr.com>
X-Zippy:  Thank god!!..  It's HENNY YOUNGMAN!!
X-Attribution:  Tom
Reply-To: tromey@cygnus.com
From: Tom Tromey <tromey@cygnus.com>
Date: 29 Nov 2000 19:52:06 -0700
In-Reply-To: Brian Jones's message of "29 Nov 2000 00:21:19 -0500"
Message-ID: <877l5mw4rt.fsf@creche.cygnus.com>
Lines: 36
X-Mailer: Gnus v5.7/Emacs 20.5
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

>>>>> "Brian" == Brian Jones <cbj@gnu.org> writes:

Brian> I think I'd rather keep it to one java class and just use
Brian> multiple public final static variables.

I thought Bryce's plan was to have one master debug flag (a constant)
and then use properties for the individual things.

    if (Configuration.DEBUG && Boolean.getBoolean ("gnu.debug.this"))
      {
        ...
      }

Brian> Configure works like a shell script.  We have to define
Brian> variables by name and set them equal to true or false.  Then we
Brian> have to use AC_SUBST(VARIABLE_NAME).  I can't think of a way to
Brian> do this generically at the moment.  I don't think I've ever
Brian> seen an example of such either.

Sure.  You can make your own autoconf macro, something like
(untested):

dnl usage: DEBUG(variable, shell code)
AC_DEFUN([DEBUG],[
  $1=false
  if $2; then
    $1=true
  fi
  AC_SUBST([$1])
])

Then you can do `DEBUG(ALWAYS, true)' or whatever.

configure.in isn't just shell code, it has m4 preprocessing as well...

Tom


From nferrier@tapsellferrier.co.uk Thu Nov 30 08:02:12 2000
Received: from tf2.tapsellferrier.co.uk ([195.7.242.227] helo=tapsellferrier.co.uk)
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 141TLo-0003rN-00; Thu, 30 Nov 2000 08:02:12 -0500
Received: from tf-Message_Server by tapsellferrier.co.uk
	with Novell_GroupWise; Thu, 30 Nov 2000 13:11:04 +0000
Message-Id: <sa2651e8.005@tapsellferrier.co.uk>
X-Mailer: Novell GroupWise 5.2
Date: Thu, 30 Nov 2000 13:11:10 +0000
From: "Nic Ferrier" <nferrier@tapsellferrier.co.uk>
To: tullmann@cs.utah.edu
Cc: cbj@gnu.org, classpath@gnu.org
Subject: Re: A java DEBUG flag
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Disposition: inline
Sender: classpath-admin@gnu.org
Errors-To: classpath-admin@gnu.org
X-BeenThere: classpath@gnu.org
X-Mailman-Version: 2.0
Precedence: bulk
List-Help: <mailto:classpath-request@gnu.org?subject=help>
List-Post: <mailto:classpath@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=subscribe>
List-Id: <classpath.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/classpath>,
	<mailto:classpath-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/classpath/>

>>> Patrick Tullmann <tullmann@cs.utah.edu> 30-Nov-00 12:51:25 AM >>>

>FYI, I use: 'gcc -E -x c -P -Wcomment -nostdinc -undef -C'.  

That's what I meant.

>Or are you asking about the ".Java" extension I use?  

No... I was suggesting that gcc's preproc could take over from
that... for non-GNU projects I understand that you might need
something that doesn't rely on GCC but I think it's okay to have a GCC
dependancy for GNU java code.


>The nice thing about using a pre-process in gcj is that 
>you guys won't have to hunt around to find out if the GNU 
>cpp is installed.  :)

Exactly!  /8-> as I said... this should be the same for all GNU java
projects since the GNU system must include GCC.


Nic


