November 12, 2001

This week in DotGNU, No. 4 (November 12)

Author: JT Smith

See for general information about DotGNU. In this issue:

* Assistant maintainers needed for GlobalLogin auth project
* New versions of Portable.NET and pnetlib
* Status of Free Software efforts on Java
Assistant maintainers needed for GlobalLogin auth project

Mason Ham is working on the final preparations for the release
(under GPL) of the fully functional GlobalLogin virtual identity
system which is implemented in Java. Since he expects a high
level of interest for this system, he is looking for three people
who are interested in helping him maintain it. For each of the
areas "database", "crypto", and "servlets" one such sub-maintainer
is sought. Mason Ham can be reached by email at
mlham AT zambit DOT com.

New versions of Portable.NET and pnetlib

Portable.NET 0.2.2 and pnetlib 0.0.6 have been released:

Web Page:

The major changes in this release are to the runtime engine and
the linker. The engine now understands most of the IL instruction
set. The main omissions are exceptions, interfaces, delegates,
and tail calls.

The engine can now run "Hello World" slightly better than the
last release. Because virtual methods now work, it can use the
stream classes in "System.IO" to do the heavy lifting on "stdout".

Status of Free Software efforts on Java

I asked about the status of the Free Software efforts on Java,
here is the answer Brian Jones gave me:

With the exception of the AWT, I would say the essential components
are mostly working. There are things like cleaning up the JNI code to
check for exceptions more appropriately and the possibility of using
APR (Apache Portable Runtime) to make that code more portable to other
platforms without much effort. A few places within the code base have
UNIX-isms that shouldn't be where they are. I don't think the
SecurityManager is integrated into every method it is supposed to be.
It would be nice to support multi-language exception messages in both
Java and native code but that might be too much overhead. I'm not
confident in our text/string/character implementation which is
different from what gcj uses... mostly because it seems to fail every
Mauve test. Finally, there are not enough Mauve tests to make me say
it works well because a lot of functionality does not yet have a Mauve

That's the ugly side... the other side is that multiple VMs use some
form of Classpath and you can actually run some applications with it.
Contributing to Mauve is really easy and if you just barely know Java
you can still make valuable contributions there.

P.S. Translations of "This week in DotGNU" into other laguages are
very welcome; please let me know about them. Currently the only
language into which "This week in DotGNU" is translated is Polish.

"This week in DotGNU" is Copyright (C) 2001 by Norbert Bollow.
Verbatim copying and distribution of this entire issue is
permitted in any medium or format, provided this notice is

Click Here!