ndt-users list created

[Matt Mathis <>]

Just to complicate the issue:  John just re-tested npad w/ -source 1.3, and it
seemes to work fine with IPv6.

He points out that using -source 1.3 just causes your current java environment
to use the 1.3 subset of the byte codes.  It does not change which runtime you
select so IPv6 works iff your runtime works.   (So if you want to use IPv6,
you may need 1.4.2, even if NPAD or NDT doesn't need it).

Given that neither NDT nor NPAD are using any 1.4 features, I still don't
understand what problem we are solving.

Perhaps there is another issue here:  NPAD's threhold for a minimally
competent user is much lower than NDT's.  "You need to upgrade your OS" is one
of things that NPAD sometimes advises.  Loosing %2 of the audience isn't
forward progress.  If there is a case were 1.4 works better than 1.3 we will
consider it.  (I also note that the "bytecode subset" issue means that we
don't have to pick the same version).

Thanks,
--MM--
-------------------------------------------
Matt Mathis      http://www.psc.edu/~mathis
Work:412.268.3319    Home/Cell:412.654.7529
-------------------------------------------
Evil is defined by mortals who think they know
"The Truth" and use force to apply it to others.

On Fri, 25 Aug 2006, Richard Carlson wrote:

> Hi Peter;
>
> In-line;
>
> At 04:40 PM 8/24/2006, Peter Van Epp wrote:
> >On Thu, Aug 24, 2006 at 10:38:12AM -0400, Richard Carlson wrote:
> > > All;
> > >
> > > By now you have all seen the email exchange between Peter Van Epp and
> > > me.  I would like some feedback from this user community before
> > deciding on
> > > a final solution to the bug Peter reported.
> > >
> > > Internet2 is now supplying an Knoppix based Live-CD with a pre-installed
> > > NDT server.  I built this server using the Java 1.4.2 SDK.
As Peter
> > > reported the jut released NDT-3.3.12 distribution fails on some Mac
> > > browsers with a run-time exception.  Peter's team reported the problem 
> > > and
> > > also supplied me with a patch (thanks Lixin).
> > >
> > > I applied this patch to my test system, generated a 3.3.13 distribution,
> > > and set this up on my test system (web100.internet2.edu) again using the
> > > 1.4.2 SDK.
> > >
> > > Peter now reports that the Safari browser on his Mac still fails, but
> > > firefox and opera work on the patched system.
> >
> >         Being finally out of
> >meetings-the-practical-alternative-to-anything-usefull :-)
> >
> >         This one turned out to be operator error, I had at some point
> > disabled
> >Java in Safari. When I enabled it all three browsers work the same.
>
> Great.  So Java 1.4 is a winner!  I'm going to change the distribution so
> it uses the "-source 1.4" option and that's how I'll build the ISO
> distribution.  This should keep the IPv6 code operational and keep the
> 1.4.2 JRE clients running too.
>
> > >
> > > One possible solution to the Safari problem is to switch to the Java 
> > > 1.5.0
> > > SDK development package.  Unfortunately, this is not a simple switch.  I
> > > have found that an Applet built with the 1.5.0 SDK will not run if the
> > > browser is using the 1.4.2 JRE.  All the user gets is a small red X in 
> > > the
> > > applet window.
> > >
> >
> >         Yes, thats what started us down this path :-). Lixin had the 1.5 
> > SDK
> >on our server and we got that exact thing on my machine. I'm still running
> >Mac
> >10.3 (while waiting for a new G5 to appear to install 10.4 on) so SDK
> >1.4.2 is
> >the last one available for 10.3 (and things like Win98 both of which are
> >admittadly old). SDK 1.5 is only available on Mac OS 10.4 according to our
> >Mac support folks.
> >
> > > There are 2 issues that I how need to address.  (1) what SDK version
> > should
> > > I use to build the Knoppix ISO?  Since this is burned onto the disk it
> > > needs to work for the majority of cases.  (2) should I recommend a SDK
> > > version for administrators who download and install the NDT tar file.
> > >
> > > Our options are:
> > >
> > > 1) do nothing and ignore those pesky Mac users .-) [not a real option]
> > > 2) apply the patches Lixin supplied and continue to use the 1.4.2
> > > SDK.  [This may mean that Mac Safari users will be unable to run tests.]
> > > 3) apply the patches and upgrade to the 1.5.0 SDK.  [This will force 
> > > user
> > > to upgrade the JRE plug-in and there needs to be a note on the NDT web
> > page
> > > telling users to do that.]
> > > 4) find out what new patches are needed to make the Safari browser run
> > with
> > > a 1.4.2 Applet.
> > >
> >
> >         Sounds like the oldest version that will support IP V6 would be a
> > good
> >option, 1.4.2 if 1.3 doesn't. It seems that later versions on the client 
> >side
> >are backward compatable to 1.4.2 (at least) and that would give the widest
> >range of possible clients out there who can connect.
> >         I'd think a note pointing out that an older SDK will be the most
> >compatible with old client OSs for the self compiling folks would also be a
> >good bet, so the builder can decide if they have all new machines or
> >not  :-).
> >
> >Peter Van Epp / Operations and Technical Support
> >Simon Fraser University, Burnaby, B.C. Canada
>
> ------------------------------------
>
>
>
> Richard A. Carlson                              e-mail: 
> 
> Network Engineer                                phone:  (734) 352-7043
> Internet2                                       fax:    (734) 913-4255
> 1000 Oakbrook Dr; Suite 300
> Ann Arbor,  MI  48104
>