New XMaxima: no socket problem with old Maxima 5.5 beta: success!

Dear Colleagues,

This a very delayed confirmation on my computer (under Windows Millen-
nium) of the fact (that Mike Clarkson first observed) that the"socket
problem"
in XMaxima when used together with Maxima 5.9.0 Release Candidate 1
(i.e. the fact that Maxima quite frequently becomes absent, it "dies", when
used with Mike's XMaxima and in such a case one has to restart Maxima
essentially losing much of his previous results) ceases being present when
Mike's XMaxima is used together with the old Maxima 5.5 beta (Dec. 2000).

Thanks to the valuable help and support of Mike Clarkson (and I am really
thankful to him for them) I have been able to construct the Greek Windows
executable (under the cp1253 encoding) of Mike's new XMaxima from the
TCL file he has been so kind to provide to me and check it here together
with
Maxima 5.5 beta. In my extensive tests I have been unable to observe the
"socket problem" any more and, therefore, I can confirm Mike's observation
of September 14, 2002 (below) that this problem (the "dying" of Maxima) is
present only with Maxima 5.9.0 Release Candidate and not with the old
Maxima 5.5 beta (under Windows Millennium I mean).

Recently, I begun to systematically use Maxima in my courses and
distribute it to some of my students here in Patras. My impressions are
quite positive and I intend to continue using Maxima (presently simultan-
eously with Mathematica). For the moment I am (and my students are)
using just Maxima 5.5 beta, but I feel that I may be able soon to move
to the use of Mike's new and so improved XMaxima (with so many new
capabilities such as the saving of the XMaxima concole and a new and
improved look of course) together with Maxima 5.5 beta.

On the other hand, I would be really very happy if I could be able to move
from Maxima 5.5 beta to the new Maxima 5.9.0 if possible, but the "socket
problem" (the "dying" of Maxima) makes it completely impractical as far
as I am concerned.

Therefore, I would be most helpful to Mike and any other colleague of
course who could pay further attention to the solution of this serious
problem, which makes my using of Maxima 5.9.0 under Windows
practically prohibitive under its present Release Candidate 1.

I just hope that this problem may have been resolved in Release
Candidate 2 of 5.9.0 or in the final Windows distribution of 5.9.0
so that I can move to Maxima 5.9.0, which has also so many
advantages over 5.5 beta and will be further developed to 6.0.

On September 19, 2002, Mike Clarkson reported that the
socket problem has to do with the command

load("foo.mac")

where foo.mac does not exist in GCL Maxima.

Your help on this point will be most gratefully appreciated.
Many sincere thanks in advance!

Mike claimed on September 15, 2002 that "Windows is 80%
of the downloads" in his software distributions. Therefore, any
further help towards the possibility of actually using the new
and so impressive XMaxima together with the new and so
improved Maxima 5.9.0 will be greatly appreciated and will
be useful not only to me and my students, but, possibly, to
other Windows user's as well.

Many sincere thanks for your having offered me the opportunity
to use Maxima and XMaxima and many thanks in advance for
any help for the solution of the "socket problem" in 5.9.0. This
will be decisive for the actual use of Maxima 5.9.0 in the
Windows environment.

Many thanks and best regards from Patras,

Nikos

----- Original Message -----
From: "Mike Clarkson" <mike at python>
To: <maxima@www.ma.utexas.edu>
Sent: Saturday, September 14, 2002 3:54 AM
Subject: Re: [Maxima] xmaxima: test impre


>
> From: "Nikolaos I. Ioakimidis" <ioakimidis at otenet>
> > In both machines error messages appear quite frequently during a
> > session and insist to be there. The error message
> >
> > Error: can not find channel named "sock76" ("sock80" too in other
cases).
> >
> > is frequent, very frequent. With the word "insist" I mean that I have
> > been finally obliged to Restart maxima from its left menu. In at least
> > one time an unjustified error of the above type appeared (and insisted)
> > although most times this happens after an error by me in the
> > command line. (In any case, insisting errors and frequent resort to
> > Restart are somewhat inconvenient.)
>
> I can confirm this, and am seeing it too, also under ME. Can anyone
> else report on seeing (or not seeing) this, with OS and version?
>
> What's happened is that maxima has died, so there's nothing on the
> end of the socket. I have checked in code to give the user a more
> informational message, but in essence you have to restart maxima.
>
> What's interesting is that I've also checked in code to make xmaxima
> work with the old 5.5 binary windows release, and am *NOT* seeing this
> dying of maxima. At least for the moment I do not suspect the xmaxima
> socket code, because when I look at this error, maxima has already
> dies from the windows process list.
>
> Mike.