Re: 3.99.3

From: Erick Gallesio <>
Date: Thu, 1 Oct 1998 14:16:26 +0200 (CEST)

Andrew Dorrell writes:
> Hi,
> Just before you went on Hols Erick you sent a bulk patch with an
> appology and a promise of a bundeled release (aka 3.99.3) including all
> the updates and fixes.

I apologize for the long delay I took but I was overloaded by courses
that I didn't expect to have to do when I said that the 3.99.3 will be
out. Furthermore, I have tried to make a big step towards the 4.0
(i.e. a version with Windows support).

> Is this coming?

I have made a snapshot of the current state and named it 3.99.3 even
if not all the items I have planned to integrate in it are herein.
This release is available from kaolin and sunsite (it is already in
the devel/lang/lisp directory).

(An announce of the new release will be in a next message)

> It would really be appreciated as I am currently having
> nightmares whenever I need to install on a new machine! Access to the
> latest set of patches has always been a problem with STk. I found the
> main reason I updated to 3.99.2 was to get rid of the bug which made the
> CPU go 100% whenever an STk script was run in the background. I'm sure
> there must have been a patch somewhere to fix that though?

No there was not a patch for that, there were only report of the
problem but no fix at all and I was not able to reproduce it with a
simple program. As soon as someone (I don't remember who) told me that
the problem was with the interpreter running in background I have made
the fix and the release has come just after (ChangeLog says that I
have fixed it the 04/26 and the 3.99.2 release was out the 06/09).
During those 6 weeks, I had a lot of works with exams to prepare for
students, and I had not a lot of time for STk. Furthermore providing a
patch was not easy since there were too much changes in STk for
isolating it in a single patch (and I must admit that I'm not very
happy with a lot a patches floating around, particularly if they are
patches which don't correspond EXACTLY to the code that will be in a
future release).

> Also, is it possible that STk is dying?

I don't know what "dying" exactly means. If you think that it is no
more developed, this is of course false. If you think that there are
no users, I don't know but there was no significant decrease in the
number of access to the home page and to the number of downloads of
STk For instance there were 105 download in the 4 last days on
kaolin. Of course I have no idea of how many download are done from
sunsite and it's mirrors (I've not took the time to have very sure
stats, this is only a grep in the xferlog file, maybe I should try to
make things more precise).

> The list is very quiet and
> there is a *lot* of activity with new toolkits such as gtk and Qt. The
> fact that there are fast responses however implies to me that there is a
> small core of familiar users, but very few new users.

That's partially true. In fact the number of people in the mailing
list is something on which I have more clear ideas. In fact, as you
say there is a core of familiar user (2/3 very probably) and a very
volatile number of new users which are in the mailing list and which
disappear after a while. For now, there are 210 people in the mailing
list (i.e. quite the same number of people which are in the
fetchmail. Don't ask me from this number, how many people are
effectively reading messages, and even more, how many people are using

In fact comparing STk with Qt and Gtk is unfair, because the
interested people is a priori bigger for general graphical toolkit
than for a Scheme based one. What should be compared is the audience
of this toolkit vs Tk or Motif, imho.

BTW, there are far less developers for STk than for these toolkits, I
don't know why and it is not that I only want a cathedral development
style for STk. This is tsimply that nobody else seems to be interested
to co-maintain it ;-)

> My impression is that STk has a relatively low profile --- you don't
> even find a mension of it on the Tcl/Tk web pages (where PerlTk and
> Python are).

My interpretation is that the rms Tcl-war which arose 5 years ago
explains that. Guile is also connected to Tk and I suppose that it is
not mentioned too ;-)

> Anyway, a 3.99.3 bundle (and rpms if possible) would be greately
> appreciated.

RPMS are also built for i386 architecture. They are avialable from
kaolin or you sunsite mirror.

                -- (sad) Erick
Received on Thu Oct 01 1998 - 14:15:50 CEST

This archive was generated by hypermail 2.3.0 : Mon Jul 21 2014 - 19:38:59 CEST