Re: New Interpreter (was RE: Marketing Xconq?)

Re: New Interpreter (was RE: Marketing Xconq?)

[Richard Hunt]

>Well, what platform are you on?
>If the platform is Linux or Solaris, then chances are >good that
>Tcl/Tk is already installed.

I am using Debian, but I am a bit anally retentive (how do you spell that :) so I do my installation entirely using dselect, which seems to avoid installing a lot of packages that I want. I tend to avoid tcl/tk programs simply because they tend to be slow on my old computer, and xconq is the only tcl/tk program I still want to run.

Another thing I just remembered: with the version that I am using (a CVS checkout from about a month ago, since I don't have internet access on my pc during term time) there is a config.cache file in the xconq directory which means that eg. the --tclconfigdir and ---tkconfigdir flags are ignored.


Richard Hunt, 0102806h@student.gla.ac.uk

Re: New Interpreter (was RE: Marketing Xconq?)

[Eric McDonald]

On Wed, 19 Nov 2003, Richard Hunt wrote:

> I am using Debian, but I am a bit anally retentive (how do you 
>spell that :) 

AR ;-)

>so I do my installation entirely using dselect, 
>which seems to avoid installing a lot of packages that I want. I 
>tend to avoid tcl/tk programs simply because they tend to be slow 
>on my old computer, and xconq is the only tcl/tk program I still 
>want to run.

I understand. Up until the beginning of last year, I was in the 
same situtation. Even now, I am very discriminating about what I 
install; I guess old habits die hard.

> Another thing I just remembered: with the version that I am 
>using (a CVS checkout from about a month ago, since I don't have 
>internet access on my pc during term time) 

That is probably better for the grades (marks), I imagine. :-)

>there is a 
>config.cache file in the xconq directory which means that eg. the 
>--tclconfigdir and ---tkconfigdir flags are ignored.

Hmmm... That probably shouldn't be there.
You should be able to do a "make distclean" to make it go away 
though. Or "rm".

Eric

RE: New Interpreter (was RE: Marketing Xconq?)

[Eric McDonald]

On Mon, 17 Nov 2003, Erik Jessen wrote:

> My thought was this:
> - add an interpreter, using a standard language (like Perl) that can
> simply access existing data-structures.
> 
> The hard part (at least to me) is: how do you decide when to execute the
> new programs?  It's easy to load data-structures at initialization time
> & then have canned code execute it (what we do today).  It's harder to
> say "execute this subroutine any time combat occurs".

Well, with the Tcl/Tk interface we access a Tcl interpreter 
from C code all the time. As long as whatever language you have 
in mind can expose its interpreter to C in some for or another, 
you can call back into the interpreter and ask it to do things on 
your behalf. And vice versa, if the interpreter can call C 
functions. Then it is simply a matter of putting hooks into the 
relevant code sections.  But my experience with Tcl and Xconq is 
that this sort of arrangement is harder to debug.

Also, with a full blown interpreter being used, one must consider 
the security aspect. Especially since Xconq has the setgid bit set  
on Unix/Linux systems....

> Alas, I've not been able to locate my examples (I hope they weren't on
> my system that had the HDD crash), but I'm hopeful of backups.

Good luck finding them (and anything else you lost).

Eric

RE: New Interpreter (was RE: Marketing Xconq?)

[Erik Jessen]

My thought was this:
- add an interpreter, using a standard language (like Perl) that can
simply access existing data-structures.

The hard part (at least to me) is: how do you decide when to execute the
new programs?  It's easy to load data-structures at initialization time
& then have canned code execute it (what we do today).  It's harder to
say "execute this subroutine any time combat occurs".

Alas, I've not been able to locate my examples (I hope they weren't on
my system that had the HDD crash), but I'm hopeful of backups.

Erik


-----Original Message-----
From: Eric McDonald [mailto:mcdonald@phy.cmich.edu] 
Sent: Monday, November 17, 2003 7:34 PM
To: Erik Jessen
Cc: 'xconq'
Subject: New Interpreter (was RE: Marketing Xconq?)

Hi Erik,

On Mon, 17 Nov 2003, Erik Jessen wrote:

> I think having an interpreted language (v. the current Lisp-format
> datafile) would put Xconq way ahead of the competition (like ADC2 et
> al).

In past discussion, with Bruno I think, I jokingly referred to 
this as GDL++. But seriously, I run into things that I would like 
to do in Xconq that GDL cannot presently handle. Foremost among my 
desires are definitions based on forumlae rather than just static 
tables. It would also be nice to script events into certain games.

However, if one writes a new interpreter, it would still need to 
work within the limitations of the backend data structures. 
Writing a new, more powerful interpreter is an ambitious project, 
but it would certainly be fun to do sometime. (As would expanding 
the standing orders syntax.)

  Regards,
   Eric

P.S. How are the Xconq example "games" coming? If you have them 
ready, we can toss them into CVS, perhaps under lib/examples.

New Interpreter (was RE: Marketing Xconq?)

[Eric McDonald]

Hi Erik,

On Mon, 17 Nov 2003, Erik Jessen wrote:

> I think having an interpreted language (v. the current Lisp-format
> datafile) would put Xconq way ahead of the competition (like ADC2 et
> al).

In past discussion, with Bruno I think, I jokingly referred to 
this as GDL++. But seriously, I run into things that I would like 
to do in Xconq that GDL cannot presently handle. Foremost among my 
desires are definitions based on forumlae rather than just static 
tables. It would also be nice to script events into certain games.

However, if one writes a new interpreter, it would still need to 
work within the limitations of the backend data structures. 
Writing a new, more powerful interpreter is an ambitious project, 
but it would certainly be fun to do sometime. (As would expanding 
the standing orders syntax.)

  Regards,
   Eric

P.S. How are the Xconq example "games" coming? If you have them 
ready, we can toss them into CVS, perhaps under lib/examples.