[Axiom-developer] RE: axiom shell script

[Page, Bill]

Tim,

It seems to me that we should avoid changes that will
outdate TeXmacs as it is now being distributed. TeXmacs
looks for something called AXIOMsys (it could be the
AXIOMsys binary for Axiom or it could be a bash script
of the same but located somewhere else). If such a
script it necessary, it should still be called AXIOMsys.
We do not want to make changes that would require that
TeXmacs be changed again to accomodate a new/old naming
convention for Axiom.

The name AXIOMsys works "right out of the box" for the
old commercial version of Axiom that many people are
still running and except for the way Axiom is configured
on Debian, it works for the other new source linux
versions. It seems to me that it is the Debian distribution
that should change to be compatible with the others. If
(for reasons that still remain very unclear to me) the
directory structure on Debian must be changed from the
what has be distributed for Axiom so far, then the minimum
that would be required is the on Debian the name AXIOMsys
refer either via a link or by a simple script to the
real executable - where ever it has to be located. Surely
this is done all the time on Debian? No?

Running clef makes no sense when interfacing between
TeXmacs and Axiom. In fact using AXIOMsys is an
improvement for the commercial version of Axiom as
well sense that extra stuff is not hanging around
in memory somewhere between TeXmacs and Axiom.

So I still vote to keep it simple and just make sure
that on every Axiom installation (just like the old
commercial version) an AXIOMsys is in the path.

Cheers,
Bill Page.


> -----Original Message-----
> From: Tim Daly [mailto:address@hidden
> Sent: Tuesday, February 24, 2004 12:45 PM
> To: address@hidden; address@hidden; address@hidden;
> address@hidden
> Cc: address@hidden
> Subject: Re: axiom shell script
> 
> 
> *,
> 
> Camm sees a need to execute the axiom shell script rather than
> directly invoking AXIOMsys, the axiom binary. 
> 
> In the past I believe we moved away from using the shell script
> because it invoked clef (the axiom readline). 
> 
> The compromise position seems to be to invoke the shell script
> but do it with the -noclef option and that the shell script
> be rewritten to handle this. Axiom used to have a -noclef option
> and will have it again in the future.
> 
> (a) are there any objections?
> (b) if there are no objections we need to:
>    (1) merge Camm's axiom approach with the current clef version
>    (2) add -noclef handline
>    (3) modify the texmacs plugin to add -noclef
>    (4) backport the new axiom shell script to debian
> 
> Comments?
> 
> Tim
>