Installation instruction and a small bug.

First, I think there is a small bug in the file todd-coxeter.lisp
The patch would be
5c5
< (proclaim '(type (vector t)  $todd_coxeter_state))
---
> (proclaim '(type (vector (t))  $todd_coxeter_state))
21c21
<     (declare (fixnum nvars) (type (vector t) multiply-table))
---
>     (declare (fixnum nvars) (type (vector (t)) multiply-table))

With the original version (vector (t)), it complains that the type (t) is not 
defined. But t is, of course, a type by itself.

Second, I have tried to compile the current source tree (version 
5.9.0pre-cvs) with cmucl-18d in a RH7.2 linux box and think there some 
interesting points: 

-In the INSTALL instructions there should be a prominent reference to the 
MAXIMA_DIRECTORY environment variable and the value it should have.

-The core dumped by the makefile is a bit bizarre. It does not accept the 
order "to_lisp();". However, it is easily corrected if one load the system 
and saves a core image without the -f user::run option in the makefile. To 
assure that maxima is started when this core is loaded, one may add the 
option: " -eval "(user::run)" to the cmucl option of maxima-run-lisp. There 
are no problems with "to_lisp();" and "(run)" with these trick. I suppose 
that the real problem lies inside cmucl implementation of the :init-function 
option when saving a core; but these hacks are enough to circunvect the 
problem.

Finally, in the documentation about emaxima, it would interesting to add that 
one should put something like
(autoload 'maxima "maxima" "Mathematical package" t)
in the .emacs file of the user if one wants to use maxima under emacs just by 
typing "M-x maxima".

That is all.
-- 
Juan Pablo Hierro Álvarez
hierro en ideafix punto litec punto csic punto es
Clave pública: 0xA8707ADF en pgp.rediris.es