bug in continued fraction expansion ? / applying patches

I was going to write that it is not at all necessary to recompile 
Maxima.  That is, one
can patch lisp programs in a  running Maxima by loading files, or even 
by typing
new definitions.
Unfortunately, the style of patches on  sourceforge is not oriented to 
doing this,
and so the simplest mechanical way of fixing stuff is to essentially 
re-build the
system.

To illustrate how much easier Lisp can handle a patch,  imagine you wish to
change the Maxima  "integrate" command.  You can do this:

:lisp (defun $integrate(a b) "I changed integrate")

or you could load a file that had that new definition.

Sourceforge patches are diffs of files, therefore requiring not just 
"the right program"

but "the wrong program" and "the file in which the wrong program lives".

Instead of producing "the right program"  it produces "the file in which 
the wrong program lives, with changes".

This is an example of the retrograde motion of computer technology.
Compare to 1966, when Macsyma was built, and bug fixes were put
in the source code and in a "fix file".  Each user loaded Macsyma, which
loaded the fix file, and so everyone had a complete current system.
  When enough fixes accumulated, a new system was built.

RJF





On 5/31/2012 8:25 AM, jcarbaut at voila.fr wrote:
> Thank you very much for the link. I, for one, would care, but maybe it 
> will be a bit difficult for me to apply a patch and recompile Maxima ;-)
> Since I have a workaround and I don't need CF for more complicated 
> quadratic numbers, I may just stick with it.
> Anyway, it's a nice occasion to look under the hood, and to try to 
> understand how Maxima interacts with Lisp.
>
> Jean-Claude Arbaut
>
>     ========================================
>     Message du 31/05/12 16:38
>     De : "Alexander Klimov"
>     A : jcarbaut at voila.fr
>     Copie ? : "maxima"
>     Objet : Re: [Maxima] bug in continued fraction expansion ?
>
>     Hi. Thu, 31 May 2012 11:23:58 +0200 (CEST) ?? jcarbaut at voila.fr: >
>     cf(sqrt(n)) returns (apparently) what is expected except for some
>     numbers: You can try to use the patch that nobody cares to apply
>     for half a year:
>
>
>
>
>
> ------------------------------------------------------------------------
> D?couvrez ces stars qui viennent tout juste d'?tre mamans sur le 
> dossier f?te des m?res de Voila.fr 
> <http://actu.voila.fr/evenementiel/fete-des-meres/stars-mamans/>;
>
>
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima