Re: [Maxima-discuss] Convolution and Plotting - bind stack overflow

[Stavros M. <mac...@al...>]

There appears to be a bug where maybe(x>NUMBER) adds a node to the assume
database, which is silly since of course asking the question adds no
information.

            -s

(%i3) :lisp (setq *print-circle* t *print-level* 50 *print-length* 50 linel
150)

150
(%i3) :lisp (symbol-plist '$%pi)

(SYSCONST T TEXWORD \pi DATA
 (#1=((MEQP $%PI
            #2=(3.141592653589793 DATA
                (#1#
                 #3=((MGRP #2#
                           #4=(2.718281828459045 DATA
                               (#5=((MGRP #4#
                                          #6=(1.618033988749895 DATA
                                              (#7=((MGRP #6#

 #8=(0.5772156649015329
                                                             DATA
                                                             (((MEQP $%GAMMA
                                                                     #8#)
                                                               CON $GLOBAL)
                                                              #7#))))
                                               ((MEQP $%PHI #6#) CON
$GLOBAL)
                                               #5#))))
                                #3# ((MEQP $%E #4#) CON $GLOBAL))))))))
      CON $GLOBAL)
  ((KIND $%PI $CONSTANT) CON $GLOBAL) ((KIND $%PI $IRRATIONAL))
  ((KIND $%PI $REAL)) ((KIND $%PI $NONINTEGER)))
 IMPLIED-QUOTEP T MPROPS (NIL $NUMER 3.141592653589793))
(%i3) maybe(x>234)$

(%i4) :lisp (symbol-plist '$%pi)

(SYSCONST T TEXWORD \pi DATA
 (#1=((MEQP $%PI
            #2=(3.141592653589793 DATA
                (#3=((MGRP
                      #4=(234 DATA (#5=((MEQP (234.0 DATA (#5#)) #4#))
#3#))   <<<<<<<< ???
                      #2#))
                 #1#
                 #6=((MGRP #2#
                           #7=(2.718281828459045 DATA
                               (#8=((MGRP #7#
                                          #9=(1.618033988749895 DATA
                                              (#10=((MGRP #9#

#11=(0.5772156649015329
                                                               DATA
                                                               (#12=((MGRP
#11#

 (0

DATA

(#12#))))
                                                                ((MEQP
$%GAMMA
                                                                       #11#)
                                                                 CON
$GLOBAL)
                                                                #10#))))
                                               ((MEQP $%PHI #9#) CON
$GLOBAL)
                                               #8#))))
                                #6# ((MEQP $%E #7#) CON $GLOBAL))))))))
      CON $GLOBAL)
  ((KIND $%PI $CONSTANT) CON $GLOBAL) ((KIND $%PI $IRRATIONAL))
  ((KIND $%PI $REAL)) ((KIND $%PI $NONINTEGER)))
 IMPLIED-QUOTEP T MPROPS (NIL $NUMER 3.141592653589793))


On Sun, May 17, 2020 at 2:30 PM Richard Fateman <fa...@cs...>
wrote:

>
>
> On 5/16/2020 8:36 PM, Stavros Macrakis wrote:
> > There is nothing illegal in any way about circular lists.
> > If a garbage collector doesn't handle them properly, it is a BUG.
> I entirely agree.  My suspicion was that in optimizing for other things,
> the GC handling of circular lists got messed up.  Most programs don't
> use circular lists, but they are supported in various ways in the Common
> Lisp standard, e.g.  *print-circle*.
>
> If the assume database is caching too much, and somehow stack space is
> getting consumed looking in it, that's also possible.
>
> For geometric predicates like   x<1  it  would seem unnecessary to also
> store x<2.  If that's what's going on.
> RJF
>
>
>
>