#438 rat/tellrat/modulus:2 can return -1 on GCL (SBCL OK)

[Stavros Macrakis]

algebraic:true$ modulus:2$
tellrat(a^2+a)$

rat(a^2) => - a

With modulus=2, -1 is supposed to simplify to 1.

This can be fixed (bizarrely) by ev:

ev(%) => a

[Wolfgang Jenkner]

Logged In: YES
user_id=581700

On both SBCL and CLISP I get the expected result

(C4) rat(a^2);
(D4)/R/ a

I'd guess this is the same GCL specific misfeature as #706562
(mod(-2,4) => -2 not 2).

[Robert Dodier]

Logged In: YES
user_id=501686

Observed w/ 5.9.3cvs / gcl 2.6.7, but not 5.9.3cvs / sbcl
0.9.9 or clisp 2.38.

[Robert Dodier]

Referred-to ticket #706562 is presumably https://sourceforge.net/p/maxima/bugs/264/ (due to system-wide renumbering several years ago). #264 is reported to be fixed so maybe this bug just needs to be retested.

[Robert Dodier]

Retested w/ not-too-old Maxima + GCL and bug is still present.

(%i3) algebraic:true$ modulus:2$

(%i5) tellrat(a^2+a)$

(%i6) 
rat(a^2);
(%o6)/R/                              - a
(%i7) ev(%);
(%o7)/R/                               a

Maxima version: "branch_5_44_base_181_geeb4b055f_dirty"
Maxima build date: "2020-12-08 08:54:55"
Host type: "i686-pc-linux-gnu"
Lisp implementation type: "GNU Common Lisp (GCL)"
Lisp implementation version: "GCL 2.6.12"
User dir: "/home/robert/.maxima"
Temp dir: "/tmp"
Object dir: "/home/robert/.maxima/binary/branch_5_44_base_181_geeb4b055f_dirty/gcl/GCL_2_6_12"
Frontend: false