#3974 rtest_to_poly_solve failure in maxima-5.46.0

[Michael Orlitzky]

Full bug report: https://bugs.gentoo.org/838202

Failure info:

Running tests in rtest_namespaces: WARNING: redefining MAXIMA::INTERN-INVERT-CASE in DEFUN
WARNING: redefining MAXIMA::DIMENSION-ATOM in DEFUN
WARNING: redefining MAXIMA:MAKESTRING in DEFUN
WARNING: redefining MAXIMA::REMVALUE in DEFUN
WARNING: redefining MAXIMA::KILL1 in DEFUN
WARNING: redefining MAXIMA:MSIZE-ATOM in DEFUN
WARNING: redefining MAXIMA::DIMENSION-PREFIX in DEFUN
WARNING: redefining MAXIMA::DIMENSION-INFIX in DEFUN
WARNING: redefining MAXIMA::DIMENSION-NARY in DEFUN
WARNING: redefining MAXIMA::DIMENSION-POSTFIX in DEFUN
WARNING: redefining MAXIMA::DIMENSION-NOFIX in DEFUN
WARNING: redefining MAXIMA::DIMENSION-MATCH in DEFUN

********************** Problem 18 (line 52) ***************
Input:
map(fundef, [f, g, hh, hh])


Result:
                                                                         2
                                                           n            x
[f(x) := foo|my_constant sin(x), g(n) := expand((f(a) + n!) ), hh(x) := --, 
                                                                        2
                                                                             2
                                                                            x
                                                                   hh(x) := --]
                                                                            2

This differed from the expected result:
                                                           n
[f(x) := foo|my_constant sin(x), g(n) := expand((f(a) + n!) ), 
                                                   1   2
                                         hh(x) := (-) x , hh(z) := f(z) - g(2)]
                                                   2

********************** Problem 22 (line 67) ***************
Input:
hh(u - v)


Result:
       2
(u - v)
--------
   2

This differed from the expected result:
                                 2
(- 1729) sin(v - u) - 2989441 sin (a) + (- 6916) sin(a) - 4

********************** Problem 25 (line 76) ***************
Input:
hh(u - v)


Result:
hh(u - v)

This differed from the expected result:
       2
(u - v)
--------
   2

82/85 tests passed (not counting 1 expected errors)

The following 3 problems failed: (18 22 25)

Running tests in rtest_to_poly_solve: 
********************** Problem 210 (line 760) ***************
Input:
nicedummies(%solve([b2 + a1 = r2, a1 b2 = r3, a2 b2 + a3 b1 = r4, 
a3 b2 + a5 + r4 = d5, a5 b1 = 0], [a1, a2, a3, a5, r2, r3, r4, b1, b2], 
'simpfuncs = ['expand]))


Result:
%union([a1 = %c0, a2 = %c1, a3 = %c2, a5 = d5 - %c2 %c3 - %c1 %c3, b1 = 0, 
b2 = %c3, r2 = %c3 + %c0, r3 = %c0 %c3, r4 = %c1 %c3], 
                d5    %c5 %c6
[a1 = %c4, a2 = --- - ------- - %c5, a3 = %c5, a5 = 0, b1 = %c6, b2 = %c7, 
                %c7     %c7
r2 = %c7 + %c4, r3 = %c4 %c7, r4 = d5 - %c5 %c7], 
                           d5
[a1 = %c8, a2 = %c9, a3 = ----, a5 = 0, b1 = %c10, b2 = 0, r2 = %c8, r3 = 0, 
                          %c10
r4 = d5])

This differed from the expected result:
%union([a1 = %c0, a2 = %c3, a3 = %c4, a5 = d5 - %c4 %c5 + (- %c3) %c5, b1 = 0, 
b2 = %c5, r2 = %c5 + %c0, r3 = %c0 %c5, r4 = %c3 %c5], 
                           d5
[a1 = %c10, a2 = %c1, a3 = ---, a5 = 0, b1 = %c2, b2 = 0, r2 = %c10, r3 = 0, 
                           %c2
                          d5    %c7 %c8
r4 = d5], [a1 = %c6, a2 = --- - ------- - %c7, a3 = %c7, a5 = 0, b1 = %c8, 
                          %c9     %c9
b2 = %c9, r2 = %c9 + %c6, r3 = %c6 %c9, r4 = d5 - %c7 %c9])

312/313 tests passed (not counting 29 expected errors)

The following 1 problem failed: (210)

Running tests in rtest_vect: 
********************** Problem 19 (line 71) ***************
Input:
ev(vectorsimp(div (u a)) - (u div a + grad u . a), expandall : true)


Result:
0

... Which was correct, but was expected to be wrong due to a known bug in
 Maxima or SBCL.
66/66 tests passed (not counting 10 expected errors)

The following 1 problem passed but was expected to fail: (19)

Error summary:
Error(s) found:
  /var/tmp/portage/sci-mathematics/maxima-5.46.0/work/maxima-5.46.0/share/contrib/namespaces/rtest_namespaces.mac problems:
    (18 22 25)
  /var/tmp/portage/sci-mathematics/maxima-5.46.0/work/maxima-5.46.0/share/to_poly_solve/rtest_to_poly_solve.mac problem:
    (210)
Tests that were expected to fail but passed:
  /var/tmp/portage/sci-mathematics/maxima-5.46.0/work/maxima-5.46.0/share/vector/rtest_vect.mac problem:
    (19)
4 tests failed out of 17,412 total tests.

[Robert Dodier]

The errors from rtest_namespaces are no longer relevant since the package namespaces has been removed.

The expected failure in rtest_vect is good news since it means something has actually been fixed.

That leaves rtest_to_poly_solve problem number 210. It seems possible the answer is correct except for a renaming of variables and/or reordering of items in the %union.