Update to maxima 5.48

[antonio-rojas]

Mostly output format changes, but it is not trivial to make test pass with both old and new versions

[antonio-rojas]

@tobiasdiez looks like this is still not available in conda, any idea who to ping to make it happen?

[github-actions]

Documentation preview for this PR (built with commit a58c077; changes) is ready! 🎉
This preview will update shortly after each push to this PR.

[tobiasdiez]

Only conda-forge/maxima-feedstock#40 needs to be merged by either @isuruf or @saraedum.

[antonio-rojas]

Reported this regression upstream at https://sourceforge.net/p/maxima/bugs/4601/

[orlitzky]

You may want to base this on #40574 in case some of the new output changes. I got started with the 5.48.x upgrade but gave up when I realized that not all of the output changes are improvements. (Thinking: we may want to support 5.47.x simultaneously rather than force people to upgrade to a version with regressions, but this is a lot of work).

[orlitzky]

IMO anything in a TESTS:: block is a good candidate to be moved to pytest, and to be checked with assert actual == expected. It's pretty dumb that we have to go through and update this output every time a new version of maxima is released. If the answer is right, the output is OK.

[antonio-rojas]

You may want to base this on #40574 in case some of the new output changes. I got started with the 5.48.x upgrade but gave up when I realized that not all of the output changes are improvements. (Thinking: we may want to support 5.47.x simultaneously rather than force people to upgrade to a version with regressions, but this is a lot of work).

Since that is already positively reviewed and this will take some time, I will rebase it after the next beta. Besides the non-trivial work of rewriting the doctests, supporting both versions also requires dealing with the pcprntd → *pcprntd* rename, which I have no idea how to do (I don't speak lisp).

[tobiasdiez]

supporting both versions also requires dealing with the pcprntd → *pcprntd* rename, which I have no idea how to do (I don't speak lisp).

A bit of dreaming: After #39030, we could have meson write the version of maxima (or actually all dependencies) into config.py. Then you should be able to conditionally use the correct ecl_eval call based on the version (and also make the doctests work for different maxima versions).

[antonio-rojas]

Actually, it looks like we may just drop that principal redefinition. It was originally added in #11987 to fix integrate(sin(x)*sin(x/3)/x^2, x, 0, oo), but that seems to work just fine nowadays without that redefinition. I don't completely understand the code though.

CORRECTION: That redefinition is there since the introduction of the maxima lib interface in 970a833, not clear why. Running CI now without it to see if anything breaks.

[antonio-rojas]

Seems to be needed:

**********************************************************************
Error: Failed example:: Got: 4/9
                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        
    integrate(1/x^3,x,-1,3)
Expected:
    Traceback (most recent call last):
    ...
    ValueError: Integral is divergent.
Got:
    4/9
**********************************************************************

[antonio-rojas]

After #40574 it looks like #20467 is back:

**********************************************************************
File "/usr/lib/python3.13/site-packages/sage/symbolic/integration/integral.py", line 1096, in sage.symbolic.integration.integral.integrate
Failed example:
    integral(sin(k*x)/x*erf(x^2), x, 0, oo, algorithm='maxima')
Exception raised:
    Traceback (most recent call last):
      File "/usr/lib/python3.13/site-packages/sage/doctest/forker.py", line 733, in _run
        self.compile_and_execute(example, compiler, test.globs)
        ~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
      File "/usr/lib/python3.13/site-packages/sage/doctest/forker.py", line 1157, in compile_and_execute
        exec(compiled, globs)
        ~~~~^^^^^^^^^^^^^^^^^
      File "<doctest sage.symbolic.integration.integral.integrate[152]>", line 1, in <module>
        integral(sin(k*x)/x*erf(x**Integer(2)), x, Integer(0), oo, algorithm='maxima')
        ~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
      File "/usr/lib/python3.13/site-packages/sage/misc/functional.py", line 787, in integral
        return x.integral(*args, **kwds)
               ~~~~~~~~~~^^^^^^^^^^^^^^^
      File "sage/symbolic/expression.pyx", line 13181, in sage.symbolic.expression.Expression.integral
        return integral(self, *args, **kwds)
      File "/usr/lib/python3.13/site-packages/sage/symbolic/integration/integral.py", line 1104, in integrate
        return integrator(expression, v, a, b)
      File "/usr/lib/python3.13/site-packages/sage/symbolic/integration/external.py", line 41, in maxima_integrator
        result = maxima.sr_integral(expression, v, a, b)
      File "/usr/lib/python3.13/site-packages/sage/interfaces/maxima_lib.py", line 802, in sr_integral
        return max_to_sr(maxima_eval(([max_integrate],
                         ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^
                                      [sr_to_max(SR(a)) for a in args])))
                                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
      File "sage/libs/ecl.pyx", line 828, in sage.libs.ecl.EclObject.__call__
        return ecl_wrap(ecl_safe_apply(self.obj, (<EclObject>lispargs).obj))
      File "sage/libs/ecl.pyx", line 356, in sage.libs.ecl.ecl_safe_apply
        raise RuntimeError("ECL says: {}".format(message))
    RuntimeError: ECL says: BINDING-STACK overflow at size 10240. Stack can probably be resized.
    Proceed with caution.
**********************************************************************

Reproducible in maxima itself (but only in 5.48):

(%i1) load(abs_integrate);
(%o1) /usr/share/maxima/5.48.1/share/contrib/integration/abs_integrate.mac
(%i2) integrate(sin(k*x)/x*erf(x^2),x,0,inf);

Maxima encountered a Lisp error:

 BINDING-STACK overflow at size 10240. Stack can probably be resized.
Proceed with caution.

Automatically continuing.
To enable the Lisp debugger set *debugger-hook* to nil.

[orlitzky]

After #40574 it looks like #20467 is back:

https://sourceforge.net/p/maxima/bugs/4603/