Update maxima to 5.42.2

[arojas]

Also change the command line argument from -r (run in interactive mode) to --batch-string (run in batch mode), which is what we want here. The only reason it worked before was a bug in maxima that made -r behave like batch mode [1]. This is fixed in 5.42, so using -r now makes the maxima interface hang waiting for input.

--batch-string also echoes the input, so we need to remove one more line before processing the output.

Tarball: ​​https://vorboss.dl.sourceforge.net/project/maxima/Maxima-source/5.42.2-source/maxima-5.42.2.tar.gz

[1] ​​https://sourceforge.net/p/maxima/mailman/message/36177637/

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​b1dc567

Update maxima to 5.42.1

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​bda2320

Update one doctest for maxima 5.42

[fbissey]

No progress concerning the patch set. That's a bit annoying.

[fbissey]

I am getting

sage -t --long /usr/lib64/python2.7/site-packages/sage/symbolic/integration/integral.py
**********************************************************************
File "/usr/lib64/python2.7/site-packages/sage/symbolic/integration/integral.py", line 728, in sage.symbolic.integration.integral.integrate
Failed example:
    integrate(f, x, 0, 2*pi)  # long time (4s on sage.math, 2012)
Expected:
    integrate(abs(sin(x)), x, 0, 2*pi)
Got:
    0
**********************************************************************

[fbissey]

Which is wrong of course.

[fbissey]

And now that I look at the code in that area I see this is #10914 coming back.

    We will not get an evaluated answer here, which is better than
    the previous (wrong) answer of zero. See :trac:`10914`::
        sage: f = abs(sin(x))
        sage: integrate(f, x, 0, 2*pi)  # long time (4s on sage.math, 2012)
        integrate(abs(sin(x)), x, 0, 2*pi)

[arojas]

Replying to fbissey:

And now that I look at the code in that area I see this is #10914 coming back.

    We will not get an evaluated answer here, which is better than
    the previous (wrong) answer of zero. See :trac:`10914`::
        sage: f = abs(sin(x))
        sage: integrate(f, x, 0, 2*pi)  # long time (4s on sage.math, 2012)
        integrate(abs(sin(x)), x, 0, 2*pi)

Weird, since maxima does give the right answer now

[fbissey]

Since I was looking from the sage-on-gentoo perspective there is always the possibility that the maxima maintainer re-introduced a bug (doubtful but possible).

[arojas]

FWIW this is the commit that breaks it

​https://sourceforge.net/p/maxima/code/ci/801a0bb782fd09b9907a599541e1bff622ea453d

I can't decrypt what it means though, or why it breaks this test in sage but not in maxima itself.

[slelievre]

Regarding integrals of abs(sin(x)) and related, see

• ​https://trac.sagemath.org/query?order=id&desc=1&summary=~abs(sin

To explore anything to do with Maxima, it's important to remember that Sage starts Maxima by setting some defaults including domain : complex and loading some packages including abs_integrate, leading to differences with running Maxima on its own.

See ​lines 171 to 173 in src/sage/interfaces/maxima_lib.py

init_code = ['besselexpand : true', 'display2d : false',
            'domain : complex', 'keepfloat : true',
            'load(to_poly_solve)', 'load(simplify_sum)',
            'load(abs_integrate)', 'load(diag)']

Bugs in Maxima can be discussed on the maxima-discuss mailing list, always recalling these specifics of running Maxima in Sage.

• ​http://maxima.sourceforge.net/maximalist.html

[arojas]

Ah indeed, after 'load(abs_integrate)' I can reproduce the issue in maxima

(%i2) load(abs_integrate);
(%o2) /usr/share/maxima/5.42.1/share/contrib/integration/abs_integrate.mac
(%i3) integrate(abs(sin(x)),x,0,2*%pi);
(%o3)                                  0

[arojas]

Upstream report: ​https://sourceforge.net/p/maxima/bugs/3519/

[slelievre]

For reference, Maxima bugs dealing with abs(sin(x)) or abs(cos(x)):

• ​https://sourceforge.net/p/maxima/bugs/search/?q="abs(sin"
• ​https://sourceforge.net/p/maxima/bugs/search/?q="abs(cos"

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​ff20568	Merge branch 'develop' of git://git.sagemath.org/sage into t/26625/update_maxima_to_5_42_1
​2530dbf	Update maxima to 5.42.2

[kcrisman]

Replying to slelievre:

For reference, Maxima bugs dealing with abs(sin(x)) or abs(cos(x)):

• ​https://sourceforge.net/p/maxima/bugs/search/?q="abs(sin"
• ​https://sourceforge.net/p/maxima/bugs/search/?q="abs(cos"

Also for reference, see #12731 and/or ​the wiki part about abs_integrate problems.

[arojas]

Adding #12731 as a dependency, since it seems unlikely this will get any attention upstream.

[fbissey]

I am putting this as "need_review" since #12731 is getting attention. I am ready to move this ticket to a positive review once #12731 is merged.

[fbissey]

#12731 is closed so this can go in.