use libgiac instead of giac for integration

[gh-mwageringel]

Now that giacpy_sage has been included into Sage by #29171, we can use it for integration instead of the Pexpect interface. This ticket adds a libgiac_integrator which replaces the giac_integrator in the default integrator (which uses Maxima, Giac, and Sympy).

This will resolve issues like #28913 or this ​ask-sage question where the Pexpect output cannot be parsed by Sage because Giac prints warning messages to the output.

The libgiac_integrator behaves slightly different than the Pexpect-based integrator, in that warning messages issued by Giac are not stripped away anymore.

sage: integrate(abs(cos(x)), x, 0, 2*pi)
Warning, integration of abs or sign assumes constant sign by intervals (correct if the argument is real):
Check [abs(cos(x))]
Discontinuities at zeroes of cos(x) were not checked
4

These messages are generally useful in interactive use. For doctests, the current branch adjusts most of the tests so that they do not depend on the presence of these messages.

Due to the problem at #31655, the first doctest in a file that uses libgiac results in some additional messages, but again the tests are written so that this does not matter. If that ticket is resolved, the second commit could be mostly reverted, but this is not strictly necessary. With the default integrator, these Giac loading messages will appear on the first integral that Maxima cannot solve.

Part of meta-ticket #16688.

[gh-mwageringel]

New commits:

​900ec88	31873: replace giac by libgiac for integration
​e023915	31873: ignore Giac loading messages in doctests, see #31655

[git]

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

​a47a360

31873: fix pyflakes warnings

[git]

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

​b591d2a

Merge tag '9.4.beta3' into #31873

[mkoeppe]

LGTM

[gh-mwageringel]

Thanks.

[vbraun]

test failures

sage -t --long --warn-long 45.1 --random-seed=0 src/sage/symbolic/integration/integral.py  # 2 doctests failed
sage -t --long --warn-long 45.1 --random-seed=0 src/doc/en/prep/Calculus.rst  # 1 doctest failed

[git]

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

​4481d50	Merge tag '9.4.beta6' into #31873
​bb15f0f	31873: fix doctests

[gh-mwageringel]

I have fixed the doctests. Now this passes ptestlong again. Also, this does not conflict with #32275.