Silence initialization of giac

[tornaria]

When sage.libs.giac.giac is imported for the first time, some noise is output to stderr. This is annoying and complicates doctesting as ... has to be added to the first time giac is used.

This output comes from the line

Pygen('add_language(1)').eval()  # Add the french keywords in the giac library language.

which doesn't seem to be necessary.

This commit removes the line, and fixes a few doctests which where expecting some output when using libgiac for the first time.

Also label # long time a doctest in src/sage/calculus/functional.py which takes 6-7 seconds. Doing this in the first place motivates the current change as the initialization noise would happen in different tests depending on whether --long or not is used.

Same for a doctest in src/sage/calculus/tests.py, and one in src/sage/symbolic/integration/integral.py. A test in src/sage/libs/giac/giac.pyx was marked # random but after this fix it seems to work ok.

📝 Checklist

• The title is concise, informative, and self-explanatory.
• The description explains in detail what this PR is about.
• There are existing tests covering the changes.

[videlec]

You have a doctest failure

sage -t --random-seed=272486796797018539112758668052593627245 sage/symbolic/relation.py
**********************************************************************
File "sage/symbolic/relation.py", line 932, in sage.symbolic.relation.solve
Failed example:
    solve([(2/3)^x-2], [x], algorithm='giac')
Expected:
    [[-log(2)/(log(3) - log(2))]]
Got:
    Warning, argument is not an equation, solving (2/3)^sageVARx-2=0
    [[-log(2)/(log(3) - log(2))]]
**********************************************************************
1 item had failures:
   1 of 124 in sage.symbolic.relation.solve
    [395 tests, 1 failure, 12.41 s]

[tornaria]

You have a doctest failure

sage -t --random-seed=272486796797018539112758668052593627245 sage/symbolic/relation.py
**********************************************************************
File "sage/symbolic/relation.py", line 932, in sage.symbolic.relation.solve
Failed example:
    solve([(2/3)^x-2], [x], algorithm='giac')
Expected:
    [[-log(2)/(log(3) - log(2))]]
Got:
    Warning, argument is not an equation, solving (2/3)^sageVARx-2=0
    [[-log(2)/(log(3) - log(2))]]
**********************************************************************
1 item had failures:
   1 of 124 in sage.symbolic.relation.solve
    [395 tests, 1 failure, 12.41 s]

It should be fixed now, my bad...

This change was not intended to go in. The ... in this doctest corresponds to a warning, which doesn't show up in my version of giac (1.9.0-43) but it still shows up in the version bundled with sage.

I only meant to remove a few ... which correspond to the initialization of giac that this PR is silencing.

[kiwifb]

@fchapoton you are credited with the change adding "French keywords" in 7d3fd94 . What was your motivation with that?

[tornaria]

@fchapoton you are credited with the change adding "French keywords" in 7d3fd94 . What was your motivation with that?

No, that's just a style change. I tracked that line all the way to the original commit that includes giacpy into sage: d095c37.

Here is the line:

sage/src/sage/libs/giac/giac.pyx

Line 6263 in d095c37

Pygen('add_language(1)').eval(); # Add the french keywords in the giac library language.

The ticket for this is #29171.

Now, this code originates in https://gitlab.math.univ-paris-diderot.fr/han/giacpy-sage/, more precisely:
https://gitlab.math.univ-paris-diderot.fr/han/giacpy-sage/-/blob/master/giacpy_sage.pyx#L6386.

And looking at the giacpy history, the file giacpy_sage.pyx is created as a 5636-line file which already contains this line in https://gitlab.math.univ-paris-diderot.fr/han/giacpy-sage/-/commit/7b1bf00d364d52d0b3f74f72c0d21f9e95375fa6.

[kiwifb]

Thanks for doing the archeology. It is just so bizarre to single out French unless there is a deep reason to it. We definitely have to expunge this.

[kiwifb]

The bots are happy and everything that needed sorting has been. Let's move on.

[tornaria]

@kiwifb I had to add a small fix to support new giac. Can you verify that everything is still ok for you?

Fix doctest so it works with old and new giac.

This `...` corresponds to a warning in the version of giac bundled with
sage, but the warning is removed in current version of giac (tested with
1.9.0-43).

Before this `...` was ok since it was matching giac initialization, but
after the last commit this is no longer the case.

Placing the `...` in the same line as the output makes it optional.

[kiwifb]

I'll try to test it later today.

[github-actions]

Documentation preview for this PR is ready! 🎉
Built with commit: 656b940

[kiwifb]

Well Volker is already merging it. It does work OK with 1.9.0-43. Although, I am not pushing for either -41 or -43. Both give me a segfault in pari when running the giac testsuite and I haven't really had time to investigate further.

[tornaria]

Well Volker is already merging it. It does work OK with 1.9.0-43. Although, I am not pushing for either -41 or -43. Both give me a segfault in pari when running the giac testsuite and I haven't really had time to investigate further.

Try this: /~https://github.com/void-linux/void-packages/blob/master/srcpkgs/giac/patches/giac-pari-2.15-test.patch

I'm running 1.9.0-43.

[kiwifb]

Well Volker is already merging it. It does work OK with 1.9.0-43. Although, I am not pushing for either -41 or -43. Both give me a segfault in pari when running the giac testsuite and I haven't really had time to investigate further.

Try this: /~https://github.com/void-linux/void-packages/blob/master/srcpkgs/giac/patches/giac-pari-2.15-test.patch

I'm running 1.9.0-43.

Thanks heaps. Helps with 1.9.0-45 too, I just tested it.