Use 'python -m sage.doctest' instead of 'sage -t'

[tobiasdiez]

Since sage -t is not working with meson. Progress towards #39875 (not sure if I got all of them).

📝 Checklist

• The title is concise and informative.
• The description explains in detail what this PR is about.
• I have linked a relevant issue or discussion.
• I have created tests covering the changes.
• I have updated the documentation and checked the documentation preview.

⌛ Dependencies

[github-actions]

Documentation preview for this PR (built with commit d0d96e3; changes) is ready! 🎉
This preview will update shortly after each push to this PR.

[user202729]

But sage -t works on my machine just fine?

Edit: okay I see the context, it's only when you install it instead of running from source tree.

[antonio-rojas]

IIRC sage-the-distro's python (and possibly other distros) doesn't ship the python→python3 symlink, so we should be using the python3 command consistently.

Side note: as long as sage-the-distro allows installing its own Python, it should be CI tested, which doesn't seem to be the case currently.

[user202729]

as long as sage -t is supported, surely it should be tested somewhere? What if some future change break it and some developers workflow is affected?

If you want to unsupport sage -t and ask them to migrate to python -m something… possible but you may want to announce it first.

[antonio-rojas]

as long as sage -t is supported, surely it should be tested somewhere? What if some future change break it and some developers workflow is affected?

If you want to unsupport sage -t and ask them to migrate to python -m something… possible but you may want to announce it first.

It is still being tested in sage/tests/cmdline.py (which is the natural place to do so). By the way, that file should not be installed by meson.

[antonio-rojas]

This fixes doctesting the sage/doctest dir for me in an installed, meson-built sage. However, the doctesting output still displays sage -t:

> python -m sage.doctest /usr/lib/python3.13/site-packages/sage/doctest 
Running doctests with ID 2025-06-10-13-55-18-e5673ad5.
Running with SAGE_LOCAL='/usr' and SAGE_VENV='/usr'
Using --optional=pip,sage
Features to be detected: 4ti2,SAGE_SRC,benzene,bliss,buckygen,conway_polynomials,coxeter3,csdp,cvxopt,cvxopt,database_cremona_ellcurve,database_cremona_mini_ellcurve,database_cubic_hecke,database_ellcurves,database_graphs,database_jones_numfield,database_knotinfo,dot2tex,dvipng,ecm,fpylll,fricas,gap_package_atlasrep,gap_package_design,gap_package_grape,gap_package_guava,gap_package_hap,gap_package_polenta,gap_package_polycyclic,gap_package_qpa,gap_package_quagroup,gfan,giac,glucose,graphviz,imagemagick,info,ipython,jmol,jupymake,jupyter_sphinx,kenzo,kissat,latte_int,lrcalc_python,lrslib,mathics,matroid_database,mcqd,meataxe,meson_editable,mpmath,msolve,nauty,networkx,numpy,palp,pandoc,pdf2svg,pdftocairo,pexpect,phitigra,pillow,plantri,polytopes_db,polytopes_db_4d,pplpy,primecountpy,ptyprocess,pycosat,pycryptosat,pynormaliz,pyparsing,python_igraph,requests,rpy2,rubiks,sage.combinat,sage.geometry.polyhedron,sage.graphs,sage.groups,sage.libs.braiding,sage.libs.ecl,sage.libs.flint,sage.libs.gap,sage.libs.giac,sage.libs.homfly,sage.libs.linbox,sage.libs.m4ri,sage.libs.ntl,sage.libs.pari,sage.libs.singular,sage.misc.cython,sage.modular,sage.modules,sage.numerical.mip,sage.plot,sage.rings.complex_double,sage.rings.finite_rings,sage.rings.function_field,sage.rings.number_field,sage.rings.padics,sage.rings.polynomial.pbori,sage.rings.real_double,sage.rings.real_mpfr,sage.sat,sage.schemes,sage.symbolic,sage_numerical_backends_coin,sagemath_doc_html,scipy,singular,sirocco,sloane_database,sphinx,symengine_py,sympy,tdlib,threejs,topcom
Doctesting 15 files.
sage -t --warn-long 5.0 --random-seed=214446839433195787860917771076279359673 /usr/lib/python3.13/site-packages/sage/doctest/test.py
    [24 tests, 0.49s wall]
sage -t --warn-long 5.0 --random-seed=214446839433195787860917771076279359673 /usr/lib/python3.13/site-packages/sage/doctest/check_tolerance.py
    [19 tests, 0.02s wall]
[...]

[antonio-rojas]

Nevermind, I forgot to pass -l when testing. There are still a few failures:

**********************************************************************
File "src/sage/src/sage/doctest/test.py", line 241, in sage.doctest.test
Failed example:
    subprocess.call(["sage", "-tp", "1000000", "--timeout=120",  # long time
         "--die_timeout=10", "--optional=sage",
         "--warn-long", "0", "99seconds.rst", "interrupt_diehard.rst"], **kwds2)
Expected:
    Running doctests...
    Doctesting 2 files using 1000000 threads...
    Killing test 99seconds.rst
    Killing test interrupt_diehard.rst
    ----------------------------------------------------------------------
    Doctests interrupted: 0/2 files tested
    ----------------------------------------------------------------------
    ...
    128
Got:
    usage: sage [-h] [-v] [-q] [--simple-prompt] [-V] [-n [{jupyter,jupyterlab}]]
                [-c [COMMAND]]
                [file]
    sage: error: unrecognized arguments: -tp --timeout=120 --die_timeout=10 --optional=sage --warn-long 0 99seconds.rst interrupt_diehard.rst
    2
**********************************************************************
File "src/sage/src/sage/doctest/test.py", line 259, in sage.doctest.test
Failed example:
    pid = int(open(F).read())
Exception raised:
    Traceback (most recent call last):
      File "/usr/lib/python3.13/site-packages/sage/doctest/forker.py", line 730, in _run
        self.compile_and_execute(example, compiler, test.globs)
        ~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
      File "/usr/lib/python3.13/site-packages/sage/doctest/forker.py", line 1154, in compile_and_execute
        exec(compiled, globs)
        ~~~~^^^^^^^^^^^^^^^^^
      File "<doctest sage.doctest.test[25]>", line 1, in <module>
        pid = int(open(F).read())
    ValueError: invalid literal for int() with base 10: ''
**********************************************************************
File "src/sage/src/sage/doctest/test.py", line 261, in sage.doctest.test
Failed example:
    os.kill(pid, signal.SIGQUIT)  # 2 seconds passed => still alive
Exception raised:
    Traceback (most recent call last):
      File "/usr/lib/python3.13/site-packages/sage/doctest/forker.py", line 730, in _run
        self.compile_and_execute(example, compiler, test.globs)
        ~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
      File "/usr/lib/python3.13/site-packages/sage/doctest/forker.py", line 1154, in compile_and_execute
        exec(compiled, globs)
        ~~~~^^^^^^^^^^^^^^^^^
      File "<doctest sage.doctest.test[27]>", line 1, in <module>
        os.kill(pid, signal.SIGQUIT)  # 2 seconds passed => still alive
                ^^^
    NameError: name 'pid' is not defined
**********************************************************************
File "src/sage/src/sage/doctest/test.py", line 263, in sage.doctest.test
Failed example:
    os.kill(pid, signal.SIGQUIT)  # 10 seconds passed => dead  # random
Expected:
    Traceback (most recent call last):
    ...
    ProcessLookupError: ...
Got:
    <BLANKLINE>
    Traceback (most recent call last):
      File "/usr/lib/python3.13/site-packages/sage/doctest/forker.py", line 730, in _run
        self.compile_and_execute(example, compiler, test.globs)
        ~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
      File "/usr/lib/python3.13/site-packages/sage/doctest/forker.py", line 1154, in compile_and_execute
        exec(compiled, globs)
        ~~~~^^^^^^^^^^^^^^^^^
      File "<doctest sage.doctest.test[29]>", line 1, in <module>
        os.kill(pid, signal.SIGQUIT)  # 10 seconds passed => dead  # random
                ^^^
    NameError: name 'pid' is not defined
**********************************************************************
1 item had failures:
   4 of  56 in sage.doctest.test
    [55 tests, 4 failures, 175.01s wall]
----------------------------------------------------------------------
sage -t --long --warn-long 30.0 --random-seed=150612886400060636501286193160487656764 src/sage/src/sage/doctest/test.py  # 4 doctests failed
----------------------------------------------------------------------

[tobiasdiez]

IIRC sage-the-distro's python (and possibly other distros) doesn't ship the python→python3 symlink, so we should be using the python3 command consistently.

Right, forgot about this one. Thanks!

as long as sage -t is supported, surely it should be tested somewhere? What if some future change break it and some developers workflow is affected?

If you want to unsupport sage -t and ask them to migrate to python -m something… possible but you may want to announce it first.

./sage -t is still used in CI to run the tests, that should give enough confidence that it's not broken. Moreover, sage -t just directs to sage-runtest which is a 3 line wrapper doing exactly the same as python -m sage.doctest.
I hope we can deprecate sage's custom doctester relatively soon and just use pytest instead.

Nevermind, I forgot to pass -l when testing. There are still a few failures:

Should be fixed now, at least partly. One of these tests uses tempfile which for some reason is not working well for the meson build (and should be replaced by built-in python functions anyway).

[antonio-rojas]

LGTM now