fix: preserve tuple trailing comma (swev-id: sphinx-doc__sphinx-9367)

[casey-brooks]

Summary

• add regression coverage for single- and nested single-element tuples in tests/test_pycode_ast.py
• ensure _UnparseVisitor.visit_Tuple preserves trailing commas for single-element tuples

Reproduction

1. export MAMBA_ROOT_PREFIX=/workspace/micromamba
2. PYTHONPATH=/workspace/sphinx micromamba run -n sphinx-py39 pytest tests/test_pycode_ast.py -q

Observed Failure

FAILED tests/test_pycode_ast.py::test_unparse[(1,)-(1,)] - AssertionError: assert '(1)' == '(1,)'
FAILED tests/test_pycode_ast.py::test_unparse[((1,),)-((1,),)] - AssertionError: assert '((1))' == '((1,),)'

Testing

• PYTHONPATH=/workspace/sphinx micromamba run -n sphinx-py39 pytest tests/test_pycode_ast.py -q

Fixes #80

[casey-brooks]

Local validation:\n- PYTHONPATH=/workspace/sphinx micromamba run -n sphinx-py39 pytest tests/test_pycode_ast.py -q → 43 passed, 8 warnings\n- micromamba run -n sphinx-py39 flake8 sphinx/pycode/ast.py tests/test_pycode_ast.py → no issues

[noa-lucent]

Looks good—_UnparseVisitor.visit_Tuple now preserves trailing commas for single-element tuples and the added regression tests cover the previously failing cases. I wasn't able to run pytest locally (pytest not installed in this environment), but the new tests target the regression directly.

[rowan-stein]

LGTM. Verified spec compliance and tests cover regression.