Fix overload selection with nested Any and object fallback

[ting-hong-shieh]

Summary

• prefer a specific overload over a later object fallback when nested gradual types create ambiguity in compatibility mode
• keep top-level Any ambiguous and preserve spec-compliant overload behavior
• add regression coverage for the original equality case and both guard cases

Root cause

Overload evaluation materialized the nested Any in the argument during step 5. The materialized argument could no longer select the bounded generic overload, so the later object fallback remained in the candidate set. Because the two overloads had incompatible return types, the call resolved to Unknown.

Validation

• cargo test test::overload (117 passed)
• python3 test.py --no-test --no-tensor-shapes --no-conformance --no-jsonschema
• git diff --check

Fixes #3977

[meta-codesync]

This pull request has been imported. If you are a Meta employee, you can view this in D115850637. (Because this pull request was imported automatically, there will not be any future comments.)

[github-actions]

Diff from mypy_primer, showing the effect of this PR on open source code:

============================================================
SUMMARY
============================================================
Total: +11 new errors, -24 fixed errors

Projects with changes (3):
  scipy: +1 -1
  scikit-learn: +0 -3
  pandas: +10 -20
============================================================

FULL DIFF DETAILS
------------------------------------------------------------

scipy (https://github.com/scipy/scipy)
- ERROR scipy/signal/_spectral_py.py:231:13-57: Implicit conversion of `numpy.bool | Unknown` to `bool` is not allowed [implicit-bool]
+ ERROR scipy/signal/_spectral_py.py:231:13-57: Implicit conversion of `numpy.bool` to `bool` is not allowed [implicit-bool]

scikit-learn (https://github.com/scikit-learn/scikit-learn)
- ERROR sklearn/decomposition/_nmf.py:395:19-57: The type of `permutation` is unknown; it is inferred as an implicit `Any` [unknown-variable-type]
- ERROR sklearn/feature_extraction/text.py:1310:21-63: The type of `j_indices` is unknown; it is inferred as an implicit `Any` [unknown-variable-type]
- ERROR sklearn/neighbors/_classification.py:306:20-59: The type of `mode` is unknown; it is inferred as an implicit `Any` [unknown-variable-type]

pandas (https://github.com/pandas-dev/pandas)
- ERROR pandas/core/arrays/sparse/array.py:2151:8-24: Implicit conversion of `NDFrame | builtins.bool | ndarray[tuple[Any, ...], dtype[numpy.bool]]` to `bool` is not allowed [implicit-bool]
- ERROR pandas/core/arrays/string_.py:492:16-30: Implicit conversion of `NDFrame | builtins.bool | ndarray[tuple[Any, ...], dtype[numpy.bool]]` to `bool` is not allowed [implicit-bool]
- ERROR pandas/core/arrays/string_.py:509:20-34: Implicit conversion of `NDFrame | builtins.bool | ndarray[tuple[Any, ...], dtype[numpy.bool]]` to `bool` is not allowed [implicit-bool]
- ERROR pandas/core/arrays/string_.py:574:16-30: Implicit conversion of `NDFrame | builtins.bool | ndarray[tuple[Any, ...], dtype[numpy.bool]]` to `bool` is not allowed [implicit-bool]
- ERROR pandas/core/arrays/string_.py:589:16-30: Implicit conversion of `NDFrame | builtins.bool | ndarray[tuple[Any, ...], dtype[numpy.bool]]` to `bool` is not allowed [implicit-bool]
- ERROR pandas/core/arrays/string_.py:589:16-73: Implicit conversion of `Literal[False] | NDFrame | numpy.bool | ndarray[tuple[Any, ...], dtype[numpy.bool]]` to `bool` is not allowed [implicit-bool]
+ ERROR pandas/core/arrays/string_.py:589:16-73: Implicit conversion of `Literal[False] | numpy.bool` to `bool` is not allowed [implicit-bool]
- ERROR pandas/core/arrays/string_.py:887:20-28: Object of class `bool` has no attribute `any` [missing-attribute]
+ ERROR pandas/core/arrays/string_.py:887:20-30: Implicit conversion of `numpy.bool` to `bool` is not allowed [implicit-bool]
- ERROR pandas/core/dtypes/cast.py:1422:8-53: Implicit conversion of `Literal[0] | NDFrame | builtins.bool | ndarray[tuple[Any, ...], dtype[numpy.bool]]` to `bool` is not allowed [implicit-bool]
+ ERROR pandas/core/dtypes/cast.py:1422:8-53: Implicit conversion of `Literal[0] | bool` to `bool` is not allowed [implicit-bool]
- ERROR pandas/core/dtypes/cast.py:1429:16-27: Implicit conversion of `NDFrame | builtins.bool | ndarray[tuple[Any, ...], dtype[numpy.bool]]` to `bool` is not allowed [implicit-bool]
- ERROR pandas/core/indexes/multi.py:4243:28-36: Implicit conversion of `signedinteger[_NBitIntP] | Unknown` to `bool` is not allowed [implicit-bool]
+ ERROR pandas/core/indexes/multi.py:4243:28-36: Implicit conversion of `signedinteger[_NBitIntP]` to `bool` is not allowed [implicit-bool]
- ERROR pandas/core/indexes/multi.py:4253:28-36: Implicit conversion of `signedinteger[_NBitIntP] | Unknown` to `bool` is not allowed [implicit-bool]
+ ERROR pandas/core/indexes/multi.py:4253:28-36: Implicit conversion of `signedinteger[_NBitIntP]` to `bool` is not allowed [implicit-bool]
- ERROR pandas/core/indexing.py:3893:16-29: Implicit conversion of `Literal[False] | numpy.bool | Unknown` to `bool` is not allowed [implicit-bool]
+ ERROR pandas/core/indexing.py:3893:16-29: Implicit conversion of `Literal[False] | numpy.bool` to `bool` is not allowed [implicit-bool]
- ERROR pandas/core/internals/construction.py:413:16-48: Implicit conversion of `Literal[False] | ndarray[tuple[Any, ...], dtype[numpy.bool]] | Unknown` to `bool` is not allowed [implicit-bool]
+ ERROR pandas/core/internals/construction.py:413:16-48: Implicit conversion of `Literal[False] | ndarray[tuple[Any, ...], dtype[numpy.bool]]` to `bool` is not allowed [implicit-bool]
- ERROR pandas/core/nanops.py:369:24-40: Implicit conversion of `NDFrame | builtins.bool | ndarray[tuple[Any, ...], dtype[numpy.bool]]` to `bool` is not allowed [implicit-bool]
- ERROR pandas/core/strings/accessor.py:731:16-34: Implicit conversion of `numpy.bool | Unknown` to `bool` is not allowed [implicit-bool]
+ ERROR pandas/core/strings/accessor.py:731:16-34: Implicit conversion of `numpy.bool` to `bool` is not allowed [implicit-bool]
- ERROR pandas/io/pytables.py:5980:8-16: Object of class `bool` has no attribute `any` [missing-attribute]
+ ERROR pandas/io/pytables.py:5980:8-18: Implicit conversion of `numpy.bool` to `bool` is not allowed [implicit-bool]
+ ERROR pandas/io/pytables.py:5980:8-38: Implicit conversion of `builtins.bool | numpy.bool | Unknown` to `bool` is not allowed [implicit-bool]
- ERROR pandas/io/pytables.py:5986:45-49: Argument `NDFrame | builtins.bool | ndarray[tuple[Any, ...], dtype[numpy.bool]]` is not assignable to parameter `mask` with type `ndarray` in function `_make_index_nan_rep` [bad-argument-type]
- ERROR pandas/io/pytables.py:5987:38-43: Unary `~` is not supported on `bool` [deprecated]
- ERROR pandas/plotting/_matplotlib/hist.py:102:18-39: The type of `values` is unknown; it is inferred as an implicit `Any` [unknown-variable-type]
- ERROR pandas/plotting/_matplotlib/hist.py:104:39-45: The type of this argument is unknown [unknown-argument-type]