ruff UP038: tool crashes

Hypothesis Nemotron 3 Super predicted to fire did not fire

isinstance(x, (int, float)) must distinguish actual int/float from bool (subclass of int) and Decimal; rewriting preserves type identity but not value semantics.

Program

Output

(no stdout)

(not run)

Judge Nemotron 3 Ultra advisory program-bug · low

The linter crashed because rule UP038 was removed from the codebase and cannot be selected, not because of any semantic change to the program. The program itself is trivial (z = True) and contains no isinstance or issubclass calls, so the rule would not have applied even if it existed. The crash is an environment/tooling issue unrelated to the rewrite's correctness.

Explanation Nemotron 3 Super unverified

The original assignment sets z to True; after applying the (nonexistent) fix from rule UP038—which has been removed—no change occurs in source text but the linter crashes due to attempting to invoke a deleted rule. Since the transformation never actually rewrites the code, the program’s behavior remains identical to before any attempted fix. However, because the tool fails to complete linting, any downstream reliance on its output (e.g., assuming the fix was applied or that analysis succeeded) may lead to incorrect conclusions about correctness, effectively altering the perceived semantics of the development process rather than the program itself. No actual runtime difference arises from the diff, as none was produced.

Reproduce

file

prog.py (the before program)

minimized program

r1-p005.py

evidence sheet

r1-p005.md

regression fixture

r1-p005.py the minimized program with a header stating the observed change, for the tool's test suite

python3 prog.py
ruff check --preview --isolated --select UP038 --fix prog.py
python3 prog.py

Upstream search GitHub issues Tavily Nemotron 3 Super + Ultra

UP038 fix · ruff UP038 unused variable · rule UP038 removal crash · linter autofix deleted rule

• related astral-sh/ruff/issues/13051 Existing report is about outdated documentation for UP038 fix safety, not a tool crash when the rule is removed.
• related astral-sh/ruff/issues/7871 Existing report is about UP038 rewriting code to be slower, not a tool crash due to rule removal.
• duplicate astral-sh/ruff/issues/18349 Existing report describes the same defect: enabling UP038 (a deprecated/removed rule) causes an error/crash because the rule no longer exists.

Details

program

Numeric Coercion and Type Checks

hazard

isinstance(x, (int, float)) must distinguish actual int/float from bool (subclass of int) and Decimal; rewriting preserves type identity but not value semantics.

oracle

lint crashed: ruff failed Cause: Rule `UP038` was removed and cannot be selected.

run

2026-10-17T21-20-52-142Z-httm-ruff-UP038

# non-pep604-isinstance (UP038)

Derived from the **pyupgrade** linter.

Fix is always available.

## Removed
This rule was removed as using [PEP 604] syntax in `isinstance` and `issubclass` calls
isn't recommended practice, and it incorrectly suggests that other typing syntaxes like [PEP 695]
would be supported by `isinstance` and `issubclass`. Using the [PEP 604] syntax
is also slightly slower.

## What it does
Checks for uses of `isinstance` and `issubclass` that take a tuple
of types for comparison.

## Why is this bad?
Since Python 3.10, `isinstance` and `issubclass` can be passed a
`|`-separated union of types, which is consistent
with the union operator introduced in [PEP 604].

Note that this results in slower code. Ignore this rule if the
performance of an `isinstance` or `issubclass` check is a
concern, e.g., in a hot loop.

## Example
```python
isinstance(x, (int, float))
```

Use instead:
```python
isinstance(x, int | float)
```

## Options
- `target-version`

## References
- [Python documentation: `isinstance`](https://docs.python.org/3/library/functions.html#isinstance)
- [Python documentation: `issubclass`](https://docs.python.org/3/library/functions.html#issubclass)

[PEP 604]: https://peps.python.org/pep-0604/
[PEP 695]: https://peps.python.org/pep-0695/