I found it somewhat misleading when I ran slotscheck mymod/ and it instead tested the installed version of mymod. This is especially since the command line options seem to imply that the argument is a file unless the -m option is provided.

I also tried this in an environment that doesn't have my module installed and it couldn't import it, which I also found surprising given that the module is there in the directory I ran slotscheck from.

Hi,

I tried configuring no-strict-imports i pyproject.toml, but running the tool via pre-commit ignored this completely. I concluded it probably doesn't read the pyproject.toml file inside pre-commit, so I insterad tried doing this in my pre-commit config:

  - repo: https://github.com/ariebovenberg/slotscheck
    rev: v0.14.1
    hooks:
      - id: slotscheck
        exclude: "test_*"
        args: [--no-strict-imports]

But the tool keeps complaining about imports. I think it simply doesn't pick the config.

Hi there,

Thanks for this useful library. I am integrating it in the starlite pipeline (https://github.com/starlite-api/starlite).

Would be awesome if I could run this as part of pre-commit. For this you will need to add a pre-commit config to the library - its not hard and will increase exposure i think.

best

Currently, slotscheck is marking classes that inherit from typing.Protocol as regular classes, and reports missing slots. However protocol classes are usually not used for actual inheritance, and defining __slots__ on them would mean a type-checker would be checking for presence of __slots__ when determining whether a type can be downcasted to given protocol class.

Here's an example of this:

from typing import Protocol, Callable, TypeVar, ParamSpec
from functools import wraps

P = ParamSpec("P")
R = TypeVar("R")

class DecoratorFunction(Protocol):
    def __call__(self, __x: Callable[P, R]) -> Callable[P, R]:
        ...
        
def log_with_markers(x: int, y: int) -> DecoratorFunction:
    def inner(func: Callable[P, R]) -> Callable[P, R]:

        @wraps(func)
        def wrapper(*args: P.args, **kwargs: P.kwargs) -> R:
            print(f"Function called, markers: {x=}, {y=}")
            return func(*args, **kwargs)

        return wrapper
    return inner

The DecoratorFunction class shouldn't be slotted, and slotscheck probably shouldn't enforce the presence of __slots__ there.

Hi there,

I'm running this library as a pre-commit hook. I introduced some changes, and do have an issue in my code I need to fix - but this causes slotscheck to go heywire:

Traceback (most recent call last):
  File "/opt/homebrew/Cellar/[email protected]/3.10.4/Frameworks/Python.framework/Versions/3.10/lib/python3.10/pkgutil.py", line 495, in find_loader
    spec = importlib.util.find_spec(fullname)
  File "/opt/homebrew/Cellar/[email protected]/3.10.4/Frameworks/Python.framework/Versions/3.10/lib/python3.10/importlib/util.py", line 94, in find_spec
    parent = __import__(parent_name, fromlist=['__path__'])
  File "/Users/naamanhirschfeld/workspace/starlite/starlite/__init__.py", line 3, in <module>
    from .app import Starlite
  File "/Users/naamanhirschfeld/workspace/starlite/starlite/app.py", line 19, in <module>
    from starlite.asgi import StarliteASGIRouter
  File "/Users/naamanhirschfeld/workspace/starlite/starlite/asgi.py", line 10, in <module>
    from starlite.routes import ASGIRoute, HTTPRoute
  File "/Users/naamanhirschfeld/workspace/starlite/starlite/routes.py", line 20, in <module>
    from starlite.controller import Controller
  File "/Users/naamanhirschfeld/workspace/starlite/starlite/controller.py", line 6, in <module>
    from starlite.handlers import BaseRouteHandler
  File "/Users/naamanhirschfeld/workspace/starlite/starlite/handlers/__init__.py", line 3, in <module>
    from .http import HTTPRouteHandler, delete, get, patch, post, put, route
  File "/Users/naamanhirschfeld/workspace/starlite/starlite/handlers/http.py", line 54, in <module>
    class HTTPRouteHandler(BaseRouteHandler):
  File "/Users/naamanhirschfeld/workspace/starlite/starlite/handlers/http.py", line 256, in HTTPRouteHandler
    def resolve_exception_handlers(self) -> Dict[Union[int, Type[Exception], ExceptionHandler]]:
  File "/opt/homebrew/Cellar/[email protected]/3.10.4/Frameworks/Python.framework/Versions/3.10/lib/python3.10/typing.py", line 312, in inner
    return func(*args, **kwds)
  File "/opt/homebrew/Cellar/[email protected]/3.10.4/Frameworks/Python.framework/Versions/3.10/lib/python3.10/typing.py", line 1143, in __getitem__
    _check_generic(self, params, self._nparams)
  File "/Users/naamanhirschfeld/.cache/pre-commit/repotvckiflq/py_env-python3.10/lib/python3.10/site-packages/typing_extensions.py", line 103, in _check_generic
    raise TypeError(f"Too {'many' if alen > elen else 'few'} parameters for {cls};"
TypeError: Too few parameters for typing.Dict; actual 1, expected 2

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/opt/homebrew/Cellar/[email protected]/3.10.4/Frameworks/Python.framework/Versions/3.10/lib/python3.10/runpy.py", line 196, in _run_module_as_main
    return _run_code(code, main_globals, None,
  File "/opt/homebrew/Cellar/[email protected]/3.10.4/Frameworks/Python.framework/Versions/3.10/lib/python3.10/runpy.py", line 86, in _run_code
    exec(code, run_globals)
  File "/Users/naamanhirschfeld/.cache/pre-commit/repotvckiflq/py_env-python3.10/lib/python3.10/site-packages/slotscheck/__main__.py", line 4, in <module>
    root()
  File "/Users/naamanhirschfeld/.cache/pre-commit/repotvckiflq/py_env-python3.10/lib/python3.10/site-packages/click/core.py", line 1130, in __call__
    return self.main(*args, **kwargs)
  File "/Users/naamanhirschfeld/.cache/pre-commit/repotvckiflq/py_env-python3.10/lib/python3.10/site-packages/click/core.py", line 1055, in main
    rv = self.invoke(ctx)
  File "/Users/naamanhirschfeld/.cache/pre-commit/repotvckiflq/py_env-python3.10/lib/python3.10/site-packages/click/core.py", line 1404, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/Users/naamanhirschfeld/.cache/pre-commit/repotvckiflq/py_env-python3.10/lib/python3.10/site-packages/click/core.py", line 760, in invoke
    return __callback(*args, **kwargs)
  File "/Users/naamanhirschfeld/.cache/pre-commit/repotvckiflq/py_env-python3.10/lib/python3.10/site-packages/slotscheck/cli.py", line 163, in root
    classes, modules = _collect(files, module, conf)
  File "/Users/naamanhirschfeld/.cache/pre-commit/repotvckiflq/py_env-python3.10/lib/python3.10/site-packages/slotscheck/cli.py", line 382, in _collect
    modules_all = consolidate(
  File "/Users/naamanhirschfeld/.cache/pre-commit/repotvckiflq/py_env-python3.10/lib/python3.10/site-packages/slotscheck/discovery.py", line 46, in consolidate
    for t in trees:
  File "/Users/naamanhirschfeld/.cache/pre-commit/repotvckiflq/py_env-python3.10/lib/python3.10/site-packages/slotscheck/discovery.py", line 158, in module_tree
    loader = pkgutil.get_loader(module)
  File "/opt/homebrew/Cellar/[email protected]/3.10.4/Frameworks/Python.framework/Versions/3.10/lib/python3.10/pkgutil.py", line 481, in get_loader
    return find_loader(fullname)
  File "/opt/homebrew/Cellar/[email protected]/3.10.4/Frameworks/Python.framework/Versions/3.10/lib/python3.10/pkgutil.py", line 501, in find_loader
    raise ImportError(msg.format(fullname, type(ex), ex)) from ex
ImportError: Error while finding loader for 'starlite.controller' (<class 'TypeError'>: Too few parameters for typing.Dict; actual 1, expected 2)

Description

Closes #52

Before merge

• [ ] tox runs successfully
• [ ] Docs updated

Before release (if applicable)

• [ ] Version updated in pyproject.toml
• [ ] Version updated in pre-commit hook example
• [ ] Version updated in changelog
• [ ] Branch merged
• [ ] Tag created and pushed
• [ ] Published

Consider the following:

from typing import TypedDict

foo = TypedDict("foo", {"bar": str})


class Foo(TypedDict):
    bar: str

slotscheck running in stric mode (require-subclass = true) will flag both foo and Foo as missing slots, although TypedDict are not used as "real class" and don't support slots:

PS D:\test> slotscheck main.py -v
ERROR: 'main:Foo' has no slots, but it could have.
       This is flagged because the strict `require-subclass` setting is active.
ERROR: 'main:foo' has no slots, but it could have.
       This is flagged because the strict `require-subclass` setting is active.
Oh no, found some problems!

Or should TypedDict support slots out of the box?

Bumps pydantic from 1.9.2 to 1.10.2.

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

Description

Since typing_extensions is only required on >3.10, attempts to run the latest release on <=3.10 fail if typing_extensions hasn't been installed into the target environment.

Protocol was added to the std typing module in 3.8 so this should only ever need to fallback to typing_extensions for Protocol on 3.7 and some 3.8 alpha releases.

I don't know why your CI is installing typing_extensions before running tests on 3.10 and 3.11.

Before merge

• [ ] tox runs successfully
• [ ] Docs updated

Before release (if applicable)

• [x] Version updated in pyproject.toml
• [x] Version updated in pre-commit hook example
• [x] Version updated in changelog
• [x] Branch merged
• [x] Tag created and pushed
• [x] Published

In this example:

@dataclass
class SemanticVersion:
    __slots__ = ("version", "prerelease", "build_metadata")

    version: tuple[int, int, int]
    prerelease: Optional[tuple[str, ...]] = None
    build_metadata: Optional[tuple[str, ...]] = None

slotscheck reports a failed import, due to prerelease (and build_metadata) being in __slots__ conflicting with class variables.

However clearly, these aren't actually class variables, they're in fact just fields of a dataclass with some default values. Slotscheck should probably be able to distinguish between actual class variables and usage in dataclasses, to prevent this false positive.

Note that sometimes, variables defined like this can actually be considered class variables by the dataclass decorator, but only if they're type-hinted as one, or if they aren't type hinted at all, see this example:

from typing import ClassVar

@dataclass
class Foobar:
    FOO: ClassVar[int] = 5  # actual class variable, not present in generated `__init__`
    BAR = 10  # another class variable, not present in generated `__init__`
    
    x: int = 2  # instance variable, with default value in `__init__`

Hi there,

So I encountered a problem that currently forced me to disable this tool in pre-commit. You can see the pertinent branch here: https://github.com/starlite-api/starlite/pull/315

The issue is that this plugin fails on import errors. In the above linked branch we are creating rust based extension and we are using a *.pyi file for its typings. Import fails because the extension is not built during pre-commit.

I'd like to suggest one of the following to handle this: stop failing on import errors. Maybe make this configurable with a command line arg, e.g. emit error, warning or continue on import error, but in any case the preferable opinion from my PoV is to to skip import errors and continue without failure.

(A bit of an obscure edge case, but worth documenting for future reference.)

The require-subclass mode warns of any classes which could be slotted, but are not.

class A(int):  # slots-required error triggered because A cloud be slotted, but isn't
    ...

class B(A):  # no slots-required error because B cannot be slotted (because one of its superclasses isn't)
    ...

However, protocols complicate things a bit:

class MyProto(Protocol):  # no slots-required error because protocols are not concrete classes
    ...

class C(MyProto):  # no slots-required error because one of its superclasses has no slots.
    ...

In the last case, we might want to emit a warning that C could be slotted if its protocol is. In the current situation, C is more-or-less silently allowed to not have slots. Note that this problem is not present in ABC because they are meant to be inherited from and do trigger slots-required themselves.

For debugging issues it'd be useful to be able to include some debug output. This could include:

• Given paths/arguments
• Settings used, and from which files they come
• installed packages, Python version etc.

slotscheck should support something like:

# slotscheck: ignore[require-superclass,overlaps]
class A(Base):
    __slots__ = ('a', 'b')

# slotscheck: ignore
class B(Base):
    __slots__ = ()

It's already documented why:

slotscheck mymodule/

doesn't work if the current directory is not explicitly in sys.path.

However, this is a corner of Python unknown to many. To prevent unnecessary frustration, a helpful message can be printed linking to the docs

Currently slotscheck has the require-subclass option, which enforces the use of __slots__ wherever slotted classes are subclassed. However, this greatly depends on whether builtin/extension classes are considered "slotted". For example, object has no __dict__. Flagging all direct object subclasses would essentially require slots for all classes in a module (that don't subclass from a non-slot class outside the module).