Slotscheck

Abstract

This page explains how we enforce the proper use of __slots__ on our classes with slotscheck tool. We go over what slotted classes, what slotscheck enforces, how to run slotscheck and how to configure it.

On top of the tools you already saw (ruff & basedpyright), we also have one more tool that performs static analysis on our code: slotscheck.

What is slotscheck

Slotscheck is a tool that focuses on enforcing proper use of __slots__ on classes.

What are slotted classes

If you aren't familiar with slotted classes, you should check the official documentation. That said, if you just want a quick overview:

• Slots allow you to explicitly declare all member attributes of a class (e.g. declaring __slots__ = ("a", "b") will make the class instances only contain variables a and b, trying to set any other attribute will result in an AttributeError).
• The reason we like using slots is the efficiency they come with. Slotted classes use up less RAM and offer a faster attribute access.

Example of a slotted class:

class FooBar:
    __slots__ = ("foo", "bar")

    def __init__(self, foo: str, bar: str) -> None:
        self.foo = foo
        self.bar = bar

x = FooBar("a", "b")
print(x.a, x.b)
x.c = 5  # AttributeError

With a low level project like mcproto, efficiency is important and __slots__ offer such efficiency at a very low cost (of simply defining them).

The purpose of slotscheck is to check that our slotted classes are using __slots__ properly, as sometimes, it is easy to make mistakes, which result in losing a lot of the efficiency that slots provide. Issues that slotscheck detects:

• Detect broken slots inheritance
• Detect overlapping slots
• Detect duplicate slots

How to use slotscheck

To run slotscheck on the codebase, you can use the following command:

slotscheck -m mcproto

Make you have an activated poetry virtual environment and you're in the project's root directory.

Configuring slotscheck

Sometimes, you may want to ignore certain files from being checked. To do so, you can modify the slotscheck configuration in pyproject.toml, under the [tool.slotscheck] option. That said, doing so should be very rare and you should have a very good reason to ignore your file instead of fixing the underlying issue.