Hacktoberfest 2026: the issues maintainers tagged for October, open and beginner-friendly. Browse Hacktoberfest issues

Suggestion: Consider changing the order for variable initialization in `__init__`

Open
#1,296 1 comment 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
5/5
Estimated time
Over a week
Newbie friendliness
25/100
Issue type
Feature
Clarity
Mostly clear
Activity status
Stale
Tech stack
python
Domain
tooling

Research direction

Start by reading the documentation section describing default factories and the generated init behavior. Then trace how attrs generates initialization for fields with supplied values, defaults, and factories, and check whether tests cover factory ordering or the NOTHING sentinel. Done means the proposed ordering is evaluated with tests and the supported behavior is documented or declined.

Written by the indexing model from the issue text.

Description

Feature

First of all thanks a lot for the awesome library! I just recently discovered it and it saved me so many headaches. Recently I was trying to do something of this sort, where either x or y has to be given to the __init__ method for this to be valid:

@define
class MyClass:
    x: int = field()
    y: int = field()
    
    @x.default
    def y_plus_one(self):
        return self.y + 1
        
    @y.default
    def x_plus_one(self):
        return self.x + 1

This however (as the docs state) fails because self.y would not be initialized during the call to y_plus_one. Even though the docs state this fails I wanted to try and fix the issue and managed to do so in a very hacky way. However in the process I thought up a non invasive way in which this may work.

If we first initialize those fields that were given a value through __init__ and then use the NOTHING sentinel to initialize the values for those attributes that were not initialized through __init__ but before calling the factories, the problem dissapears. The reordered __init__ method would look something like this (I'm not really sure how the method is autogenerated but its a rough idea):

@define
class MyClass:
    ...
    
    def __init__(self, x = NOTHING, y = NOTHING): # <- this is autogenerated
        # initialize all fields before calling factories
        if x is not NOTHING:
            self.x = x
        elif isinstance(self.__attrs_attrs__['x'].default, Factory):
            self.x = NOTHING
        elif self.__attrs_attrs__['x'].default is not None:
            self.x = self.__attrs_attrs__['x'].default

        if y is not NOTHING:
            self.y = y
        elif isinstance(self.__attrs_attrs__['y'].default, Factory):
            self.y = NOTHING
        elif self.__attrs_attrs__['y'].default is not None:
            self.y = self.__attrs_attrs__['y'].default

        # call the factories for all fields that were not given a value through init
        if self.x is NOTHING and isinstance(self.__attrs_attrs__['x'].default, Factory):
            self.x = self.__attrs_attrs__['x'].default.factory(self)
            if self.x is NOTHING:
                raise ValueError("a factory may not return the NOTHING sentinel")

        if self.y is NOTHING and isinstance(self.__attrs_attrs__['y'].default, Factory):
            self.y = self.__attrs_attrs__['y'].default.factory(self)
            if self.y is NOTHING:
                raise ValueError("a factory may not return the NOTHING sentinel")
    
    ...

This has the additional benefit that now I can throw more descriptive error messages during initialization, like:

    ...
    @x.default
    def y_plus_one(self):
        if self.y is NOTHING:
            raise ValueError("either x or y has to be initialized")
        return self.y + 1
    ...
Dominant language
Python
Stars
5.8k
Forks
490
Avg merge
3d 12h
Merged PRs (30d)
3

Getting set up

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from python-attrs/attrs

All issues in python-attrs/attrs

Similar issues

More Python issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.