python / python/cpython

`__future__` imports aren't always eager

Open
#151,577 2 comments 0 reactions 1 assignee View on GitHub

@brandtbucher is already working on this.

Since Jun 17, 2026.

3.15 3.16 interpreter-core topic-lazy-imports topic-parser type-bug
Dominant language
Python
Stars
77.2k
Forks
35.9k
PR merge metrics
PR metrics pending

Description

Bug report

It's my understanding that __future__ imports are always supposed to be eager. PEP 810 even goes so far as to specifically exclude lazy from __future__ import ... syntax:

No, future imports can’t be lazy because they’re parser/compiler directives. It’s technically possible for the runtime behavior to be lazy but there’s no real value in it.

However, the current implementation on 3.15 and main does allow __future__ imports to become lazy when lazy imports are globally enabled:

% ./python.exe -c 'from __future__ import annotations; print(globals()["annotations"])'                      
_Feature((3, 7, 0, 'beta', 1), None, 16777216)
% PYTHON_LAZY_IMPORTS=all ./python.exe -c 'from __future__ import annotations; print(globals()["annotations"])'
<lazy_import '__future__.annotations'>

I have a patch that flags these imports as not-lazy-eligible in the bytecode compiler (similar to * imports, imports in try blocks, and non-module-level imports). Backporting this fix to 3.15 would require bumping the bytecode magic number, though (it can be checked at runtime, but that likely introduces some overhead). Just thought I'd check to make sure my understanding is correct here and that this is indeed a bug.

Contributor guide

Open the contributing guide

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.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.