Reword atexit docs - #156086
Conversation
For the `atexit` module: - Move common info from `register` to the module level (a lot of this was duplicated -- less maintainable and harder to read) - Use *interpreter shutdown* consistently, introducing it as a more general term for the old docs' "program termination". - Use consistent terminology (exit handlers) - Move warning for a mitigated footgun to a change entry - Add a new warning about keeping things usable Similarly clarify docs for the `atexit` attribute in `weakref`.
Documentation build overview
|
| **Note:** The effect of registering or unregistering functions from within | ||
| a cleanup function is undefined. |
There was a problem hiding this comment.
This is defined starting in 3.15; Python will run any handlers that have been added.
There was a problem hiding this comment.
Hmm, really?
import atexit
atexit.register(print, 'run me!')
@atexit.register
def foo():
atexit.register(print, 'run me too!')$ python3.15 /tmp/demo.py
run me!In any case, should we guarantee that? What about unregistering?
Sounds like material for a follow-up that's not backported to 3.14.
There was a problem hiding this comment.
Well, it works as documented -- and better than an infinite loop with a function registering itself...
| Attempts to start a new thread or :func:`os.fork` a new process | ||
| in an exit handler now leads to :exc:`RuntimeError`. | ||
| Previously, this could cause race conditions between the main Python | ||
| runtime thread freeing thread states while internal :mod:`threading` |
There was a problem hiding this comment.
Link to the term here?
| runtime thread freeing thread states while internal :mod:`threading` | |
| runtime thread freeing :term:`thread states <thread state>` while internal :mod:`threading` |
There was a problem hiding this comment.
Clarifying which internals exactly cause your crash isn't too useful.
(I considered removing the details entirely, but kept them to illustrate the warning above.)
encukou
added
needs backport to 3.13
|
I'll merge ~Friday if there are no objections. |
|
Thanks @encukou for the PR 🌮🎉.. I'm working now to backport this PR to: 3.13, 3.14, 3.15. |
|
GH-156878 is a backport of this pull request to the 3.15 branch. |
bedevere-app
Bot
removed
the
needs backport to 3.15
|
Sorry, @encukou, I could not cleanly backport this to |
|
GH-156879 is a backport of this pull request to the 3.14 branch. |
|
GH-156889 is a backport of this pull request to the 3.13 branch. |
For the `atexit` module: - Move common info from `register` to the module level (a lot of this was duplicated -- less maintainable and harder to read) - Use *interpreter shutdown* consistently, introducing it as a more general term for the old docs' *program termination*. - Use the term *exit handler* consistently - Move warning for a mitigated footgun to a change entry - Add a new warning about keeping things usable Similarly clarify docs for the `atexit` attribute in `weakref`. (cherry picked from commit 1d28836) Co-authored-by: Petr Viktorin <encukou@gmail.com>
2 participants
For the
atexitmodule:registerto the module level (a lot of this was duplicated -- less maintainable and harder to read)Similarly clarify docs for the
atexitattribute inweakref.