On Sun, 05 Jan 2020 12:37:24 +1300 Greg Ewing <greg.ew...@canterbury.ac.nz> wrote: > On 5/01/20 11:41 am, Antoine Pitrou wrote: > > > +1 for pointing to `weakref.finalize` as a better > > alternative for things that may survive until interpreter shutdown. > > I'm not sure that weakref.finalize is much better. It can still be > called at an arbitrary time from an arbitrary thread with the rest > of the system in an unknown state, so it needs to be written very > defensively to cope with that.
I didn't say otherwise. The main thing going for `weakref.finalize` is that it is called at the beginning of the interpreter shutdown while __del__ may be called much later in the process. > I don't think a blanket warning against using __del__ would be > very helpful, because many of its pitfalls also apply to the > suggested alternatives. This is an area where there's no > substitute for an in-depth understanding of the issues. I didn't suggest a blanket warning either. I do believe there should be a discussion, somewhere in the docs, of the various options available for finalization and their respective pitfalls. However, many (most?) Python developers know about __del__, while few know about `weakref.finalize`, so it makes sense to add a pointer to the latter in the docs for the former. Regards Antoine. _______________________________________________ Python-Dev mailing list -- python-dev@python.org To unsubscribe send an email to python-dev-le...@python.org https://mail.python.org/mailman3/lists/python-dev.python.org/ Message archived at https://mail.python.org/archives/list/python-dev@python.org/message/YEMIGOLDY4KZYG3KU2UEXWMS3G6G7BC4/ Code of Conduct: http://python.org/psf/codeofconduct/
