I wanted to know the behavior of atexit.register with the same function added more than once and found out it is not explicitly documented.  I experimented and made a patch to the doc.

Maybe one of you gentlemen would like to review this.

I assume the patch is an improved description of reality. The old 'without binding the original name to ``None``.' is confusing to me in that context, so good riddance. However, I am not familiar with atexit or how it is intended to work. If you are not sure either, you might ask on pydev.

I wrote the patch after testing in a shell, so I’m confident it describes reality; I was more asking a wording/phrasing review.

+ This function returns *func* which makes it possible
add , after 'func'.

+This obviously only works with functions that ...
I think I would like this better without 'obviously' (which you inherited).

Otherwise, seems good to go to me.

Thanks.  I’ve also realized that the doc does not mention that any callable can be used; I don’t know if we should say it explicitly (I think I did a change like that recently in another file), or if we expect people to just know it from experience, or if we should add an entry for “function” in the glossary.

New changeset 55fc092dad72 by Éric Araujo in branch '3.2':
Improve doc for atexit.register and unregister (#12297)
http://hg.python.org/cpython/rev/55fc092dad72

New changeset f7163afecb97 by Éric Araujo in branch 'default':
Merge fixes for #1326113 and #12297 from 3.2
http://hg.python.org/cpython/rev/f7163afecb97

New changeset a99632426af5 by Éric Araujo in branch '2.7':
Improve doc for atexit.register (#12297)
http://hg.python.org/cpython/rev/a99632426af5

Closing as fixed, but if you have any feedback on my function vs. callable question, please share.