Message198208
| Author |
eli.bendersky |
| Recipients |
eli.bendersky, ncoghlan, scoder, serhiy.storchaka |
| Date |
2013年09月21日.14:47:30 |
| SpamBayes Score |
-1.0 |
| Marked as misclassified |
Yes |
| Message-id |
<CAF-Rda_zWj34wnbHMOGSq92hgUoSkP=z-suYf23t5OtQq9qkxQ@mail.gmail.com> |
| In-reply-to |
<CADiSq7f_vHNPyGWGaDw9u=baEvp_oCve0fHoFEn4SrMXKmC66g@mail.gmail.com> |
| Content |
On Sat, Sep 21, 2013 at 7:26 AM, Nick Coghlan <report@bugs.python.org>wrote:
>
> Nick Coghlan added the comment:
>
> Right, the minimum change needed is to prefix it with an underscore, but if
> it isn't actually needed for anything, we may as well remove it entirely.
>
Just to be clear - do we consider all non-prefixed-with-underscore
attributes of classes in the stdlib part of the public API, even when
they're explicitly undocumented? And while we're at it, all
non-prefixed-with-underscore methods as well?
I still consider this refactoring gratuitious at this point. The API is
well defined by the documentation. All the rest is implementation details. |
|