Re: Replacement PEP for namespace packages
Brett Cannon <[email protected]>
| Newsgroups | gmane.comp.python.import |
|---|---|
| Message-ID | <CAP1=2W7xYrGvyyosVzkx3sJZ-W2_OzL6hYf_V=pOO5Z_RHj-uA@mail.gmail.com> |
On Thu, Apr 12, 2012 at 08:04, Eric V. Smith <[email protected]> wrote: > On 04/09/2012 07:22 PM, Yury Selivanov wrote: > > > How is it going with the PEP? > > I'm reviewing the email threads from the last round of discussions, > making sure I have everything covered. I hope to have a draft early next > week. > > > (The reason I'm asking is because we want to backport the accepted > > approach to 3.2 and start using it...) > > Our plan is to wait until Brett merges importlib as the default import > mechanism, so that we can work entirely (or at least mostly) in Python. > That should be happening soon. Keep an eye on python-dev for an email about it once the patch lands (have one more thing to work out in default and then use in my branch before it can land). > Any port to 3.2 is going to be an entirely separate effort, all in C. > But I support it! > I don't think it necessarily has to be that way. If you either use your own importer or override builtins.__import__ with importlib.__import__ and tweak some things under the hood you can actually match behaviour at a roughly 5% import performance cost compared to what 3.3 will have for import. -Brett > > Eric. > > _______________________________________________ > Import-SIG mailing list > [email protected] > http://mail.python.org/mailman/listinfo/import-sig > _______________________________________________ Import-SIG mailing list [email protected] http://mail.python.org/mailman/listinfo/import-sig