[Python-Dev] PEP 451 (ModuleSpec) is accepted
ncoghlan at gmail.com
Fri Nov 8 22:57:59 CET 2013
On 9 Nov 2013 06:39, "Brett Cannon" <brett at python.org> wrote:
> I just want to publicly thank Eric for his hard work on this PEP. I think
Nick, Eric, and myself all came to the same conclusion independently about
needing to update the APIs used by import to make them more flexible and
able to grow in the future, but Eric is the one that took the time to
champion the changes, write the PEP, and handle the huge amount of feedback
both on the import SIG and here along with his current patch.
> I know I'm quite pleased with the way this PEP turned out and am already
looking forward to using the new APIs hopefully in Python 3.5 for
introducing a lazy loader mechanism to the stdlib.
We should even be able to finally make circular imports work more
consistently, better handle partial failures of packages that implicitly
import submodules and address some of the import traps described in PEP 395.
This change also enables better integration of runpy with the normal import
system, the opportunity for much cleaner reload support (it's currently
rife with opportunities for implicit, hard to debug, failures) and the
possibility of bringing C extension loading up to a similar level of
capability to Python module execution (e.g. allowing extension modules to
support execution via runpy)
Many of those related changes won't happen until Python 3.5, but PEP 451 is
an important step towards making them practical.
Awesome work, Eric - this is a far more significant upgrade to the import
system than my indirect import idea that originally spawned the discussion
on import-sig :)
> On Fri, Nov 8, 2013 at 2:49 PM, Eric Snow <ericsnowcurrently at gmail.com>
>> I'm pleased to announce that Brett Cannon and Nick Coghlan (the
>> co-BDFL-delegates) have accepted PEP 451 for inclusion in Python 3.4.
>> Both of them have contributed substantially to the discussions of the
>> proposal and have helped make it solid. I'm grateful for their
>> diligence, attentiveness, and expertise. I'd also like to thank PJ
>> Eby, who provided important insight on several occasions and helped
>> out with some sticky backward-compatibility situations.
>> Python-Dev mailing list
>> Python-Dev at python.org
> Python-Dev mailing list
> Python-Dev at python.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Python-Dev