News for hackers. Stuff that matters.
Re:Python 3.5.0 by baxter001 on Sunday September 13, 2015 @11:49AM (#10211748)
The scandir update also changes the underlying implementation of os.walk giving loads of production apps a huge speed increase by making use of additional data returned by the os calls.
Re:Python 3.5.0 by lamby on Sunday September 13, 2015 @02:32PM (#10212302)
Could os.walk have been changed "under the hood" without a PEP? eg. using — just for example — fast_walk_in_c in _os.so

(I understand that adding os.scandir might require a PEP)

Re:Python 3.5.0 by masklinn on Sunday September 13, 2015 @03:03PM (#10212426)
In theory, but because the change was possibly complex and could be useful if available separately, one of the first thing a core member recommended was to iterate the poc as a genuine package and provide an API to the underlying system (which is now scandir).

https://mail.python.org/pipermail/python-ideas/2012-November... is the original proposal on -ideas.

Re:Python 3.5.0 by haypo on Wednesday September 16, 2015 @10:53AM (#10227029)
No, a PEP was required because there was a lot of discussion on the API, and also non trivial choices on the implementation (pure C? pure Python? C+Python?). To have an idea of the amount of discussion, read the "Rejected ideas" of the PEP:

https://www.python.org/dev/peps/pep-0471/#rejected-ideas

Re:Python 3.5.0 by lamby on Thursday September 17, 2015 @01:33PM (#10234824)
Why would changing the underlying workings of os.walk require a PEP?