I just thought I would quickly update everyone on where things stand in terms of fixing our development process to be easier for us to keep up with patches.
First, I have given Donald Stufft and Nick Coghlan a deadline of October 31 -- Halloween in North America -- to get demo instances of their approaches for tweaking our approaches (and they were the only ones to propose anything so if you are asking "what about considering XXX?", it's because no one proposed it and stepped up to make sure it happened if their approach was chosen). Nick is proposing using Kalithea and his proposal is found at https://www.python.org/dev/peps/pep-0474/ . Donald is proposing GitHub + Phabricator -- although the Phabricator bit is optional for people -- and his proposal is at https://www.python.org/dev/peps/pep-0481/ . The plan is to get the test instances up, play with them to see what it looks like from both a core dev and contributor perspective and get opinions from various people of different levels of experience as to how the proposals feel. My hope is to make a decision by January 1 so we can switch over ancillary repositories like the devguide, devinabox, and the benchmark suite quickly and then work towards switching the cpython repository.
Second, if you find Misc/NEWS a pain when doing merges then have a look at
http://bugs.python.org/issue18967 and look at the recent posts. Larry
Hastings is proposing a wrapper around hg commit that will create
individual files per NEWS entry so that you can just pull the file forward
in a merge and thus have no issues. We probably need feedback from Windows
developers more than anyone since this would make using TortoiseHg a bigger
pain since you would need to generate the NEWS file separately before doing
your commit while everyone else has it integrated into the commit process.
The long term goal is to put a field in the issue tracker for the NEWS
entry so that it's directly attached to the issue related to the change and
makes it easier to update before a release.
So that's the current status. As always, if you want to participate in any of this you should join the core-workflow@ mailing list.
On Sep 29, 2015, at 09:36 PM, Brett Cannon wrote:
First, I have given Donald Stufft and Nick Coghlan a deadline of October 31 -- Halloween in North America -- to get demo instances of their approaches for tweaking our approaches (and they were the only ones to propose anything so if you are asking "what about considering XXX?", it's because no one proposed it and stepped up to make sure it happened if their approach was chosen). Nick is proposing using Kalithea and his proposal is found at https://www.python.org/dev/peps/pep-0474/ . Donald is proposing GitHub + Phabricator -- although the Phabricator bit is optional for people
I really appreciate all the work you guys are doing to improve our process.
I am going to ask the question anyway Brett, and depending on the answer, I might be willing to write a PEP and work on putting up a demo instance, although with my limited time, I'd sure love to work with someone on it.
Recently, the CEO of GitLab made an offer to the Debian community:
http://article.gmane.org/gmane.linux.debian.devel.general/201695
GitLab (community edition at least) is open source, so it's way more palatable to me than GitHub. And if they were willing to extend the same offer to us, then we would at least have their assistance for hosting. I'm happy to reach out to Sytse and ask - can't hurt, right?!
I have no connection to GitLab other than as a happy user.
Cheers, -Barry
On Tue, 29 Sep 2015 at 15:04 Barry Warsaw barry@python.org wrote:
On Sep 29, 2015, at 09:36 PM, Brett Cannon wrote:
First, I have given Donald Stufft and Nick Coghlan a deadline of October
31
-- Halloween in North America -- to get demo instances of their approaches for tweaking our approaches (and they were the only ones to propose anything so if you are asking "what about considering XXX?", it's because no one proposed it and stepped up to make sure it happened if their approach was chosen). Nick is proposing using Kalithea and his proposal is found at https://www.python.org/dev/peps/pep-0474/ . Donald is proposing GitHub + Phabricator -- although the Phabricator bit is optional for
people
I really appreciate all the work you guys are doing to improve our process.
I am going to ask the question anyway Brett, and depending on the answer, I might be willing to write a PEP and work on putting up a demo instance, although with my limited time, I'd sure love to work with someone on it.
Recently, the CEO of GitLab made an offer to the Debian community:
http://article.gmane.org/gmane.linux.debian.devel.general/201695
GitLab (community edition at least) is open source, so it's way more palatable to me than GitHub. And if they were willing to extend the same offer to us, then we would at least have their assistance for hosting. I'm happy to reach out to Sytse and ask - can't hurt, right?!
You have clearance from me to sneak in a GitLab proposal then if you think you can make the deadline. I leave it up to you to figure out how much you want to involve GitLab (heck, you can have them do all the work for all I care since I'm trying to be pragmatic about this so I don't care if it's fully hosted somewhere else, etc.). Just remember the over-reaching goal is to make it so we can all review and merge patches in the majority of cases from a tablet on a beach (IOW I can do a full code review and acceptance on my lunch break at work without a clone w/ SSH keys).
I have no connection to GitLab other than as a happy user.
Honestly, even if you did I wouldn't care. Key thing is that I don't have any connection to any of the proposed companies or projects.
-
Barry Warsaw -
Brett Cannon -
Brett Cannon