Description
We are closing in on 6.0 (I think). There are 35 open issues/PR targeted 6.0 and about 10 for 5.3. While usually we would stop working on 5.x after 6.0, 5.x is our LTS, so I don't want to block on it too much. Ie 5.3 can happen post 6.0 and we'll likely have a 5.4 and 5.5.
I'll use this as a meta issue to track the status of 6.0 release.
Completer API.
Notebook 5.0 is about to be released, so I don't see a point in refactoring Completion API to send extra info in metadata now. I think we can experiment with 5.1/Lab
Completer API is marked as unstable so we can still make changes afterward the old API has not been removed.
AFAICT, we have no more crashes with Jedi 0.10. I have pending PR to speed things up, but I doubt it's worth waiting for a 0.11 (we might want to bump to 0.10.1 though because of a messed up 0.10.tgz).
pip magic
Twitter is happy about it, but would like more. Do we ? Do we backport to 5.3 ?
Many issues can be bumped or are significant refactor.
I'm worried about the GeoJson display class. Which is relatively new. And whether is should be in core.
Activity
Carreau commentedon Feb 24, 2017
cc @JamiesHQ , @willingc I'd like to have 6.0 out likely by mid March , at least be mid April to have it settled for PyCon.
Carreau commentedon Feb 28, 2017
Oh, a Big thing to be careful, this release if Python 3 only.
JamiesHQ commentedon Mar 1, 2017
I've created a card on the release calendar: https://github.com/orgs/jupyter/projects/4
jtratner commentedon Apr 11, 2017
Just wanted to give you a heads up that the 6.0.0rc1 release is having a weird interaction with setuptools when unversioned
ipythonis a requirement - pypa/setuptools#1009 (even though pip still seems to do the right thing so far). I don't think this is anything you folks have control over, but I figure quite a few people will start hitting this soon.takluyver commentedon Apr 11, 2017
I guess setuptools' dependency finding thing doesn't know about the requires-python metadata which handles avoiding the version for pip. As suggested on that issue, does it work if you try
pip install -e .instead ofpython setup.py develop?I guess the most practical way forwards is going to be to patch projects that require IPython to do something like this:
jtratner commentedon Apr 11, 2017
takluyver commentedon Apr 11, 2017
Thanks :-)
Carreau commentedon Apr 19, 2017
I'm going to start the release process of 6.0.
mpacer commentedon Apr 19, 2017
Ok, @Carreau has some issues with the github api but now we're back on track for the release schedule.
note to all others: if you ever are getting 401's even from https://api.github.com, check your
.netrcbecauserequestswill look there and automatically append authentication headers to your requests & they might be out of date.Carreau commentedon Apr 19, 2017
Strangely @takluyver appears to be the one uploading zip files to GitHub :-)
https://github.com/ipython/ipython/releases/tag/6.0.0
Carreau commentedon Apr 19, 2017
Released. Drafting the announcements.
takluyver commentedon Apr 19, 2017
Weird, I don't know how it got that.
takluyver commentedon Apr 19, 2017
👍 thanks for doing the release
Carreau commentedon Apr 19, 2017
Can SO update RTD and mark 6.0 as stable, I lost my PW and the reset link seem to not want to make its way my inbox.
Hope your creds are not somewhere in the source tree.
takluyver commentedon Apr 19, 2017
I hope so too!
I've updated the 'stable' version on RTD to point to 6.0.0.
Carreau commentedon Apr 19, 2017
Cool thanks.
Carreau commentedon Apr 19, 2017
TODO:
willingc commentedon Apr 19, 2017
Nice! ☀️