Release 5.1

[gnestor]

• Close outstanding issues marked with 5.1 milestone
• Merge open PRs marked with 5.1 milestone
  • Add nbextension/serverextension commands for uninstall/disable all #2149
• Add 5.1 to changelog
• Tag release candidate, publish to PyPI, and send a message to the Google Group
• Release rc1 to PyPI
• Send out message to Google Group
• Release 5.1
  • Upload to PyPI
  • Create release on Github: https://github.com/jupyter/notebook/releases/tag/5.1.0
  • Send PR to conda-forge
    • 5.1.0 conda-forge/notebook-feedstock#18
    • Release 5.1 & re-render with conda-smithy conda-forge/notebook-feedstock#19
  • Check readthedocs branches

[takluyver]

This is debatable, but I'd say we should avoid copying static lists of issue numbers from a milestone, because we often change which issues and PRs are in a milestone.

[gnestor]

@takluyver Agreed, but I want to publish a release candidate today, so if there is anything else that we want to include in this release (like #1968), let's make sure it's marked with 5.1 and let's try to get it merged.

[takluyver]

@gnestor I think we're all set for an RC today in your time zone. :-)

[gnestor]

I just cut a release and I'm testing now: it looks like all my notebooks are not trusted and kernels are not starting! 😮

[takluyver]

Might be a good job you tested! I'm about to go to bed - hopefully you'll get it sorted while I'm asleep, but if not, post whatever you've found and I'll carry on investigating tomorrow.

[gnestor]

Ok, will do!

[gnestor]

@takluyver We still have #2747 which remains mysterious because we can't reproduce it, it only affects specific notebooks, and it goes away after refreshing the notebook. Do you think we should hold off on releasing 5.1 until we figure this out? If so, I think we should release an rc2 today. If not, I think we should release 5.1 today so that it's available for JupyterCon next week.

[takluyver]

I think we need to work it out, or at least have some confidence that we're not making a full release with a serious problem. I'm fine with making an rc2, though - maybe the issue was a bizarre accident when creating the build and it will work properly next time...

[gnestor]

Ok, sounds good. I'll publish rc2 now...

[stonebig]

final release soon ?

[gnestor]

Notebook 5.1.0rc3 was published today: https://pypi.org/project/notebook/5.1.0rc3/

[gnestor]

@takluyver Unless #2752 requires a bit more testing in master or rc4, I believe that we are ready to officially release 5.1.0. What do you say?

[takluyver]

Oh, I forgot I'd just merged that ;-). Let's try to all use it a bit today to check that nothing's obviously amiss (calling @Carreau @mpacer !). Assuming no major problems come up, go for a release this afternoon your time so we can avoid the stereotypical Friday afternoon release-and-run.

[gnestor]

Excellent!! Is there anything we can do specifically to test #2752?

[takluyver]

With the notebook server running in a terminal, shut it down by one of these methods:

1. Ctrl-C, type y, enter
2. Ctrl-C, Ctrl-C
3. Send it SIGTERM using kill -TERM [pid] (replacing [pid] with the notebook process's pid)

I've tried 1 & 2 already. I don't think there's likely to be any problems, but it can't hurt to exercise the code in some different systems.

[Carreau]

Sorry I know I should try the Rc, but haven't. I trust you. I'll edit the top comment to add "send a PR to conda-forge", and "check the readthedocs branches"

[mpacer]

I've been running off of the up-to-date master and haven't run into any obvious bugs. That said, I also haven't been trying customise it extensively or use it in weird ways… so I can't make too many guarantees. That said customization and weird bugs would seem to be candidates for patch releases anyway. So my vote is: good to go!

[gnestor]

5.1.0 is published to PyPI! I will submit a PR to conda-forge next.

@Carreau What does "Check readthedocs branches" entail?

[Carreau]

Make sure that the "stable" branch on readthedocs is 5.1.0 Right now the default doc page is the docs of 5.0.0. this used to not be automatic. it seem to be now so as soon as the repository is tagged, it should work.

Basically head to readthedocs admin page, and check the settings; though you do not seem to have admin access, what's your readthedocs name, I can add you.

[Carreau]

5.1.0 is published to PyPI! I will submit a PR to conda-forge next.

Awesome !

[gnestor]

It's gnestor

[Carreau]

It's gnestor

Done.

[Carreau]

Strangely readthedocs does not see the 5.1.0 tag. It may just be a question of time before it sees it.

[gnestor]

Ya, I'm seeing the same thing...

[Carreau]

Did you tag after pushing the commit ? Or pushing and tagging at the same time ? It's not really important we can figure that out later.

[gnestor]

I pushed the commit then the tag

[Carreau]

5.1.0 appeared since yesterday, I've hidden 5.0.0.

Also there is master and latest, should we hide one of the two ?

[Carreau]

All checkmarks are complete, feel free to close the issue and make an announce.!
Congrats !

[gnestor]

Cool, I see that in readthedocs too! Yay!! Let's announce on the mailing list and do you think this release warrants a blog post?

[Carreau]

Cool, I see that in readthedocs too! Yay!! Let's announce on the mailing list and do you think this release warrants a blog post?

It's been a while, so yes, likely a blog post on medium, now that we're moving to medium.

[gnestor]

Ah, ok. I tried logging into Ghost but no dice. Can you invite me to edit the Medium blog?

[Carreau]

Can you invite me to edit the Medium blog

You don't need to, you just write a blog as your self, and submit it to https://medium.com/jupyter-blog Look at the dropbox notes from 2 Tuesday ago, Brian left some instructions.

[gnestor]

Blog post is published: https://medium.com/jupyter-blog/jupyter-notebook-5-1-0-ddc8b0d6e989

[Carreau]

And now I have the 🎶music🎶 in my ears.