Skip to content

Conversation

@pp-mo
Copy link
Member

@pp-mo pp-mo commented Sep 15, 2021

🚀 Pull Request

Description

Final changes prior to tagging v3.1

A docs-build is available here


Consult Iris pull request check list

@bjlittle bjlittle self-assigned this Sep 15, 2021
@bjlittle bjlittle self-requested a review September 15, 2021 12:42
* Fix broken ABF link

* add cartopy downloader utility

* update netcdf4-python link

* explicitly specify features to download for ci

* added whatsnew entry

* update utility help

* utility tidy

* tidy utility

* copy with cartopy script rename

* update whatsnew to clarify as dev tool

* push fix to cartopy

* use cartopy master
@pp-mo pp-mo changed the title Finalise whatsnew and version string. Finalise whatsnew and version string; update CI to fix. Sep 15, 2021
Copy link
Member

@rcomer rcomer left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good to me. I particularly like the links from the highlights to see more details further down.

My only question is whether the "Deprecations" section should be removed, as there is nothing in it.

Copy link
Member

@bjlittle bjlittle left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@pp-mo A few comments to service. Happy for you to push back on them.

The biggest issue I have at the moment is that the final merge back from the v3.0.x release feature branch has not been included in the v3.1.x release branch.

This is a deal breaker for me. We need include those v3.0.x patch fixes into v3.1.x before we release.

@pp-mo
Copy link
Member Author

pp-mo commented Sep 15, 2021

the final merge back from the v3.0.x release feature branch has not been included in the v3.1.x release branch.

Oh, I hadn't spotted that at all.
That seems a really serious problem, and should have been in the RC

So
Can we add a mergeback now without releasing another RC ?

Personally, I have the impression that no-one has paid much attention to this one (i.e. the 'rc0'), and I'm inclined to just merge in 3.0.x and tag that as 3.1.0 straight away.
The risks should be small ?

@pp-mo
Copy link
Member Author

pp-mo commented Sep 15, 2021

Made changes addressing all the text points (I hope).
@bjlittle please re-review.

Still awaiting: mergeback from 3.0.x

@rcomer
Copy link
Member

rcomer commented Sep 15, 2021

The biggest issue I have at the moment is that the final merge back from the v3.0.x release feature branch has not been included in the v3.1.x release branch.

I thought this was done. The last commit on v3.0.x was in July, and there is a mergeback commit from 10th August in v3.1.x. Or have I got the wrong end of the stick entirely?

FB82DC1E-611F-4E2D-8B6A-9445D93541A3

@bjlittle
Copy link
Member

bjlittle commented Sep 15, 2021

@pp-mo Okay, I finally understand what's happened here... and why I'm confused.

Why did #4285 change the whatsnew history by collapsing all the v3.0.1, v3.0.2, v3.0.3 and v3.0.4 releases into the whatsnew for v3.0.0 ?

Those patches are separate releases and not part of the v3.0.0 release.

This will now be confusing post-v3.1.0, when comparing documentation with v3.0.x releases.

Specifically, our GH Discussions...
image

... and all our patch release annotations for v3.0.x...
image

... link to resources that have moved and look different in v3.1.0. So when viewing the v3.1.x docs you've now lost the convenience of easily seeing the self-contained patch changes for v3.0.x and instead force users to search.

Also, from a release managers perspective we're also adding an extra step to the release process i.e., collapse the previous release history and blend it together, all of which requires to be reviewed and checked.

Personally, I think we should break from the previous pattern of blending release history together. That was the advantage of leveraging sphinx-panels throughout v3.0.x for patch clarity.

However, there is perhaps a compromise here...

For example, for the v3.0.x release, all release and patch details go into the whatsnew/3.0.rst, but we don't blend - rather maintain the sphinx-panels separation. A release and it's patches is a purely additive process, and there is no extra work to be done for the following release manager, e.g., when v3.1.x rolls around. Plus there is no duplication as previously with 3.0.rst content in 3.0.1.rst et al

@bjlittle bjlittle mentioned this pull request Sep 16, 2021
@pp-mo
Copy link
Member Author

pp-mo commented Sep 16, 2021

Why did #4285 change the whatsnew history by collapsing all the v3.0.1, v3.0.2, v3.0.3 and v3.0.4 releases into the whatsnew for v3.0.0 ?

As explained there, essentially because we that is what we did to the 2.2.1 changes when making 2.3
- we put the bug fixes into the 2.2 release notes : see #3263 (comment)
Which after some thought, made good sense to me

@bjlittle bjlittle merged commit b7ed8fd into SciTools:v3.1.x Sep 16, 2021
@pp-mo
Copy link
Member Author

pp-mo commented Sep 16, 2021

Thanks @bjlittle .
Great stuff, I'll now attend to #4320 which I think is really valuable in this context

@bjlittle
Copy link
Member

@pp-mo Thanks for hanging-out so that we would align understanding - it was hard to easily clarify in prose 😄

@lbdreyer lbdreyer mentioned this pull request Jan 26, 2022
16 tasks
@pp-mo pp-mo deleted the final_3v1 branch April 28, 2023 13:45
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants