v3.0.x

[bjlittle]

The release manager for v3.0.x is @bjlittle

The release feature branch is v3.0.x

Major Release	Status	Milestone	Project	Date	RTD	Release	conda-forge	PyPI
v3.0.0rc0	Completed	v3.0.0	Iris v3.0.0	02/10/2020	Whatsnew	v3.0.0rc0	Available	-
v3.0.0	Completed	v3.0.0	Iris v3.0.0	25/01/2021	Whatsnew	v3.0.0	Available	Available

Patch Release	Status	Milestone	Project	Date	RTD	Release	conda-forage	PyPI
v3.0.1	Completed	-	Iris v3.0.x	27/01/2021	Whatsnew	v3.0.1	Available	Available
v3.0.2	Completed	-	Iris v3.0.x	27/05/2021	Whatsnew	v3.0.2	Available	Available
v3.0.3	Completed	-	Iris v3.0.x	07/07/2021	Whatsnew	v3.0.3	Available	Available
v3.0.4	Completed	-	Iris v3.0.x	22/07/2021	Whatsnew	v3.0.4	Available	Available

[jamesp]

This week looks like a good week for a push on getting the iris 3.0.2 release cut, @bjlittle let us know what you need from the rest of us @SciTools/iris-devs to do so 👍

[bjlittle]

@SciTools/iris-devs I'm looking for devs to review and merge the following pull-requests from @rcomer:

• Contourf workaround for Cartopy 0.19 #4127, and
• Intersection bounds fix (replacement PR) #4059

into the v3.0.x release feature branch as part of the pending iris v3.0.2 release.

Anyone with spare capacity willing to take these on? 🙏

[rcomer]

I think we first need someone to merge all of master's recent getting-the-tests-running commits into v3.0.x. BTW I have now incorporated #4127 into #4128, which targets master, as that seemed the easiest way to demonstrate that the fix actually works and the tests pass.

[bjlittle]

@rcomer Cool, thanks for the update

We may also need to address cirrus-ci on the v3.0.x release feature branch... but we'll cross that bridge when come to it.

[trexfeathers]

I merged #4128 before reading my emails 😂 So that half is done.

Need to focus on some other work this afternoon but if no one else has picked this up by tomorrow I'll try and get my head round #4059.

[rcomer]

@jamesp already had a look at #4059 and I think the main blocker is the CI issues.

@trexfeathers you need a break from reviewing!

[jamesp]

Yep I'm happy with #4059 to go in now without the tests passing, or wait if we are going to cherry pick the CI changes into the 3.0.x branch.

[rcomer]

I started making a branch to pull in the recent changes to get tests working with new dependencies. The good news is the gallery tests pass locally with latest matplotlib & cartopy. The bad news is I'm seeing about a gazillion cml failures in the main tests. It looks like all the attributes are the same, but in a different order e.g.

-        <dimCoord bounds="[[0.0, 42077.5]]" id="fb105c19" long_name="level_height" points="[21038.8]" shape="(1,)" units="Unit('m')" value_type="float32">
+        <dimCoord id="fb105c19" long_name="level_height" units="Unit('m')" value_type="float32" shape="(1,)" points="[21038.8]" bounds="[[0.0, 42077.5]]">

does this ring any bells for anyone?

[rcomer]

@jamesp there is documentation for the CLA checker here: https://github.com/SciTools/scitools.org.uk/wiki/Contributors-License-Agreement-processing-details

I'm not claiming to have read or understood it!

[jamesp]

Thanks @rcomer I was just going down that rabbit hole. Oh my, heroku and webhooks...

[jamesp]

SciTools-incubator/scitools-cla-checker#22

[bjlittle]

CLA checker is objecting to the github-actions bot. This is odd, because #4125 went through OK.

@rcomer I think I've got the privileges to fix this... I recently manually injected the necessary metadata to get the pre-commit.ci bot to pass the CLA.

Hang tight and I'll take a peek...

[rcomer]

@bjlittle thanks, but it didn’t prevent the PR getting merged. So not urgent as far as I’m aware.

[rcomer]

Can I interest anyone in reviewing some link repairs? #4132

[bjlittle]

Sure 😄

[rcomer]

Does this still need merging back into master?

[bjlittle]

Yup, thanks... that's on me, no worries 👍

[bjlittle]

@rcomer Thanks for the nudge 😄

[bjlittle]

We're hoping to unpin cftime and make this available to the community asap as iris 3.0.3, rather than keeping this until an iris 3.1 release (which is currently scheduled for Sept '21)

There is package dependency trouble brewing between iris, cftime, nc-time-axis and cf-units... so the sooner we address this, then the better 👍

[zklaus]

I realize this is a lot of uninteresting work. Thanks for doing it! It is really important!

[bjlittle]

@zklaus We're glad to help 😄

Are you guys scheduling an esmvaltool 2.2.1 patch release? If so, any idea when? That might help us prioritise at this end 👍

[zklaus]

Good question. I don't think we had planned for that, but I'll bring it up.

[bouweandela]

We're not scheduling an esmvaltool 2.2.1 patch release, but @zklaus is in the process of releasing esmvaltool 2.3 (planned this week?)

[bjlittle]

@zklaus and @bouweandela I should be able to get 3.0.3 out this week, but it also relies on a cf-units 3.0.1 patch release to support it, which is also in progress 👍

HTH

[pp-mo]

Stacked up #4230 for inclusion in v3.0.4
but not for 3.0.3, as we don't want any further risk or delay with that

[bjlittle]

iris 3.0.4 is now available. Whoop!

Thank you so much to everyone who contributed to this patch release.

A fantastic team effort! Awesome job 🎉 🙏 👏

[jamesp]

Great work everyone, great to see so much momentum on iris

[bjlittle]

I still need to merge back v3.0.x to main... this will include the stacked changes from 3.0.2, 3.0.3 and 3.0.4 😊

[bjlittle]

Post 3.0.4 release I noticed that PyPI is reporting the release as failing

This is due to the badge still pointing at master instead of main (BTW release badges have always pointed to master/main instead of the feature release branch)

The tests and release are actually all good, and I propose that we don't fix this in a specific post patch release.

Instead, we can wait for iris 3.1rc0 in August and 3.1.0 in September to clobber this. I'm totally okay with that.

[pp-mo]

Archiving = recategorise as "archived"
WAS category = "Releases" 🚩