No more sequential tests since SegFaults were not noticed anymore (knock on wood)

[valeriupredoi]

Description

Let's see if indeed SegFaults that we used to see are now gone (bar, of course, the ones - many of them - from netCDF>1.6.0 + iris). Don't merge this just yet (hence draft, I'd like to run a few of these test sets).

---

Checklist

It is the responsibility of the author to make sure the pull request is ready to review. The icons indicate whether the item will be subject to the 🛠 Technical or 🧪 Scientific review.

• 🛠 This pull request has a descriptive title and labels
• 🛠 All checks below this pull request were successful

---

To help with the number pull requests:

• 🙏 We kindly ask you to review two other open pull requests in this repository

[codecov]

Codecov Report

Merging #1819 (eb0043a) into main (5644203) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##             main    #1819   +/-   ##
=======================================
  Coverage   91.50%   91.50%           
=======================================
  Files         202      202           
  Lines       10919    10919           
=======================================
  Hits         9991     9991           
  Misses        928      928           

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[valeriupredoi]

Test batches (Circle + GA: test and monitor): 7 since commit 66c8c99 - that's a lot of tests with no SegFault 🥳

[bouweandela]

You also need to remove the sequential marker from setup.cfg

[valeriupredoi]

You also need to remove the sequential marker from setup.cfg

indeed, my dear squid! Just did that 👍

[valeriupredoi]

I don't want to be a conspiracy theorist but I have noticed this ESMF-related test fail on Tool just a few hours ago too - @zklaus any idea where it may come from?

@bouweandela would you be able to give this one last looksee and approve, please, bud? Absolutely no SegFaults from a ton of test runs, we have somehow cured that 🥳

[bouweandela]

Thanks @valeriupredoi 🥇

[bouweandela]

I don't want to be a conspiracy theorist but I have noticed this ESMF-related test fail on Tool just a few hours ago too - @zklaus any idea where it may come from?

That looks like it could be this issue: numpy/numpy#22623

[valeriupredoi]

Brilliant, cheers muchly for the review @bouweandela 🍺 @ESMValGroup/technical-lead-development-team could one kind tech soul merge this please 🍻

[valeriupredoi]

I don't want to be a conspiracy theorist but I have noticed this ESMF-related test fail on Tool just a few hours ago too - @zklaus any idea where it may come from?

That looks like it could be this issue: numpy/numpy#22623

Hmmm - that does look like ours indeed, great find, man! But why did it happen only in a very restricted time frame and we don't see it anymore (am on the bus to work on me phone, so a bit hard to check setuptools versions, will do at work)

[valeriupredoi]

I don't want to be a conspiracy theorist but I have noticed this ESMF-related test fail on Tool just a few hours ago too - @zklaus any idea where it may come from?

That looks like it could be this issue: numpy/numpy#22623

Hmmm - that does look like ours indeed, great find, man! But why did it happen only in a very restricted time frame and we don't see it anymore (am on the bus to work on me phone, so a bit hard to check setuptools versions, will do at work)

OK I looked at setuptools versions - 65.6 is problematic - and it's the one causing our fails, the other tests that pass all got 65.5 (how the solver decided to jump to 65.6 for just a few Python/OS versions is beyond me, I blame it on magic). Also this comment on the issue you linked, Bouwe numpy/numpy#22623 (comment) is very worrisome - shall we pin setuptools for now?

[bouweandela]

If I understood that comment correctly, it looks like ESMF should pin setuptools because they are the ones using the numpy build system.

[bouweandela]

Anyway, let's take this discussion to an issue?

[valeriupredoi]

yes and yes 😁

[schlunma]

Thanks both 🚀

[valeriupredoi]

cheers Manu 🍺