Move LFRic extraction tests to Python 3.14

[nmnobre]

No description provided.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.90%. Comparing base (c9c20b1) to head (2553b2c).

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #3192   +/-   ##
=======================================
  Coverage   99.90%   99.90%           
=======================================
  Files         376      376           
  Lines       53158    53158           
=======================================
  Hits        53106    53106           
  Misses         52       52           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[nmnobre]

This PR bumps Python to 3.14 for the LFRic extraction tests. These tests use fab, which calls fparser within spawned subprocesses. Python 3.14 does include changes to its concurrency model but I thought, reading this, that it wouldn't affect the functionality used in fab, namely import multiprocessing - well, I was clearly wrong... So it's up to people what to do with this; the patch here demonstrates it'd be trivial to patch fab to avoid concurrency in the identified places when we detect 3.14, or else perhaps @hiker can provide further insight as to why this is a problem. I'll leave this as a draft for somebody else to pick up.

[hiker]

I think we want to run the extraction in parallel - extraction makes python quite a bit slower (static call tree analysis). We have a ticket in Fab (MetOffice/fab#510) to update fab (and I wanted to update the extraction to the official 2.0 release as well).
Is there nay urgency in this?

[nmnobre]

I think we want to run the extraction in parallel - extraction makes python quite a bit slower (static call tree analysis). We have a ticket in Fab (MetOffice/fab#510) to update fab (and I wanted to update the extraction to the official 2.0 release as well). Is there nay urgency in this?

None. Mat put it well in MetOffice/fab#510: not needed right now, but will inevitably need addressing, that's all. :)

[hiker]

Then I'll prefer to let us look at Fab first, and then update the build to the updated Fab :)