fix: multivariate polynomial rings over zero rings

[thofma]

No description provided.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 88.07%. Comparing base (c062cc1) to head (bef95db).
Report is 13 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1910      +/-   ##
==========================================
- Coverage   88.17%   88.07%   -0.11%     
==========================================
  Files         120      120              
  Lines       30300    30421     +121     
==========================================
+ Hits        26718    26792      +74     
- Misses       3582     3629      +47     

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

[thofma]

We are now getting the "expected" errors in Oscar, for which we need oscar-system/Oscar.jl#4350

[fingolfin]

Overall this looks fine to me (ignoring whether or not this causes any upstream breakage -- but if we put this into 0.44.0, it should be OK?

The caching of is_trivial/istrivial may be unnecessary now, but it also doesn't hurt, so I don't mind!

[thofma]

Yes, I think it is fine and we also fixed all the issues in Oscar that this revealed. Oscar master branch is passing and the release branch is failing (as expected).