Fixed type instability in tensor operations

[JordiManyer]

The problem:

Every arithmetic operation on MultiValue types (scalar *, +, -, /) followed the same pattern: extract the independent components as a tuple, apply the operation element-wise, reconstruct
the result. The extraction was written as Tuple(a)[1:Li] where Li = num_indep_components(a) is a plain Int computed at runtime.

The problem is that slicing a tuple with a runtime integer is type-unstable in Julia: the compiler sees NTuple{N,T}[1:Int] and cannot determine the length of the result statically. Everything downstream — the element type, the result type, the constructor call — then cascades into uncertainty. Thus, Base.promote_op cannot infer the output type and returns Any.

The reason we slice instead of just taking the whole data is that some types (only SymTracelessTensorValue in reality) store extra components that are not independent. So the [1:Li] was load-bearing for that type specifically, while being a unnecessary for all others.

The fix:

We introduced get_indep_components(a), which replaces the type-unstable slice: It is equivalent to Tuple for all MultiValues, except for the special ones.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.83%. Comparing base (f39eac1) to head (277fe82).
⚠️ Report is 3 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1293      +/-   ##
==========================================
- Coverage   88.83%   88.83%   -0.01%     
==========================================
  Files         227      227              
  Lines       29775    29772       -3     
==========================================
- Hits        26451    26448       -3     
  Misses       3324     3324              

Flag	Coverage Δ
drivers	39.95% <100.00%> (-0.01%)	⬇️
extensions	5.12% <57.14%> (-0.01%)	⬇️
unit-adaptivity	40.09% <85.71%> (-0.02%)	⬇️
unit-basics	14.68% <100.00%> (-0.01%)	⬇️
unit-celldata	21.12% <85.71%> (-0.02%)	⬇️
unit-fespaces-1	32.35% <85.71%> (-0.02%)	⬇️
unit-fespaces-2	38.83% <85.71%> (-0.01%)	⬇️
unit-fields	17.57% <100.00%> (-0.02%)	⬇️
unit-geometry	28.77% <71.42%> (-0.02%)	⬇️
unit-multifield	30.87% <85.71%> (-0.02%)	⬇️
unit-odes	28.74% <71.42%> (-0.02%)	⬇️
unit-referencefes	34.24% <100.00%> (-0.02%)	⬇️
unit-visualization	11.90% <71.42%> (-0.02%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ 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.