Improve inference in axpy for non-inferred operator

[jishnub]

The following is type-inferred after this:

julia> d1, r1 = Legendre(), Jacobi(2,2);

julia> d2, r2 = Chebyshev(), Chebyshev();

julia> K = (Operator(I, d1) ⊗ Operator(I, d2)) → (r1 ⊗ r2);

julia> @inferred BandedBlockBandedMatrix(view(K, Block(1):Block(2), Block(1):Block(2)))
2×2-blocked 3×3 BandedBlockBandedMatrix{Float64, BlockArrays.PseudoBlockMatrix{Float64, Matrix{Float64}, Tuple{BlockArrays.BlockedUnitRange{Vector{Int64}}, BlockArrays.BlockedUnitRange{Vector{Int64}}}}, BlockArrays.BlockedUnitRange{Vector{Int64}}}:
 1.0  │  0.0  0.0     
 ─────┼───────────────
  ⋅   │  1.0  0.0     
  ⋅   │   ⋅   0.333333

[codecov]

Codecov Report

Patch coverage: 69.84% and project coverage change: -44.46% ⚠️

Comparison is base (4f33475) 71.88% compared to head (1b0e6df) 27.42%.
Report is 13 commits behind head on master.

Additional details and impacted files

@@             Coverage Diff             @@
##           master     #522       +/-   ##
===========================================
- Coverage   71.88%   27.42%   -44.46%     
===========================================
  Files          81       80        -1     
  Lines        8546     8362      -184     
===========================================
- Hits         6143     2293     -3850     
- Misses       2403     6069     +3666     

Files Changed	Coverage Δ
src/ApproxFunBase.jl	44.00% <ø> (-40.00%)	⬇️
src/Operators/SubOperator.jl	42.64% <0.00%> (-41.18%)	⬇️
src/Operators/banded/CalculusOperator.jl	17.35% <0.00%> (-58.46%)	⬇️
src/Operators/banded/Conversion.jl	22.95% <0.00%> (-63.49%)	⬇️
src/Spaces/ProductSpaceOperators.jl	0.00% <0.00%> (-65.37%)	⬇️
src/eigen.jl	0.00% <0.00%> (-96.43%)	⬇️
src/Fun.jl	36.17% <50.00%> (-31.62%)	⬇️
src/Operators/general/InterlaceOperator.jl	23.46% <66.66%> (-57.87%)	⬇️
src/LinearAlgebra/helper.jl	55.58% <100.00%> (-13.35%)	⬇️
src/Multivariate/TensorSpace.jl	39.83% <100.00%> (-45.34%)	⬇️
... and 3 more

... and 59 files with indirect coverage changes

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