Add a return type wrapper

[ChrisRackauckas]

No description provided.

[codecov]

Codecov Report

Merging #39 (c30a203) into main (d7b478b) will increase coverage by 0.04%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main      JuliaLang/julia#39      +/-   ##
==========================================
+ Coverage   85.31%   85.35%   +0.04%     
==========================================
  Files           6        6              
  Lines         320      321       +1     
==========================================
+ Hits          273      274       +1     
  Misses         47       47              

Impacted Files	Coverage Δ
src/common.jl	72.50% <100.00%> (ø)
src/factorization.jl	94.44% <100.00%> (+0.10%)	⬆️
src/pardiso.jl	83.33% <100.00%> (ø)
src/wrappers.jl	91.15% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d7b478b...c30a203. Read the comment docs.

[ChrisRackauckas]

@vpuri3 do you know why this would make v1.6 fail? 😅

[vpuri3]

it's svd ! i would see if the factorization algorithm overwrites the matrix A which would cause the test A*u \approx b to not be appropriate

[vpuri3]

so the in place solution is not the same as the one that is emitted by ldiv!(y fact, b) --> y in 1.6.4

julia> VERSION
v"1.6.4"

julia> using LinearAlgebra

julia> n=8; A=rand(n,n); u=rand(n); b=A*u; v=zero(u); F=svd(A); ldiv!(v,F,b) ≈ u, v ≈ u
(true, false)

julia> using LinearAlgebra

julia> VERSION
v"1.7.0"

julia> n=8; A=rand(n,n); u=rand(n); b=A*u; v=zero(u); F=svd(A); ldiv!(v,F,b) ≈ u, v ≈ u
(true, true)

[vpuri3]

is this expected behavior or should this be an issue in LinearAlgebra?

[ChrisRackauckas]

It looks like it was a bug that was fixed in v1.7 but should probably be backported to LTS.

[vpuri3]

ok created an issue in julialang JuliaLang/LinearAlgebra.jl#895