Remove LazyArrays.jl dependency

- Removed LazyArrays from Project.toml dependencies
- Replaced BroadcastArray(@~ .-(λ .* tu)) with simple broadcast .-(λ .* tu) in adjoint.jl
- Updated tests to check for AbstractMatrix instead of BroadcastArray
- All tests pass successfully

Author: claude

Project.toml

@@ -12,7 +12,6 @@ EnumX = "4e289a0a-7415-4d19-859d-a7e5c4648b56"
 GPUArraysCore = "46192b85-c4d5-4398-a991-12ede77f4527"
 InteractiveUtils = "b77e0a4c-d291-57a0-90e8-8db25a27a240"
 Krylov = "ba0b0d4f-ebba-5204-a429-3ac8c609bfb7"
-LazyArrays = "5078a376-72f3-5289-bfd5-ec5146d43c02"
 Libdl = "8f399da3-3557-5675-b5ff-fb832c97cbdb"
 LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
 MKL_jll = "856f044c-d86e-5d09-b602-aeab76dc8ba7"
@@ -105,7 +104,6 @@ Krylov = "0.10"
 KrylovKit = "0.10"
 KrylovPreconditioners = "0.3"
 LAPACK_jll = "3"
-LazyArrays = "2.3"
 Libdl = "1.10"
 LinearAlgebra = "1.10"
 MPI = "0.20"

src/LinearSolve.jl

@@ -15,7 +15,6 @@ using LinearAlgebra: LinearAlgebra, BlasInt, LU, Adjoint, BLAS, Bidiagonal, Bunc
                      cholesky, cholesky!, diagind, dot, inv, ldiv!, ldlt!, lu, lu!, mul!,
                      norm,
                      qr, qr!, svd, svd!
-using LazyArrays: @~, BroadcastArray
 using SciMLBase: SciMLBase, LinearAliasSpecifier, AbstractSciMLOperator,
                  init, solve!, reinit!, solve, ReturnCode, LinearProblem
 using SciMLOperators: SciMLOperators, AbstractSciMLOperator, IdentityOperator,

src/adjoint.jl

@@ -84,7 +84,7 @@ function CRC.rrule(::typeof(SciMLBase.solve), prob::LinearProblem,
         end
 
         tu = adjoint(sol.u)
-        ∂A = BroadcastArray(@~ .-(λ .* tu))
+        ∂A = .-(λ .* tu)
         ∂b = λ
         ∂prob = LinearProblem(∂A, ∂b, ∂∅)
 

test/Project.toml

@@ -0,0 +1,5 @@
+[deps]
+FiniteDiff = "6a86dc24-6348-571c-b903-95158fe2bd41"
+ForwardDiff = "f6369f11-7733-5829-9624-2563aa707210"
+RecursiveFactorization = "f2c3362d-daeb-58d1-803e-2bc74f2840b4"
+Zygote = "e88e6eb3-aa80-5325-afca-941959d7151f"

test/adjoint.jl

@@ -1,7 +1,6 @@
 using Zygote, ForwardDiff
 using LinearSolve, LinearAlgebra, Test
 using FiniteDiff, RecursiveFactorization
-using LazyArrays: BroadcastArray
 
 n = 4
 A = rand(n, n);
@@ -19,7 +18,7 @@ end
 f(A, b1) # Uses BLAS
 
 dA, db1 = Zygote.gradient(f, A, b1)
-@test dA isa BroadcastArray
+@test dA isa AbstractMatrix
 
 dA2 = ForwardDiff.gradient(x -> f(x, eltype(x).(b1)), copy(A))
 db12 = ForwardDiff.gradient(x -> f(eltype(x).(A), x), copy(b1))
@@ -37,7 +36,7 @@ _ff = (x,
 _ff(copy(A), copy(b1))
 
 dA, db1 = Zygote.gradient(_ff, copy(A), copy(b1))
-@test dA isa BroadcastArray
+@test dA isa AbstractMatrix
 
 dA2 = ForwardDiff.gradient(x -> f(x, eltype(x).(b1)), copy(A))
 db12 = ForwardDiff.gradient(x -> f(eltype(x).(A), x), copy(b1))
@@ -58,7 +57,7 @@ function f3(A, b1, b2; alg = KrylovJL_GMRES())
 end
 
 dA, db1, db2 = Zygote.gradient(f3, A, b1, b1)
-@test dA isa BroadcastArray
+@test dA isa AbstractMatrix
 
 dA2 = FiniteDiff.finite_difference_gradient(
     x -> f3(x, eltype(x).(b1), eltype(x).(b1)), copy(A))
@@ -83,7 +82,7 @@ function f4(A, b1, b2; alg = LUFactorization())
 end
 
 dA, db1, db2 = Zygote.gradient(f4, A, b1, b1)
-@test dA isa BroadcastArray
+@test dA isa AbstractMatrix
 
 dA2 = ForwardDiff.gradient(x -> f4(x, eltype(x).(b1), eltype(x).(b1)), copy(A))
 db12 = ForwardDiff.gradient(x -> f4(eltype(x).(A), x, eltype(x).(b1)), copy(b1))