About `issymmetric(REAL_SYM)`

[briochemc]

Could marking real structurally symmetric matrices (REAL_SYM) as symmetric cause any issues? Specifically this line:

Pardiso.jl/src/enums.jl

Line 15 in a645c66

LinearAlgebra.issymmetric(v::MatrixType) = v in (REAL_SYM, REAL_SYM_POSDEF, REAL_SYM_INDEF, COMPLEX_STRUCT_SYM,

where I understand REAL_SYM (type 1 in Intel docs) as "the sparsity pattern is symmetric", and thus a REAL_SYM matrix may not be symmetric.

I'm worried Pardiso.jl might ccall with the wrong parameters and thus run an incorrect branch. It's very likely I'm just confused and this is irrelevant, so please don't hesitate to close this issue if that's the case! 😅

[KristofferC]

I think the only place this is used is in

Pardiso.jl/src/Pardiso.jl

Lines 307 to 310 in a645c66

	if !issymmetric(mtype)
	T == :C && return conj(A)
	return A
	end

and I think that one is correct. But I agree it is confusing and there is no need to extend the LinearAlgebra method at all.