Eigen does not work for Double64 matrices without GenericSchur

[jaredjeya]

I'm trying to diagonalise a Double64 matrix, but Julia throws an error when I do this:

No method matching eigen!(::Symmetric{Double64, Matrix{Double64}}; sortby::Nothing)

Yet bizarrely, various other functions such as lu and even eigvals work on this input.

Apparently the solution to this is to also import GenericSchur at runtime, but this is not documented anywhere in the library -- in fact, the documentation suggests it will work with just DoubleFloats. Can you please update the documentation?

[araujoms]

This is a bug in GenericLinearAlgebra: JuliaLinearAlgebra/GenericLinearAlgebra.jl#144

You need to wrap you matrix with Hermitian instead of Symmetric:

a = Hermitian(randn(Double64,3,3))
eigen(a)

[araujoms]

@JeffreySarnoff this has been fixed in version 0.3.15 of GenericLinearAlgebra, so you might want to close this issue.

[JeffreySarnoff]

Thank you!