Why do you add a symmetric constraint rather than using var_symm?

[ubtjoe]

In solveSDP you set up the SDP solution as a var_sdp object, and then add a symmetric constraint. According to the SDPNAL+ docs, it looks like you could have used var_symm for the definition of Y, in which case you can remove the symmetric constraint here. Is there something that I am missing here?

Thanks for the implementation, and I enjoyed your talk at CVPR this year!