Error with Sum of Log

[RoyiAvital]

I get error trying to solve:

using Convex;
using ECOS;

function CVXSolver( mK :: AbstractMatrix{T} )  where {T <: AbstractFloat}
    
    numRows = size(mK, 1);
    
    vX = Variable(numRows);
    sConvProb = minimize( sum(-log(one(T) .+ mK' * vX)), [norm(vX) <= one(T)] );
    solve!(sConvProb, ECOS.Optimizer; silent = true);
    return vec(vX.value);

end

mK = randn(numRows, numCols);
mK = mK ./ reshape(norm.(eachcol(mK)), (1, numCols));

vX = CVXSolver(mK);

I get ERROR: MethodError: no method matching iterate(::Convex.MultiplyAtom).

[ericphanson]

I’m not at a computer but try omitting the . in .+

[RoyiAvital]

@ericphanson , Indeed.
I always forget not to use broadcasting in the context of Convex.jl.
Maybe it should be documented in the Expressions and Objective sections as it is documented in the Constraints section.

I don't mind do a PR for that.
Just need your confirmation it should indeed be avoided.

[ericphanson]

Yes it should be avoided, documentation improvements welcome, ty!

[RoyiAvital]

@ericphanson , See #717.