`constraints` matrix changes after running glmnetcv

[bonStats]

Is this behavior intended?

MWE:

X = rand(1000, 2)
y = X[:,1] + randn(1000)
constraints = [-1 -1; Inf Inf]
res = glmnetcv(X, y, constraints = constraints)

julia> constraints
2×2 Matrix{Float64}:
 -4.69035e-7  -4.57555e-7
 Inf          Inf