ConstraintIndex starts from 2

[WalterMadelim]

Are there any justification for labeling a very first ScalarAffineFunction constraint in an optimizer to be 2, rather than 1?

import MathOptInterface as MOI
import Gurobi
o = Gurobi.Optimizer()
x = MOI.add_variable(o)
f = MOI.ScalarAffineFunction([MOI.ScalarAffineTerm(1.,x)], 0.)
s = MOI.GreaterThan(1.)
# Notice the output: Can explain why this number starts from 2 here?
println(MOI.add_constraint(o,f,s)) # MathOptInterface.ConstraintIndex{...}(2)
# ... = MathOptInterface.ScalarAffineFunction{Float64}, MathOptInterface.GreaterThan{Float64}

[blegat]

The MOI implementations can use any constraint indices as long as they don't overlap. However, this might be unintended behavior here, see #523

[WalterMadelim]

Thanks. I'm now writing integer programming algorithms with MOI. It fulfills my need :)