They latest JuMP release is 1.9.0 so you should probably use that.
JuMP v1.X is backwards compatible, so it should work with v1.6.0
@Rick, unfortunately I cannot reproduce this. Can you show the full output of these commands in your REPL?
julia> import Pkg
julia> Pkg.status()
Status `/private/tmp/kn/Project.toml`
[4076af6c] JuMP v1.6.0
[67920dd8] KNITRO v0.13.2
julia> using JuMP, KNITRO
julia> model = Model(KNITRO.Optimizer)
##### This license is only intended for use by JuMP-dev team. #####
##### License is valid until Aug 31, 2023 #####
A JuMP Model
Feasibility problem with:
Variables: 0
Model mode: AUTOMATIC
CachingOptimizer state: EMPTY_OPTIMIZER
Solver name: Knitro
Can you copy the code I wrote above in a new Julia REPL, and provide all output (exactly like I have done). I don’t know why the status has printed twice, or what the “Replacing module” appears.
Also, you shouldn’t need the master branch installed. Run
import Pkg
Pkg.free("KNITRO")
Then restart Julia.
Finally, what do you see if you do:
julia> import KNITRO
julia> KNITRO.Optimizer()
##### This license is only intended for use by JuMP-dev team. #####
##### License is valid until Aug 31, 2023 #####
A MathOptInterface model with backend:
Knitro 13.1.0
-----------------------
Problem Characteristics
-----------------------
Objective goal: Minimize
Objective type: -1
Number of variables: 0
Number of constraints: 0
Number of nonzeros in Jacobian: 0
Number of nonzeros in Hessian: 0