Don't forget to do this in the final version

[ocots]

CTModels.jl/ext/CTModelsMadNLP.jl

Lines 49 to 55 in 5b83947

	function CTModels.extract_solver_infos(
	nlp_solution::MadNLP.MadNLPExecutionStats,
	minimize::Bool, # whether the problem is a minimization problem or not
	)
	# Get minimization flag and adjust objective sign accordingly
	# minimize = NLPModels.get_minimize(nlp)
	objective = minimize ? nlp_solution.objective : -nlp_solution.objective

Remark: actually, MadNKLP will stop to give the opposite sign in its next release I think.