Warning: "problem does not have general constraints; calling tron"

[ForceBru]

Yes indeed, my problem only has bound constraints sometimes and looks like this:

nlp = ADNLPModels.ADNLPModel(
  negative_log_likelihood, params0,
  [-Inf, 0, -1], [Inf, 1/2, 1],
)
percival(nlp)

What is it warning me about? Is it "bad" that there are no other constraints? Okay, it falls back to JSOSolvers.tron here: https://github.com/JuliaSmoothOptimizers/Percival.jl/blob/2a9f2cf88b44a1d4e3be515016099f29438bd870/src/method.jl#L36-L47

Is this a problem? Seems totally fine to me... Why is there a warning about this?

[dpo]

It's not a problem. Percival is just alerting you that there is no need to use an augmented-Lagrangian method for a bound-constrained problem. You could have just called Tron directly, and Percival is doing it for you.