update status_mapt to latest constants

[kalmarek]

Does SCS_SOLVED_INACCURATE qualify for :UserLimit?

[mlubin]

What are the conditions that could trigger that status? If it's an iteration limit, then yes, I'd say UserLimit.

By the way, we're thinking about reworking the MathProgBase status system because it's pretty broken at the moment, see the discussion at JuliaOpt/MathProgBase.jl#156.

[kalmarek]

Yest this can be an iteration limit

[mlubin]

UserLimit isn't appropriate if the solver stops because of an iteration limit which the user didn't set.

[kalmarek]

I'm pretty sure this is me being dumb, but I have no idea how to distinguish SCSSolver() and SCSSolver(iterations=20000) in code; for the first hitting its default 20000 iterations shouldn't return :UserLimit, while the second should (then what should the first one return? :Inaccurate?)

How is discussion on extending the number of solve states progressing?

[mlubin]

User-provided options are stored in SCSMathProgModel.options, so you can check if max_iters was set. If SCS stops because it hit the user-provided iteration limit, it should return :UserLimit. If it stops because it hits its internal limit without solving the problem, it should return something like :Error.

I wouldn't expect a revision of the MPB solver statuses until June.