Change caching copy logic to allow incremental solvers to use a copy_to function

[odow]

Currently, solvers such as GLPK skip copy_to via JuMP, due to the behavior of CachingOptimizer.

model = Model(GLPK.Optimizer)
@variable(model, x)
optimize!(model)

Doesn't use copy_to, even though GLPK provides a fast one.

Rather than making an ad-hoc change, we should take the time to throughly document the caching optimizer system, choose a design that makes the most sense, and then implement that. There are quite a few points at which various layers are added, dropped, emptied, and reset.

[odow]

Closing because I think we're in a better state now.