Fredrik Bagge Carlson

The number of allocations made and thus the time to solve can probably be reduced quite dramatically by switching to a StaticArrays approach for the dynamics matrices. Benchmarks [here](https://github.com/baggepinnen/LowLevelParticleFilters.jl/blob/master/docs/benchstatic.md)

https://github.com/baggepinnen/DifferentialDynamicProgramming.jl/blob/master/src/forward_pass.jl#L42

Using something from https://github.com/JuliaDiff

Make `nlp` a concretely typed field in `MadNLPSolver`

1

Before this PR, all accesses to fields on `ips.nlp` were type unstable and caused allocations.

Definition of primal solution?

3

I'm a bit confused about how to interpret the result The example https://madnlp.github.io/MadNLP.jl/dev/quickstart/ contains two variables, but ``` ips.x 4-element Vector{Float64}: -0.7921232178470456 -1.2624298435831807 0.9999999900052515 0.801605892122408 ``` has length 4. What...

Reuse existing `MadNLPSolver`

4

The construction of an `MadNLPSolver` allocates a lot of memory, and repeatedly solving similar optimization problems by defining a new `MadNLPSolver` for each problem is thus associated with a lot...

Cannot construct PDMat from Cholesky{..., ::UpperTriangular}

4

According to the readme, this should work, and there is a method for it, but it fails ```julia julia> c = Cholesky(UpperTriangular(q.R), :U, 0) Cholesky{Float64,UpperTriangular{Float64,Array{Float64,2}}} U factor: 4×4 UpperTriangular{Float64,Array{Float64,2}}: 10.0084...

Plotly.jl does not have upper bounds listed on some of it's dependencies, notably `HTTP.jl`. Upper bounds have been introduced on all dependencies in `general`, c.f., https://github.com/JuliaRegistries/General/blob/master/P/Plotly/Compat.toml#L2 This causes some confusion...

Plots overwrite each other in IJulia

1

When using IJulia and plotting with Plotly as backend to plots, plots after the first one seem to overwrite the first plot rather than creating a new plot. Example ```julia...