* This change requires the parameter tree to contain new parameters
* Updates the tests to always include a default set of newton parameters

[skip ci]

This test was initially committed by mistake. It included zero reactions on the membrane as a test. This includes the full version to be tested with the jacobian test.

The error is still on the order of magnitude of the numerical jacobian differences, so it should be fine. 
And it also makes sense that the error increases with higher random numbers

The change in e96d0f12 makes the result of this test to change. Because it was a memory bug on the domain interface, the new solution feels to be the right one. 

* In a galerkin method this does not matter as indices should be the same, but still is good to have the right semantics.

This helps to keep tests cleaner as well as centralizing logging settings.

* Is easier to reason about the solver out of the model
* For the testing of the jacobian we only need the stationary grid operator so renaming is necessary

When building this object, it was necessary but now it is not anymore.

  - lkeegan/libsbml-static -> spatial-model-editor/sme_deps_common
  - /opt/libs -> /opt/smelibs
  - C:/libs -> C:/smelibs

* Aditionally, the branch `latest` will also be pushed with a `latest` tag.

Resolve "Support 3D simulations"

Closes #31

See merge request !27

[skip-ci]

Resolve "Add simple adaptive timestepping"

Closes #29, #30, #12, #28, and #26

See merge request !25

[skip-ci]

Before, the md comparison version was not obtaining the grid functions from the latest state and led to much higher errors and a different result with respect to sd executable

- Full verbosity of one step method
- Fix code format on executables

- There was an error on the coefficients of the evolution of the gaussian equation.
- Reduced simulated time and diffusion to make bell fit almost always in the domain.

  - now uses `in` as initial state instead of `out`
  - previous behaviour was ok until commit aa186676 which changes `out` in between calls

  - closes #30

  - this is thrown if the iterative solver doesn't converge
  - e.g. "SolverAbort: breakdown in BiCGSTAB - omega -0 <= EPSILON 1e-80 after 1835.5 iterations"