* 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"

  - was doing a step from `out` -> `out`
  - this resulted in null ptr deref after a failed integration step
  - instead do a step from `prev_out` -> `out`, where `prev_out` is a temporary which stores the previous `out` value
  - closes #29

fix dt > max_step error message

  - minor change to existing code:
    - include <dune/copasi/common/filesystem.hh> instead of <filesystem>
    - use namespace `fs` instead of `std::filesystem`
    - where by default `fs` is just an alias for `std::filesystem`
  - if -DUSE_FALLBACK_FILESYSTEM=ON in CMake configuration
    - `fs` instead uses implementation from https://github.com/gulrak/filesystem
    - no longer links or depends on std::filesystem
    - needed for macOS < 10.15 which doesn't have std::filesystem support

  - mingw gcc on msys will not implicitly convert std::filesystem::path to std::string
  - `cannot convert 'path' (type 'const std::filesystem::__cxx11::path') to type 'const string&' {aka 'const std::__cxx11::basic_string<char>&'}`
  - https://stackoverflow.com/a/57377681/6465472