Correctly initialise the signal velocity of the SPH particles in the Minimal and Pressure-Entropy cases.

Implemented the changes in ls that prevent gizmo crashes in cases where the mesh is seriously distorted

Initialize the signal velocity to be the particle's sound-speed. Prevents divisions by zero in cases where a particle has 0 neighbours.

Updated plotting scripts for the hydrostatic halo case to be more SSH friendly and to read the correct parameters in the yaml file.

New addition to the test suite: A perturbed version of the 125 cells case



See merge request !341

Updated the scripts running the tests to actually crash if they fail at any iteration and not just on the last one. Code formatting. Do not compile the FFT functions if FFTW is not found.

Split the drift into hydro and gravity drift tasks. Implement FFT task and dependencies

Changes:
 - Two drift tasks, one pert particle type.
 - FFT task with gravity ghost tasks for the dependencies.

It passes my usual tests but could you make sure it does not break anything ? That should be the last big change. The rest of gravity should be contained within the tasks themselves.

See merge request !340

No limit on h in test125cells. Add more digits to the output. Perturbed version of the test125cells now works.

Conflicts:
	tests/test125cells.c