WIP:Actsfw-35 fatras integration

This merge request contains adaptions to use acts-fatras simulation. Fixes ACTSFW-35.

1. Introduces a flag USE_FATRAS if fatras the fatras example using acts-fatras should be build
2. Introduces a FatrasAlgorithm internally using the ExtrapolationAlgorithm, because extra functionality is needed. For the moment it only needs to set the random engine of the MaterialEffectsEngine.
3. Possibility to hand over MaterialEffectsEngine to ExtrapolationUtils (in this way the Fatras example can use the ExtrapolationUtils and code duplication is avoided). If no MaterialEffectsEngine is handed over it uses the default one of ats-core.
4. Adaption of RootExCellWriter to write out the differences in momentum and position (comparing first and last step)
5. The Fatras Example now uses acts-fatras and its MaterialEffectsEngine and the multiple scattering highland sampler and energyloss (due to ionization) sample - Both have been validated againt Geant4 => The FatrasExample has been validated against previous results and Geant4 after rebasing
6. Introduce FatrasValidation algorithm & example - This is a geant4 simulation writing out the difference in momentum and position which was used to validate the fatras material effects. After rebasing this example could not be run again, since the current LCG90a release has some problems with Geant4 libraries.
7. A MaterialEffectsWriter to compare different particle properties has been introduced.
8. Root scripts to write out and compare the material effects have been created.
9. Updates the submodules acts-core & acts-fatras to the current masters
10. Updates the gitlab ci job to build fatras

The validation has been done using the fatras master branch for muons and can be found in acts-data (https://gitlab.cern.ch/acts/acts-data/tree/FatrasValidation_MSC_eLoss).

This merge request depends on !69 (merged) and should be merged afterwards. It uses merge request 345 of acts-core.