#1022 Move the ParameterAtQuadraturePoint one level upward. This is just a step: the idea is at the end for them to be in FluidPressure data object...

#1022 Poromechanics: new term in PressureCOntribToRhs is now computed with a GlobalParameterOperator and a 'standard' Parameter. The reason is that the hack used previously wouldn't work for other operators such as Darcy, which do not cover at all solid displacement.

#859 GlobalParameterOperator: add the DoAllocateGradientPhi constructor argument (I removed it previously, believing wrongly it wouldn't be useful in any situation).

#1022 Toward replacing fluid pressure by expression depending on fluid mass in Darcy operator. This first approach was too naive: I can't piggyback the solid required computations of CauchyGreen on an operator which works on a completely unrelated unknown (I did it once for another operator but it worked solely because unknowns exhibit similar behaviour: solid velocity/solid displacement).

#1022 Make Darcy a template class with hyperelastic law as argument. Rename a method of ElementaryData classes.

#1022 PressureAtQuadPt is no longer a full-fledged Parameter, but a similar object which GetValue() takes more argument to give all the underlying invariantHolder needs to update directly.

#1022 First draft of PressureAtQuadPt, and use it in the pressure contrib to rhs. Present version is rather crude: parameter must be defined within the local variational operator (not pretty but we can live with that) and does not handle at all update of invariant (much more problematic: this is VERY error prone).

#1022 Introduce skeleton of the Parameter class which will include the computation on the fly of pressure at quadrature points.

#1022 Poromechanics: stop using mass matrix to evaluate rhs term of fluid pressure. This way, there are no longer any direct use of a GlobalVector related to fluid pressure.

#1022 - #1035 Introduce in poromechanics a scalar transient source, and fix remaining problems in the library about them.

#1022 FluidPressure: make global vector not publicly accessible. They were actually only used... to print the values with DevPrint. It is rather important to check that now: new scheme proposed by Bruno make the pressure disappear entirely, and be replaced by an expression in function of fluid mass expressed at quadrature points. So to make sure the transition is smooth we need to ensure fluid pressure is really used only through Parameters...

#1045 Add a new field in SCons to specify subdirectory into which build the current libraries and executables.

#1040 Poromechanics: porosity is now computed with the brand new CauchyGreen tensor (so it is available in PoroVarf).

#1040 PressureContribToRhs: use new Cauchy Green tensor (in a Parameter) instead of former one computed on the fly.

#1040 SecondPiolaStressTensor: now invariant is computed with the value stored in the parameter, rather than the formerly computed one, only in hyperelasticity case (active stress and viscosity will be modified likewise in next commit).

#1040 Now update calls are correctly done. Next step is to actually use the newly computed CauchyGreen tensor (instead of the one computed within the variational operator.