postprocess.cc variables need dependence on equations.cc variables

[david-montiel-t]

When the the post-processing variables are declared there needs to be a dependence of at least one of the field variables declared on equations.cc. This is the case even if the post-process variable does not use any of the "equations" variables. Otherwise, the resulting output of the post-processing variable will be zero.

For example the following declaration will always give zero values of the variable "ppv1":

// Variable 1
set_variable_name               (1,"ppv1");
set_variable_type               (1,SCALAR);

set_dependencies_value_term_RHS(1, "");
set_dependencies_gradient_term_RHS(1, "");

set_output_integral             (1,false);

[landinjm]

This is also the case in equations.cc. If no dependency is specified (e.g., if your RHS term was some function of the q point locations), the term is zero by default.