Different changes to FEInterfaceValues and NM::FEInterfaceValues
#17045
Conversation
Use active_fe_index Fix
peterrum
force-pushed
the
fe_interface_value_cell_accessor
branch
from
2ede70a
to
b459ab5
Compare
| template <typename CellIteratorType, typename CellNeighborIteratorType> | ||
| template <typename CellType> | ||
| void | ||
| reinit(const CellIteratorType &cell, | ||
| const unsigned int face_no, | ||
| const unsigned int sub_face_no, | ||
| const CellNeighborIteratorType &cell_neighbor, | ||
| const unsigned int face_no_neighbor, | ||
| const unsigned int sub_face_no_neighbor, | ||
| const unsigned int q_index = numbers::invalid_unsigned_int, | ||
| const unsigned int mapping_index = numbers::invalid_unsigned_int, | ||
| const unsigned int fe_index = numbers::invalid_unsigned_int); | ||
| reinit(const TriaIterator<CellType> &cell, | ||
| const unsigned int face_no, | ||
| const unsigned int sub_face_no, | ||
| const TriaIterator<CellType> &cell_neighbor, |
There was a problem hiding this comment.
I don't understand this change. First, it shouldn't be CellType as template argument, but AccessorType (or perhaps CellAccessorType), but more importantly, I'm unclear what it is you are trying to enable with the change that wasn't possible before (or what you want to prevent that accidentally compiled before but shouldn't have)?
There was a problem hiding this comment.
I'm unclear what it is you are trying to enable with the change that wasn't possible before (or what you want to prevent that accidentally compiled before but shouldn't have)?
The code evaluated the shape functions at the quadrature points AND loaded the DoF indices. You can do the latter only of DoFCellAccessors.
There was a problem hiding this comment.
Ah, so you're saying that the change to the iterator type is only because you wanted to get access to the Accessor type of the iterator? You could have gotten that via typename CellIteratorType::AccessorType, which I think makes all of this easier to read because users don't typically know what classes are behind the cell iterators.
FEInterfaceValuescan now acceptTriaIterator<CellAccessor>NM::FEInterfaceValuescan now acceptTriaIterator<CellAccessor>active_fe_indexinFEInterfaceValues(there have been places where the argument was not used)FEInterfaceValuesinNM::FEInterfaceValuesso that one does not have to store vectors ofFEInterfaceValues(I believeNM::FEValuescould be rewritten similarly: instead of usingFEValuesand would usehp::FEValuesand get the currentFEValuesinstance byhp::FEValues::get_present_fe_values(); by thestd::optionalmakes this complicated/impossible (?))