Previously the code failed for the tutorial if it is compiled with -DFULLDEBUG flag as the geometryWENO::surfIntTrans() method access elements out of bounds of the owner vector.
The case of faces on the boundary patches are now handled by giving them the owner property.
see commit: 3bad99cfd5d23039b80979944a985fec812572d1
The naming of vairables is made consistent throughout geometryWENO and WENOBase.
The volume integral is now always called volIntegrals and the alias for the triple list type is volIntegralType. For aliases or typedefs the C++11 syntax is used with using instead of typedef
The reference frame for volume integration is now adaptive to the mesh size. See commit: d49c69d86c63604964efa7476b573c1871d6cd94
In Foam::geometryWENO the class is converted to a namespace as no constructor is needed (66e5accce4d2c351b2cb54b87c91a2ab8d23e084)
The Jacobi matrix is saved in a Foam::squareMatrix<Type> as this type has built in calculation of the determinant. For this the function geometryWENO::jacobi() is written that generates the jacobi matrix.
Possible memory improvement by only allocating the volIntegral field for the fields that are accessed. Theoretical memory improvement is 50% but tests have shown that the compiler optimized this when -O3 optimization is selected. (This is the default in OpenFOAM)
Following issues have been addressed:
-DFULLDEBUGflag as thegeometryWENO::surfIntTrans()method access elements out of bounds of the owner vector. The case of faces on the boundary patches are now handled by giving them the owner property. see commit: 3bad99cfd5d23039b80979944a985fec812572d1volIntegralsand the alias for the triple list type isvolIntegralType. For aliases or typedefs the C++11 syntax is used withusinginstead oftypedefFoam::geometryWENOthe class is converted to a namespace as no constructor is needed (66e5accce4d2c351b2cb54b87c91a2ab8d23e084)Foam::squareMatrix<Type>as this type has built in calculation of the determinant. For this the functiongeometryWENO::jacobi()is written that generates the jacobi matrix.