the new general purpose index design introduced a bug in the loader related to the byte position of the byteStart field. This has been resolved. Further, I have updated the save to ascii format to provide an explicit warning that the ascii version is only for debugging purposes and not meant to be used directly.
OPAL type II tables use a 4 vector for composition, while many other opacity tables use a 2 vector of composition to index tables. In an attempt to be as general as possible I have updated the OPAT specification to allow for tables indexed by a vector of any length [1, 255]. This should allow a single spec to be used when storeing and accessing type I or type II tables.
BREAKING CHANGE: Careful thought will have to be given to how the C++ module handles this. We want the same code to work for querying data from any kind of opaticy table
opatIO can now read tables properly and retreive them in a useful manner. Future aditions will be the ability to lookup "closest" tables and a pretty printer for tables.
the polytrope module will be used as an initial guess to the solver. A skeleton of this has been imported from https://github.com/tboudreaux/FEMPolytrope
This module will need major updates still to handle 3D, proper boundary conditions, and to incorporate it with the rest of our meshing scheme
MFEM is a dependency of our code and we want to build it with our code for portability, here I add a meson subproject which fetches mfem and builds it. Because of some issues with CMake and meson I apply a patch which manually disables running tests and building examples and miniapps in mfem. This is okay as mfem here is not intenteded to be linked against other programs (though it still can be).
const needds to be accessed all throughout so it has been changed to a singleton to allow for more efficient usage
BREAKING CHANGE: Any previous loads to const will break, also constant->Constant and constants->Constants
