|
Ferrite.reinit!(vals::NTuple{N,Ferrite.Values}, args...) where N = foreach(v->reinit!(v, args...), vals) |
|
Ferrite.reinit!(vals::NamedTuple, args...) = foreach(v->reinit!(v, args...), vals) |
When working with https://github.com/Ferrite-FEM/Ferrite.jl/tree/kam/CellMultiValues it seems like this code pattern could allocate and be slow. Consider changing to map (and also change the name to avoid type piracy).
Also, the dispatch of NTuple might not work for different values, hence it should just be Tuple (which is ok once it has a separate name anyways)
Alternatively, support only the CellMultiValues once they are completed.
FerriteAssembly.jl/src/ferrite_additions.jl
Lines 17 to 18 in a05511c
When working with https://github.com/Ferrite-FEM/Ferrite.jl/tree/kam/CellMultiValues it seems like this code pattern could allocate and be slow. Consider changing to map (and also change the name to avoid type piracy).
Also, the dispatch of
NTuplemight not work for different values, hence it should just be Tuple (which is ok once it has a separate name anyways)Alternatively, support only the
CellMultiValuesonce they are completed.