JuliaApproximation / FrameFun.jl

Exploring practical possibilities of approximating functions with frames rather than with a basis
Other
23 stars 7 forks source link

Compile errors in FrameFun.jl #13

Closed rctrl11 closed 3 years ago

rctrl11 commented 3 years ago

Following the commands suggested in an example on the website--using BasisFunctions, Plots, DomainSets, FrameFun gr(), I have the following compile errors:

WARNING: Method definition *(ToeplitzMatrices.Circulant{T1<:Real, S} where S<:Number, ToeplitzMatrices.Circulant{T2<:Real, S} where S<:Number) where {T1<:Real, T2<:Real} in module ToeplitzMatrices at.....

WARNING: could not import BasisFunctions.stepsize into FrameFun WARNING: could not import BasisFunctions.dualdictionary into FrameFun WARNING: could not import BasisFunctions.quadraturenormalization into FrameFun WARNING: Method definition convert(Type{Tuple{Vararg{Int64, N}}}, Base.IteratorsMD.CartesianIndex{N}) where {N} in module GridArrays at /Users/ronchan/.julia/packages/GridArrays/60TtN/src/subgrid/maskedsubgrid.jl:35 overwritten in module FrameFun at /Users/ronchan/.julia/packages/FrameFun/c2GQe/src/sampling/subgrid.jl:32. incremental compilation may be fatally broken for this module

I have no clue what I have done wrong when I installed all the packages needed for running the example code. Currently, my revision of Julia is 1.5 and pretty sure that I have installed FrameFun perfectly via the command suggested on the website.

Would be great if you could shed some light to resolve my problem?

daanhb commented 3 years ago

Hi, thanks for your interest. Sorry to hear you're having difficulties installing - it is very likely not your fault. Did you use the registry at FrameFunVC by any chance? The problem is probably that this registry is not up to date. The BasisFunctions and FrameFun packages are not regular registered packages, I've been meaning to do something about that and then the situation will improve. Would you mind trying to clone from the master branches of these two packages? Currently, that has better chances of succeeding.

An alternative for function approximation would be ApproxFun, which is much better maintained. Consider FrameFun to be research code... (But if you have any research questions, feel free to ask about that, here or via e-mail.)