ODEProblem gives "ArgumentError: reducing over an empty collection is not allowed"

[ianwilliamson]

This package looks really nice and seems to have come a very long way since I last looked at it back in Jan/Feb. I am trying the following very simple test, based on the heat equation test script in the repo (which seems to use an outdated BC API):

using DiffEqOperators, OrdinaryDiffEq, PyPlot

N = 16
h = 2π/(N-1)
x = collect(0.0 : h : 2π)
u0 = sin.(x)

Dxx = CenteredDifference(2, 2, h, N)
Q = Dirichlet0BC(Float64)

L = Dxx * Q

eqn = ODEProblem(L, u0, (0.0, 1.0))
u = solve(eqn, Tsit5(), dense=false, tstops=0.0:0.01:1.0)

plt.figure()
plt.plot(u[:, end])

I get the following error:

ERROR: LoadError: ArgumentError: reducing over an empty collection is not allowed
Stacktrace:
 [1] _empty_reduce_error() at ./reduce.jl:220
 [2] reduce_empty(::Function, ::Type) at ./reduce.jl:230
 [3] mapreduce_empty(::typeof(identity), ::Function, ::Type) at ./reduce.jl:255
 [4] _mapreduce(::typeof(identity), ::typeof(max), ::IndexLinear, ::Array{Any,1}) at ./reduce.jl:309
 [5] _mapreduce_dim at ./reducedim.jl:312 [inlined]
 [6] #mapreduce#555 at ./reducedim.jl:307 [inlined]
 [7] mapreduce at ./reducedim.jl:307 [inlined]
 [8] _maximum at ./reducedim.jl:657 [inlined]
 [9] _maximum(::Array{Any,1}, ::Colon) at ./reducedim.jl:656
 [10] #maximum#562 at ./reducedim.jl:652 [inlined]
 [11] maximum at ./reducedim.jl:652 [inlined]
 [12] numargs(::GhostDerivativeOperator{Float64,DerivativeOperator{Float64,1,false,Float64,StaticArrays.SArray{Tuple{3},Float64,1,3},StaticArrays.SArray{Tuple{0},StaticArrays.SArray{Tuple{4},Float64,1,4},1,0},Nothing,Nothing},RobinBC{Float64}}) at /Users/___/.julia/packages/DiffEqBase/Fs7ts/src/utils.jl:13
 [13] isinplace(::GhostDerivativeOperator{Float64,DerivativeOperator{Float64,1,false,Float64,StaticArrays.SArray{Tuple{3},Float64,1,3},StaticArrays.SArray{Tuple{0},StaticArrays.SArray{Tuple{4},Float64,1,4},1,0},Nothing,Nothing},RobinBC{Float64}}, ::Int64) at /Users/___/.julia/packages/DiffEqBase/Fs7ts/src/utils.jl:26
 [14] #ODEFunction#85(::Base.Iterators.Pairs{Symbol,Nothing,NTuple{8,Symbol},NamedTuple{(:analytic, :tgrad, :jac, :Wfact, :Wfact_t, :paramjac, :syms, :colorvec),NTuple{8,Nothing}}}, ::Type{ODEFunction}, ::GhostDerivativeOperator{Float64,DerivativeOperator{Float64,1,false,Float64,StaticArrays.SArray{Tuple{3},Float64,1,3},StaticArrays.SArray{Tuple{0},StaticArrays.SArray{Tuple{4},Float64,1,4},1,0},Nothing,Nothing},RobinBC{Float64}}) at /Users/___/.julia/packages/DiffEqBase/Fs7ts/src/diffeqfunction.jl:358
 [15] Type at ./none:0 [inlined]
 [16] convert at /Users/___/.julia/packages/DiffEqBase/Fs7ts/src/diffeqfunction.jl:940 [inlined]
 [17] #ODEProblem#214(::Base.Iterators.Pairs{Union{},Union{},Tuple{},NamedTuple{(),Tuple{}}}, ::Type{ODEProblem}, ::GhostDerivativeOperator{Float64,DerivativeOperator{Float64,1,false,Float64,StaticArrays.SArray{Tuple{3},Float64,1,3},StaticArrays.SArray{Tuple{0},StaticArrays.SArray{Tuple{4},Float64,1,4},1,0},Nothing,Nothing},RobinBC{Float64}}, ::Array{Float64,1}, ::Tuple{Float64,Float64}, ::DiffEqBase.NullParameters) at /Users/___/.julia/packages/DiffEqBase/Fs7ts/src/problems/ode_problems.jl:78
 [18] Type at /Users/___/.julia/packages/DiffEqBase/Fs7ts/src/problems/ode_problems.jl:78 [inlined] (repeats 2 times)
 [19] top-level scope at /Users/___/drive/Code/test.jl:13
 [20] include at ./boot.jl:328 [inlined]
 [21] include_relative(::Module, ::String) at ./loading.jl:1094
 [22] include(::Module, ::String) at ./Base.jl:31
 [23] include(::String) at ./client.jl:431
 [24] top-level scope at REPL[4]:1
in expression starting at /Users/___/drive/Code/test.jl:13

Am I making a stupid mistake?

[ChrisRackauckas]

Not a stupid mistake. We aren't quite ready yet. We've completely transformed the interface into something that can work, but we need to iron out a lot of missing functions over the next year. But yes, we're getting there.

[ianwilliamson]

Okay, no worries. Will keep tracking the progress of this.