Promote tspan for discrete probs #928
Conversation
|
I don't think any of these failures are related, but it is quite a large list of packages with test issues right now it seems. |
|
There's lots of little things to do. |
| u0, tspan::Tuple, p = NullParameters(); | ||
| u0, tspan, p = NullParameters(); |
There was a problem hiding this comment.
Doesn't this give an ambiguity?
There was a problem hiding this comment.
With which method?
Aqua reports:
julia> Aqua.test_ambiguities(SciMLBase)
8 ambiguities found. To get a list, set `broken = false`.
Ambiguity #1
(L::SciMLBase.AffineDiffEqOperator)(du, u, p, t::Number) @ SciMLBase ~/.julia/dev/SciMLBase/src/operators/diffeq_operator.jl:49
(L::SciMLOperators.AbstractSciMLOperator)(du::Number, u::Number, p, t, args...; kwargs...) @ SciMLOperators ~/.julia/packages/SciMLOperators/KVzmP/src/interface.jl:122
Possible fix, define
(::SciMLBase.AffineDiffEqOperator)(::Number, ::Number, ::Any, ::Number)
Ambiguity #2
(L::SciMLBase.DiffEqArrayOperator)(du, u, p, t) @ SciMLBase ~/.julia/dev/SciMLBase/src/operators/basic_operators.jl:205
(L::SciMLOperators.AbstractSciMLOperator)(du::Number, u::Number, p, t, args...; kwargs...) @ SciMLOperators ~/.julia/packages/SciMLOperators/KVzmP/src/interface.jl:122
Possible fix, define
(::SciMLBase.DiffEqArrayOperator)(::Number, ::Number, ::Any, ::Any)
Ambiguity #3
(L::SciMLBase.DiffEqIdentity)(du, u, p, t) @ SciMLBase ~/.julia/dev/SciMLBase/src/operators/basic_operators.jl:205
(L::SciMLOperators.AbstractSciMLOperator)(du::Number, u::Number, p, t, args...; kwargs...) @ SciMLOperators ~/.julia/packages/SciMLOperators/KVzmP/src/interface.jl:122
Possible fix, define
(::SciMLBase.DiffEqIdentity)(::Number, ::Number, ::Any, ::Any)
Ambiguity #4
(L::SciMLBase.DiffEqScalar)(du, u, p, t) @ SciMLBase ~/.julia/dev/SciMLBase/src/operators/basic_operators.jl:205
(L::SciMLOperators.AbstractSciMLOperator)(du::Number, u::Number, p, t, args...; kwargs...) @ SciMLOperators ~/.julia/packages/SciMLOperators/KVzmP/src/interface.jl:122
Possible fix, define
(::SciMLBase.DiffEqScalar)(::Number, ::Number, ::Any, ::Any)
Ambiguity #5
(L::SciMLBase.FactorizedDiffEqArrayOperator)(du, u, p, t) @ SciMLBase ~/.julia/dev/SciMLBase/src/operators/basic_operators.jl:205
(L::SciMLOperators.AbstractSciMLOperator)(du::Number, u::Number, p, t, args...; kwargs...) @ SciMLOperators ~/.julia/packages/SciMLOperators/KVzmP/src/interface.jl:122
Possible fix, define
(::SciMLBase.FactorizedDiffEqArrayOperator)(::Number, ::Number, ::Any, ::Any)
Ambiguity #6
mul!(Y::AbstractArray, α::SciMLBase.DiffEqScalar, B::AbstractArray) @ SciMLBase ~/.julia/dev/SciMLBase/src/operators/basic_operators.jl:76
mul!(Y::AbstractVecOrMat, L::SciMLBase.AbstractDiffEqLinearOperator, B::AbstractVecOrMat) @ SciMLBase ~/.julia/dev/SciMLBase/src/operators/common_defaults.jl:43
Possible fix, define
mul!(::AbstractVecOrMat, ::SciMLBase.DiffEqScalar, ::AbstractVecOrMat)
Ambiguity #7
size(L::SciMLBase.FactorizedDiffEqArrayOperator, args...) @ SciMLBase ~/.julia/dev/SciMLBase/src/operators/basic_operators.jl:188
size(A::SciMLOperators.AbstractSciMLOperator, d::Integer) @ SciMLOperators ~/.julia/packages/SciMLOperators/KVzmP/src/interface.jl:178
Possible fix, define
size(::SciMLBase.FactorizedDiffEqArrayOperator, ::Integer)
Ambiguity #8
size(L::SciMLBase.FactorizedDiffEqArrayOperator, args...) @ SciMLBase ~/.julia/dev/SciMLBase/src/operators/basic_operators.jl:188
size(L::SciMLBase.AbstractDiffEqLinearOperator, i::Integer) @ SciMLBase ~/.julia/dev/SciMLBase/src/operators/common_defaults.jl:7
Possible fix, define
size(::SciMLBase.FactorizedDiffEqArrayOperator, ::Integer)
Test Failed at /Users/isaacsas/.julia/packages/Aqua/epbUr/src/ambiguities.jl:78
Expression: iszero(num_ambiguities)
ERROR: There was an error during testingThere was a problem hiding this comment.
Do you mean the second and third constructors when passing three arguments with the first two being nothing?
There was a problem hiding this comment.
Should I make tspan a union of vectors, tuples, and numbers? tspan::T where {T <: Union{Vector,Tuple,Number}?
There was a problem hiding this comment.
I think so. I don't think the no function or f = identity dispatch is compatible with getting rid of the tspan::Tuple or some form of a type marker, since otherwise you cannot distinguish it from p.
There was a problem hiding this comment.
Doesn't the old version have the same ambiguity though. I don't see how changing the first constructor, the only one I modified, impacts that. The ambiguity is between the second and third constructors isn't it?
I also don't get any messages about ambiguities locally on 1.11.3 when calling
p = (0.1 / 1000, 0.01)
dprob4 = DiscreteProblem{true}(nothing, nothing, p)
|
@ChrisRackauckas I've added more constructor instances. At least locally none of these result in errors about ambiguities. I don't know if this covers the particular case you are worried about though. |
|
Your test does cover the cases I was worried about, so I'm happy to merge it and see. I have so much clean up to do from the v1.11 memory changes still 😅 |
Closes #927