Cannot precompile on latest Julia master: no method matching length(::Nothing)

[alhirzel]

See: JuliaLang/julia#49171

Jameson Nash believes there may be an issue with traits/generated functions. This results in the following error on latest Julia master when precompiling OrdinaryDiffEq:

ERROR: LoadError: MethodError: no method matching length(::Nothing)

Closest candidates are:
  length(::Union{Base.KeySet, Base.ValueIterator})
   @ Base abstractdict.jl:58
  length(::Union{ArrayInterface.BidiagonalIndex, ArrayInterface.TridiagonalIndex})
   @ ArrayInterface ~/prj/20230327_bisect_julia/depot/packages/ArrayInterface/xqHUD/src/ArrayInterface.jl:804
  length(::Union{Tables.AbstractColumns, Tables.AbstractRow})
   @ Tables ~/prj/20230327_bisect_julia/depot/packages/Tables/AcRIE/src/Tables.jl:180
  ...

Stacktrace:
  [1] _similar_shape(itr::Nothing, ::Base.HasLength)
    @ Base ./array.jl:658
  [2] _array_for(::Type{Core.MethodInstance}, itr::Nothing, isz::Base.HasLength)
    @ Base ./array.jl:674
  [3] _method_instances(f_type::Any, t_type::Any)
    @ Tricks ~/prj/20230327_bisect_julia/depot/packages/Tricks/zD0qD/src/Tricks.jl:136
  [4] #s1#1
    @ ~/prj/20230327_bisect_julia/depot/packages/Tricks/zD0qD/src/Tricks.jl:20 [inlined]
  [5] var"#s1#1"(T::Any, ::Any, f::Any, t::Any)
    @ Tricks ./none:0
  [6] (::Core.GeneratedFunctionStub)(::UInt64, ::LineNumberNode, ::Any, ::Vararg{Any})
    @ Core ./boot.jl:600
  [7] promote_f
    @ ~/prj/20230327_bisect_julia/depot/packages/DiffEqBase/qPmC2/src/solve.jl:1081 [inlined]
  [8] get_concrete_problem(prob::SciMLBase.ODEProblem{Vector{Float64}, Tuple{Float64, Float64}, true, SciMLBase.NullParameters, SciMLBase.ODEFunction{true, SciMLBase.AutoSpecialize, OrdinaryDiffEq.var"#lorenz#651", LinearAlgebra.UniformScaling{Bool}, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, typeof(SciMLBase.DEFAULT_OBSERVED), Nothing, Nothing}, Base.Pairs{Symbol, Union{}, Tuple{}, NamedTuple{(), Tuple{}}}, SciMLBase.StandardODEProblem}, isadapt::Bool; kwargs::Base.Pairs{Symbol, Any, Tuple{Symbol, Symbol}, NamedTuple{(:u0, :p), Tuple{Vector{Float64}, SciMLBase.NullParameters}}})
    @ DiffEqBase ~/prj/20230327_bisect_julia/depot/packages/DiffEqBase/qPmC2/src/solve.jl:1022
  [9] get_concrete_problem
    @ ~/prj/20230327_bisect_julia/depot/packages/DiffEqBase/qPmC2/src/solve.jl:1016 [inlined]
 [10] solve_up(prob::SciMLBase.ODEProblem{Vector{Float64}, Tuple{Float64, Float64}, true, SciMLBase.NullParameters, SciMLBase.ODEFunction{true, SciMLBase.AutoSpecialize, OrdinaryDiffEq.var"#lorenz#651", LinearAlgebra.UniformScaling{Bool}, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, typeof(SciMLBase.DEFAULT_OBSERVED), Nothing, Nothing}, Base.Pairs{Symbol, Union{}, Tuple{}, NamedTuple{(), Tuple{}}}, SciMLBase.StandardODEProblem}, sensealg::Nothing, u0::Vector{Float64}, p::SciMLBase.NullParameters, args::OrdinaryDiffEq.BS3{typeof(OrdinaryDiffEq.trivial_limiter!), typeof(OrdinaryDiffEq.trivial_limiter!), Static.False}; kwargs::Base.Pairs{Symbol, Union{}, Tuple{}, NamedTuple{(), Tuple{}}})
    @ DiffEqBase ~/prj/20230327_bisect_julia/depot/packages/DiffEqBase/qPmC2/src/solve.jl:929
 [11] solve_up
    @ ~/prj/20230327_bisect_julia/depot/packages/DiffEqBase/qPmC2/src/solve.jl:905 [inlined]
 [12] solve(prob::SciMLBase.ODEProblem{Vector{Float64}, Tuple{Float64, Float64}, true, SciMLBase.NullParameters, SciMLBase.ODEFunction{true, SciMLBase.AutoSpecialize, OrdinaryDiffEq.var"#lorenz#651", LinearAlgebra.UniformScaling{Bool}, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, typeof(SciMLBase.DEFAULT_OBSERVED), Nothing, Nothing}, Base.Pairs{Symbol, Union{}, Tuple{}, NamedTuple{(), Tuple{}}}, SciMLBase.StandardODEProblem}, args::OrdinaryDiffEq.BS3{typeof(OrdinaryDiffEq.trivial_limiter!), typeof(OrdinaryDiffEq.trivial_limiter!), Static.False}; sensealg::Nothing, u0::Nothing, p::Nothing, wrap::Val{true}, kwargs::Base.Pairs{Symbol, Union{}, Tuple{}, NamedTuple{(), Tuple{}}})
    @ DiffEqBase ~/prj/20230327_bisect_julia/depot/packages/DiffEqBase/qPmC2/src/solve.jl:842
 [13] solve(prob::SciMLBase.ODEProblem{Vector{Float64}, Tuple{Float64, Float64}, true, SciMLBase.NullParameters, SciMLBase.ODEFunction{true, SciMLBase.AutoSpecialize, OrdinaryDiffEq.var"#lorenz#651", LinearAlgebra.UniformScaling{Bool}, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, Nothing, typeof(SciMLBase.DEFAULT_OBSERVED), Nothing, Nothing}, Base.Pairs{Symbol, Union{}, Tuple{}, NamedTuple{(), Tuple{}}}, SciMLBase.StandardODEProblem}, args::OrdinaryDiffEq.BS3{typeof(OrdinaryDiffEq.trivial_limiter!), typeof(OrdinaryDiffEq.trivial_limiter!), Static.False})
    @ DiffEqBase ~/prj/20230327_bisect_julia/depot/packages/DiffEqBase/qPmC2/src/solve.jl:832
 [14] macro expansion
    @ ~/prj/20230327_bisect_julia/depot/packages/OrdinaryDiffEq/FttEX/src/OrdinaryDiffEq.jl:348 [inlined]
 [15] top-level scope
    @ ~/prj/20230327_bisect_julia/depot/packages/SnoopPrecompile/1XXT1/src/SnoopPrecompile.jl:62
 [16] include
    @ ./Base.jl:468 [inlined]
 [17] include_package_for_output(pkg::Base.PkgId, input::String, depot_path::Vector{String}, dl_load_path::Vector{String}, load_path::Vector{String}, concrete_deps::Vector{Pair{Base.PkgId, UInt128}}, source::Nothing)
    @ Base ./loading.jl:2043
 [18] top-level scope
    @ stdin:2
in expression starting at /home/alhirzel/prj/20230327_bisect_julia/depot/packages/OrdinaryDiffEq/FttEX/src/OrdinaryDiffEq.jl:1
in expression starting at stdin:2
ERROR: Failed to precompile OrdinaryDiffEq [1dea7af3-3e70-54e6-95c3-0bf5283fa5ed] to "depot/compiled/v1.10/OrdinaryDiffEq/jl_2EajZX".
Stacktrace:
 [1] error(s::String)
   @ Base ./error.jl:35
 [2] compilecache(pkg::Base.PkgId, path::String, internal_stderr::IO, internal_stdout::IO, keep_loaded_modules::Bool)
   @ Base ./loading.jl:2293
 [3] compilecache
   @ ./loading.jl:2160 [inlined]
 [4] _require(pkg::Base.PkgId, env::String)
   @ Base ./loading.jl:1804
 [5] _require_prelocked(uuidkey::Base.PkgId, env::String)
   @ Base ./loading.jl:1658
 [6] macro expansion
   @ ./loading.jl:1644 [inlined]
 [7] macro expansion
   @ ./lock.jl:267 [inlined]
 [8] require(into::Module, mod::Symbol)
   @ Base ./loading.jl:1607

My version info is below. Happy to help debug the issue.

Status `~/prj/20230327_bisect_julia/depot/environments/v1.10/Project.toml`
  [1dea7af3] OrdinaryDiffEq v6.49.4

Julia Version 1.10.0-DEV.873
Commit dc3953db43* (2023-03-24 13:56 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
  CPU: 24 × AMD Ryzen 9 5900X 12-Core Processor
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-14.0.6 (ORCJIT, znver3)
  Threads: 1 on 24 virtual cores
Environment:
  JULIA_DEPOT_PATH = depot
  JULIA_EDITOR = nvim
Status `~/prj/20230327_bisect_julia/depot/environments/v1.10/Project.toml`
  [1dea7af3] OrdinaryDiffEq v6.49.4

Julia Version 1.10.0-DEV.873
Commit dc3953db43* (2023-03-24 13:56 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
  CPU: 24 × AMD Ryzen 9 5900X 12-Core Processor
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-14.0.6 (ORCJIT, znver3)
  Threads: 1 on 24 virtual cores
Environment:
  JULIA_DEPOT_PATH = depot
  JULIA_EDITOR = nvim

[ChrisRackauckas]

This was put on the wrong repo and meant for Tricks.jl?

[alhirzel]

That may be. The issue does not occur when I precompile Tricks.jl though, and I don't know how to articulate the issue for the Tricks.jl developers as I am not familiar enough with these packages and the way OrdinaryDiffEq.jl is using Tricks.jl.

I see @oxinabox is also a member of the SciML organization, so I can tag her. Perhaps she can shine a light. Happy to close this and file directly against Tricks.jl if someone confirms.

[ChrisRackauckas]

Jameson's comment is that Tricks.jl violates a compiler assumption and that's what the root cause is, so I would assume we'd address this by fixing that. @vtjnash did mention that Tricks would not be needed if certain backedges were added to Base. Were those ever added? We can just drop Tricks if they were.

[alhirzel]

Ah, I saw Traits, not Tricks! I assumed he was referencing hand coded traits within OrdinaryDiffEq. Sorry for the confusion.

[ChrisRackauckas]

I assume it was a typo, since the stack trace was pointing to issues with Tricks.jl

  [3] _method_instances(f_type::Any, t_type::Any)
    @ Tricks ~/prj/20230327_bisect_julia/depot/packages/Tricks/zD0qD/src/Tricks.jl:136
  [4] #s1#1
    @ ~/prj/20230327_bisect_julia/depot/packages/Tricks/zD0qD/src/Tricks.jl:20 [inlined]
  [5] var"#s1#1"(T::Any, ::Any, f::Any, t::Any)
    @ Tricks ./none:0

[vtjnash]

Those edges were added

[ChrisRackauckas]

@oxinabox should Tricks then just deprecate and fallback to the normal methods after that commit?

[Keno]

static_hasmethod is a hack. This package should use the regular supported hasmethod and if there's some sort of perf issue with it, file an issue with Base and we'll fix it.

[oxinabox]

Yes, now that that has been added to Base,
Tricks.jl should after that commit just do:

if version > V1.10+DEV-????
     const static_hasmethod = hasmethod
else
     # current definition
      ...
end

and a note added to the readme about "You don't need this if you only need to support Julia 1.10+"
Tricks.jl has served it's purpose: it showed this stuff was useful enough to get it into the language itself.

If someone can tell me the version of Julia that added that feature,
I can make that change in Tricks.jl.
(If someone tells me just the commit, i will have to remember how to use the thing in contrib to compute versions from commits, but i can still do it)

[Keno]

If someone can tell me the version of Julia that added that feature,

JuliaLang/julia@f9e1c6c

[Keno]

keno@deepsea2:~/julia/contrib$ ./commit-name.sh f9e1c6cb5963a26ccbd8fdf4603fdb817aebd263
1.10.0-DEV.609

[oxinabox]

this will be closed by: oxinabox/Tricks.jl#37

[alhirzel]

Tested and working for me, fwiw. Thanks all!