import NewvarNode

[niklasschmitz]

This PR fixes a bug due to NewvarNode not being imported in recurse_fwd.jl

Currently this leads to an error:

using Diffractor
x = ones(3)
v = ones(3)
f(x) = sum(abs2, x)

julia> Diffractor.PrimeDerivativeFwd(ε -> f(x + ε*v))(0.0)
ERROR: LoadError: UndefVarError: NewvarNode not defined
Stacktrace:
  [1] (::Diffractor.var"#mapstmt!#215"{Core.SimpleVector, Int64, Diffractor.var"#emit!#214"{Vector{Any}, Vector{Any}}, Vector{Int64}})(stmt::GlobalRef)
    @ Diffractor ~/.julia/dev/Diffractor.jl/src/stage1/recurse_fwd.jl:58
  [2] (::Diffractor.var"#210#216"{Diffractor.var"#emit!#214"{Vector{Any}, Vector{Any}}})(stmt::GlobalRef)
    @ Diffractor ~/.julia/dev/Diffractor.jl/src/stage1/recurse_fwd.jl:39
  [3] iterate
    @ ./generator.jl:47 [inlined]
  [4] _collect(c::Vector{Any}, itr::Base.Generator{Vector{Any}, Diffractor.var"#210#216"{Diffractor.var"#emit!#214"{Vector{Any}, Vector{Any}}}}, #unused#::Base.EltypeUnknown, isz::Base.HasShape{1})
    @ Base ./array.jl:783
  [5] collect_similar
    @ ./array.jl:698 [inlined]
  [6] map
    @ ./abstractarray.jl:2853 [inlined]
  [7] (::Diffractor.var"#mapstmt!#215"{Core.SimpleVector, Int64, Diffractor.var"#emit!#214"{Vector{Any}, Vector{Any}}, Vector{Int64}})(stmt::Expr)
    @ Diffractor ~/.julia/dev/Diffractor.jl/src/stage1/recurse_fwd.jl:38
  [8] transform_fwd!(ci::Core.CodeInfo, meth::Method, nargs::Int64, sparams::Core.SimpleVector, N::Int64)
    @ Diffractor ~/.julia/dev/Diffractor.jl/src/stage1/recurse_fwd.jl:93
  [9] perform_fwd_transform(ff::Type{Diffractor.∂☆recurse{1}}, args::Any)
    @ Diffractor ~/.julia/dev/Diffractor.jl/src/stage1/recurse_fwd.jl:139
 [10] (::Core.GeneratedFunctionStub)(::Any, ::Vararg{Any})
    @ Core ./boot.jl:580
 [11] ∂☆internal
    @ ~/.julia/dev/Diffractor.jl/src/stage1/forward.jl:115 [inlined]
 [12] ∂☆
    @ ~/.julia/dev/Diffractor.jl/src/stage1/forward.jl:134 [inlined]
 [13] (::Diffractor.PrimeDerivativeFwd{1, var"#1#2"})(x::Float64)
    @ Diffractor ~/.julia/dev/Diffractor.jl/src/interface.jl:177
 [14] top-level scope

After import:

julia> Diffractor.PrimeDerivativeFwd(ε -> f(x + ε*v))(0.0)
6.0