Version 0.6 support: Problem with @simd macro

[getzdan]

When trying to run on 0.6, I get the following error:

julia> A = rand(4,4)
4×4 Array{Float64,2}:
 0.583756  0.913072  0.142562  0.409319 
...

julia> tensortrace(A,[1,1])
WARNING: Array{T}(::Type{T}, m::Int) is deprecated, use Array{T}(m) instead.
...
ERROR: MethodError: no method matching @simd(::LineNumberNode)
Closest candidates are:
  @simd(::LineNumberNode, ::ANY) at simdloop.jl:89
Stacktrace:
 [1] @generated body at /home/dan/.julia/v0.7/TensorOperations/src/implementation/recursive.jl:29 [inlined]
 [2] trace_rec!(::TensorOperations.One, ::TensorOperations.StridedData{1,Float64,:N}, ::TensorOperations.Zero, ::TensorOperations.StridedData{1,Float64,:N}, ::Tuple{Int64}, ::Int64, ::Int64, ::Tuple{Int64}) at /home/dan/.julia/v0.7/TensorOperations/src/implementation/recursive.jl:27
 [3] trace!(::Int64, ::Array{Float64,2}, ::Type{Val{:N}}, ::Int64, ::Array{Float64,0}, ::Array{Int64,1}, ::Array{Int64,1}, ::Array{Int64,1}) at /home/dan/.julia/v0.7/TensorOperations/src/implementation/stridedarray.jl:63
 [4] tensortrace(::Array{Float64,2}, ::Array{Int64,1}, ::Array{Int64,1}) at /home/dan/.julia/v0.7/TensorOperations/src/functions/simple.jl:54
 [5] tensortrace(::Array{Float64,2}, ::Array{Int64,1}) at /home/dan/.julia/v0.7/TensorOperations/src/functions/simple.jl:51

When the @simd code is taken out of _stridedloops in meta.jl, things work again. The culprit seems to be a LineNumber nodes which enter the expression.

[getzdan]

Oops... forgot to switch to master first. Closing.

[getzdan]

Still a problem on master.

[getzdan]

Further discussion in the pull request #31 (which solves this problem, at least temporarily for me).