I’m getting an odd type instability that I don’t understand the cause of,
as illustrated in this gist
<https://gist.github.com/tlycken/18db1ed677cd9074b8fe>. It seems that the
type instability comes from the coordlookup function, but I don't
understand why that wouldn't be type stable. Perhaps easiest illustrated
with a code sample:
julia> foo(r::FloatRange, x) = (r.divisor * x - r.start) / r.step +
one(eltype(r)) # not the actual name, but exactly the actual body
foo (generic function with 1 method)
julia> @code_warntype foo(2:2.1:10, 3.2)
Variables:
r::FloatRange{Float64}
x::Float64
Body:
begin # none, line 1:
return
(Base.box)(Base.Float64,(Base.add_float)((Base.box)(Base.Float64,(Base.div_float)((Base.box)(Base.Float64,(Base.sub_float)((Base.box)(Base.Float64,(Base.mul_float)((top(getfield))(r::FloatRange{Float64},:divisor)::Float64,x::Float64)::Any)::Float64,(top(getfield))(r::FloatRange{Float64},:start)::Float64)::Any)::Float64,(top(getfield))(r::FloatRange{Float64},:step)::Float64)::Any)::Float64,(Base.box)(Float64,(Base.sitofp)(Float64,1)::Any)::Float64)::Any)::Float64
end::Float64
julia> versioninfo()
Julia Version 0.4.0-rc1
Commit e5c6964 (2015-09-09 16:07 UTC)
Platform Info:
System: Windows (x86_64-w64-mingw32)
CPU: Intel(R) Core(TM) i7-4980HQ CPU @ 2.80GHz
WORD_SIZE: 64
BLAS: libopenblas (USE64BITINT DYNAMIC_ARCH NO_AFFINITY Haswell)
LAPACK: libopenblas
LIBM: libopenlibm
LLVM: libLLVM-3.3
