pass init to prod in normalization

[jishnub]

Normalization is defined as 1/prod(dimensions), which should be well-defined for an empty dimensions (in which case the prod evaluates to 1). This PR gets around the following error:

julia> AbstractFFTs.normalization(zeros(), 1:0)
ERROR: MethodError: reducing over an empty collection is not allowed; consider supplying `init` to the reducer

and the following runtime dispatch that arises from the error:

julia> using JET

julia> @report_opt AbstractFFTs.normalization(rand(4), 1:1)
═════ 1 possible error found ═════
┌ @ /home/jishnu/Dropbox/JuliaPackages/AbstractFFTs.jl/src/definitions.jl:275 AbstractFFTs.normalization(AbstractFFTs.real(AbstractFFTs.eltype(X)), AbstractFFTs.size(X), region)
│┌ @ /home/jishnu/Dropbox/JuliaPackages/AbstractFFTs.jl/src/definitions.jl:274 AbstractFFTs.prod(Base.Generator(#91, region))
││┌ @ reduce.jl:613 Base.:(var"#prod#270")(pairs(NamedTuple()), #self#, a)
│││┌ @ reduce.jl:613 mapreduce(identity, Base.mul_prod, a)
││││┌ @ reduce.jl:302 Base.:(var"#mapreduce#263")(pairs(NamedTuple()), #self#, f, op, itr)
│││││┌ @ reduce.jl:302 mapfoldl(f, op, itr)
││││││┌ @ reduce.jl:170 Base.:(var"#mapfoldl#259")(Base._InitialValue(), #self#, f, op, itr)
│││││││┌ @ reduce.jl:170 Base.mapfoldl_impl(f, op, init, itr)
││││││││┌ @ reduce.jl:44 Base.foldl_impl(op′, nt, itr′)
│││││││││┌ @ reduce.jl:49 Base.reduce_empty_iter(op, itr)
││││││││││┌ @ reduce.jl:378 Base.reduce_empty_iter(op, itr, Base.IteratorEltype(itr))
│││││││││││┌ @ reduce.jl:379 Base.reduce_empty(op, eltype(itr))
││││││││││││┌ @ reduce.jl:356 Base.mapreduce_empty(%1, [quote], T)
│││││││││││││ runtime dispatch detected: Base.mapreduce_empty(%1::AbstractFFTs.var"#91#92"{Tuple{Int64}}, [quote]::Base.BottomRF{typeof(Base.mul_prod)}, T::Type{Int64})::Union{}
││││││││││││└─────────────────

After this PR,

julia> AbstractFFTs.normalization(zeros(), 1:0)
1.0

julia> @report_opt AbstractFFTs.normalization(rand(4), 1:1)
No errors detected

[codecov]

Codecov Report

Attention: 3 lines in your changes are missing coverage. Please review.

Comparison is base (0918c3c) 84.13% compared to head (42d157b) 94.84%.
Report is 32 commits behind head on master.

Files	Patch %	Lines
src/definitions.jl	0.00%	3 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           master      #81       +/-   ##
===========================================
+ Coverage   84.13%   94.84%   +10.70%     
===========================================
  Files           2        5        +3     
  Lines         208      446      +238     
===========================================
+ Hits          175      423      +248     
+ Misses         33       23       -10     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[jishnub]

Gentle bump, is there anything else that needs to be done here?

[jishnub]

Gentle bump