zero dims gives zero length for img to write

[buthanoid]

The following code fails:

writefits!("/tmp/toto.fits", FitsHeader(), fill(1e0))
ERROR: FitsError(302): "column number < 1 or > tfields"
Stacktrace:
  [1] check
    @ ~/Documents/EasyFITS.jl-antoine/src/utils.jl:34 [inlined]
  [2] #write#75
    @ ~/Documents/EasyFITS.jl-antoine/src/images.jl:444 [inlined]
  [3] write
    @ ~/Documents/EasyFITS.jl-antoine/src/images.jl:398 [inlined]
  [4] write(file::FitsFile, hdr::FitsHeader, arr::Array{Float64, 0})
    @ EasyFITS ~/Documents/EasyFITS.jl-antoine/src/images.jl:365
  [5] #41
    @ ~/Documents/EasyFITS.jl-antoine/src/files.jl:190 [inlined]
  [6] FitsFile(func::EasyFITS.var"#41#42"{Tuple{FitsHeader, Array{Float64, 0}}}, filename::String, mode::String; kwds::Base.Pairs{Symbol, Union{}, Tuple{}, NamedTuple{(), Tuple{}}})
    @ EasyFITS ~/Documents/EasyFITS.jl-antoine/src/files.jl:40
  [7] FitsFile(func::Function, filename::String, mode::String)
    @ EasyFITS ~/Documents/EasyFITS.jl-antoine/src/files.jl:36
  [8] #openfits#23
    @ ~/Documents/EasyFITS.jl-antoine/src/files.jl:57 [inlined]
  [9] openfits
    @ ~/Documents/EasyFITS.jl-antoine/src/files.jl:55 [inlined]
 [10] writefits(::String, ::FitsHeader, ::Vararg{Any}; overwrite::Bool, kwds::Base.Pairs{Symbol, Union{}, Tuple{}, NamedTuple{(), Tuple{}}})
    @ EasyFITS ~/Documents/EasyFITS.jl-antoine/src/files.jl:189
 [11] writefits!(::String, ::FitsHeader, ::Vararg{Any}; kwds::Base.Pairs{Symbol, Union{}, Tuple{}, NamedTuple{(), Tuple{}}})
    @ EasyFITS ~/Documents/EasyFITS.jl-antoine/src/files.jl:226
 [12] writefits!(::String, ::FitsHeader, ::Vararg{Any})
    @ EasyFITS ~/Documents/EasyFITS.jl-antoine/src/files.jl:225
 [13] top-level scope
    @ REPL[3]:1
 [14] top-level scope
    @ ~/julia-1.9.4/share/julia/stdlib/v1.9/REPL/src/REPL.jl:1416

The problem is that in Julia, the length of a zero dimensional array is one. In the write() function, we want len to be the number of scalar value to write, so it should be zero when the array is zero dimensional.

In Julia, a zero dimensional array contains a value, whereas in FITS standard it means no data at all.

I suggest to favorise FITS meaning, that is a zero dimensional array has no data, and to disregard the value contained in the Julia zero dimensional array. This suggestion relies on the fact that this is for now, the only way of writing a zero dimensional fits file in EasyFITS (apart from it being bugged)

[buthanoid]

Looking again at the code of write(), the code of this PR may not be the right way to resolve the problem, maybe a better way is to add a first if testing dimensionality and exiting early if it is zero