Returns dramatically wrong results (from Formatting.jl)

[goerz]

It seems that this port is bringing over the following rather serious correctness bug from Formatting.jl:

JuliaIO/Formatting.jl#108

julia> using Format
julia> fspec = FormatSpec(".1e")
julia> printfmt(fspec, 0.0003)
2.10e-04

[jariji]

Any chance this could be addressed?

[ScottPJones]

Yes - I intend to do a full rewrite of some of the formatting code (which currently has 3 different formatting paths!), to be based on the code that is already based on Ryu, which will address this issue.