Formatting.jl is broken on v1.0

[dmitry-v-vlasov]

    fmtrfunc = generate_formatter("%11.9e")
    buf = IOBuffer()
    for i = 1:N, j = 1:N
      print(buf, fmtrfunc(Mˡ[i, j]))
      if i <= N && j < N
        print(buf, "  ")
      elseif i <= N && j == N
        print(buf, "\n")
      end
    end

[ Info: Saving the transformation matrix to 'output1-transformation-matrix-general.dsv'
ERROR: LoadError: MethodError: no method matching (::getfield(Formatting, Symbol("##5#6")))(::Float64)
The applicable method may be too new: running in world age 25141, while current world is 25142.
Closest candidates are:
#5(::Any) at /home/dmitry/.julia/packages/Formatting/kSPYy/src/cformat.jl:18 (method too new to be called from this world context.)
Stacktrace:
[1] saveMatrixList(::Array{Float64,1}, ::Array{Array{Float64,2},1}, ::String) at /home/****/.julia/packages/Diabatizem/qEDnV/src/data/data.jl:107

[jmkuhn]

It's hard for me to see what is going on here without a complete example. If I define the missing variables as:

N = 2
Mˡ = [1.2 3.4; 5.6 7.8]

and run the rest of the code, it works for me.

[jmkuhn]

OK, I have been able to reproduce this by putting the code into a function.

[jmkuhn]

As a temporary workaround add

import Base: invokelatest

to your application then change your print line to:

print(buf, invokelatest(fmtrfunc, Mˡ[i, j]))

This should be fixable from within Formatting but my initial attempts to fix it result in a huge number of allocations and terrible runtime performance in the tests.

I will be out of the country without a laptop for 2 weeks so I won't have time to work on this until Jan 7.