Broadcasting Imputed Values

Hello,
I’m new to Turing so this might just be a beginners error but when I try following code

using Turing

@model function gdemo(x, y)
    s² ~ InverseGamma(2, 3)
    m ~ Normal(0, sqrt(s²))
    x .~ Normal(m, sqrt(s²))
    y ~ Normal(m, sqrt(s²))
end

mdl = gdemo(missing,missing)

chn = sample(mdl,NUTS(0.65),1000)

I get the following error

ERROR: LoadError: MethodError: no method matching dot_assume(::Random._GLOBAL_RNG, ::DynamicPPL.SampleFromUniform, ::Normal{Float64}, ::AbstractPPL.VarName{:x, Tuple{}}, ::Missing, ::DynamicPPL.UntypedVarInfo{DynamicPPL.Metadata{Dict{AbstractPPL.VarName, Int64}, Vector{Distribution}, Vector{AbstractPPL.VarName}, Vector{Real}, Vector{Set{DynamicPPL.Selector}}}, Float64})
Closest candidates are:
  dot_assume(::Any, ::Union{DynamicPPL.SampleFromPrior, DynamicPPL.SampleFromUniform}, ::Union{AbstractArray{var"#s79", N} where {var"#s79"<:Distribution, N}, Distribution}, ::AbstractArray{var"#s78", N} where {var"#s78"<:AbstractPPL.VarName, N}, ::AbstractArray, ::Any) at /home/waweros/.julia/packages/DynamicPPL/RcfQU/src/context_implementations.jl:458
  dot_assume(::Any, ::DynamicPPL.Sampler{var"#s152"} where var"#s152"<:MH, ::Union{AbstractArray{var"#s151", N} where {var"#s151"<:Distribution, N}, Distribution}, ::AbstractPPL.VarName, ::AbstractArray, ::Any) at /home/waweros/.julia/packages/Turing/uMQmD/src/inference/mh.jl:472
  dot_assume(::Any, ::DynamicPPL.Sampler, ::Any, ::AbstractArray{var"#s79", N} where {var"#s79"<:AbstractPPL.VarName, N}, ::Any, ::Any) at /home/waweros/.julia/packages/DynamicPPL/RcfQU/src/context_implementations.jl:471

With a stack trace leading back to the sampler if I remove the .~ then everything’s fine. The issue is that with the code I originally found the behavior with it’s a lot harder to remove all the broadcasting when I want to sample the prior.
Does anyone have any idea what I should do?